Oceanography And Marine Biology

OCEANOGRAPHY AND MARINE BIOLOGY

AN ANNUAL REVIEW Volume 25

OCEANOGRAPHY AND MARINE BIOLOGY AN ANNUAL REVIEW Volume 25

HAROLD BARNES, Founder Editor MARGARET BARNES, Editor The Dunstaffnage Marine Research Laboratory Oban, Argyll, Scotland

ABERDEEN UNIVERSITY PRESS

FIRST PUBLISHED IN 1987 This edition published in the Taylor & Francis e-Library, 2005. “To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.” This book is copyright under the Berne Convention. All rights reserved. Apart from any fair dealing for the purpose of private study, research, criticism or review, as permitted under the Copyright Act, 1956, no part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, electrical, chemical, mechanical, optical, photocopying, recording or otherwise, without the prior permission of the copyright owner. Enquiries should be addressed to the Publishers. © Aberdeen University Press 1987 British Library Cataloguing in Publication Data Oceanography and marine biology: an annual review.—Vol. 25 1. Oceanography—Periodicals 2. Marine biology—Periodicals 551.46′005 GC1 ISBN 0-203-40068-2 Master e-book ISBN

ISBN 0-203-70892-X (Adobe eReader Format) ISBN 0-08-035066-8 (Print Edition) ISSN 0075-3218

PREFACE

1987 sees the publication of the 25th volume of Oceanography and Marine Biology: An Annual Review. The series was started in 1963 by Dr Harold Barnes because he thought the time was appropriate to draw together the work being done in all the marine sciences. The aim was stated in the preface to volume 1: “… to consider annually basic aspects of the field, returning to each at appropriate intervals, to deal with subjects of especial and topical importance, and to add new ones as they arise.” As far as possible this aim has been fulfilled. In the 25 years 254 articles have been published involving over 300 authors, some of whom have contributed more than once. At no time has there been any controversy between the editor and authors; the willingness with which they have acceded to editorial requests has always been appreciated and has made the editor’s task a pleasure rather than a tedious trial. The Annual Review has also been well served by its publishers who have attended to its production with meticulous care and so far have always managed to maintain the annual schedule. Manuscripts continue to be submitted to this series; many experts are still willing and even anxious to accept invitations to contribute to it. The desire to publish in it must reflect its importance and value to marine scientists in general. During the years since the death of Harold Barnes in 1978 I have been fortunate in having the advice of many friends and colleagues including, in particular, Drs A.D.Ansell, R.N.Gibson, and T.H.Pearson. Their help has been, and still is, greatly appreciated. It is hoped that this series of Annual Reviews will continue for many more years to fulfil the aims of its founder editor.

CONTENTS

PREFACE

iv

Phytoplankton Dynamics in Marginal Ice Zones WALKER O.SMITH JR

1

Sampling and the Description of Spatial Pattern in Marine Ecology N.L.ANDREWAND B.D.MAPSTONE

26

Flumes: Theoretical and Experimental Considerations for Simulation of Benthic Environments ARTHUR R.M.NOWELL AND PETER A.JUMARS

70

Larval Settlement of Soft-sediment Invertebrates: the Spatial Scales of Pattern explained by Active Habitat Selection and the emerging Rôle of Hydrodynamical Processes CHERYL ANN BUTMAN

89

Aplysia: its Biology and Ecology THOMAS H.CAREFOOT

139

A Review of the Comparative Anatomy of the Males in Cirripedes WALTRAUD KLEPAL

250

The Benguela Ecosystem. Part IV. The Major Fish and Invertebrate Resources R.J.M.CRAWFORD , L.V.SHANNON AND D.E.POLLOCK

305

The Association between Gobiid Fishes and Burrowing Alpheid Shrimps ILAN KARPLUS

458

The Ecological Impact of Salmonid Farming in Coastal Waters: A Review R.J.GOWEN AND N.B.BRADBURY

508

AUTHOR INDEX

520

SYSTEMATIC INDEX

547

SUBJECT INDEX

559

PHYTOPLANKTON DYNAMICS IN MARGINAL ICE ZONES WALKER O.SMITH, JR Botany Department and Graduate Program in Ecology, University of Tennessee, Knoxville, Tennessee 37996, U.S.A.

INTRODUCTION The marginal ice zone is an oceanographic front in which a transition from dense (those waters completely covered with ice) pack ice to one completely free of ice occurs (Fig. 1). The marginal ice zone is dynamic, responding rapidly to physical forcing; hence, the transition from 10/10 ice cover to open water can be abrupt or occur over hundreds of kilometres. The position of the ice edge can vary widely, with mesoscale variations occurring over the time scale of days and large-scale changes occurring seasonally (see Fig. 2). Furthermore, significant interannual variations occur (Niebauer, 1980; Zwally et al., 1983) which are related to global variations in air-sea interactions. An understanding of the processes which create changes in the ice-edge position will allow successful modelling of the marginal ice zone (both its physical and biological processes) as well as prediction of spatial and temporal variations in its position and gradients of ice concentration within it. This in turn will allow human activities in the region (offshore petroleum exploration, transportation, fisheries development) to increase. Polar regions are characterized by environmental extremes. An obvious major seasonal change occurs in the quantity of incident light, with long periods of darkness followed by a rapid change to continuous irradiance. As a result of the low amounts of incident radiation received annually, the local heat budget of polar regions is negative (i.e. a net flux of heat occurs to the atmosphere), which results in very low seawater temperatures, often near the freezing point. Both light, by virtue of its direct and indirect effects on phytoplankton growth, and temperature, via its overall control of microbial growth (Eppley, 1972), strongly influence autotrophic processes in polar regions. Despite the constraints that these factors (and potentially others) place on phytoplankton growth, they in and of themselves do not produce an environment conceptually different from temperate regions. Recent investigations into the physical and biological characteristics of marginal ice zones have, however, clearly shown that the marginal ice zone possesses attributes which are unique and which generate features which distinguish the ice-edge region from either polar or temperate areas. It has been noted for many years that the ice edge is a locus for activity of all trophic levels. For example, in the Antarctic, krill (Euphausia superba) seem to be associated with the marginal ice zone (Marr, 1962), as are many species of whales which feed upon krill (Mackintosh, 1970). Many species of birds are highly concentrated in the marginal ice zone, such as penguins (Spheniscidae) and snow petrels (Pagodroma nivea) (Ainley & Jacobs, 1981). In the Arctic polar bears frequent the ice edge, feeding on seals which surface in the leads during summer. The causes of the elevated higher trophic level biomass and activity in the marginal ice zone can be related to food abundance, using ice as a refugia from predation, and using ice as a nesting and breeding site (Ainley & Jacobs, 1981). Because the food web at the ice edge is ultimately

2

WALKER O.SMITH, JR

Fig. 1.—An aerial image of the marginal ice zone in the Fram Strait: the image was constructed by photography from repeated airplane overpasses; each ‘strip’ is approximately 10 km wide; note the irregular features along the ice edge, including a prominent eddy, ice bands, and meanders; the centre of the eddy is located at approximately 78°48′ N: 2°37′ W, and the mosaic is orientated in a north-south direction; composite provided by Dr R.Shuchman, Environmental Research Institute of Michigan, Ann Arbor, Michigan.

PHYTOPLANKTON IN MARGINAL ICE ZONES

3

dependent on the primary producers, such elevated standing stocks and activities of higher trophic levels must be supported by enhanced primary production within the marginal ice zone. Carbon fixation is conducted by algae attached to the ice and by phytoplankton. Ice algae contribute biogenic material early in the growing season and can reach large concentrations in a spatially confined zone, but the annual primary production associated with the ice when compared with that of the water column plankton is low (Whitaker, 1977; Garrison, Sullivan & Ackley, 1986). Ice-edge phytoplankton blooms (defined as standing stocks greater than those present in the absence of ice) have clearly been shown to be consistent features of marginal ice zones (Hart, 1934; Marshall, 1957; El-Sayed, 1971; McRoy & Goering, 1974; Alexander & Niebauer, 1981; El-Sayed & Taguchi, 1981; Schandelmeier & Alexander, 1981; Smith & Nelson, 1985a, b; Smith, Smith, Codispoti & Wilson, 1985). It is the purpose of this paper to discuss the features of marginal ice zones which make them unique among polar systems and how the oceanographic processes that occur within them influence phytoplankton growth and accumulation. PHYSICAL-BIOLOGICAL INTERACTIONS WITHIN THE MARGINAL ICE ZONE The major feature of the marginal ice zone is the physical presence of ice. The effects of ice are numerous. For example, during its seasonal advance and retreat, the freeze-thaw cycle changes the surface layer’s salinity and density, thereby changing the stability of the upper water column. Melting ice has been shown to create large density gradients in the Arctic (e.g. Marshall, 1957; McRoy & Goering, 1974; Alexander & Niebauer, 1981; Smith et al., 1985) and the Antarctic (e.g. Jacobs, Gordon & Ardai, 1979; Jacobs, Huppert, Holdsworth & Drewry, 1981; Smith & Nelson, 1985a); in contrast, freezing ice and subsequent brine rejection leads to instability and deep vertical mixing (Foster, 1968; Matthews, 1981). Changes in the stability of the upper layers of the ocean can have a rapid impact on plankton growth. The effect of stability on phytoplankton growth was first described by Riley (1942) and Sverdrup (1953), who mathematically derived the relationship between critical depth and the depth of vertical mixing to predict the onset of a temperate spring phytoplankton bloom. At a receding ice edge, low salinity (hence low density) melt-water is released, and the vertical stratification is greatly increased (Marshall, 1957; Alexander & Niebauer, 1981; Smith & Nelson, 1985a). This maintains the phytoplankton in the upper portion of the water column and provides sufficient light for growth to occur. In contrast, increased vertical mixing caused by brine rejection should decrease phytoplankton growth; such an effect has never been directly observed. The effect of melting ice on vertical stability appears to be the major factor in the initiation of a phytoplankton bloom within the marginal ice zone. Ice also modifies the quantity of surface irradiance which reaches the water column. Ice within the marginal ice zone can be either annual (produced within the past winter season) or multi-year ice (present for more than one winter). The optical characteristics of the two types are substantially different, but both greatly attenuate light. In general, pack ice (including any snow cover layer) will absorb from 80–99–95% of surface irradiance (Maykut & Grenfell, 1975; Sullivan, Palmisano & SooHoo, 1984). Any microbial community on or within the ice will further contribute to light attenuation. Despite the fact that the compensation light intensity of phytoplankton photosynthesis in polar waters is low (less than 1 µE·m−2·s−1; El-Sayed, Biggs & Holm-Hansen, 1983; Palmisano et al., 1985), phytoplankton photosynthesis in areas of heavy ice cover is severely restricted by the attenuation of light by sea ice per se. Ice can also influence the properties of the water column by modifying the air-sea interaction. For example, ice shields the water column from the wind’s energy, thereby reducing vertical mixing in areas with extensive ice cover. In areas with less ice cover, the ice moves more rapidly in response to wind than does the ocean’s upper layer, since the frictional drag on the rougher surface is greater than that on the

4

WALKER O.SMITH, JR

ocean. Another effect of ice is the generation of ice-edge upwelling. Upwelling has been observed within the marginal ice zone in the Bering Sea (Alexander & Niebauer, 1981) and in the East Greenland Sea (Buckley et al., 1979; Johannessen et al., 1983), although never in the Southern Ocean. The proximate cause of iceedge upwelling appears to be wind-driven off-ice Ekman transport (Johannessen et al., 1983; Niebauer & Alexander, 1985), and analytical (Gammelsrød, Mork & Røed, 1975) and numerical (Niebauer, 1982) models confirm this. Røed & O’Brien (1983), however, developed a model with a movable ice sheet which predicted surface convergence (rather than divergence) under the same conditions as the models of Gammelsrød et al. and Niebauer. This discrepancy may be the result of the sensitivity of all the models to various physical coefficients describing air-sea-ice interactions (Røed, 1983). Regardless of the conditions which give rise to upwelling, it is clear that such an event would have a significant impact on phytoplankton growth in the marginal ice zone of the Arctic by injecting nutrients into the euphotic zone, particularly during the summer months after nutrients had been depleted. Any occurrence of ice-edge upwelling in the marginal ice zone of the Southern Ocean, however, might be counter-productive, in that nutrients are rarely limiting, even in the densest blooms (Nelson & Smith, 1986). Upwelling in the Antarctic would reduce vertical stability caused by ice-melt, increase vertical mixing, and thereby result in a reduced light environment for phytoplankton. Therefore, if ice-edge upwelling were to occur in the Antarctic, a decrease in phytoplankton productivity and biomass might result. The ice edge also plays a rôle in the formation of eddies. They have been frequently observed in the marginal ice zone of the Fram Strait region (Johannessen et al., 1987; Shuchman et al., 1987; Manley et al., 1987), and there have been a number of mechanisms proposed to explain their genesis. For example, Hakkinen (1986) modelled the generation of eddies via a variable ice cover along an ice edge, where eddies are generated by a differential wind-induced circulation of water and ice. Topographical anomalies also have been shown to induce eddy formation (Smith, Morison, Johannessen & Untersteiner, 1984), and because the bottom topography in the vicinity of the ice edge of the Fram Strait is complex, eddies are frequently shed. Eddies are important biologically to the marginal ice zone of the Fram Strait for a number of reasons. First, they transfer heat into and out of the ice, increasing the rate of ice ablation and altering the stability of the water column. Secondly, they can move parcels of water under the ice (Manley et al., 1987), thereby reducing the amount of light available in the water column and light-dependent phytoplankton productivity. Finally, eddies induce vertical motion within their boundaries. The location within the eddy of the vertical motion is dependent on the direction of horizontal flow (cyclonic vs. anticyclonic) and whether the eddy is accelerating or decelerating, but in a manner similar to ice-edge upwelling, eddies will inject nutrients into the euphotic zone, and result in a significant stimulation of phytoplankton growth and accumulation during certain periods of the year. FACTORS INFLUENCING PHYTOPLANKTON GROWTH IN THE MARGINAL ICE ZONE Because marginal ice zones occur in different oceanic basins, each with unique physical, chemical, and biological features, it is difficult to generalize among the various ice-edge systems. Therefore, a comparative analysis of different marginal ice zones may be instructive. Four regions with well studied ice edges are included in this analysis: the Bering and Chuckchi Seas, the Fram Strait and the Barents Sea, the Weddell Sea, and the Ross Sea.

PHYTOPLANKTON IN MARGINAL ICE ZONES

5

MARGINAL ICE ZONES OF THE ARCTIC Bering and Chuckchi Seas The Bering Sea is a shallow continental shelf which is normally completely covered during the winter, but completely ice free by early June (Fig. 2). In some years the ice cover retreats over 1000 km (Niebauer, 1982), with the mean distance of melt-back along 170° W being 920 km (Konishi & Saito, 1974). The retreat occurs over a period of three months, so that the mean rate of ice retreat is 10 km per day. There are also variations in rate of retreat during the three months; mean distances of retreat are 135 km in March, 460 in April, and 325 in May (Konishi & Saito, 1974). It has been shown that a large phytoplankton bloom occurs within the Bering Sea marginal ice zone (Alexander & Niebauer, 1981; Schandelmeier & Alexander, 1981), and that this bloom is related to the creation of a vertically stable region in the vicinity of the receding ice. Extremely high concentrations of phytoplankton biomass occur in the bloom (chlorophyll a concentrations of more than 25 µg·l−1) and completely deplete the nitrate within the euphotic zone (Alexander & Niebauer, 1981). The bloom appears to be restricted to the zone of increased stability, but the distance of this stratified region is variable, ranging from 25 km (Alexander & Niebauer, 1981) to 100 km from the ice edge (Niebauer, Alexander & Cooney, 1981). The difference in horizontal extent is a function of the rate of meltwater input, the wind direction and speed, and the local current patterns. Given that there is nutrient depletion within the bloom, further new production (sensu Dugdale & Goering, 1967; Eppley & Peterson, 1979) requires introduction of nitrate into the surface waters. This apparently is accomplished by two mechanisms in the Bering Sea. The first is ice-edge upwelling (Alexander & Niebauer, 1981; Niebauer, 1983; Niebauer & Alexander, 1985), and the second is deep vertical mixing, often enhanced at the 50 m isobath by tidal forces (Niebauer & Alexander, 1985). Without the introduction of nitrate into the euphotic zone after the ice-edge bloom, significant primary production would not be expected to be sustained. That is, the ice-edge bloom would change the strength and timing of primary productivity, but would not increase the total annual carbon fixation. It is clear that mixing and upwelling do occur; their frequency is, however, unknown, and hence an estimate of their impact on phytoplankton blooms in the marginal ice zone is not yet possible. McRoy & Goering (1976) estimated the importance of primary production within the ice-edge system, and concluded that it accounts for approximately 40% of the annual production on the continental shelf. More recently, Alexander (unpubl.) has measured primary production within the marginal ice zone (Table I); her data show the intensity of the bloom during ice retreat and also the expected seasonal cycle in the open waters. When the ice-edge and open-water productivity is integrated through the entire year, approximately 50% of the total production is contributed by the ice-edge system (Table I). Recent estimates of primary productivity in the western portion of the Bering Sea by Sambrotto, Goering & McRoy (1984) indicated that production in that portion of the continental shelf had been significantly under-estimated because the stimulation of growth by nutrient input via cross-shelf flow had not been taken into account. Regardless of the absolute level of open-water production, it is clear that ice-edge phytoplankton blooms in the Bering Sea provide a highly significant contribution to the annual rates of carbon fixation in the region. It is also possible to estimate the impact of marginal ice zone-related blooms at one location by combining data on the extent of the bloom, the duration, the area covered, and the primary productivity within the bloom (Smith & Nelson, 1986). Because the magnitude of the production is variable through time (Table I), the localized impact is also variable. At a location on the outer shelf, primary production is limited to only the early melt-back period (March), and hence production is small (about 0·32 g C·m−2·day −1).

6

WALKER O.SMITH, JR

Fig. 2.—The average maximum and minimum extent of pack ice in the Northern and Southern Hemispheres: data for Southern Hemisphere ice limits from Zwally et al. (1983); data for Northern Hemisphere after Dietrich, Kalle, Krauss & Siedler (1980).

PHYTOPLANKTON IN MARGINAL ICE ZONES

7

TABLE I Primary productivity of open water locations and within the ice-edge phytoplankton bloom: values in brackets are linear extrapolations; data are means of 14C-productivity measurements taken on the continental shelf south of 60° N; data by courtesy of V.Alexander, University of Alaska Primary productivity Ice edge Daily C·m−2·day−1)

Month

(g

January February March April May June July August September October November December Total (g C·m−2·yr−1)

– – 0·32 3·16 6·60 – – – – – – –

Open ocean Monthly (g

C·m−2)

– – 4·03 94·80 204·60 – – – – – – – 309·32 (50·1%)

Daily (g

C·m−2·day−1)

[0·10] [0·11] 0·13 2·05 4·00 1·50 [1·10] 0·69 [0·48] [0·27] 0·06 [0·07]

Monthly (g C·m−2) [2·85] [3·15] 4·03 61·50 124·00 61·50 [33·95] 21·39 [14·40] [8·37] 1·80 [2·40] 308·44 (49·9%)

Because ice retreat occurs at a rate of 10 km·day−1 and the maximum extent of the bloom is 100 km (Niebauer et al., 1981), a bloom would be maintained at one location at most for ten days, and primary productivity contributed by the bloom would be 7·35 g C·m−2. This calculation assumes that the major factor in initiating and sustaining ice-edge blooms is the introduction of melt-water from the retreating ice edge (Alexander & Niebauer, 1981; Smith & Nelson, 1986). Similarly, for a mid-shelf location influenced by the ice-edge bloom in April, the primary productivity at one point would be 20·6 g C·m−2, and for a shallow inner-shelf area which supports a bloom in May, the primary productivity would equal 63·0 g C·m −2. Given that total production on the continental shelf is estimated to be at least 300 g C·m−2·yr−1 (Sambrotto et al., 1984), ice-related production is a modest source of organic matter on the outer shelf. Clearly, as the bloom intensifies it becomes more significant to the yearly production for a single location, and the major impact occurs in areas uncovered during late spring. Significant interannual variations in the extent and concentration of ice cover in the Bering Sea have been reported by Niebauer (1980). These are apparently related to basin-wide air-sea interaction anomalies, in that during years when significant ENSO (El Niño-Southern Oscillation) events occur (Barber & Smith, 1981), ice cover is much less extensive than in ‘normal’ years. During ‘warm’ years the Aleutian low pressure system shifts southward, bringing warmer air temperatures over the Bering Sea which to a large degree restrict the development of ice. A one- or two-year time lag occurs between an ENSO event and the Bering Sea temperature anomaly, indicating that such large scale interactions are of substantial magnitude (Niebauer, 1984). Because the ice-edge blooms in the Bering Sea are related to vertical stability induced by melt-water, it is clear that during warm years the effect of these blooms on the annual productivity cycle is lessened; the magnitude of the interannual productivity fluctuations is, however, unknown.

8

WALKER O.SMITH, JR

The ice-edge system of the Chuckchi Sea has been investigated to a lesser extent than that of the Bering Sea. Significant ice-melt does not begin until June, and large portions of the region remain ice-covered the entire year. Those areas that are uncovered during ice-melt exhibit strong stratification (Fleming & Heggarty, 1966; Hameedi, 1978), and nutrient levels are extremely low during the growing season. For example, nitrate was virtually undetectable within the euphotic zone during the summer at stations near the ice edge (Hameedi, 1978), and surface chlorophyll a levels were moderate (ranging from 0·3–0·8 µg·l−1). An extreme chlorophyll maximum was observed at nearly all stations in the marginal ice zone, being 1–2 orders of magnitude greater than surface levels. This maximum was associated with the base of the pycnocline; in view of the surface layer’s depletion of inorganic nitrogen, the maximum probably formed from surface biomass production and subsequent passive sinking of intact, metabolically active cells. The mean integrated euphotic zone chlorophyll a value was 147·1 mg·m−2, which represents nearly the entire water column in this shallow continental shelf region. Primary productivity ranged from 150 to 6857 mg C·m−2·day−1, with a mean of 1914 mg C·m−2·day−1 (all values reported in Hameedi, 1978, were doubled, since the original data were given on a half-day basis). These values are among the largest ever reported for a marginal ice zone and may be partially reflective of the elevated temperatures at which incubations were conducted. For example, one station had water column temperatures ranging from 0·9 to –1·6°C, but productivity incubations were conducted at 10·2°C. In view of the susceptibility of polar phytoplankton to significant influences of temperature (Neori & Holm-Hansen, 1982), the productivity data of Hameedi (1978) should be viewed with caution until substantiated. From Strait and Barents Sea Although geographically close, the marginal ice zones of the Fram Strait and the Barents Sea are in many ways dissimilar. For example, the Barents Sea is quite shallow (less than 100 m), and its southern portion is permanently ice-free due to the advection into the area of warm (greater than 3°C) North Atlantic water (Sverdrup, Johnson & Fleming, 1943). It also is influenced by freshwater input from rivers. The Fram Strait region, however, is isolated from continental influences and is at least 2000 m deep in most places. The bottom topography is complex and apparently generates a quasi-permanent eddy centred around the Malloy Deep (Wadhams & Squire, 1983; Smith et al., 1984). The area is influenced not only by the northerly flowing North Atlantic water but also by the southerly outflow from the Arctic Basin, the East Greenland Current (Paquette, Bourke, Newton & Perdue, 1985). The boundary between these waters, the Polar Front, is a region of rapidly changing temperature and also separates distinct biological communities (Smith et al., 1985). The marginal ice zone of each region is also different. The ice edge in the Barents Sea begins breakup in May and retreats approximately 475 kilometres by early August (Rey & Loeng, 1985); ice retreat is controlled by localized solar heating. The Fram Strait marginal ice zone is relatively invariant in space, being controlled by the position of the Polar Front (Paquette et al., 1985). When ice is blown or advected over the warmer North Atlantic water, melting occurs independent of the time of year; similarly, if the ice retreats so that it covers only waters originating from the East Greenland Current, little ablation will occur. The distance between the mean minimum and maximum extent of ice in the Fram Strait at 80° N is 120 km (Vinje, 1977). During September-October primary production within the marginal ice zone in the Barents Sea averaged 23·9 mg C·m−2·h−1 (equivalent to 287 mg C·m−2·day−1 if a 12-h day-length is assumed; Heimdal, 1983). During this study the surface nitrate values were generally low, ranging from 0·1–1·6 µM; phosphate and silicate were also low. All stations exhibited marked vertical stability, with the pycnocline generally located near 20 m, although there was some variation in the depth of the mixed layer (from about 12–25 m).

PHYTOPLANKTON IN MARGINAL ICE ZONES

9

Chlorophyll a concentrations were always less than 2·0 µg·l−1. Because pre-bloom concentrations of nitrate in the region are approximately 10 µM (Anderson & Dyrssen, 1981), it is obvious that substantial production and export had occurred prior to the study. No discernible increases in nutrient concentrations (via ice-edge upwelling), phytoplankton biomass or primary production were noted within the marginal ice zone, although the upward movement of isopycnals was observed in at least one section from which biological data were taken (Johannessen et al., 1983). Surface productivity and chlorophyll values were noted to range from 0·8–57·8 mg C·m−3·day−1 and 0.2–1.6 µg·l−1, respectively, in the western Barents Sea (Vedernikov & Solov’yeva, 1972); however, the stations occupied were very close to the coast and may have been influenced by continental processes. Rey & Loeng (1985) also studied the production within the marginal ice zone of the Barents Sea. They found that primary productivity was greatest during the spring (1465 mg C·m−2·day−1) but decreased to much lower rates in summer (239–495 mg C·m−2·day−1) as a result of the restriction of nutrient input due to strong vertical stability. No biological manifestations of upwelling were noted, probably because the sampling period encompassed only eight days in the period May through October. A phytoplankton bloom associated with the ice edge was observed and found to follow the retreat of the ice through the entire summer; the quantity and vertical distribution of phytoplankton biomass was, however, much different in the summer than in the spring, in that the summer distribution was characterized by marked subsurface maxima. This study was one of the few that was able to follow the temporal progression of a bloom associated with the marginal ice zone, and confirmed for this region the overriding importance of vertical stability not only to the timing of bloom initiation but on the limitation of primary production by restriction of nutrient influx. The Fram Strait marginal ice zone is, as stated, different in many respects from that of the Barents Sea. It is, however, similar in that initial nutrient concentrations prior to a spring bloom are alike (Anderson & Dyrssen, 1981), chlorophyll a concentrations within the bloom are somewhat similar (Smith et al., 1985; Smith, Baumann, Wilson & Aletsee, 1987), and the degree of vertical stability is similar in both. Smith et al. (1985) studied the distribution of nutrients and phytoplankton in relation to physical processes and found that nitrate concentrations were reduced in the euphotic zone to below 0·3 µM when adequate light was available for phytoplankton growth. Chlorophyll a concentrations reached a maximum of 11 µg·l−1. Based on the observed chlorophyll-nitrate relationship, a number of stations appeared anomalous (i.e. ‘excess’ chlorophyll a was present relative to the nitrate levels found). This suggested the introduction of nitrate into surface waters by some mechanism which subsequently stimulated phytoplankton growth and/or accumulation. In a study at the same location approximately one year later, Smith et al. (unpubl.) found slightly elevated levels of chlorophyll a in the same area, and that these were associated with lowered concentrations of nitrate (Fig. 3). That area is the site of a quasi-permanent eddy described by Wadhams & Squire (1983) centred in the vicinity of the Malloy Deep (a topographic depression). Because eddies can impart significant vertical motion within their structure, it was concluded that the increased chlorophyll was a result of growth enhanced by eddy-induced vertical flux of nutrients. Furthermore, eddies are a common feature of the marginal ice zone in the Fram Strait (Shuchman et al., 1987), apparently being generated not only via conservation of vorticity (i.e. induced by topographic anomalies; Smith et al., 1984) but by winds (Hakkinen, 1986), and variations in ice concentration (Hakkinen, 1986). The duration of the transient eddies in the vicinity of the ice edge appears to be at least 20 days (Shuchman et al., 1987), which is definitely long enough for an increase in phytoplankton growth and biomass to occur. The spatial extent of the eddies is variable. The Malloy Deep eddy appears to be approximately 100 km in diameter at the surface (although its shape is by no means symmetrical; Fig. 3), and its horizontal motion extends to the bottom (Johannessen et al., 1987). The transient eddies, those associated with the ice edge or formed by instabilities generated by

10

WALKER O.SMITH, JR

the Malloy Deep flow, are smaller (about 30 km in diameter) and may have their vertical and horizontal motions confined to the upper 200 m (Manley et al., 1987). Determinations of primary production in the Fram Strait region are scarce and the seasonal cycle of production is poorly known. Smith et al. (1987) found large variations in productivity across the ice edge, with the largest values (1718 mg C·m−2·day−1) occurring within 10 km of the ice edge. The mean productivity was 426 mg C·m−2·day−1, with minimum values of 29 mg C·m−2·day−1 at a station 60 km from the ice edge in waters totally covered with ice and 178 mg C·m−2·day−1 at a station 50 km from the ice edge in open water. Mesoscale variations in phytoplankton biomass were also noted (Fig. 4) and appeared to be in part related to the movement of the ice edge through time. The data do, however, clearly show that the ice edge consistently is a locus of phytoplankton biomass in the region. Interannual variations in phytoplankton productivity in the Fram Strait marginal ice zone (as well as all ice-edge systems) are basically unknown, despite the well-known yearly variations in ice extent and concentration (e.g. Zwally et al., 1983). Two studies completed in 1983 and 1984 (Smith et al., 1985; Smith et al., 1987), however, suggest the potential for significant differences between years. The 1983 study took place in July-August, whereas the 1984 study was conducted in June-July, so that some seasonal bias in the comparison of data is unavoidable. Chlorophyll a concentrations within the Malloy Deep eddy in 1983 were greater than 300 mg·m−2; in 1984 integrated chlorophyll a values never exceeded 100 mg·m−2 in the same region (both data sets integrated through 100 m). Nutrient values were slightly higher in 1984, but still were less than 1 µM in surface waters during both years. The major difference noted was a difference in temperature; temperatures in 1983 within the eddy were greater than 4°C, whereas in 1984 temperatures were less than 2°C. Vertical density gradients were large in both years. It is unclear whether this difference was related to differences between the two years or whether it was a manifestation of a seasonal effect. Fig. 3 cont. The Fram Strait region, like the Bering Sea, is interesting in that ice-melt results in active phytoplankton growth which ultimately becomes nutrient-limited, and that further new production is dependent on introduction of nitrate into the euphotic zone. Therefore, new productivity is dependent on physical processes such as upwelling or eddy formation, and the annual primary productivity budget, particularly during the summer (and presumably nutrient-limited) months, is dependent on the frequency of these physical events. Because the ice completely disappears by June in the Bering Sea, other processes not related to the presence of ice such as tidal mixing and storm events must replenish nutrients. In the Fram Strait ice-related processes, however, occur throughout the entire growing season. A complete understanding of the phytoplankton dynamics of the Fram Strait will require knowledge of the duration, extent, and magnitude of vertical fluxes generated by the important physical processes (eddies, ice-edge upwelling) found in the marginal ice zone of this region. MARGINAL ICE ZONES OF THE SOUTHERN OCEAN Weddell Sea The Weddell Sea is completely ice-covered for much of the year, with the pack ice extending to near the Antarctic Convergence. The circulation within the Weddell Sea is essentially a large cyclonic gyre, and significant mixing with other water masses (Drake Passage water, Scotia Sea water) occurs at the gyre’s northern extension. Ice retreat begins in October, with ice-melt and compaction occurring at the northern edge and in some years in the centre of the gyre (Zwally et al., 1983). At the minimum extent of the ice

PHYTOPLANKTON IN MARGINAL ICE ZONES

11

Fig. 3.—The distribution of temperature (a), σt (b), nitrate (c), and chlorophyll a (d) in a section taken across the Fram Strait: the influence of the eddy centred over the Malloy Deep can be seen in the density, nutrient and chlorophyll distributions (from Smith et al., unpubl.).

during February, ice remains along the coast of the Antarctic Peninsula (see Fig. 2), but the rest of the region is generally ice-free. The biology of the Weddell Sea marginal ice zone has been investigated since the early DISCOVERY expeditions (Hart, 1934, 1942). Hart (1942) separated various phytoplankton taxa into different groups, one of which included ice-edge forms. He also noted that this group included species whose relative contribution to the seasonal phytoplankton maximum was greatest. Thus, even early descriptive works indicated the biological importance of marginal ice zones. In 1968 an extensive bloom within the pack ice was observed by El-Sayed (1971). Chlorophyll a concentrations of 190 µg·l−1 were measured at the surface, and the bloom was overwhelmingly comprised of the diatom Thallasiosira tumida. The area covered by the

12

WALKER O.SMITH, JR

bloom was estimated to be 15500 km2. Unfortunately, only two stations were occupied which had chlorophyll a values at the surface greater than 2·0 µg·l−1, and water column characteristics were reported for only one of those. It is clear that the accumulation was not due to in situ growth, since nitrate concentrations even if quantitatively converted into phytoplankton biomass could not result in such large chlorophyll a levels. There must have been some concentration of cells from the ice and/or water column during the study. It is most probable that the massive accumulations of diatoms were the result of freezing ice which concentrated the algae from within the top 10 m of the water column by the mechanism proposed by Garrison, Ackley & Buck (1983). Even if ice crystals concentrated all the cells from the upper 10 m to the top few cm of the water column, the surface layer still must have had a chlorophyll a concentration of approximately 15–20 µg·l−1. It should also be noted that El-Sayed (1971) reported large concentrations of pancake ice in the area studied, indicating that water was actively freezing. Therefore the bloom was indeed unusual in its mode of formation, but its importance to seasonal production was probably small.

PHYTOPLANKTON IN MARGINAL ICE ZONES

13

Fig. 4.—The distribution of integrated (from 0–100 m) chlorophyll a in the marginal ice zone of the Fram Strait, June, 1984: the ice-edge position indicated by a dashed line; the marginal ice zone is the major site of growth and accumulation of phytoplankton biomass in this region.

The first detailed study of the marginal ice zone in the Southern Ocean was conducted in 1977 by ElSayed & Taguchi (1981). The cruise investigated the variations in phytoplankton biomass and productivity along the ice edge in relation to local physical and chemical factors. Large differences in autotrophic biomass and productivity were noted between the northern (north of 72° S) and southern sectors; specifically, mean integrated chlorophyll a concentrations were seven times greater in the southern sector (31·6 compared with 4·4 mg·m−2), and primary productivities were four times as great (0·41 compared with 0·10 mg C·m−2·day−1). No difference between the degree of ice cover between the two sectors was observed, and the vertical stability profiles showed no obvious systematic difference. The large-scale biological differences were attributed to water column stability, grazing, and proximity to land masses. During the same study, an analysis of the quantity, distribution, and taxonomic composition of ice-algae was also completed (Ackley, Buck & Taguchi, 1979). In general, low amounts of algal biomass within ice were found; maximum chlorophyll a concentrations were 4·3 µg·l−1 (that value represents the integrated pigment concentration for an entire ice core in which subsamples had been melted and quantified). The potential contribution of ice-algae to overall autotrophic biomass appeared to be larger in the northern portion of the Weddell Sea, not as a result of increased growth and production of ice-algae, but rather as a result of lower levels of phytoplankton biomass observed within the water column. El-Sayed & Taguchi (1981) did, however, describe the potential for biological interactions between the communities in the ice and those in the water column. The phytoplankton dynamics of the Weddell Sea marginal ice zone has recently been studied as part of AMBRIEZ (Antarctic Marine Ecosystem Research at the Ice-Edge Zone). The purpose of this project is to describe the seasonal patterns at various trophic levels within one marginal ice zone. In November– December, 1983, the biomass and productivity of phytoplankton was determined in transects normal to the ice edge to determine the level of activity at the onset of ice retreat. In March, 1986, similar transects were occupied to compare the standing stocks and growth of an ice edge which is advancing with one which had been retreating (Fig. 5). The first cruise started approximately six weeks prior to the maximum solar angle (most direct radiation), and the second approximately nine weeks after the equinox. Vertical density differences observed during the 1983 cruise were not as great as those in 1986 (Nelson et al., 1987; Muench

14

WALKER O.SMITH, JR

& Husby, unpubl.). The mean depth of the mixed layer in spring was 55 m; in autumn it averaged 35 m. In fact, the vertical stratification observed in the autumn was nearly constant and independent of ice cover (Fig. 5), whereas the mixed layer depths in spring were highly variable, ranging from 15 to 130 m, and spatially correlated with the retreating ice edge. These conditions persisted despite (or as a result of) the surface freezing and melting that occurred during the autumn and spring, respectively. Chlorophyll a concentrations were much greater in spring than in autumn, with mean euphotic zone values in the spring of 3·1 µg·l−1 (maximum value 12·9 µg·l−1), and in the autumn the euphotic zone average being 0·18 µg·l−1 (maximum observed value 1·80 µg·l−1). Maximum standing stocks in the spring occurred in the vicinity of the ice edge, but no clear spatial correlation between the vertical stability and chlorophyll distributions in autumn was found. Primary productivity values paralleled the biomass distribution (Fig. 5), with integrated (through the depth to which 0·1% light penetrated) means for spring and autumn being 571 and 200 mg C·m −2·day−1, respectively. Thus, the conceptual model that an ice-edge phytoplankton bloom is initiated by melt-water induced stratification and is dissipated by the breakdown of this stability appears only partially correct for the Weddell Sea ice edge, in that strong stability was still observed in autumn but elevated phytoplankton standing stocks were absent. The water column characteristics were qualitatively similar to those found in the Scotia Sea (Rönner, Sörensson & Holm-Hansen, 1983; Glibert, Biggs & McCarthy, 1982), where mixed layer depths of approximately 50 m occurred in the absence of ice. The low biomass and primary productivity encountered in autumn, 1986 may have resulted from low ambient light conditions, substantial grazing pressure or the dependence on ammonium as a nitrogen source for growth; regardless of the cause(s), it is clear that an ice-edge bloom in this region did not persist into the austral autumn. Ross Sea The Ross Sea is hydrographically less complex than the Weddell Sea in that it is somewhat more isolated from the large-scale circulation patterns of the Fig. 5 cont. Fig. 5 cont. entire Southern Ocean. Water movement within the Ross Sea can be described as a cyclonic gyre (Ainley & Jacobs, 1981), with water flowing along the coast of Victoria Land towards the equator and along the Ross Ice Shelf towards Victoria Land. The entire sea is ice-covered until late October, when open-water areas begin to appear along the edge of the Ross Ice Shelf at about 175° E (Zwally et al., 1983). The open areas continue to expand equatorward and towards Victoria Land until the entire region is ice-free east of 160° E. Ainley & Jacobs (1981) reported that in December areas with little or no ice exhibited the colouration and light attenuation coefficients characteristic of water with large accumulations of phytoplankton (no chlorophyll a measurements were made), and found that these bloom areas were confined to shelf waters which had been recently uncovered via ice-melt. Holm-Hansen, El-Sayed, Franceschini & Cuhel (1977) occupied a number of stations in the Ross Sea, but none of them can be considered to be within the marginal ice zone. Their data do provide, however, a good estimate of the background (i.e. areas not recently influenced by ice) biomass and productivity in the Antarctic region. Mean integrated (through 200 m) chlorophyll a and carbon assimilation values were found to be 10·4 mg·m−2 (range 0·73–30·8 mg·m−2; n=16) and 140 mg C·m−2·day−1 (range 40–294 mg C·m−2·day −1; n=23), respectively. El-Sayed, Biggs & Holm-Hansen (1983) investigated the nutrient and phytoplankton distributions in the vicinity of the Ross Ice Shelf during January but, again, these locations cannot be considered to be part of the marginal ice zone at the time of the study because any effects of the pack ice had been dissipated. The mean integrated (through the euphotic zone) chlorophyll a value was 19·1 mg·m−2,

PHYTOPLANKTON IN MARGINAL ICE ZONES

15

Fig. 5.—The distribution in 1983 (a, c, e) and 1986 (b, d, f) of density (as sigma-t), chlorophyll a, and primary productivity from sections normal to the ice edge in the Weddell Sea: the 1983 section was conducted in the austral spring (November) and was orientated along a north-south axis at about 61° S; the 1986 section was completed in autumn (March) and was orientated in an east-west direction at 64° S: data for a and c from Nelson et al. (1987); data in b courtesy of Drs R.Muench and D. Husby; data in d, e and f from W.Smith (unpubl.).

16

WALKER O.SMITH, JR

slightly higher than that reported by Holm-Hansen et al. (1977). El-Sayed & Turner (1977) reported a mean of 20·2 mg·m−2 for chlorophyll a concentrations within the euphotic zone in the Ross Sea.

PHYTOPLANKTON IN MARGINAL ICE ZONES

17

The distribution of phytoplankton biomass, productivity, and nutrient uptake at the ice edge off Victoria Land in January–February, 1983 was investigated by Smith & Nelson (1985a, b). Transects occupied

18

WALKER O.SMITH, JR

normal to the ice edge found that a large freshwater lens was present at the surface which created a large density gradient within the water column, reducing the mixed layer depth to less than 20 m. The spatial distribution of phytoplankton biomass (as evidenced by chlorophyll a, particulate carbon, particulate nitrogen, biogenic silica, and cell numbers) was highly correlated with the spatial extent of the zone of reduced salinity (and density). At locations removed from the ice-melt influence, vertical stability and phytoplankton biomass decreased markedly. The bloom was confined to an area about 250 km from the ice edge. Some variations in biomass along the ice edge were noted (Smith & Nelson, 1985b) and were attributed to fast-ice effects on the source water. The stratification induced by ice-melt was observed in all transects occupied. Because a number of estimates of biomass were measured concurrently in this study, it provided insights into what may be unusual adaptations of some species of Antarctic phytoplankton. Chlorophyll concentrations in the mixed layer of the bloom averaged 3·7 µg·l−1 (Smith & Nelson, 1986). Particulate carbon levels for the same samples averaged 39·5 µmol·l−1; therefore, the carbon/chlorophyll (w/w) ratio was 118. Such a value is extremely high when compared with temperate and tropical systems, and may reflect the low light conditions within this bloom (due to self-shading). Other studies, however, have also observed elevated C/Chl. ratios (Li, 1980; Sakshaug & Holm-Hansen, 1984), so it is possible that this is an adaptation to the low temperatures encountered in polar waters. The biogenic silica concentrations (a measure of the diatom, silicoflagellate, and radiolarian opaline material) were the highest ever measured in any ocean, averaging 24·4 µmol·l−1 within the bloom’s mixed layer (Smith & Nelson, 1986); in fact, the concentrations observed were higher by a factor of two than the previous maxima (Smith & Nelson, 1985a). The biogenic material was composed almost entirely of diatoms (Wilson, Smith & Nelson, 1986). The silica/carbon molar ratio was 0·60; normal oceanic diatoms have ratios of approximately 0·12 (Brzezinski, 1985). Therefore, the bloom contained only moderate levels of chlorophyll, very high levels of particulate carbon, and massive amounts of biogenic silica. If observations had consisted of only chlorophyll, an erroneous impression of the magnitude of the bloom would have been obtained. Microscopic examination of whole water samples indicated that extremely low levels of recognizable detrital material were present; nearly all of the observable particles were intact diatoms. Primary productivity within the bloom averaged 960 mg C·m−2·day−1 (Wilson et al., 1986), and outside of the high stability region it averaged 617 mg C·m−2·day−1. Integrated (through the euphotic zone) chlorophyll a concentrations for the bloom stations averaged 63·7 mg·m−2 (Smith & Nelson, 1985b). Species composition within and outside the bloom was similar (Smith & Nelson, 1985a), and growth rates within and outside the bloom were also similar. This was interpreted as evidence that at the edge of the bloom, the depth of vertical mixing increased, distributing the biomass through the entire water column, and hence over short time periods no difference in species composition would be noted. Furthermore, growth rates were similar because both areas were low-light environments; within the bloom the high standing stocks of phytoplankton caused self-shading, and outside the bloom deeper vertical mixing (the average mixed layer depth outside the bloom was 40 m) reduced the average light environment encountered by the phytoplankton. Few studies have investigated the form of inorganic nitrogen removed by phytoplankton within marginal ice zones of the Southern Ocean. Outside the marginal ice zone it has been found that nitrate provides anywhere from 5 to 50% of the total nitrogen required for growth (Rönner et al., 1983). It is this percentage of primary production that can be converted to higher trophic level biomass and is available for export (Dugdale & Goering, 1967; Eppley & Peterson, 1979). In view of the massive amounts of nitrate that are almost always present in surface waters in the Antarctic, such extensive use of recycled nitrogen (ammonium) is somewhat surprising, although it has been suggested on the basis of nutrient utilization

PHYTOPLANKTON IN MARGINAL ICE ZONES

19

patterns (Kamykowski & Zentara, 1985). Nelson & Smith (1986) measured ammonium and nitrate uptake within the marginal ice zone of the Ross Sea and found that on average 65% (range 32–95%) of the primary production was supported by nitrate. Such extensive amounts of new production (i.e. exportable to higher trophic levels or as particulate flux out of the euphotic zone) are among the highest measured values to date in the ocean, and may help explain why marginal ice zones are often the site of large accumulations of herbivores (krill, copepods, baleen whales, salps), birds (penguins, fulmars, petrels), fishes (myctophids), and marine mammals (seals, whales). IMPACT OF ICE-ALGAL COMMUNITY ON ICE-EDGE PHYTOPLANKTON BLOOMS Numerous species of algae grow on and within the pack ice of both polar regions (Horner, 1976; Horner & Schrader, 1982; Palmisano & Sullivan, 1983; Garrison, Sullivan & Ackley, 1986). It is now clearly established that the ice-algae begin active growth prior to any phytoplankton bloom by virtue of their relative position in the light field (Matheke & Horner, 1974), and upon the disintegration of the ice, the algae are released into the water column. It has been suggested that the ice-algae, by virtue of their extreme biomass (chlorophyll a concentrations within a small section of ice often are greater than 100 µg·1−1; Bunt & Lee, 1970; Whitaker, 1977; Palmisano & Sullivan, 1983), can give rise to bloom conditions within the marginal ice zone (Meguro, Ito & Fukushima, 1967). Recent observations have indicated that ice-algae sink very rapidly (Horner & Schrader, 1982) and that the residence time for most species within the water column is short. The fact that most ice-algae do not remain suspended for long time periods does not, however, preclude selected species from actively growing in the highly stratified upper layers of marginal ice zones. Wilson et al. (1986) found that Nitzschia curta, a pennate diatom usually found as a member of the ice-algal community, dominated the bloom in the western Ross Sea. N. curta also occurred in the ice, but it was not as common in the ice as it was in the water. Furthermore, micro-autoradiographic studies confirmed that it was actively photosynthesizing in the water column. Wilson et al. (1986) suggested that N. curta was released during ice-melt, selectively seeded the stratified surface waters, and accumulated to produce the large biomass observed in the study. Garrison & Buck (1985) used a statistical approach to determine the similarity in taxonomic composition of ice-algal communities and water column phytoplankton. They found that there was a significant overlap in taxa between ice samples and plankton taken from water directly below the ice, and suggested that the under-ice pelagic populations are derived from those in the ice. In a similar fashion, phytoplankton are concentrated from the water column during the formation of frazil ice (Garrison et al., 1983). Ice in the Weddell Sea consists of 50–70% frazil ice (Clarke & Ackley, 1984) which forms as the surface loses heat to the atmosphere i.e. where polynias and leads remain open due to wind action. Therefore, substantial amounts of frazil ice can form in energetic environments. As ice crystals form, they rise to the surface (their formation is limited to approximately the upper 10 m), scavenging and concentrating algae as they rise. In such conditions a strong species similarity would be expected. During conditions of ice retreat, a bloom forms however, in the wake of the ice; that is, there should be a spatial separation of the ice and the maximum concentration of bloom biomass. The magnitude of the spatial separation should depend on the rate of ice retreat and water-column seeding. In the Bering Sea, where rates of retreat are rapid, ice-algal communities and bloom communities are strikingly dissimilar (Schandelmeier & Alexander, 1981), whereas in the Ross Sea there seemed to be more overlap (Wilson et al., 1986). The degree to which icealgae serve as innocula for the surface waters undoubtedly depends not only on the input rate of algal cells, but on the. physical characteristics of the ice and the surrounding oceanographic environment.

20

WALKER O.SMITH, JR

PARTITIONING OF PRIMARY PRODUCTIVITY WITHIN MARGINAL ICE ZONES Ice-edge phytoplankton blooms are conspicuous mesoscale features of marginal ice zones, but an assessment of their importance to food-web dynamics and biogeochemical cycles has just begun. Just as numerous physical, chemical, and biological differences exist among the marginal ice zones studied to date, there is no reason to expect that the fate of ice-edge primary productivity will be the same among these various systems. The Bering Sea ice-edge bloom appears to contribute to a demersal food web. Little phytoplankton biomass appears to be utilized by herbivores within the water column during the life of the bloom (Alexander, 1980; Alexander & Niebauer, 1981), but rather sinks to the benthos where it is ingested, remineralized and/or resuspended during storm events. Walsh et al. (1985) suggest that a significant portion of primary production of the Bering Sea is advected and deposited in the sediments of the Chuckchi Sea, but the proportion of the ice-edge production being deposited is unknown. In the Fram Strait region, total annual marginal ice zone primary production appears to be less than in other systems, and zooplankton (large calanoids such as Calanus hyperboreus and C. finmarchicus) appear to remove 80% of the daily primary productivity despite their relatively low biomass (S.Smith, pers. comm.). The region also supports an active pelagic microbial food web (nanoplankton, bacterioplankton, and heterotrophic flagellates). Some biogenic material is lost from the euphotic zone via sinking of particles. Wefer & Honjo (1985) found that at 2100 m the maximum vertical flux occurred in September to October and was not temporally correlated with the maximum period of surface productivity. They also found that the flux of organic matter in autumn reached 7 mg·m−2·day−1, which is approximately 2·4% of the surface production. During the spring and summer seasons, a much smaller percentage of organic matter reaches the deep sea, indicating that surface production is being efficiently cycled at most times of the year. The Ross Sea marginal ice zone may be quite unusual among ice-edge systems in its manner of partitioning of biogenic material. There are large deposits of biogenic particulates in the region, and it has been assumed that these deposits reflect a large surface productivity (Noriki, Harada & Tsunogai, 1985). These deposits of siliceous material are large enough to make the area one of the largest locations for silica removal in the global silica budget (DeMaster, 1981). Carbon productivity for much of the growing season is low, with the exception of the ice-edge bloom which presumably occurs when- and wherever vertical stability results from melting ice. When the annual bloom-related carbon production is calculated from the data of Wilson, Smith & Nelson (1986) and compared with the total yearly production calculated by HolmHansen et al. (1977), the bloom supplies about 67% of the annual (bloom+non-bloom) productivity. When compared with carbon contents and deposition rates of the sediments below (Dunbar, Anderson & Domack, 1985; Ledford-Hoffman, DeMaster & Nittrouer, 1986), it is clear that most (about 90%) of the surface organic production is remineralized within the water column or at the surface of the sediments. When a similar calculation is made for biogenic silica, nearly all of the silica produced at the surface is being deposited in the benthos. This implies that opaline sediment accumulation in the Ross Sea occurs not because of extraordinarily large surface production but by a decoupling of organic and siliceous cycles either within the water column during particle flux or at the sediment surface, and that the marginal ice zone has an extremely important rôle in the global biogeochemical cycle of silica. Micro-paleontological data also indicate the importance of the ice edge to the benthos in the Ross Sea. Nitzschia curta, the overwhelmingly dominant species in the ice-edge bloom (Wilson et al., 1986), is also the dominant form found in the sedimentary record (Truesdale & Kellogg, 1979). Furthermore, it has been deposited consistently throughout the past 18000 years, indicating that the surface bloom is indeed a predictable feature. The contribution of N. curta increases in the biogenic sediments closer to the coast of Victoria Land (Truesdale & Kellogg, 1979), which is consistent with the idea that an ice-edge bloom is

PHYTOPLANKTON IN MARGINAL ICE ZONES

21

initially seeded by ice-algae but the taxonomic composition of the bloom will change in space and time as a result of selective growth of released species. Concentrations of herbivores, particularly euphausids, and other pelagic higher trophic levels may be somewhat less in the Ross Sea than in the Weddell Sea (Marr, 1962; Everson, 1977). It is possible that a large proportion of the biogenic material is utilized in the benthos rather than the water column. Sedimenttrap collections indicate that much of the material involved in vertical flux is whole phytoplankton cells rather than faecal material (Dunbar et al., 1985); therefore, the Ross Sea may function like the Bering Sea and support an active demersal food web. Little quantitative data exist (other than the geochemical evidence of siliceous deposition) to substantiate this hypothesis. The dependence of the Weddell Sea food web on the marginal ice zone production also has not been quantified. Much greater standing stocks of krill (Euphausia superba) have, however, been observed in the South Atlantic sector of the Southern Ocean than any other (Marr, 1962), including massive swarms in the vicinity of Elephant Island (Shulenberger, Wormuth & Loeb, 1984). The numbers of pelagic birds, penguins, and marine mammals also seems to be greater in the Weddell Sea than in other regions (Everson, 1977; Ainley, O’Connor & Boekelheide, 1984; Fraser & Ainley, 1986), but the degree of dependence on production associated with the marginal ice zone is unknown. Extremely high rates of biogenic silica accumulation have been measured in the South Atlantic sector near the northern limit of the Antarctic Convergence (DeMaster, 1981); the rate of organic matter deposition is unknown. Studies of the food web of the Weddell Sea ice-edge system are at present in progress, so that the uncertainties involved in the trophic relationships of this area should become resolved. CONCLUSIONS Marginal ice zones are unusual regions within polar seas; they not only have a distinctive physical setting, being covered by a moving ice pack during portions of the year, but have a characteristic biological system associated with them. The ice edge is a physically dynamic region, with eddies, fronts, jets, and other mesoscale features often present. They are also the sites of certain oceanographic processes such as ice melting or freezing, and upwelling, all of which have immediate impacts on biological communities. Communities in physically dynamic habitats are often coupled to the energy provided by the physical processes in the environment (Margalef, 1978); those in marginal ice zones, and phytoplankton assemblages in particular, are no exception. Recently developed techniques for the study of phytoplankton distributions need to be applied to marginal ice zones. For example, the Coastal Zone Color Scanner (CZCS) has been used in temperate and tropical regions to document the meso- and large-scale distributions of chlorophyll in the ocean. Although problems remain in its application within regions with a high solar angle, preliminary studies of ice-edge systems have been made (Maynard, 1986). Further use of airborne and satellite-based imaging techniques will greatly stimulate traditional biological oceanographic studies, as they have in other regions. Continuous monitoring of biological properties (e.g. optical transmission, fluorescence, bioluminescence) should also be utilized and will enable researchers to have a greater appreciation of the spatial and temporal variations of these properties in marginal ice zones, as well as the interactions between physical and biological processes in these systems. The phytoplankton assemblages of ice-edge regions are often much more productive than those removed from the ice edge or those within the heavy ice pack, and hence a potential exists for the marginal ice zone to be a locus of activity for the entire food web and for the biogeochemical cycles of the region. In many ice-edge systems this has been shown to be true, but the physical and biological processes differ among ice-

22

WALKER O.SMITH, JR

edge systems. Much remains to be learnt about the trophic structure within these ecologically important regions. For example, the annual progression of primary production in a pelagic environment has not been well documented, so that the quantitative impact of marginal ice zones on carbon cycles cannot be accurately estimated. The relationship of the water column plankton to the ice-algal community of the pack ice remains elusive, as does the quantitative significance of nanoplankton and microbial grazing. Finally, the relationship of ice cover (and the biology of the marginal ice zones) to large-scale variations in global circulation is unknown, despite evidence that such teleconnections exist. Given that so much about marginal ice zones remains to be learnt, future research efforts directed at understanding these regions will undoubtedly contribute much to our knowledge of the structure and function of polar ocean systems. ACKNOWLEDGEMENTS Financial support for research in marginal ice zones has been provided by the Office of Naval Research and the Division of Polar Programs, National Science Foundation. Many colleagues and students have collaborated on various aspects of this research and have contributed generously to the synthesis of concepts presented in this manuscript. I would especially like to acknowledge Drs D.Nelson, H.J.Niebauer, S.Smith, and L.Codispoti. REFERENCES Ackley, S.F., Buck, K.R. & Taguchi, S., 1979. Deep-Sea Res., 26, 269–282. Ainley, D.G. & Jacobs, S.S., 1981. Deep-Sea Res., 28, 1173–1185. Ainley, D.G., O’Connor, E.F. & Boekelheide, E.F., 1984. Ornithol. Monogr. 32, Amer. Ornithol. Union, Washington, D.C., 97 pp. Alexander, V., 1980. Cold Reg. Science Tech., 2, 157–178. Alexander, V. & Niebauer, H.J., 1981. Limnol. Oceanogr., 26, 1111–1125. Anderson, L. & Dyrssen, D., 1981. Oceanol. Acta, 4, 305–311. Barber, R.T. & Smith, R.L., 1981. In, Analysis of Marine Ecosystems, edited by A.R.Longhurst, Academic Press, New York, pp. 31–68. Brzezinski, M.A., 1985. J. Phycol., 21, 347–357. Buckley, J.R., Gammelsrød, R., Johannessen, J.A., Johannessen, O.M. & Røed, L.P., 1979. Science, 203, 165–167. Bunt, J.S. & Lee, C.C., 1970. J. mar. Res., 28, 304–320. Clarke, D.B. & Ackley, S.F., 1984. J. geophys. Res., 89, 2087–2095. DeMaster, D.J., 1981. Geochim. Cosmochim. Acta, 45, 1715–1732. Dietrich, G., Kalle, K., Krauss, W. & Siedler, G., 1980. General Oceanography. J. Wiley & Sons, New York, 626 pp. Dugdale, R.C. & Goering, J.J., 1967. Limnol. Oceanogr., 12, 196–206. Dunbar, R.B., Anderson, J.B. & Domack, E.W., 1985. In, Oceanology of the Antarctic Continental Shelf, edited by S.S.Jacobs, American Geophysical Union, Washington, D.C., pp. 291–312. El-Sayed, S.Z., 1971. In, Biology of the Antarctic Seas, edited by G.Llano & I. Wallen, American Geophysical Union, Washington, D.C., pp. 301–312. El-Sayed, S.Z., Biggs, D.C. & Holm-Hansen, O., 1983. Deep-Sea Res., 30, 871–886. El-Sayed, S.Z. & Taguchi, S., 1981. Deep-Sea Res., 28, 1017–1032. El-Sayed, S.Z. & Turner, J.T., 1977. In, Polar Oceans, edited by M.Dunbar, Arctic Institute of North America, Montreal, pp. 463–503. Eppley, R.W. 1972. Fish. Bull. NOAA, 70, 1063–1085. Eppley, R.W. & Peterson, B.J., 1979. Nature, Land., 282, 677–680. Everson, I., 1977. The Living Resources of the Southern Ocean. FAO, Rome, 156 pp.

PHYTOPLANKTON IN MARGINAL ICE ZONES

23

Fleming, R.H. & Heggarty, D., 1966. In, Environment of the Cape Thompson Region, Alaska, edited by N.J.Wilimovsky & J.N.Wolfe, U.S. Atomic Energy Commission, Oak Ridge, TN, pp. 697–754. Foster, T.D., 1968. J. geophys. Res., 73, 1933–1938. Fraser, W.R. & Ainley, D.G., 1986. BioScience, 36, 258–263. Gammelsrød, T., Mork, M. & Røed, L.P., 1975. Mar. Science Comm., 1, 115–145. Garrison, D.L., Ackley, S.F. & Buck, K.R., 1983. Nature, Land., 306, 363–365. Garrison, D.L. & Buck, K.R., 1985. In, Marine Biology of Polar Regions and Effects of Stress on Marine Organisms, Proc. 18th Europ. Mar. Biol. Symp., edited by J.S.Gray & M.E.Christiansen, J.Wiley & Sons, Chichester, pp. 103–122. Garrison, D.L., Sullivan, C.W. & Ackley, S.F., 1986. BioScience, 36, 243–250. Glibert, P.M., Biggs, D.C. & McCarthy, J.J., 1982. Deep-Sea Res., 29, 837–850. Hakkinen, S., 1986. J. geophys. Res., 91, 819–832. Hameedi, M.J., 1978, Mar. Biol., 48, 37–46. Hart, T.J., 1934. ‘Discovery’ Rep., 8, 1–268. Hart, T.J., 1942. ‘Discovery’ Rep., 21, 261–356. Heimdal, B.R., 1983. J. Plankton Res., 5, 901–918. Holm-Hansen, O., El-Sayed, S.Z., Franceschini, G.A. & Cuhel, R.L., 1977. In, Adaptations within Antarctic Ecosystems, edited by G.Llano, Gulf Publ. Co., Houston, pp. 11–50. Horner, R.A., 1976. Oceanogr. Mar. Biol. Ann. Rev., 14, 167–182. Horner, R.A. & Schrader, G.C., 1982. Arctic, 35, 485–502. Jacobs, S.S., Gordon, A.L. & Ardai, J.L., 1979. Science, 203, 439–444. Jacobs, S.S., Huppert, H.E., Holdsworth, G. & Drewry, D.J., 1981. J. geophys. Res., 86, 6547–6555. Johannessen, O.M., Johannessen, J.A., Morison, J., Farrelly, B.A. & Svendsen, E.A.S., 1983. J.geophys. Res., 88, 2755–2769. Johannessen, O.M., Johannessen, J.A., Svendsen, E.A.S., Shuchman, R.A., Campbell, W.J. & Josberger, E., 1987. Science, in press. Kamykowski, D. & Zentara, S.-J., 1985. Mar. Ecol. Prog. Ser., 26, 47–59. Konishi, R. & Saito, M., 1974. In, Oceanography of the Bering Sea with Emphasis on Renewable Resources, edited by D.Hood & E.J.Kelley, University of Alaska, Fairbanks, pp. 425–450. Ledford-Hoffman, P.A., DeMaster, D.J. & Nittrouer, C.A., 1986. Geochim. Cosmochim. Acta, 50, 2099–2110. Li, W.K.W., 1980. In, Primary Productivity in the Sea, edited by P.Falkowski, Plenum Press, New York, pp. 259–279. Mackintosh, N.A., 1970. In, Antarctic Ecology, edited by M.W.Holdgate, Academic Press, London, pp. 195–212. Manley, T.O., Villanueva, J.Z., Gascard, J.C., Jeannin, P.F., Hunkins, K.L. & van Leer, J., 1987. Science, in press. Margalef, R., 1978. Oceanol. Acta, 1, 493–509. Marr, J.W.S., 1962. ‘Discovery’ Rep., 32, 33–464. Marshall, P.T., 1957. J. Cons. perm. int. Explor. Mer, 23, 173–177. Matheke, G.E.M. & Horner, R., 1974. J.Fish. Res. Bd Can., 31, 1779–1786. Matthews, J.B., 1981. J. geophys. Res., 86, 6653–6660. Maykut, G.A. & Grenfell, T.C., 1975. Limnol. Oceanogr., 20, 554–563. Maynard, N.G., 1986. Mar. Tech. Soc. J., 20, 14–27. McRoy, C.P. & Goering, J.J., 1974. In, Oceanography of the Bering Sea with Emphasis on Renewable Resources, edited by D.Hood & E.J.Kelley, University of Alaska, Fairbanks, pp. 403–421. McRoy, C.P. & Goering, J.J., 1976. Mar. Science Comm., 2, 255–267. Meguro, H., Ito, K. & Fukushima, H., 1967. Arctic, 20, 114–133. Nelson, D.M. & Smith, Jr, W.O., 1986. Deep-Sea Res., 33, 1389–1412. Nelson, D.M., Smith, Jr, W.O., Gordon, L.I. & Huber, B., 1987. J. geophys. Res., in press. Neori, A. & Holm-Hansen, O., 1982. Polar Biol., 1, 33–38. Niebauer, H.J., 1980. J. geophys. Res., 85, 7507–7515. Niebauer, H.J., 1982. Cont. Shelf Res., 1, 49–98.

24

WALKER O.SMITH, JR

Niebauer, H.J., 1983. J. geophys. Res., 88, 2733–2742. Niebauer, H.J., 1984. Bull. Am. meteor. Soc., 65, 472–473. Niebauer, H.J. & Alexander, V., 1985. Cont. Shelf Res., 4, 367–388. Niebauer, H.J., Alexander, V. & Cooney, R.T., 1981. In, Eastern Bering Sea Shelf: Oceanography and Resources, Vol. II, edited by D.W.Hood & J.A.Calder, U.S. Dept Commerce, Washington, D.C., pp. 763–772. Noriki, S., Harada, K. & Tsunogai, S., 1985. In, Marine and Estuarine Geochemistry, edited by A.C.Sigleo & A.Hattori, Lewis Publ., Chelsea, MI, pp. 161–170. Palmisano, A.C., SooHoo, J.B., White, D.C., Smith, G.A., Stanton, G.R. & Burkle, L.H., 1985. J. Phycol., 21, 664–667. Palmisano, A.C. & Sullivan, C.W., 1983. Polar Biol., 2, 171–177. Paquette, R.G., Bourke, R.H., Newton, J.F. & Perdue, W.F., 1985. J. geophys. Res., 90, 4866–4882. Rey, F. & Loeng, H., 1985. In, Marine Biology of Polar Regions and Effects of Stress on Marine Organisms, Proc. 18th Europ. Mar. Biol. Symp., edited by J.S.Gray & M.E.Christiansen, J.Wiley & Sons, Chichester, pp. 49–64. Riley, G.A., 1942. J. mar. Res., 5, 67–87. Røed, L.P., 1983. J. geophys. Res., 88, 6039–6042. Røed, L.P. & O’Brien, J.J., 1983. J. geophys. Res., 88, 2863–2872. Rönner, U., Sörensson, F. & Holm-Hansen, O., 1983. Polar Biol., 2, 137–147. Sakshaug, E. & Holm-Hansen, O., 1984. In, Marine Phytoplankton and Productivity, edited by O.Holm-Hansen et al., Springer-Verlag, Berlin, pp. 1–18. Sambrotto, R.N., Goering, J.J. & McRoy, C.P., 1984. Science, 225, 1147–1150. Schandelmeier, L. & Alexander, V., 1981. Limnol. Oceanogr., 26, 935–943. Shuchman, R.A., Burns, B.A., Johannessen, O.M., Josberger, E.G., Campbell, W.J., Manley, T. & Lannelongue, N., 1987. Science, in press. Shulenberger, E., Wormuth, J.H. & Loeb, V.J., 1984. J. crust. Biol., 4, 75–95. Smith, D.C., Morison, J.H., Johannessen, J.A. & Untersteiner, N., 1984. J. geophys. Res., 89, 8205–8208. Smith, S.L., Smith, Jr, W.O., Codispoti, L.A. & Wilson, D.L., 1985. J. mar. Res., 43, 693–717. Smith, Jr, W.O., Baumann, M., Wilson, D.L. & Aletsee, L., 1987. J. geophys. Res., in press. Smith, Jr, W.O. & Nelson, D.M., 1985a. Science, 227, 163–166. Smith, Jr, W.O. & Nelson, D.M., 1985b. In, Antarctic Nutrient Cycles and Food Webs, edited by W.R.Siegfried et al., Springer-Verlag, Berlin, pp. 70–77. Smith, Jr, W.O. & Nelson, D.M., 1986. BioScience, 36, 251–257. Sullivan, C.W., Palmisano, A.C. & SooHoo, J.B., 1984. In, Ocean Optics VII, edited by M.A.Blizard, SPIE, Bellingham, WA, pp. 159–165. Sverdrup, H.U., 1953. J. Cons. perm. int. Explor. Mer, 18, 287–295. Sverdrup, H.U., Johnson, M.W. & Fleming, R.H., 1942. The Oceans. Prentice-Hall, Englewood Cliffs, New Jersey, 1087 pp. Truesdale, R.S. & Kellogg, T.B., 1979. Mar. Micropaleon., 4, 13–31. Vedernikov, V.I. & Solov’yeva, A.A., 1972. Oceanology, 12, 559–565. Vinje, T.E., 1977. Norsk Polarinst. Arbok 1975, 163–174. Wadhams, P. & Squire, V.A., 1983. J. geophys. Res., 88, 2770–2780. Walsh, J.J., Premuzic, E.T., Gaffney, J.J., Rowe, G.T., Harbottle, G., Stoenner, R., Balsam, W.L., Betzer, P.R. & Macko, S.A., 1985. Deep-Sea Res., 32, 853– 884. Wefer, G. & Honjo, S., 1985. EOS, 66, 1271 only. Whitaker, T.M., 1977. In, Adaptations within Antarctic Ecosystems, edited by G.A. Llano, Gulf Publ. Co., Houston, pp. 75–83. Wilson, D.L., Smith, Jr., W.O. & Nelson, D.M., 1986. Deep-Sea Res., 33, 1375– 1387.

PHYTOPLANKTON IN MARGINAL ICE ZONES

25

Zwally, H.J., Comiso, J.C., Parkinson, C.L., Campbell, W.J., Carsey, F.D. & Gloersen, P., 1983. Antarctic Sea Ice, 1973–1976: Satellite Passive-Microwave Observations, NASA Sp. Publ. 459, Washington, D.C., 206 pp.

Oceanogr. Mar. Biol. Ann. Rev., 1987, 25, 39–90 Margaret Barnes, Ed. Aberdeen University Press

SAMPLING AND THE DESCRIPTION OF SPATIAL PATTERN IN MARINE ECOLOGY N.L.ANDREW and B.D.MAPSTONE* Institute of Marine Ecology, Zoology Building, School of Biological Sciences, University of Sydney, Sydney, N.S.W. 2006, Australia

INTRODUCTION The description of pattern is of fundamental importance in ecology. Irrespective of the field of study, all marine ecologists are faced with the problem of establishing and quantifying patterns in nature. Observed patterns are the building blocks of the models from which we generate hypotheses, both about the patterns themselves and about processes that may govern them. Predictive hypotheses about processes suggested to explain observed patterns are tested by experiments and thus experiments are dependent on the patterns we perceive. More subtly, our perceptions of patterns often colour the sorts of questions we ask, thereby canalizing the design of experiments and the sorts of answers we get from them. Furthermore, observed patterns frequently provide the context within which the results of experiments are interpreted. Information on the distribution and abundance of organisms is often the sole basis for ecological and management decisions. The accurate and precise description of pattern is, therefore, essential to most aspects of ecology. Our reading of the literature suggests that the adequacy of sampling methods and designs has not often been demonstrated. Most studies apparently proceed more by custom and tradition than by careful consideration of potential biases and problems inherent in sampling different organisms. Certain procedures or methods are so popular in some fields that they have become “standards”; a good example is the use of the 0·25 m2 quadrat in intertidal studies. In other fields enormously divergent procedures have been used, apparently without justification. For example, in studies of tropical fish, a great variety of methods, ranging m transects, have been used to quantify from counts per unit time of observation to counts per abundances of fish. Many authors and/or editors of journals may consider demonstrations of the adequacy of sampling methods to be too preliminary or mundane to be incorporated into papers. Hence, this information may lie unpublished in filing cabinets or in inaccessible theses or reports. Alternatively, the appropriateness of the design may be determined completely informally, with intuition and experience guiding the design in an unstructured manner. Uncritical acceptance of standard techniques is, however, apparently widespread. Popular methods were often used despite overwhelming evidence of great mprecision and consequent low reliability and repeatability (see Downing, 1979; Resh, 1979). There are obviously many different methods of gathering data on the abundance of organisms. These can vary from the use of microscopy to count bacteria, to the use of cores to collect meiofauna, to aerial surveys of cetaceans. Each method has a special set of conditions and problems. Some methods are definitely more

*Order of authorship determined alphabetically.

SAMPLING AND SPATIAL PATTERN

27

limiting than others: it is much easier to estimate the abundance of algae on a shore than it is to count deepsea fish. Despite the many and varied methods used, the data so collected have two features in common: (1) all are subject to problems of inaccuracy and imprecision arising from the application of the sampling methods; and (2) all ecologists are constrained by time and funding, which restrict the placement and number of samples taken. Although many problems of sampling are specific to the particular methods used, there are considerations in design that have broad generality. Our aim in this review is to assess how marine ecologists have addressed the description of patterns in the arrangement and abundance of organisms and to discuss principles that have general applicability across all disciplines. We shall discuss two aspects of the description of pattern. The term “spatial pattern” will be used to refer to the arrangement of individuals within an area and “patterns in abundance” will refer to the abundance of organisms within and among areas. We have not used the more common term dispersion in our discussion of the arrangement of organisms in space to avoid confusion with the mathematical usage of the term. The term dispersion is reserved to describe the spread, or variability, of estimates about their mean (Sokal & Rohlf, 1981). We shall discuss the design of sampling programmes involving univariate data gathered to answer specific questions about spatial pattern or patterns in abundance of organisms. For example, how many bivalves are there in an estuary and are they non-randomly arranged? Are there more in one estuary than in others along the coast? We offer nothing more novel than the synthesis and interpretation of a large and scattered literature. We hope that this synthesis will promote the careful design of sampling programmes that are tailored to the demands of particular studies. While the review focuses on how marine ecologists have approached the description of pattern, relevant literature about terrestrial or freshwater organisms will be discussed where necessary; the principles are, after all, the same. For convenience, we shall concentrate most of our discussion on the estimation of abundance, but the discussion may be applied to most other variables, e.g. estimates of biomass or behavioural data. Two different aims may be identified when discussing the determination of abundance. One is to estimate the number of organisms in one area, and the second is to examine differences among areas. The design of sampling programmes to fulfil these different aims follows different guidelines. The procedures that best estimate the total number of organisms in an area may require uneven replication within sub-areas, adjusted to the exact requirements of the area (stratified sampling). On the other hand, when comparing estimates among areas, the requirements of the analysis are often best met by having equal numbers of replicates in each area (balanced multi-stage sampling). We discuss the best ways to analyse data collected for these different purposes in the relevant sections. Because we are primarily interested in univariate data sets gathered to answer the sorts of questions posed above, we have not considered multivariate techniques such as classification and ordination. Such techniques are more exploratory or hypothesis-generating in nature and are concerned with answering fundamentally different questions from those that will be discussed here (Green, 1979). Studies concerned with the description of whole communities or species diversity, and measures of association between species are also outside the scope of this review. We introduce sampling programmes designed to assess the abundance of organisms via a discussion of some general considerations such as the formulation of objectives and the interaction between statistical adequacy and biological realism. A discussion of the concepts of accuracy and precision leads on to more specific problems such as the determination of appropriate sizes of sampling units, replication, optimization procedures and the design of single-stage programmes to estimate the number of organisms in one area. This is then extended to more complex designs such as stratified random sampling and multi-stage

28

N.L.ANDREW AND B.D.MAPSTONE

sampling. Power analysis is reviewed as a tool for designing sampling programmes and for analysing the sensitivity of tests. In most marine studies, the description of spatial pattern entails the demonstration of non-random pattern. The various indices used to detect non-random pattern are reviewed along with their application to marine studies. The techniques used to describe non-random pattern in more complex situations, and in more detail, are then briefly outlined. In reviewing the literature we were primarily interested in those papers that concentrated on the description of spatial pattern or patterns of abundance, and especially those that were concerned with procedural aspects, i.e., studies that introduced new tests or justified the use of methods. Journals were searched from 1965 (or their inception) to June 1986; a list of the journals searched is provided in Table I. We have made selective forays into other journals and the earlier literature where necessary, usually as a result of citations from papers found in the above search. We have deliberately avoided citing unpublished papers and those in the “grey” literature, such as local or internal reports, because these are generally not available to most readers. We apologize to anyone who feels a paper of his/hers has been unjustly omitted. TERMS AND CONCEPTS In general we shall keep to a minimum the use of mathematics in this review. There are, however, a number of mathematical terms used that should be explicitly defined to avoid confusion. As far as possible we shall maintain consistency with terminology already used in the ecological literature. The term “population” will generally be used in a biological sense to refer to the local assemblage of organisms being studied (Caughley, 1977; Elliott, 1977). Where “population” is used in a statistical sense, clear qualifiers will be used to avoid ambiguity (e.g. population of values from which sample data are drawn). TABLE I List of journals searched American Naturalist Annual Review of Ecology and Systematics Canadian Journal of Fisheries and Aquatic Sciences (formerly Journal of the Fisheries Research Board of Canada) Coral Reefs Ecology Ecological Monographs Estuarine and Coastal Shelf Science (formerly Estuarine and Coastal Marine Science) Journal of Animal Ecology Journal of Applied Ecology Journal of Ecology Journal of Experimental Marine Biology and Ecology Journal of the Marine Biological Association of United Kingdom Journal of Wildlife Management Limnology and Oceanography Marine Biology

SAMPLING AND SPATIAL PATTERN

29

Marine Ecology Progress Series Oceanography and Marine Biology: an Annual Review Oecologia (Berlin) Oikos

It is unusual in biological, especially ecological, research to be able to count or measure the entire population being studied (Cochran & Cox, 1957; Cochran, 1963; Southwood, 1966; Elliott, 1977; Seber, 1982). Consequently, data are collected from a subset, or “sample”, of the population and inferences about the population are made based on these sample data. The size of the sample is usually very small relative to the size of the population (Caughley, 1977; Elliott, 1977; Green, 1979). In those cases where a relatively large proportion (>5–10%: Cochran, 1963; Elliott, 1977; Green, 1979; Seber, 1982) of the population is being measured, corrections for finite sampling may be necessary, details of which can be found in one of the many texts that treat sampling design (Cochran, 1963; Scheaffer, Mendenhall & Ott, 1979; Snedecor & Cochran, 1980). We use the term “sampling method” to describe the procedures used to obtain samples (e.g. coring, visual census, vertically hauled plankton nets). The actual device, implement or operation used to obtain mm core, m transect, 250each reading in a sample will be called the “sampling unit” (e.g. mm diameter plankton net). The way in which sampling units are allocated in space and/or time will determine the sampling design. To illustrate: a population of meiofauna might be sampled by taking cores (the sampling method) using a simple random sampling design, with ten replicate cores of 20 mm diameter mm depth (the sample unit) being collected. In this example, each core might be subsampled to economize counting time, in which case a similar statement could be made about the method, design, unit size, and replication of the sub-sampling procedure. “Variables” are the characteristics of the organisms or population being measured. Examples would include standard length, test diameter, density or TABLE II A list of symbols and their meaning as used throughout this review Symbol

Meaning

Description

n x µ

Sample size Variate Population mean Sample mean Population variance Sample variance Population standard deviation Sample standard deviation Standard error

Number of replicate units taken Single measure of variable, e.g. number of individuals in quadrat Mean of the population Mean of sample: estimator of µ Variance of the population Variance of sample: estimator of σ2 Standard deviation of population Standard deviation of sample: estimator of σ Standard deviation of

σ2 s2 σ s SE

distance from nearest neighbour. “Parameters” are mathematical or numerical values used to describe some characteristics of the entire population of the variable being measured (Winer, 1971; Scheaffer et al., 1979; Snedecor & Cochran, 1980). In most situations, the values of parameters are never known since they are the exact or “true” measures of an entire population of values. Normally we only have a sample of those values.

30

N.L.ANDREW AND B.D.MAPSTONE

For example, the true mean and variance of the weights of a population of snails are parameters of the population. Estimates of parameters are represented by the corresponding values obtained from a sample of the population; e.g. the arithmetic values of the mean and variance of the weights of a sample of snails taken from the above population. Winer (1971) and Elliott (1977) refer to estimates obtained from samples as “statistics”. We shall use “statistics” in two specific senses: sample or descriptive statistics being synonymous with the sample estimates, as in the above usage; and “test statistic” meaning the derived numeric value (e.g. F-value, t-value) used in a statistical test of a hypothesis about the data. Clearly, when an entire population is measured, the sample will be the population, and the parameters and their estimates will be identical. As noted above, this is unlikely in ecological contexts. The symbols and their meanings used throughout this review are given in Table II. Note that we have listed only the arithmetic mean as a measure of location and standard deviation, variance and standard error as the associated measures of dispersion. These are, by far, the most often used parameters or estimators in the ecological literature. ESTIMATING ABUNDANCE Virtually all ecologists are faced at some stage with the task of estimating abundances of organisms in the field. In recent years there have appeared a few publications specifically concerned with the design of sampling programmes and/or experimental studies (Holme & McIntyre, 1971; Caughley, 1977; Green, 1979; Underwood, 1981; Hurlbert, 1984; Underwood & Denley, 1984; Millard, Yearsley & Lettenmaier, 1985; Millard & Lettenmaier, 1986). These recent publications differ from excellent works such as Cochran & Cox (1957), Cochran (1963), Winer (1971), Scheaffer et al. (1979), Snedecor & Cochran (1980) and Seber (1982) in that the former were more concerned with ecological problems than the latter texts, which tended to concentrate on statistical procedures, agricultural experiments or sample survey design for human populations. Our purpose in this section is to outline the procedures that have been, or might be, followed by researchers intending to sample marine organisms. This will mostly involve reiterating the points made in the above publications. Green (1979), in particular, has thoroughly outlined the proper development of sampling programmes for environmental impact studies. Our discussion will largely parallel his “ten principles”. QUESTIONS, SCALE AND GENERAL CONSIDERATIONS As Green (1979) has pointed out, the interpretation of the results of a sampling exercise will be only as clear as the statement of the objectives of the study. It cannot be emphasized too strongly that before a sampling programme can be designed efficiently, all the questions to be addressed must be stated clearly and explicitly. Ambiguity in the intentions can only result in corresponding weaknesses in the design of the programme and the results that come from it. For example, if the objective of a study is to estimate the total number of holothurians on a particular coral reef, then a stratified sampling design should be used. If, however, the objective is to compare abundances among reefs, then a multi-stage sampling design should be used. It is most important to identify the sources of variation that might influence the data and, hence, the perception of pattern. Sources of variation in sampling programmes are often associated with phenomena operating at various scales, or can be estimated at several scales. For example, Caffey (1985) studied the recruitment of barnacles at several sites on each of several rocky shores at several locations along the New South Wales coast and thus accounted for variation at four spatial scales. The spatial scale at which a study

SAMPLING AND SPATIAL PATTERN

31

is done can have major consequences for both the results obtained and their subsequent interpretation (recent references include Harris, 1980; Denman & Gargett, 1983; Dayton & Tegner, 1984; Irish & Clarke, 1984; Millard et al., 1985; Wiens, 1986; Wiens, Addicott, Case & Diamond, 1986). Perceptions of “scale” are entrenched in our perception of problems, and subsequent generation of hypotheses. As Resh (1979) and Harris (1980) pointed out, however, we must deal not only with the convenient scales of our own perceptions and methods, but also with the functional scales at which the organisms are likely to respond to their environment. Clearly, it is logistically impossible to deal with all scales and/or sources of variation and careful choices have to be made. Results cannot be interpreted at scales, or for sources of variation, that the sampling programme did not address. Careful consideration has to be given also to the nature of the factors being investigated: are the questions being posited for only specific situations (e.g. on a particular coral reef) or is more generality required (the whole Great Barrier Reef)? Resolution of this question influences not only the formulation of testable hypotheses and the interpretation of results, but also the selection of study conditions (such as locations): studies of a single coral reef in one region cannot be used to infer patterns for the whole Great Barrier Reef. Resolution will come partially from a greater attention to the framing of questions and partly from a recognition that many ecological questions need answering at more than one spatial scale. An explicit statement of the questions to be addressed by a sampling programme will not always be straightforward. The questions must be expressed sufficiently clearly that specific null hypotheses can be stated. The hypotheses to be tested will incorporate the original objectives of the study and decisions about the scales of interest. They will strictly determine the subsequent design of the sampling programme. Once a hypothesis is stated, the researcher has implicitly defined the ‘right’ and ‘wrong’ scales at which sampling should be done, so statement must be clear and explicit. Decisions about the appropriate spatial scales and sources of variation have to be made from such intangibles as intuition and experience, in addition to existing natural history and distributional data. These decisions will be easier if the objectives are clearly stated. Incorrect decisions are likely to result in a design that is inappropriate to the objectives of the study and inaccurate perceptions of real patterns. At this stage of the design process, the researcher should have a clear concept of the general structure of the sampling programme. It is imperative to anticipate the analysis of the data and to check that the proposed programme conforms to the assumptions of those analyses. For example, if a programme is designed with the expectation that a three-way Analysis of Variance will be used to analyse the results, the form of the Analysis of Variance—the F ratios and their degrees of freedom—should be examined before the data are collected. If, for example, a fully factorial analysis is anticipated in which two of the factors are considered random factors, then there will be no valid F test for the main effect of the fixed factor, unless at least one of the first order interactions containing that factor is clearly non-significant (P>0·25, by convention, Winer, 1971) so that a pooled denominator mean square can be calculated (see Winer, 1971; Snedecor & Cochran, 1980; Underwood, 1981). Changing a priori the designation of one of the random factors to ‘fixed’ may correct the problem of analysis but will also fundamentally alter the original hypotheses, the way in which sampling is done and the interpretation of the results (Underwood, 1981). Problems of this kind must be sorted out a priori rather than a posteriori. As Green (1979) has commented, the time for statistical advice is when the programme is being designed, not after the data are in hand and one is wondering what can be done with them. The investigator must next give careful consideration to the way in which samples will be collected. In many situations, the identity and spacing of the basic unit of replication is relatively clear e.g. discrete units such as patch reefs, tide-pools or islands. This is less clear, however, in situations where the sample unit is arbitrary (e.g. a quadrat) and the area being studied cannot be divided, a priori, into natural units. For

32

N.L.ANDREW AND B.D.MAPSTONE

example, if a large and relatively uniform area has to be arbitrarily divided, then how far apart should the basic units be: 1, 10, 100 or 1000 metres? The determination of the spatial scales at which a study will be done involves an explicit judgement about the relative importance of variation at different scales. Variability among these scales, and its importance, will vary with the species and physical environment studied but can be estimated before the main study by doing multi-stage pilot studies. Replication at all levels to be considered is essential (Green, 1979; Hurlbert, 1984). Furthermore, Green (1979), Bernstein & Zalinski (1983), Hurlbert (1984), and Underwood & Denley (1984), among others, have all discussed several aspects of confounding in experimental or sampling designs. All have emphasized, that data must measure variability within as well as among the levels of a sampling programme and that the sources of variation about which hypotheses are posed must be unambiguously identifiable in the data. Hurlbert (1984) presented an extensive discussion of the ways in which designs have often been confounded. In general, random collection of sample units within the lowest level (smallest source of variation considered) of a design is recommended to ensure independence of the data and to avoid systematic error. True randomization is often difficult to implement but great care should be exercised in the substitution of ‘haphazard’ sampling for random sampling: ‘haphazard’ sampling may introduce subtle, unanticipated biases. Millard et al. (1985) have clearly demonstrated the effects of non-independence of sample data. It is also important to consider the proportion of an area or population that will be sampled by some chosen number of replicates, particularly where random samples are to be taken from an area or set of areas on a number of occasions. If the proportion of the local population or area sampled each time (the sampling fraction) is greater than 5–10%, then the estimate of abundance may need to be adjusted for finite sampling (Cochran, 1963; Elliott, 1977; Seber, 1982). Where the sampling fraction is larger, there may be a high probability of samples substantially overlapping and introducing unplanned confounding in the data. Replicates will not be independent and variability will be under-estimated. When comparisons of abundances are of interest, it is also desirable that the sampling programme remains balanced—that is, that replication at each level of the programme is consistent across all treatments. Comparative analyses of unbalanced data are possible for many situations, but in general the results of unbalanced analyses are less reliable than those of a corresponding balanced analysis (Bradley, 1968). When an initially balanced sampling programme becomes unbalanced it may be preferable to rebalance the data set by omitting some values prior to analysis. When the objective of a sampling programme is to estimate the total number of organisms in an area (e.g. the population of cockles in a bay), however, the best design may be an unbalanced stratified sampling design. Where sampling is designed to measure an effect of some perturbation or treatment, it is essential that control situations are also sampled so that the effects of treatments can be distinguished from stochastic processes or other, unspecified events that affect only some sampled areas (Green, 1979; Millard & Lettenmaier, 1986). Bernstein & Zalinski (1983) have elaborated on Green’s statements and have emphasized the need to consider various scales of change in both control and treatment conditions. Finally, after the sampling programme is completed, the data should be examined to ensure that they conform to the assumptions of the proposed statistical analyses and they should be transformed if necessary. This subject has been covered repeatedly, both in the statistical literature (Winer, 1971; Snedecor & Cochran, 1980; Sokal & Rohlf, 1981) and in the ecological literature (Downing, 1979, 1980, 1981; Green, 1979; L.R.Taylor, 1980; W. D.Taylor, 1980; Underwood, 1981; Morin, 1985) and will not be laboured here.

SAMPLING AND SPATIAL PATTERN

33

ACCURACY AND PRECISION In most standard English dictionaries, accuracy and precision are defined in terms of each other and are often noted as synonyms (Gove, 1969; Sykes, 1982). In scientific and engineering fields, however, the terms have come to represent quite different concepts (Cochran & Cox, 1957; Cochran, 1963; Lafedes, 1978; Sokal & Rohlf, 1981; Lincoln, Boxhall & Clark, 1982). The difference between common and technical usage has apparently led to inconsistent use of the two terms in much of the ecological literature. Clear distinction between the terms seems, however, to be both useful and desirable. Accuracy is defined as the closeness of a measurement or estimate to the true value of the variable being measured or parameter being estimated (Cochran & Cox, 1957: Cochran, 1963; Sokal & Rohlf, 1981; Lincoln et al., 1982). Thus, accuracy refers to the location of an estimate relative to the location of the true value. For example, if the true mean standard length (SL) of a population of fish is 43·7 mm, then a sample mean of 43·6 mm SL is a more accurate estimate of the population mean than a sample mean of 40·5 mm SL. A method that gives estimates that are repeatedly and predictably inaccurate is said to be biased. Bias, then, is the systematic deviation of an estimate from the true value and is caused by artefacts of the method used to obtain the estimate. Precision refers to the degree of concordance among a number of measurements or estimates for the same population (Cochran & Cox, 1957; Cochran, 1963; Sokal & Rohlf, 1981; Lincoln et al., 1982). Precision is reflected by the variability of an estimate. In the above example, if a number of samples of the population of fish returned a set of estimates of mean SL that ranged from 40·3 mm to 40·7 mm, then those estimates would have greater precision than a set with range 46·5 mm SL-52·0 mm SL. Note that accuracy cannot generally be inferred from precision. The preceding two sets of samples were both inaccurate (because the true mean was 43·7 mm) but the first was far more precise than the second. Clearly, then, estimates can be inaccurate but precise, both accurate and precise and so on. Precision and accuracy are truly synonymous only when no methodological biases exist (Cochran & Cox, 1957). The terms have been used in this sense (e.g. Winer, 1971; Underwood, 1981; Seber, 1982) but it is unlikely that bias will be absent in biological research and we will assume that some bias is always, at least potentially, present. We thus use the terms strictly and separately. Two other terms that have often been used with accuracy and precision are reliability and repeatability. These terms are apparently less clearly defined than accuracy and precision but have sometimes been used as synonyms for accuracy or precision, respectively (Cochran & Cox, 1957; Cochran, 1963). We shall use the terms reliability and repeatability to describe something of both accuracy and precision. Repeatability will be used to refer to the degree to which a sampling design or method can be expected to give consistently estimates with the same accuracy and precision. Reliability refers to the confidence that we can have in a given method or procedure: a reliable estimate would be both accurate and precise and an unreliable estimate would be neither. Accuracy and methods Inaccuracy in estimates can be attributed to two main sources: (1) inappropriate design of the sampling programme; and (2) biases inherent in the sampling methods. The first of these sources arises because the design of a sampling programme is inappropriate to the question(s) being investigated (see pp. 44–47). The second is a source of error that is systematically implicated in all sampling programmes. All estimators are dependent on the methods used to obtain the sample data and will thus incorporate the biases of those methods. Bias may arise from many sources, including errors made by observers, loss of organisms from sample units, over- or under-estimation of the areas sampled, disturbance caused by sampling, avoidance of

34

N.L.ANDREW AND B.D.MAPSTONE

the sampling device by organisms, and others. Our main interest in this section is not the cause of methodological biases, but the ways in which researchers have investigated the accuracy of their methods. Although many authors have apparently ignored questions of accuracy, several have emphasized the importance of methodological bias in determining the results of sampling programmes (Wiebe & Holland, 1968; McIntyre, 1971; Wiebe, 1971; Russell, Talbot, Anderson & Goldman, 1978; Resh, 1979; Southward & Barry, 1980; Sale, 1980; Sale & Douglas, 1981; Weinberg, 1981; Omori & Hamner, 1982; Fairweather & Underwood, 1983; Sale & Sharp, 1983; Kennelly & Underwood, 1984, 1985; Pot, Noakes, Ferguson & Coker, 1984). Practical tests of accuracy will be specific to each methodology, and will usually rely on attempts to assess absolute abundance by an independent method. This independent assessment must be accurate, otherwise one is only comparing two sampling methods, both with unknown biases (Kinzie & Snider, 1978). There have been a number of reports where accuracy has been thoroughly tested, usually by knowing a priori the absolute abundance of organisms in the sampled population or enumerating completely a population after sampling (Haury, 1973; Weinberg, 1981; Pihl & Rosenberg, 1982; Bell, Craik, Pollard & Russell, 1985; Andrew & Stocker, 1986). In many systems, such as plankton or bacterial communities, the accuracy of sampling and/or processing methods is extremely difficult, if not impossible, to assess because direct, independent enumeration of populations is impossible. In these, and many less problematic instances, the relative accuracy of different methods has been assessed by sampling the same populations by different methods and then comparing the values of the estimates obtained. Where estimates differ one method can only be said to be more accurate than another on the basis of intuition, auxiliary knowledge, and argument rather than on unequivocal evidence that it gave a truer account of the population. When methodologies have been compared, it has often been argued that the greatest estimate of abundance is the most accurate (Leatherwood, Gilbert & Chapman, 1978; Sale & Douglas, 1981; Sale & Sharp, 1983; Stretch, 1985; Gray & Bell, 1986). This argument is based on the assumption that you do not count what is not present, so over-estimation is extremely unlikely (Caughley, 1977). Exclusion or loss of organisms from samples is far more likely than inclusion. Some authors, however, have pointed out that boundary effects, where the inclusion of individuals in a sampling unit such as a quadrat is somewhat subjective, often lead to over-estimation (references in Downing & Anderson, 1985; Downing & Cyr, 1985). For sessile or slow-moving organisms, such boundary effects can be avoided by procedures such as including all ‘borderline’ individuals on two sides of a quadrat and excluding them along the other two sides. The problem is more difficult to circumvent when the organisms being counted move rapidly across sample unit boundaries and the observer has to make immediate decisions about whether an organism was inside or outside the unit at the instant when it was first sighted. Ideally, inclusions and exclusions of such marginal individuals should be equi-probable but inclusion seems more likely than exclusion (references in Downing & Anderson, 1985). For example, when visually censusing fish along belt transects, inclusion of individuals that are seen to swim into the transect is more likely than exclusion of fish that are observed leaving the transect. A number of recent studies have attempted either to quantify absolute accuracy or quantitatively to compare the accuracy of alternative methods. Jones (1974) showed that direct counts of bacteria using epifluorescence techniques varied dramatically when the bacteria were treated with different stains. Youngbluth (1982) came to a similar conclusion about the effect of the design of emergence traps on estimates of abundance of demersal zooplankton. Mackay, Cooling & Berrie (1984) found considerable disparities among five methods of estimating primary production in estuarine angiosperms. Gray & Bell (1986) reported that catches of vagile macrofauna from sea-grass meadows varied greatly with the method used (poison compared with beam trawl). There have been several comparisons of methods for assessing the

SAMPLING AND SPATIAL PATTERN

35

abundance of fish, either destructively or non-destructively, and all report major differences in the estimates returned by different methods (Russell et al., 1978; Brock, 1982; DeMartini & Roberts, 1982; Pot et al., 1984; Kimmell, 1985). By contrast, Bouchon (1981) found that quadrat and line transect methods gave similar estimates of coral cover and diversity. Nie & Vijverberg (1985) found no differences between catches of plankton from a Friedinger sampler and those from a Schindler sampler. Leatherwood et al. (1978) obtained similar, although very imprecise, estimates of numbers of dolphins using different methods of aerial survey. They noted that estimates varied with observers for all methods. Kennelly & Underwood (1984) found no difference between estimates of abundance of epilithic micro-organisms obtained in situ using an underwater microscope and from chips of substratum examined with superior microscopes in the laboratory. The underwater microscope, however, detected only about 25% of the micro-organisms that could be collected with a small suction sampler (Kennelly & Underwood, 1985). In a few instances, computer simulations have been used to model the sampling characteristics of different methods and consequently determine their biases and accuracy. Kinzie & Snider (1978) found that four visual survey techniques commonly used in field studies of corals, consistently gave inaccurate results when simulated with coral populations with varying characteristics. Wiebe & Holland (1968) and Wiebe (1971) found in simulations that estimates of the abundance of plankton were affected significantly by both the diameter of the plankton nets used and the distance over which the nets were towed. Furthermore, several of the more inaccurate estimates were more precise than most accurate estimates, illustrating that precision cannot be used to infer accuracy (e.g. Keast & Harker, 1977). Unless such simulations are constructed on a sound knowledge about the distribution and behaviour of the organisms in natural conditions and of the properties of the sampling device, extrapolations from computer to nature should be treated cautiously. Wiebe & Holland (1968) found good agreement between the precision of their simulated samples and that of several field studies and Wiebe (1972) tested the predictions of his earlier simulation study in the field. Despite the problems involved a number of researchers have attempted to assess absolute accuracy. Van Vleet & Williams (1980) tested the accuracy of a number of methods of sampling organic films on the sea surface by sampling, in the laboratory, surfaces that had been contaminated only with measured quantities of known compounds. Haury (1973) experimentally seeded a Longhurst-Hardy plankton recorder during operation with known quantities of real and model organisms in order to quantify biases arising from the variable residence time of plankton in the catch net leading the recorder. Weinberg (1981) tested the accuracy of seven methods of surveying coral communities by estimating abundances of corals by each method in an area that had previously been mapped in detail. The methods differed greatly in accuracy. An example of careful assessment of bias and accuracy of a sampling method has been presented by Pihl & Rosenberg (1982). By sampling known populations of vagile decapods and flatfish within fenced areas of a shallow sediment flat, Pihl & Rosenberg were able to compare the estimates of abundance obtained from the use of their drop trap with the known true values. They assumed that sampling fenced organisms was equivalent to sampling unfenced organisms. They attempted to address this problem by observing the responses of many individual organisms to the movements of the trap carriers and the operation of the trap. In most cases, the true assessment of accuracy will be very problematic and almost always extremely expensive in time and effort. In many studies, it may be unwarranted because knowledge of absolute abundance is not of primary interest (Caughley, 1977). Where comparisons are the aim of a study, estimates of relative abundance are often sufficient, although such comparisons assume equal sampling bias for all sites or conditions being compared. Unknown variations in the bias of a method with its use in different circumstances (e.g. habitats) is particularly problematic and difficult to assess. For example, artefacts of

36

N.L.ANDREW AND B.D.MAPSTONE

visual censuses of fish may change with changes in the complexity of the habitat surveyed. Similarly, Kennelly & Underwood (1985) found their suction sampler to be reliable when sampling clean, hard substrata of low complexity but to be of little use on more complex substrata such as silted surfaces or articulated coralline algae. Although estimates that are known to be inaccurate may be useful when accurate ones are not obtainable, some knowledge of the accuracy of the sampling methods used is always desirable. PRECISION The usefulness of an estimate is dependent on its precision as well as its accuracy (Cochran, 1963; Caughley, 1977; Green, 1979). Unlike accuracy, precision can be assessed relatively easily from characteristics of the sample data. Consequently, the well-established statistical tools for assessing precision are not specific to methodologies and can be considered as general principles. Precision is a function of the variance of the sample estimate: precision increases as the variance of the estimate decreases (Cochran & Cox, 1957; Cochran, 1963; Elliott, 1977; Eberhardt, 1978a, b). Precision is related to the confidence we have in the estimate, such as mean abundance. An important distinction must be drawn here between variation among sample data and the variation of the sample estimate, by which precision is measured. Variation in the sample data is an estimate of the variation in the arrangement of organisms in the real world. The spatial arrangement of individuals is inherently variable in space and time. In most populations, organisms are not regularly or randomly distributed but aggregated (Southwood, 1966; Pielou, 1969; Elliott, 1977; Downing, 1979; Resh, 1979). As we will discuss later (see p. 70), the variation among sample data depends on the size of the sampling unit relative to the scale of aggregation. Variations in the numbers of organisms in sampling units represent real phenomena and, if the methods used are reasonably reliable, sample variance should be relatively invariate with changes in sample size (=number of replicates). Variation in the sample estimate is, by contrast, implicitly dependent on the size of the sample. Consider the earlier example of estimating the mean size of fish. If a large number of fish (say 200) were measured for each repeated sample, then the sample means are more likely to be similar than if each of the samples included only 20 fish. Precision is, therefore, a characteristic of the sampling procedure rather than a reflection of some characteristic of the population being sampled. The Standard Error (SE) of a sample mean is the estimated standard deviation of a population of sample means from samples of that size. It is estimated from the standard deviation (s) of the sample data by:

where n is the sample size. Thus, for any s, which will be determined by the spatial arrangement of organisms and the size of units used to sample them, SE will decrease as replication increases. It follows that precision is a function of the SE of a sample. Precision has, however, been measured in terms of several sample statistics. The variance/mean ratio has sometimes been used in this context in addition to its use as a measure of the spatial arrangement of organisms (see p. 71). The ‘coefficient of of the data is also an expression often used in discussions of precision (Wiebe & variation’ Holland, 1968; Wiebe, 1971; Elliott, 1977; Resh, 1979; Pringle, 1984). These measures are not indicators of precision but of the relative variability of the sample data, standardized for the magnitude of the sample mean (Snedecor & Cochran, 1980). In other studies, precision is appropriately measured by comparing the 95% confidence intervals of the sample mean to that mean (Resh, 1979) or, more commonly, by the ratio (Southwood, 1966; Downing, 1979; Pihl & Rosenberg, 1982; Pringle, 1984; Downing & Anderson,

SAMPLING AND SPATIAL PATTERN

37

1985; Downing & Cyr, 1985; Morin, 1985). Unfortunately, the ratio has also been referred to as the ‘coefficient of variation’ (Irish & Clarke, 1983; Pringle, 1984) and in several papers it was not clear or was used to calculate precision. whether Precision is inversely related to the values of the ratios used to measure it. Precision is great when the SE is small relative to the mean and precision decreases as the ratio increases. The precision required for an estimate should be set a priori by the researcher and will be largely determined by the questions being to be 0·5 whilst another may demand greater precision, for asked. One researcher may require The important point to note here, is that by setting desired precision, the researcher example can then design a sampling programme to achieve that precision as economically as possible. OPTIMIZATION AND PILOT STUDIES It is usually desirable to describe patterns with the maximal precision and resolution possible with the available resources. Almost all studies are constrained by logistic and economic considerations (Cochran, 1963; Holme & McIntyre, 1971; Lewis, 1976; Saila, Pikanowski & Vaughan, 1976; Green, 1979; Resh, 1979; Morin, 1985; Millard & Lettenmaier, 1986). Optimization of the design of sampling programmes is achieved by determining the most efficient allocation of resources—i.e., minimizing decreases in precision and/or resolution imposed by cost or by logistical constraints (Cochran, 1963; Saila et al., 1976; Scheaffer, Mendenhall & Ott, 1979; Irish & Clarke, 1984; Downing & Anderson, 1985; Downing & Cyr, 1985; Morin, 1985). The notion of optimizing the design of sampling methods is certainly not new (see references in Cochran, 1963; Cochran & Cox, 1957; Snedecor & Cochran, 1980). Statistical methods for choosing appropriate replication and cost-efficient allocation of resources have been available for several decades. In this section we shall review the use of optimization procedures for designing sampling programmes in marine studies. Apart from considerations, such as ensuring that sampling programmes are appropriate to the question(s) being asked, there are two main procedural questions that need to be addressed empirically if a sampling programme is to be efficiently designed: (1) how big should each sample unit be, and (2) how many replicates are needed? For all sampling designs, optimization procedures require estimates of variances and/ or means before the main programme has commenced. These estimates can be obtained from three sources: (1) pilot studies; (2) previous studies; and (3) published data. If pilot studies are not possible, or if the same organisms have been studied in the same system then one of the other two methods may be used to obtain estimates of mean abundance, spatial pattern and variance. Clearly, recent prior studies done in the same system are far more likely to give useful estimates than studies done in a different system. Numerous authors have emphasized that pilot studies are preferable to the other two sources of estimates (Cochran, 1963; Gray, 1971; Holme & McIntyre, 1971, Hulings & Gray, 1971; Downing, 1979; Green, 1979; Resh, 1979; Underwood, 1981; Downing & Anderson, 1985; Downing & Cyr, 1985). Green (1979, p. 31) made the point emphatically: “Those who skip this step because they do not have enough time, usually end up losing time.” We shall discuss the implementation of pilot studies and draw attention to some of the problems likely to be encountered. In the first instance (see below) we shall assume that sample units are placed randomly over an area to obtain a single estimate for a population—Simple Random Sampling. We shall then discuss the more complex Stratified and Multi-stage designs.

38

N.L.ANDREW AND B.D.MAPSTONE

The size and number of sampling units The two aspects of sampling that most affect the precision of sample estimates are the size of the sampling units and the number of replicates collected (Cochran, 1963; Greig-Smith, 1964; Southwood, 1966; Snedecor & Cochran, 1980; Elliott, 1977; Green, 1979). The size and shape of the sampling units used have been repeatedly shown, both theoretically and empirically, to have a great impact on the precision of an estimate (Wiebe & Holland, 1968; Stoddart, 1969; Gray, 1971; McIntyre, 1971; Wiebe, 1971; Elliott, 1977; Resh, 1979; Sale & Douglas, 1981; Sale & Sharp, 1983; Pringle, 1984; Downing & Anderson, 1985 and references therein; Downing & Cyr, 1985; Morin, 1985). The choice of the size of the sample unit is fundamentally related to characteristics of the population being sampled such as spatial arrangement. This is particularly true when organisms are aggregated. Units that are smaller than or equal to the scale of aggregation will often give more variable estimates of density than those that are large with respect to the scale of aggregation of the organisms (Smith, 1938; Pechanec & Stewart, 1940; Wiebe & Holland, 1968; Wiebe, 1971; Elliott, 1977; Watling, Kinner & Maurer, 1978; Green, 1979; Helshe & Ritchey, 1984). Estimates of abundance will be most variable when the size of the sample unit is approximately equal to the average distance between aggregations (Southwood, 1966; Elliott, 1977) (see p. 71). This occurs because these and smaller units are likely either to miss aggregations and so contain few or no organisms or to include an aggregation, and contain many organisms. Larger units will be likely to include part of at least one aggregation and so very small or zero counts are unlikely to occur. Thus, estimates of average abundance obtained from large sampling units will be less affected by the patterns in the spatial arrangement of the organisms. Consequently, for a given sample size, the precision of a sample estimate is likely to increase with increasing size of sampling units (Elliott, 1977; Downing, 1979; Resh, 1979). The rate of increase in precision with increasing size of sample unit will initially be great but will rapidly decline as the size of the unit exceeds the average distance between aggregations in the population. Resh (1979), however, cautioned that this trend is dependent on sample units being located wholly within habitats or other physical clines. If sampling units correspond to the scale of such factors as natural micro-habitat units, great variability may result from factors not simply related to the spatial pattern of the organisms. The shapes of sampling units may also affect the precision of estimates of abundance. Where boundary effects are important, the amount of boundary relative to the area or volume of the sample unit should be minimized (references in Downing & Anderson, 1985). Boundary effects are a function of the shape as well as the size of the sampling unit and may be particularly important where the shape and size of sampling units correspond to those of aggregations of organisms or topographic features to which organisms respond (Resh, 1979). Furthermore, the minimum linear dimension of a unit may be more critical than the other linear dimensions, area or volume. For example, Caughley (1977) and Sale & Sharp (1983) have demonstrated the importance of the widths of transects when using transect survey techniques for visual censuses. Hargrave & Burns (1979 and references therein) found that both diameter and aspect (diameter, depth of sampler) of sediment traps significantly affected their performance and that the effects varied with the turbulence of the water in which they were used. Four characteristics of the size and shape of sample units can be considered: length, area, volume, and time. The relative importance of each of these measurements to the precision and accuracy of the units will depend on the particular application. In general it would seem expedient to compare sampling units of a number of sizes at the beginning of a study (Southwood, 1966). Where there is some prior knowledge of the spatial arrangement of individuals in the population (see p. 70), the smallest unit considered should, if feasible, be larger than the average spacing among aggregations. Where the spatial arrangement of the organisms is unknown or unimportant, then the smallest unit should be at least one order of magnitude larger than the size of the largest organism being

SAMPLING AND SPATIAL PATTERN

39

counted; Green (1979) suggested at least 20 times larger. The largest unit tested should be as large as logistic and cost constraints will allow, with the provisos that it is not so large that it approximates or exceeds the physico-environmental scales of interest or causes overlap among replicates when sampling a limited area. It is difficult to suggest how many sizes of sampling unit should be compared, but three should be the minimum and more than six or seven is likely to be too costly for a pilot study. In the simplest case (simple random sampling), an equal number of (at least three) replicate sampling units of each size should be collected randomly within the area to be studied. The randomization of all units within one area is intended to minimize the chance that systematic biases caused by variations within the area will affect only the replicates of a unit of particular size. For example, if all units of one size are placed in one area and all units of another size are placed in another area, then the effects of unit size will be confounded with pre-existing differences between the areas (Hurlbert, 1984). The analysis of pilot data should consider two aspects of the sample units tested—their relative accuracy and their relative precision. Relative accuracy is assessed by comparing the estimates of mean density (number of organisms/size of unit) obtained from the units of various sizes. Significant differences among the standardized means indicate differences in the relative accuracies of at least some of the unit sizes tested, although it will not necessarily be clear which unit size is the most accurate unless the true density of the population is known. Differences among means do not necessarily indicate differences in absolute accuracy; it is possible that two significantly different means are equally inaccurate, one being larger than the true population mean and the other being smaller by about the same amount. Decisions about which size of sample unit is the most accurate can only be made in the same way as decisions about the accuracy of different sampling methods (see pp. 48–50). Sale & Sharp (1983) studied the effects of changing the width of transects used in visual censuses of coral-reef fish. They found significant linear relationships between width of transect and estimated abundances of a number of species and advocated the use of such relationships to predict the expected true density of fish. Using a separate method (which they argued was more accurate than censusing along transects) they independently estimated the abundance of one of the species and compared this estimate with the ‘true’ value predicted from their regression of abundance on width of transect. There was no significant difference between the two values, but nor was there a difference between the independent estimate and the estimate obtained from the narrowest transect used for that species. can be calculated A number of avenues can be followed to compare precision. In the simplest case, for units of each size and the values compared. The unit with the smallest value will give the most precise estimates. When desired precision (p) is set,

By re-arrangement of terms, necessary replication (n) can be calculated from:

In their detailed computer simulations, Wiebe & Holland (1968) and Wiebe (1971) considered the implications of sampling patchily distributed plankton populations with nets of various sizes towed over a number of distances. They concluded that, in general, nets of larger diameter gave estimates of abundance that were both more accurate and more precise than did smaller nets. They also found that increasing the length of the tow significantly increased both accuracy and precision for all net sizes. In both papers,

40

N.L.ANDREW AND B.D.MAPSTONE

however, it was emphasized that both precision and accuracy increased with the size of net and length of tow at a decreasing rate. Thus, the most efficient sample units are not always the largest. For example, precision increased by about 26% with an increase in net diameter from 25 to 100 cm, but a further fourfold increase in the diameter of nets caused only an additional 8·2% improvement in precision. Pihl & Rosenberg (1982) took increasing After arbitrarily setting their desired precision at numbers of samples with a sampling unit of set size until they attained the required precision for estimates of abundance for the four species they studied (two fish and two crustaceans). Although they found that only 10, 15, 20, and 25 samples would be required for each of the four species, respectively, they conservatively chose to use 30 replicates in their main study. This study represents an example of the empirical rather than mathematical determination of appropriate sample size. With more complex but more potentially informative designs, a number of estimates of precision can be obtained for units of all sizes and the mean precisions compared using analysis of variance with the replicate measures of precision as data. This is possible where the pilot study is repeated at a number of similar sites or over a short time or when sufficient replicates of each sized unit are collected to allow subdivision of the data into ‘b’ subsets of n/b replicates for each size of unit (Scheffe, 1959). Measures of precision are calculated for each data set for each size of unit and the values analysed. Significant differences among the means of estimates of precision indicate better average precision for one or more unit sizes than for others. If the costs of using different sized units are similar (e.g. it might be only marginally more time consuming to count urchins in a 1 m2 quadrat than in a 0·25 m2 quadrat), then the analysis could (greatest precision) is the best choice if costs are stop here. The size of unit with the smallest value for not considered because fewer units of that size than of the other sizes will be required for any chosen level of precision. When differences in the costs among units are negligible, the cost of sampling is proportional only to sample size. In such a case, the choice of the size of the unit to be used is greatly simplified and the unit with the smallest CV should be used since that unit will yield greatest precision for any sample size. Green (1979) suggests that when precision is the same for units of all sizes then larger rather than smaller units should be favoured. It is unlikely, however, that all units will cost the same to use and hence efficiency (cost of attaining required precision) will be likely to favour the choice of one unit over others. For any size of sampling unit, precision will increase with sample size because the standard error and confidence intervals decrease with increasing replication. Increased precision is offset by the increased cost (time, effort, money) of obtaining and processing large samples. As with increases in the size of sampling units, the rate of increase in precision with increasing sample size is initially great but declines as sample size becomes large (Cochran & Cox, 1957; Cochran, 1963; Green, 1979; Scheaffer, Mendenhall & Ott, 1979). The declining rate of increase in precision suggests a situation of diminishing return for effort. It is usually economically desirable to use the smallest possible size of unit, but the cost of obtaining sufficient of them to obtain a required precision must also be considered. The total cost of sampling will be the cost per unit multiplied by the number of units required, plus any overhead costs, such as getting to the study site. To , fewer larger units will usually be required than smaller obtain a given precision, expressed as, say, units. In studies from several fields, it has been found that it is generally more economical to use many smaller sample units than a few larger ones (Gerard & Berthet, 1971; Elliott, 1977; Downing, 1979; Pringle, 1984; Downing & Anderson, 1985; Downing & Cyr, 1985; Morin, 1985). This result is not, however, universal and several authors have recommended the use of larger rather than smaller sampling units (Dennison & Hay, 1967; Wiebe & Holland, 1968; Gray, 1971; Wiebe, 1971; Kenchington, 1978). The relative economy of using sample units of various sizes should not be the sole arbiter of optimality. Particular biological, behavioural, physical or other factors will also be important in choosing units of appropriate, optimum size. For example, observer fatigue may increase bias and decrease precision when

SAMPLING AND SPATIAL PATTERN

41

large numbers of organisms are being counted (Caughley, 1977; Green, 1979; Kirchman, Sigda, Kapuscinski & Mitchell, 1982). It is also likely that small quadrats will be searched more thoroughly than large quadrats. Wiebe (1971) discusses, but does not analyse, the possible effects of costs and logistics, such as ship time and sorting time, on the choice of optimum size sampling units for studies of plankton. He also re-emphasizes suggestions by other authors that, in such studies, the choice of larger rather than smaller nets is likely to be an important means of decreasing avoidance of nets by real zooplankton in the field, an important source of inaccuracy in estimates of abundance of plankton. Such considerations will be specific to different fields. Where both cost and precision vary with unit size it has been repeatedly shown that efficiency is not always simply a function of the size of sampling units (Downing, 1979; Pringle, 1984; Downing & Anderson, 1985; Downing & Cyr, 1985; Morin, 1985). The estimate with the smallest variation will often be from the largest and probably the most costly unit. The efficiency (cost of obtaining a desired precision) of units of various sizes can be compared by cost-benefit analyses (Cochran, 1963; Saila et al., 1976; Snedecor & Cochran, 1980; Underwood, 1981; Irish & Clarke, 1984). In this case the cost-benefit analysis is the determination of the cost of using sufficient units of each size to estimate abundance with the desired precision. From the equation on page 55, number, n, of units of size u necessary for desired precision, p, is:

The toptal costs, Ct, of sampling with units of size u will be the sum of the overhead costs, co, and nu times the cost per unit, cu:

Evaluation of this expression for units of each size allows a direct comparison of their efficiency and, hence, allows the choice of optimum size of unit for the population being sampled. If overhead costs are constant and independent of size of unit, then only the last term in the above equation need be calculated. Note also that if a general cost function is desired, cost per unit can be expressed as a function of one or or size of unit (Downing & Anderson, 1985; Downing & Cyr, 1985; Morin, more variables, such as 1985). In most fields of ecology the relationships among the size of sampling units, replication, and efficiency have been examined only infrequently. Pringle (1984) compared the sampling efficiency (=cost per unit precision) of six quadrat sizes (0·25 m2−4·0 m2) used for estimating the biomass of a seaweed and found that although replication required for given precision varied linearly with size of sample units (quadrats), the relationship between total sampling time and size of quadrat was curvilinear. The relationship between cost and benefit (in terms of precision) was not a simple linear function of the size and number of quadrats used. Intermediate sized quadrats (1·0 m2 and 1.56 m2) were least efficient whilst the smallest quadrat tested was the most efficient. This means that it is more economical to take a larger number of the smallest units than a smaller number of larger units to estimate biomass with any given level of precision. Several Canadian scientists have recently reported similar investigations of optimal numbers and sizes of sampling units for the study of benthic organisms in freshwater lakes and streams. Downing (1979) and Morin (1985) analysed published data for a number of benthic animals from lakes and large rivers, and from streams, respectively. Downing & Anderson (1985) analysed original and published data for estimated

42

N.L.ANDREW AND B.D.MAPSTONE

biomass of macrophytes; Downing & Cyr (1985) did likewise for epiphytic invertebrates. These authors derived empirical relationships between variance, mean abundance and the size of sample units based on the principles of Taylor’s power law (see p. 75) and the expected change in variance with changes in unit size. The relationships were developed by multiple regression of the logarithm of variance on the logarithms of mean density and sample unit size. In all cases the relationships were highly significant and accounted for between 88 and 95% of the variation in the large sets of data fitted. These empirically derived relationships and the terms re-arranged to give an were then substituted into the formula for precision, measured as expression for n (sample size) in terms of mean abundance, size of sample unit and desired precision. The conclusions from these syntheses of large bodies of published data were relatively consistent across all of the studies. In all cases, the sample size required to achieve a given level of precision was a tight function of population density and the size of the sample units. The sample size needed to achieve the desired precision declined as the density of the population being sampled increased and as larger sample units were used. In both cases, the rate of change in replication necessary for desired precision also declined. W.D. Taylor (1980) has criticized the approach of Downing (1979) arguing that conclusions taken from the synthesis of such a variety of species, systems, and methods are of doubtful use for studies of single species. For further comment see also Downing (1980, 1981) and L.R.Taylor (1980). When functions describing the costs of obtaining and processing samples were applied to the precision formulae in the above four studies the results diverged. Downing (1979) and Downing & Anderson (1985) concluded that it was always more efficient to use many small units than fewer large ones. In the first study, the cost was estimated by the total area of sediment that would have to be collected and sorted for the required replication of units of a given size. In the second study, costs per sample were measured when data were being collected for comparisons of performance of different sized sample units (quadrats). Downing & Cyr (1985) measured the costs of collecting and processing samples using quadrats of five different sizes and concluded that a quadrat of intermediate size (500 cm2) was the most efficient. Morin (1985) concluded that when population density was great, small sample units were most efficient but that larger units were more cost effective for sampling populations of low density. These studies illustrate the importance of replication and the size of sample units to sampling efficiency and precision. The relationships identified by Downing (1979), Downing & Anderson (1985), Downing & Cyr (1985), and Morin (1985) are useful indicators for the calculation of optimum sample size but encompass data from a wide range of taxa, methods, and locations. As these authors stress, pilot studies are the best means of fine tuning methods, size of sample units, and sample size for a particular study. Stratified sampling In the preceding section, we discussed optimization procedures on the assumption that one sample was to be collected to give a single estimate for a population. Estimates from simple random sampling are most likely to be reliable when the characteristics (e.g. aggregation, density) of the population are relatively consistent throughout the area sampled. Often the population or area being sampled is obviously not homogeneous and such characteristics as density and patterns of aggregation vary from place to place. In such cases, overall abundance or average population density can be estimated more precisely if a number of subsections of the population are sampled separately. In other words, various strata of the environment or population are identified and the organisms sampled in a way that is appropriate for each stratum. This is referred to as stratified sampling (Cochran, 1963; Scheaffer et al., 1979; Snedecor & Cochran, 1980). As with singlestage sampling, the objective is the estimation of the total number of organisms in the whole area (e.g. the

SAMPLING AND SPATIAL PATTERN

43

number of scallops in a bay). Precise estimates of parameters are obtained from the weighted combination of the estimates from all strata. There are several ways to design a stratified sampling programme. In the simplest case (Stratified Simple Random Sampling) little is known about the population and strata are identified on the basis of some auxiliary variable(s) (e.g. depth, type of sediment or habitat). Equal numbers of sample units are then collected randomly from each stratum. Alternatively, the total number of replicates feasible for the programme is allocated to the strata in proportion to their areas (Proportional Stratified Sampling). In both procedures, the researcher need know nothing about the population prior to the main programme, but requires a map of the area to be sampled. In practice, optimal sampling procedures may differ among strata. The most precise estimates of the total population will be obtained when sampling is optimal in all strata. To obtain approximately similar precision for all strata, sampling effort is weighted among strata in proportion to the variance of the data obtained from each stratum. Strata with very variable population densities are more intensively sampled than those with relatively homogeneous data. Here, the allocation of samples is proportional to characteristics of the subpopulations, rather than the areas of the strata. Sample allocation can be optimized according to three sets of criteria: (1) minimize the variance of the estimate for fixed total cost; (2) minimize the total cost of obtaining an estimate with a given variance, where costs per unit are equal for all strata; and (3) minimize the total cost of obtaining an estimate with a given variance, where costs per unit vary among strata. Formulae and full discussions of these may be found in Cochran (1963), Scheaffer et al. (1979), and Snedecor & Cochran (1980). To obtain estimates of identical precision from each stratum, each stratum should be treated as a single sample and the replication needed to give desired precision calculated as discussed on pages 52–58. In this case (Stratified Sampling with Optimal Allocation), the total number of replicates is the sum of the numbers of replicates from each stratum. If this total effort is beyond the limits of the programme, then one of the proportional allocation procedures must be used or the level of precision relaxed. It may also be desirable to optimize sampling methods and/or the size of sampling units for each stratum separately, provided that this does not introduce biases that differ among strata. For optimal allocation or proportional allocation based on variance, estimates of variation are required prior to the main study. These are best obtained from pilot studies in which a small sample is recorded from each stratum. Cochran (1963) has pointed out that stratification with sample allocation based only on auxiliary variables often fails to give estimates that are more precise than those from a Simple Random Sample design. Stratified Simple Random Sampling is the stratified design most commonly used in marine ecological studies. It is common, for example, to stratify sampling by habitat. In most cases, however, this design is used to obtain data for comparison of abundances in different habitats rather than to estimate the total number of organisms in all habitats sampled. Other stratified sampling designs, better suited to estimating total numbers, have apparently not been common in marine studies. Cuff & Coleman (1979) used a Proportional Stratified Sampling design to estimate the abundances of molluscs, polychaetes and crustaceans in Western Port Bay, Australia. They allocated samples according to the areas of the strata and used the data to calculate the expected variance of the estimated populations of the three taxa for three

44

N.L.ANDREW AND B.D.MAPSTONE

designs of stratified sampling and for Simple Random Sampling. They found that the precision of the estimates was not improved by stratifying the sampling and concluded that stratification was generally no better than Simple Random Sampling. Green (1980) critically assessed Cuff & Coleman’s study and argued that they were not justified in extending their results, from only one study in one bay, to general recommendations for the design of future sampling programmes elsewhere. The failure of stratification to increase precision may well have arisen from the way in which Cuff & Coleman divided the bay into strata. More informed allocation may well have produced a more precise estimate than Simple Random Sampling (Cochran, 1963). Furthermore, Cuff & Coleman pooled counts of species and analysed the data by phyla. Green (1980) noted that this level of taxonomic resolution may be an inappropriate basis for stratification because it is unlikely that all species in a phylum will respond to environmental variables in the same way and have similar patterns of abundance. Heisig & Hoenig (1986) described a method of optimizing sampling design for estimating secondary productivity from size-frequency data. The allocation of effort among different times of the year (=strata) was optimized according to the expected abundance of organisms at each time. Multi-stage sampling Many studies are more concerned with examining patterns in abundance than with estimating the total number of organisms in an area. It may be important to assess whether a population can be considered homogeneous at various spatial scales, or whether differences in population parameters occur among, for example, different habitats. These situations have been referred to as mensurative experiments (Hurlbert, 1984). The simplest design for such comparisons is a stratified Simple Random Sampling design. In most instances, stratified sampling with proportional or optimal allocation of replicates would be more efficient for obtaining a single estimate for a population. Because of the unequal sample sizes, however, such designs are not as well suited to statistical comparisons among strata as Stratified Simple Random Sampling. Stratified designs can involve two or more hierarchically arranged levels of replication. Cuff & Coleman (1979), for example, collected replicate grab samples at each of several stations within each stratum of a bay. When replication is balanced among strata and at each level within strata, these hierarchical, multi-stage designs are perhaps the most powerful designs for comparing estimates. In general this class of designs is concerned with the estimation of differences among means for a single variable. The analysis of this type of data is often best accomplished through techniques of analysis of variance. No other analyses can formally test for both hierarchical and interactive effects within the same design. Formal introductions to these techniques of analysis may be gained from standard texts such as Winer (1971), Snedecor & Cochran (1980), and Sokal & Rohlf (1981). Underwood (1981) gave an introduction to techniques of analysis of variance and reviewed their use in marine ecology. Multi-stage designs allow the researcher to estimate variability at different scales within strata (such as habitats) and provide a stronger basis for generalization. For example, in the investigation of patchiness in the abundance of plankton, replicate samples should be taken in close proximity at sites separated by several kilometres, with several sites being sampled at stations separated by hundreds of kilometres. Another example might be the comparison of numbers of fish in kelp forests with the numbers in sea-grass beds. In order to draw general conclusions it would be necessary to sample each habitat at more than one location (separated by perhaps several kilometres) and to sample more than one site within each habitat at each location. The problem with complex designs such as these, is how to allocate effort efficiently among the various levels of the programme—how many locations, sites, and replicates give optimum precision to the estimates of abundances of fish in the two habitats.

SAMPLING AND SPATIAL PATTERN

45

Methods are readily available for such multi-stage cost-benefit analyses (see Cochran, 1963; Snedecor & Cochran, 1980; Sokal & Rohlf, 1981). Saila et al. (1976) and Underwood (1981) discussed the procedures and advantages of multi-stage cost-benefit analysis. Based on the hierarchical (or nested) analysis of variance, this form of cost-benefit analysis uses estimates of the proportion of variation explained by each level of sampling (such as locations, sites, and replicates), and the expected costs of replication at each level, to allocate sampling effort among levels so that overall variation of estimates of means is minimized. Bernstein & Zalinski (1983) have recommended the use of these procedures for planning environmental monitoring studies. Perhaps the most frequent application of multi-stage sampling is in studies of micro-organisms where hierarchical subsampling is employed to decrease the effort and expense of counting the large numbers of organisms collected in each field sample. Subsampling is routinely used in studies of plankton, benthic infauna, and bacteria. Well-developed and tested methods of subsampling and accurately splitting the samples into subunits exist, but there has been little attention to optimization. Subsampling has been done mostly by traditional rather than tested practices (Alden, Dahiya & Young, 1982; Kirchman et al., 1982; Montagna, 1982; Irish & Clarke, 1984). Kirchman et al. (1982) used cost-benefit analyses to determine the optimum number of subsamples, aliquots (=filters) and fields of view to be taken when using hierarchical Subsampling to estimate bacterial abundance in bodies of water (note that they refer to the field samples as subsamples of the water body). Irish & Clarke (1984) used cost-benefit procedures to determine optimal designs for sampling, and Subsampling, a tarn to estimate chlorophyll a concentrations and abundance of phytoplankton. Kirchman et al. (1982) found that by optimizing the design of the Subsampling procedures, “accuracy” (actually precision) of the estimates of abundance for the water bodies sampled increased by 20–50% compared with the traditionally used designs with only marginal increases in costs. They emphasize that most customary sampling designs fail to estimate variation at all levels (e.g. among fields or among filters) of the Subsampling procedure and, in the worst cases, cannot provide any statement of precision for the estimate of the number of bacteria in the sample. Montagna (1982) also found that the optimum design for Subsampling sediment samples to count bacteria involved replication of subsamples, filters, and fields of view. He also pointed out that the customary practice of counting 15–20 fields of view on only one filter per sample was inadequate. Venrick (1971) has also stressed that it is important to estimate the amount of variation at each level of Subsampling procedures. She gave a full explanation of the components of variation at each level and used a computer simulation to examine the relative effects of different subsample fractions and replication at each level on the estimation of the variation in the estimate of abundance at the highest level. She did not, however, deal with cost-benefit analyses. In other fields, spatial and/or temporal variation have only infrequently been investigated usefully by explicitly incorporating spatial scale into the design of sampling programmes. Replicates within levels (representing different scales) of hierarchical sampling programmes have often been combined with the result that variation at those scales could not be estimated (but see Lee & McAlice, 1979; Schwiegert & Sibert, 1983; Hankin, 1984; Kennelly & Underwood, 1984, 1985; Schwiegert, Haegele & Stocker, 1985; Andrew, 1986). Schwiegert & Sibert (1983) examined optimum designs for sampling Pacific herring from commercial catches to estimate the age structure and abundance of the fished population. The optimal designs varied among age classes of fish and among the different fishing ports sampled. In some instances, the analyses indicated that up to 100 fish should have been taken from several subsamples of each load that entered a port. In other cases, replication of subsamples was unnecessary and the optimal design was a two-stage

46

N.L.ANDREW AND B.D.MAPSTONE

design (a number of fish taken from several replicate loads). For further discussion, see Smith (1984) and Schwiegert & Sibert (1984). Schwiegert et al. (1985) compared the precision of estimates of the abundance of herring spawn derived from several two-stage designs and a simple random sampling design. They found that carefully designed two-stage sampling gave more precise estimates than Simple Random Sampling but that other two-stage designs were no better than Simple Random Sampling. Hankin (1984) also compared a number of methods of allocating samples among levels of a two-stage sampling design implemented to estimate the abundance of fish in freshwater streams. He found that the most efficient design was one in which primary units (sections of stream) were variable in size and corresponded to discrete habitat units, such as riffles or pools. Random secondary units were then taken within each of these primary units. Hankin’s study essentially represented an empirical demonstration of the advantages of carefully designed stratified sampling programmes for estimating the total abundance of organisms in a heterogeneous habitat. Saila, Pikanowski & Vaughan (1976) used data from a prior survey in single and multi-stage cost-benefit analyses to estimate optimal designs for sampling several species of benthic invertebrates in the New York Bight. Kennelly & Underwood (1984, 1985) used pilot studies and cost-benefit analyses to derive the optimal replication at three spatial scales for sampling micro-organisms in a temperate sub-littoral kelp forest. All of the above studies have used the minimization of sample variance, and maximization of precision of estimated means, as the criterion of ‘benefit’ in cost-benefit analyses. Millard & Lettenmaier (1986) have discussed the use of multi-stage cost-benefit analyses based on balancing statistical power against cost rather than simply considering cost-precision relationships. They argue that in many situations optimal design based on the precision of means is inadequate since no account is taken of the likelihood that a hypothesized effect will be detected. Several authors have emphasized that patterns and processes can be better understood if estimates of variability at various scales are explicitly included in the design of sampling programmes. Caffey (1985) is a good example of such a study. The shortage of studies that have done this perhaps reflects attitudes of research workers to variability and its manifestation at various scales of investigation. As with more simple designs, pilot studies are required for optimization procedures for stratified and multi-stage sampling. Ideally, different sized units and methods should be examined in a number of the habitats, sites, etc., involved in these designs, although it may not always be possible to do this at all levels of the intended programme. Tests done under the extremes of density, habitat complexity or other independent variables may suffice. In pilot studies for multi-stage sampling it is important that each level of replication in the proposed programme be addressed by replication at corresponding levels in the pilot study. Note that cost-benefit analyses in this context can only be applied to hierarchically arranged factors. Replication at higher level orthogonal factors, when not fixed implicitly by the objectives of the study, can only be designated after estimates of replication at all lower levels have been calculated (Underwood, 1981). For example, in the earlier example of estimating the abundance of fish in kelp forests and in sea-grass beds, cost-benefit analyses would allow the determination of optimal numbers of replicate transects and sites in each habitat at each location. The number of locations to be sampled can then be calculated by dividing the total resources (time, money, etc.) available by the amount of resources required to sample the optimum number of sites and transects in each habitat. When the number of treatments in orthogonal or higher level factors is predetermined by the objectives of the programme, the total affordable cost is divided equally among the orthogonal treatments. The replication of nested factors within those treatments is then determined by cost-benefit analysis based on the sub-divided cost limits (Schwiegert et al., 1985).

SAMPLING AND SPATIAL PATTERN

47

POWER ANALYSIS Most sampling programmes are suboptimal because of logistic and/or cost constraints. Procedures such as cost-benefit analysis will ensure that the best use is made of limited resources, but the question remains whether the design is good enough to do the job asked of it. Recognition of this problem makes it desirable to know, a priori, how large a difference among means could be detected or, a posteriori, how sensitive a test was. Power analysis makes such statements possible. Statistical power is defined as the probability that a test will lead to the correct rejection of the null hypothesis; that is, that the null will be rejected when the alternative is true (Sokal & Rohlf, 1981). Power is, the probability of falsely retaining conceptually and probabilistically, the complement of Type II error (Yamane, 1967; Winer, 1971; Cohen, 1977; Sokal & a null hypothesis, and accordingly is expressed as Rohlf, 1981). A clear and extensive treatment of power analysis is given by Cohen (1977) who has provided tests, F tests, analysis of tables for the determination of power for correlation coefficients, t tests, variance for fixed effects and their interactions (for a limited number of designs) and analysis of covariance. A more statistically demanding treatment may be found in Winer (1971), who also considered the power of tests involving random factors in analysis of variance. Recent discussions in the ecological literature include Underwood (1981), Bernstein & Zalinski (1983), Toft & Shea (1983), and Rotenberry & Wiens (1985). Cohen (1977) has identified four parameters of statistical inference useful in power analysis : (1) power, (3) sample size (n), and (4) effect size. Knowledge of any three will allow the (2) significance level determination of the fourth (Cohen, 1977). The relationship between these parameters provides the research worker with the opportunity to exercise greater control over the design and analysis of sampling programmes and allows better and more careful interpretation of results (Bernstein & Zalinski, 1983; Toft & Shea, 1983). Power analysis is the analysis of the sensitivity of tests of null hypotheses against specified, quantified alternatives. It may be used either in the design phase of a programme or as a post hoc test of the adequacy of a design (Winer, 1971; Cohen, 1977; Sokal & Rohlf, 1981; Underwood, 1981; Bernstein & Zalinski, 1983; Toft & Shea, 1983; Rotenberry & Wiens, 1985; Millard & Lettenmaier, 1986). For example, the researcher may wish to know what sample size is required to detect an effect of given size subject to chosen probabilities of both Type I and Type II error. Alternatively, where a test failed to reject the null hypothesis, the sensitivity of the test can be examined: was the design too weak to detect anything but very large differences (Toft & Shea, 1983)? Despite this great potential, power analysis has remained a much under-utilized technique in ecological research (Toft & Shea, 1983). Several reasons for this lack of use have been suggested. Toft & Shea (1983) have argued that it is a reflection of our pre-occupation with Type I error and also a lack of awareness of Type II error and its importance. The reluctance (or inability) of ecologists to specify exact alternative hypotheses has been cited by Underwood (1981) and Rotenberry & Wiens (1985) as another possible reason for the lack of use of power analysis. By convention, we accept a 5% chance of committing Type I error without any real idea of the probability of Type II error. It is usually argued that this is acceptable because it is better to know the probability that an effect will be incorrectly claimed than it is to know the probability of not demonstrating a real effect. This is not necessarily so, particularly when strong assertions are made from the lack of a statistically significant result—when a null hypothesis is retained (Toft & Shea, 1983). These authors have argued that where a positive conclusion is drawn from a ‘negative’ (non-significant) statistical result, it should be exposed to the same restrictive standards as a positive conclusion drawn from a ‘positive’ (significant) statistical result (see also Rotenberry & Wiens, 1985).

48

N.L.ANDREW AND B.D.MAPSTONE

In many situations, Type II error may have ramifications at least as important as Type I error, yet Type II error is rarely considered. As an ecological example, consider the situation in which research workers wish to assess the effects of commercial fishing on the numbers of adult fish. Areas that are regularly fished are sampled and mean abundance of the fished species in these areas is compared with that in areas that have been protected from fishing. A Type I error (asserting that fishing has depleted stocks when, in fact, it has not) is to be guarded against because such a decision would subject the fishermen to the imposition of strict quotas and the under-utilization of the resource. Further consideration, however, reveals that a Type II error is potentially at least as dangerous. The incorrect assertion that the stock is being fished at or below its sustainable yield (that is, existing fishing has no significant biological effect on stock size) could result in the total collapse of the fishery with consequences more severe than the imposition of quotas. Knowledge of the power of tests is particularly important in the description of pattern. Experiments test hypotheses concerning processes that might explain patterns. The pattern, or observation is taken for granted. In contrast, when hypotheses about patterns are tested, the observation itself is being questioned. The test is not concerned with why there are more gastropods here than over there, but rather whether it is true that there are more here than there. Incorrectly claiming no difference in abundance between the two areas will mean that the ecological processes influencing the density of gastropods will be assumed to be the same in both places. This could seriously delay understanding the processes that influence the number of gastropods in the two areas. of falsely In such cases, knowledge of is perhaps more important than knowing the probability rejecting the null hypothesis. Certainly, serious consideration should be given to the relative weighting of Type I and Type II errors. Setting power at a given level makes an explicit statement about the relative costs of Type I and Type II error (Winer, 1971; Cohen, 1977; Underwood, 1981; Toft & Shea, 1983). Winer (1971) has noted that the conventions of using 0·05 and 0·01 as acceptable probabilities of Type I error have little scientific or logical basis. Greater consideration of the relative costs of Type I and II error may lead to the de-sanctification of 0·05 and 0·01 as inviolate standards of Type I error. is not a The relationship between power and the more familiar probability of making a Type I error power will increase with simple one. If the size of an effect is held constant then, for a fixed value of increasing sample size. Conversely, if power and effect are fixed, then α decreases as sample size increases. That is, Type I error is less likely with larger samples than with smaller samples when power is set to a desired level. If sample size is fixed, as is often the case because of cost constraints, α and power will be directly related. More stringent requirements for power will be accompanied by decreased probability of power and effect size Type II error but increased probability of Type I error. The relations among for a two sample case are shown in Figure 1. Power analysis for planning If power analysis is to be used in the design stage of a sampling programme, then several alternative and n. The approaches may be taken. One is to determine the power of a test given effect size, are familiar and directly determination of appropriate sample size and significance level for Type I error under the research worker’s control. The determination of an appropriate effect size is much less clear, and great caution should be exercised when deciding the size of effect that is considered important (Cohen, 1977; Underwood, 1981; Rotenberry & Wiens, 1985). Effect size is defined as the standard deviation of means divided by a common standard deviation (Cohen, 1977). Where only two means are compared, effect size is the difference between the two means, expressed as a proportion of one of them. Where more than two means are being compared, effect size has no immediately intuitive meaning because a significant

SAMPLING AND SPATIAL PATTERN

49

result can be the product of differences among one or several of the means. Furthermore, the way in which similar means may be grouped and the number of groups may affect the interpretation of effect size and the power of the test. Winer (1971) and Cohen (1977) have provided various procedures by which an appropriate effect size can be calculated with reference to existing information. An appropriate effect size may be determined from previous studies in the system. Pilot studies should provide an estimate of error variance which can be used to calculate the effect size that a future sampling programme would be expected to detect. In such a case, after setting power, the only unknown becomes the numerator in the effect size ratio. Sweatman (1985) used the magnitude of differences among means that were significantly different in a similar study as the effect size criterion for calculating the power of non-significant results. In addition, Cohen has provided operational definitions of small, medium, and large effect sizes. He stressed, however, that these are relative measures, the specific value of which will be unique to the investigation. What constitutes a small effect in one system might be of overwhelming biological importance in another (Cohen, 1977; Underwood, 1981). It should be remembered that power analysis is a statistical procedure. Even if a sampling programme has great power, say 0·95 for all tests, it will be of little use if it is set within an unrealistic biological context. Millard & Lettenmaier (1986) have discussed the use of power analysis in designing environmental monitoring programmes. They advocate the a priori statement of an effect size considered to be important, possibly by legislation, and the use of cost-benefit analyses to design programmes that optimize the power of detecting an effect of that size. Unfortunately, they only consider the power of detecting main effects in complex designs. It is often the case that the interaction of these main effects (e.g. Event status×Time) are of greater interest in detecting environmental change than the main effects alone (Green, 1979; Bernstein & Zalinski, 1983, but see Hurlbert, 1984). Rotenberry & Wiens (1985) have argued that the suitability of a design be determined, not from the calculation of power for a given effect size, and n, but from the calculation of detectable effect size from and n. Decisions about the adequacy of the design would then be made after consideration of known a priori, the suitability of that effect size to the biological hypothesis being tested. By assigning subjectivity is shifted from the critical, a priori estimation of an unknown effect size, to the evaluation of suitability of a known one. That is, rather than stating an explicit, numeric alternative hypothesis, the relative importance of Type I and Type II error are stated and the worker decides upon the suitability of a calculated effect size that is expected to be detectable. Post hoc power analysis The second major use of power analysis is as a post hoc procedure to determine the sensitivity of an already completed test. The analysis here is essentially dependent on decisions already made by the research worker when the sampling programme was designed. Consequently, post hoc power analysis can only indicate the probability of erroneous retention of the null hypothesis; i.e. it can only be meaningfully applied to nonsignificant results. The only three published instances of post-hoc power analysis that we encountered (Sweatman, 1985; Andrew, 1986; Doherty & Sale, 1986) all sought to explain non-significant results. In both Andrew (1986) and Doherty & Sale (1986) the analyses revealed important deficiencies in design. A non-significant F ratio had 70–77% chance of being in error. Power analysis should not be used to justify the post hoc elevation of non-significant trends in the data to ‘pseudo-significance’, rather, it should point towards better design for future studies. It seems, therefore, that while power analysis opens the door to greater control over sampling design, the research worker should tread carefully. The greater control brings with it a greater responsibility in terms of

50

N.L.ANDREW AND B.D.MAPSTONE

Fig. 1.—The relations among Power Effect Size (E), Sample Size (n), Type I and Type II error for a hypothetical two-sample situation. In each diagram, the two curves represent the distribution of the means of samples of size n taken from two populations. and are the grand means of the two sets of estimates and the best estimates of the population means. If the null-hypothesis is true, sample means greater than will result in the incorrect rejection of a true null-hypothesis (Type I error). Thus as a percentage of all values under curve i, is represented by the area under curve i to the right of in each diagram. Conversely, if the null-hypothesis is false, values of less than will result in the incorrect retention of the null-hypothesis (Type II error). In this case, the area under curve ii to the left of in each diagram will represent (as percentage of all results described by curve ii). The area under the remainder of curve ii (to the right of ) represents the correct rejection of the (false) null-hypothesis, i.e. Power It can be seen from diagram B, that increasing a with unchanged n and effect size, results in decrease in and increased power. Diagrams C-E show the effects of increasing sample size (n) and so increasing the precision of the estimates In C, is arbitrarily fixed at the same level as in A, with the result that decreases and power increases, and power are both fixed in D as in A and, clearly, declines. Finally, in diagram E, and power are set at the same level as in A and the detectable effect size is seen to diminish considerably.

SAMPLING AND SPATIAL PATTERN

51

the realistic setting of ß and effect size. The more specifically the hypotheses are stated, the greater the usefulness of the technique. It should be noted that tests of power are inherently one-tailed; tests have no power if any effect occurred that was opposite to that proposed (Cohen, 1977). Clearly in two-tailed tests of significance, Type II error is possible at either tail of the null distribution, but cannot occur at both. Consequently, the power of a test is meaningful only when it is discussed with respect to a single alternative hypothesis. This point has been treated more fully by Sokal & Rohlf (1981). CONCLUDING REMARKS It is important to know the limitations of any sampling programme. The concept of a pilot study is one of doing a smaller, cheaper version of the anticipated main study so that the main study can be optimized. It would, therefore, be pointless expending a large portion of the total budget, time, and resources on a pilot study and subsequently severely constraining the main programme. The amount of time and resources spent on the pilot study will be specific to each project and generalizations cannot be made. The cost of a pilot study, however, will generally be outweighed in the long run by the improved efficiency, accuracy, and precision in the main sampling programme (Green, 1979). Pilot studies along the lines of those suggested on pp. 44–52 could also be designed to compare methods of sampling, with different methods being substituted for the different sized units. More elaborately, various dimensions of units (e.g. core diameter and depth, transect length and width) can be varied orthogonally in pilot studies to establish not only the optimum size (volume or area) but also the optimum dimensions of sampling units. Different methods can also be compared across a range of unit sizes by testing unit size and method orthogonally in the one pilot study. Although orthogonally testing a suite of variables is more powerful and informative, to do so may prove prohibitively expensive. Many possibilities exist, but cost will limit the extent of a pilot study. It will generally be cheaper and often sufficient to evaluate methods, unit size and replication at various levels in a set of sequential, separate pilot studies than in a single big one. Some arguments against pilot studies have been made. Smith (1984) argued that estimates of variance obtained in pilot studies are frequently biased by small sample sizes (Cochran, 1963) and that the apparent benefits of pilot studies may be misleading when a population is sampled at a number of times: a design that is optimal at one time may be very poor at another time because of changes in the characteristics of the population being sampled. Schwiegert & Sibert (1984) replied that some indication of possible improvements in design is better than consistently sampling in ignorance. If temporal variation in population characteristics is likely to cause changes in the optimum design of a programme, then small pilot studies may be required at the beginning of each sampling occasion (Kennelly & Underwood, 1985). With respect to the problem of biased estimates caused by small sample sizes, Schwiegert & Sibert (1984) pointed out that Cochran (1963) detailed methods of estimating confidence limits for sample sizes chosen on the basis of pilot studies. In general, precision remains relatively stable for sample sizes within these confidence limits. In the absence of careful design procedures, there is the risk that methods and designs for sampling will be chosen simply by reference to previous studies. In some cases, such as successive studies of a system, this may be justified. In others, however, it leads down the potentially blind alley of standardization by convention. If a large number of independent studies come to similar conclusions and widespread standardization of sampling procedures in some systems is a result, then the use of standard procedures is to be encouraged. Standardization for its own sake, against evidence of variation in the usefulness of methods and designs when used under differing conditions, is, however, to be discouraged (Gray, 1971; Loya, 1978; Omori & Hamner, 1982).

52

N.L.ANDREW AND B.D.MAPSTONE

From our discussion of power analysis, it should be clear that much is to be gained from the use of power analysis to examine the likely usefulness of the planned sampling programme and analyses. With data from the pilot studies, the proposed analyses can be thoroughly tested for their sensitivity with the expected nature of the data. It is at this stage that careful decisions have to be made about the relative importance of Type I and Type II errors. It may be that the desired levels of these errors are unattainable simultaneously with the design and replication proposed for the main sampling programme. In such a case, critical reexamination of the sampling design, of the desired levels for Type I and Type II errors, of costs of increasing replication (to increase power), and even of the basic questions being asked will be necessary if the programme is going to fulfil its intended purpose. The adjustments that have to be made will be entirely specific to each study. SPATIAL PATTERN The arrangement of organisms in space may range from aggregated (contagious, over-dispersed, clumped) through a random pattern to being, more rarely, regular (under-dispersed, uniform). In this context it is worth repeating Cassie’s (1963) explanation of the terms over- and under-dispersion. The terms refer to the distribution of density estimates about the mean rather than the spatial arrangement of individuals. In an aggregated population there will be a greater occurrence of small and large densities, hence the frequency distribution would be over-dispersed. In contrast, if the biological population were uniformly arranged in space, then density estimates would be less variable and their frequency distribution under-dispersed. Knowledge of the spatial patterns of organisms is a necessary prerequisite in establishing what processes might underlie the observed arrangement of organisms. If, for example, individuals were regularly spaced throughout an area then competition might be invoked and tested for experimentally. If they were aggregated, then a range of processes leading to that arrangement might be tested. The extent to which the arrangement of organisms deviates from a random pattern is also implicitly considered in the design of sampling programmes used to estimate their abundance. Clark & Evans (1954, p. 446) have provided a definition of spatial randomness: “In a random distribution of a set of points on a given area, it is assumed that any point has had the same chance of occurring on any sub-area as any other point, that any sub-area has had the same chance of receiving a point as any other subarea of that size, and that the placement of each point has not been influenced by that of any other point.” In this review, we are concerned only with the description of pattern and do not wish to infer processes to account for any identified patterns. The first step in the description of the spatial pattern of a population is to test the null hypothesis of spatial randomness. A null hypothesis of spatial randomness is advantageous because tests of this hypothesis are two-tailed: rejection unambiguously suggests one of the two possible alternatives, either clumped or regular. If the test fails to reject the null, then little more needs to be said about the arrangement of those organisms: random is random. The need for more detailed description, or explanation, only arises if there is evidence of non-random pattern. Furthermore, although the great majority of populations in nature are clumped, random and regular populations have both been described, (e.g. Holme, 1950; Barnes & Marshall, 1951; Stimson, 1974; Jumars, 1975a; Seapy & Kitting, 1978; Ebert & McMaster, 1981). The null hypothesis that the population under study may be random is realistic. DETECTING NON-RANDOM PATTERN Indices of spatial pattern fall into two general categories: those derived from estimates gained from some chosen sample unit, such as quadrats or cores; and those based on measurements of distances between

SAMPLING AND SPATIAL PATTERN

53

organisms. We shall consider several of the most commonly used indices from each category and discuss are based on estimates the limitations of their use. The coefficient of dispersion and Morisita’s (1959) derived from unit sampling and Clark & Evans’ (1954, 1979) R, Pielou’s (1959) and Mountford’s (1961) and Johnson & Zimmer’s (1985) I are based on estimates of distances between organisms. Measures from sample units Variance/mean ratio (Coefficient of dispersion). The variance/mean ratio is the simplest index of spatial pattern and perhaps the one most commonly used in the marine ecological literature. It is based on the characteristic that, for a population of individuals that are randomly arranged in space, the variance of a sample will equal the mean density. Because equality of the variance and mean is a defining characteristic of the Poisson distribution, deviation from this expectation provides the basis for a test of spatial randomness (Pielou, 1969). Under the null hypothesis of random spatial pattern therefore, the estimated variance (s2) and obtained from a sample of randomly placed units should be equal. mean The variance/mean ratio may be tested for significant departure from the expectation (under a Poisson process) of unity by calculation of its confidence intervals (Greig-Smith, 1964). Examples of this approach include Kosler’s (1968) investigation of the patterns of spatial pattern of meiofauna in the Baltic Sea and Dayton’s (1973) analysis of the rocky intertidal alga, Postelsia palmaeformis. More usually, departures from unity are tested for statistical significance after conversion to the statistic:

which is compared with a distribution with degrees of freedom (Southwood, 1966; Pielou, 1969; Fisher, 1970). Tests of significant departure of the variance/mean ratio from unity have been used in many studies of species ranging from polychaetes and molluscs (Rosenberg, 1974) to asteroids (Scheibling, 1980). In several papers reviewed it was unclear whether the variance/mean ratio was incorrectly compared distribution or was first converted to its test statistic I. Partial explanation for this may lie directly with a in semantic confusion, as the terms “variance/mean ratio”, “coefficient of dispersion” and “index of dispersion” appear to have slipped into synonymy (e.g. Southwood, 1966; Elliott, 1977; Diggle, 1983). The test statistic as used above (Fisher, term “index of dispersion (I)” has traditionally meant the 1970). The “coefficient of dispersion”, as originally proposed by Blackman (1942, p. 352), referred to the variance/mean ratio: “This estimate of dispersion, which might be termed a ‘reduced index of dispersion’, since it is the index of dispersion divided by the degrees of freedom, has been called by Clapham (1936) ‘relative variance’.” Blackman, therefore, clearly distinguished the “coefficient of dispersion” from the “index of dispersion”, as have many authors since then (e.g. Cassie, 1963; Greig-Smith, 1964; Pielou, 1969; Jumars, 1975a). The distinction between these terms should be maintained. Several authors have investigated the power of the “coefficient of dispersion” to detect non-random pattern (Bateman, 1950; Kathirgamatamby, 1953; Diggle, 1979; Perry & Mead, 1979; Helshe & Ritchey, 1984). In general, this statistic was a powerful test for non-randomness, even when the organisms were only mildly aggregated. The “coefficient of dispersion” was not, however, as powerful in detecting regular patterns. Bateman (1950) has shown that the “coefficient of dispersion” provided a powerful test of randomness only when sample size was greater than five. Bateman (1950) considered that if the mean density of organisms per sample unit was less than one, then the results of the test were invalid. Means of

54

N.L.ANDREW AND B.D.MAPSTONE

between one and five organisms per unit produced conservative results, there being only a small probability of rejecting the null hypothesis of random spatial pattern (see also Cassie, 1963). This caution seems to be supported by results of studies in which “coefficients of dispersion” based on mean densities of less than five were tested for significant departure from unity (e.g. Holme, 1950; Barnes & Marshall, 1951; Kosler, 1968; Gage & Geekie, 1973a). In these studies there was a strong trend towards detection of significant departure from randomness with increasing density, and in many samples with a mean density of less than five there was no significant evidence for non-random pattern. The variance/mean ratio is a useful descriptive statistic. It is perhaps best regarded as an exploratory descriptive device, capable of detecting non-random pattern, but subject to a number of limitations. Its mean, and n make it of little use as a comparative test statistic when strong dependence on unit size, these variables differ among samples (Elliott, 1977). Goodness-of-fit tests. A test for non-random pattern may also be made by directly comparing the observed frequency distribution of estimates of density to the Poisson distribution with the same mean. The test, the more recently recommended comparison is made by a goodness-of-fit test such as the traditional G test, or their nonparametric equivalent, the Kolmogorov-Smirnov test (Sokal & Rohlf, 1981). Gage & Geekie (1973a) analysed a number of data sets from an investigation of benthic fauna with both the test and tests rejected the null hypothesis in a the Kolmogorov-Smirnov test. They found that, although the two similar percentage of tests the two tests produced conflicting results for some data sets. Fifteen of 43 tests test were not significant by the Kolmogorov-Smirnov test. Gage & judged to be significant by the Geekie argued that it is generally assumed that discrepancies arise in such cases from the different power of test, size classes with low expected frequency have to be combined, thus decreasing the two tests. In the the degrees of freedom of the test. Because such pooling is not required for the Kolmogorov-Smirnov test, it would not be similarly affected at the tails of the distribution. This index is based on an analysis of the proportion of the total number of organisms Morisita’s index found in each replicate sample unit. The index is calculated from:

where Σx is the sum of individuals found in all replicates (Morisita, 1959, 1962, 1971; Southwood, 1966; for maximum regularity and n when all Elliott, 1977). Its value can range between individuals are in the same replicate (maximum aggregation). It has a value of one for a random pattern, is, therefore, strongly dependent on sample size (Green, 1966; Elliott, 1977). The significance of may be tested against the F dis-tribution by calculating

(Southwood, 1966). Alternatively, the test statistic

is distributed with degrees of freedom (Elliott, 1977). Morisita (1959), Hairston, Hill & Ritte (1971), and Elliott (1977) have investigated the behaviour of under differing spatial patterns and found it to be strongly influenced by the size of sampling unit. For an aggregated pattern with individuals randomly gives a large and constant value with increasing unit size until the size of the arranged within clumps,

SAMPLING AND SPATIAL PATTERN

55

sample unit approximates the size of the clumps. Beyond this it declines towards one. If individuals are will increase with increasing unit size before uniformly arranged within clumps then the value of declining as above. The index gave unpredictable results when applied to a pattern that was known to be has been shown to be relatively invariant under conditions of uniform (see Fig. 18 in Elliott, 1977). different abundances (Morisita, 1959; Gage & Geekie, 1973b; Elliott, 1977). is relatively invariant with Gage & Geekie (1973b) and Colby & Fonesca (1984) used the fact that differing densities in their comparisons of spatial pattern between samples of infaunal molluscs and sand crabs, respectively. Interestingly, Gage & Geekie (1973b) found considerable concordance between the “index of dispersion” (I) and in the detection of significant departures from randomness in a number of was found to be slightly more conservative than I, but both tests gave more significant their samples, goodness-of-fit test or the Kolmogorov-Smirnov test. results than tests of a Poisson model using either a Other indices based on sample units. Other indices derived from measures of variance of density estimated with sample units have been discussed in several texts (Greig-Smith, 1964; Southwood, 1966; Pielou, 1969; Elliott, 1977). These indices have received little use in the marine ecological literature and will not be discussed here. Most are variants of the variance/mean ratio and are subject to the same limitations in interpretation (Elliott, 1977). One variant that deserves some mention is the index proposed by Green (1966). The index is given by:

and is valued at for maximum regularity, 0 for random pattern, and 1 for maximum and and may, therefore, be used as a aggregation (Green, 1966). Cx is independent of variations in n, comparative index even when these values vary between samples (Green, 1966). The usefulness of Cx as a comparative statistic is limited because there is at present no test of significance for departures from randomness. Hogue (1978) and Findlay (1981) have used Cx as a descriptive statistic in studies of pattern in benthic meiofauna. Jumars (1975a) has extended the use of the “index of dispersion” to consider the spatial pattern of more than one species. By comparing the summed “indices of dispersion” with the “index of dispersion” calculated from the summed data, differences in the spatial pattern of the species could be detected. If all species are similarly arranged then the two will be similar. If species tend to be segregated in sample units then they will differ. Two studies of deep-sea meiofauna have found evidence for patchiness in the arrangement of species using this test (Jumars, 1975b; Bernstein & Meador, 1979). Cautions for the use of indices based on sample units. The widespread use of these indices probably stems from the fact that data obtained to estimate abundance can be used to test for non-random pattern. Although this simplicity may be attractive, several points should be kept in mind when interpreting results. The first and most important point is that, for aggregated populations, the results gained may be an unpredictable consequence of the size and shape of the sampling units used (Skellam, 1952; Southwood, 1966; Payandeh, 1970; Perry & Mead, 1979; Diggle, 1983; Helshe & Ritchey, 1984). The observed pattern may vary from random, to aggregated and finally regular depending on the scale at which the population was sampled relative to the scale at which it was aggregated. This sampling artefact may arise from at least two sources. The relationship between variance and mean may change due to variations in abundance alone or through variations in the variance without attendant shifts in abundance, or both. Perry & Mead (1979) have shown that the capacity of significance tests to detect significant departures from random pattern for the variance/mean ratio increases with increasing density, and suggested that this is more the result of an increase in variance with increasing density than it is due to increases in density per se.

56

N.L.ANDREW AND B.D.MAPSTONE

For aggregated populations of many different types of organisms, there is strong empirical evidence that the variance of a sample increases disproportionately with increases in its mean (Clapham, 1936; Clark & Milne, 1955; Taylor, 1961, 1971; Kosler, 1968; Gage & Geekie, 1973a; Gagnon & Lacroix, 1982). This relationship is known as Taylor’s power law, and is essentially a statement that the spatial behaviour of organisms is density-dependent (Taylor, Woiwod & Perry, 1978). Under these conditions, as the size of the sampling unit changes, aggregation, as estimated by a unit-based index may change as a function of mean density, even if the variance remains constant. Alternatively, apparent constancy of the variance/mean ratio may arise from changes in both elements of the ratio acting to cancel out each other (Downing, 1979). Indices based on distances between organisms From the above discussion, it will be clear that in the description of the arrangement of organisms in space, there are constraints imposed by the use of artificially defined sampling units. The use of quadrats or other sampling units imposes an order on the community being sampled that may mask true spatial relations among organisms. Consideration of these problems led to the development, by plant ecologists in the 1950s, of indices based on distances between organisms rather than on estimates of abundance gained from unit-based sampling. The data collected are, therefore, free from artefacts introduced by the sampling units. The general technique has been termed plotless sampling, or nearest-neighbour (N-N) analysis. Clark and Evans’ R. The earliest index based on distances between organisms to gain common usage was that proposed by Clark & Evans (1954). The measure is based on the fact that under the null hypothesis of random spatial pattern the expected mean distance between nearest neighbours (rE) can be calculated from knowledge of the density of organisms. where is an independent estimate of the density of organisms. The index R is the ratio of the observed mean distance (r0) and the expected number. R is tested for significance by the standard normal variate Z:

where

N being the total number of individuals in area A. R has been used in several studies of spatial pattern in marine organisms. An early example is Connell’s (1963) laboratory study of the spatial arrangement of colonizing Erichthonius braziliensis, a tube-dwelling amphipod. Immediately after settlement no non-random pattern was discernible, but then a regular pattern became established. Other examples include Stimson’s (1974) analysis of the spatial arrangement of corals and Wilson’s (1976) study of the bivalve Tellina tennis. Clark & Evans (1979) have extended their original index to include considerations of the spatial pattern of organisms in k dimensions. Practically, this allows the analysis of patterns along a line (one dimension) and in three dimensions and removes the restrictive assumption that organisms be positioned in two dimensions only. The extension of an analysis of spatial pattern to three dimensions might be useful in the study of infauna, or organisms inhabiting branching corals. Rohlf & Archie (1978) briefly discussed practical means of plotting the positions of organisms in three dimensions. The three dimensional form of R has apparently not yet been used in marine ecology.

SAMPLING AND SPATIAL PATTERN

57

Pielou’s and Mountford’s α Pielou (1959) has developed a different measure derived from the distribution of distances between randomly chosen points and the nearest individual rather than N-N distances. If individuals in a population are randomly arranged, then the distance between random points and nearest neighbours will be the same as N-N distances. The calculation of Pielou’s index requires a set of such distances and the density of the population. The index is given by:

where d is the density of organisms and is the mean, squared distance between random points and values greater than one indicating nearest-neighbours. Random pattern will be indicated by clumpedness, and values less than one indicating regularity. may be tested for significant departure from 1 by the test statistic

where n is the number of distances measured. A calculated value greater than the tabulated at a significance level of (usually) 0·05 indicates significant aggregation, while a value less than the (usually) 0·95 tabulated value indicates a regular pattern. Mountford (1961) has demonstrated that when density is estimated rather than known absolutely (by enumeration of the entire population) then may result in excessive numbers of erroneous departures from randomness, compared with that expected for a randomly arranged population. He provided a corrected version of the test statistic for that takes into account the estimation (rather than absolute knowledge) of density. has been used to detect non-random spatial pattern in studies of intertidal gastropods and subtidal sea urchins and gastropods (Underwood, 1976; Andrew & Stocker, 1986; Choat & Andrew, 1986). Mountford’s corrected test statistic was not used in these studies because the whole population used in the calculation of was counted. Density was therefore known, not estimated. Johnson and Zimmer’s I. Johnson & Zimmer (1985) have recently proposed a new index of spatial pattern (I). This index seems to offer several advantages over previous indices because all that is needed for its calculation is a set of distances between random points and the nearest individual. The calculation of I does not require an independent estimate of density. It can be used to consider the arrangement of organisms in three dimensions, as can R. When applied to sets of real data Johnson & Zimmer’s index compared favourably with results given by Fisher’s “index of dispersion” and Pielou’s and Mountford’s (Johnson & Zimmer, 1985). When tested with simulated data, it proved to be more powerful than in detecting regular patterns and an aggregated pattern that followed the Negative Binomial distribution. Although the behaviour of I under different spatial patterns and sampling properties has yet to be fully explored, it would seem to have great potential. No accounts of its use have so far appeared in the marine ecological literature. Cautions for the use of indices based on distances. The three indices considered are based on the assumption that an infinite population is being sampled, i.e. there are no boundaries to the area. Edge or boundary effects will arise if the individual (or random point) from which distances are measured is close to the edge of the area considered because there is a higher probability that its nearest neighbour will be outside the area of study. This is less likely to be true if the base individual were in the centre of the area. Selection of the closest individual within the area may therefore introduce bias by over-estimating the distance to nearest-neighbour.

58

N.L.ANDREW AND B.D.MAPSTONE

Sinclair (1985) has shown that boundary effects may have a large impact on R in the above form, biasing results in favour of regularity by over-estimating rE. He went on to discuss a correction factor proposed by Donnelly (1978, in Sinclair, 1985) that counters this bias. Sinclair demonstrated that when the corrected test indicated a significantly regular pattern, with a probability of Type I error of 0·05, the probability of Type I error in the uncorrected test was actually 0·40. This enormous discrepancy arose from a population bounded within a square, which is the most common shape of quadrat. The magnitude of this bias will depend on the shape of the area considered with longer boundaries relative to the area enclosed meaning a greater bias. An alternative method of avoiding biases introduced by edge effects is to create a buffer, or border, around the edges of the area. Individuals in this area may be included as neighbours but may not be used as base individuals. Although this tactic was suggested by Clark & Evans (1954) in their original paper it has rarely been used (Sinclair, 1985). An appropriate size for such a buffer may be calculated as that distance within which a large proportion of N-N distances fall. Anderson & Kendziorek (1982) set that proportion to be 0·9 and found the required distance (d) by solving the equation

where is the estimated density within the area. The formula may be used for any proportion by appropriate substitution. Similarly, Harvey, Ryland & Hayward (1976, their p. 101) set the size of the buffer around a circular area “…such that individuals lying at the perimeter in a randomly distributed population would have a neighbour nearer than the edge of the disk on 95% of occasions…”. They arrived at that distance by solving the formula

for the square of the required width of the buffer, is the density. They noted that for aggregated populations, this distance would be an over-estimate. Several studies, e.g. Underwood (1976) and Andrew & Stocker (1985) have used buffers to avoid biases introduced by edge effects. Kinzie & Snider (1978) and Simberloff (1979) have considered the effect that the assumption of point processes has on the results gained from analyses of spatial pattern based on N-N distances. Both papers have demonstrated that the assumption that individuals were points in space, i.e. they had no area (or volume), can cause misleading results. The magnitude of the errors introduced will increase with the size of the individuals relative to the distances between them. Simberloff (1979) provided solutions to this problem for Clark & Evans’ R. He provided an approximate algorithm to correct this bias for situations in which the average diameter of organisms (assuming they are circular) is less than half of the expected mean N-N distance. Simberloff suggested the use of simulation procedures to estimate the magnitude of the bias for organisms with diameters greater than half the expected mean N-N distance. Simberloff’s consideration of these biases was extended to the three dimensional form of R. Simberloff re-analysed Connell’s (1963) data, taking into account the size of the amphipods, and although Connell’s interpretations were confirmed, the probability that the significant regularity was an error was increased. A preliminary re-analysis of Stimson’s (1974) and Wilson’s (1976) data suggested that results from both studies may have been similarly affected by the assumption that the individuals under study had no area. The expected mean N-N distance in the two studies was dependent on the density of the organisms, and ranged between 2·9 and 6·7 cm and between 16·7 and 28·8 cm, respectively. The magnitude of bias introduced would depend on the size of the organisms; the larger the organisms relative to these N-N

SAMPLING AND SPATIAL PATTERN

59

distances the greater the bias. Although the sizes of the corals and bivalves were not given, it would appear likely that, at least at the higher densities considered, the expected mean N-N distances would have been under-estimated, therefore favouring regularity. Stimson found that tests of R consistently indicated significant regularity. Interestingly this result was in contrast to a random pattern as indicated by the “index of dispersion” (but see p. 73 for limitations of that index). Wilson did not reject the null hypothesis of random pattern for all densities measured in the field. In the laboratory there was some evidence for aggregation but only in the lower two-thirds of the density range considered. In the absence of size estimates, the extent to which the statistical significance of these results, and interpretation, are in error, is a matter of conjecture. Anderson & Kendziorek (1982) have made use of the modified form of R in their detailed study of the spacing patterns of tube-dwelling polychaetes. They found significant evidence for regularity in the spacing of individuals. Pielou (1959) has demonstrated the importance of randomly selecting the base individual when using Clark & Evans’ R. If individuals are chosen by some short-cut method, such as choice of random areas followed by random selection of individuals within those areas, then the results of the test will be biased in favour of regularity. Assuming an aggregated population, randomly choosing small areas, and then individuals within them, would decrease the probability of selecting individuals within aggregations. True randomization would require the identification of all the organisms in the population and the selection of a sample using random numbers—a task that would be enormous if not impossible in many instances. Clark & Evans (1979) noted that this was one of the major practical limitations of their technique. Pielou (1969) has also pointed out an important distinction between those indices based on N-N distances, such as R, and those derived from estimation of distances between random points and nearest-neighbours. Pielou has identified two aspects of spatial pattern: intensity and grain. The intensity of a pattern may be thought of as the variability in density from place to place (Pielou, 1969). The grain of spatial pattern may be thought of as the way that variability in density is arranged in space. For example, a coarse-grained pattern may take the form of aggregations that are widely spaced with large areas of low density between aggregations. Alternatively, if there are large fluctuations in density over short distances, then the pattern may be considered to be fine-grained. The grain of a pattern may be independent of its intensity (Pielou, 1969). Indices of spatial pattern based on N-N distances measure only the intensity of pattern (Pielou, 1969). This is because they are derived from distances between organisms and in an aggregated population the majority of distances measured will be within clumps. The contribution from individuals outside clumps will be less than that from isolated individuals if random point to nearest-neighbour distances were measured. The latter indices are therefore influenced by the grain of the pattern to a greater degree than the N-N indices (Pielou, 1969). Advantages of the various indices It is not easy to compare the relative merits of tests for non-random spatial pattern based on numbers of individuals in sampling units with those based on measurements of distances. Both have advantages, whose relative worth will depend on the situation considered. Where distances between individuals can be measured, the distance techniques would seem to be preferable because they are less subject to artefacts introduced by the sampling units. For many research workers, however, the ability to measure distances among individuals may be an unobtainable luxury, e.g. those working on plankton ecology or on deep-sea fishes. Measures based on numbers of individuals per sample unit are the only option open under such conditions. Interpretations should always be constrained by the limits of the technique of sampling and

60

N.L.ANDREW AND B.D.MAPSTONE

analysis. If indices based on data from sampling units are used, then explicit statements should be made about the scale at which any deviation from random pattern was observed. Detecting spatial pattern in heterogeneous areas The discussion so far has assumed that samples were taken from homogeneous areas, i.e. the probability of an individual occurring at any point was constant throughout the area. An ecological problem that has generated some recent work concerns the determination of boundaries of occurrence of species along gradients. The assumption of homogeneity clearly does not apply, so different probability rules are necessary. Pielou & Routledge (1976) applied Pielou’s (1975, as quoted in Pielou & Routledge, 1976) test for the independent assortment of boundaries to the occurrence of plant species along a gradient in a salt marsh community. The test is based on the determination of the probability of a boundary (i.e. presence and absence of a species in adjacent quadrats) occurring along a continuous transect of quadrats running along the gradient. Gardiner & Haedrich (1978) and Underwood (1978) have independently shown that the probability theory used by Pielou to establish the expected occurrence of boundaries within quadrats biased the test in favour of regularity (but see Pielou, 1979). A corrected version of the test based on probability rules that distinguish among species was provided in both papers. Gardiner & Haedrich applied this modified test to an analysis of boundaries of the occurrence of deep-sea fauna along a single transect. Underwood (1978) considered the problem of species’ boundaries in replicated transects on a rocky intertidal shore. Chaloupka & Hall (1984) also examined the occurrence of species’ boundaries with increasing height up intertidal shores. These authors have extended the Gardiner & Haedrich and Underwood test to consider situations in which the number of species boundaries per quadrat is restricted. Abel, Williams, Sammarco & Bunt (1983) have applied the same probability rules as Pielou & Routledge for the determination of the probability of a species boundary occurring in any given quadrat, but have constrained the number of boundaries that can occur within a single quadrat. They applied the test to a study of the distribution of corals in the Caribbean Sea. DESCRIBING NON-RANDOM PATTERN The indices so far described are sufficient only to detect deviations from spatial randomness, and provide no information on the details of those non-random patterns. For example, they cannot detect any underlying periodicity or constancy to the pattern or whether it takes the form of a mosaic of patches of high and low density; Jumars, Thistle & Jones (1977) provide an illustration of this point. Furthermore, the indices give no indication of scales at which pattern might be found other than the scale at which the data used to calculate the index are gathered (Hill, 1973). In Pielou’s (1969) terminology, the indices tell us only about the intensity (variability in density from place to place) of pattern and, although some are influenced by grain (Pielou, 1969), they say nothing explicit about the grain of a non-random pattern. Additional levels of structure may overlay pattern at the lowest level, i.e. aggregations of aggregations or large scale uniformity in the data. Pattern may be most evident at the finest scale (among-individual), and the knowledge that non-random pattern exists at that scale may often be all that is required (see earlier references). In some instances, however, this is not sufficient, and the investigator may wish to describe pattern more comprehensively. The description of pattern and the detection of higher orders of pattern are more difficult than demonstrations of non-randomness. The data required are more exhaustive and the analyses used more

SAMPLING AND SPATIAL PATTERN

61

complex (and consequently more open to mis-interpretation). Two techniques that have been widely used in the ecological literature are Spectral Analysis and Pattern Analysis. Both require intensive, systematic sampling, but only techniques based on this sort of data will provide extensive information on the nature of pattern in a community. The point estimates introduced by Jumars et al. (1977) provide some information based on a limited amount of systematic sampling, but if detailed descriptions of non-random pattern are required then extensive systematic sampling is required. Of course the optimal, but rarely possible, solution is to map the positions of all organisms in the entire area. Although detailed discussions of Spectral Analysis and Pattern Analysis will not be attempted here, a very brief introduction to the literature is appropriate. Spectral Analysis describes observed patterns by fitting a linear combination of wave forms to variability in data collected as a series of counts along a transect (Ripley, 1978). The use of Spectral Analysis in ecology has been well reviewed by Platt & Denham (1975), Fasham (1978), and Ripley (1978). Despite the enthusiasm Platt & Denham (1975) had for the future of the technique as a source of a new theoretical framework for ecological research, the only area in which it has been used extensively in the marine ecological literature has been in the study of patchiness of plankton. Fasham (1978) has reviewed this literature and so we shall not discuss its use in this field here. Ripley (1978) cautioned that some experience is required if Spectral Analysis is to be interpreted correctly. Pattern Analysis, or “contiguous quadrat analysis”, has been used extensively by ecologists studying terrestrial plants (Greig-Smith, 1952; Kershaw, 1957; see Greig-Smith, 1979, for review), yet has been virtually ignored by marine ecologists. The following brief discussion is intended more as an introduction to the literature than a review of usage. As first outlined by Greig-Smith (1952), Pattern Analysis involved the comparison of counts from contiguous quadrats of increasing size. The area under study was divided into a grid of contiguous quadrats and density estimated within each. Numbers from adjacent quadrats were then summed, producing estimates from quadrats double the original size, which were then doubled, and so on. Thus the size of quadrats considered increased as powers of two. Variances of the mean counts for each size of quadrat were then plotted against quadrat size. Peaks in the plots occur at scales at which there is great variability in estimates of density. These peaks are interpreted as being the scales at which non-random pattern occurred. Extensive descriptions of the technique may be found in Greig-Smith (1964, 1979). This original formulation has been shown to have several limitations (Thompson, 1958; Pielou, 1969; Kershaw, 1970; Errington, 1973; Mead, 1974; Usher, 1975; Ludwig & Goodall, 1978; Upton, 1984). The most serious among these arose from the non-independence of estimates of variance among sizes of quadrat. This reduced the usefulness of the technique because the F ratios used to test for the significance of peaks were based on non-independent mean square estimates, and therefore were invalid. Other problems lay in the critical dependence of the starting point of the transect on the estimation of the scale at which the pattern (s) occurred. This was especially a problem when there was a periodic pattern of variability in density along a transect. If the transect started in an area of low density, then the estimated distance between peaks in variance would differ from that given if the start coincided with a peak. Another problem was that the length of the transect increased by powers of two and, therefore, there were often relatively few data points to analyse (Pielou, 1969; Goodall, 1974). There have been a number of derivations of the original technique of Pattern Analysis proposed to circumvent the problems posed above, but none has been entirely successful. Mead (1974) modified the test test for anomalous groupings (i.e. groupings of high to compare quadruplet groupings of quadrats by a or low density). Upton (1984) has subsequently demonstrated that Mead’s test is subject to the same problem of critical dependence on starting point as the original formulation.

62

N.L.ANDREW AND B.D.MAPSTONE

The only study we came across that had used classical Pattern Analysis in a marine ecological study was Grassle’s (1973) study of pattern in a coral community. Our failure to find more studies using this technique seems likely to be more a reflection of our limited search than a true picture of the popularity of the technique. It seems improbable that Grassle’s paper is the only instance where some form of Pattern Analysis has been applied to a marine problem, particularly in the phycological literature. Fitting contagious distributions Where significant departures from random spatial pattern have been demonstrated, many different statistical distributions have been fitted to summarize data. In particular, the negative binomial distribution, has proved popular as a model for contagious populations, e.g. Gärdefors & Orrhage (1968), Oviatt & Nixon (1973), Fasham, Angel & Roe (1974), and Todd (1978). See Cassie (1963) for a review. The negative binomial distribution is described by two parameters, mean density and the exponent k (Elliott, 1977; Taylor, Woiwod & Perry, 1979, and references therein). The reciprocal of k has been used as a measure of clumping of individuals: as the population tends to a random (Poisson) distribution 1/k tends to zero. Taylor et al. (1978, 1979) have investigated the utility of the negative binomial distribution as a model for the spatial pattern of organisms. In order for k to be a useful descriptor of spatial arrangement, it must behave in a predictable way when density and the degree of aggregation changes. It fails to do this. Taylor et al.’s (1979) analysis suggests that k has a rather dim future as a measure of dispersion. They found that even assuming that the negative binomial distribution adequately fitted the data, it had a number of severe limitations in its application to ecological problems. Numerous authors have cautioned against drawing ecological conclusions from the apparent fit of statistical models such as the negative binomial (Pielou, 1969; Bliss, 1971; Sokal, p. 374 in Taylor, 1971; Taylor et al., 1978, 1979; Todd, 1978). Although the data may fit a given distribution, it cannot then be assumed that the mathematical processes that determined that particular function are adequate descriptors of processes that determined the spatial arrangement of the organisms. Goodall (1974) suggested that the fitting of such contagious models has not been helpful in generating interesting and testable hypotheses about the arrangement of organisms in space. Taylor et al. (1979, their p. 301) concluded that the study of aggregation might best proceed if “based solidly on real data unprejudiced by preconceptions from models having little correspondence with reality”. CONCLUDING REMARKS If statements about spatial patterns are based on counts from a sample of randomly placed units, then only limited statements can be made about the intensity of the non-random pattern (Pielou, 1969). Non-random pattern can be claimed with more confidence from sets of distances between organisms, but unless the data contain detailed information about the spatial relations among individuals, i.e. the positions of organisms are mapped, then little more than a demonstration of non-random pattern is possible. Rapid methods for plotting the positions of organisms in the field may be found in Underwood (1977), Rohlf & Archie (1978), and Weinberg (1981). Ripley (1978) and Diggle (1983) have provided detailed discussions of the types of analyses that may be applied to mapped patterns. A review of the literature suggests that data of this completeness have only rarely been collected in marine systems. The overwhelming majority of studies concerning the spatial patterns of organisms seem content with testing for non-random pattern. As models attempting to explain the organization of communities become more complex, information more detailed

SAMPLING AND SPATIAL PATTERN

63

than demonstrations of non-random pattern will be required to test hypotheses related to the spatial relations of organisms. GENERAL CONCLUSIONS It will be evident that there are no easy prescriptive solutions to the problems of designing a sampling programme. To paraphrase Williams (1978), slightly out of context, there are only vague answers to such precise questions as: where?, how big? and how many? The best way to sample the abundance of an organism has to be determined for each individual study. We have reviewed the application of sampling procedures in the marine ecological literature and found considerable evidence of standardization of methodologies and of adherence to customary ways of doing things. Following a discussion of well-established criteria used to design better, more efficient sampling programmes, we have joined others (e.g. Underwood, 1981; Hurlbert, 1984) in seeking to attract attention to the advantages to be gained from good design. The reappraisal of methodologies continues to be a topic of much research and discussion in other fields. For example, in the literature on wildlife management there are many papers dealing with the types of issues we have discussed here. The great logistic difficulties facing many studies in marine ecology should prompt similar discussion in this field (see also Omori & Hamner, 1982). Our recommendations should not be misconstrued as a plea for the complete breakdown of established techniques. Hopefully, popular techniques will prove to have a sound basis justifying their common usage. The important point is that their use should depend upon their demonstrated suitability to each study, rather than simply on convention. It may be argued that the use of different sampling units will make comparative work difficult because the estimates will be expressed in different units. The ability to standardize estimates of abundance among samples from units of different sizes, circumvents this objection. The standardization is, however, subject to several assumptions and limitations. Assuming that different estimates of mean abundance are equally accurate and, therefore, that sampling units of different size will yield proportional results, data may be scaled to a different size of unit. Variance estimates can be scaled as the square of the change in unit size, i.e. if the unit has to be doubled to the standard unit then the variance of the mean should be increased fourfold. The description of spatial pattern from estimates based on units cannot be standardized so simply. As discussed previously, if the population is aggregated, the relationship between the mean and variance of a sample will vary unpredictably with the size of the sample unit. Unless the relationship between the variance and mean is known for the density range considered, the degree to which variance will change with a given change in mean cannot be calculated. Measures of spatial pattern derived from adjusted estimates of abundance will, therefore, generally be invalid. We suggest that the advantages to be gained from designing programmes so as to gain accurate and precise estimates, with minimal cost, far outweigh any disadvantages. It should be borne in mind that, given the evidence we have presented, the advantages of using a standardized method for comparative work may be illusory because estimates of abundance may be poor and, hence, of little comparative value in any event. If no statements about the accuracy and precision of estimates are made in studies that are to be compared then it can only be assumed that the estimates are equally reliable and therefore truly comparable. We do not suggest that extensive discussions of sampling methodologies be published in every paper. All that is required is a brief justification of why the organisms were sampled the way they were. Such brief statements will also be an aid to those who are designing similar programmes by providing a range of

64

N.L.ANDREW AND B.D.MAPSTONE

possible methods to be considered. Toft & Shea (1983) and Underwood (1981) have made similar pleas for the publication of power of tests, or at least the provision of sufficient details for others to calculate power. Details of replication and sources of variation are now generally required in describing experimental studies. There is no lesser need in the description of sampling programmes. Ecologists have long recognized the importance of justifying their sampling methods. The German naturalist Victor Hensen, one of the first ecologists to estimate the abundance of plankton, was concerned with establishing the validity and reproducibility of his results (Hensen, 1911, 1912, as quoted in Lussenhop, 1974). The development of more sophisticated methodologies for the collection of data and the increasingly large array of statistical procedures for their analysis make the need to evaluate the accuracy and precision of the sampling even greater today. Fortunately, the tools to do that evaluation have also been developed considerably since Hensen’s day, so it is also easier. Optimization techniques, such as cost-benefit analysis and power analysis, put the modern marine ecologist in a good position to design sampling programmes that provide accurate and precise descriptions of natural systems. Our impression is that the description of pattern has become the poor cousin of experimental work in ecological studies. We do not take issue with the need for manipulative experiments to investigate processes in ecology. Experiments, however, will be difficult to design and interpret without extensive and reliable natural history and distributional information. We hope our discussion will prompt a greater attention to how marine ecologists describe spatial pattern. ACKNOWLEDGEMENTS We thank A.J.Underwood for his advice and extensive criticisms of the manuscript. We are grateful to J.H.Choat, P.G.Fairweather, K.McGuiness, P.F.Sale and H.P.A.Sweatman for discussion and constructive comments on the manuscript. We hope they will accept the path we chose through their sometimes conflicting criticisms. We are especially indebted to L.J.Stocker for her many contributions, in formulating ideas, giving us advice and encouragement, and in the preparation of the manuscript. Funding for this project was provided by a Commonwealth Scholarship and Fellowship Plan Award (to N.L.A.), a Commonwealth Postgraduate Research Award, and grants from the Great Barrier Reef Marine Park Authority and from the Australian Coral Reef Society (to B.D.M.). We are grateful for the support provided by the School of Biological Sciences, University of Sydney. REFERENCES Abel, D.J., Williams, W.T., Sammarco, P.W. & Bunt, J.S., 1983. Mar. Ecol. Prog. Ser., 12, 257–265. Alden, R.W., Dahiya, R.C. & Young, R.J., 1982. J. exp. mar. Biol. Ecol., 59, 185– 206. Anderson, D.J. & Kendziorek, M., 1982. J. exp. mar. Biol. Ecol., 58, 193–205. Andrew, N.L., 1986. J. exp. mar. Biol. Ecol., 97, 63–79. Andrew, N.L. & Stocker, L.J., 1986. J. exp. mar. Biol. Ecol., 100, 11–23. Barnes, H. & Marshall, S.M., 1951. J. mar. biol. Ass. U.K., 30, 233–263. Bateman, G.I., 1950. Biometrika, 37, 59–63. Bell, J.D., Craik, G.J.S., Pollard, D.A. & Russell, B.C., 1985. Coral Reefs, 4, 41– 44. Bernstein, B.B. & Meador, J.P., 1979. Mar. Biol., 51, 179–183. Bernstein, B.B. & Zalinski, J., 1983. J. Environ. Mgmt, 16, 35–43. Blackman, G.E., 1942. Ann. Bot. N.S., 6, 351–370. Bliss, C.I., 1971. In, Statistical Ecology, Vol 1, Spatial Patterns and Statistical Distributions, edited by G.P.Patil et al., Pennsylvania State University Press, Pennsylvania, pp. 311–335.

SAMPLING AND SPATIAL PATTERN

65

Bouchon, C., 1981. Mar. Ecol. Prog. Ser., 4, 273–288. Bradley, J.V., 1968. Distribution-free Statistical Tests. Prentice Hall, London, 388 pp. Brock, R.E., 1982. Bull. mar. Sci., 32, 269–276. Caffey, H.M., 1985. Ecol. Monogr., 55, 313–332. Cassie, R.M., 1963. Oceanogr. Mar. Biol. Ann. Rev., 1, 223–252. Caughley, G., 1977. Analysis of Vertebrate Populations. John Wiley & Sons, New York, 234 pp. Chaloupka, M.Y. & Hall, D.N., 1984. J. exp. mar. Biol. Ecol., 84, 133–143. Choat, J.H. & Andrew, N.L., 1986. Oecologia (Berl.), 68, 387–394. Clapham, A.R., 1936. J. Ecol., 24, 232–251. Clark, P.J. & Evans, F.C., 1954. Ecology, 35, 445–453. Clark, P.J. & Evans, F.C., 1979. Ecology, 60, 316–317. Clark, R.B. & Milne, A., 1955. J. mar. biol. Ass. U.K., 34, 161–180. Cochran, W.G., 1963. Sampling Techniques. John Wiley & Sons, New York, 2nd edition, 413 pp. Cochran, W.G. & Cox, G.M., 1957. Experimental Designs. John Wiley & Sons, New York, 2nd edition, 611 pp. Cohen, J., 1977. Statistical Power Analysis for the Behavioural Sciences. Academic Press, New York, revised edition, 474 pp. Colby, D.R. & Fonseca, M.S., 1984. Mar. Ecol. Prog. Ser., 16, 269–279. Connell, J.H., 1963. Res. Popul. Ecol., 5, 87–101. Cuff, W. & Coleman, N., 1979. J. Fish. Res. Bd Can., 36, 351–361. Dayton, P.K., 1973. Ecology, 54, 433–438. Dayton, P.K. & Tegner, M., 1984. In, A New Ecology: Novel Approaches to Interactive Systems, edited by P.W.Price et al., John Wiley & Sons, New York, pp. 457–481. DeMartini, E.E. & Roberts, D., 1982. Mar. Biol., 70, 129–134. Denman, K.L. & Gargett, A.E., 1983. Limnol. Oceanogr., 28, 801–815. Dennison, J.M. & Hay, W.W., 1967. J. Paleont., 41, 706–708. Diggle, P.J., 1979. Biometrics, 35, 87–101. Diggle, P.J., 1983. Statistical Analysis of Spatial Point Patterns. Academic Press, Inc., London, 148 pp. Doherty, P.J. & Sale, P.F., 1986. Coral Reefs, 4, 225–234. Downing, J.A., 1979. J. Fish. Res. Bd Can., 36, 1454–1463. Downing, J.A., 1980. Can. J. Fish. Aquat. Sci., 37, 1329–1330. Downing, J.A., 1981. Can. J. Fish. Aquat. Sci., 38, 127–129. Downing, J.A. & Anderson, M.R., 1985. Can. J. Fish. Aquat. Sci., 42, 1860–1869. Downing, J.A. & Cyr, H., 1985. Can. J. Fish. Aquat. Sci., 42, 1570–1579. Eberhardt, L.L., 1978a. J. Wildl. Mgmt, 42, 1–31. Eberhardt, L.L., 1978b. J. Wildl. Mgmt, 42, 207–238. Ebert, T.A. & McMaster, G.S., 1981. J. Ecol., 69, 559–564. Elliott, J.M., 1977. Methods for the Analysis of Samples of Benthic Invertebrates. Freshwater Biological Association, Scient. Publ. No. 25, Ferry House, U.K., 160 pp. Errington, J.C., 1973. J. Ecol., 61, 99–105. Fairweather, P.G. & Underwood, A.J., 1983. Oecologia (Berl.), 56, 169–179. Fasham, M.J.R., 1978. Oceanogr. Mar. Biol. Ann. Rev., 16, 43–79. Fasham, M.J.R., Angel, M.V. & Roe, H.S.J., 1974. J. exp. mar. Biol. Ecol., 16, 93–112. Findlay, S.E.G., 1981. Estuar. cstl shelf Sci., 12, 471–484. Fisher, R.A., 1970. Statistical Methods for Research Workers. Oliver & Boyd, Edinburgh, 14th edition, 362 pp. Gage, J. & Geekie, A.D., 1973a. Mar. Biol., 19, 41–53. Gage, J. & Geekie, A.D., 1973b. Mar. Biol., 20, 89–100. Gagnon, M. & Lacroix, G., 1982. J. exp. mar. Biol. Ecol., 56, 9–22. Gärdefors, D. & Orrhage, L., 1968. Oikos, 19, 311–321. Gardiner, F.P. & Haedrich, R.L., 1978. Oecologia (Berl.), 31, 311–317.

66

N.L.ANDREW AND B.D.MAPSTONE

Gerard, G. & Berthet, P., 1971. In, International Symposium on Statistical Ecology, Vol. 1. Spatial Patterns and Statistical Distributions, edited by G.P.Patil et al., Pennsylvania State University Press, Pennsylvania, pp. 59–67. Goodall, D.W., 1974. Vegetatio, 29, 135–146. Gove, P.B., 1969. Editor. Websters Third International Dictionary of the English Language. Merriam, Massachusetts, 2662 pp. Grassle, J.F., 1973. In, Biology and Geology of Coral Reefs, edited by R.Endean & O.A.Jones, Academic Press, New York, pp. 247–270. Gray, C.A. & Bell, J.D., 1986. Mar. Ecol. Prog. Ser., 28, 43–48. Gray, J.S., 1971. In, Proc. 1st Int. Conf. Meiofauna, Smithson. Contr. Zool. No. 76, edited by N.C.Huling, Smithsonian Institute Press, Washington, pp. 191–197. Green, R.H., 1966. Res. Popul. Ecol., 8, 1–7. Green, R.H., 1979. Sampling Design and Statistical Methods for Environmental Biologists. John Wiley & Sons, New York, 257 pp. Green, R.H., 1980. Can. J. Fish. Aquat. Sci., 37, 296 only. Greig-Smith, P., 1952. Ann. Bot. N.S., 16, 293–316. Greig-Smith, P., 1964. Quantitative Plant Ecology. Butterworths, London, 2nd edition, 256 pp. Greig-Smith, P., 1979. J. Ecol., 67, 755–779. Hairston, N.G., Hill, R.W. & Ritte, U., 1971. In, International Symposium on Statistical Ecology Vol. 1. Spatial Patterns and Statistical Distributions, edited by G.P.Patil et al., Pennsylvania State University Press, Pennsylvania, pp. 337–356. Hankin, D.G., 1984, Can. J. Fish. Aquat. Sci., 41, 1575–1591. Hargrave, B.T. & Burns, N.M., 1979. Limnol. Oceanogr., 24, 1124–1136. Harris, G.P., 1980. Can. J. Fish. Aquat. Sci., 37, 877–900. Harvey, P.H., Ryland, J.S. & Hayward, P.J., 1976. J. exp. mar. Biol. Ecol., 21, 99– 108. Haury, L.R., 1973. Limnol. Oceanogr., 18, 500–506. Heisig, D.M. & Hoenig, J.M., 1986. Limnol. Oceanogr., 31, 211–215. Helshe, J.F. & Ritchey, T.A., 1984. Biometrics, 40, 877–885. Hill, O., 1973. J. Ecol., 61, 225–235. Hogue, E.W., 1978. Mar. Biol., 49, 211–222. Holme, N.A., 1950. J. mar. biol. Ass. U.K., 29, 267–280. Holme, N.A. & McIntyre, A.D., 1971. Methods for the Study of Marine Benthos, IBP Handbook No. 16, Blackwell, Oxford, 334 pp. Hulings, N.C. & Gray, J.S., 1971. In, A Manual for the Study of Meiofauna, Smithson. Contr. Zool., No. 78, edited by N.C.Hulings & J.S.Gray, Smithsonian Institute Press, Washington, pp. 16–24. Hurlbert, S.H., 1984. Ecol. Monogr., 54, 187–211. Irish, A.E. & Clarke, R.T., 1984. Brit. Phycol. J., 19, 57–66. Johnson, R.B. & Zimmer, W.J., 1985. Ecology, 66, 1669–1675. Jones, J.G., 1974. Limnol. Oceanogr., 19, 540–543. Jumars, P.A., 1975a. Mar. Biol., 30, 253–266. Jumars, P.A., 1975b. Mar. Biol., 30, 245–252. Jumars, P.A., Thistle, D. & Jones, M.L., 1977. Oecologia (Berl), 28, 109–123. Kathirgamatamby, N., 1953. Biometrika, 40, 225–228. Keast, A. & Harker, J., 1977. Environ. Biol. Fish., 1, 181–188. Kenchington, R.A., 1978. In, Coral Reefs: Research Methods, Monogr. Oceanogr. Methodol. No. 5, edited by D.R.Stoddart & R.G.Johannes, UNESCO, pp. 149–161 . Kennelly, S.J. & Underwood, A.J., 1984. J. exp. mar. Biol. Ecol., 76, 67–78. Kennelly, S.J. & Underwood, A.J., 1985. J. exp. mar. Biol. Ecol., 89, 55–67. Kershaw, K.A., 1957. Ecology, 38, 291–299. Kershaw, K.A., 1970. Ecology, 51, 729–734.

SAMPLING AND SPATIAL PATTERN

67

Kimmel, J.J., 1985. Environ. Biol. Fish., 12, 23–32. Kinzie, R.A. & Snider, R.H., 1978. In, Coral Reefs: Research Methods, Monogr. Oceanogr. Methodol. No. 5, edited by D.R.Stoddart & R.G.Johannes, UNESCO, pp. 231–250. Kirchman, D., Sigda, J., Kapuscinski, R. & Mitchell, R., 1982. Appl. Env. Microbiol., 44, 376–382. Kosler, A., 1968. Mar. Biol., 1, 266–268. Lafedes, R.N., 1978. Editor, McGraw-Hill Dictionary of Scientific and Technical Terms. McGraw-Hill, New York, 1829 pp. Leatherwood, S., Gilbert, J.R. & Chapman, D.G., 1978. J. Wildl. Mgmt, 42, 239– 250. Lee, W.Y. & McAlice, B.J., 1979. Estuar. cstl mar. Sci., 8, 565–582. Lewis, J.R., 1976. Oceanogr. Mar. Biol. Ann. Rev., 14, 371–390. Lincoln, R.J., Boxhall, G.A. & Clark, P.F., 1982. A Dictionary of Ecology, Evolution and Systematics. Cambridge University Press, Cambridge, 298 pp. Loya, Y., 1978. In, Coral Reefs: Research Methods, Monogr. Oceanogr. Methodol. No. 5, edited by D.R.Stoddart & R.G.Johannes, UNESCO, pp. 197–217. Ludwig, J.A. & Goodall, D., 1978. Vegetatio, 38, 49–59. Lussenhop, J., 1974. J. Hist. Biol., 7, 319–337. Mackay, A.P., Cooling, D.A. & Berrie, A.D., 1984. J. appl. Ecol., 21, 515–534. McIntyre, A.D., 1971. In, Proc. 1st int. Conf. Meiofauna, Smithson. Contr. Zool., No. 76, edited by N.C.Huling, Smithsonian Institute Press, Washington, pp. 149–154. Mead, R., 1974. Biometrics, 30, 295–307. Millard, S.P. & Lettenmaier, D.P., 1986. Estuar. cstl shelf Sci., 22, 637–656. Millard, S.P., Yearsley, J.R. & Lettenmaier, D.P., 1985. Can. J. Fish. Aquat. Sci., 42, 1391–1400. Montagna, P.A., 1982. Appl. Environ. Microbiol., 43, 1366–1372. Morin, A., 1985. Can. J. Fish. Aquat. Sci., 42, 1530–1534. Morisita, M., 1959. Mem. Fac. Sci., Kyushu Univ., Ser. E (Biol.), 2, 215–235. Morisita, M., 1962. Res. Popul. Ecol. Kyoto Univ., 4, 1–7. Morisita, M., 1971. Res. Popul. Ecol. Kyoto Univ., 13, 1–27. Mountford, M.D., 1961. J. Ecol., 49, 271–275. Nie, H.W. de & Vijverberg, J., 1985. Hydrobiologia, 124, 3–11. Omori, M. & Hamner, W.M., 1982. Mar. Biol., 72, 193–200. Oviatt, C.A. & Nixon, S.W., 1973. Estuar. cstl mar. Sci., 1, 361–378. Payandeh, B., 1970. For. Sci., 16, 312–317. Pechanec, J.F. & Stewart, G., 1940. J. Am. Soc. Agron., 32, 669–682. Perry, J.N. & Mead, R., 1979. Biometrics, 35, 613–622. Pielou, E.C., 1959. J. Ecol., 47, 607–613. Pielou, E.C., 1969. An Introduction to Mathematical Ecology. Wiley, New York, 286 pp. Pielou, E.C., 1979. Oecologia (Berl.), 44, 143–144. Pielou, E.C. & Routledge, R.D., 1976. Oecologia (Berl.), 24, 311–321. Pihl, L. & Rosenberg, R., 1982. J. exp. mar. Biol. Ecol., 57, 273–301. Platt, T. & Denman, K.L., 1975. Ann. Rev. Ecol. Syst., 6, 189–210. Pot, W., Noakes, D.L.G., Ferguson, M.M. & Coker, G., 1984. Hydrobiologia, 114, 249–254. Pringle, J.D., 1984. Can. J. Fish. Aquat. Sci., 41, 1485–1489. Resh, V.H., 1979. J. Fish. Res. Bd Can., 36, 290–311. Ripley, B.D., 1978. J. Ecol., 66, 965–981. Rohlf, F.J. & Archie, J.W., 1978. Ecology, 59, 126–132. Rosenberg, R., 1974. J. exp. mar. Biol. Ecol., 15, 69–80. Rotenberry, J.T. & Wiens, J.A., 1985. Am. Nat., 125, 164–168. Russell, B.C., Talbot, F.H., Anderson, G.R.V. & Goldman, B., 1978. In, Coral Reefs: Research Methods, Monogr. Oceanogr. Methodol., No. 5, edited by D.R.Stoddart & R.G.Johannes, UNESCO, pp. 329–345.

68

N.L.ANDREW AND B.D.MAPSTONE

Saila, S.B., Pikanowski, R.A. & Vaughan, D.S., 1976. Estuar. cstl mar. Sci., 4, 119– 128. Sale, P.F., 1980. Oceanogr. Mar. Biol. Ann. Rev., 18, 367–421. Sale, P.F. & Douglas, W.A., 1981. Environ. Biol. Fish., 6, 333–339. Sale, P.F. & Sharp, B.J., 1983. Coral Reefs, 2, 37–42. Scheaffer, R.L., Mendenhall, W. & Ott, L., 1979. Elementary Survey Sampling. Wadsworth Publishing Company, Belmont, California, 2nd edition, 278 pp. Scheffe, H., 1959. The Analysis of Variance. Wiley, New York, 477 pp. Scheibling, R.E., 1980. Mar. Ecol. Prog. Ser., 2, 321–327. Schweigert, J.F. & Sibert, J.R., 1983. Can. J. Fish. Aquat. Sci., 40, 588–597. Schweigert, J.F. & Sibert, J.R., 1984. Can. J. Fish. Aquat. Sci., 41, 827–828. Schweigert, J.F., Haegele, C.W. & Stocker, M., 1985. Can. J. Fish. Aquat. Sci., 42, 1806–1814. Seapy, R.R. & Kitting, C.L., 1978. Mar. Biol., 46, 137–145. Seber, G.A.F., 1982. The Estimation of Animal Abundance. Griffin, London, 654 pp. Simberloff, D., 1979. Ecology, 60, 679–685. Sinclair, D.F., 1985. Ecology, 66, 1084–1085. Skellam, J.G., 1952. Biometrika, 39, 346–362. Smith, H.F., 1938. J. agric. Sci., 28, 1–23. Smith, S.J., 1984. Can. J. Fish. Aquat. Sci., 41, 826–827. Snedecor, G.W. & Cochran, W.G., 1980. Statistical Methods. The Iowa State University Press, Iowa, 7th edition, 507 pp. Sokal, R.R. & Rohlf, F.J., 1981. Biometry. Freeman Co., New York, 2nd edition, 859 pp. Southward, A.J. & Bary, B.M., 1980. J. mar. biol. Ass. U.K., 60, 295–311. Southwood, T.R.E., 1966. Ecological Methods. Chapman & Hall, London, 391 pp. Stimson, J., 1974. Ecology, 55, 445–449. Stoddart, D.R., 1969. Biol. Rev., 44, 433–498. Stretch, J.J., 1985. J. exp. mar. Biol. Ecol., 91, 125–136. Sweatman, H.P.A., 1985. Ecol. Monogr., 55, 469–485. Sykes, J.B., 1982. Editor. The Shorter Oxford English Dictionary. Oxford University Press, Oxford, 7th edition, 1368 pp. Taylor, L.R., 1961. Nature, Land., 189, 732–735. Taylor, L.R., 1971. In, International Symposium on Statistical Ecology. Vol. 1. Spatial Patterns and Statistical Distributions, edited by G.P.Patil et al., Pennsylvania State University Press , Pennsylvania, pp. 357–377. Taylor, L.R., 1980. Can. J. Fish. Aquat. Sci., 37, 1330–1332. Taylor, L.R., Woiwod, I.P. & Perry, J.N., 1978. J. anim. Ecol., 47, 383–406. Taylor, L.R., Woiwod, I.P. & Perry, J.N., 1979. J. anim. Ecol., 48, 289–304. Taylor, W.D., 1980. Can. J. Fish. Aquat. Sci., 37, 1328–1329. Thompson, H.R., 1958. Aust. J. Bot., 6, 322–342. Todd, C.D., 1978. J. anim. Ecol., 47, 189–203. Toft, C.A. & Shea, P.J., 1983. Am. Nat., 122, 618–625. Underwood, A.J., 1976. Oecologia (Berl.), 26, 257–266. Underwood, A.J., 1977. J. exp. mar. Biol. Ecol., 26, 191–201. Underwood, A.J., 1978. Oecologia (Berl.), 36, 317–326. Underwood, A.J., 1981. Oceanogr. Mar. Biol. Ann. Rev., 19, 513–605. Underwood, A.J. & Denley, E.J., 1984. In, Ecological Communities: Conceptual Issues and the Evidence, edited by D.R.Strong et al., Princeton University Press, Princeton, pp. 151–180. Upton, G.J.G., 1984. Biometrics, 40, 760–766. Usher, M.B., 1975. J. Ecol., 63, 569–586. Van Vleet, E.S. & Williams, P.M., 1980. Limnol. Oceanogr., 25, 764–770. Venrick, E.L., 1971. Limnol. Oceanogr., 16, 811–818.

SAMPLING AND SPATIAL PATTERN

69

Watling, L., Kinner, P.C. & Maurer, D., 1978. J. exp. mar. Biol. Ecol., 35, 109–118. Weinberg, S., 1981. Bijdragen tot de Dierkunde, 51, 199–218. Wiebe, P.H., 1971. Limnol. Oceanogr., 16, 29–38. Wiebe, P.H., 1972. J. Cons. perm. int. Explor. Mer, 34, 268–275. Wiebe, P.H. & Holland, W.R., 1968. Limnol. Oceanogr., 13, 315–321. Wiens, J.A., 1986. In, Community Ecology, edited by J.Diamond & T.J.Case, Harper & Row, New York, pp. 154–172. Wiens, J.A., Addicott, J.F., Case, T.J. & Diamond, J., 1986. In, Community Ecology, edited by J.Diamond & T.J.Case, Harper & Row, New York, pp. 145–153. Williams, B., 1978. A Sampler on Sampling. John Wiley & Sons, New York, 254 pp. Wilson, J.G., 1976. Mar. Biol., 37, 371–376. Winer, B.J., 1971. Statistical Principles in Experimental Design. McGraw-Hill Kogakusha, Tokyo, 2nd edition, 907 pp. Yamane, T., 1967. Statistics: an Introductory Analysis. Harper & Row, New York and John Weatherhill Inc., Tokyo, 919 pp. Youngbluth, M.J., 1982. J. exp. mar. Biol. Ecol, 61, 111–124.

Oceanogr. Mar. Biol. Ann. Rev., 1987, 25, 91–112 Margaret Barnes, Ed. Aberdeen University Press

FLUMES: THEORETICAL AND EXPERIMENTAL CONSIDERATIONS FOR SIMULATION OF BENTHIC ENVIRONMENTS* ARTHUR R.M.NOWELL and PETER A.JUMARS School of Oceanography, University of Washington, Seattle, WA 98195, U.S.A.

INTRODUCTION Flumes have been used to transport water since Roman times. The word flume is derived from Latin “fluere… to flow” and today means “an inclined channel for conveying water from a distance”. Flumes have been extensively used in sedimentary geology and civil engineering for approximately a century to examine the modes and rates of sediment transport. Their use in benthic biology resulted from the recognition that a variety of processes of biological interest were strongly influenced by water motions at the boundary with solid surfaces and that these processes in the bottom boundary layer can be modelled in flumes. The benthic boundary layer in the transport of scalar quantities such as nutrients plays an important rôle not only to benthic biologists and geochemists, but also to sedimentary geologists and boundary-layer fluid mechanicians. Our primary objective will be to review the characteristics of such flows and what constraints must then be placed on the design and operation of laboratory models of benthic boundary layers irrespective of the discipline to which the flume experiments are applied. As the application of flume techniques to biological problems in particular is not straightforward, our goal in this review is to present the fluid dynamic issues and compromises that govern the applicability and operation of laboratory flumes. Previous articles on flumes have focused on engineering considerations and construction techniques (Williams, 1971), on the differing types of flumes that are routinely used in sedimentary geology (Middleton & Southard, 1984), and on descriptions of specific flumes (for example, Vogel & LaBarbara, 1978; Nowell, Jumars & Eckman, 1981). Use of open-channel flumes in sediment transport studies ranges from early work by Osborne Reynolds in the 1880s, to the classical work by Gilbert (1914) and Shields (1935) to more recent studies by Guy et al. (1966), Grass (1971), and Sumer & Deigaard (1981). All these studies were concerned with the complex case of boundary-layer flow over a deformable (soft-sediment) bed, and the response of the bed to fluid forcing. All used fresh water and paid no attention to biota. Apart from the effort involved in filling the flumes, there were few limits on the sizes of the devices. Thus flumes used by these authors ranged up to 80 m in length. Once one incorporates a concern for the biota, especially for either controlling microbial abundance or for implanting a community of macrofauna and meiofauna, a severe pragmatic limitation is imposed on the size of the device. We take it as given that the purpose of a flume is not only to simulate realistic field conditions near the sea bottom, but also to simplify them so that the flow characteristics can be summarized in a small number

*University of Washington Contribution No. 1649.

FLUMES: SIMULATION OF BENTHIC ENVIRONMENTS

71

of parameters (i.e. those presented by Nowell & Jumars, 1984). Only then can the results in terms of effects of changes in those parameters be both replicated and generalized. There is no benthic flow environment simple enough to be characterized by a single variable, such as flow velocity (volumetric discharge divided by cross-sectional area or streamwise velocity at a given depth) along a flume or in nature, so there is no point in reviewing biological observations made where this was the only variable measured or used in matching field and laboratory flow regimes. The epitome of irreproducibility in this context is the taking of small-scale physical (e.g. Statzner & Holm, 1982) or chemical (e.g. Jørgensen & Revsbech, 1985) measurements about organisms without measuring or controlling the larger-scale fluid motions that determine these small-scale patterns. Nowell & Jumars (1984) review the important parameters, but do not show how to control them in a laboratory setting. Rules of thumb in flume design have been given previously in the marine ecological literature (Jumars & Nowell, 1984; Muschenheim, Grant & Mills, 1986) but their derivations have not. Our primary aim in this review is to expose the principles underlying the rules of thumb utilized in flume design, and to do so highly selectively. The problem of flume design can be broken into four parts, i.e. producing specific entrance conditions, providing certain exit conditions, driving the flow (including, in some cases, return of fluid with or without its particle load to the flume entrance), and lastly but most important tailoring the flow in the test section. Flows in the entrance, exit and drive sections are complex, defying simple and general parameterization, but conditions there are for the most part irrelevant to biological measurements of interest. Furthermore, the miniaturization often required of flumes for biological purposes usually makes no special demands here; the rules of thumb are reasonably independent of physical scale. For these reasons, we give brief treatment of these design considerations which are often satisfied with simple, empirically derived guidelines. Control of fluid dynamics in the test section is another matter. Here is where the serious and inevitable compromises of flume scale must be made. It is impossible to make them intelligently without understanding the underlying principles. We review those principles (of conservation of mass and momentum) by recourse to the basic Navier-Stokes equations, to which most oceanographers have at one or several times been exposed. Very readable introductions to such material, which focus on phenomenological understanding rather than mathematics, are presented by Bradshaw (1971), Francis (1975), van Dyke (1982), and Allen (1985). Finally, we exercise those principles by applying them to specific biological problems, including ones where it becomes obvious that laboratory flume work is impossible or impractical. We limit our scope explicitly to flumes for simulation of bottom boundary layers. Specifically we are excluding from consideration flow devices where the flow goes into solid-body rotation, such models generally being used to simulate geostrophic flows. We also exclude enclosed water tunnels for, although such devices are useful for measuring drag on bodies (e.g. fish fins or model submarine hulls) suspended far from a boundary, they are not useful without extensive corrections in simulating bottom boundary layers or in making drag measurements on objects attached to the bed or sidewalls. Thus, our major emphasis will be on flumes which permit us to study small-scale, viscous, and turbulent flows, most often flows with a single dominant velocity gradient. We consider in this review flows with a free surface, and in which stratification and rotation play negligible rôles. While the oceans are clearly stratified, rotating shear flows, the flow close to a body, and flows on the scales of importance to a single organism, or a local segment of a community, may be viewed as dominated by the relative balance of inertial, viscous, and body forces alone. As Shakespeare points out about life, and as more emphatically noted in “Trobriand Cricket”, by our entrances and exits shall we be remembered. The entrance conditions as well as the exit conditions can affect the nature of flow in the test section of a flume. What specific characteristics of fluid motion we need are specified by the question that we are trying to answer. How well we can answer the question will be

72

ARTHUR R.M.NOWELL AND PETER A.JUMARS

determined by how well we can tailor the boundary-layer flow in our flume to match the field situation. Our ability to mimic a field situation in the laboratory depends on selection of an appropriate flume design. ENTRANCE CONDITIONS The entrance to the flume is crucial for the development of simple flows. The problem can be viewed as getting the flow to ‘forget’ its recent history. Because the water usually is delivered by a pipeline which most likely has bends in it, the flow entering the flume often will have strong circulation from going round a curve. In addition the flow may have to diverge as it enters the channel. As an example of these problems consider the flow entering a flume from a 20 cm (diam.) pipe having just turned through 90° and diverging rapidly into a channel that is 30 cm wide. Clearly the flow from the pipe will form a centre jet which will take many pipe diameters to decay (based on many observations downstream of obstacles the flow will ‘remember’ for about 20 times the critical length scale which in this case would be pipe diameter). We could, therefore, not expect to develop a simple boundary layer for over 4 m. A smooth, gradual divergence at the entrance is required to avoid separation of the flow. Experience from the design of ship hulls shows that expansion angles exceeding 7° often show flow separation (Chang, 1970), so this angle should not be exceeded without detailed examination of the consequences. Rapid expansions (e.g. O’Brien, Tay & Zwart, 1986) lead to free jets and complex flows that are virtually impossible to describe in detail, again diminishing the value of bringing the problem into the laboratory. One common solution to the problem of removing the effects of the pipe flow and upstream delivery of the fluid is to force it through a diffuser aimed at the far upstream wall of the flume. The motion is broken into small scales and the flow distributed across the full width of the flume. Honeycomb grids downstream of the diffuser will break up remaining scales of motion that are slightly bigger than the grid scale, but such grids cannot markedly affect flow non-uniformity on the scale of the whole channel diameter. The use of grids for generating small-scale turbulence has a long and well-recorded history (Laws & Livesy, 1978) and it has become almost a shibboleth of flume building to include a screen at the entrance. Laws & Livesy (1978) detail the calculations required to estimate the effects of screens on the flow, but we note that in essence screens only help to tailor an already well-cut flow; they cannot generate smooth boundary layer velocity profiles out of non-uniform jets. Moreover in free-surface flows such as flumes, the more dense the screen, the bigger the pressure drop across it; consequently, a very dense or a very long section of mesh will result in a downstream hydraulic jump that propagates surface waves downstream, causing non-uniformity in the boundary layer. As a general rule the tubes comprising the mesh should be about 20 times longer than the mesh diameter in order to apply a sufficiently strong strain to the fluid to alter the incoming turbulence. Elegant mesh designs have been developed for wind tunnels in order to pre-shape the flow into a boundarylayer profile, but such designs require considerable care and many hours of computing. Furthermore, they are well suited to but a single flow rate. If a fully rough-turbulent boundary layer is desired, its development can be accelerated or ‘tripped’ by placing small-scale roughness elements across the flume bed on the downstream side of the honeycomb grid. EXIT CONDITIONS The exit problem can be viewed as not letting the flow know what is coming, i.e. of making a smooth exit without breaking cadence. At the exit of the test section the flow can either fall freely, be retained by a weir, or be directed through tail gates. In the case of a weir, especially if the flow is supercritical (with respect to Froude number defined below), upstream effects are clearly visible. The flow is suddenly deepened having

FLUMES: SIMULATION OF BENTHIC ENVIRONMENTS

73

to go over the gate, and as a consequence a hydraulic jump may occur just upstream of the gate, or a backwater slope is developed (Henderson, 1966). But upstream effects also exist in subcritical situations and the weir causes the flow to diverge from the bed upstream of the gate. This effect clearly imposes extra length on our flume as it reduces the length of the useful test section. Similarly, a free overfall has an upstream effect at subcritical Froude numbers as the flow close to the bed is accelerated on approaching the overfall. As a compromise to the two extremes of a weir and a free-overfall, louvred gates, much like venetian blinds (set either vertically or horizontally) provide a minimum of upstream interference. A good picture of such gates is provided by Yalin (1977). DRIVING THE FLOW Flumes are used because we wish to control the mass or momentum flux or the boundary stress. Such control is achieved by manipulating discharge (volume per cross-sectional area per time) through the flume. Discharge is regulated either by using a pump and returning the fluid to the entrance, or by using a constanthead tank. For the types of velocities that are most commonly needed in oceanographic studies, a constanthead tank provides an inexpensive and effective method of providing a regulated, constant discharge. Pumps are often necessary for high discharges but require considerable attention to avoid problems of unsteadiness at low velocities (for example picking up the blade frequency from the impellor) and varying discharge at high rates when small voltage changes may cause large changes in effective discharge rate. Most flows of concern have boundary layer Reynolds numbers (free stream velocity times boundary layer thickness divided by kinematic viscosity) greater than 105, and because the fluid is well mixed through the return pipes and pumps we rarely have to worry about fluid stratification. At low-flow Reynolds numbers, however, fluid convection can cause numerous problems in the laboratory. Low Reynolds numbers of settling particles, such as larvae, provide a challenging design problem. Obtaining a totally still body of water is more complex than usually imagined and many settling columns are plagued with problems of convection. Although the velocities of convection are low, so are the velocities of larval settling. Temperature changes can also occur in flumes due to heating of the water as it passes through the pump; because of the high thermal capacity of water this only becomes a problem in very small flumes when short recirculation times may occur. Temperature variations should always be monitored for small changes in temperature can often markedly affect velocity sensors using heat transfer rate to infer velocity. TEST SECTIONS At last the stage is set. Well characterized, reproducible flow in the test section is the goal for the design of any flume. The test section should not be considered as the distance left over after we have moved far enough away from the entrance to avoid non-uniformity or advective effects and far enough upstream to miss the results of the flow leaving the flume. Rather the test section should be long enough to allow simulation of the flow field of interest. Unfortunately, unless one has carried out the calculations as to one’s needs before one either builds the flume or tries to use an already existing device, one may end up with no useful test section at all. MASS AND MOMENTUM BALANCE To obtain a valid simulation of any flow there is a hierarchy of similarity that needs to be considered. At the simplest level we wish to maintain geometric similarity; that is, the shape of the boundary (the bed in most

74

ARTHUR R.M.NOWELL AND PETER A.JUMARS

cases) should be the same from field to laboratory. Because the perimeter of a body in a flow where the noslip condition (zero velocity immediately at the body’s surface) applies is also a streamline, we will maintain kinematic similarity if we keep the velocity the same between laboratory and field. We usually wish, however, not only to maintain geometric and kinematic similarity but also to model the fluxes of material and momentum, in other words we wish to maintain similarity of forces, or dynamic similarity. To do so reliably we shall treat the equations that express the balance of forces within the flow; we shall use Newton’s second law. The governing equations are the continuity equation for an incompressible fluid (1) and the conservation of momentum equation (2) which may be written out in its scalar form as

The velocity is represented by ui, which has three components u, v and w, t is time, p pressure, g gravity, is density and and v are the dynamic and kinematic viscosity (and ) The terms on the left of the Navier-Stokes Equation 2 represent, respectively, the local and advective rate of change of momentum. They are balanced on the right hand side of the equation, respectively, by the pressure gradient, the viscous forces and the body force (which we take to be gravity). Equation 2 is just the familiar force balance F=ma written for a fluid parcel. Another way to conceptualize each term is as a force (MLT−2) per unit volume (L −3) of the water parcel, again yielding the proper units for each term (ML−2T−2). General solutions of these equations are not available, and simplification of them is achieved only if we can apply certain restrictions. To understand the physics represented by these equations we want to transform each of the terms into a nondimensional form. Let us define

FLUMES: SIMULATION OF BENTHIC ENVIRONMENTS

75

where L is an independent length scale, V is an independent velocity scale and is an independent time for dimensional uniformity. We substitute these terms into the scale and we divide the pressure by Equation 2:

and multiplying through by L/V2

Each of the terms is now dimensionless, whereas in Equation 2 it had the dimension of mass/unit volume through multiplying the various times acceleration. There are four dimensionless groups, labelled terms. Term one, a frequency parameter, is called the Strouhal number, and is used as a measure of the unsteadiness of the flow. Most commonly it is used to relate periodic vortex shedding from cylinders to the velocity incident upon and diameter of the cylinder (Tritton, 1977), but its general interpretation is as a measure of unsteadiness. When the Strouhal number is zero the flow is steady and values rarely exceed unity for even the most unsteady cases. For example, in a tidal flow in an estuary where the tidal frequency seconds, the maximum velocity of order 50 cm·s-1 and the flow depth say 10 m, the Strouhal is thus modelling such a boundary-layer flow in the laboratory could be done using a number is quasi-steady flow especially if we were interested in the response of the bed, which is nearly instantaneous to changes in stress. As another example, consider shallow-water wind waves in the intertidal at a depth of 1 m where the mean flow velocity may be of order 20 cm·s−1, and wind waves may have a period of 5 s. Such parameters will result in a Strouhal number of 1·0. Clearly it would be unwise to attempt to model in the laboratory diffusional processes in such regions without including the unsteady term. By keeping the Strouhal number approximately the same between laboratory and field we ensure that we are correctly reproducing the kinematics of interest from the field. Term two is the ratio of gravitational to inertial forces, and is the inverse of the Froude number. It is important whenever we must consider free-surface effects, i.e. when the boundary layer extends through the complete flow depth, or when surface waves enter the problem. The Froude number, which as Lu (1977) notes is named after William Froude who ‘retired’ at age 36 to do research on ship rolling and resistance, also represents the ratio of the mean flow velocity to the phase speed of a shallow-water wave. As such it distinguishes between two regimes. Subcritical or tranquil flow has a Froude number less than unity, and the flow is deep and slow; information and boundary layer effects from downstream may be transmitted upstream by surface waves (viz. Lu, 1977). When the Froude number is greater than one, supercritical or shooting flow exists and information is transmitted only downstream. The Froude number is thus crucial when we attempt to model flows in estuarine and intertidal areas. Flows over most tidal flats are subcritical most of the time, although occasionally in run-off channels the flow is critical or supercritical, and standing waves may be observed at the transition.

76

ARTHUR R.M.NOWELL AND PETER A.JUMARS

The third term is the familiar Reynolds number (actually its inverse) which is the ratio of inertial to viscous forces. The ratio of terms two and four in Equation 2 namely the inertial to viscous terms may be written as

which dimensionally is

We see that changing the Reynolds number then represents changing the balance of force from two terms in the Navier Stokes equation. In order to model flows dynamically we must maintain similarity of forces; hence all the dimensionless groups must be kept numerically similar from field to laboratory. The fourth dimensionless group, called the Euler number, is a measure of the pressure gradient, and is of consequence mainly in flows where gravity plays only a small rôle and the flow is being driven by a pressure drop, such as in flow through a horizontal pipe. We thus have to consider three dimensionless groups when we attempt to model flow in the laboratory. If we consider only steady flow (no changes with mean flow parameters with time, however), then we can dispense with consideration of the Strouhal number. Flumes are traditionally very long relative to their widths, and customarily are operated with flow depths which are shallow relative to their widths. The reasons for this design can be seen if we take the Navier Stokes equations and simplify them. Let us consider a steady flow with no gradients in the cross-stream direction, much like the flow in a wide, straight river or the flow over a wide sand flat. With no gradients in the y-direction the steady-flow equations reduce to conservation equations for mass (3) and, dividing through by and ignoring gravity, for momentum (4) (3)

(4)

If we consider a flow over a flat bed and apply the no-slip condition at the wall (at z=0, U=0) then the only geometric length scale is the distance x from the leading edge. Observation of such flows shows that the is very small, so that If we let the x component of velocity be thickness of the boundary layer of order u, and d/dx be of order 1/x then du/dx is of order u/x and from the continuity equation dw/dz must be of the same order. Now as w is smaller than u and since d/dz is larger than d/dx then the order of dw/dz may

FLUMES: SIMULATION OF BENTHIC ENVIRONMENTS

be met by considering w to be of order order

and d/dz of order

77

Then using the symbol 0( ) to mean of

we consider the order of the terms in the steady two-dimensional equations as (5)

(6) Equations 5 and 6 are termed the thin shear layer equations (Cebeci & Bradshaw, 1973) and can be solved to give the growth rates of boundary layers. They can be integrated over the boundary layer thickness to give the flux of momentum to the bed, that is the momentum integral yields the flux of momentum to the bed, or the stress acting on the bed. Now the two acceleration terms are of the same order in each equation, but the first viscous term in Equation 5 is very much smaller than the second viscous term so can be neglected. Since fluid parcels may be accelerated in boundary layers, and since strong viscous effects exist (having imposed the no-slip condition at the wall) the dominant viscous term is assumed to be of the same order of magnitude as the inertial terms, i.e.

This relation can be rearranged so that we see

Thus from a purely order-of-magnitude analysis we deduce that a viscous boundary layer thickness increases All the terms in Equation 6 are smaller than in Equation 5 and so to order may be ignored. as The shear-layer equations then become (7)

(8) The exact growth rate of a boundary layer can be predicted by using Equations 7 and 8 and solving for any imposed velocity profile, as done in many textbooks (cf. Shames, 1967) exactly for laminar flow and in an

78

ARTHUR R.M.NOWELL AND PETER A.JUMARS

approximate manner for turbulent flows. The resulting growth laws show that boundary layers are very thin relative to their lengths of development, and thus we need flumes that are long in order to develop boundary layers of reasonable thickness. Table I gives some typical thickness for viscous and turbulent boundary layers for commonly used velocities. Because our flume must have sides, the numerical artifice of saying that we have purely two-dimensional flow is not empirically obtainable. Boundary layers grow just as quickly on the sides of the flume as on the bed, as long as the roughness is the same. Because in the field the assumption of two-dimensional flow may often be fulfilled, we must attempt to make our flumes as wide as feasible. Increasing the width of the flume puts immediate demands on the pumping capacity, because the boundary layer must be fairly thick in TABLE I Upper part gives boundary layer thickness (in cm) for various lengths of channel, lower part gives corresponding boundary layer Reynolds numbers : the viscous cases, in which Re <3,000, was computed using the exact Blasius solution to Equations (7) and (8), namely for the turbulent cases, the approximation was used; note that in some cases the boundary layer goes turbulent at a considerable distance down channel; also note the caveat on using Re=3,000, mentioned in the text; *calculated using turbulent flow formula Free-stream velocity cm s−1

Downstream distances (x)

50

100

200

500

1000

1 5 10 20 50 1 5 10 20 50

3·3 1·5 1·1 0·73 1·6* 328 734 1038 1467 7826

4·6 2·1 1·5 3·3* 2·7* 464 1037 1467 6546 13625

6·6 2·9 2·1 5·7* 4·7* 656 1450 2075 11398 23724

10·4 4·6 3·3 11·9* 9·9* 1037 2320 3281 23723 49379

14·7 6·6 23·7* 20·6* 17·2* 1467 3281 23724 41306 85973

order to accommodate, for example, macrofauna. How narrow can the flume be? Clearly, if the flume were twice as wide as deep, we would have flow at the midpoint of the flume width that was equidistant from each solid surface, which certainly is not a two-dimensional approximation. None the less, many if not most flow devices are less than twice as wide as deep (e.g. McIntire, Garrison, Phinney & Warren, 1964) and the measurements and corrections required to back-calculate bottom effects without sidewalls are considerable (e.g. Grant, Boyer & Sanford, 1982). In addition, because we only have three solid surfaces and one liquid-gas interface we shall develop asymmetries in the resistance to flow, resulting in secondary (non-downstream) circulation across the channel. To simplify the situation as much as possible we want the flume wide so that we are very much closer to the bed than to the sides in the measurement region of interest. A minimum value for the width-to-depth ratio then is 5, to reduce in the channel centre the effects of secondary circulation and the boundary layers from the sidewalls to less than 5% of the mean velocity. It is wise to try to get a ratio of 10:1, but designing and operating flumes is full of compromises between conflicting dimensionless groups and pragmatic limitations such as pump capacity. As an example of the need to compromise we usually want the Reynolds number to be high, and because we have only a finite depth of water in our flume we usually increase the velocity. But as we increase the

FLUMES: SIMULATION OF BENTHIC ENVIRONMENTS

79

Reynolds number we are also increasing the Froude number, and we usually have to compromise to a lower Reynolds number in order to keep the Froude number in the correct range (i.e. in the region of subcritical flow). So far in dealing with the equations we have said (except for the Strouhal number) nothing about boundary conditions or whether the flow is viscous or turbulent. While there is no single Reynolds number which delineates the two flow regimes, it is easy to observe when the flow is turbulent. Boundary layers over natural sediment surfaces are virtually always turbulent: as a rough guideline, good to a factor of 5, turbulence occurs when the mean flow Reynolds number exceeds 3000. This number can only be a guideline because it is chiefly the boundary condition, namely the roughness which determines the onset of turbulence for any imposed mean velocity. To see where turbulence is generated we take the Navier-Stokes equations, perform Reynolds averaging on them and then identify the extra term in the equation. Formally, we say that the instantaneous velocity is made up of a mean and a fluctuating component where the mean of the fluctuations is zero,

The overbar indicates mean quantities and the prime indicates fluctuating quantities. When we introduce these terms into the Navier-Stokes equations and average (cf. Hinze, 1975 or Tennekes & Lumley, 1972) we obtain

Comparing this equation with Equation 2 there is one extra term, which may be identified as a turbulent advection term, that is, it is identical to the mean advection term next to it in the equation but applies only to the fluctuating velocities. By using Equation 1 and the rule of differentiation of a product we may, however, rewrite the equation as

The new term, now on the right hand side, is a stress, and is called the Reynolds stress. It is diffusing momentum within the flow and is apparent to us by rapid mixing and nearly random motion. By analogy to the viscous stress, we write the term in the form of a viscosity multiplied by a mean velocity gradient. The concept of the eddy viscosity coefficient (which has the same dimensions as the coefficient for viscous momentum diffusion, i.e. L2T−1) has much appeal for boundary layers because it is simple to compute, intuitively easy to understand and above all works remarkably well. Thus, turbulent flows have more stresses, resulting in greater resistance to flow, and to higher mixing. This higher mixing means that we mix higher-momentum fluid closer to the bed, so that boundary layers in turbulent flows exhibit stronger velocity gradients than those in laminar flows. In order to use the eddy viscosity concept we need to specify a length (L) times a velocity scale (LT−1). The length scale for the size of the eddies responsible for mixing is regarded as the same size as distance to the bed (z), that is the largest vertical scale of eddy that could be accommodated, so

80

ARTHUR R.M.NOWELL AND PETER A.JUMARS

where is von Karman’s constant (0·41) and the velocity scale is taken as the square root of the boundary stress, divided by density, that is

so our eddy viscosity in the lower 30% of the boundary layer is given as

A more general formulation valid throughout the boundary layer is given by Businger & Arya (1974) and Long (1981). To examine the flow close to the bed we can use the parameters that we have already defined to characterize the velocity. The velocity gradient close to the bed must depend on the turbulent velocity scale, Thus, we write the gradient as represented dimensionally by u*, and the mixing length

and if we integrate this equation we obtain

This relationship is called the law of the wall and describes the velocity profile in the region close to the wall where the momentum is transferred by the turbulent stresses. Because the left-hand side is dimensionless, the right-hand side must also be dimensionless, and thus z0 is a length scale, termed the roughness length. The law of the wall clearly applies only when we are far enough away from the bed that viscosity is not the main mechanism responsible for transferring momentum. The scales of the viscous sublayer, when the viscous stresses dominate and where the velocity profile has the form

have been given in Jumars & Nowell (1984). A full description of the velocity profiles is given in Dyer (1986). In order of distance away from the wall there is an inner region, usually millimetres thick called the viscous sublayer, a region above this called the logarithic layer which extends to approximately 30% of the boundary layer thickness, and an outer flow region termed the velocity defect region (Coles, 1956). Within the viscous sublayer, there is an even thinner region called the diffusive sublayer. While the viscous sublayer is the region where the flux of momentum is determined by the velocity gradient times the viscosity, the diffusive sublayer is the region where a scalar quantity such as oxygen is diffused by molecular diffusion. As the molecular diffusion coefficient for any chemical species is smaller than the molecular diffusion coefficient for momentum (which is the dynamic viscosity), the diffusive sublayer will be thinner than the viscous sublayer. While momentum can be transferred from molecule to molecule,

FLUMES: SIMULATION OF BENTHIC ENVIRONMENTS

81

chemical properties are transferred only via the migration of chemical species, making chemical transfer inherently less efficient than momentum transfer. OTHER NECESSARY SCALINGS There are three other dimensionless groups that must be considered if we are to model flow over a deformable bed, i.e. where the sediment on the bed can be moved either as bedload or as suspended load. Clearly the roughness of the bed will affect the local transfer of momentum to the bed, and the scale of the particles, i.e. the sediment grain size (ks), can be considered one such appropriate scale. The shear velocity, which is the square root of the boundary shear stress (the force acting on the bed), gives us the appropriate velocity scale, and we can create a roughness Reynolds number,

which is a measure of the bed and its interaction with the flow. When the roughness Reynolds number is less than about 3 the particles on the bed are subsumed within a viscous sublayer, and the flow is termed hydraulically smooth or smooth-turbulent. When the roughness Reynolds number is greater than about 70 the flow is called hydraulically rough or rough-turbulent and there is no continuous viscous sublayer. Under such circumstances the diffusion of nutrients or chemical species will be determined entirely by the turbulent eddy viscosity and will occur many times faster than when there is a continuous viscous sublayer. As with the Froude number there are classes of flow; at a minimum it is necessary to ensure our laboratory roughness Reynolds number falls in the same class as that from the field. Once we are concerned with moving material we need to consider the dimensionless force required to move the particles, namely

and the ratio of the settling velocity (ws) of the particle to the boundary shear velocity (u*). This ratio is termed the Rouse number; small values (1/20) indicate rapid diffusion of the particles throughout the boundary layer, whereas large values (around unity) indicate strong concentration gradients of larger material close to the bed. The significance of these terms has been outlined by Jumars & Nowell (1984) and is discussed in detail by Southard, Boguchwal & Romia (1980). There are two further complications that must be considered in modelling the natural marine environment, namely non-uniform roughness and wave effects. Rarely in the field will the bed be composed of a semi-infinite flat bed of uniform roughness; patchiness on many scales is more usual. When one wants to study, for example, flow around seagrasses one is faced with asking what characteristics one wishes to simulate. It is unlikely that there will be a flume wide enough to accommodate a natural patch, and if we make the patch the same width as the flume we have forced the flow to go over and through the seagrass, whereas in the field it is likely that horizontal cross-stream deflection of momentum is very important. Modelling flows over non-uniform roughness requires special care. Work by Antonia & Luxton (1971) and references covered in Eckman & Nowell (1984) and Paola, Gust & Southard (1986) indicate the types of problem that must be addressed. An abrupt change in bottom roughness results in an internal boundary layer growing from the leading edge of the roughness. This horizontal non-uniformity in the vertical direction means that we must exercise great care to ensure first that our measurements in the vertical refer to local

82

ARTHUR R.M.NOWELL AND PETER A.JUMARS

effects and not to effects advected from upstream (Townsend, 1969; Peterson, 1971). Secondly, we must now maintain similarity of two sets of dimensionless groups, those referring to the upstream conditions and those referring to the local bed. This double requirement is especially important with reference to the Froude number. Rarely is the marine environment devoid of waves, yet for many problems we can deal with steady flow as a first approximation. Especially in the intertidal and on the continental shelf, however, waves are a zero order effect and we must determine the appropriate dimensionless groups. In a series of papers Grant & Madsen (1979, 1984, 1986) have very elegantly demonstrated the mechanics of the wave boundary layer, and shown that one can model the combined effects of waves and currents on sediment movement as a large steady stress. Although this work includes neither strong local acceleration (which may be important on sloping beaches) nor breaking waves, the work is of great general relevance especially to soft-bottom environments. Grant & Madsen demonstrate how the waves and currents interact and show how to use simple dimensionless scalings based on the maximum combined wave and current stress to parameterize the effects of waves. Strict analog modelling of such flows in the laboratory requires a very different type of flume from that used to model steady, unidirectional flow (Jonsson, 1966; Long, 1981). In summary, theoretical considerations lead us to suggest that for boundary layer flows over a modifiable where is the boundary layer bed we will require that the boundary layer Reynolds number is the free stream velocity, the Froude number the Strouhal number thickness and where n is the frequency, the roughness Reynolds number the Shields parameter and the all be comparable between the laboratory and field. In addition, we shall Rouse parameter need more dimensionless groups if the bed is non-uniform, or if we wish to consider the combined effects of a steady current with waves superimposed. The thin shear layer equations give us guidelines on the scale of the flume we shall require. If we are concerned only with small obstacles on the bed, we have no further concerns other than how to measure the flow, but often the body we want to place in the flow is of a predetermined size. Target organisms rarely come in a complete size range and thus if we wish to examine flow around, for example, a bryozoan there will be a minimum size for the organism. Such a size will not only determine how thick the boundary layer flow over it must be in order to maintain appropriate dynamic similarity, but in the flume we also have to be concerned about flow blockage. If the obstacle occupies more than about 35% of the flow depth (derived from potential flow theory) in the laboratory the flow pattern may be very different from that in the field. If the flow pattern about a seagrass bed is to be studied we have to be careful not to constrain the flow by forcing it to go through the seagrass bed in the flume whereas in the field the flow might diverge around the patch. EXPERIMENTAL CONSIDERATIONS This extensive list of factors that must be considered leads to two obvious but often ignored conclusions: (1) there is no such item as a universally useful flume; and (2) we must define the problem that we wish to study in the laboratory quite specifically before we can identify the crucial elements in the flume that will permit our laboratory measurements to simulate accurately the field environment. While many flumes exist at present it is unwise a priori to expect them to be suitable for specific questions. For example, the flume described by Vogel & LaBarbara (1978) is totally unsuited for boundary layer work because of its size, its three dimensional flow, and its entrance conditions, but is suitable for making drag measurements on bodies far from a solid surface. As a second example, the flume described by Nowell, Jumars & Eckman (1981), and the similar flume described by Muschenheim, Grant & Mills (1986)

FLUMES: SIMULATION OF BENTHIC ENVIRONMENTS

83

are completely inadequate to undertake continuous sediment-transport measurements but are appropriate for making initial-motion measurements and studying organism responses to boundary-layer fluid stresses. As two further examples, the annular flume described by Taghon, Nowell & Jumars (1984) is suitable for transport distance questions, but it is unsuited for precise sediment transport rate problems due to the inherent secondary circulation in an annulus; the flume described by Nowell, Jumars, Self & Southard (in press) is suitable for sediment transport rate measurements under steady and tidal flows, but cannot readily simulate wave effects. Most water tables and fluid containers are complex geometric shapes and so we can expect that any flow within them will have complex velocity fields. Hubbe (1981) provides a useful summary of flow in commonly used laboratory containers and gives approximations for back-calculating approximate values of boundary shear stress. Such methods are clearly approximations but prove very useful when trying to put previous work into its appropriate fluid mechanical context. A series of questions may help to distinguish the properties of the flume that is needed. (1) Is the problem one where unsteadiness dominates or can a quasi-steady approximation be used (Strouhal number)? (2) What are the field values of the Reynolds number and Froude number and can they be appropriately matched in the flume? Do we need other scaling groups such as the roughness Reynolds number, Shields parameter or Rouse number? (3) Are we dealing with an initial motion problem or do we have to maintain constancy of transport not only of the fluid but also the sediment or some other diffused quantity such as larvae (determining the drive or recirculation mechanism)? (4) Is the region of concern approximately uniform or is patchiness important (one set of scalings or many)? (5) What measurement accuracy do we need from the laboratory to answer the question we are posing? FOUR CLASSES OF EXAMPLES There are three major categories of flow channel, namely a straight-through tilting channel where the flow and the sediment may either flow through just once or be recirculated via a pump (Middleton & Southard, 1984), a closed horizontal system such as the race-track flume described by Nowell et al. (in press), and a Ushaped channel for studying oscillatory boundary layers (Jonsson, 1966). Within each category there are wide ranges of flume size and configuration, but the parameters listed will permit us to select not only the type of flume that is appropriate for the problem, but also to decide whether the flumes available fulfil the required scaling lengths. In this section we shall use four examples to highlight some of the questions and challenges involved in designing a laboratory flume study. The problems cover a wide range and end with an example in which it is easier to remain in the field than to try to simulate field conditions in the laboratory. We take first a simple problem to illustrate the use of the equations outlined above. Recent studies (Miller & Jumars, 1986; Hannan, pers. comm.) indicate that feeding rate is dramatically altered for some deposit-feeding polychaetes when their faecal pellets are removed from their feeding ambits. At what shear velocities (u*) are the pellets eroded from the bed? We assume a bed without ripples. As a caveat we emphasize that preliminary observations of the organism under current regimes in the flume will be required not only to determine a realistic bed roughness, but also to permit one to specify when or if the pellets are deposited. For example, Pseudopolydora kempi japonica releases the pellets from its palps while they are extended above the bed. In still water, the pellets accumulate as a mound. In moving water, the mound may

84

ARTHUR R.M.NOWELL AND PETER A.JUMARS

never develop. Assuming faecal piles do accumulate around the tube we need to ensure that the roughness Reynolds number of the surrounding bed is in the same range as that observed in the field, that the flow Froude number is also of the appropriate range and that the bed material has the same hydrodynamic characteristics as at the field site, i.e. the setting velocity of the sediment particles that comprise the bed are the same. In this case we note that by keeping the roughness Reynolds number the same we remove from consideration the bulk flow Reynolds number (which is based on mean flow velocity and boundary layer thickness). We can ignore the latter Reynolds number as we are dealing with an initial motion condition on the boundary and the roughness Reynolds number will set the appropriate scaling, but we still must ensure that laboratory and field Froude number are in the same range. Provided we have a well-constructed flume so that the flow is approximately two dimensional and the test section is of sufficient length that we may obtain a logarithmic layer which is thick relative to the size of the faecal mound the problem is reasonably formulated. We may now review the questions posed at the end of the last section. Because we are dealing with an initial motion problem, and because the particles respond instantaneously to a change in shear stress, we may use a steady flow approximation and ignore the Strouhal number. The problem is concerned with initial motion and we thus must ensure that the upstream roughness and the roughness of the faecal pellets is similar between laboratory and field. Because of this requirement the bulk flow Reynolds number is not relevant, and because we shall use sediment of the same size and density in the laboratory as in the field then the Shields parameter will be similar. It is unlikely we would want to change the geometric properties of the sediment as we know that organisms are size selective in their feeding. In this initial motion problem once the faecal pellets move, then we have answered our question and unless the organisms are very densely crowded in the field we may deal with the problem as one where patchiness is unimportant. If the faecal mound is 0·5 cm high, and the pellets comprised of fine sand, then a flume approximately 2 m long will permit a boundary layer to grow to approximately 6 cm thickness (Table I) so that the logarithmic region is much larger than the faecal mound. By using a hot-film anemometer we could measure the velocity profile to obtain the boundary shear stress. So even for this simple problem of initial motion we require consideration of three dynamic scalings, namely the roughness Reynolds number, the Froude number, and the particle settling velocity; the flume requirements are very modest because we are dealing with a small obstacle, and are not concerned with the flux of particles through the system, only their initial motion. Consider the problem of a suspension-feeder which may occur as an isolated individual or roughness element. Assume we are concerned with studying particle selection and rates of capture. The problem consists of two interrelated parts: getting the boundary conditions correct for the momentum field, and getting the particle concentration field and particle flux field correct. In nature the organism attaches to a hard substratum and we note that at the field site the roughness is much smaller than the scale of the organism (a simplifying assumption for this example, but check the effects of upstream roughness noted by Eckman & Nowell, 1984). The current is predominantly tidal with a maximum velocity of 40 cm·s−1 in a flow depth of 15 m. Particles in suspension are silt and smaller with a settling velocity of less than 0·1 cm·s −1. Rather than attempt to simulate a tidal velocity signal, the problem can be approximated in a series of velocity steps because the boundary layer adjusts quickly enough to changes in the free-stream velocity. Calculating the field Froude number shows that we are always in the subcritical regime. The most important step is to estimate the values of u* for the field, and to select steps in u* that will allow us to evaluate the to the suspended material concentrations. One simple although not overly importance of the ratio precise method would be to use a current meter in the field at 1 m above the bed to get the tidal velocity range and to use the drag coefficient that relates u* to U100 (Sternberg, 1967; Nowell et al., 1981) to calculate the approximate range for the laboratory study. Using u* and the ambient bed sediment size will result

FLUMES: SIMULATION OF BENTHIC ENVIRONMENTS

85

in a roughness Reynolds number which is low because the organism is itself roughening the bed, but it will act as a guide as to whether we can ignore the smooth turbulent regime. The maximum u* and grain size and show whether the bed is being modified; will yield the highest entrainment parameter if the sea floor is rock, an appropriate length scale which describes the height of the roughness will be needed. The flow is now adequately prescribed and the following requirements describe the flume flow. Froude number: 0·03. Roughness Reynolds number: maximum 75. Shields parameter: not relevant if the bed is not modifiable. Organism height: 1 cm (so one requires a laboratory log layer thickness of at least 3 cm; as log layer is 30% of total boundary layer thickness, boundary layer 10 cm). Flow depth: 10 cm, flow width 50 cm. Flume length required to achieve 10 cm thick turbulent boundary layer: approximately 5 m. Maximum discharge: 20 l·s−1. The flume for this study will thus be quite large: at least 6 m long and 50 cm wide with a pump capacity of 40 gallons per minute. The concentration field is modelled appropriately if the correct settling velocity of particles is selected, the particles placed on the flume bed and the boundary shear velocity is varied over the field range. Such variation will show that at some velocities a strong vertical concentration gradient will develop (see Middleton & Southard, 1984) and thus the challenge of making accurate suspended sediment concentration measurements must be addressed. Isokinetic sampling (sampling by sucking into an upstream facing tube at the same velocity as the flow velocity at the tube opening in the boundary layer) is the only accurate method (see Fuchs, 1975). Making accurate measurements of suspension-feeding rates in realistic concentration gradients and flux regimes is thus seen to be very challenging. The biggest problem is developing and measuring the vertical suspended particle profile, and measuring the velocity field on very small scales close to the bed. The problem is even more daunting if one intends to examine particle selection. With such an array of variables to be modelled it is clear that there is plenty of scope for accurate laboratory measurements in the future of suspension-feeding research. The third example illustrates the problems of measuring exchange across the sediment-water interface. Consider the problem of larval recruitment to a bed that is covered by an array of polychaete tubes. The ambient sediment is a fine sand which is adhesively bound so that the dominant roughness is the organism tubes and the polychaete faecal pellets. The region is subtidal, with a maximum velocity of 30 cm·s−1. The larvae have a settling velocity of approximately 0·1 cm·s−1. The problem is now appropriately constrained; the only question remaining being at what scale we want to observe the magnitude and patterns of larval settlement. For the purposes of this example we assume that we are looking at gross settlement patterns, rather than the patterns around a single tube. The recent work of Eckman (see Eckman & Nowell, 1984; Eckman, pers. comm.) deals with details of recruitment around a single tube. Larval settlement problems also necessitate being very careful about the flux of larvae; in a closed recirculating system increasing the velocity increases the flux without changing the standing stock. Moreover, the larvae cannot be put through a pump where the strong shears around the propellers or in the fluid jet may alter the larvae! The major questions in this example revolve around determining the appropriate scale of boundary roughness and ensuring a measurable and controllable flux of larvae. For this example a complete flume width would need to be covered with tube mimics; if only a small patch were to be covered the flow patterns would be complex diverging around the patch of tubes and resulting in a secondary circulation across the flume because of the asymmetry of bed roughness. In addition to the dimensionless groups used in the previous

86

ARTHUR R.M.NOWELL AND PETER A.JUMARS

example, we would need to decide how to alter velocity and particle flux among treatments and how to describe the variability in bed shear stress caused by having tubes distributed over the bed. We note that evaluating settlement into a tube patch of arbitrary size is a much tougher problem than the example detailed above. The fourth category of example highlights two types of problems, one where it is impossible to bring the subject back to the laboratory undisturbed; the other where it is easier, although not easy, to work in the field. The entrainment of soft sediments which has been modified by organisms has attracted the attention of geologists and benthic biologists alike (Rhoads, 1974; Rhoads & Boyer, 1982). It is frequently suggested that it is worthwhile bringing box cores back from the field and running them in a flume to establish the critical entrainment conditions for the sediment surface. One major problem invalidates this simplistic approach, namely that the sediment and water have differing densities. Thus, when the core is being recovered and transported, as the fluid oscillates back and forth over the sediment surface, the bed will be scoured. Even with a closed fluid-filled container the density differences will produce such interfacial stresses, and it would be naïve to believe that a core can be transported up from the sea floor back to the laboratory without such disturbance, not to mention effects on biota. To overcome such problems, in situ flumes have been built. Pioneering work by Southard and Young (Young, 1977) led to the construction of a small straight through inverted flume (SeaFlume) that is suitable for studying initial motion of sands and coarser material which moves as bedload. For such material a critical shear stress can be defined by increasing the imposed velocity until individual grains start to move. For silt and finer materials, which comprise the majority of the sea floor, the SeaFlume is inappropriate; for these sediments we require a closed recirculating system, so the rate of entrainment evaluated as a change in suspended sediment concentration can be measured. A new computer-controlled in situ flume (SeaDuct) designed specifically for studying entrainment of biologically modified marine salts and clays has been tested in shallow water (Nowell, McCave & Hollister, 1985) and is at present being used as part of the High Energy Benthic Boundary Layer Experiment. Devices such as SeaDuct will permit manipulative experiments on fine sediments which could not otherwise be carried out in the laboratory. The final problem we address is that of the effect of breaking waves on a steeply sloping rock face. Clearly the simplifications to the shear layer equations will break down; unsteadiness, surface tension, cavitation due to air entrapment by the waves and gross flow non-uniformity all indicate that it would be nearly impossible to model such a system properly in the laboratory. Even generating the correct wave spectrum and wave amplitude would be daunting, and thus such a problem would more likely yield to careful and innovative field measurements of the relevant fluid forces (Denny, 1982). CONCLUSIONS The most obvious conclusion of this analysis is that no one flume design is applicable to a wide spectrum of benthic biological problems of current interest. The smaller the flume and the smaller the range of discharges and flow depths, the less flexible it will be for applications outside the one for which it is specifically tailored. Conversely, a flume larger than the one needed for the problem at hand will add work and expense not only in capital construction but also in constant care and provisioning with organisms. Flume design and operation is a compromise in every sense, and the optimization problems are not always easy ones. It is especially important that the compromises be made knowingly and that the non-ideal behaviour of the flume be clearly known. With these important caveats, we can expect to see a continuing renaissance of quantitative natural history studies in flumes. The heyday of descriptive natural history was in the late nineteenth century, when most organisms were observed in still water and quantification was not an overriding issue. The welter of

FLUMES: SIMULATION OF BENTHIC ENVIRONMENTS

87

behaviours newly observed or re-interpreted under flow (e.g. Taghon, Jumars & Nowell, 1980; Merz, 1984; Palmer & Brandt, 1981; Butman, 1986b) suggests that the problems to be addressed will not soon be exhausted. Flumes will find an expanding rôle in the quantitative interpretation of material and energy flows in marine ecosystems. They will continue to be used directly in interpreting catches by sediment traps (cf. Butman, 1986a; and Butman, Grant & Stolzenbach, 1986) and will find growing use in determining the interactions between vertical and horizontal particle transport in making resources available to individual organisms (Miller, Jumars & Nowell, 1984). This class of use extends to the prediction and measurement of chemical fluxes into and out of the sea bed (Santschi et al., 1983; Ray & Aller, 1985). Past attention has focused on net vertical fluxes, while new flume capabilities and field observations are shifting attention to the usually much larger horizontal fluxes and gross vertical exchanges, as well as the time dependence and horizontal inhomogeneity. ACKNOWLEDGEMENTS Our work is supported by NSF Grant 86–01543 and ONR Contract NO 14– 003–012. REFERENCES Allen, J.R.L., 1985. Principles of Physical Sedimentology. Allen & Unwin, London, 272 pp. Antonia, R.E. & Luxton, J.A., 1971. J. Fluid Mech., 48, 721–761. Bradshaw, P., 1971. An Introduction to Turbulence and Its Measurement. Pergamon Press, Oxford, 218 pp. Businger, J.A. & Arya, S.P.S., 1974. Adv. in Geophys., 18, 73–92. Butman, C.A., 1986a, J. mar. Res., 44, 645–693. Butman, C.A., 1986b. In, Marine Interfaces: Ecohydrodynamics, edited by J.C.J. Nihoul, Elsevier, Amsterdam, pp. 487–513. Butman, C.A., Grant, W.D. & Stolzenbach, K.D., 1986. J. mar. Res., 44, 601–644. Cebeci, T. & Bradshaw, P., 1973. Momentum Transfer in Boundary Layers. Hemisphere Publishing, London, 391 pp. Chang, P.K., 1970. Separation of Flow. Pergamon Press, Oxford, 777 pp. Coles, J., 1956. J. Fluid Mech., 1, 191–226. Denny, M.W., 1982. Limnol. Oceanogr., 27, 178–183. Dyer, K.R., 1986. Coastal and Estuarine Sediment Dynamics. Wiley, New York, 342 pp. Eckman, J.E. & Nowell, A.R.M., 1984. Sedimentology, 31, 851–862. Francis, J.R.D., 1975. Fluid Mechanics for Engineering Students. Arnold, London, 370 pp. Fuchs, N.A., 1975. Atmos. Env., 9, 697–707. Gilbert, G.K., 1914. The Transportation of Debris by Running Water. U.S.Geol. Survey Prof. Paper 86, 263 pp. Grant, W.D., Boyer, L.F. & Sanford, L.P., 1982. J. mar. Res., 40, 659–677. Grant, W.D. & Madsen, O.S., 1979. J. geophys. Res., 84, 1797–1808. Grant, W.D. & Madsen, O.S., 1984. J. geophys. Res., 87, 469–481. Grant, W.D. & Madsen, O.S., 1986. Ann. Rev. Fluid Mech., 18, 265–305. Grass, A.J., 1971. J. Fluid Mech., 50, 233–255. Guy, H.P., Simons, D.B. & Richardson, E.V., 1966. Summary of Alluvial Channel Data from Flume Experiments. U.S. Geol. Survey Prof. Paper 462–1. 96 pp. Henderson, F.M., 1966. Open Channel Flow. Macmillan Press, London, 522 pp. Hinze, J.O., 1975. Turbulence: An Introduction to its Mechanisms. McGraw-Hill, London, 790 pp. Hubbe, M.A., 1981. Progr. Surface Sci., 11, 65–138. Jonsson, I.G., 1966. Proc. 10th Coastal Engr. Conf., 127–148.

88

ARTHUR R.M.NOWELL AND PETER A.JUMARS

Jørgensen, B.B. & Revsbech, N.P., 1985. Limnol. Oceanogr., 30, 111–122. Jumars, P.A. & Nowell, A.R.M., 1984. Am. Zool., 24, 45–55. Laws, E.M. & Livesy, J.L., 1978. Ann. Rev. Fluid Mech., 10, 247–266. Long, C.E., 1981. A Simple Model for Time-Dependent Stably Stratified Turbulent Boundary layers. Special Rept 95, Dept of Oceanogr., Univ. of Washington, Seattle, 170 pp. Lu, P-C., 1977. Introduction to the Mechanics of Viscous Fluids. Hemisphere Publishing, London, 440 pp. McIntire, C.D., Garrison, R.L., Phinney, H.K. & Warren, C.E., 1964. Limnol. Oceanogr., 9, 92–102. Merz, R.A., 1984. Biol. Bull. mar. biol. Lab., Woods Hole, 167, 200–209. Middleton, G.V. & Southard, J.B., 1984. Mechanics of Sediment Movement. S.E.P.M. Notes for Short Course No. 3, Providence, Rhode Island, 401 pp. Miller, D.C. & Jumars, P.A., 1986. J. Exp. Mar. Biol. Ecol., 99, 1–17. Miller, D.C., Jumars, P.A. & Nowell, A.R.M., 1984. Limnol. Oceanogr., 29, 1202– 1217. Muschenheim, D.K., Grant, J. & Mills, E.L., 1986. Mar. Ecol. Progr. Ser., 28, 185– 196. Nowell, A.R.M. & Jumars, P.A., 1984. Ann. Rev. Ecol. Syst., 15, 303–328. Nowell, A.R.M., Jumars, P.A. & Eckman, J.E., 1981. Mar. Geol., 42, 133–153. Nowell, A.R.M., Jumars, P.A., Self, R.F.L. & Southard, J.B., in press. In, New Perspectives on Deposit Feeding, edited by G.L.Lopez & G.L.Taghon, Springer Verlag, New York. Nowell, A.R.M., McCave, I.N. & Hollister, C.D., 1985. Mar. Geol., 66, 397–409. O’Brien, D.P., Tay, D. & Zwart, P.R., 1986. Mar. Biol., 90, 517–527. Palmer, M.A. & Brandt, R.R., 1981. Mar. Ecol. Progr. Ser., 4, 207–212. Paola, C., Gust, G. & Southard, J.B., 1986. Sedimentology, 33, 279–293. Peterson, E.H., 1971. J. atmos. Sci., 20, 12–22. Ray, A.J. & Aller, R.C., 1985. Mar. Geol., 62, 371–379. Rhoads, D.C., 1974. Oceanogr. Mar. Biol. Ann. Rev., 12, 263–300. Rhoads, D.C. & Boyer, L.F., 1982. In, Animal-Sediment Relations, edited by R.L. McCall & M.J.S.Teresz, Plenum Press, New York, pp. 3–52. Santschi, P.H., Bower, P., Nyffleler, U.P., Azevedo, A. & Broecker, W.S., 1983. Limnol. Oceanogr., 28, 899–912. Shames, I.H., 1967. Mechanics of Fluids. McGraw-Hill, New York, 558 pp. Shields, A., 1935. Versuch. Wasserbau und Schiff, 26, 26 pp. Southard, J.B., Boguchwal, L.A. & Romia, R.D., 1980. Earth Surf. Proc., 5, 17– 23. Statzner, B. & Holm, T.R., 1982. Oecologia (Berl.), 53, 290–292. Sternberg, R.W., 1967. Mar. Geol., 6, 243–260. Sumer, B.M. & Deigaard, R., 1981. J. Fluid Mech., 109, 311–337. Taghon, G.L., Jumars, P.A. & Nowell, A.R.M., 1980. Science, 210, 562–564. Taghon, G.L., Nowell, A.R.M. & Jumars, P.A., 1984. Limnol. Oceanogr., 29, 64–72. Tennekes, H. & Lumley, J.L., 1972. A First Course in Turbulence. MIT Press, Boston, Mass., 300 pp. Townsend, A.A., 1969. J. Fluid Mech., 22, 241–252. Tritton, D.J., 1977. Physical Fluid Dynamics. Van Nostrand Rheinhold, New York, 362 pp. van Dyke, M., 1982. An Album of Fluid Motion. Parabolic Press, Stanford, California, 176pp. Vogel, S. & LaBarbara, M., 1978. BioSci., 28, 638–643. Williams, G.P., 1971. Aids in Designing Laboratory Flumes. U.S. Geol. Survey Open File Report, 294 pp. Yalin, M.S., 1977. Mechanics of Sediment Transport. Pergamon Press, London, 298 pp. Young, R.A., 1977. Mar. Geol., 23, M11-M18.

Oceanogr. Mar. Biol. Ann. Rev., 1987, 25, 113–165 Margaret Barnes, Ed. Aberdeen University Press

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES: THE SPATIAL SCALES OF PATTERN EXPLAINED BY ACTIVE HABITAT SELECTION AND THE EMERGING RÔLE OF HYDRODYNAMICAL PROCESSES1,2 CHERYL ANN BUTMAN3 Department of Ocean Engineering, Woods Hole Oceanographic Institution, Woods Hole, MA 02543, U.S.A. INTRODUCTION While Robertson Davies (1985) was not discoursing on the mechanisms controlling larval settlement of soft-sediment invertebrates when he wrote the poignant phrase, “…science is the theology of our time, and like the old theology it’s a muddle of conflicting assertions”, a perusal of the literature on larval settlement from the last half-decade probably would only fortify this point of view. Since the alternative hypotheses of active habitat selection and passive deposition were identified, an adverse relationship between these mechanisms has been perpetuated in the literature with, seemingly, “conflicting assertions” as to which process is actually responsible for creating the observed distributions of postlarval and adult infauna. There are several notable exceptions, however, where a truce in the war between alternative hypotheses has been proposed by discussions of the space and time scales likely to be associated with each process and the implications to settlement. The present review expands on this theme. The process-orientated literature on larval settlement (primarily studies of active habitat selection) is summarized in terms of the scales of distribution that can be explained. The emerging rôle of bottom boundary-layer flows during settlement is discussed, again with a focus on the applicable scales of the processes. The intention is to illustrate that active habitat selection and passive deposition need not be competing, but are likely complementary, hypotheses by providing examples or proposed scenarios where both mechanisms would be operating, but separated in space and time. Thus, in retort to Davies’ somewhat pessimistic view of the rigour of scientific explanations, I hold that when seemingly conflicting assertions are closely scrutinized they may all be valid replies but, in fact, to different questions. The rôle of larval settlement in determining the distribution and abundance of soft-substratum organisms is still largely unknown (see Connell, 1985, for similar conclusions regarding the hard-substratum case). While factors responsible for inducing larval settlement and metamorphosis have been identified primarily through laboratory experiments in still water, the importance of these factors as causative agents creating the observed infaunal distributions in the field can only be deduced. Little is known about how initial patterns of larval settlement relate to the eventual distributions of adults. Field studies to determine mechanisms controlling benthic community structure usually did not consider larval settlement phenomena. Even when larval settlement was included in analyses, the studies were rarely designed so that differential

1

In honour of the contributions of Douglas P.Wilson. Contribution No. 6303 from Woods Hole Oceanographic Institution. 3 Previously published as Cheryl Ann Hannan. 2

90

CHERYL ANN BUTMAN

larval settlement could be distinguished from differential post-settlement mortality. Studies designed to determine mechanisms controlling larval settlement overwhelmingly favour the active habitat selection hypothesis; most of these studies did not, however, consider or test the alternative hypothesis that larvae are passively deposited onto the sea bed. Whether or not larval settlement sites in the field are determined, even in part, by boundary-layer flow processes has been largely a matter of data interpretation. The few field studies which experimentally explored the rôle of physical processes during settlement, however, demonstrate that hydrodynamical hypotheses are feasible alternative explanations for patterns of larval settlement and infaunal recruitment. This paper reviews the process-orientated literature on soft-substratum larval settlement to establish the scales of observed pattern of infaunal distribution which can be explained by a given process. From this organization of the existing data, some new working hypotheses concerning the rôles of biology and physics during settlement are generated. In addition, conceptual and research gaps in the literature are identified. This is not an examination of the specific cues demonstrated to affect or effect settlement and metamorphosis (Table I), but is designed to complement other larval-settlement reviews by focusing on the scales of pattern and processes. This review is limited to settlement of larvae of soft-sediment infaunal invertebrates, but the literature on hard substrata is tapped periodically to illustrate a particular point that has not been studied in a softsubstratum system. Aspects of the settlement of larvae onto soft and hard substrata may be similar, but there are sufficient differences or potential differences in biological and physical features of these habitats, to which the larvae may respond, that separate discussions are warranted. Hard substrata are basically twodimensional and organisms must attach to the surface, while sediments are three-dimensional and organisms can escape flow forces and other surface phenomena by burrowing into this third dimension (see Woodin & Jackson, 1979). In addition, characteristics of flows over soft substrata may differ from flows over hard substrata because of the nature of the roughness elements and because moveable-bed effects (e.g. Smith & McLean, 1977; Grant & Madsen, 1982) apply only to sediments. Crisp (1984) and Connell (1985) provide recent reviews of factors controlling settlement onto hard substrata although there is not yet a formal treatment of the rôle of hydrodynamical processes (but see Nowell & Jumars, 1984; Wethey, 1986). The physical behaviour of planktonic stages of organisms in fluid flows is probably segregated more by size and taxonomic position than by the nature of the substratum on or in which the organisms live. For example, crustacean larvae generally swim an order of magnitude faster than polychaete larvae (see reviews of Mileikovsky, 1973; Chia, Buckland-Nicks & Young, 1984). TABLE I Reviews of factors (e.g. cues) affecting or inducing settlement or metamorphosis of benthic organisms Reference

Group or taxa reviewed

Thorson (1957, 1966) Wilson (1958) Bergquist, Sinclair & Hogg (1970) Meadows & Campbell (1972a) Campbell (1974) Crisp (1974, 1976, 1984) Fell (1974) Gray (1974) Scheltema (1974)

Soft-substratum organisms Soft- and hard-substratum organisms Demosponges Soft- and hard-substratum organisms, as well as freshwater organisms Cnidarians Soft- and hard-substratum organisms, with special emphasis on barnacles Sponges Soft-substratum organisms Soft- and hard-substratum organisms

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

Reference

Group or taxa reviewed

Berrill (1975) Schroeder & Hermans (1975) Beeman (1977) Chia & Bickell (1978) Eckelbarger (1978) Hadfield (1978a) Lewis (1978) Strathmann (1978) Switzer-Dunlap (1978) Andrews (1979) Pearse (1979) Sastry (1979) Guérin (1982) Burke (1983) Day & McEdward (1984)

Tunicates Polychaetes Opisthobranch gastropods Coelenterates Sabellariid polychaetes Molluscs Cirripedes Echinoderms Aplysiid gastropods Oysters Chitons Bivalves (excluding oysters) Soft- and hard-substratum organisms Hard-substratum organisms Soft- and hard-substratum organisms

91

Concerning the effects of flow physics on sinking and settlement of larvae, the size of the organism and its fall velocity are the most important considerations (see Hannan, 1984a). Thus, in discussions of physical processes, I often include studies of any organism within the size range of infaunal larvae. For clarity, the definitions of some terms commonly used in larval studies are reiterated here. Planktonic larvae of infaunal invertebrates undergo larval development in the water column during the dispersal phase of their life history. Dispersal is entirely a water column phenomenon; larvae are generally considered to be developing and not ready to settle during large-scale dispersal. Dispersing planktonic larvae are easy to identify in certain groups, such as the Crustacea, where several distinct larval stages precede a final larval form that can settle onto the substratum. In other groups, such as the Polychaeta, larval development usually involves the gradual addition of segments and loss of ciliation; cessation of planktonic development, and thus, of the dispersal phase, may be more difficult to identify in these groups. Dispersal is largely regarded as passive transport by water currents because the scales of horizontal water motion are so much greater than the swimming speeds of larvae (Mileikovsky, 1973). Active behavioural and physiological responses of larvae (e.g. phototaxis, geotaxis, and responses to salinity changes) may, however, displace them vertically into water masses where fluid motion may not be large compared with the movements of the larvae (Mileikovsky, 1973, but see also Butman, 1986a). Thus, there is an active component to dispersal in many organisms because they can make vertical migrations into particular water masses (e.g. see review by Sulkin, 1984, for decapod larvae). The relative contribution of active larval behaviours compared with passive transport processes for the retention of larvae in estuaries has been debated for several decades (see Kennedy, 1982, for recent papers on this topic); contemporary ideas on larval dispersal processes can be found in the overview by Scheltema (1986). Planktonic larval development was separated into two periods for the gastropod, Nassarius obsoletus, by Scheltema (1967): a development period during which growth and differentiation of the larva occurs and a delay period during which there is a gradual decrease in growth and the larva is physiologically capable of metamorphosis. During metamorphosis, an organism undergoes certain morphological changes that “portend a new way of life” (Scheltema 1974, p. 263). For organisms with a planktonic larval stage,

92

CHERYL ANN BUTMAN

settlement is defined as “the termination of a pelagic, larval existence and the assumption of a sessile or nonsessile sedentary life”, Scheltema (1974, p. 263). Usually larvae considered available for settlement include only larvae that have entered the delay period of their planktonic development and, thus, are competent to metamorphose. For many species this is difficult to determine, except ex post facto; even in laboratory experiments, discerning if an unmetamorphosed larva is capable of metamorphosis is difficult because once the organism has passed through the development period, it may require a specific cue for metamorphosis to occur (see reviews listed in Table I). None the less, in this review larvae considered to be available for settlement are only those within the delay period of their planktonic larval development (i.e. competent larvae). It is important to distinguish clearly between metamorphosis, settlement, and recruitment. Metamorphosis may precede, coincide with, or follow settlement and refers to an irreversible set of anatomical and physiological changes in the organism presumably “coordinated through an endocrine mechanism” (Scheltema, 1974, p. 263). Metamorphosed larvae are referred to as postlarvae in this review. Settlement “denotes a responsive behaviour” and is “presumed to be under nervous control” (Scheltema, 1974, p. 263). While this definition could imply an active choice by the larva to settle, the interpretation used here is that, during settlement, the organism takes up activities or behaviours (e.g. burrowing and tube building) which are indicative of the benthic life history stage. Thus, whereas metamorphosis involves morphological and physiological changes from a larval to a postlarval form, settlement involves a change in venue from a planktonic to a benthic existence. Unfortunately, this definition of settlement implies that once settled, the organism will not re-appear in the water column but will reside entirely in or on sediments. In this regard, Scheltema’s (1974) definition of settlement must be modified because there are mounting records of the occurrences of postlarval and adult benthic organisms in the water column (see Table II), exclusive of swarming behaviour for spawning and reproduction. TABLE II Observations of benthic postlarval and adult polychaetes, molluscs, and meiofauna in the water column: some studies may have collected larval forms along with immature and adult organisms; this information usually was not given in the paper; when information on the state of maturation of the collected organisms was given in the paper, the name of the organism is followed by: A=adult, F=female with eggs, I=immatures, NE=not epitokous, NS=non-spawners, PM=post-metamorphic, S=some individuals may have been spawning; all species listed are polychaetes, unless noted as B=bivalve Reference

Method of observation

Organisms observed

Bayne (1964) Emery (1968) Williams & Porter (1971)

Plankton tows Plankton tows; suction devices Plankton tows

Seymour (1972)

Laboratory observations, but cites Page & Legendre (1927) for field observations

Mytilus edulis (B) Nereids and other polychaetes Ensis directus (B, PM) Tagelus divisus (B, PM) Solemya (=Solenomya) velum (B, PM) Solen viridis (B, PM) Donax variablis (B, PM) Petricola pholadiformis (B, PM) Spisula raveneli (B, PM) Arenicola marina (NS)

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

Reference

Method of observation

Organisms observed

Thomas & Jelley (1972)

Emergence traps

Beukema (1973) Porter (1974) Hobson & Chess (1976) Alldredge & King (1977)

Plankton tows Plankton tows Plankton tows Emergence traps; plankton tows

Porter & Porter (1977) Porter, Porter & Batac-Catalan (1977) Dean (1978a)

Emergence traps Emergence traps

Eteone lactea Glycera dibranchiata Nereis succinea (I, A, S) Nereis virens (I, A, S) Pherusa affinis Scoloplos fragilis (F, S) Macoma balthica (B) Polychaetes (A) Polychaetes Polychaetes from the families: Syllidae, Orbiniidae, Opheliidae Polychaetes Polychaetes

Dean (1978b)

Direct observations of surface waters at night with search-light; specimens collected with dipnet Direct observations of surface waters at night with search-lights; buoyed and anchored nets

Nereis virens (I)

Glycera dibranchiata (I)

Glycera capitata (I) Eteone longa (F) Nephtys discors Glycera sp. (I) Nereis virens Pherusa affinis Graham & Creaser (1978) Beukema & DeVlas (1979) Hobson & Chess (1979) Alldredge & King (1980)

Buoyed and anchored plankton nets Plankton tows Emergence traps Emergence traps

Glycera dibranchiata (A, NS) Arenicola marina (I) Polychaetes Brania sp. (F, S) Sphaerosyllis hystrix (F, S) Armandia brevis (A, NE) Aricidea sp. (A) Prionospio heterobranchiata newportensis (A) Pseudoeurythoe sp. (A) Gyptis brevipalpa (A) Protodorvillea gracilis (A)

93

94

CHERYL ANN BUTMAN

Reference

Method of observation

Hammer & Zimmerman (1979)

Emergence traps

Bell &Sherman (1980)

Samples of water overlying sediments Plankton tows; direct observations of surface waters Containers filled with sediment raised 0·5-m above bottom

Dauer, Ewing, Tourtellotte & Barker (1980) Santos & Simon (1980a)

Bhaud, Aubin & Duhamel (1981)

Near-bottom sediment traps

Hagerman & Rieger (1981)

Sediment traps

Hammer (1981)

Emergence traps

Levin & Greenblatt (1981)

Plankton tows

McWilliam, Sale & Anderson (1981)

Emergence traps

Organisms observed Nematonereis unicornis (A) Syllidae, sp. A Alciopidae, sp. A Autolytus sp. Diopatra ornata Eunoe sp. Exogone sp. Lumbrineris sp. Odontosyllis sp. Schistomeringos longicornis Platynereis bicanaliculata Meiofauna Meiofauna Scolecolepides viridis (I, A, S) Nereis succinea (A) Gyptis vittata (A) Parahesione luteola Stylochus sp. Phyllodoce sp. (I or A) Polydora antennata (I or A) Spio martinensis (I or A) Amphictenidae (I or A) Polyophthalmus pictus (I or A) Ophryotrocha sp. (I or A) Lamellibranchs (B, I or A) Meiofauna Meiobenthic polychaetes Schistomeringos longicornis Minuspio spp. Odontosyllis sp. Hesionids Loimia sp. (I) Exogone sp. (I) Armandia sp. (I)

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

Reference

Method of observation

Tranter et al. (1981) Ohlhorst (1982) Dobbs & Vozarik (1983)

Downward-directed light traps Emergence traps Filtrate from power plant cooling system

Levin (1984)

One-gallon plankton jar samples

Hannan (1984b) Palmer & Gust (1985) Levin (1986)

Near-bottom sediment traps Pump samples of near-bottom water One-gallon plankton jar samples

95

Organisms observed Other polychaetes Polychaetes Polychaetes Polychaetes (47 species) Bivalves (7 species) Gastropods (12 species) Exogone lourei (A, S, F) Odontosyllis phosphorea (A, S) Harpacticoid copepods Mediomastus ambiseta (A, I) Meiofauna Streblospio benedicti (I) Rhynchospio arenincola (I)

In contrast to dispersal, metamorphosis, and settlement, recruitment is not a physiologically or behaviourally distinct stage in the life history of the organism, but is observer-defined; the organisms surviving to a size collected by the sampler are considered recruited individuals (Keough & Downes, 1982). Because recruitment generally is defined by the sieve screen size and the sampling interval in infaunal studies, recruited organisms can be unmetamorphosed larvae, postlarvae, juveniles, or even adult organisms. Note, however, that settlement refers only to larvae or, when larvae metamorphose prior to settlement, postlarvae. By these definitions, in order to study settlement, the first larval stages to reach the sea bed and begin living as benthic organisms must be sampled; to emphasize this, I often refer to these as “initially settled larvae” in this review. As Keough & Downes (1982) and Connell (1985) recently elucidated for hardsubstratum organisms, most studies which claim to measure larval settlement have actually measured recruitment. PATTERNS OF JUVENILE AND ADULT DISTRIBUTION Research on factors that determine settlement sites for infaunal larvae was motivated by early documentation (e.g. Petersen, 1918) of distinct faunal assemblages that vary spatially. The procession of benthic survey studies which followed further revealed that species distributions are often well correlated with distributions of particular sediment grain sizes (Table III and summaries of the early work by Thorson, 1955, 1957; but see also conclusions of McNulty, Work & Moore, 1962b; Santos & Simon, 1974). The spatial TABLE III Benthic survey studies where infaunal communities and sediment grain sizes were documented over areas of the sea floor: the intention of this table is to give a historical perspective of the sampling scales, sampling intervals and sieve sizes used in a selection of survey studies where benthic communities and sediments were sampled simultaneously; the list is intended to be illustrative, not exhaustive; minimum distances between stations were usually estimated from plots

96

CHERYL ANN BUTMAN

of station locations on maps of the study sites; NG=information not given in the reference; sampled once=each station was sampled once only and not necessarily simultaneously with the sampling of the other stations Reference

Location of study

Depth (m) Sieve screen size (µm)

Minimum sampling Minimum distance interval between stations (m)

Ford (1923) Spärck (1933)

Plymouth, England Franz Joseph Fjord, East Greenland Northern North Sea, Scotland Scoresby Sound Fjord Complex, East Greenland Mouth of Exe Estuary, England Long Island Sound, NY Buzzards Bay, MA Puget Sound, WA Biscayne Bay, FL

NG 27–780

“series of screens” NG

Sampled once Sampled once

123 730

0–200

1500, 2000

Sampled once

NG

20–530

NG

Sampled once

1760

Intertidal

1000

Sampled once

30

6–31

300, 2000

2 months

3220

7–20 Intertidal 2–10

500 Not sieved 1000

Sampled once Sampled once Sampled once

1850 NG 230

Barnstable Harbor, MA

Intertidal

Unscreened, 750

1 yr

NG

Coast of Northumberland, North Sea South of Martha’s Vineyard, MA Manukau Harbor, North Island, New Zealand

10–90

NG

Sampled once

1800

40–567

74, 1000

Sampled once

1800

Intertidal

2500

Sampled once

800

Stephen (1933) Thorson & Ussing (1934) Holme (1949) Sanders (1956) Sanders (1958) Wieser (1959) McNulty, Work & Moore (1962a) Sanders, Goudsmit, Mills & Hampson (1962) Buchanan (1963)

Wigley & McIntyre (1964) Cassie & Michael (1968)

Gray (1968) Lie (1968) Gibbs (1969) Lie & Kisker (1970) Nichols (1970) Day, Field & Montgomery (1971) Hughes & Thomas (1971) Johnson (1971)

Eagle Cove, San Juan Island, WA Puget Sound, WA Plymouth Sound, England Juan de Fuca Strait and off Washington coast, USA Port Madison, Puget Sound, WA Beaufort Shelf, NC Biddeford River, Prince Edward Island, Canada Tomales Bay, CA

Intertidal

“fine plankton net” 1·5 months

10

12–200 2–12 13–317

1000 500 1000

2 months Sampled once Sampled once

2000 50 1800

2–34

1000

Sampled once

75

0–200

1000

2 months

1110

Intertidal

500

Sampled once

5

0–18

1500

NG

NG

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

Pearson (1971) Bloom, Simon & Hunter (1972) Gage (1972a, b)

Lochs Linnhe and Eil, Scotland Tampa Bay, FL

Lochs Etive and Crenan, Scotland Hughes, Peer & Mann (1972) St. Margarets Bay, Nova Scotia Eagle (1973) Liverpool Bay, England Gage & Geekie (1973b) Loch Etive, Scotland Santos & Simon (1974) Tampa Bay, FL Crumb (1977) Delaware River, NJ Mountford, Holland & Chesapeake Bay, MD Mihursky (1977) Tyler & Banner (1977) Oxwich Bay, Bristol Channel, Wales Whitlatch (1977) Barnstable Harbor, MA Larsen (1979) Sheepscot Estuary, ME Flint & Holland (1980) Gulf of Mexico, TX Shin & Thompson (1982) Coastal waters of Hong Kong Georges Bank, MA Maciolek & Grassle (in press)

0–>90

1000

Sampled once

250

Intertidal

1000

4 months

20

0–117

1000

Sampled once

710

18–>60

800

Sampled once

175

5–11 500, 1000 20–60 1000 Intertidal 500 Intertidal 500, 1000 3–9 1000

5 months 310 Sampled once 100 3 months 90 1 month 4000 3 months 280

5–20

1000

Sampled once

55

Intertidal 0–9 22–131 13–70

250 1000 500 400

1 month Sampled once 1 month 2 months

65 NG 6390 500

38–168

300

3 months

500

97

scales (tens of metres to tens of kilometres) separating distinct assemblages and sediment types were, in part, dictated by the manoeuvrability of the sampling vessel and the accuracy of shipboard-operated navigational equipment; minimum distances between subtidal stations ranged from 50 m to 64 km, while intertidal communities could be sampled at closer intervals of 5 m to 800 m (Table III). In addition, once a relationship between species and sediment composition was observed, delimiting species distributions in relation to sediment type became the primary purpose of most benthic surveys, so relatively large distances between stations were desirable, since significant differences in bulk properties of sediments (e.g. grain size) could be easily detected at these spatial scales. As the topic of organism-sediment relations was experimentally dissected through the years, nearly all of the field and a good portion of the laboratory research was on the favourability of particular habitats to adults and on the interactions between different trophic and mobility types (Gray, 1974; Rhoads, 1974). Detailed studies of the feeding and mobility types of the infauna revealed that functional groups of organisms occurred in distinct types of sediment. Most authors did not speculate on larval settlement mechanisms which could have produced these patterns of distribution (in fact, larval settlement is not discussed at all in Rhoads’, 1974, review), but only discussed the favourability of these particular habitats to adults. Initially, the most popular explanation for these assemblages concerned the availability of food resources. For example, Sanders (1958) hypothesized that deposit-feeders dominate clays because these sediments are also rich in organics and microbes, while filter-feeders occur in sandier environments because the higher near-bottom flows deliver suspended particulates to the organisms at faster rates. Later experimental manipulations showed that interactions between functional groups are also important. Rhoads & Young’s (1970) classic “trophic group amensalism” hypothesis, for example, states that activities

98

CHERYL ANN BUTMAN

of deposit-feeding organisms interfere with the establishment and maintenance of populations of suspensionfeeders and that such amensalistic interactions are intimately related to the nature of the sedimentary environment (e.g. the degree of substratum motion). Recently, more complex interactions between the feeding and locomotive activities of benthic organisms and the structure of the bottom boundary-layer flow and sediment-transport regimes have been identified, stipulating a re-evaluation of the effects of functional groups on sediments and sediment transport (Jumars, Nowell & Self, 1981; Jumars & Nowell, 1984a) and of fluid- and sediment-dynamic effects on community structure (reviewed by Jumars & Nowell, 1984b). Few of these studies considered how the functional groups of organisms are initially established. The studies usually did not ask if the distinct assemblages resulted from differential larval settlement or differential post-settlement mortality, nor did they consider how the mechanisms controlling larval settlement (e.g. active habitat selection or passive deposition) would affect the establishment and maintenance of the assemblages (see also Dayton & Oliver, 1980). The licence to focus primarily on adults may have resulted because, concurrent with the early survey studies, a relatively small core of biologists (e.g. see studies cited in Table V, see pp. 128–9) conducted meticulous laboratory experiments on infaunal larvae and meiofauna, demonstrating that the organisms can actively choose between microhabitats. Thus, the rôle of larval settlement in creating the observed organism-sediment relations was generally assumed to be through active habitat selection (e.g. Thorson, 1957; Wilson, 1958; Meadows & Campbell, 1972a; Gray, 1974), even though scant direct evidence from the field was available to support this tenet (see later discussion, pp. 139–141). In fact, passive deposition of larvae also could have produced the observed patterns of organism distribution if, (1) larvae were deposited over broad areas, but differentially survived only in hospitable adult habitats (corresponding to particular sedimentary environments), or (2) speciesspecific larval fall velocities corresponded with particular sediment fall velocities so that hydrodynamically similar particles and larvae were deposited in the same environment. Interpreting the importance of amensalism or other interactions between established infauna and the flow or sediment environment to benthic community structure requires knowledge of the rôle of larval settlement processes (see also Jumars & Nowell, 1984b). For example, the trophic group amensalism hypothesis requires that initial distributions of larvae on the sea bed result from differential larval settlement, due to active habitat selection, or to differential post-settlement survival. If differential larval settlement results from passive deposition (i.e. settlement patterns depend on larval fall velocities and on the near-bottom flow regime), then it may not be necessary to evoke complex amensalistic interactions to explain the distributions of the adults. Thus, for example, suspension-feeders may not co-occur with deposit-feeders simply because the two functional groups have larvae with different fall velocities that are passively deposited in different fluid-dynamic environments. In both the survey and the process-orientated studies of soft-bottom community structure, the importance of larval ecology cannot be assessed a posteriori because larvae were rarely quantitatively collected in samples (Tables III and IV). Two methodological problems have especially prohibited an adequate consideration of the larval stages; Dayton & Oliver (1980), Santos & Simon (1980a), and Williams (1980) have discussed these problems. (1) Field sampling was usually too infrequent (monthly or even biweekly) to record initial settlement prior to post-settlement interactions. (2) The sieve screen size (500 µm) commonly used in recent benthic studies is too large to retain newly settled larvae of most invertebrate species. Even though the sieve screen size used in faunal surveys has decreased over time (note that Thorson, 1966, defined macrofauna as those organisms retained on a 2-mm sieve and meiofauna was originally defined by Mare, 1942, as organisms with body lengths between 0·2 and 2·0 mm), so that 300-µm screens are used in some contemporary survey studies (e.g. Grassle et al., 1985; Thistle, Yingst & Fauchald, 1985; Maciolek &

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

99

Grassle, in press), 60- or 100-µm screens often are required to retain newly settled larval (or postlarval) stages (Eckman, 1979; Gallagher, Jumars & Trueblood, 1983; Hannan, 1984a). THE RÔLE OF LARVAL SETTLEMENT In most discussions of the rôle of larval settlement in soft-substratum community ecology (e.g. Thorson, 1946, 1950, 1957, 1966; Smidt, 1951; Muus, TABLE IV Process-orientated field studies of factors controlling soft-substratum community structure: studies are arranged by the process under investigation; included in this table are studies of processes structuring macrofaunal communities, but not studies of single species populations; this list includes the commonly cited studies in the English literature and is not intended to be comprehensive; NS=some of the samples were not sieved Reference Colonization, succession, response to disturbance Grassle & Grassle (1974) Boesch, Diaz & Virnstein (1976) Dauer& Simon (1976) McCall (1977) Rees, Nicholaidou & Laskaridou (1977) Rhoads, Aller & Goldhaber (1977) VanBlaricom (1982) Woodin (1978) Oliver et al. (1980) Santos& Bloom (1980) Santos& Simon (1980a) Santos & Simon (1980b) Arntz & Rumohr (1982) Zajac & Whitlatch (1982a, b) Gallagher, Jumars & Trueblood (1983) Watzin (1983, 1986) Ambrose (1984b) Levin (1984) McGrorty & Reading (1984) Predation Young, Buzas & Young (1976) Reise (1978) Virnstein (1978) Arntz (1980) Holland et al. (1980) Hulberg& Oliver (1980) Mahoney & Livingston (1982) Animal-sediment relations

Minimum sampling interval

Sieve screen size (µm)

3 days 3 months 1 month 10 days 1 month 2 months 1 wk 1 month 1 month 1 month 1 wk 1 month 2 months 2 wk 2 days 7 days 2 months 3 days 6 months

297 500, 1000 500 297, 1000 1000 300, 1000 500 1000 500 500 250, 500 500 125, 500, 1000 297 63 63 500 250 500

1 month 2 wk 2 months 2 months 2 months 2 months 1 month

1000 500, 1000, NS 500 1000 500 250, 500 500

100

CHERYL ANN BUTMAN

Rhoads & Young (1970) Young & Rhoads (1971) Levinton (1977) Myers (1977a, b) Orth (1977) Wilson (1979) Brenchley (1981) Wilson (1981) Competition Woodin (1974) Peterson (1977) Weinberg (1979) Peterson & Andre (1980) Wilson (1983) Temporal variability Muus (1967) Lie (1968) Dauer& Simon (1975) Peterson (1975) Holland & Polgar (1976) Holland, Mountford & Mihursky (1977) Whitlatch (1977) Buchanan, Sheader & Kingston (1978) Ziegelmeier (1978)

1 month Sampled once 1 yr 1 wk 1 month 1 month 7 days 15 days

500, 1000 1000 2000 500 1000 500 500 500

1 month 1 day 1 month 55 days 1 wk

500, 1000 2300 500 2300 500

1 month 2 months 3 months 4 months 3 months 3 months 1 month 2 months 6 months

700, 1000, NS 1000 500 3200 1000 1000 250 500 3200

1973; Gray, 1974; Woodin, 1976, 1979, 1985; Oliver, 1979; Woodin & Jackson, 1979; Dayton & Oliver, 1980; Watzin, 1986), active habitat selection by larvae is the favoured mechanism for establishing benthic communities. Support for this hypothesis comes primarily from the numerous laboratory experiments where larvae were given a choice of substrata in which to settle (Tables I and V). Post-settlement mortality may also determine (e.g. Levinton & Bambach, 1970; Luckenbach, 1984) or further restrict the distribution of adults (e.g. Thorson, 1966; Muus, 1973; Oliver, 1979; Peterson, 1986; Watzin, 1986). It is not surprising that active habitat selection has been the favoured larval-settlement process because the clear evidence from the early laboratory studies (see Table V) is difficult to ignore. In a few notable discussions, however, reservations were raised regarding the application of these laboratory results, where experiments were conducted at very small scales and in still water, to the field, where the scales of processes are much larger. While Thorson is frequently credited as an early advocate of active habitat selection, because of his observations of settling larvae associated with particular sediment types in “bottle collectors” (Thorson, 1946), he was, in fact, consistently cautious when applying results of laboratory

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

101

experiments in still water to the field. For example, regarding the choice experiments of Wilson (1952, 1953a, b), Thorson (1966, p. 275) noted that the experiments were done, “… in petri-dishes, where the larvae by swimming 1 or 2 centimeters only had a chance to discriminate between heaps of sand which might be more or less attractive, neutral or more or less repellent to them as a future substratum. In nature, the larvae will not get a similar opportunity to compare a series of substrata by swimming a short distance only…. Far from questioning Wilson’s main thesis: That the larvae may discriminate between attractive and non-attractive substrata, a fact shown so convincingly that it can be accepted as a ‘biological rule’, we have, however, to find out what will happen in nature, when a larval swarm ready to metamorphose and drifting along the bottom will for the first time meet a substratum which they might ‘accept’, although it is far from ideal for their settling. The larvae cannot know, that if they continued to drift over the bottom for perhaps 10–20 kilometers more, they might meet a much more attractive substratum. It seems reasonable to assume, that such larvae, at least if they have already postponed their metamorphosis for some time and are in their less critical phase, will accept and accordingly settle in a bottom substratum much less attractive than the one they would have preferred, had they been given a ‘free choice’. “The consequence of this must be, that the distributional pattern of larvae on the natural bottom substrata must be much less delicate, i.e. much more coarse, than in the experiments undertaken in the laboratory.” Thorson was impressed with the behaviour of dispersing and settling larvae, but he also acknowledged that test sites for larvae on the sea bed were probably dictated by near-bed currents, so he tended to under- rather than over-state the rôle of active habitat selection. The evidence that larvae can test the substratum and have preferred habitats simply indicates that “…their chance of finding a suitable place for settling is much better than hitherto believed” (Thorson, 1950, p. 36). One of the most lucid commentaries on the extent to which active habitat selection determines the distribution of benthic marine organisms is the brief (and infrequently cited) paper by Moore (1975), which was written in response to the views of Meadows & Campbell (1972a, b) and Meadows & Mitchell (1973). Moore (1975) proposed “habitat availability” and “ecological opportunity” as alternative arguments to active habitat selection and animal behaviour for explaining the “local” distribution of organisms in the sea. He reemphasized one of Thorson’s (1966) points, that organisms may not have the same kinds of “free choices” in the field as they have been given in the laboratory. During dispersal, planktonic larvae are restricted to particular localities by passive transport processes so that larvae may never even encounter preferred substrata (as determined in laboratory experiments) in the field. Post-settlement mortality or passive deposition of larvae may then shape species distributions. Moore (1975) also reiterates the postulate of Pratt (1953; discussed more later, see pp. 141–3) that correlations between the composition of softsubstratum communities and sediment type may also result from passive sorting of both larvae and sediments by hydrodynamical processes. Furthermore, Moore (1975) emphasizes the importance of scale in directly applying the habitat-selection results to the field, stating that “local” to a behaviouralist (e.g. Meadows & Campbell, 1972a, b; Meadows & Mitchell, 1973) may refer to a much smaller scale (i.e. onthe-order-of the organism) than the “local” of an ecologist, which generally refers to more geographicaltype scales; the disparity between these scales decreases, however, with increasing organism size and their ability to independently traverse large distances. His concluding remarks (Moore, 1975, p. 100) raise questions that are still relevant, and largely unanswered, today: “Re-examining the generality of Meadows and Campbell’s statement that habitat selection largely determines the local distribution of animals in the sea, a number of issues appear conditional, (i) how is ‘determine’ construed? (ii) how is the concept of ‘local’ envisioned? and (iii) which type of ‘animal’ is involved with reference to (ii)? But in any circumstances, to regard habitat selection as ‘largely’ determining local distribution would seem to be an overstatement of the case.”

102

CHERYL ANN BUTMAN

A small number of benthic studies (e.g. Baggerman, 1953; Pratt, 1953; Fager, 1964; Tyler & Banner, 1977) have favoured the passive deposition, rather than the active habitat selection, hypothesis to account for some or all of the observed patterns of infaunal species distribution. Curiously, the passive deposition hypothesis was suggested to these authors by the same kinds of correlations between sediment and species composition, that led most other authors (cited previously) to conclude that larvae actively select for particular sediment or sedimentary environments. Support for the passive deposition hypothesis was largely correlative in these early studies (but see later discussion of experimental manipulations by Baggerman, 1953). Later studies have, however, shown that hydrodynamical null hypotheses are feasible explanations for the observed patterns of distribution; Jumars & Nowell (1984b) review some of this work. Now that the stage has been set with the alternative hypotheses of active habitat selection compared with passive deposition for creating observed patterns of species distributions, it is fruitful to examine closely the data base substantiating each of these views to determine the plausibility and scales of cause and effect. THE ACTIVE HABITAT SELECTION HYPOTHESIS Laboratory experiments on larval settlement can be divided roughly into two groups: (1) studies of habitat selection (i.e. where larvae were given a choice of substrata) and (2) studies of environmental or biological factors that induce metamorphosis. Results from experiments in the first category can provide direct evidence of habitat selection, while selection is only implied by results from experiments in the second category. There is some confusion in the literature as to which studies actually provide direct evidence of habitat selection (through choice experiments), so these are listed in Table V and will be discussed separately from the metamorphosis experiments. All laboratory studies of active habitat selection (i.e. the choice experiments) were done in still water, except one (Cuomo, 1985), so that relevance of these results to settlement in field flows is at present obscure. The response of settling larvae to water motion was qualitatively investigated in the laboratory for several infaunal larvae and meiofauna species. The polychaete larvae of Ophelia bicornis and Polydora ciliata were stimulated to attach to sand grains when subjected to water motion (“squirting” water on larvae placed in a Petri dish in the case of Wilson) in the studies of Wilson (1948) and Whitelegge (1890), respectively. Wilson (1948) also reported that Ophelia could “use” the current in order to detach from an unpreferred substratum to re-enter the flow. Wilson (1968) and Eckelbarger (1975, 1976) induced settlement of larvae of sabellariid polychaetes which live in habitats subjected to waves as adults, by stirring the water in the experimental container. Boaden (1963, 1968) and Gray (1966b) observed behaviour of meiofauna in water flowing through a small space between parallel plates and through clear tubing. They found that, at low current speeds, some species were rheotactic, moving upstream toward the source of the current, but all of the organisms were simply washed downstream above some higher current speeds. These studies of water motion relative to some aspects of the behaviour of settling infaunal larvae or meiofauna were not designed to mimic a particular, realistic boundary-layer flow regime. At most, the mean current speed (i.e. TABLE V Laboratory experiments on substratum selection by soft-substratum invertebrate larvae, juvenile or adult macrofauna, epifauna and meiofauna: dimensions of treatments and distances between treatments are rough estimates, taken from the information available in the reference; A=archiannelid; B=bivalve; C=cumacean; CR=crab; G=gastrotrich; GA=gammarid amphipod; H=harpacticoid copepod; I=isopod; L=lancelet; LO=lobster; N=nematode;

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

103

O=ophisthobranch gastropod; P=polychaete; S=shrimp; T=turbellarian; TO=tubificid oligochaete; TP=thin partition between adjacent sediment treatments; NG=information not given in paper Reference

Organism(s) studied

Studies of macrofauna or epifauna larvae Wilson (1948) Ophelia bicornis (P) Wilson (1952, 1953a, Ophelia bicornis (P) b, 1954, 1955) Wilson (1970a) Sabellaria alveolata (P) Wilson (1970b) Sabellaria spinulosa (P) Wilson (1977) Lygdamis muratus (P) Keck, Mauer & Mercenaria Malouf (1974) mercenaria (B) Botero & Atema Homarus (1982) americanus (LO) Cuomo (1985) Capitella sp. I (P) J.P.Grassle (pers. Capitella sp. I and II comm.) (P) McCann (in press) Streblospio benedicti and L.A.Levin (pers. (P) comm.) Studies of meiofauna Wieser (1956) Cumella vulgaris (C) Gray (1966a) Protodrilus symbioticus (A) Gray (1966b) Protodrilus symbioticus (A) Gray (1966c) Protodrilus symbioticus (A) Gray (1967a) Protodrilus rubropharyngeus (A) Gray (1967b) Protodrilus hypoleucus (A) Jansson (1967a) Parastenocaris vicesima (H) Jansson (1967b) Coelogynopora schulzii (T) Aktedrilus monospermatecus (TO)

Maximum dimension Maximum dimension Maximum distance of experimental of treatment (cm) between treatments container (cm) (cm) 9·0 7·0

1·5 0·75

4·5 “a few cm”

6·5

1·7

3·0

6·5

1·7

3·0

6·5

2·3

2·0

NG

NG

NG

55·0

27·5

TP

60·0 13·3

7·5 3·9

3·0 5·5

25·5

7·6

1·0

NG 7·0

NG 1·0

NG 5·0

15·0

NG

NG

7·0

1·0

5·0

7·0

1·0

5·0

7·0

1·0

13·0

NG

0·8

NG

NG

0·8

NG

104

CHERYL ANN BUTMAN

Gray (1968)

Leptastacus constrictus (H) Gray & Johnson Turbanella hyalina (1970) (G) Jensen (1981) Chromadorita tenuis (N) Klauser (1986) Convoluta sp. (T) Studies of macrofauna or epifauna Teal (1958) Uca minax (CR) Uca pugilator (CR) Uca pugnax (CR) Webb & Hill (1958) Branchiostoma nigeriense (L) Williams (1958) Penaeus setiferus (S) Penaeus aztecus (S) Penaeus duorarum (S) Meadows (1964a) Corophium volutator (GA) Corophium arenarium (GA) Meadows (1964b) Corophium volutator (GA) Meadows (1964c) Corophium volutator (GA) Corophium arenarium (GA) Croker (1967) Parahaustorius longimerus (GA) Neohaustorius schmitzi (GA) Lepidactylus dytiscus (GA) Haustorius sp. (GA) Acanthohaustorius sp. (GA) Lewis (1968) Fabricia sabella (P) Sameoto (1969) Haustorius canadensis (GA) Neohaustorius biarticulatus (GA) Acanthohaustorius millsi (GA) Parahaustorius longinerus (GA)

7·0

1·0

5·0

15·0

1·0

13·0

10·0

10·0

6·0

5·0

1·0

3·0

75·0

37·5

TP

24·0

7·5

10·6

243·0

45·7

137·2

9·0

9·0

16·0

34·0

17·0

TP

12·0

6·0

TP

“Large finger bowls”

“Divided in half”

TP

9·0 9·5

6·4 6·7

TP TP

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

Hadl et al. (1970) Jones (1970)

Morgan (1970) Gray (1971) Phillips (1971)

Protohaustorius deichmannae (GA) Microhedyle milaschewitchii (O) Eurydice pulchra (I) Eurydice affinis (I) Pectenogammarus plancrurus (GA) Scolelepsis fuliginosa (P) Callianassa jamaicense louisianesis (S) Callianassa islandgrande (S)

6·0

1·5

0·2

“Large circular tank”

“Crystallizing dishes”

“At equal intervals around perimeter”

15·2

2·5

10·2

30·0

7·5

15·0

32·0

16·0

TP

105

from the average fluid-discharge rate) was measured. Relevant aspects of the boundary-layer flow regime (e.g. the shear or boundary shear stress, see pp. 145–8) were quantified, relative to settlement, in only one published study to date, that of Crisp (1955) on barnacle cyprids. The experiments were conducted in clear glass tubing and the animals were stimulated to attach over a range of low shear (the change in velocity with distance above the surface), but were prevented from attachment beyond some threshold value. All the water flow compared with attachment or settlement observations mentioned above indicate the potential sensitivity of larvae to moving fluid and the likelihood of passive transport very close to the sea bed, although the limiting values of boundary-layer flow parameters for which this would occur have yet to be quantified for soft-substratum organisms (but see theoretical calculations of Butman, 1986a). LABORATORY STUDIES OF HABITAT SELECTION Laboratory choice experiments of settling larvae were pioneered by Wilson in an extensive series of substratum-selection experiments on Ophelia bicornis (Wilson, 1948, 1952, 1953a, b, 1954, 1955). The studies were done in small Petri dishes (3–9 cm in diameter), where larvae were allowed to choose between small piles (0·75–1·5 cm in diameter) of sediment separated by several centimetres. These are the smallestscale experiments conducted on active habitat selection. A similar experimental design was used by Wieser (1956), Gray (1966a, b, c, 1967a, 1968), Croker (1967), Gray & Johnson (1970), Hadl, Kothbauer, Peter & Wawra (1970), Wilson (1970a, b, 1977), and Klauser (1986) (see Table V). Similar-sized dishes were used in the studies of Wilson (1948), Meadows (1964a), Lewis (1968), and Sameoto (1969), with the various treatments separated into pie-shaped sections by narrow vertical barriers (e.g. glass slides), so they were essentially adjacent. Very small-scale experiments also appear to have been done in the studies of Jansson (1967a, b), but only the treatment dimension (0·5 cm3) is given in the paper. The experiments were done in a “simple alternative chamber” made of plastic tubing, where the treatment patches were placed in either end. In all the studies cited above, the entire experiment was conducted at the scale of centimetres: in containers 9 cm in diameter (except, perhaps, Jansson, 1967a, b), with maximum treatment dimensions of <7 cm, and maximum distances between treatments of 5 cm (Table V). In the remaining choice studies

106

CHERYL ANN BUTMAN

(Table V), containers and distances between treatments were of the order of tens of centimetres, except for one experiment (Williams, 1958) conducted in a relatively large chamber (a trough 243 cm long by 61 cm wide), with five adjacent sediment treatments. The studies by Meadows (1964b, c), Gray (1966b, 1971), Gray & Johnson (1970), Jones (1970), and Morgan (1970) were similar to the Wilson design, but in dishes of larger diameter (12–34 cm). Webb & Hill (1958), Phillips (1971) and J.P.Grassle (pers. comm.) conducted experiments in shallow, square or rectangular containers subdivided into equal-sized compartments which contained the sediment treatments. The studies of Teal (1958), Botero & Atema (1982), and McCann (1986) were done in aquaria. The sediment treatments were separated in the vertical by Jensen (1981); a dish was suspended 2 cm below the water surface in an aquarium and different sediment treatments were placed in the dish and on the bottom of the aquarium. Finally, Cuomo (1985) continuously supplied polychaete larvae to a sea-water table containing square dishes, 15 of each of two sediment treatments. This is the only study where larvae entered the treatment area via moving fluid, although the experiments were not designed to mimic any particular field flow regime. The direct choice studies generally have shown that the organisms preferentially settle or accumulate in sediment treatments that characterize their natural adult habitat. Some of the specific attractive elements of a particular sediment treatment have been reduced by experimentation. The “attractive factors” are, for example, the microorganism population on the sediment particles (Wilson, 1955) or the cement secreted by conspecific adults (Wilson, 1970a, b). Thus, the potential for organisms actively to select preferred habitats is established by these (except Cuomo, 1985) laboratory experiments in still water, but only over spatial scales of centimetres to tens of centimetres (and up to ≈ 200 cm in the case of Williams, 1958). Note, also, that most of the experiments were on meiofauna and postlarval macrofauna or epifauna. Choice experiments for settling larvae were limited to studies of only eight infaunal species and lobster larvae (Table V). LABORATORY STUDIES OF METAMORPHOSIS The first experiment to induce metamorphosis of a planktonic larva was conducted by Mortensen (1921) on echinoderm larvae, but again Wilson was responsible for much of the detailed work on infauna which immediately followed (e.g. Wilson, 1932, 1937, 1948, 1951, 1953a, b, 1954, 1955, 1958, 1968, 1970a, b; Day & Wilson, 1934). From these experiments, larvae competent (i.e. physiologically capable) of settlement were introduced into separate dishes containing various sediment and water-column treatments. After a period of time, the dishes were scored for the number of metamorphosed larvae. While these kinds of experiments do not directly demonstrate habitat selection, they are useful for resolving the specific components of the attractive factors. For example, using dialysis membranes, Highsmith (1982) and Suer & Phillips (1983) determined the approximate molecular weight of the “scent” promoting metamorphosis in a sand dollar and an echiuran worm, respectively, allowing further characterization of the chemical nature of the substance. Most factors that promote metamorphosis are organic and often they are species-specific (Burke, 1983; Crisp, 1984). Cuomo (1985), however, recently showed that dissolved hydrogen sulphide originating from sediments or in the water column strongly promotes settlement and metamorphosis of Capitella sp. I larvae. Many organisms which metamorphose in response to a particular treatment will delay metamorphosis in the absence of that factor (see especially the reviews of Strathmann, 1978, and Crisp, 1984). Crisp (1984) attributed a revolution in the way of thinking about the rôle of larval settlement in establishing benthic communities to the discovery of delayed metamorphosis because it indicates that certain species are not “forced” to settle in an inhospitable environment, but have time to search for a preferred habitat. Since

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

107

research in invertebrate zoology in the 1930s was primarily on embryology and development, Crisp (1984) noted that the importance of temperature- and time-dependent processes overshadowed the potential rôle of external factors in the developmental sequence. Thus, the discovery of delayed metamorphosis invoked the participation of the environment, and of the ecologist, in the process of larval settlement. Now, however, we may have come full-circle, as the rôle of developmental constraints on the duration of pelagic life is again being stressed. Pechenik (1980) suggested that a limit to the delay period for dispersing larvae may be programmed into development so that larvae would be capable, for example, of delaying metamorphosis for a longer time in cold water. Subsequent experiments on the relationship between development, metamorphosis and temperature (e.g. Jackson & Strathmann, 1981; Pechenik, 1984; Lima & Pechenik, 1985) lend support to this hypothesis. In addition, there are reports of deterioration or regression of the larva (Henderson & Lucas, 1971; Caldwell, 1972; Yamaguchi, 1974) and a decline in selectivity (e.g. Scheltema, 1961; Gray, 1967a; Caldwell, 1972; Grassle, 1980) over the delay period and some species metamorphose in the water column (Day, 1937; Thorson, 1946; Rasmussen, 1956; Baggerman, 1953; Sarvala, 1971; Lacalli, 1980; Peckenik, 1980; Levin & Greenblatt, 1981). Recently Kempf & Hadfield (1985) found, however, that the lecithotrophic larvae of a nudibranch will eventually feed in the plankton if they are deprived of a settling stimulus for a sufficiently long time; thus, they extend their competent period and enhance the probability of encountering a suitable settlement cue. Furthermore, Richmond (1985) has shown complete reversible metamorphosis in the planula larva of a coral species in response to disturbance, with subsequent resettlement and successful metamorphosis. SITE PERUSAL, CUE DETECTION, AND SITE SELECTION The procedure used by a larva to select a habitat and the method of cue detection are still largely the subject of speculation for infaunal organisms. There has been considerably more research on the chemosensory response for hard-substratum (and especially, fouling) organisms (Crisp, 1984); even so, Burke (1983) concluded that there is still only “circumstantial evidence” for the involvement of particular sensory structures in perceiving cues which induce metamorphosis, and that while a neurological and behavioural model of stimulus and response can be ascribed to the induction of metamorphosis, there is still no substantive information on how various neural and endocrine mechanisms actually control the metamorphic sequence. Crisp & Meadows (1963) coined the phrase “tactile chemical sense” to describe the process of chemoreception in settling barnacles, where the cyprid must make direct contact with the chemically treated surface to detect the cue; presumably an organ in the antennules is the site of the chemoreceptors in this group (e.g. Gibson & Nott, 1971). A tactile requirement for perception of the chemical cue is also supported by most of the data available for infaunal and epifaunal larvae. Observations of the behaviour of the organism during settlement generally indicate that the larva must contact the surface to perceive the cue (e.g. Wilson, 1968; Caldwell, 1972; Eckelbarger, 1978; Rice, 1978); some animals actually burrow into the substratum, without metamorphosing, and then swim away (e.g. Wilson, 1955; Rice, 1978). Eckelbarger (1978) discusses the potential sensory function of anterior ciliary tufts on sabellariid polychaetes just prior to settlement, during the searching stage. The results of Suer & Phillips (1983) directly support the tactile chemical sense in an infaunal organism (the echiuran, Urechis caupo) because the chemical factor promoting metamorphosis was effective only if it was absorbed onto a surface. The chemotaxis hypothesis (here meaning movement toward or away from a waterborne cue) has received only intermittent support through the years. As Crisp (1974) reiterated, dilution of the cue in the water as it diffused and was mixed by flow turbulence above the bed, is an obvious problem with this

108

CHERYL ANN BUTMAN

hypothesis. Organisms would have to perceive the relatively undiluted cue while they were practically sitting on the sea bed (e.g. see Crisp & Meadows, 1962; Butman, 1986a) to settle in close proximity to the source; otherwise, once they perceive the diluted cue at some distance away from the bed, settlement could occur over a broad region of the bottom which may or may not contain the source of the cue. Reports of organisms responding to cues “at a distance” from the bed are rare (Crisp, 1974), although much more research is needed. Most of the evidence for the chemotaxis hypothesis is circumstantial: the larvae metamorphosed when a substance was added to the water (Crisp, 1974). To be direct evidence of chemotaxis, it must be shown that the substance added did not adhere to a surface (e.g. the walls of the container) that the larva then tested; for example see the criticisms of Scheltema’s (1961) study in Crisp (1974) and Scheltema (1974). The chemotaxis hypothesis is supported by the recent experiments of Cuomo (1985), who suggested that a threshold concentration of hydrogen sulphide in the water above the sediment was responsible for eliciting the settlement response in a capitellid polychaete. The nuchal organ may be the site of chemoreception, as speculated by Bhup & Marsden (1982). In addition, Highsmith (1982) showed that surface textures or films were not involved in eliciting metamorphosis in the sand dollar, Dendraster excentricus, and suggested that the larvae can detect a concentration gradient of the inducer, released from the adult sand dollar bed. Further substantiation of a waterborne inducer for metamorphorsis in this species is given in Burke (1984), where > 90% metamorphosis occurred in aqueous extracts from water overlying the sand in which the adult pheromone was released, but only 5% metamorphosis occurred in extracts overlying sand outside the sand dollar bed. Like the direct choice studies and the metamorphosis experiments, the logistics of active habitat selection whether by the tactile chemical sense or by chemotaxis, are poorly explored for infaunal larvae settling in moving fluid; all experiments have been done in still water. If an entirely waterborne cue can elicit the response while the larva is still in the plankton, as the results of Highsmith (1982), Burke (1984), and Cuomo (1985) suggest, then it is particularly critical to do laboratory studies in simulated field flows. As mentioned earlier, by the time cues advected and mixed by flow turbulence are perceived by a planktonic larva, the organism may end up on the sea bed in a habitat from which the cue did not emanate (e.g. Cameron & Rumrill, 1982). Doyle (1975) proposed a settlement model for active habitat selection where the probability of a larva responding to a given cue in the water can be only zero or one, i.e., a threshold level of the stimulus evokes the response. This is an attractive theory, particularly if the competent larvae drift in water very close to (i.e. within centimetres of) the sea bed during the cue-detection stage, because it minimizes errors in site selection and requires a relatively simple behaviour response. While accurate site location would be improved if a larva could swim upstream along a cue concentration gradient, this possibility appears to be limited to very weak near-bed flow regimes, due to the relatively slow swimming capabilities of most infaunal larvae (Mileikovsky, 1973; Mann & Wolfe, 1983; Chia, Buckland-Nicks & Young, 1984) compared with velocities very close to the sea bed (Butman, 1986a). Even if cues must be adsorbed to a surface that the larva can test, as the bulk of the evidence to date suggests, test sites also may be specified by bottom boundary-layer flow conditions (Butman, 1986a; see later discussion, pp. 148–51). Thus, while larvae of many infaunal invertebrates are clearly capable of discriminating between microhabitats and metamorphosing in response to specific cues, the field conditions wherein active habitat selection actually determines patterns of recruitment are unknown.

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

109

FIELD EXPERIMENTS ABOVE OR ON THE SEA FLOOR Field studies on active habitat selection may have the advantage of being realistic from a fluid-dynamic point of view, but other aspects of field conditions are limiting to experimentation. For example, it is nearly impossible to sample initial settlement onto the sea bed and to measure subsequent early postlarval mortality in soft-substratum systems. Even in hard-substratum systems, the newly settled stages have been identified and followed for only a few species (Connell, 1985). Experimental studies in the field can be classified either as manipulations above the bottom or direct manipulations of the sea floor. The literature is reviewed with the primary emphasis on identifying the spatial scales involved in the experiments and on separating cases where habitat selection actually was demonstrated in the study from indirect or inconclusive evidence of selection. While experiments generally are required to determine processes, direct sampling of the unmanipulated sea floor has also provided useful information on patterns of recruitment and significant correlates, especially when sampling was both frequent (days to weeks) and rigorous (using appropriately small sieves to sample newly settled organisms) and when the water column was sampled simultaneously (e.g. Muus, 1966, 1973; Oliver, 1979; Hannan, 1980; Luckenbach, 1984; Webb, 1984). Results of the detailed field study of Muus (1966, 1973) on bivalve larval availability in the plankton (e.g. from Fosshagen, 1965, which overlapped with the first year of Muus’ study) and recruitment in two localities (at 18 m and 27 m), indicate that both habitat preferences in settling larvae and early postlarval mortality shape adult distributions. This is a somewhat unusual study in that the two field sites were separated by only 1 km and differed markedly in faunal composition, but differences in the bottom sediments (dominance by the 64 to 250-µm fraction at the 18-m site and dominance by the 64 to 125-µm fraction at the 27-m site) probably were not hydrodynamically meaningful (i.e. did not represent a large enough change in bed roughness to alter the structure of the near-bed flow; see pp. 148–54. These relatively small sediment differences certainly may be biologically meaningful. Muus acknowledged that observed patterns of recruitment imply active habitat selection only if the supply of larvae to the two sites was equivalent, but provides reasonable arguments, based on known circulation patterns in the Øresund, that “the same water masses and same larval swarms” probably pass over the two localities (Muus, 1973, p. 103). To avoid problems associated with direct sampling of the sea bed (e.g. processes operating at the sediment-water interface that may obscure initial settlement patterns), several manipulative field studies have been done in structures raised above the sea floor (Table VI). In nearly all cases, larval settlement differed among the various treatments deployed simultaneously and the authors concluded that larvae actively select their settlement sites. All field studies which compared collections in artificial structures with collections from the natural sea bed may, however, have suffered from “trapping artifacts” —physical, chemical, and biological differences between the micro-environment of the trap and the natural bottom— which complicate interpretation of the results (Oliver, 1979; Hannan, 1981). Unless the collection characteristics of the traps for passive inert particles (e.g. sediments) can be defined (Hargrave & Burns, 1979; Gardner, 1980; Butman, 1986b; Butman, Grant & Stolzenbach, 1986), collections resulting from biological processes (e.g. active habitat selection behaviours of the larvae) cannot be separated from collections resulting entirely from hydrodynamical processes. Hannan (1981) was unable to distinguish between these possibilities to account for the differences (orders of magnitude) in numbers of postlarvae collected in traps placed 1 m above the sea bed compared with those in cores of the natural bottom. Oliver (1979; some results are also reported in Dayton & Oliver, 1980) used relatively “tall” and “short” plastic cups filled with the same amount of sediment (to the rim of the short cup) to simulate physical conditions of deposition and resuspension, respectively. Relatively more “Capitella capitata” (the sibling species, sensu Grassle & Grassle, 1976, may be similar to Capitella sp. Ia, as in Hannan’s, 1981, study

110

CHERYL ANN BUTMAN

conducted nearby) were collected in the tall than in the short cups, while another polychaete species, Armandia brevis, was not differentially collected by the two trap designs. Oliver (1979) suggested that Capitella actively selected the depositional environment in the short traps, but that Armandia was less selective in its settlement requirements; both behaviours are consistent with the distributional patterns of the adults and the responses of the populations to disturbance (Oliver, Slattery, Hulberg & Nybakken, 1980). Passive accumulation of Capitella larvae or postlarvae in the depositional environment is, however, also consistent with the results. Differences between species in hydrodynamic properties (e.g. fall velocities), swimming abilities, and periods of larval availability relative to flow processes, could account for the different patterns of collection by the traps for the two species. In the only study (Hannan, 1984a, b) where hydrodynamical properties of larvae and collection characteristics of traps (Butman, 1986b) were defined in the laboratory before field deployments, nearly all of the abundant infaunal organisms (in three invertebrate phyla) were collected in the relative abundances predicted for passive particle collections by traps. It is valid to compare collections in different sediment treatments placed in the same type of artificial structure raised above the sea floor, when the treatments are exposed to the same flow regime (e.g. Oliver, 1979; Levin, 1981, 1984; Watzin, 1983, 1986), but considerations of possible “edge effects” and other differences between treatments due to position within the structure TABLE VI Field experiments on larval settlement or early recruitment of infauna: experimental studies where sampling intervals were 1 month or sieve screen sizes ≥1 mm are not included in this table because initial settlement or early recruitment are unlikely to be detected with these methods; distances and dimensions were estimated, when possible, from the information given in the paper; *=in studies where settlement into one type of box, tray or trap (usually containing defaunated sediment) was compared with settlement onto the natural sea bed, the distance between “treatments” would be between the structure and where the sea bed was sampled; this usually was not given in the study but probably is no greater than tens of centimetres or metres; NG=information not given in the paper; NA=not applicable to this study; OSO=one station only was sampled; P=polychaete; B=bivalve Reference

Experimental Organisms approach studied

Thorson (1946)

“Bottle collectors” moored above sea bed; plankton and bottom samples Manipulatio ns of the sea bed “Sediment bottle collectors” Different substrata placed in suspended

Baggerman (1953) Reish (1961) Hermans (1964)

Minimum sampling interval

Sieve screen Minimum size (µm) distance between stations (m)

Maximum treatment dimension (cm)

Maximum distance between treatments (cm)

All infauna

6 wk for bottles, 2 wk for plankton

NG for bottles, 83 for plankton

NG

Bottle dimensions not given

NG*

Bivalves

12 h

500

OSO

45

NG

All infauna

28 days

246

710

NG

2 wk

NG

OSO

“Gallon jar” dimensions not given “Thorson” bottles, dimensions not given

Armandia brevis (P)

NG

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

Reference

Richter & Sarnthein (1977), but technical layout in Sarnthein & Richter (1974) Grassle & Grassle (1974)

Guérin & Massé (1978), Massé & Guérin (1978)

McCall (1977)

VanBlarico m (1978)

Experimental Organisms approach studied

bottle collectors Different substrata placed in trays moored above sea floor Boxes of defaunated sediment made flush with the sea bed; bottom samples Different substrata placed in three designs of collectors moored on the bottom or above the bottom Boxes of defaunated sediment placed on the sea bed; bottom samples Containers with thin layer of sediment moored above the sea bed; bottom samples

Minimum sampling interval

Sieve screen Minimum size (µm) distance between stations (m)

Maximum treatment dimension (cm)

Maximum distance between treatments (cm)

Molluscs

2 wk

63

80

71

118

All infauna

3 days

297

4600

100

NG*

Polychaetes and molluscs

1 month

1000 “on diagonal” (=700 if mesh is square)

NG

8·6

180

All infauna

10 days

297

9000

3700

NG*

All infauna

13 days for containers, 1 month for bottom

250 for containers, 500 for bottom

OSO

10 for containers

NG*

111

112

CHERYL ANN BUTMAN

Eckman (1979) Oliver (1979), Dayton & Oliver (1980)

Santos & Simon (1980a)

Williams (1980) Bhaud, Aubin & Duhamel (1981)

Hannan (1981)

Levin (1981)

Manipulatio ns of the sea bed (a) different treatments filled containers held in racks above sea bed; bottom samples (b) manipulatio ns of the sea bed Containers filled with sediment placed in rack above sea bed; plankton and bottom samples Manipulatio ns of the sea bed Collectors filled with sediment placed above sea bed; bottom samples with epibenthic sledge Collectors with thin layer of sediment moored above sea bed; bottom samples Different sediment treatments filled containers

All abundant infauna Polychaetes

11 days

61

OSO

100

700

(a) 6 days

(a) 250 for containers, 500 (250 for “a few cores”) for bottom

(a) OSO

(a) 14 for containers

(a) 42 for containers

(b) 1 month

(b) 500 (250 for “a few cores”)

(b) 1000

(b) 2000

(b) NG

All infauna

7 days for containers, “irregularly ” for plankton, 1 month for bottom

OSO

250 for containers, 144 for plankton, 500 for bottom

5 for containers

NG*

Tapes japonica (B) Polychaetes and bivalves

7 days

149

OSO

150

750

5 days

NG

OSO

“2 litre capacity” collectors, dimensions not given

NG*

Armandia brevis (P), Capitella spp. (P), Nothria elegans (P), Prionospio pygmaea (P) Streblospio benedicti (P) Pseudopoly dora

7 days

250

400

10

NG*

2 wk

Worms were visually counted under

OSO

9

26

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

VanBlarico m (1982) Zajac & Whitlatch (1982a)

Eckman (1983) Gallagher, Jumars & Trueblood (1983) Watzin (1983, 1986)

Hannan (1984a, b)

placed directly on sea bed; bottom samples Manipulatio ns of the sea bed Buckets of defaunated sediment made flush or protruding above the sea bed; bottom samples Manipulatio ns of the sea bed Manipulatio ns of the sea bed

paucibranc hiata (P)

Different sediment treatments filled containers held in racks raised above sea bed Different sediment trap designs, with known passive particle collection characteristi cs, moored

dissecting microscope

All infauna

7 days

500

OSO

30

NG

All infauna

14 days

297

620

48

NG*

All abundant infauna All abundant infauna

2 days

61

OSO

30

1970

2 days

63

OSO

3·7

500

All infauna

7 days

63

OSO

14

50

All abundant infauna

1 day

100

OSO

14·7

2400

113

114

CHERYL ANN BUTMAN

Reference

Levin (1984)

Luckenbach (1984)

Bonsdorff & Österman (1985)

Whitlatch & Zajac (1985)

Woodin (1985)

Experimental Organisms approach studied

above sea bed Manipulatio ns of the sea bed; containers filled with sediment placed on sea bed; plankton samples Sampled four areas of sea bed representing natural sediment treatments; plankton and bottom samples for initial availability Trays filled with sediment placed on sea bed; bottom samples Different sediment treatments in cores held in racks above sea bed Different sediment treatments in cores implanted in bottom

Minimum sampling interval

Sieve screen Minimum size (µm) distance between stations (m)

Maximum treatment dimension (cm)

Maximum distance between treatments (cm)

Polychaetes

3 days for manipulated sediments, 1 month otherwise

250 for containers and sediments, 63 for plankton

100

63 for bottom, 9 for containers

100

Mulinia lateralis (B)

1 day for initial availability; 4 days for initial settlement; 1 month for recruitment

105

OSO

10

NG*

All infauna

2 wk

500 for macrofauna ; 500, 200, 63 for meiofauna

OSO

40

NG*

All infauna

10 days

180, 300

OSO

5

100

Spionid polychaetes

9 days

250

OSO

11

1500

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

115

must be carefully analysed (see Nowell & Jumars, 1984). While it may not be possible to define the hydrodynamic conditions above these sediment treatments, as long as conditions are constant among treatments, between-treatment differences in settlement or recruitment can be assessed. Results of the field manipulations above the sea floor are strongly suggestive of active habitat selection by many infaunal larvae or postlarvae on scales of tens of centimetres to metres (Table VI), with the caveat that hydrodynamic alternative hypotheses usually were not considered or tested. Data from many of these manipulative field studies will remain equivocal until the possibility of differential passive deposition or accumulation between structures or between structures and the bottom can be discounted. The strongest results are for studies where hydrodynamic conditions were held constant among treatments, although the physical characteristics of the flows over these sediments are undefined. Directly manipulating bottom sediments to test the active habitat selection hypothesis alleviates the problems with structures. If the bottom roughness scales remain unchanged between manipulated and control sediments, then flow characteristics should be similar for all treatments (see Nowell & Jumars, 1984). The results of such studies (e.g. Oliver, 1979; Williams, 1980; Gallagher, Jumars & Trueblood, 1983) show that recruited postlarvae are associated with distinct habitats on the scale of metres (Table VI). Because of problems in sampling initially settled larvae and subsequent early mortality, mentioned above, it is not clear if the pattern results from active selection when the larvae first reach the sea floor or from a re-distribution of the postlarvae after initial settlement; it is also possible that observed distributions resulted from very early postlarval mortality of settled larvae that were originally evenly distributed among the sediment treatments. In cases where bottom sediments were manipulated specifically to change the nature of the nearbed flow regime, with accompanying a priori predictions of hydrodynamic effects on recruiting infaunal postlarvae or meiofauna (Eckman, 1979, 1983; Hogue & Miller, 1981), the hydrodynamic null hypotheses could not be falsified (see pp. 141–5). PATTERN OF DISTRIBUTION AND ACTIVE HABITAT SELECTION: A PROBLEM OF SPATIAL SCALES The spatial scales (centimetres to tens of centimetres) for which active habitat selection has been conclusively demonstrated in laboratory experiments in still water (Table V) are one to six orders of magnitude smaller than the spatial scales (tens of metres to tens of kilometres) over which species and sediment composition are significantly correlated in the field (see Table III, pp. 120–1). Thus, the process of active habitat selection, as demonstrated by these laboratory results, cannot account for the observed field distributions, due to the mismatch in spatial scales. Field experiments on processes controlling larval settlement were conducted at spatial scales of tens of centimetres to tens of metres (Table VI); while active habitat selection was strongly implied by the results of many of these studies, this interpretation remains equivocal because the alternative hypothesis of passive deposition was usually neither considered nor tested. When patterns of community composition and structure have been delimited at small spatial scales, e.g. of the order of 1 to 10 m in Jones (1962), of 0·1 to 1 m in Angel & Angel (1967) and Grassle et al. (1975), of 0·01 to 1 m in Reise (1979), of 100 cm in Gärdefors & Orrhage (1968) and Jumars (1976), of 10 m in Gage & Geekie (1973a), and of 10 cm in Olsson & Eriksson (1974), sediment samples were not taken at each infaunal sampling location, except in one case (Angel & Angel, 1967). The entire area sampled in these small-scale dispersion studies was usually considered homogeneous in its bulk sediment characteristics, based on one to a few sediment samples from the area. Thus, the spatial patterns and scales of diversity

116

CHERYL ANN BUTMAN

detected in these studies were usually attributed to processes other than those directly related to bulk properties of sediments. The small-scale patterns of species distribution and diversity detected in these studies are the only patterns to which results of habitat selection experiments can be applied directly. Even though bulk properties of sediments were presumed to be constant within the areas sampled in these studies, Angel & Angel (1967) and Jumars (1976) briefly discussed the potential importance of small-scale variability in sediment characteristics; Jumars & Eckman (1983) provide a more detailed discussion of this topic. Local heterogeneity in sediment topography (e.g. due to geological or biological processes) can cause significant small-scale (centimetres to metres) variations in sediment grain size because of local changes in the nearbottom flow regime; patchiness in infaunal distributions at these small scales can be attributed to this sediment heterogeneity (e.g. Rhoads & Young, 1970; Eckman, 1979, 1983; Thistle, 1983). In addition, detailed analyses of sediment characteristics using micro-scopic methods and staining techniques (e.g. Whitlatch & Johnson, 1974) indicate that bulk sediment analyses obscure variation in sediment properties (e.g. protein, carbohydrate, and lipid contents, as well as grain size) to which organisms may respond (Whitlatch, 1974, 1980). Many laboratory studies of habitat selection have demonstrated that there are chemical and biological substances (e.g. chemical conditioning of sediments by adults or the abundance and composition of bacterial populations) in sediments which augment grain size as attractive factors to stimulate or enhance larval settlement. Thus, within an area of homogeneous sediment type (based on analysis of grain size), larvae may actively select for microhabitats based on these other aspects of sediments. For example, Thistle, Reidenauer, Findlay & Waldo (1984) and Eckman (1985) have shown that there is local enhancement of bacterial abundances around vertical protrusions (seagrass shoots or animal tubes) from the sea bed and that infauna are concentrated in these regions. In summary, larvae may select for microhabitats at small spatial scales (centimetres to tens of centimetres) based on sediment characteristics other than just grain size (as determined from bulk sediment analyses). The capability of larvae to distinguish between and actively select for habitats with distinctly different grain sizes and separated by large distances (tens of metres to tens of kilometres) is yet to be demonstrated. The passive deposition hypothesis may resolve this problem because it specifies that larvae are deposited at the same spatial scales as apply to sediment transport and deposition (see p. 144). At this time, passive deposition of larvae represents the simplest and most feasible mechanism for creating initial large-scale distributions of larvae in the field. Active habitat selection may be confined to only very small spatial scales. THE PASSIVE DEPOSITION HYPOTHESIS To my knowledge, the passive deposition hypothesis was first formally proposed to account for patterns of initial larval settlement or recruitment of infaunal species in the studies of Baggerman (1953) and Pratt (1953). Prior to these, brief, qualitative discussions of the rôle, or potential rôle, of “currents” in controlling larval dispersal and in determining settlement sites were given in Orton (1937), Kreger (1940), Thorson (1946, 1950), and Verwey (1952). For hard-substratum habitats, experiments on the rôle of hydrodynamical processes in settlement occurred much earlier. Observations and experiments on flows which permit or inhibit settlement of fouling organisms date from the 1940s (McDougall, 1943; Smith, 1946; Doochin & Smith, 1951; Crisp, 1955; Wood, 1955) due, at least in part, to the important applied aspects of this problem (i.e. the commercial need for developing methods to inhibit biofouling). Likewise, probably the most extensive studies, to date, of the rôles of both biological and physical processes in the dispersal and settlement of any single species were done on barnacles (Bousfield, 1955; de Wolf, 1973).

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

117

Strictly speaking, the passive deposition hypothesis stipulates that competent planktonic larvae initially reach the sea floor at sites where passively sinking particulates, with fall velocities similar to larvae, initially settle (Hannan, 1984a, b). As indicated in Hannan (1984b), this hypothesis does not specify that the deposited organisms will accumulate at these locales, as the geological definition of “deposits” implies, but refers only to the process controlling where the larvae will initially come to rest on the sea bed. Then, other biological or physical processes may re-distribute the organisms (see later discussion, pp. 154–5). Note also that, “deposited larvae may or may not have ‘settled’ according to the biological definition of Scheltema” (Hannan, 1984b, p. 1109). The passive deposition hypothesis has never been tested directly because it requires simultaneous sampling of initially deposited larvae and passive particles with fall velocities similar to larvae. This eventually may be possible in a laboratory flume, where realistic field flow regimes could be simulated (see Nowell & Jumars, 1987), and the distributions of inert particles with known fall velocities could be compared with the distributions of larvae or postlarvae when they first reach the bottom. The chances of testing the passive deposition hypothesis in the field seem remote, due to problems in actually sampling initial distributions of larvae and particles prior to interference by benthic biological and physical processes, and to problems of defining the fall velocities of initially settled particulates in their naturally occurring states (e.g. flocculated or biologically aggregated). Support for the passive deposition hypothesis comes from studies of passive accumulation, passive sinking, and passive resuspension and transport of larvae, postlarvae or meiofauna, which are discussed separately below. It is important to distinguish the passive deposition hypothesis from the earlier notion that larvae fall in a “random rain” onto the sea bed. which was once considered the alternative hypothesis to active habitat selection (e.g. see discussion of these early ideas in Thorson, 1957). Random deposition explicitly states that there is an equal probability that individual larvae will fall onto any bed location. This hypothesis is synonymous with the passive deposition hypothesis only for a homogeneous suspension of larvae and particles falling through still water. In moving water, for an infinite water mass with a uniform particle supply distributed homogeneously in the water column, and with a steady and non-varying physical regime, the initial distribution of particles on the sea bed would be random. For temporally and/or spatially varying flow regimes, particle abundances, and particle distributions in the water column, the particles will not, however, fall at random onto the sea bed. In these cases, the sites for initial settlement of particles are determined by the hydrodynamical processes and the particle characteristics. Thus, for the physical regimes of interest in most marine studies, a random rain of larvae to the sea bed is not the appropriate null hypothesis for testing the importance of physical processes, since particle deposition is not expected to be random. In fact, a random pattern of initial larval settlement would, in most cases, falsify the passive deposition hypothesis. If larvae physically behave in a flow like passive particles, then it is their fall velocity and hydrodynamical processes which determine when and where the larvae will reach the sea bed. Thus, passive deposition is the appropriate physical null hypothesis against which biological (i.e. active habitat selection) hypotheses can be tested. PASSIVE ACCUMULATION The correspondence of distributional patterns of a cockle (Baggerman, 1953), two species of bivalve (Pratt, 1953), and several echinoderm species (Tyler & Banner, 1977) with modern fine-sediment distributions was attributed to the passive accumulation of settling larvae and fine sediments in similar locales. Illuminated by discussions with Baggerman, Verwey (1952) and Kristensen (1957) suggested that some bivalve populations may result from passive accumulation in “sheltered” or “weak current” areas. Orton (1937), Segerstråle (1960, but see also 1962), and Carriker (1961) also indicated the potential importance of strong

118

CHERYL ANN BUTMAN

near-bed currents maintaining larvae in suspension (e.g. in gyres or swift tidal channels) and weak currents allowing the spat to settle onto the bed. Fager (1964) attributed the presence of an unusually shallow, very dense, and oddly shaped (elliptical, with the long axis parallel to shore) bed of the polychaete, Owenia fusiformis, to such physical processes. He suggested that larval settlement was concentrated in this locale due to the coincidence of a water mass containing large numbers of competent larvae with a rip current at the site. The hydrodynamics associated with the rip would allow larvae to accumulate passively in the unusual bed configuration. On a much smaller scale, Birkeland & Chia (1971) suggested that early recruitment of sand dollars may be more successful in patches of sand within a cobble field compared with a sand flat, because the cobbles act like “break-waters” for the flow over the sand; it is, however, unclear if the authors were implying passive deposition or enhanced retention in these slow-flow regions. In all of these studies, evidence of the rôle of hydrodynamical processes is largely correlative, but the novelty of these interpretations when other, similar, correlative evidence invoked the active habitat selection hypothesis (see pp. 123–7), is striking. The manipulative field experiments of Baggerman (1953), Eckman (1979, 1983), and Hogue & Miller (1981) provide substantive support for the passive accumulation hypothesis. Baggerman (1953) placed vertical barriers (screens) on the sea bed and sampled for cockle spat near and away from the screens. She also determined that a range in sizes of cockle spat were likely to be transported and deposited like fine sediments by showing that measured gravitational fall velocities of spat were similar to the measured fall velocities of the sediments transported at the study sites. She did not, however, determine, a priori, how the vertical screens would affect the near-bottom flow at her study sites; she assumed that the region of low flow developing in the lee of the screens would be sufficient to trap sediments and passively falling spat. Eckman (1983) made specific a priori hypotheses on how the artificial tubes he placed in sediments would affect both the fluid flux to the bed and the boundary shear stress because he did laboratory flume experiments to measure these physical effects. In another study, Eckman (1979) placed artificial tubes at regular intervals in sediments and, taking contiguous samples over the area, determined the spatial scales of organism distributions and compared them with the spatial scale of the physical effects resulting from this manipulation. Hogue & Miller (1981) repeated Eckman’s (1979) experiments in a different intertidal area, but studied dispersion patterns of nematodes, rather than recruitment of infauna. In all these studies, recruitment patterns were consistent with predictions based on hydrodynamical criteria; that is, the null hypothesis of passive accumulation could not be falsified. Indirect support for passive deposition and accumulation comes from the numerous reports of higher postlarval or adult infaunal abundances in depressions on the sea floor (e.g. Chapman & Newell, 1949; Pratt, 1953; Pamatmat, 1968; Sameoto, 1969; Howard & Dörjes, 1972; Farke, de Wilde & Berghuis, 1979; VanBlaricom, 1982; McLusky, Anderson & Wolfe-Murphy, 1983; Levin, 1984) or in seagrass beds that baffle water motion (e.g. Orth, 1977; Scheibling, 1980; Peterson, Summerson & Duncan, 1984) than in adjacent sandflats. The pattern of distribution for newly settled larvae has, however, yet to be measured. These patterns of enhanced abundances in areas of relatively slow flow need not arise at the time of settlement, but may result from differential post-settlement mortality. To determine at what stage in the life history the pattern of enhanced abundances of the hard clam, Mercenaria mercenaria, in seagrass beds compared with adjacent sandflats is established, Peterson (1986) computed the ratio of organism densities between the two habitats for the 0-year class and all subsequent year classes. Because these ratios were ) for the older year classes, Peterson (1986) concluded that post-settlement considerably larger (by phenomena, such as competition and predation, were at least as important as settlement phenomena in creating the pattern.

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

119

For all the passive accumulation studies, a problem with unambiguously interpreting the process responsible for the observed pattern of enhanced recruitment in regions of slow flow is that fine sediment and detritus also will accumulate in these areas. Thus, the alternative hypothesis that larvae actively select sites where fine sediments and detritus accumulate or preferentially survive in these areas cannot be discounted. PASSIVE SINKING Organisms are unlikely to be passively deposited onto the sea floor unless they sink through near-bottom waters like passive particles. Hannan (1984a, b) tested this passive sinking hypothesis for larvae falling through turbulent field flows using several groups of geometrically different sediment trap designs (see Table VI, p. 138). A priori predictions regarding the rank order that the various traps would collect larvae in the field were dictated from laboratory flume experiments to determine particle collection efficiencies of the traps in flows dynamically similar to average conditions at the field site studied. The flume flow was seeded with particles having fall velocities similar to those measured in the laboratory for non-swimming polychaete larvae. In these experiments, nearly all of the abundant organisms (polychaete, bivalve, and enteropneust postlarvae) were collected by traps in the patterns predicted for passive particle collections. Thus, the passive sinking hypothesis could not be falsified. PASSIVE RESUSPENSION AND TRANSPORT Indirect evidence that organisms living at the sediment surface may be resuspended and transported comes from studies where the water column and the sea bed were sampled simultaneously, or where the bottom was sampled intensively, throughout storm events (Hagerman & Rieger, 1981; Hogue, 1982; Dobbs & Vozarik, 1983); organisms either were missing from the sea bed or were present in the water column during the storms. The sampling studies of Bell & Sherman (1980) and Palmer & Brandt (1981) suggested that even tidal velocities may be sufficient to resuspend and transport meiofauna (but see also Grant, 1981). Palmer & Gust (1985) quantified this effect by measuring the bottom shear stress over a tidal cycle, when simultaneous water column and bottom samples also were collected. The a priori hypothesis was that meiofauna would be resuspended with the surface sediments only when the bottom shear velocity exceeded the critical erosion velocity for the sediments. They found that organism abundances in the water directly above (within tens of centimetres of) the sea bed were highest when the bottom shear velocity exceeded the threshold value. Furthermore, from laboratory experiments, Palmer (1984) showed that the organisms probably were not actively entering the water, although certain behaviours (i.e. remaining at the sediment surface rather than burrowing) increased a given organism’s probability of being resuspended. Indirect support for passive resuspension and transport of surface- or near surface-dwelling infauna comes from the numerous reports of postlarval and adult organisms in the water column (see Table II, pp. 117–9) and of post-settlement migrations (e.g. Chapman & Newell, 1949; Baggerman, 1953; Kristensen, 1957; Sigurdsson, Titman & Davies, 1976; Farke et al., 1979). SUMMARY The hypotheses that hydrodynamical processes determine accumulation, sinking or resuspension and transport of larvae, postlarvae or meiofauna could not be falsified in the experimental studies conducted thus far. These results provide support for the passive deposition hypothesis, but direct tests for initially

120

CHERYL ANN BUTMAN

settled larvae are lacking. A limitation to interpreting results from the passive accumulation experiments is that fine sediments and detritus tend to accumulate in regions of slow flow so that the observed enhanced abundances of organisms in these areas, presumed to be the result of passive accumulation, could also result from active habitat selection for these detritalrich zones or from enhanced early postlarval survival. Experimental manipulations are needed to distinguish among these possibilities. LARVAL SETTLEMENT IN THE BOTTOM BOUNDARY LAYER Results of the passive accumulation, sinking, and resuspension and transport studies stipulate that physical processes cannot be discounted in considerations of larval settlement phenomena, so it is worthwhile to discuss briefly characteristics of the bottom boundary-layer flow environment, where settlement takes place. The near-bed flow regime determines the spatial scales applicable to passive deposition and also the hydrodynamical constraints for successful active habitat selection. Other discussions of bottom boundarylayer processes relevant to benthic ecology, and written for a general audience, can be found in Wimbush (1976), Vogel (1981), Nowell (1983), Nowell & Jumars (1984), and Butman (1986a); the recent review of Grant & Madsen (1986), written primarily for fluid dynamicists, summarizes many important aspects of boundary-layer flows in continential-shelf environments. The following discussion is limited to steady, uniform (in the horizontal) flow over a bottom which is also uniform over large horizontal distances, relative to the height off the bottom. The purpose is to provide some basic fluid-dynamical perspective on larval settlement, while retaining the essential physics. GENERAL FEATURES OF BOUNDARY-LAYER FLOWS OVER SOFT SUBSTRATA where u is the As water flows over the sea bed, a region of shear (the slope of the velocity profile, horizontal velocity component and z is the perpendicular distance from the bed; see Fig. 1) develops as a result of the retarding effect (drag) of the boundary on the flow. This region of shear near the bed is called the boundary layer. Within the boundary layer, current speed goes from zero at the bed to the mean-stream the boundary-layer thickness). For heights velocity (U) at the top of the boundary layer (where the bottom no longer has a significant effect on the flow; this is called the region of exceeding potential or frictionless flow and, in the absence of other flow processes (e.g. surface wind stress or other in this region. When the shear near the bed is sources of flow turbulence) and for a constant density, sufficiently large, turbulent eddies are generated that mix lower-momentum fluid close to the bed with higher-momentum fluid away from the bed; this thickens the boundary layer and reduces the mean velocities at a given height above the bed (especially close to the bottom). The shape of the velocity profile in the boundary layer depends on flow properties (e.g. the flow Reynolds number, the background turbulence and accelerations), fluid properties (e.g. stratification induced by temperature, salinity and suspended sediment), and boundary characteristics (e.g. the bed roughness and the cohesiveness of sediments). Velocity profiles have been measured for controlled laboratory flows and their characteristics have been determined theoretically under certain conditions. For the steady, uniform flow case considered here, two shapes of the velocity profile are well known, a parabolic shape for laminar boundary layers and a logarithmic shape for turbulent boundary layers. The boundary layer will be laminar or turbulent, depending on the flow Reynolds number, a dimensionless parameter which is the ratio of inertial forces to viscous forces in the flow. The Reynolds number (VL/v) depends on a length (L) and a velocity (V) scale for the flow, as well as on the fluid

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

121

Fig. 1.—Diagram of a turbulent boundary layer plotted on a linear scale for both axes, showing the relative positions of the viscous sublayer, the log layer, and the log-deficit layer: taken from Butman (1986a).

kinematic viscosity (v). Laminar boundary layers occur at low Reynolds numbers; molecular viscosity dominates as inertial forces are relatively unimportant for these conditions. Laminar boundary layers are very stable in the downstream direction; any disturbance to the layer (caused by flow over a bump, for example) will be quickly dissipated by viscosity, restoring the velocity profile to the undisturbed state. Thus, in laminar boundary layers, the flow is parallel to the bottom. Turbulent boundary layers occur at high Reynolds numbers and thus inertial forces (or turbulence) dominate over molecular viscosity. The velocity is composed of a mean component plus a fluctuating (turbulent) component. Transfer of mass and momentum within the layer is caused by these turbulent eddies. While the time-averaged flow velocity is in the horizontal, as in the laminar case, turbulent eddies have velocity components in all directions. Descriptions of laminar and turbulent boundary layers can be found in Clauser (1956), Schlichting (1979), and Yaglom (1979); features most relevant to problems in benthic ecology are indicated in Nowell & Jumars (1984). Laminar boundary layers are rare in the ocean, so that subsequent discussion will be for the turbulent case. Turbulent flows are classified as smooth, rough, or transitional (e.g. Schlichting, 1979), depending on the which is, again, a dimensionless ratio of inertial to viscous roughness Reynolds number (Re*= forces in the flow, but in this case it depends on the shear in the flow (u*, the bottom shear velocity, which where µ is the molecular viscosity of the fluid) and on the physical bed roughness (kb), as well is as on kinematic viscosity. In the immediate vicinity of the bottom, molecular viscosity is primarily responsible for dissipating flow energy. Outside the viscous sublayer, turbulent eddies mechanically dissipate flow energy as they break down into smaller and smaller eddies until, ultimately, energy is again dissipated by viscosity. A pronounced viscous sublayer (see Fig. 1) may develop in the case of flow over hydrodynamically smooth bottoms occurring at low Re* (e.g. Eckelmann, 1974). Over hydrodynamically rough bottoms (high Re*), viscosity still acts at the boundary, but no distinct well-behaved sublayer forms and eddies may penetrate to within tenths of a millimetre of the bed; thus, in rough-turbulent flow, the velocity structure close to the bed is complicated (e.g. Nowell & Church, 1979) and not well known. For intermediate Re*, transitional flow occurs, with characteristics intermediate between smooth-and roughturbulent. In the field, smooth-turbulent profiles have been measured by Chriss & Caldwell (1982) and W.D.Grant (pers. comm., see Butman, 1986a) and rough-turbulent profiles by Smith & McLean (1977), Cacchione & Drake (1982), Gross & Nowell (1983), and Grant, Williams & Glenn (1984). At a given site, the flow can be smooth-turbulent under one flow condition and rough-turbulent under another, for example, due to changes in bed roughness by rippling during storms or by bioturbation (see Grant & Madsen, 1986).

122

CHERYL ANN BUTMAN

Based on empirical studies and scaling arguments (Clauser, 1956), turbulent boundary layers can be divided into three regions (Fig. 1). Adjacent to the boundary, in the viscous sublayer, velocity (u) varies linearly with distance from the bottom. Above this, u varies with ln z in what is known as the log layer. The region farthest from the boundary is known as the log-deficit layer because the deficit velocity (U— u) varies with In z. The remainder of this discussion will focus on the qualitative and quantitative features of the log layer and the viscous sublayer, because their characteristics are relatively well known (e.g. Clauser, 1956; Yaglom, 1979; Nowell, 1983; Grant & Madsen, 1986) and they are the regions most relevant to larval settlement. The total thickness of the bottom boundary layer depends on the bottom shear velocity (u*) and inversely on the forcing frequency for the flow. On the continental shelf, at a latitude of 40°, for a flow periodicity stipulated by the Coriolis force, a u* of about 1 cm·s−1, and in the absence of stratification, the bottom boundary layer would be about 40 m thick (Grant & Madsen, 1986). The boundary layer grows all the way to the water surface in the smooth-turbulent, tidally driven flows at 10-m depth in Buzzards Bay, Massachusetts (U.S.A.), and for u* between 0.4 and 0.6 cm·s−1 (flow speeds of about 10 to 15 cm·s−1 at z=50 cm); the boundary layer fills half the water column for u*=0·2 cm·s−1 (a flow speed of about 5 cm·s−1 at z=50 cm) (Butman, 1986a). Boundary layers resulting from forcing due to surface waves are very thin (centimetres to tens of centimetres), however, because of the high-frequency nature of these flows (Grant & Madsen, 1986). In the field, the log layer is known to be about 10–15% of the total boundary layer (Clauser, 1956; Nowell & Church, 1979; Grant & Madsen, 1986), so the log-layer thickness varies between centimetres (wave boundary layer) to about a metre (tidal boundary layer) to several metres (planetary boundary layer), in the examples above. For smooth-turbulent flows, the viscous sublayer can be estimated by 10 v/u*; for u* between 0·1 and 1·0 cm·s−1 (typical values for smooth-turbulent flow) and v=0·01 cm2·s−1, the viscous sublayer thickness will be from 0·1 to 1·0 cm. In summary, a larva beginning its descent through the water column in the region of potential flow will and then will experience a sheared flow, experience a constant horizontal velocity until it reaches where the velocity decreases approaching the bed. At some distance close to the bottom, the horizontal velocity becomes vanishingly small (since u=0 at the sea bed), so the organism would be free to manoeuvre in basically still water. A question relevant to larval settlement in general, and active habitat selection in particular, is: in what region above the sea bed are flow speeds sufficiently low such that settling organisms could effectively manoeuvre (e.g. swim among test sites)? Such hydrodynamical constraints for active habitat selection are discussed below. If the larvae sink through the water and are deposited onto the sea bed like passive particles, then parameters of the boundary-layer flow and the gravitational fall velocities of the organisms determine where they will initially reach the sea floor and where they are likely to accumulate. In this case, sediment transport theory can be used to predict depositional or accumulation sites for larvae on the sea floor. Physical considerations involved in such predictions are also discussed below. HYDRODYNAMICAL CONSTRAINTS ON ACTIVE HABITAT SELECTION Hydrodynamical constraints on active habitat selection depend on how settlement cues are perceived by the organisms (see pp. 132–4) and on their swimming behaviours and speeds. If larvae respond to waterborne cues, then the boundary-layer flow determines the extent of mixing (and thus, of dilution) of the cue by the time the larva perceives it. The manner in which the larva responds to the cue (e.g. does it suddenly quit swimming and sink or does it actively swim straight down to the bed?) and the structure of the near-bed flow regime determine how far the larva will be advected downstream before it reaches the bottom. If larvae must make direct contact with the sea bed in order to perceive a settlement cue then, again, potential test

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

123

Fig. 2.—Turbulent velocity profiles constructed in Butman (1986a), plotted on a log-linear scale: for the roughturbulent profiles, the dashed portion represents (10) (z0); below this level, the accuracy of predictions of velocity by the log-layer function are unknown; for the smooth-turbulent profile, the curved region is the viscous sublayer; the line is dashed at the interface between the log layer and the viscous sublayer because the actual function predicting velocities in this region is unknown; the profiles were constructed for a flow speed of u=15 cm·s−1 at z=50 cm, but for different values of bottom roughness (see Table I in Butman, 1986a).

sites on the bed depend on how they conduct a search (e.g. do they swim horizontally among sites or do they swim or sink down to a site and then reject it by swimming straight up?) and on the boundary-layer flow regime. Most of the existing laboratory data suggest that larvae must make direct contact with a surface bearing the cue in order to perceive it; this sensing mechanism is assumed, for the sake of argument, in the following discussion. To determine the flow velocities that larvae experience as they approach the sea floor, Butman (1986a) constructed boundary-layer velocity profiles, based on near-bottom current observations from a shallow (10m depth), subtidal site in the coastal embayment of Buzzards Bay. The flows at this site are primarily driven by the semi-diurnal tides and current speeds measured one metre above the bottom ranged from 0 to 22 cm·s−1. Velocity profiles in the log layer were calculated, assuming both smooth- and rough-turbulent flow, and for different flow speeds. These velocities were compared with the maximum horizontal swimming speed measured for polychaete larvae (from the review of Chia, Buckland-Nicks & Young, 1984). The surprising result of this study was that horizontal flow velocities considerably exceed larval swimming speeds, even at only several larval body lengths above the bed, for most of the flow conditions used in the analysis. At near-peak ebb or flood tide (when u=15 cm·s−1 at z=50 cm), the flow speed is 1 cm·s−1 at distances of about 300 µm (smooth-turbulent), 500 µm (rough-turbulent, u*=0·82 cm·s−1) and

124

CHERYL ANN BUTMAN

Fig. 3.—Smooth-turbulent velocity profiles for three flow speeds (representing three stages of the tidal cycle in Buzzards Bay), as constructed in Butman (1986a) and plotted on a log-linear scale: only the viscous sublayer is shown on the Figure; Line A is for u=15 cm·s−1 at z=50 cm; Line B is for u=10 cm·s−1 at z=50 cm; Line C is for u=5 cm·s−1 at z=50 cm.

1500 µm (rough-turbulent, u*= 0·98 cm·s−1) above the bed (Fig. 2). Because maximum measured swimming speeds of polychaete larvae are only 5 mm·s−1, they would have a difficult time manoeuvring horizontally (e.g. to swim between potential test sites) in any of these flows. Swimming full-speed against the flow at about two body lengths above the bottom, the larvae would still be advected downstream at 5 mm·s−1! Plots of smooth-turbulent velocity profiles for various current speeds (stages of the tide for the Buzzards Bay case) (Fig. 3) indicate that larvae could effectively manoeuvre via horizontal swimming at distances of several body lengths above the bed during near-slack tide (line C in Fig. 3, where u=5 cm·s−1 at z=50 cm) and for slower forcing flows. Figure 2 also shows that, at a given height above the bed and for the same forcing flow at the top of the log layer, the mean horizontal velocity close to the sea bed will be substantially slower in rough- than in smooth-turbulent flow due to the more efficient mixing of high- and low-momentum fluid by eddies in the rough-turbulent flow. Larvae experience, however, only horizontal flow velocities within the viscous sublayer in smooth-turbulent flow, whereas for rough-turbulent flow, they experience the mean horizontal flow speed plus the fluctuating velocity components in all directions, as eddies regularly penetrate the viscous sublayer. Thus, while a larva may encounter unmanageable flow velocities for effective manoeuvring in the horizontal, it can swim up and down unperturbed by vertical flow velocity within the viscous sublayer for smooth-turbulent, but not for rough-turbulent flow. From this quantitative analysis of bottom boundary-layer velocity profiles in a realistic field flow environment, it appears that polychaete larvae probably do not actively swim horizontally among test sites, except under very low-flow conditions (i.e. around slack tides in the Buzzards Bay case). It seems more likely that larvae test habitats by sinking or swimming down to the bed and reject a site by swimming back up into the water column, although the potential effectiveness of this behaviour for rough-turbulent flow is unclear. Since near-bed velocities would carry the larvae over a suite of potential test sites, the habitats presented for their perusal are hydrodynamically constrained. Note that while the sites that a drifting larva may inspect are hydrodynamically determined, the larva may be carried over a wide range of habitats (at 1 cm·s−1, the larva is carried about 1 km·day−1), much farther than it can swim in the same amount of time. Sinking at a

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

125

rate of 0·1 to 1·0 cm·s−1 (as measured in Hannan, 1984a, b), the larva would, however, hit bottom after being advected only centimetres, so it would have to swim up at speeds greater than or equal to its fall velocity to stay above the bottom while drifting. It is possible that larvae do not select habitats by swimming among them; once they reach the sea floor, they may simply crawl between microhabitats, in which case the spatial scales for active habitat selection are very small indeed. Given that near-bed flow velocities over a relatively smooth, flat bottom may allow for very limited manoeuvring by larvae, any flow region with substantially lower velocities (e.g. in the lee of a relatively large roughness element, such as a rock or a tube, or in a dense canopy, such as a seagrass bed) may be particularly important to settling larvae. Whether they actively leave the flow (i.e. by swimming down or sinking) to enter such regions, or simply get deposited there (see pp. 151–4), they may be able to investigate actively such areas without significant interference from the flow regime. The velocity profiles constructed in Butman (1986a) are discussed in detail here because they are unique to the present day literature in larval ecology; they represent, however, conditions for but one class of flow environment (steady, uniform, tidally driven flows in shallow, coastal embayments) and for one group of infauna (the polychaetes). Furthermore, the analysis is limited by the lack of biological information, for example, on how larvae actually peruse available sites even in still water (i.e. is it by horizontal or vertical swimming, or some combination of the two, and from what height above the bed?) and on the relative swimming speeds and fall velocities of the organisms and the changes in these speeds over their pelagic life. While it is clear that hydrodynamics may limit the active habitat selection options for settling larvae, the capabilities of the larvae to overcome or utilize these flow obstacles are not clear. PASSIVE PARTICLE TRANSPORT AND DEPOSITION The long-held tenets that larval dispersal is primarily passive, via ocean currents, but that larval settlement is controlled by active larval behaviours have assumed that flows very close to the sea bed (e.g. within the viscous sublayer) were sufficiently slow to allow for searching and active habitat selection by the larvae. The likelihood that larval settlement is controlled, at least in part, by hydrodynamical processes is strengthened by the results of Butman (1986a) that relatively large (compared with larval swimming speeds, but see also Herrmann, 1979, and Lee, 1984) horizontal flow speeds occur within larval body lengths of the sea bed. If boundary-layer flow processes are transporting and depositing larvae, then the body of literature on sediment transport and deposition can be used to generate a priori predictions of depositional and accumulation sites for the organisms on the sea bed. The trajectory of a particle falling from the water surface to the bottom is determined by the horizontal displacement caused by the flow (advection) and by gravitational sinking of the particle. Once it reaches the a force per unit bottom bottom, the particle will settle on the sea bed if the bottom shear stress area, where =fluid density) does not exceed the critical value for suspension of the particle. This critical stress is usually reported in units of velocity as u*s, the critical suspension velocity. The ratio of particle fall <0·8 (where =von velocity (w) to u*s determines if the particle remains in suspension; when Karman’s constant of 0·4), then the particle will move as suspended-sediment transport (Smith & Hopkins, >0·8, the particle will fall to the sea bed, but will move as bedload transport if 1972). For where u*c is the critical shear velocity for the initiation of particle motion. Bedload transport involves sliding, rolling or hopping of particles along the sea bed. From detailed laboratory measurements, curves have been constructed which allow prediction of u*c for given particle characteristics (diameter and density) and fluid characteristics (density and viscosity). The most common relationship used is Shields’ curve (Shields, 1936) or a subsequent modification (e.g. Miller, McCave & Komar; Yalin, 1977); these curves

126

CHERYL ANN BUTMAN

were, however, constructed from measurements on abiotic, non-cohesive sediments 100 in size and spread in homogeneous size classes (i.e. not size-class mixtures) on the bottom. Results for initiation of motion or suspension of fine, biotic or cohesive sediments and sediment mixtures (e.g. Nowell, Jumars & Eckman, 1981; Grant, Boyer & Sanford, 1982; Lick, 1982; McCave, 1985; Partheneides, 1986) have not yet been integrated into formal predictive functions, at the level of Shields’ curve, for example. The shear velocity of the flow (u*) and the particle fall velocity (w) are involved in all estimates or predictions of particle deposition and transport. Measurements of w now are reliable and routine for a wide range of non-aggregated particles; as mentioned earlier, determining the fall velocities of naturally occurring aggregates is still, however, troublesome (but see new in situ techniques in Bartz et al., 1985). Estimating u* for the suite of complex flow environments occurring in the field has been a primary focus in sediment-transport modelling over the last decade (Grant, 1977; Smith, 1977; Smith & McLean, 1977; Grant & Madsen, 1979, 1982; Grant & Glenn, 1983; McLean, 1985). From detailed field measurements of velocity at several heights within the log layer, it is possible to estimate u* from the slope of the line relating where z0=the bottom roughness parameter); the u and ln z (because, within the log layer, correlation between the two variables must, however, be extremely high (generally >0·990) for reasonable limits (e.g. <20%) to the error in such u* estimates (Gross & Nowell, 1983). This is because a well-behaved log layer generated from a single source of flow forcing is actually rare in the field. Instead, forcing from several sources (e.g. tides, internal waves, and surface waves generated by winds) occurs simultaneously to produce several log layers superimposed on one another. The effects are not simply additive between steady (due to currents) and non-steady (due to waves) flows, so a considerable theoretical modelling effort has been placed on predicting u* from field data in these cases (Smith & McLean, 1977; Grant & Madsen, 1979, 1982). The wide range of values which are possible for u* in the field, even over time at a single location, means that particles may move almost continuously or intermittently, depending on the physical forcing. Bottom topography also plays a major rôle in determining the eventual accumulation sites for a given particle class, by altering the flow regime. Through these processes, the large-scale (tens of metres to tens of kilometres or more) distribution of sediment grain sizes are established. The sediment distributions documented by regional surveys generally reflect a long-term dynamic equilibrium between the physical processes which transport and deposit the sediments. A recent overview of the spatial and temporal scales of physical processes which produce various features of the sedimentary and sediment-transport environment at a single, well-studied locale (Georges Bank, Massachusetts, U.S.A.), and written for a general audience, is given in Butman (in press). Smaller-scale (centimetres to metres) changes in sediment texture result from small-scale variability in bottom topography. The scale of the morphological feature in the sea bed sets the scale of sediment grainsize patchiness. For example, in an area of coarse sand, fine sediments often accumulate in the feeding pits of rays, which are tens of centimetres in diameter (Grant, 1981; VanBlaricom, 1982). For unevenly distributed flow obstacles (or relatively large roughness elements) such as rocks, biogenic mounds, or tubes of infauna which are relatively far apart, the sediment environment is affected at spatial scales of the order of the diameter of the obstacle (e.g. Eckman & Nowell, 1984). For more densely packed elements, effects on sedimentation are a function of the height and packing of the elements and of the flow regime (e.g. Wooding, Bradley & Marshall, 1973; Nowell & Church, 1979; Eckman, 1983), so that laboratory flume studies may be required to predict specific effects (see especially Eckman, 1983). Passive deposition and accumulation of larvae is expected at the spatial scales which apply to sediments with similar fall velocities. The quantitative data for polychaete larvae indicates that their fall velocities (0·1 to 1·0 cm·s−1) are within the range of fine quartz sediments (silts) 10–80 µm in diameter (Hannan, 1984a,

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

127

b). Pratt (1953) and Tyler & Banner (1977) found that bivalve and echinoderm postlarval distributions were well correlated with distributions of the “fine” sediment fraction over spatial scales of kilometres. Settling larvae may thus accumulate in regions where fine sediments accumulate; i.e. both groups of “particles” are in dynamic equilibrium with the physical environment. On the largest scales (tens of metres to tens of kilometres), these accumulation zones are set by large-scale topography and the flow during the time of larval settlement. Settling larvae, however, differ from settling particulates in that they may become sticky once they reach the bed or they may burrow into it. If these factors are significant, then perhaps only the initial distribution of larvae on the sea bed would be determined by physical processes. Finally, larvae may also accumulate like passive particles on small spatial scales (centimetres to metres) due to microtopography effects, as suggested by the results of Baggerman (1953), Eckman (1979, 1983) and Hogue& Miller (1981). In summary, the increasing literature on sediment transport allows predictions of erosion, transport, deposition, and accumulation of particles of a given fall-velocity class, given important information regarding the field flow regime. Such models can be used to generate predicted distributions for passively settling larvae, once the larval fall velocity is known. Both large- and small-scale patterns of initial larval settlement and subsequent accumulation could be explained by physical processes, since variability in sediment distributions can occur on several scales. It is, however, most intriguing that the observed correlations between infaunal species composition and bulk sediment distributions at large spatial scales (e.g. see Table III, pp. 120–1) may simply occur because larval settlement is determined by the same physical processes which maintain the distribution of sediments, which have fall velocities similar to larvae. The passive deposition hypothesis may also explain some of the often extreme variability in larval settlement in different years. Although the surficial sediment distributions are the cumulative result of many sediment transport events over a long time, the larvae experience only a subset of these events at the time of settlement. This wide range of flow events (storms, spring or neap tides, and run-off, for example) could change the ‘average’ deposition pattern for larvae by altering the advective, as well as the depositional regimes, for the short time that the organisms are in this critical stage of their life history. In addition, attractive characteristics of surface sediments depend, in part, on rates of sedimentation and the hydrodynamics of the region, such that appropriate settlement cues may be emitted from a given area only intermittently (Chia & Crawford, 1973). The sporadic availability of depositional environments or attractive surface sediments should be particularly important for species that have sharply limited, rather than extended, reproductive seasons. In fact, Todd & Doyle (1981) have proposed the “settlement-timing hypothesis” to account for the reproductive behaviour of benthic invertebrates, where the behaviour of a given species in a particular environmental settling is a compromise for that area, between the optimal time for spawning and the optimal time for settling, both of which are energetically constrained. COMPATIBILITY OF THE ALTERNATIVE HYPOTHESES Active habitat selection by and passive deposition of larvae need not be considered mutually exclusive alternative hypotheses. The processes may both operate, but over different scales of space and time. The relative importance of the two processes may also vary for different species (depending, for example, on the fall velocity and swimming capabilities of the organisms), for different flow environments, or depending on how long the organisms have been in the plankton. In addition, the composition of infaunal communities may be determined, in part, by dispersal and deposition (active or passive) of postlarvae or adults (but see also Santos & Simon, 1980a), as evidenced by the mounting records of benthic organisms in the water column (Table II, but see also Ambrose, 1984a). Surface or near-surface dwelling organisms may

128

CHERYL ANN BUTMAN

periodically enter the water with sediments during rigorous, storm-induced resuspension events (Hagerman & Rieger, 1981; Hogue, 1982; Dobbs & Vozarik, 1983), or regularly, during tidal resuspension (Bell & Sherman, 1980; Palmer & Brandt, 1981; Palmer & Gust, 1985). There is also indirect evidence (from sampling of the sea bed) that infauna or meiofauna actively or passively enter the water column to migrate to new locations (Baggerman, 1953; Trueman, 1971; Dauer & Simon, 1975; Farke, de Wilde & Berghuis, 1979; Grant, 1981). Thus, individual members of the infaunal community may be much more dynamic than heretofore believed. Evidence from the literature on benthic ecology and considerations of physical phenomena suggest that larvae may be passively deposited and accumulate at the large spatial scales (tens of metres to tens of kilometres) which apply to sediment transport and deposition and that active habitat selection occurs over much smaller scales (centimetres to metres) within these broad, depositional areas. Local distributions of settling or settled organisms also could be determined by small-scale physical phenomena, such as changes in flow induced by microtopography of the sea bed. In addition, larvae may be permitted to select actively habitats over larger spatial scales during time periods when flows near the bed are very slow (e.g. surrounding slack tide) or in areas of sluggish circulation, in general. The extent to which organisms can actively select habitats within their flow environments depends on their method of perceiving, testing and locating habitats. Certain species, morphological types and/or developmental stages of larvae are more capable of controlling their position in the water column than others. Crustaceans, in general, possess much better swimming and position regulating capabilities than most other infauna (see Mileikovsky, 1973; Mann & Wolf, 1983; Chia, Buckland-Nicks & Young, 1984; Sulkin, 1984), so they may be expected to manoeuvre more effectively in flows. Likewise, during development swimming abilities may increase (e.g. Herrmann, 1979; Lee, 1984) or decrease (e.g. Konstantinova, 1969; Miller & Hadfield, 1986; author’s unpubl. data), so larvae may become more or less effective in locating preferred habitats. The evidence for decreasing substratum selectivity as competent larvae spend more time in the plankton suggests a finite period for active habitat selection, following which, passive deposition may occur. Delay of metamorphosis by many species may allow the organisms to be passively advected until they reach flow environments which are sufficiently sluggish that active searches for preferred substrata are possible. Species with developmental constraints on their pelagic period may only be capable of actively selecting habitats early in their competent period, if flows permit, whereas later on the organisms may be passively deposited as the larvae are “forced” to metamorphose and settle. Hadfield (1978b), however, reported an interesting case of “partial metamorphosis” in the larvae of an enteropneust, where the larvae lose all pelagic characteristics except the teletroch, which is retained for locomotion until the organisms reach a bottom habitat that is soft enough for burrowing. These options for the compatibility of the active habitat selection and passive deposition hypotheses are posed to stimulate new research in this area. Certainly there are other options, in addition to the few proposed here. As technological advances increasingly provide biologists with the tools necessary to study organisms within their natural habitats (simulated in the laboratory or in situ), so we may expect advances in our ecological insight of how organisms perceive, cope with, and are controlled by their environment. Nearly thirty years ago, Wilson (1958) spoke of the interdisciplinary nature of larval settlement studies, urging interaction between biologists and chemists in defining attractive factors of sediments and other settlement cues. His remarks (Wilson, 1958, p. 96) are still relevant today, only “other sciences” now includes physics, as well: “As for the larval-settlement problems, so here the zoologist has been brought to the borders of other sciences, and without collaborators from these sciences to help him [and her] along he [and she] is [are] not likely to make much further progress.”

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

129

ACKNOWLEDGEMENTS I thank J.F.Grassle for his insights, untiring interaction and remarkable knowledge of and appreciation for the literature. The early evolution of these thoughts on larval settlement was greatly influenced by conversations with J.S.Oliver and J.T.Carlton. I am particularly grateful for many years of patient instruction by W.D.Grant on boundary-layer flow and sediment-transport processes and deeply regret that this gratitude is extended post mortem. Ideas on the physics applicable to larval settlement from B.Butman, P.A.Jumars, A.R.M.Nowell and K.D.Stolzenbach also contributed to this review. I thank B.Butman, J.F.Grassle, J.P.Grassle, L.A.Levin, C.H.Peterson, and C.M.Webb for helpful comments on the manuscript, C.M.Fuller for technical assistance in all aspects of my own research, and B.Butman for multifaceted support. The manuscript was skilfully typed by G.McManamin, for which I am grateful. Parts of this review constituted the first chapter of my dissertation and were supported by the Coastal Research Center at Woods Hole Oceanographic Institution, the Diving Equipment Manufacturer’s Association, an Association for Women in Science Predoctoral Award, a National Science Foundation Dissertation Improvement Grant (No. OCE81–19865), a National Ocean Survey/Sea Grant Fellowship (NOAA NA80-AA-D00077) and the Woods Hole Oceanographic Institution Education Program; the expansion published here was supported by the National Science Foundation (Grant No. OCE85–000875) and a PEW Memorial Trust fellowship to the Ocean Engineering Department at Woods Hole Oceanographic Institution. REFERENCES Alldredge, A.L. & King, J.M., 1977. Mar. Biol, 41, 317–333. Alldredge, A.L. & King, J.M., 1980. J. exp. mar. Biol. Ecol., 44, 133–156. Ambrose, Jr, W.G., 1984a. J. exp. mar. Biol. Ecol., 80, 67–75. Ambrose, Jr, W.G., 1984b. J. mar. Res., 42, 633–654. Andrews, J.A., 1979. In, Reproduction in Marine Invertebrates, Volume 5, Molluscs: Pelecypods and Lesser Classes, edited by A.C.Giese & J.S.Pearse, Academic Press, New York, pp. 293–352. Angel, H.H. & Angel, M.V., 1967. Helgol. wiss. Meeresforsch., 15, 445–454. Arntz, W.E., 1980. In, Marine Benthic Dynamics, edited by K.R.Tenore & B.C. Coull, University of South Carolina Press, Columbia, South Carolina, pp. 121– 149. Arntz, W.E. & Rumohr, H., 1982. J. exp. mar. Biol. Ecol., 64, 17–45. Baggerman, B., 1953. Archs néerl. Zool., 10, 315–342. Bartz, R., Zaneveld, J.R.V., McCave, I.N., Hess, F.R. & Nowell, A.R.M., 1985. Mar. Geol., 66, 381–395. Bayne, B.L., 1964. J. Anim. Ecol., 33, 513–523. Beeman, R.D., 1977. In, Reproduction in Marine Invertebrates, Volume 4, Molluscs: Gastropods and Cephalopods, edited by A.C.Giese & J.S.Pearse, Academic Press, New York, pp. 115–179. Bell, S.S. & Sherman, K.M., 1980. Mar. Ecol. Progr. Ser., 3, 245–249. Bergquist, P.R., Sinclair, M. & Hogg, J.J., 1970. In, The Biology of the Porifera, edited by W.G.Fry, Symp. Zool. Soc. Land., 25, pp. 247–271. Berrill, N.J., 1975. In, Reproduction in Marine Invertebrates, Volume 2, Entoprocts and Lesser Coelomates, edited by A.C.Giese & J.S.Pearse, Academic Press, New York, pp. 241–282. Beukema, J.J., 1973. Neth. J. Zool., 23, 356–357. Beukema, J.J. & DeVlas, J., 1979. Neth. J. Sea Res., 13, 331–353. Bhaud, M., Aubin, D. & Duhamel, G., 1981. Oceanis, 7, 97–113. Bhup, R. & Marsden, J.R., 1982. Can. J. Zool., 60, 2284–2295. Birkeland, C. & Chia, F.-S., 1971. J. exp. mar. Biol. Ecol., 6, 265–278. Bloom, S.A., Simon, J.L. & Hunter, V.D., 1972. Mar. Biol., 13, 43–56.

130

CHERYL ANN BUTMAN

Boaden, P.J.S., 1963. J. mar. biol. Ass. U.K., 43, 239–250. Boaden, P.J.S., 1968. Sarsia, 34, 125–136. Boesch, D.F., Diaz, R.J. & Virnstein, R.W., 1976. Chesapeake Sci., 17, 246–259. Bonsdorff, E. & Österman, C.-S., 1985. In, Proc. 19th Europ. Mar. Biol. Symp., edited by P.E.Gibbs, Cambridge University Press, Cambridge, pp. 287–297. Botero, L. & Atema, J., 1982. J. crust. Biol., 2, 59–69. Bousfield, E.L., 1955. Bull. natn. Mus. Can., No. 137, 1–69. Brenchley, G.A., 1981. J. mar. Res., 39, 767–790. Buchanan, J.B., 1963. Oikos, 14, 154–175. Buchanan, J.B., Sheader, M. & Kingston, P.F., 1978. J. mar. biol. Ass. U.K., 58, 191–209. Burke, R.D., 1983. Can. J. Zool., 61, 1701–1719. Burke, R.D., 1984. Science, 225, 442–443. Butman, B., in press. In, Georges Bank, edited by R.H.Backus & D.W.Bourne, MIT Press, Cambridge, Massachusetts. Butman, C.A., 1986a. In, Marine Interfaces Ecohydrodynamics, edited by J.C.J. Nihoul, Elsevier Oceanography Series, 42, Elsevier, Amsterdam, pp. 487–513. Butman, C.A., 1986b. J. mar. Res., 44, 645–693. Butman, C.A., Grant, W.D. & Stolzenbach, K.D., 1986. J. mar. Res., 44, 601–644. Cacchione, D.A. & Drake, D.E., 1982. J. geophys. Res., 87, 1952–1960. Caldwell, J.W., 1972. Masters thesis, University of Florida, Tallahassee, 63 pp. Cameron, R.A. & Rumrill, S.S., 1982. Mar. Biol., 71, 197–202. Campbell, R.D., 1974. In, Reproduction in Marine Invertebrates, Volume 1, Acoelomate and Pseudocoelomate Metazoans, edited by A.C.Giese & J.S.Pearse, Academic Press, New York, pp. 133–199. Carriker, M.R., 1961. J. Elisha Mitchell Sci. Soc., 77, 168–241. Cassie, R.M. & Michael, A.D., 1968. J. exp. mar. Biol. Ecol., 2, 1–23. Chapman, G. & Newell, G.E., 1949. J. mar. biol. Ass. U.K., 28, 627–634. Chia, F.-S. & Bickell, L.R., 1978. In, Settlement and Metamorphosis of Marine Invertebrate Larvae, edited by F.-S. Chia & M.E.Rice, Elsevier, New York, pp. 1–12. Chia, F.-S., Buckland-Nicks, J. & Young, C.M., 1984. Can. J. Zool., 62, 1205–1222. Chia, F.-S. & Crawford, B.J., 1973. Mar. Biol., 23, 73–82. Chriss, T.M. & Caldwell, D.R., 1982. J. geophys. Res., 87, 4148–4154. Clauser, F.H., 1956. Adv. appl. Math., 4, 1–51. Connell, J.H., 1985. J. exp. mar. Biol. Ecol., 93, 11–45. Crisp, D.J., 1955. J. exp. Biol., 32, 569–590. Crisp, D.J., 1974. In, Chemoreception in Marine Organisms, edited by P.T.Grant & A.M.Mackie, Academic Press, New York, pp. 177–265. Crisp, D.J., 1976. In, Adaptations to Environment: Essays on the Physiology of Marine Animals, edited by R.C.Newell, Butterworths, London, pp. 83–124. Crisp, D.J., 1984. In, Marine Biodeterioration: An Interdisciplinary Study, edited by J.D.Costlow & R.C.Tipper, Naval Institute Press, Annapolis, pp. 103–126. Crisp, D.J. & Meadows, P.S., 1962. Proc. R. Soc. Land. Ser. B., 156, 500–520. Crisp, D.J. & Meadows, P.S., 1963. Proc. R. Soc. Land. Ser. B., 158, 364–387. Croker, R.A., 1967. Ecol. Monogr., 37, 173–200. Crumb, S.E., 1977. Chesapeake Sci., 18, 253–265. Cuomo, M.C., 1985. Biogeochemistry, 1, 169–181. Dauer, D.M., Ewing, R.M., Tourtellotte, G.H. & Barker, Jr, H.R., 1980. Estuaries, 3, 148–149. Dauer, D.M. & Simon, J.L., 1975. Mar. Biol., 31, 363–370. Dauer, D.M. & Simon, J.L., 1976. Mar. Biol., 37, 169–177. Davies, R., 1985. What’s Bred in the Bone. Viking Penguin, New York, see p. 16. Day, J.H., 1937. Kept Dove mar. Lab., Ser. 3, No. 4, 31–71.

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

131

Day, J.H. & Wilson, D.P., 1934. J. exp. mar. Biol. Ass. U.K., 19, 655–662. Day, J.H., Field, J.G. & Montgomery, M.P., 1971. J. Anim. Ecol, 40, 93–125. Day, R. & McEdward, L., 1984. In, Marine Plankton Life Cycle Strategies, edited by K.A.Stiedinger & L.M.Walker, CRC Press, Boca Raton, Florida, pp. 93– 120. Dayton, P.K. & Oliver, J.S., 1980. In, Marine Benthic Dynamics, edited by K.R. Tenore & B.C.Coull, University of South Carolina Press, Columbia, South Carolina, pp. 93–120. Dean, D., 1978a. Mar. Biol., 45, 165–173. Dean, D., 1978b. Mar. Biol., 48, 99–104. Dobbs, F.C. & Vozarik, J.M., 1983. Mar. Ecol. Progr. Ser., 11, 273–279. Doochin, H. & Smith, F.G.W., 1951. Bull. mar. Sci. Gulf Caribb., 1, 196–208. Doyle, R.W., 1975. Am. Nat., 109, 113–126. Eagle, R.A., 1973. Estuar. cstl mar. Sci., 1, 285–299. Eckelbarger, K.J., 1975. Mar. Biol., 30, 137–149. Eckelbarger, K.J., 1976. Bull. mar. Sci., 26, 117–132. Eckelbarger, K.J., 1978. In, Settlement and Metamorphosis of Marine Invertebrate Larvae, edited by F.-S.Chia & M.E.Rice, Elsevier, New York, pp. 145–164. Eckelmann, H., 1974. J. Fluid Mech., 65, 439–459. Eckman, J.E., 1979. J. mar. Res., 37, 437–457. Eckman, J.E., 1983. Limnol. Oceanogr., 28, 241–257. Eckman, J.E., 1985. J. mar. Res., 43, 419–435. Eckman, J.E. & Nowell, A.R.M., 1984. Sedimentology, 31, 851–862. Emery, A.R., 1968. Limnol. Oceanogr., 13, 293–303. Fage, L. & Legendre, R., 1927. Archs Zool. exp. gén., 67, 23–222. Fager, E.W., 1964. Science, 143, 356–359. Farke, H., Wilde, P.A.W.J. de & Berghuis, E.M., 1979. Neth. J. Sea Res., 13, 354– 361. Fell, P.E., 1974. In, Reproduction in Marine Invertebrates, Volume 1, Acoelomate and Pseudocoelomate Metazoans, edited by A.C.Giese & J.S.Pearse, Academic Press, New York, pp. 51–132. Flint, R.W. & Holland, J.S., 1980. Estuar. cstl mar. Sci., 10, 1–14. Ford, E., 1923. J. mar. biol. Ass. U.K., 13, 164–224. Fosshagen, A., 1965. Thesis, University of Bergen, Norway, 70 pp. Gage, J., 1972a. J. mar. biol. Ass. U.K., 52, 237–276. Gage, J., 1972b. Mar. Biol., 14, 281–297. Gage, J. & Geekie, A.D., 1973a. Mar. Biol., 19, 41–53. Gage, J. & Geekie, A.D., 1973b. Mar. Biol., 20, 89–100. Gallagher, E.D., Jumars, P.A. & Trueblood, D.D., 1983. Ecology, 64, 1200–1216. Gärdefors, D. & Orrhage, L., 1968. Oikos, 19, 311–322. Gardner, W.D., 1980. J. mar. Res., 38, 17–39. Gibbs, P.E., 1969. J. mar. biol. Ass. U.K., 49, 311–326. Gibson, P.H. & Nott, J.A., 1971. In, Proc. 4th Europ. Mar. Biol. Symp., edited by D.J.Crisp, Cambridge University Press, Cambridge, pp. 227–236. Graham, J.J. & Creaser, Jr, E.P., 1978. Fishery Bull. N.O.A.A., 76, 480–483. Grant, J., 1981. Mar. Ecol. Progr. Ser., 6, 249–255. Grant, W.D., 1977. Sc. D. thesis, Massachusetts Institute of Technology, Cambridge, 275 pp. Grant, W.D., Boyer, L.F. & Sanford, L.P., 1982. J. mar. Res., 40, 659–677. Grant, W.D. & Glenn, S.M., 1983. Continental Shelf Bottom Boundary Layer Model: Theoretical Model, Vol. 1, Tech. Rept Am. Gas Assoc., PR-153–126, May 31, 1983, 163pp. Grant, W.D. & Madsen, O.S., 1979. J. geophys. Res., 84, 1797–1808. Grant, W.D. & Madsen, O.S., 1982. J. geophys. Res., 87, 469–481. Grant, W.D. & Madsen, O.S., 1986. Ann. Rev. Fluid Mech., 18, 265–305.

132

CHERYL ANN BUTMAN

Grant, W.D., Williams, III, A.J. & Glenn, S.M., 1984. J. phys. Oceanogr., 14, 506– 527. Grassle, J.F. & Grassle, J.P., 1974. J. mar. Res., 32, 253–284. Grassle, J.F., Grassle, J.P., Brown-Leger, L.S., Petrecca, R.F. & Copley, N.J., 1985. In, Marine Biology of Polar Regions and Effects of Stress on Marine Organisms, Proc. 18th Europ. Mar. Biol. Symp., edited by J.S.Gray & M.E. Christiansen, John Wiley & Sons, Chichester, pp. 421–434. Grassle, J.F., Sanders, H.L., Hessler, R.R., Rowe, G.T. & McLellan, T., 1975. Deep-Sea Res., 22, 457–481. Grassle, J.P., 1980. Am. Zool., 20, 752 only. Grassle, J.P. & Grassle, J.F., 1976. Science, 192, 567–569. Gray, J.S., 1966a. J. mar. biol. Ass. U.K., 46, 627–645. Gray, J.S., 1966b. J. Anim. Ecol., 35, 435–442. Gray, J.S., 1966c. Veröff Inst. Meeresforsch. Bremerh., Sonderb. II, 105–116. Gray, J.S., 1967a. Helgol. wiss. Meeresunters., 15, 253–269. Gray, J.S., 1967b. J. exp. mar. Biol. Ecol., 1, 47–54. Gray, J.S., 1968. J. exp. mar. Biol. Ecol., 2, 278–292. Gray, J.S., 1971. Vie Milieu, Suppl. No. 22, 707–722. Gray, J.S., 1974. Oceanogr. Mar. Biol. Ann. Rev., 12, 223–261. Gray, J.S. & Johnson, R.M., 1970. J. exp. mar. Biol. Ecol., 4, 119–133. Gross, T.F. & Nowell, A.R.M., 1983. Cont. Shelf Res., 2, 109–126. Guérin, J.-P., 1982. Oceanis B, 5, 389–404. Guérin, J.-P. & Massé, H., 1978. Téthys, 8, 151–168. Hadfield, M.G., 1978a. In, Settlement and Metamorphosis of Marine Invertebrate Larvae, edited by F.-S.Chia & M.E.Rice, Elsevier, New York, pp. 165–176. Hadfield, M.G., 1978b. In, Settlement and Metamorphosis of Marine Invertebrate Larvae, edited by F.-S.Chia & M.E.Rice, Elsevier, New York, pp. 247–254. Hadl, G., Kothbauer, H., Peter, R. & Wawra, E., 1970. Oecologia (Berl.), 4, 74–82. Hagerman, Jr, G.M. & Rieger, R.M., 1981. P.S.Z.N.I. Mar. Ecol., 2, 245–270. Hammer, R.M., 1981. Mar. Biol., 62, 275–280. Hammer, R.M. & Zimmerman, R.C., 1979. Bull. Sth. Calif. Acad. Sci., 78, 199– 206. Hannan, C.A., 1980. Masters thesis, San Jose State University and Moss Landing Marine Laboratories, California, 118 pp. Hannan, C.A., 1984a. Doctoral dissertation, Woods Hole Oceanographic Institution and Massachusetts Institute of Technology Joint Program, Woods Hole, 534 pp. Hannan, C.A., 1981. Limnol. Oceanogr., 26, 159–171. Hannan, C.A., 1984b. Limnol. Oceanogr., 29, 1108–1116. Hargrave, B.T. & Burns, N.M., 1979. Limnol. Oceanogr., 24, 1124–1136. Henderson, J.A. & Lucas, J.S., 1971. Nature, Land., 232, 655–657. Hermans, C.O., 1964. Masters thesis, University of Washington, Seattle, 132 pp. Herrmann, K., 1979. Helgol. wiss. Meeresunters., 32, 550–581. Highsmith, R.C., 1982. Ecology, 63, 329–337. Hobson, E.S. & Chess, J.R., 1976. Fishery Bull. N.O.A.A., 74, 567–598. Hobson, E.S. & Chess, J.R., 1979. Fishery Bull. N.O.A.A., 77, 275–280. Hogue, E.W., 1982. J. mar. Res., 40, 551–573. Hogue, E.W., & Miller, C.B., 1981. J. exp. mar. Biol. Ecol., 53, 181–191. Holland, A.F., Mountford, N.K., Hiegel, M.H., Kaumeyer, K.R. & Mihursky, J. A., 1980. Mar. Biol., 57, 221–235. Holland, A.F., Mountford, N.K. & Mihursky, J.A., 1977. Chesapeake Sci., 18, 370–378. Holland, A.F. & Polgar, T.T., 1976. Mar. Biol., 37, 341–348. Holme, N.A., 1949. J. mar. biol. Ass. U.K., 28, 189–237. Howard, J.D. & Dörjes, J., 1972. J. Sediment. Petrol., 42, 608–623. Hughes, R.N., Peer, D.L. & Mann, K.H., 1972. Limnol. Oceanogr., 17, 111–121.

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

133

Hughes, R.N., & Thomas, M.L.H., 1971. J. exp. mar. Biol. Ecol., 7, 1–39. Hulberg, L.W. & Oliver, J.S., 1980. Can. J. Fish. Aquat. Sci., 37, 1130–1139. Jackson, G.A. & Strathmann, R.R., 1981. Am. Nat., 118, 16–26. Jansson, B.-O., 1967a. Helgol. wiss. Meeresunters., 15, 41–58. Jansson, B.-O., 1967b. Oikos, 18, 311–322. Jensen, P., 1981. Mar. Ecol. Progr. Ser., 4, 203–206. Johnson, R.G., 1971. Mar. Geol., 11, 93–104. Jones, D.A., 1970. J. Anim. Ecol., 39, 455–470. Jones, M.L., 1962. Univ. Calif. Publ. Zool., 67, 219–320. Jumars, P.A., 1976. J. mar. Res., 34, 217–246. Jumars, P.A. & Eckman, J.E., 1983. In, Deep-Sea Biology, edited by G.T.Rowe, John Wiley & Sons, New York, pp. 399–451. Jumars, P.A. & Nowell, A.R.M., 1984a. Cont. Shelf Res., 3, 115–130. Jumars, P.A. & Nowell, A.R.M., 1984b. Am. Zool., 24, 45–55. Jumars, P.A., Nowell, A.R.M. & Self, R.F.L., 1981. Mar. Geol., 42, 155–172. Keck, R., Maurer, D. & Malouf, R., 1974. Proc. natl Shellfish Ass., 64, 59–67. Kempf, S.C. & Hadfield, M.G., 1985. Biol. Bull. mar. biol. Lab., Woods Hole, 169, 119–130. Kennedy, V.S., 1982. Editor. Estuarine Comparisons, Academic Press, New York, 709 pp. Keough, M.J. & Downes, B.J., 1982. Oecologia (Berl.), 54, 348–352. Klauser, M.D., 1986. J. exp. mar. Biol. Ecol., 97, 123–133. Konstantinova, M.I., 1969. Dokl. Akad. Nauk SSSR, 188, 942–945. Kreger, D., 1940. Arch, néerl. Zool., 4, 157–200. Kristensen, I., 1957. Arch. néerl. Zool., 12, 351–453. Lacalli, T.C., 1980. Can. Tech. Rept Fish. Aquat. Sci., No. 940, 27 pp. Larsen, P.F., 1979. Mar. Biol., 55, 69–78. Lee, H., 1984. Am. Zool., 24, 131A only. Levin, L.A., 1981. J. mar. Res., 39, 99–117. Levin, L.A., 1984. Ecology, 65, 1185–1200. Levin, L.A., 1986. Bull. mar. Sci., 39, 224–233. Levin, L.A. & Greenblatt, P.R., 1981. Bull. Sth. Calif. Acad. Sci., 80, 131–133. Levinton, J.S., 1977. In, Ecology of Marine Benthos, edited by B.C.Coull, University of South Carolina Press, Columbia, South Carolina, pp. 191–227. Levinton, J.S. & Bambach, R.K., 1970. Am. J. Sci., 268, 97–112. Lewis, C.A., 1978. In, Settlement and Metamorphosis of Marine Invertebrate Larvae, edited by F.-S.Chia & M.E.Rice, Elsevier, New York, pp. 207–218. Lewis, D.B., 1968. J. Linn. Soc. Zool., 47, 515–526. Lick, W., 1982. Hydrobiologia, 91, 31–40. Lie, U., 1968. Fiskdir. Skr. Ser. Havunders., 14, 229–556. Lie, U. & Kisker, D.S., 1970. J. Fish. Res. Bd Can., 27, 2273–2285. Lima, G.M. & Pechenik, J.A., 1985. J. exp. mar. Biol. Ecol., 90, 55–71. Luckenbach, M.W., 1984. Mar. Ecol. Progr. Ser., 17, 245–250. Maciolek, N.J. & Grassle, J.F., in press. In, Georges Bank, edited by R.H.Backus & D.W.Bourne, MIT Press, Cambridge, Massachusetts. Mahoney, B.M.S. & Livingston, R.J., 1982. Mar. Biol., 69, 207–213. Mann, R. & Wolf, C.C., 1983. Mar. Ecol. Progr. Ser., 13, 211–218. Mare, M.F., 1942. J. mar. biol. Ass. U.K., 25, 517–554. Massé, H. & Guérin, J.-P., 1978. Téthys, 8, 283–294. McCall, P.L., 1977. J. mar. Res., 35, 221–266. McCann, L.D., 1986. Am. Zool. 26 (4), 129A only.

134

CHERYL ANN BUTMAN

McCave, I.N., 1985. In, Fine-Grained Sediments: Deep-Water Processes and Facies, edited by D.A.V.Stow & D.J.W.Piper, Blackwell Scientific, Boston, pp. 35– 69. McDougall, K.D., 1943. Ecol. Monogr., 13, 321–374. McGrorty, S. & Reading, C.J., 1984. Estuar. cstl shelf Sci., 19, 303–319. McLean, S.R., 1985. Mar. Geol., 66, 243–265. McLusky, D.S., Anderson, F.E. & Wolfe-Murphy, S., 1983. Mar. Ecol. Progr. Ser., 11, 173–179. McNulty, J.K., Work, R.C. & Moore, H.B., 1962a. Bull. mar. Sci. Gulf Caribb., 12, 204–233. McNulty, J.K., Work, R.C. & Moore, H.B., 1962b. Bull. mar. Sci. Gulf Caribb., 12, 322–332. McWilliam, P.S., Sale, P.F. & Anderson, D.T., 1981. J. exp. mar. Biol. Ecol., 52, 185–203. Meadows, P.S., 1964a. J. Anim. Ecol., 33, 387–394. Meadows, P.S., 1964b. J. exp. Biol., 41, 677–687. Meadows, P.S., 1964c. J. exp. Biol., 41, 499–511. Meadows, P.S. & Campbell, J.I., 1972a. Adv. mar. Biol., 10, 271–382. Meadows, P.S. & Campbell, J.I., 1972b. Proc. R. Soc. Edinb. Ser. B, 73, 145–157. Meadows, P.S. & Mitchell, K.A., 1973. Mar. Behav. Physiol, 2, 187–196. Mileikovsky, S.A., 1973. Mar. Biol., 23, 11–17. Miller, M.C., McCave, I.N. & Komar, P.D., 1977. Sedimentology, 24, 507–527. Miller, S.E. & Hadfield, M.G., 1986. J. exp. mar. Biol. Ecol., 97, 95–112. Moore, P.G., 1975. Mar. Behav. Physiol., 3, 97–100. Morgan, E., 1970. J. mar. biol. Ass. U.K., 50, 769–785. Mortensen, T., 1921. Studies of the Development and Larval Forms of Echinoderms, G.E.C.Gad, Copenhagen, 261 pp. Mountford, N.K., Holland, A.F. & Mihursky, J.A., 1977. Chesapeake Sci., 18, 360–369. Muus, B.J., 1967. Meddr. Kommn Danm. Fisk.-og Havunders. N.S., 5, Nr. 1, 1–316. Muus, K., 1966. Veröff. Inst. Meeresforsch. Bremerh. Sonderb. II, 289–292. Muus, K., 1973. Ophelia, 12, 79–116. Myers, A.C., 1977a. J. mar. Res., 35, 609–632. Myers, A.C., 1977b. J. mar. Res., 35, 633–647. Nichols, F.H., 1970. Mar. Biol., 6, 48–57. Nowell, A.R.M., 1983. Rev. Geophys. Space Phys., 21, 1181–1192. Nowell, A.R.M. & Church, M., 1979. J. geophys. Res., 84, 4816–4824. Nowell, A.R.M. & Jumars, P.A., 1984. Ann. Rev. Ecol. Syst., 15, 303–328. Nowell, A.R.M. & Jumars, P.A., 1987. Oceanogr. Mar. Biol. Ann. Rev., 25, 91–112. Nowell, A.R.M., Jumars, P.A. & Eckman, J.E., 1981. Mar. Geol., 42, 133–153. Ohlhorst, S.L., 1982. J. exp. mar. Biol. Ecol., 60, 1–15. Oliver, J.S., 1979. Doctoral Dissertation, University of California, San Diego, 300 pp. Oliver, J.S., Slattery, P.N., Hulberg, L.W. & Nybakken, J.W., 1980. Fishery Bull. N.O.A.A., 78, 437–154. Olsson, I. & Eriksson, B., 1974. Zoon, 2, 67–84. Orth, R.J., 1977. In, Ecology of Marine Benthos, edited by B.C.Coull, University of South Carolina Press, Columbia, South Carolina, pp. 281–300. Orton, J.H., 1937. Nature, Lond., 140, 505–506. Palmer, M.A., 1984. Mar. Behav. Physiol, 10, 235–253. Palmer, M.A. & Brandt, R.R., 1981. Mar. Ecol. Progr. Ser., 4, 207–212. Palmer, M.A. & Gust, G., 1985. J. mar. Res., 43, 179–210. Pamatmat, M.M., 1968. Int. Revue ges. Hydrobiol., 53, 211–298. Partheneides, E., 1986. In, Marine Interfaces Ecohydrodynamics, edited by J.C.J. Nihoul, Elsevier Oceanography Series, 42, Elsevier, Amsterdam, pp. 515–550. Pearse, J.S., 1979. In, Reproduction in Marine Invertebrates, Volume 5, Molluscs: Pelecypods and Lesser Classes, edited by A.C.Giese & J.S.Pearse, Academic Press, New York, pp. 27–85. Pearson, T.H., 1971. Vie Milieu, Suppl. No. 22, 53–91.

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

135

Pechenik, J.A., 1980. J. exp. mar. Biol. Ecol., 44, 1–28. Pechenik, J.A., 1984. J. exp. mar. Biol. Ecol., 74, 241–257. Petersen, C.G.J., 1918. Rept Dan. biol. Stn, 25, 1–62. Peterson, C.H., 1975. Ecology, 56, 958–965. Peterson, C.H., 1977. Mar. Biol., 43, 343–359. Peterson, C.H., 1986. Limnol. Oceanogr., 31, 200–205. Peterson, C.H. & Andre, S.V., 1980. Ecology, 61, 129–139. Peterson, C.H., Summerson, H.C. & Duncan, P.B., 1984. J. mar. Res., 42, 123– 138. Phillips, P.J., 1971. Gulf Res. Rept, 3, 165–196. Porter, J.W., 1974. Proc. 2nd Int. Symp. I. Coral Reef, Great Barrier Reef Committee, Brisbane, Australia, October 1974, pp. 111–125. Porter, J.W. & Porter, K.G., 1977. Limnol. Oceanogr., 22, 553–556. Porter, J.W., Porter, K.G. & Batac-Catalan, Z., 1977. Proc. 3rd Int. Coral Reef Symp., Miami, Florida, May, 1977, pp. 105–112. Pratt, D.M., 1953. J. mar. Res., 12, 60–74. Rasmussen, E., 1956. Biol. Meddel. Kongl. Dan. Vidensk. Selsk., 23(1), 1–84. Rees, E.I.S., Nicholaidou, A. & Laskaridou, P., 1977. In, Biology of Benthic Organisms, Proc. 11th Europ. Mar. Biol. Symp., edited by B.F.Keegan et al., Pergamon Press, Oxford, pp. 465–474. Reise, K., 1978. Helgol. wiss. Meeresunters., 31, 55–101. Reise, K., 1979. Helgol. wiss. Meeresunters., 32, 55–72. Reish, D.J., 1961. Calif. Fish Game, 47, 261–272. Rhoads, D.C., 1974. Oceanogr. Mar. Biol. Ann. Rev., 12, 263–300. Rhoads, D.C., Aller, R.C. & Goldhaber, M.B., 1977. In, Ecology of Marine Benthos, edited by B.C.Coull, University of South Carolina Press, Columbia, South Carolina, pp. 113–138. Rhoads, D.C. & Young, D.K., 1970. J. mar. Res., 28, 150–178. Rice, M.E., 1978. In, Settlement and Metamorphosis of Marine Invertebrate Larvae, edited by F.-S.Chia & M.E.Rice, Elsevier, New York, pp. 83–102. Richmond, R.H., 1985. Mar. Ecol. Progr. Ser., 22, 181–186. Richter, W. & Sarnthein, M., 1977. In, Biology of Benthic Organisms, Proc. 11th Europ. Mar. Biol. Symp., edited by B.F.Keegan et al., Pergamon Press, Oxford, pp. 531–539. Sameoto, D.D., 1969. J. Fish. Res. Bd Can., 26, 2283–2298. Sanders, H.L., 1956. Bull. Bingham oceanogr. Coll., 15, 345–414. Sanders, H.L., 1958. Limnol. Oceanogr., 3, 245–258. Sanders, H.L., Goudsmit, E.M., Mills, E.L. & Hampson, G.E., 1962. Limnol. Oceanogr., 7, 63–79. Santos, S.L. & Bloom, S.A., 1980. Oecologia (Berl.), 46, 290–294. Santos, S.L. & Simon, J.L., 1974. Bull. mar. Sci., 24, 669–689. Santos, S.L. & Simon, J.L., 1980a. Mar. Ecol. Progr. Ser., 2, 235–241. Santos, S.L. & Simon, J.L., 1980b. Mar. Ecol. Progr. Ser., 3, 347–355. Sarnthein, M. & Richter, W., 1974. Mar. Biol., 28, 159–164. Sarvala, J., 1971. Ann. Zool. Fennici, 8, 231–309. Sastry, A.N., 1979. In, Reproduction in Marine Invertebrates, Volume 5, Molluscs: Pelecypods and Lesser Classes, edited by A.C.Giese & J.S.Pearse, Academic Press, New York, pp. 113–292. Scheibling, R.E., 1980. Mar. Biol., 57, 95–105. Scheltema, R.S., 1961. Biol. Bull. mar. biol. Lab., Woods Hole, 120, 92–109. Scheltema, R.S., 1967. Biol. Bull. mar. biol. Lab., Woods Hole, 132, 253–265. Scheltema, R.S., 1974. Thalassia jugosl., 10, 263–296. Scheltema, R.S., 1986. Bull. mar. Sci., 39, 290–322. Schlichting, H., 1979. Boundary-Layer Theory, McGraw-Hill, New York, 817 pp.

136

CHERYL ANN BUTMAN

Schroeder, P.C. & Hermans, C.O., 1975. In, Reproduction in Marine Invertebrates, Volume 3, Annelids andEchiurans, edited by A.C.Giese & J.S.Pearse, Academic Press , New York, pp. 1–214. Segerstråle, S.G., 1960. Soc. Scient. Fenn. Comm. Biol, 23(2), 1–72. Segerstråle, S.G., 1962. Soc. Scient. Fenn. Comm. Biol., 24(7), 1–26. Seymour, M.K., 1972. Comp. Biochem. Physiol., 41A, 285–288. Shields, A., 1936. Calif. Inst. Tech., W.M.Kecklab of Hydraulics and Water Resources, Rept No. 167, translated by W.P.Ott & J.C.vanUchelen, 20 pp. Shin, P.K.S. & Thompson, G.B., 1982. Mar. Ecol. Progr. Ser., 10, 37–47. Sigurdsson, J.B., Titman, C.W. & Davies, P.A., 1976. Nature, Land., 262, 386–387. Smidt, E.L.B., 1951. Meddr Komm Danm. Fisk.-og Havunders., Ser. Fiskeri, 11(6), 1–151. Smith, F.G.W., 1946. Biol. Bull. mar. biol. Lab., Woods Hole, 90, 51–70. Smith, J.D., 1977. In, Marine Modeling, The Sea, Vol. 6, edited by E.D.Goldberg et al., John Wiley & Sons, New York, pp. 539–577. Smith, J.D. & Hopkins, T.S., 1972. In, Shelf Sediment Transport, edited by D.J.P. Swift et al., Dowden, Hutchison & Ross, Stroudsburg, pp. 143–180. Smith, J.D. & McLean, S.R., 1977. J. geophys. Res., 82, 1735–1746. Spärck, R., 1933. Meddr Grønland, 100(1), 38 pp. Stephen, A.C., 1933. Trans. R. Soc. Edinb., 57, 601–616. Strathmann, R.R., 1978. In, Settlement and Metamorphosis of Marine Invertebrate Larvae, edited by F.-S.Chia & M.E.Rice, Elsevier, New York, pp. 235–246. Suer, A.L. & Phillips, D.W., 1983. J. exp. mar. Biol. Ecol., 67, 243–259. Sulkin, S.D., 1984. Mar. Ecol. Progr. Ser., 15, 181–205. Switzer-Dunlap, M., 1978. In, Settlement and Metamorphosis of Marine Invertebrate Larvae , edited by F.-S.Chia & M.E.Rice, Elsevier, New York, pp. 197–206. Teal, J.M., 1958. Ecology, 39, 185–193. Thistle, D., 1983. Deep-Sea Res., 30, 1235–1245. Thistle, D., Yingst, J.Y. & Fauchald, K., 1985. Mar. Geol., 66, 91–112. Thistle, D., Reidenauer, J.A., Findlay, R.H. & Waldo, R., 1984. Oecologia (Berl.), 63, 295–299. Thomas, M.L.H. & Jelley, E., 1972. J. Fish. Res. Bd Can., 29, 1234–1237. Thorson, G., 1946. Meddr Kommn Danm. Fisk.-og Havunders., Ser. Plankton, 4, 1– 523. Thorson, G., 1950. Biol. Rev., 25, 1–45. Thorson, G., 1955. J. mar. Res., 14, 387–397. Thorson, G., 1957. In, Treatise on Marine Ecology and Paleocology. Vol. 1. Ecology, edited by J.W.Hedgpeth, Geol. Soc. Am. Memoir, No. 67, 461–534. Thorson, G., 1966. Neth. J. Sea. Res., 3, 267–293. Thorson, G. & Ussing, H., 1934. Meddr. Grønland, 100(3), 68 pp. Todd, C.D. & Doyle, R.W., 1981. Mar. Ecol. Progr. Ser., 4, 75–83. Tranter, D.J., Bulleid, N.C., Campbell, R., Higgins, H.W., Rowe, F., Tranter, H. A. & Smith, D.F., 1981. Mar. Biol., 61, 317–326. Trueman, E.R., 1971. J. Zool, 165, 453–469. Tyler, P.A. & Banner, F.T., 1977. Estuar. cstl mar. Sci., 5, 293–308. VanBlaricom, G.R., 1978. Doctoral dissertation, University of California, San Diego, 348 pp. VanBlaricom, G.R., 1982. Ecol. Monogr., 52, 283–305. Verwey, J., 1952. Arch, néerl. Zool., 10, 171–239. Virnstein, R.W., 1978. In, Estuarine Interactions, edited by M.L.Wiley, Academic Press, New York, pp. 261–273. Vogel, S., 1981. Life in Moving Fluids. Princeton University Press, Princeton, 352 pp. Watzin, M.C., 1983. Oecologia (Berl.), 59, 163–166. Watzin, M.C., 1986. J. exp. mar. Biol. Ecol., 98, 65–113. Webb, C., 1984. Doctoral dissertation, University of Wales, 272 pp.

LARVAL SETTLEMENT OF SOFT-SEDIMENT INVERTEBRATES

137

Webb, J.E. & Hill, M.B., 1958. Phil. Trans. R. Soc. Ser. B, 241, 355–391. Weinberg, J.R., 1979. Mar. Ecol. Progr. Ser., 1, 301–314. Wethey, D.S., 1986. Bull. mar. Sci., 39, 393–400. Whitelegge, T., 1890. Rec. Aust. Mus., 1, 41–54. Whitlatch, R.B., 1974. Biol. Bull. mar. biol. Lab., Woods Hole, 147, 227–235. Whitlatch, R.B., 1977. Biol. Bull. mar. biol. Lab., Woods Hole, 152, 275–294. Whitlatch, R.B., 1980. J. mar. Res., 38, 743–765. Whitlatch, R.B. & Johnson, R.G., 1974. J. Sediment. Petrol., 44, 1310–1312. Whitlatch, R.B. & Zajac, R.N., 1985. Mar. Ecol. Progr. Ser., 21, 299–311. Wieser, W., 1956. Limnol. Oceanogr., 1, 274–285. Wieser, W., 1959. Limnol. Oceanogr., 4, 181–194. Wigley, R.L. & McIntyre, A.D., 1964. Limnol. Oceanogr., 9, 485–493. Williams, A.B., 1958. Limnol. Oceanogr., 3, 283–290. Williams, A.B. & Porter, H.J., 1971. Chesapeake Sci., 12, 26–32. Williams, J.G., 1980. J. mar. Res., 38, 729–741. Wilson, D.P., 1932. Phil. Trans. R. Soc. Ser. B., 221, 231–234. Wilson, D.P., 1937. J. mar. biol. Ass. U.K., 22, 227–243. Wilson, D.P., 1948. J. mar. biol. Ass. U.K., 27, 723–760. Wilson, D.P., 1951. Année Biol., 55, 491–501. Wilson, D.P., 1952. Ann. Inst. océanogr. Monaco, 27, 49–156. Wilson, D.P., 1953a. J. mar. biol. Ass. U.K., 31, 413–438. Wilson, D.P., 1953b. J. mar. biol. Ass. U.K., 32, 209–233. Wilson, D.P., 1954. J. mar. biol. Ass. U.K., 33, 361–380. Wilson, D.P., 1955. J. mar. biol. Ass. U.K., 34, 531–543. Wilson, D.P., 1958. In, Perspectives in Marine Biology, edited by A.A.Buzzati-Traverso, University of California Press, Berkeley, pp. 87–103. Wilson, D.P., 1968. J. mar. biol. Ass. U.K., 48, 387–435. Wilson, D.P., 1970a. J. mar. biol. Ass. U.K., 50, 1–31. Wilson, D.P., 1970b. J. mar. biol. Ass. U.K., 50, 33–52. Wilson, D.P., 1977. J. mar. biol. Ass. U.K., 57, 761–792. Wilson, Jr, W.H., 1979. J. mar. Res., 37, 623–641. Wilson, Jr, W.H., 1981. J. mar. Res., 39, 735–748. Wilson, Jr, W.H., 1983. Ecology, 64, 295–306. Wimbush, M., 1976. In, The Benthic Boundary Layer, edited by I.N.McCave, Plenum Press, New York, pp. 3–10. Wolf, P. de, 1973. Neth. J. Sea Res., 6, 1–129. Wood, E.J.F., 1955. Aust. J. Sci., 18, 34–37. Woodin, S.A., 1974. Ecol. Monogr., 44, 171–187. Woodin, S.A., 1976. J. mar. Res., 34, 25–41. Woodin, S.A., 1978. Ecology, 59, 274–284. Woodin, S.A., 1979. In, Reproductive Ecology of Marine Invertebrates, edited by S. E.Stancyk, University of South Carolina Press, Columbia, South Carolina, pp. 99–106. Woodin, S.A., 1985. J. exp. mar. Biol. Ecol., 87, 119–132. Woodin, S.A. & Jackson, J.B.C., 1979. Am. Zool, 19, 1029–1043. Wooding, R.A., Bradley, E.F. & Marshall, J.K., 1973. Boundary-Layer Metero., 5, 285–308. Yaglom, A.M., 1979. Ann. Rev. Fluid Mech., 11, 505–540. Yalin, M.S., 1977. Mechanics of Sediment Transport. Pergamon Press, New York, 2nd edition, 298 pp. Yamaguchi, M., 1974. Micronesica, 10, 57–64. Young, D.K., Buzas, M.A. & Young, M.W., 1976. J. mar. Res., 34, 577–592. Young, D.K. & Rhoads, D.C., 1971. Mar. Biol., 11, 242–254.

138

CHERYL ANN BUTMAN

Zajac, R.N. & Whitlatch, R.B., 1982a. Mar. Ecol. Progr. Ser., 10, 1–14. Zajac, R.N. & Whitlatch, R.B., 1982b. Mar. Ecol. Progr. Ser., 10, 15–27. Ziegelmeier, E., 1978. Rapp. P.-v. Réun. Cons. int. Explor. Mer, 172, 432–444.

NOTE ADDED IN PROOF Regretfully, most of the papers from the Proceedings of the Invertebrate Larval Biology Workshop held at the Friday Harbor Laboratories, University of Washington, 26–30 March 1985, and published in Bull. Mar. Sci., 39(2) in 1986, were not available to the author at the time this manuscript was written. Many of these papers are relevant to the ideas discussed in this review.

Oceanogr. Mar. Biol. Ann. Rev., 1987, 25, 167–284 Margaret Barnes, Ed. Aberdeen University Press

APLYSIA: ITS BIOLOGY AND ECOLOGY THOMAS H.CAREFOOT Department of Zoology, University of British Columbia, Vancouver, B.C., Canada V6T 2A9

INTRODUCTION The sea hare Aplysia has been known in writings for 2000 years. From Pliny’s imaginative observations in the first century AD, which credited this novel but innocuous “fish” with a variety of poisonous and other abhorrent characteristics (Bostock & Riley, 1856, 1857), to the fanciful description by the Swedish ecclesiastic Olaus Magnus in the sixteenth century, “Lepus marinus formidabilis: The sea hare is of divers kinds in the ocean, but in that he resembles a hare by any token, so soon as he is caught he is let loose, for he is suspected to be venomous. He hath four ridges behind his head, two whose motion is as the fins along the length of a fish, and these are long, like to a hare’s ears; and two again from his back whose motion is as the fins on the belly of a fish, with which he beareth up the weight of his head. This hare doth cause terror in the sea; on land he is as the poor little hare, fearful and atrembling (Olaus Magnus, 1555)” to a plethora of contemporary papers on its neurobiology and behaviour, the genus has excited interest. The common name, sea hare, is derived from Pliny’s original designation Lepus marinus (Eales, 1921), so named because of the animal’s prominent rhinophores and hunched posture when resting (Fig. 1). Representatives of the genus inhabit shallow marine waters world-wide. Their diet is principally seaweeds, although it may include some angiosperms. Their choice of seaweeds is generally broad, their eating habits prodigious, and their growth rates fast. Most of their behaviour involves eating, copulating, and laying eggs. Aplysia is one of only a few marine invertebrates possessing a long-lived planktotrophic larva to be successfully cultured in the laboratory. This feature, combined with the species’ enormous fecundity and fast growth rate, promises a future abundance of experimental stock. Aplysia has been used as a research model in a variety of disciplines including the study of development, circadian rhythms, and hormonal control of reproduction. Neurobiologists have used Aplysia extensively in the study of nerve function and the neuronal bases of learning and behaviour. Its value in this respect lies in its neatly ordered, comparatively simple nervous system, which contains cells large enough to be penetrated easily with microelectrodes for neurophysiological study. A few relatively simple behavioural acts, such as the siphon- and gill-withdrawal reflexes, ink discharge, escape locomotion, and feeding, coupled with an accessible nervous system, have led to a number of fundamentally important discoveries on the neural integration of behaviour and on neural functions in learning (Kandel, 1979).

140

THOMAS H.CAREFOOT

Fig. 1.—The sea hare, Aplysia.

Two major reviews of Aplysia have been written. The first, by Eales in 1921, is a classic treatise on all aspects of the biology of the genus, from notes on natural history to detailed descriptions of fine morphological structure. This fine book has aided all of us in our dissections of sea hares. The second review, by Kandel in 1979, provides a comprehensive overview of neurobiological research as it pertains to Aplysia and other opisthobranch molluscs. The author successfully bridges two fields of study—nerve function and integration, and behaviour—using Aplysia as the chief model. The aim of integrating neuronal structure and function with an animal’s behavioural biology is foremost in the minds of all neurophysiologists, yet rarely has the scientific community witnessed advances in a research discipline so rapid or so exciting as those arising from neurophysiological studies on behaviour and learning in Aplysia (see Barnes, 1986). With these fine monographs on Aplysia, how does one justify a further review? The time seems to have come for a broad approach, to view the sea hare not strictly from the form and function standpoint of its organ systems, nor as a nerve-synapse model for neurobiological research, but as a whole organism interacting with its environment. It is my intention to draw from material on the general biology and ecology of Aplysia treated in these earlier reviews along with considerable work that was omitted from them or which appeared since 1978. Foremost examples of material not previously covered include swimming and the rôle of celestial cues in navigation, nutritional aspects of diet, and the chemical ecology of sea hares, although the list could be much extended. Major topics in the present review are distribution, life cycles, feeding and nutrition, growth, energetics, locomotion, predators and defence, and parasites. Neurophysiological topics are generally omitted, save where their exclusion would obscure understanding or deny insight into mechanisms underlying special aspects of the biology or ecology of sea hares. In this review I shall be treating only the genus Aplysia under the designation sea hares, not other members of the Family Aplysiidae such as Bursatella, Dolabella, and Stylocheilus which are sometimes referred to by the same common name.

APLYSIA

141

DISTRIBUTION WORLD-WIDE Some 37 species of Aplysia live in the world’s oceans (Eales, 1960; Bebbington, 1974, 1975, 1977). Many are endemic, occupying relatively small geographical areas (e.g. A. vaccaria in California and Baja California, and A. gracilis in the Red Sea), or are restricted to defined oceanic basins (e.g. A. brasiliana=willcoxi— synonymized by Strenth & Blankenship, 1978b—in the tropical Atlantic Oceans, Gulf of Mexico, and Caribbean Sea). Only a few are truly cosmopolitan. Most species are restricted to warm waters ranging from approximately 40° N to 40° S (Fig. 2). A. punctata is unusual in that its distribution extends into the Arctic Circle in Norway, making it the most polar of all sea hares (Eales, 1960). No species is known from Antarctica. Some 26 species occur in the Pacific Ocean regions, nine in the Indian Ocean, 13 in the Atlantic Ocean, and five each in the Mediterranean and Red Seas, with overlap between them. Only three species are cosmopolitan throughout the world. These are A. dactylomela, A. parvula, and A. juliana. A. juliana possesses the greatest distributional range of all sea hares, circumtropical, from 42° N-46° S, and is found in all major ocean basins in the world (Macnae, 1955; Eales, 1960; Bebbington, 1975; Willan & Morton, 1984). Australia, with 14 species, appears to have the richest diversity. New Zealand has at least six species, including A. dura, which has an interesting distribution in that it is known only from Cook Strait in New Zealand and the island of Tristan da Cunha in the southern Atlantic Ocean, both locations falling on or about the 40° S line, but separated by several thousand kilometres of open ocean. What factors govern the distribution of sea hares? The species occupy similar habitats, eat the same sorts of foods, have similarly fast growth rates and annual life cycles. They are all relatively large with concomitant high fecundity which, coupled with the possession of a long-lived planktotrophic veliger larva, provides the potential for widespread dissemination in all species. It is no surprise, then, to find several species with distributions encompassing oceanic basins or along areas subject to strong current transport. The distribution of A. brasiliana=willcoxi, for example, may be readily explained by larval transport in major current systems such as the Florida and Gulf Stream Currents. It inhabits a broad area on the eastern and southeastern coasts of North America from southern Texas, around the coast of Florida, north to Woods Hole and Martha’s Vineyard in Masssachusetts (Sanford, 1922; Merriman, 1937), and south to Guadaloupe in the West Indies (Krakauer, 1969). Similarly, the existence of A. dura at widely disparate sites at Tristan da Cunha and Cook Strait in New Zealand (Eales, 1960) can be explained by the location of both regions in an almost contiguous water mass and the possible effect of the northernmost boundary of the Antarctic Circumpolar Current in transporting larvae. Of the five common aplysiids in Hawaii, three are represented by the cosmopolitan A. dactylomela, A. parvula, and A. juliana. The two remaining species, A. pulmonica and A. oculifera, have widespread distributions throughout the Pacific area (Kay, 1964). Bermuda has three species, A. parvula, A. dactylomela, and A. morio, the first two with wide distributions throughout the world’s oceans, the third, with a distribution extending from the eastern seaboard of the U.S. through the Gulf of Mexico and Caribbean Sea to southern Brazil. With regard to the distribution of Bermudan opisthobranchs, Clark (1984) notes that the densest populations tend generally to be species with short pelagic development. In comparison, ones with extended pelagic development, including sea hares, are sparsely spread but rich in numbers of species. Clark suggests that recruitment of species with long-lived pelagic larvae in Bermuda may be via larvae originating in Bahamian or other Caribbean islands and, in fact, the Bermudan populations may never undergo self-recruitment because of offshore transport of larvae. Larval transport in ocean currents may not be the only means of dispersion in sea hares. Eales (1960) has noted that A. euchlora can hitch rides on floating patches of seaweeds or other debris. Perhaps juveniles or

Fig 2.—World distributions of sea hares: see Table I for designation of each symbol.

142 THOMAS H.CAREFOOT

APLYSIA

143

adults of other species can do the same. Seven or eight species are capable of swimming, although energetic demands would preclude any but short bouts of this activity (cf. Hamilton & Ambrose, 1975). Larval transport in ballast-water of ships, and possibly even in aircraft (seaplanes) in modern times, has been documented by Carlton (1985) for many species of benthic marine invertebrates. This disseminating mechanism offers an intriguing alternative explanation for the world-wide distribution of various ‘cosmopolitan’ species of invertebrates including, perhaps, sea hares. One would therefore anticipate their wide-spread dispersion, but this does not explain sympatry of circumglobal and local species. Eales (1960) addresses this question and, while noting that no ready explanation is available, suggests that more extensive collecting would probably correct a number of anomalies in the distributional records of presumed local species. The jointure of the Mediterranean and Red Seas through the Suez Canal has stimulated thoughtful discussion of possible migratory routes in the evolution of Aplysia by Eales (1960, 1970) and Bebbington (1975). Five species occur in the Mediterranean and five in the Red Sea and Gulf of Aden, with overlap between A. fasciata and A. parvula. Based on the distributional patterns of these two species, Eales (1960, 1970) suggests that A. fasciata may have reached the Red Sea from the north, while A. parvula, commonly represented in all reaches of the Indian Ocean (Eales, 1944; Bebbington, 1974), may have migrated in the reverse direction, passing through the Red Sea into the Mediterranean. In theory then, both species migrated through the Suez Canal, probably as larvae, although other means of transport such as on ships’ bottoms would not be out of the question. Interestingly, such migrations may have been comparatively slow from the Canal’s opening in 1869, but more rapid in recent years (Eales, 1970). Originally, the highly saline Great Bitter Lake (about 68‰ in 1870; Oren, 1970) offered a natural barrier to all but the most hardy migrants. With time, and mainly through slow dissolution and eventual loss of a thick crust of salt at the bottom of the lake, the salinity has been reduced to about 47‰, offering less of a barrier to such migratory movements (Oren, 1970). At the same time, through the reduced flow of the River Nile following completion of the Aswan Dam, the salinity of the eastern Mediterranean increased, thereby narrowing the difference in salinity between the Mediterranean and Red Seas; the salinity of these seas is now approximately respectively. This may have reduced the salinity shock to potential marine invertebrate migrants, particularly in their more susceptible larval stages, and possibly have facilitated more recent migratory movements (Eales, 1970). Bebbington (1975), on the other hand, offers a counter-proposal for the distributional patterns of Mediterranean and Red Sea Aplysia. He contends that A. parvula may be a relict species from the ancient Tethys Sea (notwithstanding the contention by Hsü, 1972, that the Mediterranean Sea was completely dry about six million years ago) or, alternatively, that it could have migrated into the Mediterranean Sea from both west, through the Strait of Gibraltar, and east, via the Suez Canal. The issue may never be sorted out satisfactorily since Bebbington (1975) notes that, in the past, the similarities between A. parvula and its close relatives, such as A. punctata, have led to a number of misidentifications in early collecting records (e.g. Swennen, 1961). This confusion, combined with the incompleteness of these records, precludes a thorough historical investigation and, unfortunately, the truth may forever have been obscured. LOCAL HABITATS Sea hares are characteristically found in shallow marine waters less than 5 m in depth. There appears to be no general preference for intertidal over sub tidal habitats and many species occupy both. As can be seen from the habitat descriptions in Table I, data are complete for only a few species. Where these are detailed and cover a major part of the geographical range of a species, as for example A. californica, A. dactylomela,

144

THOMAS H.CAREFOOT

and A. parvula, it is apparent that habitat preferences are broad. A. dactylomela, for example, is found in such diverse habitats as rocky shores (Usuki, 1970b), subtidal coral rubble and sand (Carefoot, 1970, 1985), eelgrass beds (Kay, 1964; Neck, 1976; Strenth & Blankenship, 1977), and intertidal pools (Willan, 1979). This wide preference for habitats is also true for A. californica and A. parvula. Thus, sea hares with specific habitat requirements seem to be rare. Some species do appear, however, to have unusual preferences. For example, Eales (1960) notes that A. euchlora favours a habitat of floating algae, although perhaps too few specimens have been studied to consider this the species’ only habitat. Similarly, both Kay (1964) and Lance (1971) ascribe a retiring habit to A. parvula. In Hawaii and the Gulf of California, respectively, populations are found buried in sand-algal mats or entangled in long strands of algae, and are rarely found crawling openly, although general habitat preferences within this restricted environment may be broad for this species: amongst the eelgrass Zostera (Bebbington, 1977), in coral fragments (Marcus, 1958), on waveexposed rocky shores (Usuki, 1970b), on wave-exposed rocky coasts amongst the red algae Laurencia spp. and Plocamium costatum (Willan & Morton, 1984), and clinging to algae in wave-washed environments (Switzer-Dunlap, pers. comm.). Finally, Aplysia juliana is always found in close association with its food, the green alga Ulva and yet, given this constraint, habitat occupation by this species may likewise be rather broad (Table I). Zostera and other eelgrass habitats seem to be favoured by a number of species (e.g. Aplysia brasiliana=willcoxi, A. californica, A. cervina, A. dactylomela, A. extraordinaria, A. maculata, and A. tanzanensis), even though other habitats may be occupied by them. Finally, the data in Table I support to some extent the hypothesis put forward by Achituv & Susswein (1985) that the closely related species A. cedrosensis, A. depilans, A. dura, A. juliana, A. nigra, and A. vaccaria are found in more rough-water habitats compared with other species. Of the three representatives of this group for which habitat characteristics are reasonably well known, two appear to favour such wave-exposed environments (A. depilans and A. vaccaria), while A. juliana is occasionally found in them. Some otherwise shallow-water species appear to have set considerable depth records (Table I). Kay (1964), for example, reports that a specimen of A. dactylomela was dredged from 182 m off the coast of Hawaii, and Marcus (1972) reports that A. brasiliana=willcoxi and A. morio were trawled from depths of 350 m off the São Paulo coast. The digestive tracts of these bathophilic Brazilian sea hares were apparently filled with freshly ingested seaweeds (the red algae Acrochaetium, Polysiphonia, Hypoglossum, and Dasya, and the brown alga Dictyota; Marcus, 1972). The occurrence at such great depths of healthy algal communities (let alone the sea hares themselves), is questionable in that it would represent a new depth record for seaweeds, some 80 metres deeper than any macrophyte yet recorded (Littler, Littler, Blair & Norris, 1985). Burrowing appears to be a common behaviour in certain species, namely Aplysia brasiliana=willcoxi (Aspey & Blankenship, 1976a, b; Krakauer, 1969), A. dactylomela (Carefoot, 1970), A. geographica (Willan, pers. comm.), and A. juliana (Frings & Frings, 1965; Sarver, 1978) and is, of course, associated with a habitat of soft or sandy sediments. The function of burrowing is not clear. In A. dactylomela its association with nocturnal activity and daytime quiescence (in under-rock crevices or buried, or both, in populations in Barbados and Hawaii), suggests that it may be a way to avoid possible harmful effects of ultraviolet or other radiation in its often extremely shallow habitats (Switzer-Dunlap, pers. comm.). Susswein, Gev, Achituv & Markovich (1984a) also noted that A. fasciata, which is not an habitual burrower, may do so to escape the effect of light; it never burrows at night. A further observation in this regard is that A. dactylomela in subtidal locations near Wakayama, Japan, may sit quietly on open rock surfaces during daylight hours in late spring, but only at depths well below extreme low water (Carefoot, pers. obs.). Burrowing may also protect the sea hare from especially high wave surge during storms (Hamilton, Russell & Ambrose, 1982), from desiccating effects of air during exposure at low tide (A.

APLYSIA

145

californica: Wright, D.L., 1960), or may be related in some way to reproduction (Aspey & Blankenship, 1976a). Possible factors influencing habitat selection by sea hares include food, degree of wave exposure, predators, sunlight, crevices, desiccation, substratum type (e.g. sediments for burrowing), and salinity. Of these, the influence of food is most extensively documented. It is now well known that certain marine algae which provide for good nutrition and growth of postmetamorphic sea hares act as settlement inducers for these same species as larvae (Kriegstein, Castellucci & Kandel, 1974; Switzer-Dunlap & Hadfield, 1977; SwitzerDunlap, 1978). Thus, settlement and recruitment are enhanced TABLE I Habitat and depth distributions of Aplysia: the “general references” pertain to world distributions as shown in Fig. 1; A. willcoxi has been synonymized with A. brasiliana (Strenth & Blankenship, 1978b); A. rehderi is cited by Eales (1960) as being from Monterey, California (a single specimen), the validity of this record has been questioned by Beeman (1963); the “average” and “maximum” depths presumably relate to a zero datum point, but this is almost never mentioned by the authors Species

Symbol used in Fig. 1

Habitat

Average depth (m) Maximum depth (m)

General references

brasiliana

Intertidal amongst Spyridia and Ulva (Texas: Cobbs & Pinsker, 1982); protected rock jetties (South Carolina: Fox & Ruppert, 1985)

Intertidal to 14

350(?) 9–14

californica

Rocky intertidal areas, tide-pools, channels, protected bays (Southern California, Baja: Kupfermann & Carew, 1974); rocky coast areas, bays, estuaries (California: Beeman, 1961); shallow lagoon (California: Winkler & Dawson, 1963); wave-exposed rocky coast areas or sandy beaches or sheltered

Intertidal to 5

30

Marcus & Marcus, 1955; Eales, 1957, 1960; Marcus, 1958, 1972, 1976; Breuer, 1962; Krakauer, 1969; Tunnell & Chaney, 1970; Hamilton & Ambrose, 1975; Strenth & Blankenship, 1977 Cooper, 1863; MacGinitie, 1934; Eales, 1957, 1960; Winkler, 1958b, c, 1959c; Wright, H.O. 1960; Beeman, 1961, 1963; Marcus, 1961; Audesirk, 1976, 1979; Bebbington, 1977

146

THOMAS H.CAREFOOT

Species

Symbol used in Fig. 1

Habitat

Average depth (m) Maximum depth (m)

General references

mudflats (California: MacFarland, 1966; Ricketts & Calvin, 1968); tide-pools (California: Chapman & Fox, 1969); low intertidal Phyllospadix torreyi beds during mating (California: Audesirk, 1976) cedrosensis

cervina

Sea-grass flats in large bays (Columbia: Bandel, 1976)

“Shallow”

Eel-grass beds, gravelly or silty sand with Padina or red algae (Hawaii: Kay,

Intertidal to 8

corrigera

cronullae dactylomela

1964); rocky shores (Japan Sea: Usuki, 1970b); intertidal eel grass beds (Texas: Neck, 1976; Strenth & Blankenship, 1977); shallow-water Thalassia beds (Bimini: Lederhendler & Tobach, 1977; Puerto Rico:

182(?) 17

Bartsch & Rehder, 1939; Eales, 1960; Beeman, 1963; Bebbington, 1977 Eales, 1957, 1960; Marcus & Marcus, 1958, 1960, 1962; Warmke & Abbott, 1961; Marcus, 1972, 1976; Bandel, 1976 Eales, 1944, 1957, 1960; Bebbington, 1974, 1977 Eales, 1960; Bebbington, 1977 Engel, 1928; Baba, 1936, 1937, 1970; Eales, 1944, 1957, 1960; Macnae, 1955,

1957; Marcus & Marcus, 1955, 1960, 1962; Breuer, 1962; Marcus, 1972; Bebbington, 1974, 1975; Marcus & Hughes, 1974; Neck, 1976; Tobach, 1978; Willan, 1979; Wells & Hill, 1980, 1985

APLYSIA

147

Lederhendler et al., 1977); rocky, intertidal beach with Laurencia papillosa and Sargassum sp. (Lederhendler, 1977); shallow, sandy reef lagoon with Thalassia, Ulva, Acanthophora (Puerto Rico: Tobach, 1978); intertidal pools (New Zealand: Willan, 1979); shallow sea-grass flats (Florida Keys: DiMatteo, 1981b); shallow flats with Thalassia testudinum and Champia parvula (Florida Keys: DiMatteo, 1982); amongst coral rocks in areas of loose sand and fine rubble (Barbados: Carefoot, 1985) denisoni depilans

Exposed and protected rocky shores, rock platforms, ports (Israel: Achituv & Susswein, 1985)

dura euchlora extraordinaria fasciata

Zostera (New South Wales, Australia: Bebbington, 1977) Estuarine flats (South England: Grigg, 1949); exposed and protected rocky shores, rock platforms, ports (Israel: Susswein et al., 1984a; Achituv & Susswein, 1985)

gigantea gracilis inca juliana

Amongst Ulva, Gracilaria (Brazil: Marcus & Marcus, 1955); exposed, freshwater run-off or seepage, with Ulva (Hawaii: Van Weel, 1957; Sarver, 1978); associated

Eales, 1960; Bebbington, 1977 1·5–10 “Shallow” 20–30 Grigg, 1949; Pruvot-Fol, 1954; Eales, 1957, 1960; Bebbington, 1972, 1975; Thompson, 1976; Gev et al., 1984 Eales, 1960; Bebbington, 1977 Eales, 1960; Baba, 1970; Bebbington, 1977 Allan, 1932; Eales, 1957, 1960 Intertidal 0·5–1·5 9–24 20–30 Grigg, 1949; Pruvot-Fol, 1954; Eales, 1957, 1960, 1970; Barash & Danin, 1971; Bebbington, 1972, 1975; Marcus, 1972; Bebbington & Hughes, 1973; Thompson, 1976; Barash & Zenziper, 1980; Gev et al., 1984 Eales, 1960; Bebbington, 1977 Eales, 1960 Eales, 1957, 1960; Bebbington, 1977 1–1·5 “Shallow subtidal” Edmundson, 1946; Pilsbry, 1951; Macnae, 1955; Marcus & Marcus, 1955; Eales, 1957, 1960; Engel & Eales, 1957; Marcus, 1958, 1972, 1979;

148

THOMAS H.CAREFOOT

TABLE I—continued Species

Symbol used in Fig. 1 Habitat

Average depth (m) Maximum depth (m)

Saito & Nakamura, 1961; Bebbington, 1972, 1977; Baba, 1970; Barash & Danin, 1971; Sarver, 1978; Marcus & Hughes, 1974; Willan & Morton, 1984

with Ulva and Enteromorpha (Hawaii: Kay, 1964); near freshwater source, on Ulva-covered rocks (Barbados: Carefoot, 1970); open rocky coasts, tide-pools (Japan Sea: Usuki, 1970b); between rocks (Israel: Barash & Danin, 1971); juveniles on Monostroma angicava in intertidal area and on Enteromorpha compressa in rock-pools (Japan Sea: Usuki, 1981a) keraudreni

kurodai

maculata

morio

3·5

Open or sheltered rocky coasts (Japan Sea: Usuki, 1970b); shallow subtidal area amongst red algae, protected rocky embayment (Japan Sea, also Wakayama and Ise: Carefoot, pers. obs.) Estuaries (South Africa: Macnae, 1955); amongst Zostera (Zanzibar: Bebbington, 1974)

General references

18

“Shallow subtidal”

Eales, 1960; Bebbington, 1977; Willan & Morton, 1984 Eales, 1957, 1960; Baba, 1970; Bebbington, 1977

Eales, 1957, 1960; Bebbington, 1974

350(?) 42

Eales, 1960; Marcus, 1972;

APLYSIA

Species

Symbol used in Fig. 1 Habitat

Average depth (m) Maximum depth (m)

nigra oculifera

Frontal edges of waveswept limestone beaches (Hawaii: Kay, 1964)

parvula

Buried in sandalgal mat or in Sargassum, Hypnea or Spyridia (Hawaii: Kay, 1964); open rocky coasts (Japan Sea: Usuki, 1970b); buried in sand with Padina durvillaei (Gulf of California: Lance, 1971); on holdfast of Laminaria digitata, subtidal (South England: Bebbington &

pulmonica

punctata

Brown, 1975); subtidal in mixed algal assemblage under kelp canopy (New Zealand: Willan, 1979); open rocky areas, on Laurencia spp. in pools on the middle shore, subtidally on Plocamium costatum (New Zealand: Willan & Morton, 1984) Seaward edge of fringing reef, frontal slope of waveswept detrital beach (Hawaii: Kay, 1964) Intertidal on Fucus serratus, subtidal on Dictyota and Desmarestia (Irish Sea: Miller, 1960); subtidal on Plocamium cartilagineum (Irish Sea: Carefoot, 1967c); intertidal, rocky areas with Fucus, Enteromorpha and

“Shallow intertidal 0–10

40 80(?)

“Shallow”

0–1·5 13 below MLWS

149

General references Strenth & Blankenship, 1977 Eales, 1960; Bebbington, 1977 Eales 1944, 1957, 1960; Macnae, 1957; Kay, 1964; Baba, 1970; Usuki, 1970b; Bebbington, 1977 Baba, 1937, 1970; Eales, 1957, 1960, 1970; Marcus, 1958, 1972; Warmke & Abbott, 1961; Kay, 1964; Barash & Danin, 1971; Bebbington, 1972, 1975, 1977; Bebbington & Brown, 1975; Willan, 1979; Clark, 1984; Gev et al., 1984; Willan & Morton, 1984

Eales, 1957, 1960; Bebbington, 1977; SwitzerDunlap (pers. comm.) 18–29 4·3

Thorson, 1946; Grigg, 1949; Eales, 1957, 1960; Miller, 1960; Bruce et al., 1963; Carefoot, 1967c; Barash & Danin, 1971; Bebbington, 1972, 1975; Poizat & Vicente, 1977

150

THOMAS H.CAREFOOT

Ulva predominant (France, Atlantic: Otsuka et al., 1980) rehderi

Eales, 1960; Beeman, 1963; Bebbington, 1977 Eales, 1960; Bebbington, 1977 Beeman, 1960, 1961, 1963; Bebbington, 1977 Eales, 1960; Bebbington, 1977 Bebbington, 1974 Eales, 1960; Baba, 1970; Bebbington, 1977 Eales, 1960; Bebbington, 1977 Eales, 1960; Bebbington, 1977 Bebbington, 1974

reticulata reticulopoda

Subtidal zone (Southern California: Beeman, 1960)

2

robertsi rudmani sagamiana sowerbyi sydneyensis tanzanensis vaccaria

willcoxi

winneba

Eel grass (Tanzania: Bebbington, 1974) Rocky coasts in kelp beds (California: Beeman, 1961); Egregia beds (Southern California: Winkler & Dawson, 1963); rocky intertidal area (Gulf of California: Lance, 1967) Eel-grass beds (South Texas: Neck, 1976)

1–3

12–15

Winkler, 1955; Beeman, 1963; Eales, 1960; Wright, D.L., 1960; Winkler & Dawson, 1963; Limbaugh in Winkler & Dawson, 1963; Lance, 1967; Bebbington, 1977 Sanford, 1922; Merriman, 1937; Hackney, 1944; Zinn, 1950; Eales, 1960; Moore, 1961; Breuer, 1962; Tunnell & Chaney, 1970 Eales, 1960

in areas rich in the pertinent foodstuff and may be virtually lacking in areas without. Yet, only rarely are sea hares restricted for food to their original settlement-inducing algal species (e.g. as is A. juliana). Broad preferences for food are the rule, not the exception, and the tastes of some species seem to broaden as the animals grow. A specific food is, therefore, unlikely to be an important factor in influencing habitat selection in most species of Aplysia. In any case, ready access to different habitats is denied most species as adults, even supposing that they were capable of perceiving other distant, richer feeding grounds. Where suitable foods are absent or over-grazed, swimming may provide the means in a few species to sample new environments (Achituv & Susswein, 1985). This suggestion is consistent with the observation that starvation induces swimming in several species of Aplysia (A. fasciata: Susswein, 1984; A. pulmonica: Switzer-Dunlap, pers. comm.). Indeed, a fully-fed A. pulmonica appears to be unable to swim at all; whether this is due to the actual load of food being carried or to a suppression of swimming behaviour when satiated is not known. Food was initially implicated as the stimulus for daily migration in a small population of A. dactylomela in Barbados from a red-algal dominated assemblage some 8–10 m offshore where the animals rested during

APLYSIA

151

daylight hours, to a green-algal dominated assemblage close inshore where the animals spent their pre-dawn hours feeding (Carefoot, 1985). Comparison of energy content and nutritional value of the seaweeds (as measured by their growth-promoting quality) in each area suggested, however, that the animals would have done better to remain in the offshore habitat. For whatever reason, the animals made the daily migration to the feeding grounds. Absence of resting niches and risk of air exposure or wave damage during extreme low tides in the close-inshore habitat may ultimately have forced their pre-dawn migration. Less attention has been given to other factors which might influence habitat selection in sea hares. The animals generally avoid areas of extremely low salinity due to either intolerance of them or to lack of suitable foods there. Through a sensory ending in the osphradium (Stinnakre & Tauc, 1969), Aplysia has the ability to sense and possibly to distinguish between waters of different salinities. The response of Aplysia to diluted sea-water bathing the osphradium is an inhibition of firing of cell R15 in the abdominal ganglion (Stinnakre & Tauc, 1969). This special cell is known to secrete an antidiuretic substance (Kupfermann & Weiss, 1976) which may act as a hormone to regulate water balance. While the receptors in the osphradium which trigger the R15 response are quite sensitive to a diminution in salinity, they do not appear to be as sensitive to increases in salinity (Stinnakre & Tauc, 1969). Thus, while a possible mechanism for regulating water or ionic content has been identified in Aplysia, its precise function remains unclear. Experiments to determine osmoregulatory capabilities in sea hares have led to conclusions that they are either classic “osmo-adjusters” (A. punctata: Bethe, 1929, 1930, 1934) or that they can only weakly osmoregulate (A. juliana: van Weel, 1957; see also Kupfermann & Weiss, 1976; Lombardini, Pang & Griffith, 1979). The particular ecological characteristics associated with the two species used may explain why these studies result in different conclusions. Whereas A. punctata is not normally found in areas with brackish sea water, A. juliana is always found associated with its food, Ulva spp. (Sarver, 1978), which grows, characteristically, near freshwater run-off or seepage. Thus, osmoregulation might be selected for in this species over Aplysia punctata, which would be unlikely to encounter dilute sea water so regularly. Even though van Weel (1957) demonstrated a weak ability in A. juliana to regulate the concentration of its body fluids, he also found that exposure of A. juliana for only 5·5 h to 80% sea water inflicted irreparable tissue damage. Intertidal habitats are neither favoured nor avoided by Aplysia. Thus, although a number of species may be found commonly in isolated tidal pools or amongst tumbled rocks and seaweeds at low tide (e.g. A. punctata: Carefoot, 1967c; A. kurodai: Usuki, 1970b; A. californica: Kupfermann & Carew, 1974; A. brasiliana=willcoxi: Cobbs & Pinsker, 1982; see also Table I), other species only rarely occupy this habitat. For example, of 1098 A. fasciata observed in shallow-water habitats in Israel by Susswein et al. (1984a), only eight were found in tidepools or exposed to air. These eight animals were all dead or dying. The effects of desiccation are minimized where seaweeds are present, but death can occur within a few hours when animals are stranded on areas of open sand beach or in eelgrass beds in summer (e.g. A. brasiliana=willcoxi: Krakauer, 1969). As part of a comprehensive field study of behavioural patterns of A. californica in the La Jolla region of southern California, Kupfermann & Carew (1974) noted that 50% of the intertidal animals observed were exposed to air for more than 4 h and 16% were exposed for more than 6 h. No data are available on the effect of desiccation on mortality in sea hares. Lack of oxygen during exposure to air is another factor which may be critical in the survival of Aplysia stranded during periods of low tide. Bradycardia, noted in sea hares as a response to the stress of exposure to air, may reduce metabolic activity (Dieringer, Koester & Weiss, 1978). The onset of bradycardia occurs within 5 min of exposure or less and can reduce the heart rate by 16·5 and 39–42% in laboratory A. californica and A. brasiliana=willcoxi, respectively, and by up to 40% in field A. brasiliana=willcoxi (Pinsker, Feinstein & Gooden, 1974; Feinstein, Pinsker, Schmale & Gooden, 1977; see also Feinstein,

152

THOMAS H.CAREFOOT

Fig. 3.—Patterns of daily activity of A. fasciata in Israel and of A. dactylomela in Barbados: percentage of activity in A. fasciata calculated from data in Susswein et al. (1983) by subtracting “% immobile” from 100 for each time indicated; data for A. dactylomela represent total number of animals observed in a visual sweep 1 m on either side of a 40-m long transect line.

Aspey & Schmale, 1976). Unfortunately, these studies on heart rate in Aplysia have been conducted over short periods of a few minutes only, and it is not known whether the response would last over the several hours of possible air exposure occurring during a normal tidal cycle. An ability to take up oxygen directly from air has not been reported for Aplysia. Neither the ctenidium nor the structure of the mantle cavity area seem to be particularly well suited for this function. Distinct habitat preferences have been shown for Mediterranean populations of A. fasciata and A. depilans by Achituv & Susswein (1985). On the Mediterranean coast of Israel these species coexist for at least part of their lives in such habitats as exposed and protected rocky shores, rock platforms, and protected port areas. They eat the same foods, are found amongst the same rocks and crevices, and spawn in the same areas. With these same general requirements the possibility for competition exists, and the authors have shown that their habitats segregate along a gradient of wave-action, with A. depilans dominating in areas of greater wave action and A. fasciata in areas of calmer water. Although direct evidence of competition was not presented in the study, the authors did note that in late summer, when A. depilans was absent, A. fasciata entered some habitats previously dominated by the former species. Competition for space was noted between A. dactylomela and the sea urchin Evechinus chloroticus by Willan (1979) in New Zealand. Both species seek out the same refuge space in depressions and beneath ledges and stones especially during rough weather. Willan believes that the small size and ‘spinyness’ of the urchin makes it a superior competitor for the few shelters available in the area studied. A number of sea hares are active nocturnally and quiescent during the day. This behaviour is well documented for Aplysia dactylomela (Carefoot, 1970; Switzer-Dunlap & Hadfield, 1979) and A. fasciata (Susswein et al., 1983, 1984a; Susswein, 1984), and has also been observed in A. juliana (Carefoot, 1970; Sarver, 1978; Switzer-Dunlap & Hadfield, 1979), A. vaccaria (Eales, 1960), and A. brasiliana=willcoxi (Krakauer, 1969; Cobbs & Pinsker, 1982). In this regard, A. dactylomela and A. fasciata are remarkably similar in their patterns of behaviour (Fig. 3). Both species are active at night and inactive to varying extents during daylight hours. In Barbados, A. dactylomela is completely quiescent during the day, hiding under rocks and in crevices. In Israel, A. fasciata is generally inactive during the day but does crawl, swim, and feed a little, especially just after dawn, and begins to copulate in the afternoon during breeding season. Both species

APLYSIA

153

copulate extensively just after dark, although for A. fasciata this is just an extension of late afternoon copulatory activity. Finally, both species feed mostly in the hours preceding dawn, with A. fasciata extending this behaviour into the morning hours but at much lower intensity. There seems to be no explanation for a nocturnal as opposed to a diurnal activity pattern in sea hares. Avoidance of daytime predators is one obvious possibility, but sea hares appear to have no major predators (see pp. 250– 252). For animals living in shallow water in subtropical or tropical areas (5 cm–2 m for A. fasciata in Israel: Susswein, Gev, Feldman & Markovich, 1983; Susswein et al., 1984a; 1–2 m for A. dactylomela in Barbados: Carefoot, 1985) avoidance of harmful effects of sunlight is another possibility. Yet, while A. dactylomela in Barbados actively seek out protective crevices and areas under rocks in which to hide during the day, often becoming completely silted over in the process, A. fasciata in Israel appear not to engage in such wholesale secretive behaviour. For example, in a population of 190 A. fasciata observed in early afternoon by Susswein et al. (1984a), only nine individuals were found to be partially or completely buried in sand. Also, many other species of Aplysia in tropical and subtropical habitats are active during the day in shallow waters and are openly exposed to sunlight (e.g. A. brasiliana=willcoxi, A. californica, and A. oculifera). Too few data are available at the present time on activity patterns in field Aplysia to allow more than just speculation on this intriguing question. LIFE CYCLE Sea hares have basically an annual life cycle (Miller, 1960; Carefoot, 1967c; Krakauer, 1969; Usuki, 1970b; Audesirk, 1976, 1979). Life spans may, however, vary from 3–7 months in some species (Sarver, 1978; Usuki, 1970b; Switzer-Dunlap & Hadfield, 1979; Gev, Achituv & Susswein, 1984) to as much as 16 months, the longer time depending on whether recruitment is in summer or autumn (Carefoot, 1967c). The general pattern of life cycle is similar for all Aplysia. During the breeding season several or many animals may come together to form aggregations. Sea hares are hermaphroditic but function as either sperm donors or sperm recipients during copulation. Eggs are deposited through the warmer months from March-October in the Northern Hemisphere. After 6–15 days, depending on temperature and other local factors, the eggs hatch into planktotrophic larvae which swim and grow in the plankton for a minimum of about four weeks (based on data from laboratory culture). After this time the larvae become competent (i.e. capable of settling and metamorphosing) and they settle and metamorphose when currents carry them near the plant species which will comprise their principal food in the juvenile and often later adult stages. Growth is rapid and reproductive maturity is reached early. Most of the active hours after maturity are spent eating, copulating, and laying eggs. Little is known of the factors leading to senescence in sea hares. BREEDING AGGREGATIONS Animals may congregate in sometimes large groups during breeding season. It is not clear whether all such aggregations relate strictly to breeding, as sometimes huge aggregations happen after a particularly successful settlement, when the large number of individuals eat all available food and start moving towards new sources. Breeding aggregations may be only a few animals or may consist of several tens of individuals (MacGinitie, 1935; Wright, H.O., 1960; Kupfermann & Carew, 1974; Tobach, 1978; Audesirk, 1979; Susswein et al., 1984a; Achituv & Susswein, 1985). Groups initiate generally in the springtime at the onset of breeding and may extend through the summer (Audesirk, 1979), or may break up and re-form later in the autumn (Usuki, 1970b). In some species, however, winter breeding has been observed or suggested (A. californica: Audesirk, 1976; Dudek, Soutar & Tobe, 1980b; A. kurodai: Usuki, 1970b). Animals sometimes

154

THOMAS H.CAREFOOT

remain for a week or more in such aggregations (Kupfermann & Carew, 1974) and some localized migratory movements may be involved to bring the animals together. For example, at sites on Catalina Island studied by Audesirk (1979), A. californica aggregate in beds of surf-grass Phyllospadix torreyi in early spring. This apparently involves some movement from deeper sites offshore. By midsummer (JulyAugust) in the Catalina Island area, over 80% of the animals are aggregated into groups of 2–19 individuals. Such aggregations last until early autumn (Audesirk, 1979). Seasonal determinants of reproduction may be largely exogenous, including such factors as salinity and food availability, and perhaps increasing day-length and water temperature associated with changing season, although there is a suggestion that endogenous cueing may also be present (Audesirk, 1976). In laboratory experiments to test the effects of light and temperature on reproductive maturation in Aplysia, non-reproductive winter animals were induced to lay eggs when temperature and light levels were elevated to simulate spring-summer levels (Smith & Carefoot, 1967; Usuki, 1970b; Audesirk, 1976; Ferguson, Parsons, ter Maat & Pinsker, 1984; Pinsker & Parsons, 1985). Audesirk (1976) further showed that of these influences, only temperature was important in A. californica, changes in day-length neither operating alone to influence gonadal development nor enhancing in any significant way the effect of temperature. Nevertheless, it seems clear that changing conditions in springtime provide the synchronizing cues for initiation of gonadal development (Smith & Carefoot, 1967; Audesirk, 1976). Minimum temperatures for spawning vary with different species; for example, 9–9·5°C for A. punctata in the Irish Sea (Carefoot, 1967c; Smith & Carefoot, 1967), and 9°C for A. juliana, 12°C for A. parvula, and 12·5–14°C for A. kurodai in the Japan Sea (Usuki, 1970b). Availability of food, especially the green alga Ulva lactuca, is thought by Gev et al. (1984) to be an important determinant of the life cycle of Aplysia depilans and A. fasciata in the eastern Mediterranean. Population numbers of both species were high during seasons of maximum food abundance and declined as the Ulva population declined. Pheromones have been implicated in several aspects of the reproductive behaviour of Aplysia (Aspey & Blankenship, 1975, 1976b; Audesirk, 1977; Audesirk & Audesirk, 1977; Lederhendler, Herriges & Tobach, 1977). They may be involved, first, in the formation and maintenance of breeding aggregations in the field; secondly, in inducing non-breeding animals to undertake mating and egg-laying activities; thirdly, in enhancing sexual attractiveness; and, finally, in synchronizing egg-laying amongst different individuals. Such chemical cues appear to be released from the egg-laying animals (Audesirk, 1977; Aspey & Blankenship, 1976b; Jahan-Parwar, 1976), as well as from the eggs (Aspey & Blankenship, 1976b; JahanParwar, 1976). For example, Aspey & Blankenship (1975, 1976b) found that buried A. brasiliana=willcoxi were highly sensitive to egg-laying conspecifics. The introduction of egg-laying animals or their eggs into aquarium tanks containing buried conspecifics caused a termination of burrowing and often elicited sexual behaviour. An interesting result found by Lederhendler et al. (1977) in their studies on intraspecific attractiveness in A. dactylomela was that while non-copulating A. dactylomela were generally attractive to solitary conspecifics, thus indieating the presence of water-borne stimuli (pheromones), a copulating pair was no more attractive than was a single animal or a group of six animals. There may, therefore, be several different pheromones which mediate reproductive activities in Aplysia. The function of such pheromones clearly is to attract initially and then to stimulate and maintain animals in a breeding state. Their chemical identities or tissues of origin have not yet been identified. There is, however, some evidence that the rhinophore may be the sensory organ mediating such interactions. In neurophysiological studies on the rhinophore of A. californica, Chase (1979a) confirmed the general sensitivity of the rhinophore to waterborne stimuli originating from conspecifics, but noted that the response was not species specific.

APLYSIA

155

COPULATION AND COPULATORY RÔLES Copulation in Aplysia involves one individual acting as a sperm donor (the functional male) and another acting as a sperm recipient (the functional female), and is not usually reciprocal. The copulatory interaction involves five phases of behaviour: (1) approach, (2) contact, (3) pre-copulatory contact, (4) intromission, and (5) separation (Lederhendler & Tobach, unpubl.). Usually several animals participate to form copulatory chains. Seven to ten animals are not uncommon and often more than this participate. I have seen A. punctata in the field in chains of over 30 animals and once in a complete ring of 22 individuals. A ‘ring’ of two animals (reciprocal copulation) is rare in most species. It has been reported in A. juliana about 2% of the time in laboratory studies (Switzer-Dunlap, Meyers-Schulte & Gardner, 1984), but is apparently common in A. brasiliana=willcoxi when animals are kept in pairs in the laboratory (Blankenship et al., 1983). When in a copulatory chain the lead animal may continue to move over the substratum, given that the others move in accompanying procession, and the lead animal may continue to feed or may even spawn (Eales, 1921; Bandel, 1976; Audesirk, 1977; Blankenship et al., 1983; Susswein et al., 1984a; Lederhendler & Tobach, unpubl.). The rôle adopted by a given individual in the copulatory sequence and the factors affecting it has excited much interest amongst behavioural scientists. In theory, as a simultaneous hermaphrodite, a given individual of Aplysia can act as either a sperm donor or sperm recipient at any given time. In fact, its rôle can be influenced by size, age, and copulatory experience, but not apparently by past egg-laying history. In laboratory studies on copulatory behaviour in A. brasiliana=willcoxi, Blankenship et al. (1983) noted that animals in pairs copulated as sperm donors 52% of the time and as sperm recipients 48% of the time. In addition, paired animals spent about 45% of their copulating time engaged in reciprocal transfer of sperm. Past copulatory experience has been shown by Susswein et al. (1984a) to be important in influencing subsequent choice of copulatory rôle in A. fasciata. “Aroused” animals, waving their heads and actively crawling to others of the same species, almost always mated as males. The sperm-recipient then often became “aroused” in turn, would begin to wave its head and become motile, and subsequently would mate as a male. Field studies on A. dactylomela in Barbados indicate that animals spend about equal amounts of time as donors and recipients (51 and 49%, respectively), but with no simultaneous sperm donation and reception between pairs of animals as observed for A. brasiliana=willcoxi (Carefoot, unpubl.). In these field studies on A. dactylomela the animals were not marked in any way. Consequently, it was not possible to distinguish whether a given individual was alternating equally between being a sperm donor and sperm recipient on successive nights of observation, or whether some animals acted more often as donors and others more often as recipients, with the two activities simply balancing each other. In fact, Lederhendler & Tobach (1977) showed in laboratory studies that while most A. dactylomela have no preferences as to copulatory rôle, a few individuals do act mostly as sperm donors and a few others, mostly as sperm recipients. Thus, of 26 individuals which copulated five or more times over a three-week period in their study, 17 showed no preference as to copulatory rôle, four were sperm recipients more often than donors, and five were sperm donors more often than recipients. The authors further discovered that the nine animals which showed a preference copulated more frequently and with a wider range of partners as compared with the group medians for all 26 animals. In addition, the four “sperm-recipient” individuals tended to copulate with more partners than did the five “sperm-donor” individuals. Lederhendler & Tobach (1977) offered an ingenious explanation for this behaviour. They suggested that selection would favour, however slight the balance, a copulatory pattern which allowed the egg-producing individual to copulate widely but with least expenditure of energy. In this way, energy would be expended by a few sperm donors in searching for recipients, and energy husbanded in the sperm recipients for allocation to egg production. Lederhendler & Tobach (1977) make two assumptions in their argument: that searching is done more by the sperm donor

156

THOMAS H.CAREFOOT

than by the recipient, and that eggs are more likely to be laid by sperm recipients than by donors. Field observations of A. fasciata and A. dactylomela lend credence to the former since the sperm recipient is usually the one approached and mounted by the sperm donor (Susswein et al., 1984a; Carefoot, unpubl.). In addition, Lederhendler & Tobach (unpubl.) have shown that past copulatory experience in A. dactylomela tends to make these same individuals more likely to initiate contact and to act as sperm donors in subsequent matings. The second assumption, that eggs are more likely to be laid by sperm recipients than by donors, is certainly true during the act of copulation because sperm donors are not able to copulate and lay eggs at the same time. Also, Blankenship et al. (1983) noted that on 67 out of 101 occasions when A. brasiliana=willcoxi were observed laying eggs they were also copulating as sperm-recipients. When not copulating, the individuals inclined to be sperm donors would deposit fewer fertile eggs simply because, on average, they would have less allosperm (from another individual) available to fertilize them. Finally, an habitual sperm donor would have less time to eat than an habitual sperm recipient but, at the same time, would need less energy to manufacture reproductive products than would a sperm recipient. A spermrecipient is able to copulate and eat at the same time. A further prediction from the foregoing argument would be that younger, smaller animals would act more often as sperm donors than as recipients. Otsuka, Rouger & Tobach (1980) found this to be true for A. punctata, where smaller animals tended to be sperm donors. Unfortunately, the authors had no way of determining age and, as a consequence, no way of assuring that smaller animals were actually younger animals. Switzer-Dunlap et al. (1984), on the other hand, could show no effect of size on the initial copulatory rôle chosen in young A. juliana. They point out that young individuals of any species would have to act as sperm recipients for at least some time prior to their first egg-layings, in order that their own eggs be fertilized. Clearly, further work is required on this subject. The use of laboratory-cultured animals, as in the study by Switzer-Dunlap et al. (1984), holds most promise in this regard, as both age and size can be carefully monitored and controlled. Age has been shown by Switzer-Dunlap et al. (1984) to be important in A. juliana in influencing copulatory choice. Older animals exhibit no preference as to initial copulatory rôle as a group, but over half of the individuals demonstrate a consistent and more or less equal choice of one rôle over the other. In comparison, as noted, young animals show no preference either as a group or individually. Finally, production and laying of an egg mass from one copulation to the next does not appear to affect the copulatory rôle adopted in a subsequent copulation (Switzer-Dunlap et al., 1984). The conclusion to be drawn from these studies on copulatory rôles in Aplysia is that many factors operate to influence copulatory choice. Of these, age and size have been investigated, but work is still need on such possible influences as time of year, nutritional status, and possible effects of water movements and depth. Audesirk (1979) notes that A. californica lose weight during periods of intense breeding activity through loss of foraging time due to time spent in reproductive activities, to energy costs of producing and depositing the eggs, as well as to actual physical loss of eggs at laying. Both Kupfermann & Carew (1974) and Audesirk (1979) have reported that A. californica eat little or not at all when in breeding aggregations. A number of time-budgets have been determined for Aplysia in both field and laboratory situations. These show the duration of time allocated to locomotion, feeding, and reproduction, and other behavioural activities. The reproductive components of these budgets are shown in Table II, expressed as percentage of time allocated by various sea hares to copulating, egg-laying, and to other reproductive activities. Under the last category is included the behaviour termed as “foreplay” by Kupfermann & Carew (1974), and “courtship” by Susswein et al. (1983) and Susswein (1984). Because of their descriptive nature, such terms as foreplay and courtship may exaggerate the significance of various interactions between sea hares prior to copulation, hence, should probably be termed “pre-mating” behaviour. Since A. dactylomela, however,

APLYSIA

157

shows interactive behaviour which does not lead to copulation, something apparently not exhibited by either A. californica or A. fasciata (i.e. all intraspecific interactions between members of these species seem to lead to copulation: Kupfermann & Carew, 1974; Susswein et al., 1983; Susswein, 1984), all non-mating interactions between sea hares of a given species are termed “non-copulatory interactions” for presentation in Table II. The data represent observations on animals in their peak periods of breeding and should, therefore, reflect high intensities of reproductive activity. Several points are evident from the data given in Table II. First, the greatest amount of time spent in reproductive activities was shown by A. dactylomela (66·5% of the total observation period: Carefoot, unpubl); the least, by A. TABLE II Time spent on reproductive activities by Aplysia: F=field; L=laboratory % time spent Species

Locatio n

Field or Time of lab. year

Numbe Time of r of day animals observe d

Total hours observe d

In non- Copulat Eggcopulat ing laying ory interact ions

Total % Referen ces

brasilia na =willco xi califor nica

Texas

L

June

20

0615– 1430

8·9

0

65

2·6

65b

S. Califor nia

F

July– Aug.

2

0645– 1940

18·2

12·0

8·5

0

20·5

dactylo mela

Barbad os

F

May– June

64

1830– 2030

23·6

7·3

59·1

0·1

66·5

May– June June– July

76

0300– 0500 0500– 1300

22·2

2·1

0

0

2·1

fasciat aa

fasciat a

Israel

F

20– 402

L

June– July

20– 402

L

June– Aug.

6–10

Blanke nship et al., 1983 Kupfer mann & Carew, 1974 Carefo ot, unpubl.

Susswe in et al., 1983

1050– 1530 1830– 0500 all day

3·3

21·6

0·4

25·3

3·3

25·3

2·0

30·6

3·3

25·4

2·1

30·8

Israel with food

all day

484·3

Susswe in, 1984

158

THOMAS H.CAREFOOT

% time spent Species

fasciat a

Locatio n

Field or Time of lab. year

Numbe Time of r of day animals observe d

Total hours observe d

In non- Copulat Eggcopulat ing laying ory interact ions

Total % Referen ces

no food Israel

L F

6–10 192

435·2

4·2 35

58·3 35

July July 1980– Aug. 1983

all day 1000– 1600

47·9

6·2 0

Susswe in et al., 1984a

(includes co and non-co interact copulatory copulatory interactions) aA

tally system of expressing data was employed in this study, where numbers of animals engaged in a specific activity were recorded at given times of the day and expressed as percentages; for purposes of comparison these percentages have been equated to percentage time spent doing each activity. bThe total is not 67·6% because egg-laying occurred simultaneously with copulation.

californica (20·5%, with the major portion of this representing “foreplay”: Kupfermann & Carew, 1974). The value of 2·1% shown by A. dactylomela during 0300–0500 h reflects an unusual pattern of behaviour in this nocturnally active species whereby the animals appear to copulate after emergence from day-time hiding places, feed just prior to dawn, then hide away in crevices again until dusk. What cannot be entered into such a determination, however, is the possibility that while hidden, the animals may be copulating and laying eggs. The energetic implications of the behaviour of A. dactylomela will be considered more fully in a later section (pp. 241–242). The second point to emerge from these data is the remarkably short duration of time devoted to egg-laying as part of overall reproductive activity (mean of 1·5% of total time for all species and conditions). Large amounts of spawn are produced by sea hares (up to 72% of total energy of production in A. juliana and 32% in A. dactylomela: Carefoot, 1970; see Table IV, p. 196), and even though much of the activity involved in preparing the spawn, such as encapsulating the eggs and packaging the capsules into an egg string, is hidden from view, one expects a larger proportion of time to be occupied in this activity. Finally, Susswein (1984) was able to show two clear effects of food deprivation on timebudgeting of reproductive activity in A. fasciata: (1) in the absence of food, egg-laying was as likely to occur throughout the day as at night; when food was present, egg-laying was strictly a nocturnal activity; and (2) when food was absent, a total of 58·3% of total time was spent on reproductive activities as compared with 30·8% when food was present. This seems a characteristic feature of opisthobranchs; when food is absent, they turn to reproduction—a behaviour, as noted by Susswein (1984), of considerable selective advantage. There appears to be no explanation as to why copulation should take up such a large portion of the timebudget of sea hares. In comparison to the few minutes or even hours noted for copulation in other opisthobranchs (see Hadfield & Switzer-Dunlap, 1984), the many hours or even days (Eales, 1921) spent in copulation by sea hares seems a remarkable commitment indeed. The most obvious explanation is that sperm transfer is slow, but there is no direct evidence for this. The only comparative information on other opisthobranchs appears to be that of Schmekel (1971), Rutowski (1983), and Longley & Longley (1984)

APLYSIA

159

who indicate that sperm transfer in nudibranchs may last for a few seconds to minutes only. In those species which copulate for long periods, for example, up to several hours, Schmekel believes that actual transfer of sperm occurs just before the partners separate. The sperm in nudibranchs is apparently passed as a glued ball of prostrate secretion, not as a continuous flow, and is never in the form of true spermatophores (Schmekel, 1971). When the long duration of the act is combined with its frequent occurrence in sea hares, the day-to-day commitment in their time budgets becomes extraordinarily great. The direct energy costs of copulation in sea hares are actually low compared with other activities (Carefoot, unpubl.), but indirect costs, specifically in the denial in most instances of the opportunity to feed, may ultimately be high. Sea hares eat huge amounts of foods often poor in nutritional value. Viewed in this way, the allocation of so much time to copulation and related activities would seem to be an unfavourable trade-off resulting in a loss of potential food energy and nutritional benefits. Switzer-Dunlap (pers. comm.) questions whether the long time spent in copulation by Aplysia may be a way for the sperm donor to keep the recipient from copulating with another and thus raising the likelihood that its sperm will fertilize the next egg mass of the recipient. Little is known of sperm competition in Aplysia, nor the time required after reception of sperm for it to become capable of fertilizing eggs. EGG-LAYING Sea hares are capable of laying eggs at a relatively young age. The actual time of reproductive maturity presumably depends on factors such as temperature, light, nutritional history, and so on, although there appear to be no studies which have investigated these effects specifically. In laboratory culture, A. juliana matures 5–6 weeks after metamorphosis (at 25–27°C: Switzer-Dunlap & Hadfield, 1979; Switzer-Dunlap et al., 1984), A. dactylomela 9–10 weeks after metamorphosis (at 25–27°C: Switzer-Dunlap & Hadfield, 1979), and A. californica 17 weeks after metamorphosis (at 22°C: Kriegstein, 1977b). Sea hares reproduce several or many times during their life-time. Death usually comes at the end of a long breeding season, but may occur early in the season depending on the age of the animal (e.g. A. kurodai: Usuki, 1970b). Just prior to or at the time of fertilization, albumen is coated on the eggs, after which they are encapsulated and embedded in layers of mucopolysaccharide jelly and formed into a long cylindrical string which issues from the common genital groove near the head (Hadfield & Switzer-Dunlap, 1984). By rhythmic movements of the head, the egg string is attached to the substratum. Three different movements are described by Cobbs & Pinsker (1982) for A. brasiliana=willcoxi. Up-and-down “undulations” may prepare the substratum, side-to-side “weaves” distribute the egg cordon, and in-and-out “tamps” attach the cordon to the substratum. These head movements are probably the same as the “nodding”, “weaving”, and “tucking” described for A. californica by Arch & Smock (1977). The egg string is festooned as a tangled mass onto seaweeds or onto other solid objects on the sea bottom. Willan (1979) observed in one study site in New Zealand that A. parvula preferentially deposited its eggs on the red alga Plocamium costatum; only rarely were eggs laid on other algal species. Eggs of Aplysia punctata have been recorded from Plocamium cartilagineum, Delesseria sanguinea, and Cryptopleura ramosa in the sublittoral regions in the Irish Sea (Carefoot, 1967c); from rocks and from Desmarestia aculeata and other seaweeds in Port Erin Bay (Miller, 1960), and also from Dictyota sp. and Desmarestia sp. in Port Erin Bay (Bruce, Colman & Jones, 1963). Rey & Stoner (1984) observed that while most of the egg masses of Aplysia brasiliana=willcoxi in the Indian River Lagoon area of Florida were attached to small rocks, a few were attached to the prop roots of mangrove trees, and some appeared to be simply resting on the sandy bottom. Similarly, Sawaya & Leahy (1971) noted that the eggs of A. brasiliana=willcoxi and A. dactylomela in areas near São Paulo, Brazil were attached to the roots and pneumatophores of mangroves, as well as to rocks and pieces of wood.

160

THOMAS H.CAREFOOT

Switzer-Dunlap (pers. comm.) has found that the eggs of A. juliana and A. dactylomela in Hawaii are usually attached to the undersides of rocks. The egg jelly must itself be sticky to allow the egg string to be fastened securely to almost any type of substratum. This stickiness has not been described. It must also be a transient phenomenon because the exposed surfaces of the egg strings are not sticky to the touch, the egg mass remains clean even in silty or sandy habitats, and the egg-laying animal does not stick to its own eggs. The rounded egg capsules contain from one (A. oculifera: Usuki, 1970b; A. parvula: Lance, 1971; Usuki, 1970b; Ostergaard, 1950) to 237 eggs (A. californica: MacGinitie, 1934), with great variation occurring among individuals, and even within an egg mass (e.g. 15–135 in A. kurodai: Usuki, 1970b; see Table III). Egg diameter and number of eggs per capsule has been predicted to show an inverse relationship amongst species of Aplysia (Bridges, 1975), with many small eggs or fewer large eggs being contained in a given capsule. This was investigated using the data shown in Table III for 10 species of Aplysia by plotting the mean values for each parameter. A scattering of points resulted, with no evident relationship. Possibly the variability was too great within a species for such a relationship to hold true for a broad comparison of means among species. It could well be true for a single species, however, and would be worth investigating. As well as noting that the number of capsules per linear centimetre of string was more or less constant in A. californica (at 39 capsules per cm), MacGinitie (1934) remarked that the great variation in number of eggs contained in a capsule had no effect on its size (each value in the range 87–237 eggs per capsule given by the author represented an average; hence, the total variability was actually greater than this). Kandel & Capo (1979) have investigated in detail the effect of size of the egg-laying animal on the number of eggs per capsule in A. californica. They noted that capsule volume was constant regardless of egg number only in animals heavier than 400–600 g in live weight (the animal featured in MacGinitie’s study weighed in excess of 2600 g). Kandel & Capo (1979) also found that the number of eggs per capsule was directly related to size, and thus presumably to age, of an animal. Small individuals packaged fewer eggs per capsule (about 15 for an animal 150 g in live weight) than did larger animals (about 75 per capsule for a 1200 g individual). Unfortunately, the authors gave no indication of variability in numbers of eggs per capsule between different capsules in a single egg string; hence, there is no way to relate MacGinitie’s data on variability within a single individual (87–237 eggs per capsule) to the between-individual variability shown by Kandel & Capo (1979). (Interestingly, although MacGinitie himself did not indicate what weight change, if any, occurred in his animal over 17 weeks of egg-laying, by interpolating between the 2600 g starting weight that he recorded and the data given by Kandel & Capo, we can predict that the animal’s final weight would have been approximately 1900 g.) Toevs (1966) noted that while all eggs in a given capsule in A. californica develop at the same rate, development of eggs in different capsules is not synchronized. Finally, from the data given in Table III on size of a species, and on size of eggs produced by a species, it appears that larger species produce larger eggs. This has been noted previously for aplysiids by SwitzerDunlap & Hadfield (1977). Egg-laying and copulation tend to occur simultaneously or within a few hours or days of one another (Audesirk, 1977; Blankenship et al., 1983; Susswein et al., 1983). Because eggs are, however, commonly laid in the absence of copulation (Kupfermann, 1970; Blankenship et al., 1983; Hadfield

APLYSIA

161

TABLE III Information on life span, breeding season, eggs, and larvae of Aplysia Maximum adult size

Eggs

Species Live Length Life Breedin Numb weight (cm) span g season er in (g) (month capsul s) e brasili ana =willc oxi

608

27

califor nica

6800

48

cervin a

10

12

Aug.Dec. (S. Hemis phere: Marcus & Marcus , 1955) Mar.Apr. (Kraka uer, 1969) Jan.June (Rey & Stoner, 1984)

8–16 24

Apr.Sept., peak in July/ Sept. (Audes irk, 1977)

87– 237 150

10–15

Hatching time Diam. (µm)

77 80

Length of larval life

Colour

Numb er of days

Temp. Days °C

Temp. General °C referenc es

Bluegreen (Kraka uer, 1969); dark green, red, yellow (Strent h& Blanke nship, 1978a) ; lavend er or purple (Pinske r& Parson s, 1985) Yellow (MacG initie, 1934); yellow throug h greenis h-tan to tan (Wrigh t, H.O., 1960); green or blue (Pinske r& Parson s, 1985) Whitis h-blue

8–9 7–10 11 8–10

20 23 28 21–25

33–36 30

21–25

Krakau er, 1969; Strenth & Blanke nship, 1976, 1978a; Paige, 1981

9–11

22

34

22

MacGi nitie, 1934, 1935; Wright , H. O., 1960; Audesi rk, 1976, 1977, 1979; Kriegst ein, 1977b; Kriegst ein et al., 1974

6

Bandel , 1976

162

THOMAS H.CAREFOOT

Maximum adult size

Eggs

Species Live Length Life Breedin Numb weight (cm) span g season er in (g) (month capsul s) e

Hatching time Diam. (µm)

Oct. (Kay, 1964) Nov.May (N. Zealan d: Willan, 1979)

4–7 2–8

90

40

Mar.Aug. (Lo Bianco , 1888, 1909) late summe r (Thom pson, 1976)

20–30 25

93– 103 50–96

40

All year, but especia lly summe r (Lo Bianco , 1888, 1909)

36–48 40–50 50

90– 114 50–96

dactyl omela

1400

41

depila ns

380

fasciat a

1710

10–11

Colour

to greenis hyellow (Bande l, 1976) Golden to yellowi shgreen (Switz erDunlap & Hadfiel d, 1977)

Yellow (Baras h& Zenzip er, 1980)

Length of larval life

Numb er of days

Temp. Days °C

Temp. General °C referenc es

8–9

24–26

24–26

14–16

25

14–16

25

30

Moritz, 1936; Kay, 1964; Switze rDunlap & Hadfiel d, 1977, 1979; Willan, 1979 Lo Bianco , 1888, 1909; Thomp son & Bebbin gton, 1969; Thomp son, 1976; Gev et al., 1984 Lo Bianco , 1888, 1909; Carazz i, 1900; Saunde rs & Poole, 1910;

APLYSIA

Maximum adult size

Eggs

Species Live Length Life Breedin Numb weight (cm) span g season er in (g) (month capsul s) e

Hatching time Diam. (µm)

Colour

Numb er of days

163

Length of larval life

Temp. Days °C

Temp. General °C referenc es Thomp son &

JuneAug. (Barash & Danin, 1971)

43

juliana

468

30

5–7 (Hawai i) 6–12 (Japan)

Feb.Mar., JulyAug. (Kay, 1964); Mar.June/ July (Usuki, 1970a, b)

10–55 25–50

75–80

kurodai

590

34

7–12

Mar.July (Baba et al., 1956); Apr.July/ Aug., Nov./ Dec. (Usuki, 1970a, b)

15–30 15– 135

73–85

White to pale yellow (Switze rDunlap & Hadfiel d, 1977); yellow (Usuki, 1970b) Yellow to pink (Baba et al., 1956; Usuki, 1970b)

7–8 10–11

24–26 18–20

8–9

20–23

28

24–26

Bebbin gton, 1969; Barash & Danin, 1971; Barash & Zenzipe r, 1980; Gev et al., 1984 Kay, 1964; Usuki, 1970a, b; Switzer -Dunlap & Hadfiel d, 1977, 1979; Sarver, 1978 Usuki, 1970a, b; Baba et al., 1956; Nishiwa ki et al., 1975

164

THOMAS H.CAREFOOT

oculifer a

parvula

41

pulmon ica

punctat a

13

12

19

12

Mar.June (Baba et al., 1956) Aug./ Sept. (Usuki, 1970a, b) Mar. (Osterg aard, 1950); Mar.May/ July (Baba et al., 1956); Dec. (Kay, 1964); Apr.July (Usuki, 1970a, b); JulyFeb. (N. Zealand : Willan 1979)

1 2–3

70–75

Greenis hyellow (Baba et al., 1956; Usuki, 1970b)

8–10

24–27

Usuki, 1970a, b; Baba et al., 1956

1 1–3 1–4 2–3

70–80

Yellow to pink (Baba et al., 1956; Usuki, 1970b); light green (Kay, 1964)

7–10

19–24

Osterga ard, 1950; Baba et al., 1956; Kay, 1964; Usuki, 1970a, b; Lance, 1971; Willan, 1979

15

130

20

12–125

Feb.July (Lo Bianco, 1909) springlate summer (Eales, 1921); Mar.Aug. (Miller, 1960); MayOct. (Carefo

3–4 4 7 6–8

93– 103

Yellow, pink, mauve (Carefo ot, 1967a); orange or pink (Eales, 1921)

20–22 15 (April, Naples) 6 12–15

15 15–16 20–25

24

24–26

52

20

Switzer Dunlap, 1978 Lo Bianco, 1909; Saunder s& Poole, 1910; Eales, 1921; Miller, 1960; Carefoo t, 1967c; Thomps on & Bebbin

APLYSIA

ot, 1967c)

vaccari a

a

15,900

76

Feb.Mar. (Winkle r, 1955)

165

gton, 1969; Thomps on, 1976; Vicente & Poizat, 1977; Otsuka et al., 1981 Winkler , 1955; Winkler & Dawson , 1963

Weight estimated from a SCUBA sighting (Limbaugh, in Winkler & Dawson, 1963).

& Switzer-Dunlap, 1984), the precise relationship between the two events is unclear. Considerable research has been done and is now underway on the possible interactive rôles of peptides from the atrial glands and egg-laying hormone (ELH) from the bag cell neurones of the abdominal ganglion, in triggering egg-laying behaviour in sea hares (Kupfermann, 1967, 1970, 1972; Strumwasser, Jacklet & Alvarez, 1969; Arch & Smock, 1977; Pinsker & Dudek, 1977; Arch, Smock, Gurvis & McCarthy, 1978; Blankenship, 1980; Dudek et al., 1980a; Cobbs & Pinsker, 1982; Scheller et al., 1982; Blankenship, Rock & Schlesinger, 1982; Blankenship et al., 1983; Rothman, Wier & Dudek, 1983; Ferguson et al., 1984; Nagle, Painter, Kelner & Blankenship, 1985; for a review see Hadfield & Switzer-Dunlap, 1984). Bag-cell secretions have also been found to cause vaso-constriction in the anterior and gastroesophageal arteries in A. californica, thus decreasing blood flow to relatively inactive (during egg-laying) tissues involved with digestion and locomotion, and thus possibly enhancing blood circulation to the tissues involved in production of eggs (ovotestis and oviduct: Ligman & Brownell, 1985). In laboratory studies on the effect of past egg layings on subsequent egg-laying behaviour in A. californica, Dudek et al. (1980b) found that 30% of all layings occurred within two days of a previous egg-laying episode, and a separation of more than 11 days between two spawning bouts was rare (representing only 11 % of 294 episodes of egg laying). As expected, the volume of eggs laid increased in direct proportion to the number of days separating any two bouts of egg laying (up to 10 days until the experiment was terminated). In Hawaii, A. juliana and A. dactylomela are reported to spawn every 1–2 or 2– 4 days, respectively (at 25–27°C: Switzer-Dunlap & Hadfield, 1979). Exogenous factors which might act as cues for spawning in sea hares are not well known. Possible factors mentioned by Hadfield & Switzer-Dunlap (1984) are changes in illumination, effects of tidal cycle or lunar cycle, and relative abundance of food. Because sufficient sperm are stored to fertilize numerous egg masses, spawning may continue for days or even weeks thereafter without recourse to further copulation. In this regard, MacGinitie (1934) noted that a fertile eggs in 15 separate layings over a period of 11·3 single A. californica laid approximately weeks before its supply of sperm was exhausted. All eggs laid after this time were infertile. Krakauer (1969) found that sperm in A. brasiliana=willcoxi remained viable for, or were completely exhausted after, only 11 days. As part of their study on copulatory and egg-laying behaviour in A. brasiliana=willcoxi,

166

THOMAS H.CAREFOOT

Blankenship et al. (1983) recorded a frequency of egg deposition of one egg mass every 29 hours over a 14day period. The frequency of egg-laying was also slightly greater in solitary animals than in paired ones. Aplysia in mating aggregations may be surrounded by masses of eggs (Kupfermann & Carew, 1974; Susswein et al., 1984a), the presence of one egg mass stimulating other animals to deposit eggs in overlapping layers (Willan, 1979; also possibly with some intermixture between species such as that observed in mingled populations of A. fasciata and A. depilans by Achituv & Susswein, 1985). It is not clear what advantage accrues from overlapping depositions of eggs by successive animals. Such egg-masses can be fist-sized or larger and consist of many millions of eggs. Gas exchange for respiration must be somewhat curtailed deep in the egg cluster, and egg-eating predators (e.g. starfish; MacGinitie, 1934; Willan, 1979) would possibly benefit from the presence of such a convenient pre-packed meal. Possible explanations for this egg-laying behaviour may relate to search patterns of predators, or it may simply bring about less potential food wastage by adult sea hares. The eggs are often attached to algae and do not appear to be preferentially deposited on non-food species (but see Audesirk, 1979). Because the eggs are never, or rarely ever (see Saway & Leahy, 1971), eaten by adult sea hares, the algae to which the egg masses are attached are not consumed as food; hence, less food is wasted in providing attachment sites for the eggs. The egg masses are coloured from yellow to various shades of red-purples and delicate mauves, presumably through incorporation of pigments from the algal foods (see Table III). The pigment is situated in the eggs or embryos (Bandel, 1976; Switzer-Dunlap & Hadfield, 1977), and not in the gelatinous matrix portion of the spawn. Nishibori (1960) identified the pigment in the eggs of A. kurodai as a free xanthophyllic carotenoid. The only direct evidence that links the colour of the egg mass with the food being eaten is based on the observations by Chapman & Fox (1969). These authors report that coloured derivatives of phycoerythrins are taken up from red algal diets and incorporated into the egg masses of A. californica. In A. punctata, a change in diet from red- to green-coloured seaweeds leads to a change in colour of the egg string within just a few millimetres of newly laid string (Carefoot, 1967a). So delicate are the shadings, and so strict is the fidelity of the colour to the food being eaten by the sea hares, that a given alga (out of a selection of six algal species: two green and four red) in the diet of A. punctata, if eaten alone over a few days or more, can be identified from the colour of the spawn produced shortly after being eaten (Carefoot, 1967a). Achituv & Susswein (1985) could readily differentiate egg masses of A. fasciata and A. depilans through their bright orange-yellow and white colourations, respectively. Because the adults were eating a similar diet of the green alga Ulva lactuca (Susswein et al., 1984a; Achituv & Susswein, 1985), this suggests an inherent difference in the ability of the two species to metabolize and incorporate the pigments from a single alga. Similar results were obtained for Aplysia brasiliana=willcoxi and A. californica by Pinsker & Parsons (1985). The two species were eating the same brand of commercially marketed dried laver (Porphyra sp.), yet produced eggs of different colours. As the embryos mature and their shells develop, the egg mass loses its delicate colouration and becomes more brownish in hue because of the amber colour of the shells until, finally, just prior to hatching, the mass is a uniform brown colour. No function has yet been shown for these colours. A camouflage function seems most likely, as the egg masses tend to match the colour of the seaweed being eaten, but a physiological rôle should not be overlooked (Switzer-Dunlap & Hadfield, 1977). Many small invertebrates find the closely packed egg strings and dark interstices of Aplysia egg masses ideal as places to live. In a comprehensive study of the egg-mass associates of A. brasiliana=willcoxi in Florida, Rey & Stoner (1984) recorded 31 species of invertebrates which inhabited the egg masses. The bulk of these (representing over 90% of the total number of individuals collected) were made up of two species of amphipods and one species of mollusc. The inhabitants of the egg masses appeared not to be a special fauna,

APLYSIA

167

but were just part of the surrounding faunal assemblage. They were using the egg masses as temporary resting and foraging sites, and possibly as refuges from predators (Rey & Stoner, 1984). Most of the species were detritivores, possibly subsisting on detritus materials caught up in the egg strings. The authors did not mention whether any of the 31 species actually ate the Aplysia eggs, nor whether hatching success of the veligers was changed by their presence. Eggs are laid by most species over fairly well-defined seasonal periods. Spring through early summer is the commonest period (Table III), but spawning can extend into early autumn for some species (e.g. A. californica, A. dactylomela, A. fasciata, and A. punctata). Winter spawning has been observed in A. brasiliana=willcoxi, A. californica, A. fasciata, A. juliana, A. kurodai, A. parvula, and A. vaccaria. In the Japan Sea area, A. kurodai may spawn twice in the year: once in early spring and again in late autumnwinter (Usuki, 1970a, b). The late-breeding animals are ones which probably settled early in the spawning season of the previous spring and grew to a size sufficient to spawn in the winter before the water became too cool. It is not surprising to find that the species with the most extended breeding seasons are tropical (e.g. A. dactylomela, A. juliana, and A. parvula). Sarver (1978, 1979) has recorded spawning and recruitment of A. juliana throughout the year in Hawaii. The total number of eggs in an egg mass may be enormous. Barash & Zenziper (1980) have estimated the In a number of eggs in a single mass of A. fasciata (representing 40 metres of egg string) to be eggs and, in a single egg mass of single egg mass of A. depilans (size unspecified) there may be A. punctata, 135,000 eggs (Thompson & Bebbington, 1969). Thorson (1946) calculated that a one-metre eggs. MacGinitie egg string of A. punctata contained 200,000 eggs and a single egg mass, up to eggs. (1935) calculated that a 6·1-metre long egg string laid by a single A. californica contained Krakauer (1969) estimated that an egg string of A. brasiliana=willcoxi one metre in length contained 120, eggs produced by an A. californica of 2600 g live weight over 000 eggs. The oft-quoted value of a period of 18·3 weeks in 27 separate layings, gives testimony to the prodigious reproductive potential of sea hares (MacGinitie, 1934). MacGinitie (1934) also observed an individual of A. californica depositing its eggs at a rate of 41,000 per minute. Switzer-Dunlap & Hadfield (1977) calculated that an average egg mass eggs, and one of A. juliana, eggs. Life-time production of of A. dactylomela contained eggs in laboratory cultures of these two species in Hawaii was estimated by Switzer-Dunlap & Hadfield for A. dactylomela (over a period of seven months for an individual reaching a maximum (1979) to be for A. juliana (over a period of 5–6 months for an individual mean live weight of 245 g) and reaching a maximum mean live weight of 210 g). For A. juliana this works out to an average of eggs produced per day over a five-month period of reproduction. Life-time production of eggs by A. juliana in Japan, as measured by Usuki (1970b), representing an average period of 124 days, was about 255 g, or eggs using the conversion factor of eggs·g spawn−1 provided by Switzer-Dunlap about & Hadfield (1979) for Hawaiian representatives of this species. The reasons for such differences in fecundity among different species (comparing similar sizes of individuals), or within a species in different areas, are unknown. REPRODUCTIVE EFFORT Egg production clearly represents a large allocation of energy. Various estimates of reproductive effort, or the proportion of energy allocated to reproduction (Pr), are given for several sea hares in Table IV. These where Ps is the energy allocated to somatic growth and estimates are presented as: allocated to spawn; as where the equation represents the percentage of total production or the percentage of absorbed food energy (A) converted to energy of reproductive

168

THOMAS H.CAREFOOT

products; and, for a few instances where respiratory energy costs (R) are known, as or the percentage of non-respired energy represented by energy of spawn. For the last, a value of 100% is the theoretical maximum when all available energy is being converted into equivalent energy of reproduction. In fact, this theoretical value is never reached because of competing costs such as replacement of tissues in maintenance, and enzyme and mucus production (Carefoot, in press). The effect of temperature on reproductive effort is not known for sea hares, nor for other gastropods (Carefoot, in press); hence, temperature differences are omitted in the comparisons to follow. Two major points are evident from the data on reproductive effort in Table IV. First, there is a marked species difference in reproductive effort expressed as the percentage of total production represented by Overall means for each species are 66% for A. juliana, 23% for A. energy of eggs dactylomela, and 33% for A. punctata on all diets. These differences are reduced somewhat when reproductive effort is expressed as a percentage of absorbed energy actually converted into spawn; means of 4.6, 8.2, and 11.6%, respectively, for the three species shown in Table IV). The second point regarding the data in Table IV is that diet appears to affect reproductive effort. This is more apparent where diets differ in quality, but not in quantity. Diet quality shows an effect in A. dactylomela and A. punctata, with highest values for reproductive effort, expressed as a percentage of total being exhibited by animals eating their favoured foods production (Enteromorpha, Ulva, and Plocamium). For Aplysia punctata, somewhat lower values were found for animals eating three red algae (Heterosiphonia, Cryptopleura, and Delesseria: 32, 32, and 37%, respectively) that neither promote good growth nor represent preferred foods. The significance of these results cannot be interpreted without knowing a great deal more about the nutritional requirements of sea hares and, more specifically, whether different nutrients are required in the production of somatic as opposed to reproductive tissues. The different values for reproductive effort on different diets suggest that this may be true, but alternatively, size may affect reproductive effort, with smaller animals exhibiting lower levels of reproductive effort than larger ones. Clearly, allocation of nutrients and energy into production of somatic and reproductive tissues varies over the life of the animal, from 0% being allocated to reproduction initially, to close to 100% later in life. Aspects of energy allocation are discussed more fully in a later section (see pp. 238–243). As mentioned previously, the effect of size on reproductive effort has not been specifically investigated in Aplysia. The lowest value for reproductive effort measured as in the dietary series shown in Table IV for A. punctata (18%; Carefoot, 1967a) TABLE IV Data on reproductive effort in Aplysia: A=absorption; C=Chlorophyceae; P=Phaeophyceae; Pr=reproduction; Ps= somatic growth; R=Rhodophyceae; R=respiration Reproductive effort Species

dactylomel a

C C

Diet

Mean live starting weight (g)

Days kept Temp. °C % of total production (Pr·[Ps+Pr] −1×100%)

% of absorbed food energy (Pr·A −1×100%)

Cladophor a sp. Enteromor pha sp.

48·9

14

28·5

16

4·0

51·9

14

28·5

32

6·4

% of nonrespired absorbed energy (Pr·[A−R] −1×100%)

References

Carefoot, 1970

APLYSIA

169

Reproductive effort Species

C R juliana

C C C C

punctata

C

C R

R

R

R P R

Diet

Mean live starting weight (g)

Days kept Temp. °C % of total production (Pr·[Ps+Pr] −1×100%)

% of absorbed food energy (Pr·A −1×100%)

Ulva fasciata Laurencia papillosa Enteromor pha sp. Ulva fasciata Ulva lactuca Ulva lactuca low rationa medium ration high ration Enteromor pha intestinalis Ulva lactuca Plocamiu m cartilagine um Heterosiph onia plumosa Cryptople ura ramosa Delesseria sanguinea Laminaria digitata Plocamiu m cartilagine um

68·2

15

28·5

23

4·9

50·1

14

28·5

21

3·0

25·1

9

28·5

72

4·7

32·8

15

28·5

60

6·5

0·5

104

25·0

66

14·7

40

137

25·0

41·0 41·5

8·4 6·4

2·0

80

15·0

47·5 53

8·2 16·2

2·0

80

15·0

41

9·3

2·0

80

15·0

40

14·5

2·0

60

15·0

32

6·7

2·0

80

15·0

32

5·0

2·0

60

15·0

37

22·4

2·0

35

15·0

18

8·0

0·9

80

15·0

13

5·2

% of nonrespired absorbed energy (Pr·[A−R] −1×100%)

References

Carefoot, 1970

17·2

Sarver, 1978

Carefoot, 1967a

19·4

6·5

Carefoot, 1967b

170

THOMAS H.CAREFOOT

Reproductive effort Species

a

Diet

Mean live starting weight (g)

Days kept Temp. °C % of total production (Pr·[Ps+Pr] −1×100%)

% of absorbed food energy (Pr·A −1×100%)

% of nonrespired absorbed energy (Pr·[A−R] −1×100%)

References

“Low”, “medium”, and “high” rations were equivalent to 5, 10, and 15 g fresh algae·animal−1·day−1 for 137-day periods.

is shown by animals eating the brown alga Laminaria, a diet which hardly sustains growth and which is normally rejected as a food. In comparison, diet quantity shows little effect on reproductive effort in Aplysia. Sarver (1978) maintained groups of five Aplysia juliana each on “low”, “medium”, and “high” rations of Ulva lactuca for 137 days and measured spawn production and growth. His data show that reproductive effort, expressed as varied only from 41–47·5%, and that reproductive a percentage of total production varied only from 6·4–8·4% effort expressed as a percentage of absorbed energy (Table IV). The animals grew to quite different sizes on the different rations (means of about 68, 98, and 140 g on the “low”, “medium”, and “high” rations, respectively), and these sizes were maintained at more or less constant levels over the final 100 days of the study. Essentially all production of a non-maintenance nature was at this time being allocated to manufacture of eggs. Because of this, reproductive effort expressed as a percentage of total production should have been almost 100% on all three diets. That it was in fact less than half of this value (Table IV), was presumably attributable to the first 37 days of the study when the animals were allocating much more of their energy into production of somatic tissue than into reproduction. for other herbivorous and carnivorous gastropods range from Published values for 20–100% (Hughes & Roberts, 1980; Perron, 1982), with no evident difference between major trophic categories. The 100% value was recorded by Hughes & Roberts (1980) for 8 to 12-year old Littorina spp., and simply meant that all energy production in these old, reproductively mature animals was going into reproductive products. The authors were able to show a marked effect of age on reproductive effort, with young animals (2 to 4 years old) allocating as little as 20% of their total energy production to reproduction. No similar data on age effects on reproductive effort are available for sea hares. range from 10–27% for several species of limpets and a winkle Published values for (Grahame, 1973; Parry, 1982), somewhat higher than the comparable values for Aplysia. Too few data are available for reproductive effort, expressed as a percentage of non-respired absorbed energy to make any generalizations. The difference in values for this expression for A. punctata eating Plocamium cartilagineum in the two studies listed in Table IV (6·5% and 19·4%; Carefoot, 1967a, b) may, however, relate to differences in age. The starting size of Aplysia punctata in the first experiments was less than half that in the second; hence, the latter reached reproductive size much earlier are and remained reproductively active longer than the former. Only a few values of available for other gastropods: these include 21% obtained by Burky (1971) for the freshwater limpet Ferrissia rivularis; and 92% reported by Ansell (1982) for the carnivorous moonsnail Polinices alderi.

APLYSIA

171

LARVAL LIFE The successful culturing of Aplysia californica by Kriegstein, Castellucci & Kandel (1974) and of A. dactylomela and A. juliana by Switzer-Dunlap & Hadfield (1977) represented exciting advances in the field of larval biology. Not only did these accomplishments provide specific information about features of larval life, substratum selection, metamorphosis, and early growth in Aplysia, but they added to our general understanding of metamorphic competency and delay of settling in species with long-lived planktotrophic larvae. As a further bonus, the application of these fairly simple and reliable techniques to mass culture of sea hares provides the potential for the supply of large numbers of specimens for neurophysiological and other studies. The only drawback has been that our knowledge of developmental rates and patterns of development of larval sea hares has come from animals in laboratory culture. Ironically, our knowledge of the natural biology of Aplysia larvae is almost non-existent. Veliger larvae hatch from the egg capsules after 1–3 weeks on average (see Table III, p. 190). There is a positive relationship between egg size and length of embryonic period (Switzer-Dunlap & Hadfield, 1977). Thus, data in Table III show that at 22–25°C species with smaller eggs (e.g. 72·5 and 75 µm for A. oculifera and A. parvula, respectively) hatch after about 8·5–9 days, whereas ones with larger eggs (e.g. 98 and 102 µm for A. depilans and A. fasciata, respectively) hatch after about 15 days. Most eggs are viable in a single mass and hatching success is usually well over 90%. There appears to be no information on the mechanism of hatching, nor on the factors, either exogenous or endogenous, which render the mucopolysaccharide matrix enclosing the egg capsules sufficiently soft to allow easy transit of the larvae from the capsule to the outside. At hatching the shell lengths of laboratory-cultured Aplysia species range from 125–144 µm (Kriegstein et al., 1974; Kriegstein, 1977b; Switzer-Dunlap & Hadfield, 1977). At metamorphosis the shell lengths are 310–400 µm representing a 260% increase in length (Kriegstein et al., 1974; Switzer-Dunlap & Hadfield, 1977). Interestingly, the size at metamorphosis of A. parvula veligers obtained from plankton hauls by Switzer-Dunlap (1978) was 500 µm substantially larger than the largest larvae of other species cultured in the laboratory (400 µm). Switzer-Dunlap & Hadfield (1977) note a two-phase pattern of growth in Aplysia larvae. The first phase occupies about two-thirds of the larval life span and is one of almost linear growth to maximum larval size (see also Kriegstein, 1977b). This is accompanied by other morphological changes such as enlargement of the velar lobes, development of the heart, and appearance of the eyes. The second phase occupies the remaining one-third of the larval life span and involves no further increase in shell length but, rather, an increase in tissue mass along with other morphological changes leading to competency (Kempf, 1981). Only after this second period is the larva ready for, and capable of, settling. After hatching the veliger larvae swim upwards. The cueing mechanism for this behaviour has not been investigated experimentally but, as the larvae possess functional statocysts at hatching, a possible stimulus for upward swimming is an innate negative geotaxis (Hadfield & Switzer-Dunlap, 1984). Two functions are served by this behaviour. First, it places the larvae in the upper lighted surface layers of the sea where phytoplanktonic food is abundant. Secondly, it increases the likelihood of the larvae being caught up in horizontally moving currents to be transported from their sites of origin. The length of free-living larval life in a marine invertebrate will vary depending on a number of factors such as temperature, food, and salinity. None of these factors appears to have been studied in sea hares. The 28 to 36-day duration of larval life given in Table III is derived from data on animals in laboratory culture and represents minimum times. Delay of metamorphosis through less than optimal conditions, including absence of suitable settlement sites, is to be expected but has not yet been reported for Aplysia in the field. In laboratory culture, Kriegstein et al. (1974) were able to maintain A. californica veligers for four weeks beyond their predicted time of metamorphosis. Of 90 animals, only 5–10 died during this time, and 85% of

172

THOMAS H.CAREFOOT

the remainder successfully settled and metamorphosed when presented with their normal settling substratum of Laurencia pacifica after four weeks. Similarly, Strenth & Blankenship (1976, 1978a) and Paige (1981) held larvae of Aplysia brasiliana=willcoxi for 40 days and 100 days, respectively, past their normal times of metamorphosis before eliciting successful settlement and metamorphosis on exposure of the veligers to various red algae. Finally, Kempf (1981) maintained larvae of A. juliana in competent state for up to 311 days in laboratory culture, well beyond their normal free-living term of 28 days. None of these long-term larvae of A. juliana underwent spontaneous metamorphosis, yet were equally as capable of normal settlement and metamorphosis as were much younger larvae. Kempf (1981) points out that transport over long distances, including an estimated 77-day transit in ocean currents from Japan to Hawaii, is thus well within the survival capabilities of larvae of A. juliana. This observation further supports the theory of larval transport in ocean currents to explain the widespread dispersion of various cosmopolitan Aplysia species. Settlement from the plankton ends the first of the three post-hatching phases described by Kriegstein et al. (1974), the planktonic phase. The veliger now enters the second, or metamorphic phase, which quickly leads to the third, or juvenile stage. Its behaviour at this time changes from almost constant swimming and filter-feeding to a benthonic phase of exploration. At this time the larva searches for a substratum on which to attach and metamorphose. Mortality is obviously great during larval life. The tiny veligers fall prey to innumerable filter-feeders and other predators; they die through lack of food, low salinity conditions, or through absence of suitable sites on which to settle; or they drift away in offshore currents and die later. The fate of Aplysia veligers in this regard has never been explored specifically, but much has been written on the general topic of larval survival to enable parallels to be drawn (e.g. Thorson, 1946; Mileikovsky, 1974). Hadfield & SwitzerDunlap (1984) have extrapolated from data on density, fecundity, and recruitment presented by Sarver (1979) for A. juliana in Hawaii, and have calculated that only about 0·002% of the larvae produced in the area could survive pelagic life to settle. An assumption implicit in this reasoning is that the larvae produced by the population of A. juliana will remain in, or return to, the same nearshore area after several weeks of planktonic life. The point is, however, well made; larval mortality is enormous in sea hares. SETTLEMENT AND METAMORPHOSIS During many years of unsuccessful attempts at culturing Aplysia it was suspected, but not fully understood, that chemical inducers were required to stimulate settlement and to initiate metamorphosis in the larvae (Ostergaard, 1950; Bebbington & Thompson, 1969; Krakauer, 1969; Kupfermann & Carew, 1974; Carefoot: unpubl.). Not until this was recognized and suitable inducers in the form of seaweeds (usually representing the food of the juveniles and later adults) presented to competent sea hare veligers, were the larvae successfully cultured through metamorphosis. Such settlement inducers have been found to be the red algae Laurencia pacifica, Neoagardheilla baileyi, and Gracilaria sp. for Aplysia californica (Kriegstein et al., 1974; Kriegstein, 1977b; Capo, Perritt & Berg, 1979), various Ulva spp. for Aplysia juliana (SwitzerDunlap & Hadfield, 1977), several red algae including the genera Chondrococcus, Gelidium, Laurencia, Martensia, Polysiphonia, and Spyridia for Aplysia dactylomela (the best being Laurencia; Switzer-Dunlap & Hadfield, 1977), the red alga Chondrococcus hornemanni for Aplysia parvula (Switzer-Dunlap, 1978), the red alga Callithamnion byssoides and other species in the Order Ceramiales for Aplysia brasiliana=willcoxi (Strenth & Blankenship, 1976, 1978a; Paige, 1981), and the red alga Lomentaria articulata for Aplysia punctata (Otsuka, Oliver, Rouger & Tobach, 1981). The chemical identity of the inducer substances is not known.

APLYSIA

173

As in most other benthic marine invertebrate larvae for which details are known, an active choice is made by the veliger larva as to where to settle. When Kriegstein et al. (1974) gave a choice of six algae (Plocamium sp., Laurencia pacifica, Polysiphonia sp., Dasya sp., Chondrus sp., or Ulva sp.) to Aplysia californica veligers, the larvae crawled upon Laurencia preferentially and only rarely settled on the others. The veligers only metamorphosed on Laurencia pacifica. Veligers of Aplysia brasiliana=willcoxi which metamorphose readily on tufts of the red alga Callithamnion byssoides would also metamorphose on at least one species of Polysiphonia, but with much less success (Strenth & Blankenship, 1978a). Similarly, larvae of Aplysia juliana which, as adults almost exclusively eat Ulva spp., could be also metamorphosed on Enteromorpha sp., Caulerpa racemosa, and Halimeda opuntia, but in lower numbers and with poor, or no, postlarval growth (Switzer-Dunlap & Hadfield, 1977). No Aplysia juliana veligers metamorphosed in the presence of the red algae Spyridia filamentosa or Acanthophora spicifera, or the brown alga Dictyota crenulata (Switzer-Dunlap & Hadfield, 1977). Willan (1979) noted that recruitment of Aplysia dactylomela in New Zealand was principally on an intertidal species of Laurencia; no recruitment occurred subtidally even though Laurencia distichophylla was abundant, nor was this latter species chosen as a settlement site by “intertidal” animals. After attachment to an alga the larva begins to metamorphose. At this time several special larval features (such as the larval heart and velum) disappear and are replaced by adult structures (such as the radula, ctenidium, and definitive heart), and the behaviour of the larva changes to one of crawling (for details of metamorphosis see Kriegstein et al., 1974; Kriegstein, 1977a, b; Switzer-Dunlap, 1978; Switzer-Dunlap & Hadfield, 1977; Vicente & Poizat, 1977; Hadfield & Switzer-Dunlap, 1984). About 4–6 days after initial settlement the juvenile sea hare begins to feed on its algal food (Kriegstein et al., 1974; Strenth & Blankenship, 1976, 1978a). Recruitment times vary between different species of Aplysia. For A. punctata in the Irish Sea and for A. californica, major recruitment occurs in early autumn (Miller, 1960; Carefoot, 1967c; Audesirk, 1976, 1979). In the eastern Mediterranean, recruitment of A. depilans occurs during February-April and A. fasciata during most of the spring and summer (Gev, Achituv & Susswein, 1984). In Florida, Krakauer (1969) observed two waves of settling of A. brasiliana=willcoxi. In New Zealand, the bulk of the recruitment of A. dactylomela was in January–March (i.e. in summer), but some recruitment lasted through to July (Willan, 1979). For A. kurodai in the Japan Sea area, recruitment was in May and June (Usuki, 1970b). In comparison with these rather specific times in other parts of the world, Sarver (1979) recorded recruitment of A. juliana throughout the year in Hawaii, although with slightly lower rates in late winter and spring. The principal factors influencing recruitment in Sarver’s study were variations in larval abundance and in biomass of Ulva spp. Correlation with Ulva biomass was not so much seasonal as geographical. In summary, there appears to be no widespread correlation of recruitment with season and, therefore, no general correlation of recruitment with seasonal abundance of seaweed food as suggested by Gev et al. (1984) for Aplysia depilans and A. fasciata in the eastern Mediterranean. Because recruitment is often in the spring, and adults from the previous generation may live through the summer and early autumn, any sampling in between may show two distinct age classes (e.g. in A. punctata: Carefoot, 1967c). Usuki (1970b) suggests that a springtime sampling of A. kurodai may show two concurrent generations, one from the previous summer, the other from a late autumn-winter spawning. This implies that the winter eggs hatch into veligers which feed, become competent and metamorphose, all at winter temperatures of 9–14°C.

174

THOMAS H.CAREFOOT

SENESCENCE AND DEATH After a lifespan varying from 3–16 months the animals die. As noted previously this may be related to lack of food (Gev et al., 1984) or to season (Carefoot, 1967c; Usuki, 1970b; Audesirk, 1976, 1979), or perhaps to both. Spawning does not in itself usually lead to death, even though some authors associate the two events (e.g. Bandel, 1976). After a peak period of spawning which usually accompanies the attainment of maximum size, the animals, however, begin to lose weight, spawning declines, and death occurs soon thereafter. Senescing animals may exhibit signs of physical deterioration including erosion of parapodia, loss of body pigment (Audesirk, 1979), and bloating of the footsole (Grigg, 1949). In addition to physical degradation, ageing A. californica show discrete behavioural changes. Foremost amongst these is a diminished responsiveness in certain aspects of the defensive siphon-withdrawal reflex (specifically, in a slowness to habituate and failure to sensitize; Bailey, Castellucci, Koester & Chen, 1983). In addition, the ageing sea hare’s heart becomes less aroused in response to food stimuli (Bailey et al., 1983), and there are other neurological changes (Papka, Peretz, Tudor & Becker, 1981; Rattan & Peretz, 1981). Dead or dying animals may be found commonly in the field at a season’s end (e.g. A. punctata between September and December; Carefoot, 1967c), sometimes coincidental with mortality in laboratory populations (Audesirk, 1979). The different processes of ageing in laboratory and field populations of Aplysia are clearly shown by the study of Hirsch & Peretz (1984). These authors found that the median postmetamorphic age of death in a small laboratory population of A. californica was only about 5 months, and that growth rates were continuous up to the time of death (Peretz & Adkins, 1982; Hirsch & Peretz, 1984). From these observations, Hirsch & Peretz concluded that senescence and growth proceed simultaneously in Aplysia, but they failed to present any actual evidence of ageing in their animals. It is possible that the animals in this small laboratory population were dying from any number of laboratory-related causes: water quality, disease, parasites, and so on— factors unrelated to normal processes of senescence. Field animals of this species are known to live for many months in excess of the mean period of survival cited by Hirsch & Peretz. Audesirk (1976, 1979), for example, showed that a field population of A. californica at Santa Catalina Island had a postmetamorphic life span of about 9–11 months, that older animals lost weight before dying (as has been found for several other species: see pp. 237– 238), and that senescent animals showed signs of physical deterioration in the form of tissue erosion. PERIODICITY OF OCCURRENCE A characteristic feature of life cycles of opisthobranchs in general, and of Aplysia in particular, is their periodicity of occurrence. Often there are long gaps when no sea hares can be found in areas where they normally reside (Krakauer, 1969; Lederhendler, Bell & Tobach, 1975; Sarver, 1978; Audesirk, 1979; Achituv & Susswein, 1985). This has been noted most evidently for species which exhibit a life span shorter than one year. For example, along the Israeli coast of the Mediterranean, all traces of eggs and adults of A. fasciata disappear a minimum of four months prior to the new season, and for A. depilans there is at least a six months’ gap before the first appearance of new recruits (Gev et al., 1984). Based on results of laboratory culturing studies mentioned earlier, this is at least four times longer than the minimum time needed to produce competent veligers. The likeliest explanations for such gaps in seasonal occurrences of Aplysia are that juveniles are often hard to see, particularly amongst thick growth of algae, that recruitment may depend on larval transport from distant populations, themselves out of phase with the local populations, or that the larvae delay their metamorphosis until cued by other events, such as changing

APLYSIA

175

season, or until current vortices drop them at their original sites. A less likely explanation for such gaps is migration. The notion of migration as part of the life cycle of sea hares originated with Garstang (1890), but possibly stemmed from earlier ideas expressed by Hunt (1878) and Lo Bianco (1888). It was later repeated by Eales (1921, 1960), Yonge (1949), and Hardy (1959), prompted in part by the observations that bright rosecoloured juveniles of A. punctata in Britain (originally misnamed A. rosea) could be collected by dredging amongst red algae some distance offshore, whereas intertidal collections yielded mainly olive-green, brown, or chocolate-brown coloured adult specimens. In addition, Garstang (1890) maintained a single specimen of Aplysia sp. in a tank at the laboratory of the Marine Biological Association in Plymouth, fed it on an unspecified type of food, and observed over a period of 6·6 weeks that it changed from a bright pink-red to a deep red-brown colour. From these observations Garstang inferred that the life cycle of Aplysia involved settlement in deeper water offshore, followed by a migration shorewards through successive zones of red, brown, and finally olive-green coloured algae in the intertidal zone. He further intimated, and was later supported in this idea by Eales (1921), that Aplysia adopted protective camouflage by changing colour to match each seaweed substratum during this migration. This theory of migration has fallen into disfavour from lack of supporting evidence (Miller, 1960; Carefoot, 1967c; Krakauer, 1969; Audesirk, 1979), although a number of authors continue to acknowledge its possibility (Grigg, 1949; Kay, 1964; Marcus, 1972; Nishiwaki, Ueda & Makioka, 1975; Rey & Stoner, 1984). Not only does A. punctata, the animal around which the original theory evolved, not eat Fucus spp. (the presumed olive-green seaweed in Garstang’s study), but maintenance in later experiments of subtidally collected A. punctata on several red, green, and brown seaweeds for 11 weeks produced only a darkening of their normal red colour, with no change in hue (Carefoot, 1967c). It is possible that other factors, such as temperature, light, and pressure could be involved in colour change in Aplysia, but at present none of these has been investigated. A basic assumption in the migration theory is that reproduction necessitates, or is somehow facilitated by, an inshore migration. Yet, sublittoral populations of A. punctata in the Irish Sea were observed to undergo complete annual cycles in fairly confined geographical regions without the necessity for shoreward migrations (Miller, 1960; Carefoot, 1967c). Furthermore, it does not follow that survival of adults and eggs, hatching success, and early survival of larvae would necessarily be enhanced in an inshore (intertidal) habitat. Foods might be more abundant in an intertidal area for some species, but not for A. punctata (Carefoot, 1967c). No long-term studies of populations of sea hares have conclusively demonstrated that migrations from deeper offshore areas to shallow inshore areas are an integral part of their life cycles. Finally, no studies in which movements of tagged sea hares have been monitored have demonstrated uniform, seasonal, directional mass movements of a population which could be construed as a migration (e.g. Audesirk, 1979). There are, however, certain localized movements of sea hares. These have been associated with breeding aggregations in A. californica (Audesirk, 1976; but not considered as “migrations”: Audesirk, 1979), with possible shoreward displacement during storms or wave action in A. punctata (Eales, 1921; Carefoot, 1967c) and in A. brasiliana=willcoxi (Krakauer, 1969), with localized feeding excursions in A. dactylomela (Carefoot, 1985; see also pp. 172–181), and with mass strandings of swimming A. brasiliana=willcoxi (suggested by Hamilton, Russell & Ambrose, 1982, as possibly being associated with reproduction). Such strandings of A. brasiliana=willcoxi were probably not just displacements by wave action, as fewer animals were actually stranded during storms than during calmer weather (Hamilton et al., 1982). To account for sudden disappearances of A. brasiliana=willcoxi in shallow-water habitats in southern Florida, especially those coincidental with the seasonal onset of lethal or near-lethal water temperatures,

176

THOMAS H.CAREFOOT

Hamilton (1985, 1986) hypothesizes that a swimming migration may occur as a regular part of the life cycle. He speculates that juvenile sea hares migrate in summer to deeper-water habitats to avoid the high temperatures of shallow-water areas, a behaviour thought also to occur in this same species in areas of Brazil (Sawaya & Leahy, 1971). Several months later in early spring, the adults are suspected of swimming inshore, timing their migration to coincide with peak algal abundance. But for scale in time and distance, this hypothesized migration would parallel feeding excursions seen in other species of sea hares (e.g. A. dactylomela: Carefoot, 1985). The theory is also attractive in that it accounts for seasonal gaps in abundance without the necessity of invoking a reproductive migration. FOOD AND FEEDING Seaweeds comprise the major foods of sea hares. Green and red algae are generally favoured over brown algae, and the preferred species are often ones that give best growth and spawn production. Food species are perceived and identified through chemical signals received by the rhinophores and oral tentacles. Food is broken into bite-sized pieces by the radula and jaws and is stored in a capacious crop. A muscular gizzard helps to macerate the food and final digestion occurs in a large digestive gland. Sea hares eat prodigiously and spend several hours each day browsing for food. They may eat up to one-third of their body weight each day in algae. Because of this their effects on modifying abundances and distributions of benthic algae must be great, but have not yet been fully investigated. Little is known of the nutritional requirements of Aplysia. FIELD DIETS Field diets of several Aplysia are listed in Table V. The data were obtained through observations of animals eating in their natural habitats, by examining crop contents of freshly collected animals, and by analyses of faecal composition. It is apparent that feeding preferences in the field are broad for most species, with various combinations of green and red algae making up the bulk of the diets. The most striking example of this is Niell’s (1977) study on A. punctata, in which 44 species of seaweeds were identified in the crops of 33 individuals from a single collection in the Ria de Vigo area of Spain. Of the 44 food species, four were green, six were brown, and the remainder red. Aplysia eat brown algae only rarely, as for example Egregia and Macrocystis by Aplysia californica (Leighton, 1966; MacFarland, 1966), and this may relate more to the fact that brown algae may be the only foods available in a certain habitat, rather than to any special dietary preference. Kupfermann & Carew (1974) note that only large A. californica seemed to eat brown algae in the field, and then only infrequently; small animals were never observed to eat them. Scrutiny of the data in Table V shows a general and expected trend whereby the best studied species appear to have the broadest dietary range (i.e. A. californica, A. dactylomela, and A. punctata). Species which are poorly studied, such as A. cervina and A. vaccaria, have apparently narrow dietary ranges. It may be that further investigation will broaden these ranges; it is, however, commonly recognized that there are well-studied species, such as A. juliana, that have natural narrow selections of foods. A. juliana eats only green algae and, of these, Ulva spp. are the favoured choices and are the foods which provide best nutrition (Carefoot, 1970; Usuki, 1970b; Switzer-Dunlap & Hadfield, 1979; Vitalis, 1981). Zostera and other angiosperms, for example Diplanthera wrightii and Syringodium filiformis, are eaten by a few species (e.g. Aplysia brasiliana=willcoxi and A. californica) but it is unclear whether these nonalgal foods are a preferred dietary item. In some cases they may be ingested incidentally with the normal algal foods or may be only eaten at certain times or in certain areas where regular seaweed foods are scarce or absent.

APLYSIA

177

Finally, carnivory has been observed in some species (e.g. sponge spicules and small amphipods in the gut of A. dactylomela: Engel, 1928; Carefoot, pers. obs.; and hydroids, bryozoans, small gastropods, amphipods, and copepods in the gut of A. brasiliana=willcoxi: Krakauer, 1969). Such materials are probably ingested accidentally while the animal browses amongst its algal foods. Kupfermann & Carew (1974) reported instances of cannibalism and other carnivory in starving laboratory A. californica, but did not observe this behaviour in the field. Sawaya & Leahy (1971) reported that starving A. brasiliana=willcoxi would eat their own eggs, but this type of cannibalism is rarely observed in other species of sea hares. SPECIAL LABORATORY FOODS Laboratory animals can be fed on dried algal foods such as laver (“nori”: Porphyra spp.; Kupfermann & Pinsker, 1968; Preston & Lee, 1973; Kupfermann, 1974b; Susswein & Kupfermann, 1975b; Aspey & Blankenship, 1976a, b; Susswein, Kupfermann & Weiss, 1976b; Dieringer, Koester & Weiss, 1978; Strenth & Blankenship, 1978b; Advokat, 1980; von der Porten, Redmann, Rothman & Pinsker, 1980) and dulse (Rhodymenia palmata: Jahan-Parwar, 1972a). Susswein & Markovich (1983) and Schwarz & Susswein (1984) have used freshly thawed Ulva lactuca as food for Aplysia californica, A. depilans, A. fasciata, and A. oculifera. It is possible to use certain garden vegetables as laboratory foods for Aplysiai. These include celery tops and parsley leaves (Winkler, 1959b; Winkler & Tilton, 1961), and fresh romaine lettuce (Peretz & Adkins, 1982). It is not known whether these garden plants are nutritionally complete for sea hares. FOOD FOR LARVAE Larval diets are known only from laboratory cultures. There is no information on the natural diets of veligers in the field. The most common laboratory foods for Aplysia larvae are unicellular phytoflagellates, including Pavlova lutheri (formerly Monochrysis lutheri) and Isochrysis galbana (Krakauer, 1969; Kriegstein et al., 1974; Switzer-Dunlap & Hadfield, 1981; Strenth & Blankenship, 1976, 1978a; Kempf, 1981). Otsuka et al. (1981) initially provided Aplysia punctata veligers with a mixed diet of Monochrysis sp., Phaeodactylum sp., and Chaetoceros sp. The authors did not indicate whether the two diatom species were eaten, but as they modified the diet after one week to include only Monochrysis sp., this seems unlikely. Switzer-Dunlap & Hadfield (1977) tested the value of several algal species in promoting growth and survival of veligers of Aplysia juliana and A. dactylomela, including Pavlova lutheri, Isochrysis galbana, Dunaliella tertiolecta, Pyraminonas sp., TABLE V Foods of Aplysia Species

Foods

References

brasiliana =willcoxi

Champia parvula, Chondria leptacremon, Diplanthera wrightii (angiosperm), Gracilaria, Hypnea musciformis, Laurencia gemmifera, Syringodium filiformis (angiosperm) Ulva Enteromorpha and Zostera Ceramium eatonianum, Gigartina canaliculata, Plocamium cartilagineum Enteromorpha, Gracilaria, Ulva, occas. Zostera marina

Krakauer, 1969, 1971

californica

Sawaya & Leahy, 1971 MacGinitie, 1935 Winkler, 1959b Beeman, 1961

178

THOMAS H.CAREFOOT

Species

cervina depilans

fasciata

dactylomela

juliana

keraudreni kurodai parvula

Foods

References

Codium and Zostera Acrosorium uncinatum, Centroceras clavulatum, Ceramium eatonianum, Chondria californica, Corallina gracilis, C. pinnatifolia, C. vancouveriensis, Ectocarpus sp., Gelidium coulteri, Gelidium spp., Gigartina canaliculata, Hypnea valentiae, Jania tenella, Laurencia pacifica, Laurencia sp., Lithothrix aspergillum, Lophosiphonia sp., Phyllospadix sp., Plocamium cartilagineum, Pterocladia sp., Pterosiphonia sp., Sphacelaria sp., Tiffaniella snyderae, Ulva sp. Egregia, Gigartina, Macrocystis Zostera, red algae, or brown algae in pools Laurencia pacifica Codium sp., Colpomenia sp., Eisenia sp., Gigartina sp., Laminaria sp., Laurencia sp., Macrocystis sp., Plocamium sp., Ulva sp. Laurencia and Plocamium Red algae Ulva lactuca Ulva lactuca Zostera Ulva lactuca, some phaeophytes

Marcus, 1961 Winkler & Dawson, 1963

Ulva lactuca Ulva lactuca Enteromorpha sp. and Ulva lactuca Bryopsis adriatica, Enteromorpha compressa, E. intestinalis, E. linza, Jania rubens, Pterocladia capillacea, Sargassum vulgare, Ulva lactuca Corallina, Laurencia spp., Ulva spp. Centroceras clavulatum, Cladophora sp., Gracilaria sp., Laurencia papillosa Acanthophora spicifera, Laurencia sp., and Spyridia filamentosa as juveniles; Ulva spp. and red algae as adults Laurencia spp. Enteromorpha compressa, E. linza, Ulva pertusa Enteromorpha sp., Ulva fasciata Ulva fasciata, U. lactuca, U. reticulata Ulva fasciata, U. reticulata Enteromorpha sp., Ulva fasciata, U. reticulata Enteromorpha sp., Monostroma sp. Corallina, Laurencia spp., Ulva spp. Enteromorpha compressa, E. linza, Ulva pertusa Corallina, Laurencia spp., Ulva spp.

Leighton, 1966 MacFarland, 1966 Chapman & Fox, 1969 Kupfermann & Carew, 1974

Audesirk, 1977, 1979 Marcus & Marcus, 1958 Jordan, 1917 Ghiretti et al., 1959 Hughes & Tauc, 1962 Achituv & Susswein, 1985

Jordan, 1917 Ghiretti et al., 1959 Gev et al., 1984 Susswein et al., 1984a

Morton & Miller, 1968 Carefoot, 1970, 1985 Switzer-Dunlap & Hadfield, 1979 Willan, 1979 Usuki, 1970b Carefoot, 1970 Sarver, 1978, 1979 Switzer-Dunlap & Hadfield, 1979 Vitalis, 1981 Usuki, 1981a Morton & Miller, 1968 Usuki, 1970b Morton & Miller, 1968

APLYSIA

punctata

vaccaria

Laurencia spp., Plocamium costatum Delesseria sp., Fucus spp., Laminaria sp., Ulva lactuca Subtidal: Cryptopleura ramosa, Delesseria sanguinea, Heterosiphonia plumosa, Plocamium cartilagineum Intertidal: Enteromorpha sp., Ulva lactuca Enteromorpha sp., Gelidium pusillum (also as pulchellum), Gigartina acicularis, Halopteris scoparia, Laurencia hybrida, L. pinnatifida, Ulva gigantea+37 other red and green species Egregia

179

Willan, 1979 Eales, 1921 Carefoot, 1967a, c Carefoot, 1967c Niell, 1977

Winkler & Dawson, 1963

and Nannochloris sp. The best food for larval Aplysia Juliana was found to be Pavlova lutheri, while Aplysia dactylomela, A. parvula, and A. pulmonica did about equally well on either Pavlova lutheri or Isochrysis galbana (Switzer-Dunlap & Hadfield, 1981). PERCEPTION OF FOOD Aplysia uses its rhinophores for distance perception of foods, and the anterior tentacular area of the oral veil mainly for contact perception of foods (Jahan-Parvar, 1969; Jahan-Parwar, 1972a; Jahan-Parwar, Smith & von Baumgarten, 1969; Preston & Lee, 1973; Wells, Jahan-Parwar & Fredman, 1974; Audesirk, 1975a, b; Emery, 1976; see also Emery & Audesirk, 1978). The oral tentacle and mouth regions are, however, highly sensitive to waterborne “food odours”, as shown in tests of A. juliana by Frings & Frings (1965), who demonstrated responses using 10−7 dilutions of sea-water extracts of Ulva lactuca, and in tests of Aplysia californica by Jahan-Parwar (1972b), who used water extracts of seaweeds as well as dilutions of various amino acids. Audesirk (1975a, b) further showed in “Y-maze” tests with A. californica that the oral tentacles alone were sufficient to allow food to be located. The rôle of the osphradium in chemical perception in sea hares is unclear. In other gastropods it has been implicated in such functions as detection of sediments and long-distance chemical reception (Yonge, 1947; Kohn, 1961). In Aplysia, it has been given an oxygen-sensing function by Eales (1921) and an osmoticperceiving function by Stinnakre & Tauc (1969). Frings & Frings (1965) could demonstrate no particular sensitivity of the osphradium in A. juliana to droplets of extracts of Ulva administered through finely drawn pipettes, nor could Preston & Lee (1973) show any behavioural response in Aplysia californica to stimulation of the osphradium with a fine stream of extract of laver (Porphyra sp.). In contrast, neurophysiological studies on Aplysia californica by Jahan-Parwar et al. (1969) and Downey & Jahan-Parwar (1972) showed clear responses of neurosecretory cells in the abdominal ganglion to chemical and osmotic stimulation of the osphradium, and some neurological preparations responded to mechanical stimulation of the osphradium. It is therefore possible that the osphradium serves several functions. A number of behavioural and electrophysiological studies on the mouth-opening and biting responses of A. californica have confirmed the general responsiveness of this species to sea-water extracts of red algae (Rhodymenia palmata) applied dropwise to the mouth region (Jahan-Parwar, 1972a). Also, a fine sensitivity exists in sea hares to dilute sea-water solutions of amino acids, especially to L-aspartic and L-glutamic acids (Jahan-Parwar, 1972a). These amino acids elicit mouth-opening responses in Aplysia californica at concentrations of 10−7 to 10−5 M, several orders of magnitude less than the lowest effective concentrations of other amino acids tested. Jahan-Parwar has recorded from receptor neurones in the anterior tentacular groove of the oral veil of A. californica and demonstrated that this area is most sensitive to food substances (especially L-glutamic acid) compared with other stimuli tested.

180

THOMAS H.CAREFOOT

The generally sensitive chemosensory response to amino acids and other low molecular-weight materials in marine invertebrates (see reviews by Kohn, 1961; Laverack, 1968; Bardach, 1975), including reactions of sea hares to L-glutamic and L-aspartic acids, raises the question as to whether these or other amino acids may be used by sea hares in long-distance perception of food (Jahan-Parwar, 1972a; Jahan-Parwar, Wells & Fredman, 1974; Carefoot, 1982b). In this regard, Jahan-Parwar (1975) has shown in Y-maze experiments with A. californica that L-glutamic and L-aspartic acids elicit feeding behaviour in much lower concentrations (10−9 to 10−8 M) than concentrations of other L-amino acids which elicit similar behaviour (e.g. asparagine, methionine, phenylalanine, serine, and tyrosine, which are effective only in concentrations of 10−3 M and greater). L-glutamic and L-aspartic acids were found to be in higher free-state concentrations in comparison with all other free amino acids in several common orders of seaweeds, including the Ulvales, Cladophorales, Rhodymeniales, Ceramiales, and Gigartinales, which contain many of the favoured foods of sea hares (e.g. Ulva spp., Enteromorpha spp., Laurencia spp., and Plocamium spp.; Carefoot, 1982a). The cause and effect relationship of these two observations, if any, is not known, but it may provide a starting point for study on the possible rôle of extracellular substances in seaweeds in governing long-distance perception and recognition of foods in sea hares. Food stimuli generally elicit a complex behavioural sequence, including locomotion and head-waving, which serve to orientate the animal to its food and, following contact with food, biting and swallowing, which lead to ingestion (Frings & Frings, 1965; Preston & Lee, 1973; Kupfermann, 1974a, b; Susswein & Kupfermann, 1974; Susswein et al., 1976b; Susswein, Weiss & Kupfermann, 1978; Weiss et al., 1981; Susswein & Markovich, 1983; see also Susswein et al., 1984b). The duration of food-arousal activity in Aplysia californica has been shown to depend on several factors, including level of satiation and strength of the food stimulus presented (Susswein, Kupfermann & Weiss, 1976a; Susswein et al., 1978). Animals which are partially satiated or which are presented with weak food stimuli are slower to arouse than are hungry animals or animals presented with strong stimuli (Susswein et al., 1976a, 1978). Both tactile and chemical stimuli appear to be necessary to evoke repeated and regular biting responses in A. californica (Rosen, Weiss & Kupfermann, 1982), and repeated presentation of food stimuli decreases the time required for them to reach an aroused state (Kupfermann, 1974b). Food stimuli are known to increase heart rate, blood pressure, and blood flow in Aplysia (Dieringer et al., 1978; Weiss et al., 1981; Koch & Koester, 1982) and changes also occur in the pattern of blood flow to various organs during the feeding process. It has been found that blood flow is maximal to the digestive organs during the retraction phase of biting (pulling of the alga into the buccal cavity) and maximal to the head during the protraction or grasping phase (Koch & Koester, 1982; see also Koch, Koester & Weiss, 1984). This time-sharing of heart output is thought by Koch & Koester (1982) to ensure that the various organs work at maximum efficiency, especially in providing adequate perfusion of blood to the muscles that generate the biting movements. In fact, a reduction in blood flow to the head by mechanical means significantly impairs the biting action of the buccal apparatus (Koch & Koester, 1982). PREFERRED FOODS Experiments on feeding preferences in Aplysia have shown that a wide range of foods are chosen. The favoured algae are likely to be ones which provide best growth and spawn production. This is shown in Table VI and is most clearly illustrated for A. punctata (Carefoot, 1967a). When small A. punctata (less than 7 g live weight) from a sublittoral habitat were presented with a choice of algae that they would normally encounter, they ate them in an order of preference that directly corresponded to the value each

APLYSIA

181

provided for growth (Plocamium catilagineum Heterosiphonia plumosa > Cryptopleura ramosa > Delesseria sanguinea; rs=1·00, P=0·05, Spearman Rank Correlation Coefficient Analysis). When other algal species were included in the comparison for Aplysia punctata, even though they would normally not have been encountered by sublittoral animals, the green alga Enteromorpha intestinalis was preferred over all others, the green alga Ulva lactuca was preferred after Plocamium, and the brown alga Laminaria digitata was consistently refused when any of the other seaweeds were available to be eaten (Table VI). Not only was there good correlation of growth and spawn production in the sea hares with food preference but, for the sublittoral red algae, this order reflected the order of distribution of animals in their sublittoral habitat (Carefoot, 1967c). Thus, of 641 animals collected sublittorally for which the type of seaweed on which each animal was found was recorded, 96% were found on Plocamium, about 2% on each of Heterosiphonia and Cryptopleura, and less than 0·5% on Delesseria. In comparison with these small sublittoral animals, large intertidal Aplysia punctata clearly preferred Enteromorpha and Ulva over other seaweeds. Laminaria, contrary to other reports (e.g. Eales, 1921), was completely refused even though this alga would presumably be encountered regularly in the intertidal habitat (Carefoot, 1967c). Kupfermann & Carew (1974) used SCUBA diving and snorkelling to observe Aplysia californica in its natural habitat to determine which seaweeds were preferred. They found: (1) that certain seaweeds, as for example Laurencia sp. and Gigartina sp., were distinctly favoured, whereas others were eaten only infrequently (see Table V), (2) that the size of an animal determined whether certain brown seaweeds would be eaten (smaller animals tended to avoid these; see also Winkler & Dawson, 1963), and (3) that some common red seaweeds such as Pterocladia sp. were consistently avoided, whereas Laurencia, even when rare, was consistently eaten. In a comprehensive survey of algal foods of Aplysia californica, Winkler & Dawson (1963) noted that different populations had their own distinct preferences, which varied depending on the algae present in a given area. Red algae were observed to be eaten most by A. californica including, surprisingly, several species of corallines (Table V). The preferred foods of several populations were the red algae Hypnea valentiae, Plocamium cartilagineum, Laurencia pacifica, and Ceramium eatonianum. Winkler & Dawson listed a number of red and brown seaweeds which were common in one area (Lunada Bay, near Palos Verdes), but never eaten by Aplysia californica. Other than these few data on A. punctata and A. californica, there have been no field studies which have comprehensively related food choice in sea hares with availability of seaweeds in a species’ habitat. The general preference by sea hares for green algae has been documented in laboratory studies. Winkler & Dawson (1963) observed that A. californica, while feeding almost exclusively on red algae in the field, exhibited a strong preference for Ulva and Enteromorpha in the laboratory. Saito & Nakamura (1961) found that both Aplysia Juliana and A. kurodai preferred these green seaweeds in laboratory situations to what the authors considered were their normal foods of brown and red algae, respectively. (The contention by Saito & Nakamura that A. juliana normally eats the brown alga Undaria pinnatifida is difficult to accept. All other studies on A. juliana have shown that green algae, principally Ulva spp., but including Enteromorpha, and some Cladophora, are the only foods eaten: Frings & Frings, 1965; Carefoot, 1970; Usuki, 1970b; Sarver, 1978, 1979; Switzer-Dunlap & Hadfield, 1979; Vitalis, 1981.) Laboratory-held Aplysia dactylomela in Barbados were observed to eat preferentially Enteromorpha, Ulva, and Cladophora when given a choice of several seaweeds (Carefoot, 1970). Sawaya & Leahy (1971) have noted that field Aplysia brasiliana=willcoxi and A. dactylomela in the São Paulo area of Brazil principally eat Ulva spp. Finally, both Aplysia depilans and A. fasciata appear to favour a diet of Ulva (Jordan, 1917; Susswein, et al., 1984a; Achituv & Susswein, 1985). It is not known what characteristics make sea hares prefer one seaweed over another. Factors such as energy content and nutritional value have doubtless been important in the evolution of optimal diet in

182

THOMAS H.CAREFOOT

Aplysia, but probably no more so than texture, how easily it can be manipulated, its palatability and digestibility, and its availability to the animal. Perhaps it is not surprising that despite a wealth of information on feeding preferences in sea hares and other opisthobranch gastropods (Stehouwer, 1952; Braams & Geelen, 1953; Cook, 1962; Carefoot, 1967a, 1970; Edmunds, Potts, Swinfen & Waters, 1974; and others), there is not one instance where the factor or factors governing the selection of a food can be precisely identified (Carefoot, in press). As an example, the series of dietary preferences in Table VI for A. punctata, which shows such good correlation with growth and spawn production, does not show significant correlation with absorption of total dry matter on each diet, nor with absorption of total nitrogen, specific amino acids, or total carbohydrates (Carefoot, 1967a). Moreover, an experiment using a mechanical device to simulate the triturating action of the A. punctata gizzard showed that the two algal species most susceptible to being broken down by such treatment were the red alga Delesseria sanguinea and the brown alga Laminaria digitata. These were the two seaweeds Aplysia punctata preferred least and represented two of the three species giving poorest growth and spawn production on ad libitum diets (Carefoot, 1967a). Similarly, scrutiny of Niell’s (1977) comprehensive data on foods eaten by A. punctata in the Ria de Vigo area of Spain showed no apparent influence of texture on food choice. The 11 algae (of 44 species in total) found most frequently in the guts of 33 individuals represented seven different morphological categories of seaweeds (characterized by membranous blades, soft filaments, soft branches, rigid branches, firm rounded cylinders, flattened cylinders, and flattened segments with dichotomous branches), and had greatly differing textures. Of these 11 species, one was green, two were brown, and eight were red. It is clear that sea hares’ feeding preferences vary with habitat. This is most TABLE VI Feeding preferences of Aplysia: data were obtained from differences in amounts of seaweeds consumed when animals were presented with a choice of two or more algal species; the exceptions to this were the studies on A. californica by Brady & Young (unpubl.), which represented the difference in distribution of animals and different food substrates when presented with a choice of six different algae, and by Chapman & Fox (1969), which represented qualitative estimates of amounts of each alga consumed; A=angiosperm; C=Chlorophyceae; P=Phaeophyceae; R=Rhodophyceae Species

californic a

P P R P P

P

Food

Rank of Rank of feeding value for preference growth s

Egregia menziesii Macrocyst is pyrifera Gigartina armata Eisenia arborea Pterygoph ora californica Laminaria farlowii

1

3 4 5

6

Rank of value for spawn productio n

Rank of distributio n of animals on each food in the field

Rank of larval settling preference (lab. or field)

Rank of algal abundanc e in the field

References

Leighton, 1966 2

APLYSIA

Species

P

californic a

R

R

R

californic a (animals on a red algae-free diet)

P

P P P P P P P

C californic a

R

R

Food

Rank of Rank of feeding value for preference growth s

Cystoseira osmundac ea Laurencia pacifica

7

Plocamiu m cartilagine um Pterocladi a pyramidal e Egregia menziesii

2

Eisenia arborea Laminaria sp. Petalonia debilis Macrocyst is pyrifera Pelvetia fastigiata Zonaria farlowii Dictyopter is zonarioide s Codium fragile Plocamiu m cartilagine um Laurenica pacifica

3·5

Rank of value for spawn productio n

Rank of distributio n of animals on each food in the field

Rank of larval settling preference (lab. or field)

1

Rank of algal abundanc e in the field

References

Chapman & Fox, 1969

3

1

Chapman & Fox, 1969

3·5 3·5 3·5 7·5 7·5 7·5

7·5 1·5

1·5

1

183

Brady & Young, unpubl.

184

THOMAS H.CAREFOOT

Species

C

dactylomela

dactylomela

juliana

juliana

kurodai

Food

Rank of Rank of feeding value for preference growth s

Enteromor pha intestinalis

5·5

C R R P A C C C R R R C C R R R P P C C P C

Codium fragile Gelidium purpurascens Pterocladia capillacea Macrocystis pyrifera Phyllospadix torreyi Enteromorpha sp. Ulva fasciata Cladophora sp. Laurencia papillosa Galaxaura oblongata Laurencia spp. Enteromorpha sp. Ulva lactuca Plocamium costatum Champia laingii Herposiphonia sp. Undaria pinnatifida Ecklonia cava Enteromorpha sp. Ulva pertusa Endarachne binghamiae Ulva fasciata

5·5 5·5 5·5 5·5 5·5 1 2 3 4 5 1 2 3 4 5 6 1 2 4 4 4 1

C C C C C C R R

Ulva reticulata Enteromorpha sp. Ulva fasciata Enteromorphia sp. Enteromorpha sp. Ulva pertusa Grateloupia filicina G. okamurai

2 3 1 2 1·5 1·5 4 4

Rank of value for spawn productio n

1 2 4 3 5 1 2

1 2 4 3 5

Rank of distributio n of animals on each food in the field

Rank of larval settling preference (lab. or field)

Rank of algal abundanc e in the field

References

Carefoot, 1970, 1985 2 1 1

Willan, 1979

3 Saito & Nakamura, 1961

1

1·5

2 3 1 2

1·5 3 2 1

Vitalis, 1981; Switzer-Dunlap & Hadfield, 1977 Carefoot, 1970

1 Saito & Nakamura, 1961

APLYSIA

punctata

R C R C R R R P

G. tsurutsuru Enteromorpha intestinalis Plocamium cartilagineum Ulva lactuca Heterosiphonia plumosa Cryptopleura ramosa Delesseria sanguinea Laminaria digitata

4 1 2 3 4 5 6 7

2 1 3 4 5 6 7

1 2 3 4 6 5 7

185

Carefoot, 1967a, c 1 2·5 2·5 4

1

1 2 3 4

evident from examination of the data given for A. californica in Table VI, representing studies by Brady & Young (unpubl.) and Leighton (1966). Feeding preferences appear to differ markedly in populations from the different areas. Because one study tested the preferences of sublittoral animals for algae prevalent in kelp beds (near La Jolla, California: Leighton, 1966), while the other tested the preferences of shallowwater animals for algae associated with this type of habitat (Catalina Island, California: Brady & Young, unpubl.), perhaps the differences in preference-ranking were, however, predictable. In addition to habitat, feeding preferences could be predicted to vary with season and with changing nutritional needs associated with age, sex, and reproductive state. None of these latter has been investigated in sea hares, nor is it known whether sea hares would show a predictably higher degree of selectivity in their choice of foods when satiated or when food is common, and be less discriminating when starved or when food is scarce (Emlen, 1966). Differences in food selectivity in relation to size have been mentioned by Winkler & Dawson (1963) and Kupfermann & Carew (1974), but as yet no one has investigated this topic. The broad range of feeding modes in sea hares, from almost strict monophagy (e.g. A. juliana) to polyphagy (e.g. A. californica, A. dactylomela, and A. punctata), would provide a challenging basis for a further study of feeding preferences in this group. Food choice in sea hares is also affected by past dietary history. “Ingestive conditioning”, or the enhanced response of an animal to its food based on previous exposure to the same food, was observed in A. punctata in animals maintained for 80 days on single seaweed species (Carefoot, 1967a). The effect of this conditioning was short-lived, and within a few days of exposure to other seaweed foods the conditioned animals reverted to their normal level of preference. In studies on the effect of previous dietary history on food choice by A. californica, Brady & Young (unpubl.) also demonstrated ingestive conditioning. They found that animals previously fed on Plocamium cartilagineum for one month strongly preferred this alga over Laurencia pacifica when tested against both species in a Y-maze. Conversely, animals fed for as short a period as four days on Laurencia showed a measurable (but non-significant) preference for this alga over Plocamium. Internal food stimuli were shown by Susswein, Weiss & Kupfermann (1984b) to enhance feeding behaviour in Aplysia californica. By feeding experimental animals on small amounts of dried seaweeds (species unspecified) and control animals on filter paper, or by just stimulating the control animals by touching seaweed to their mouths then testing biting responses of both experimental and control groups to seaweeds, the authors were able to demonstrate a significant decrease in latency to bite as a response to eating in A. californica. In other words, after having eaten, the animals were quicker to respond on subsequent encounter with food. The effects lasted for up to 80 minutes following the stimuli for all groups. Other types of non-associative learning by Aplysia in connection with food and feeding are: (1) a decrease in feeding response after repeated stimulation with a non-food object (forceps and glass rods: Lickey, 1968; Lickey & Berry, 1966) and, (2) an inhibition of the feeding response following an electrical

186

THOMAS H.CAREFOOT

shock (Kupfermann & Pinsker, 1968). There are, therefore, a number of examples of non-associative learning or habituation responses in connection with food and feeding in Aplysia. In comparison, only a few examples of associative learning have been described for sea hares in this context. Jahan-Parvar (1970) described an instance of classical conditioning in A. californica, using seaweed as the unconditioned stimulus and light as the conditioned stimulus to train animals to show typical food-seeking behaviour, but this could not be successfully repeated by Kupfermann (1974b). Susswein & Schwarz (1983) and Susswein & Markovich (1983) trained A. fasciata and A. californica not to eat Ulva lactuca nor leaves of the lily Hemerocallis fulva which were wrapped in plastic netting. The animals could taste the food through the holes in the net and would initially bite and attempt to swallow the preparation. An essential component of the training exercise was, however, the tendency of the net-bound food to become stuck in the animal’s buccal cavity causing the animal to gag, which may have acted as a kind of punishment. After the animals had learned not to respond to the net-enclosed food, they could still be induced to eat nonnetted food (Schwarz & Susswein, 1982; Susswein & Markovich, 1983; see also Schwarz & Susswein, 1984). Some memory was retained by Aplysia fasciata 24 hours following training. Recently, Cook & Carew (1986) demonstrated a similar type of associative learning in Aplysia, that of operant conditioning. In their experiments the authors showed that A. californica could be operantly trained to change their explorative head-waving, which is normally side-to-side as when they are sensing food, to one side more than another in order to terminate the shining on them of an aversive strong light. The existence of a negative ingestive conditioning, in this case a post-ingestive “learning aversion” to foods, has been investigated in A. dactylomela using artificial diets with large imbalances of amino acids (Carefoot & Switzer-Dunlap, unpubl.). Preliminary results have suggested that individuals previously fed on such a nutritionally poor diet can recognize it on a subsequent encounter, remember their past experience, and avoid it. Such post-ingestive learned responses, both preference and aversion, may have played important rôles in modifying feeding behaviour in sea hares, not only in day-to-day selection of foods, but also in the longterm evolution of optimal diets. FEEDING AND MOVEMENT OF FOOD THROUGH THE GUT The mechanism of action of the radula and jaws is well described in Eales (1921) and Howells (1942), and will not be repeated here. Bite-sized pieces of algae are stored in a voluminous crop which, when full in A. dactylomela, can hold up to 10% of the live body weight (Carefoot, 1985). In A. californica, the total weight of anterior gut contents (oesophagus, crop, and gizzard) may represent 20% of the total body weight in satiated animals (Susswein & Kupfermann, 1975b). Comparable values for weight of anterior gut contents in satiated A. fasciata, A. depilans, and A. oculifera are 7·9, 11·9, and 10·8%, respectively (Susswein & Markovich, 1983). From the crop, the food moves to the two-part gizzard, each part containing a number of chitinized teeth (see Winkler, 1960; Beeman, 1969; Arnould & Jeuniaux, 1977 for information on the chemical characteristics of the teeth). Here, the food is macerated and the slurry, barely recognizable as seaweeds, is passed into the stomach (Howells, 1942). Eales (1921) believes the gizzard teeth to be of little use for grinding, but rather thinks their function is to compress and strain the food in preparation for action by digestive enzymes. During the process of digestion the crop and gizzard show rhythmical contractions which can continue for some time in in vivo preparations (Bottazzi, 1897). The possible rôles of neurotransmitter substances such as acetylcholine and serotonin in controlling these rhythmical movements have been investigated in A. dactylomela by Wells & Hill (1980, 1985). The effect of FMRFamide (Phe-Met-Arg-Phe-NH2) on

APLYSIA

187

contraction of the gizzard has been studied in A. californica (Austin, Weiss & Lukowiak, 1983), as has that of atropine in connection with oesophageal contractions in the same animal (Winkler & Tilton, 1961). Chemical digestion occurs in all regions of the fore- and mid-gut regions, and is regulated by enzymatic secretions produced in the salivary and digestive glands. The food moves from the stomach to the caecum, where digestion is completed, then travels via the intestine to the rectum. The faeces emerge from the anus into the mantle cavity area as loose, irregular pellets. From time to time the animal may forcibly expel water from the mantle cavity, which helps to carry these pellets to the outside. The large volume of food processed each day, combined with its high roughage content, results in a large production of faecal matter, to the extent that animals in confined laboratory conditions sometimes bury themselves totally in their own faeces. Sand may comprise a large portion of the gut contents in some species (e.g. A. californica: Winkler, 1961; A. dactylomela: Carefoot, 1970). In populations of A. dactylomela in Barbados sand can represent up to 28% of the dry weight of the crop contents in freshly collected and recently fed animals, and may account for a large part of the voided faecal residues (Carefoot, 1985). In fact, a minimum estimate of annual turnover of sand by A. dactylomela (of 60 g mean live weight) per linear km of suitable coastal habitat in Barbados is one metric ton (Carefoot, 1985). Its rôle in digestion, if any, is not known. It could act in digestion as a bulk carrier, by aiding in movement of food through the gut, or it could act as an adjunct to the gizzard, by aiding in the mechanical breakdown of algal tissues. Alternatively, its occurrence could be incidental, a simple consequence of feeding in shallow sand-swept areas on tightly meshed, sand-infested algae such as Cladophora sp. In Barbados, a shallow-inshore population of Aplysia dactylomela was observed to feed preferentially in such a Cladophora-dominated habitat. Interestingly, this area, almost in the wave-break part of the shore and characterized by extensive movements of sand, was chosen by the bulk of the population to feed in over a less sand-swept offshore area dominated by red algae. The animals used the latter area for their copulations, as well as to hide in during the day. Based on energetic and nutritional properties of the different algal foods in the two areas, and on the fact that the inshore Cladophora habitat offered few daytime retreats, therefore requiring that most animals undertake a return daily feeding excursion of several metres to and from the feeding area, the animals would have appeared to be better off in the offshore red-algal habitat. A comparison of amount of sand eaten by Aplysia dactylomela in the two areas, however, disclosed little difference (Carefoot, 1985). Since the offshore red algae (mainly Gracilaria sp. and Laurencia papillosa) contained visibly less entrapped sand than did Cladophora, this showed that sand might be important in digestion as the animals may have been eating it directly. FEEDING RATES Feeding rates of sea hares are thought to be influenced by a number of intrinsic and extrinsic factors, only a few of which have been studied. Chief amongst these are likely to be temperature, salinity, body size, type of food, time of day, state of tide, season, density of animals, reproductive status, and past feeding history. Data on feeding rates of sea hares are generally derived from laboratory animals, and the question arises as to whether these rates would differ from rates measured directly in the field. A test of this using bagenclosed Aplysia dactylomela in the field in Barbados showed no significant differences between laboratory and field rates when Cladophora sp. and Ulva fasciata were the foods, but did show a significant difference S.E. equivalent Joules eaten-live g−1·day−1, as when Laurencia papillosa was being eaten (field rate: S.E. equivalent J eaten·live g−1·day−1; Carefoot, 1985). compared with laboratory rate: Ingestion rates for Aplysia are given in Table VII and are shown graphically for a number of species in Figure 4. In this Figure the data are expressed as dry mg food eaten-dry g animal−1·day−1 as a function of live weight for two arbitrary temperature ranges, 15–17°C and 18–28°C. The data are varied but the two

188

THOMAS H.CAREFOOT

sets best fit the relationship where a and b are constants. This is consistent with observations on feeding in other gastropods where consumption (C) is related to body weight (W) by the (Edwards & Huebner, 1977; Bayne & Scullard, 1978). equation, The effect of temperature on feeding rates in A. juliana has been investigated by Saito & Nakamura (1961), although the results are difficult to interpret owing to variable past feeding histories and to variable weights of animals used in the experiment. In general, however, the rates of consumption increased with increasing temperature as expected. At 7–9°C animals on a diet of the brown alga Undaria pinnatifida ate about 1 % of their live body weight in fresh weight of algae per day, increasing to a value of about 11 % at 22–24°C. At temperatures higher than 24°C, feeding rates declined by an order of magnitude, and the animals died three days into the four-day experiment. Little is known of the factors that stimulate sea hares to feed. The only natural phagostimulant known for Aplysia is crude water-extract of seaweed. This was first demonstrated by Frings & Frings (1965) in A. juliana using water-extracts of Ulva lactuca, and confirmed later for Aplysia dactylomela using extracts of Ulva fasciata (Carefoot, 1979, 1980). Susswein et al. (1976b) induced Aplysia californica to bite by applying water extracts of dried laver (Porphyra sp.) to the animals’ mouths. A graded response was obtained depending on the concentration of extract and degree of satiation of the animals, although the chemical identities of these water-extracted phagostimulants are unknown. Frings & Frings (1965) determined that the extract from Ulva lactuca was heat-stable, insoluble in ether, and acted only on receptor sites near or in the mouth and on the oral tentacles, but not on the rhinophores. In later studies, Sakata et al. (1985; 1986) found that ether-extracted substances from U. pertusa, identified as glycerolipids, were highly attractive to Aplysia juliana, and the authors proposed a phagostimulatory rôle for these substances. Phagostimulatory properties of various chemical TABLE VII Feeding rates of Aplysia: C=Chlorophyceae; P=Phaeophyceae; R=Rhodophyceae Species

dactylomela

C C C R

juliana

C

P C

Food

Body weight Temp. °C No. of days Amount of (mean live g) food eaten (fresh g·animal−1 ·day−1)

% of mean body wt eaten·day−1

References

Cladophora sp. Enteromorp ha sp. Ulva fasciata Laurencia spp. Ulva pertusa

61·8

28

14

11·7

18·9

Carefoot, 1970

77·7

28

14

14·5

18·7

82·6

28

15

4·4

5·4

25·5

17

37

3·0

11·9

Willan, 1979

210·0

18–21

6

21·0

10·0

Saito & Nakamura, 1961

336·0

21

4

25·4

7·6

46·4

28

15

3·4

7·3

Undaria pinnatifida Ulva fasciata

Carefoot, 1970

APLYSIA

Species

C

Food

Body weight Temp. °C No. of days Amount of (mean live g) food eaten (fresh g·animal−1 ·day−1)

% of mean body wt eaten·day−1

29·2

28

9

10·2

35·0

110·0 239·0

25 16–20

104 5

23·1 14·8

21·0 6·2

kurodai

C C

Enteromorp ha sp. Ulva lactuca Ulva pertusa

punctata

C

Ulva lactuca

8·0

15

15

0·3

3·7

C

Enteromorp ha intestinalis Plocamium cartilagineu m Plocamium cartilagineu m Plocamium cartilagineu m Plocamium cartilagineu m Plocamium cartilagineu m

13·0

15

15

1·1

8·5

14·4

15

15

0·7

4·9

3·1

15

20

0·4

12·9

7·1

15

20

0·8

11·3

12·5

15

20

0·9

7·2

17·3

15

20

0·9

5·2

R

punctata

R

R

R

R

189

References

Sarver, 1978 Saito & Nakamura, 1961 Carefoot, 1967a

Carefoot, 1967b

substances were also investigated in A. dactylomela and A. kurodai (Carefoot, 1982b). Greatest phagostimulatory responses were elicited by starch, L-glutamic and L-aspartic acids, maltose, oleic acid, and combinations of certain vitamins. Interestingly, combinations of these phagostimulatory materials did not enhance feeding activity in A. kurodai and A. dactylomela; rather, their effects were negatively synergistic, resulting in a “masking” of the phagostimulatory properties of the components (Carefoot, 1982b). An important ramification of this discovery with regard to nutritional studies on sea hares using artificial diets of chemicals set in agar, is that “super diets” of irresistible palatability cannot apparently be created by combining several individually phagostimulatory materials. Satiation can be induced in sea hares in both the laboratory (Kupfermann, 1974b; Susswein & Kupfermann, 1974, 1975a, b; Susswein et al., 1976a, b; Kuslansky et al., 1978; Susswein & Markovich, 1983) and the field (Kupfermann & Carew, 1974). It is not known for certain that sea hares eat to satiation under natural conditions in the field. Because Kupfermann & Carew (1974), however, observed behaviour in field A. californica that was consistent with behaviour of satiated animals in the laboratory, field animals probably do eat to satiation. The quantity of food needed to produce satiation in A. californica appears to be determined by two inputs: (1) an external cue related to palatability (Susswein & Kupfermann, 1974) and,

190

THOMAS H.CAREFOOT

Fig. 4.—Rates of food consumption by sea hares: the data are arbitrarily divided into two groups representing the temperature regimes 15–17°C and 18–28ºC; equation of regressions, for 15for 18–28°C, the number by each point identifies the reference; 1, Saito & Nakamura (1961); 2, Carefoot (1967a); 3, Carefoot (1967b); 4, Carefoot (1970); 5, Sarver (1978); 6, Willan (1979).

(2) an internal cue related to the bulk properties of the food (Susswein & Kupfermann, 1975a, b). To this extent, the situation in Aplysia parallels that in vertebrates; what has not yet been identified in sea hares is the third major input noted by Susswein & Kupfermann (1975a, b) to be present in vertebrates: that of metabolic cues involved in the regulation of long-term energy and nutritional needs. An interesting example of satiation or near-satiation occurs with A. dactylomela in Barbados, where animals fill their crops in pre-dawn bouts of intense feeding, then hide for 12 hours or more during daytime to digest the food (Carefoot, 1985). Studies on A. californica in the field suggest that the animals normally partition their feeding into discrete meals (Kupfermann & Carew, 1974). The meals are separated by periods of time when portions of food are transported out of the crop (Susswein et al., 1978). Because satiation in A. californica can be induced by a non-nutritive bulk material such as filter paper, or silicone- and polyacrylamide-based gels, it is likely that some kind of mechanoreceptor in the crop is responsible for terminating feeding and not a feed-back signal based on chemical characteristics of the food (Susswein & Kupfermann, 1975a, b; Kuslansky et al., 1978). This feed-back signal from presumed stretch receptors in the crop occurs in A. californica only after a delay of some 15–20 minutes following satiation by the artificial introduction of a seaweed mash (laver: Porphyra sp.) or non-nutritive polyacrylamide gel (Kuslansky et al., 1978). As noted by Kuslansky et al. (1978), a delay of this duration opens the possibility that the signal from the gut receptors indicating satiation is hormonally (neurosecretory?) rather than neuronally mediated. LARVAL FEEDING RATES The veliger larva of Aplysia feeds on unicellular phytoplankton which it captures through the filtering action of the velum, a densely ciliated organ which also provides for locomotion. Gallager & Mann (1980) measured the feeding rates of larval sea hares and found that the highest rate of grazing on the flagellate

APLYSIA

191

Isochrysis galbana by veliger larvae of Aplysia californica (about 105·3 cells·ind−1·day−1) occurred at algal cell concentrations of 105·ml−1. At concentrations higher or lower than this optimum level, grazing rates were markedly lower. For example, at a flagellate concentration of 104 cells·ml−1, a level customarily employed in larval cultures of A. californica and other sea hares (Kriegstein, Castellucci & Kandel, 1974; Switzer-Dunlap & Hadfield, 1977, 1981; Otsuka et al., 1981), rates of grazing were about 20 times less than at the optimal concentration. On the other hand, at flagellate densities of 106·ml−1, the lowest concentration employed in the culturing of A. brasiliana=willcoxi by Strenth & Blankenship (1978a), rates of grazing by A. californica veligers as measured by Gallager & Mann (1980) were about six times less than at the optimal concentration of 105 cells·ml−1. DIGESTION AND ABSORPTION Sea hares appear to digest mainly starches and simple sugars. Chemical digestion takes place primarily in the stomach and digestive gland, but may be initiated in the crop when digestive juices are regurgitated from the stomach (Howells, 1942). Howells reported that amylases are active in secretions from the salivary and digestive glands, and that the latter organ secretes a number of enzymes which hydrolyse sucrose, lactose, and maltose. A protease is present in all regions of the gut, including secretions from the salivary glands, and lipase activity has been identified (Howells, 1942; Cho, Pyeun, Byun & Kim, 1983). While Howells noted that a cellulase was absent or only weakly present in A. punctata, this enzyme was later positively identified in the crop and digestive gland secretions of this species by Stone & Morton (1958). Koningsor & Hunsaker (1971) and Koningsor, McLean & Hunsaker (1972) also found a cellulase in the crop juices of A. vaccaria. In none of these instances, however, has an exogenous origin of the enzyme, as for example from symbiotic bacteria in the gut, been entirely ruled out. The fate of digested food particles in sea hares, particularly the sites of absorption and types of transport mechanisms involved for such materials as sugars, peptides, and amino acids, are not well known. In comparison, transport mechanisms of sodium, chloride, and other ions across the intestinal epithelia in sea hares, especially A. californica, have been extensively investigated by Gerencser (1978, 1979a, b, 1981a, b, 1982, 1983, 1984a, b), Gerencser & Hong (1977), Gerencser & White (1980), Gerencser & Loughlin (1983), and Gerencser & Lee (1985). The value of a food to Aplysia should be reflected in the efficiency with which it is digested and absorbed. Highest values would be expected on diets giving the best growth and in instances of monophagy (e.g. A. juliana), where strong selection for traits which would maximize the capacity for digesting and absorbing a (where single foodstuff would be expected. Values for absorption efficiency, expressed as C=food consumed and F=faeces produced), are given in Table VIII for sea hares eating a variety of diets. It may be seen that absorption of total dry matter ranges from 15– 84%, and highest values are predictably displayed by the monophagous A. juliana eating its favoured diet of Ulva spp. (73–84%; Carefoot, 1970; Sarver, 1978). In Aplysia punctata, absorption of total nitrogen varies from 54–79%, with no correlation with diet. Similarly, absorption of carbohydrates varies from 55–84%, again with no correlation with diet, but paralleling the pattern found for absorption of nitrogen. When the actual amount absorbed of each of these materials is, however, calculated over a known period of time for the various diets shown in Table VIII for A. punctata, the values obtained correlate positively and almost perfectly with the quality of each diet as reflected by its growth-promoting ability (Carefoot, 1967a). Uptake of specific amino acids by A. punctata was found to be highly variable, ranging from 0–100% for a variety of seaweed diets (Carefoot, 1967a). For animals in a sublittoral area studied in the Irish Sea, highest mean uptake of 17 amino acids was realized on a diet of the red alga Plocamium cartilagineum

192

THOMAS H.CAREFOOT

(mean of ). This seaweed gave the best growth and was the species on which the animals were most commonly found (Carefoot, 1967a, c). Overall, the highest mean value for absorption of amino acids was found for Aplysia punctata eating three species of red algae used in the study (75% as noted above for for animals eating Heterosiphonia plumosa, and for ones eating Delesseria Plocamium, sanguinea). In comparison, absorption of amino acids from diets of the green seaweeds Enteromorpha and respectively). intestinalis and Ulva lactuca was markedly lower ( A special feature of green algae which may bear relevance to their usefulness as foods for Aplysia, is their generally low content of amino acids (e.g. 3·5 and 5·1% of total dry weight in Enteromorpha intestinalis and Ulva lactuca, respectively, as compared with 18·2–21·2% for the three species of red algae listed above; Carefoot, 1967a, b). When the two factors of poor absorption and low concentration of amino acids are combined, a picture emerges which TABLE VIII Absorption and growth efficiencies for Aplysia: A=absorption; C=Chlorophyceae; CHO=carbohydrates; F=faeces; N=nitrogen; P=Phaeophyceae; R=Rhodophyceae; C=consumption; P=production Absorption efficiency (C−F/C×100%) Species dactylomela

R

dactylomela

C C C R R

juliana

C C C C C C C

punctata

R

punctata

R

N

Growth efficiencies

Food

Temp. °C Total dry mattera

CHO Gross, K1 Net, K2 References (P/C)×100% P/A×100%

Laurencia spp. Cladophora sp. Enteromorph a sp. Ulva fasciata Galaxaura oblongata Laurencia papillosa Enteromorph a sp. Ulva fasciata Cladophora Ulva lactuca Ulva lactuca: low rationb Ulva lactuca: medium ration Ulva lactuca: high ration Plocamium cartilagineum Plocamium cartilagineum

17

79

22

28

Willan, 1979

28

35

29

84

Carefoot, 1970

28

68

45

67

28 28

62 24

27

43

28

67

33

49

28

69

14

20

28 28 25 25

84 15 73

28

33

16 15

22

Sarver, 1978

Carefoot, 1967b Carefoot, 1967a

25

11

25

12

15

67

74

73

21

31

15

65

74

72

23

35

Carefoot, 1970

APLYSIA

Absorption efficiency (C−F/C×100%) Species C C R R R P a b

193

Growth efficiencies

Food

Temp. °C Total dry mattera

N

CHO Gross, K1 Net, K2 References (P/C)×100% P/A×100%

Enteromorph a intestinalis Ulva lactuca Heterosiphoni a plumosa Cryptopleura ramosa Delesseria sanguinea Laminaria digitata

15

59

68

69

17

29

15 15

75 71

79 74

84 76

18 15

24 21

15

71

73

76

11

15

15

45

54

57

33

74

15

53

57

55

20

38

Expressed in Joules in some studies. See Table IV for ration levels used in this experiment.

suggests that green algae may not, in fact, be particularly good sources of amino acids for Aplysia. This may explain why Ulva spp., which are otherwise favoured foods for sea hares and which are eaten well in laboratory studies, actually promote only poor growth in all species but Aplysia juliana (Carefoot, 1967a, 1970; see also pp. 234–236). Enteromorpha seems to be a superior diet for sea hares (Carefoot, 1967a, 1970), but measurements of consumption rates in ad libitum laboratory studies on Aplysia punctata have shown that 50–100% more of this green alga must be eaten to yield growth rates comparable with rates attained on the best red-algal diets (Carefoot, 1967a). It would be interesting to know whether absorption of amino acids is more efficient in A. juliana to match the overall high absorption of dry matter and high rates of growth exhibited by this species on its specialized diet. In addition to food quality, other factors which may affect absorption efficiency in Aplysia are temperature, feeding rate, age, reproductive state, and density of individuals. All are known to affect absorption efficiencies in other gastropods (Carefoot, 1987) but, save for feeding rate, have not been investigated in sea hares. As expected, feeding rate is negatively correlated with absorption efficiency in A. punctata (Carefoot, 1967b). Also, sea hares rapidly eating green algae may digest and absorb their foods so poorly that their faeces are coloured bright green (see also Winkler & Dawson, 1963). This suggests that in such circumstances of “superfluous feeding” (e.g. Ryther, 1954) the faeces may contain substantially more unused nutrients than when animals feed more slowly. After such bouts of quick feeding, A. punctata may occasionally eat its own faeces (Carefoot, 1967a). Coprophagy, although apparently rare in sea hares, may play a rôle in nutrition. NUTRITION Despite considerable research on diets, feeding preferences, and growth in sea hares, knowledge of their specific nutritional requirements is scanty (Carefoot, 1967a, b, 1970; Sarver, 1978). Reasons for this partly relate to difficulty in ascertaining the precise nutritional quality of the seaweed foods and to the absence of any radiotracer studies to monitor uptake and assimilation. In addition, large seasonal changes in chemical composition of seaweeds, diffusion of nutrients from algal tissues during growth, and loss of nutrients

194

THOMAS H.CAREFOOT

during the mechanical processes of feeding create special problems in determining the precise nutritional needs of sea hares. The development of artificial diets, made up entirely of chemicals bound in agar, has solved some of these problems and has provided a means for preliminary assessment of nutritional needs in Aplysia (Carefoot, 1979, 1980). Experimental diets are eaten well by some species, but not by others. They sustain some growth and spawn production in A. dactylomela, maintain constant weight in A. kurodai, but are not eaten at all by A. juliana. When juvenile A. kurodai were fed on such chemically defined diets, each diet deficient in a different single amino acid, no significant weight losses occurred after 24 days on any of the diets (Carefoot, 1981b). Assuming that this period was sufficiently long for amino acid deficiencies to become apparent in these fast-growing animals, the results suggest that either A. kurodai do not require the usually recognized 10 essential amino acids (as shown for the rat), or that they obtain required amino acids from another source. Save for the possibility that uptake of dissolved organic matter might provide some of the missing amino acids, the most likely source is from symbiotic bacteria in the gut. In this regard, Ghiretti, Ghiretti-Magaldi & Tosi (1959) isolated several strains of bacteria from the digestive glands of A. depilans and A. fasciata, and Vitalis (1981) identified some twelve different colony types of bacteria from the crop and stomach regions of A. juliana. Six of these types in A. juliana were abundant and appeared consistently; the others were sporadic and always occurred in small numbers. A preliminary study by Spence (unpubl.) on the effect of diet on the composition of bacterial flora, their distribution through the gut, and the effect of antibiotics on survival of the bacteria in the guts of A. dactylomela showed the following: (1) that up to 23 bacterial strains could be identified in the gut of A. dactylomela, (2) that the bacteria existed more or less evenly throughout the crop, gizzard, and digestive gland, both in numbers of species and numbers of individuals to bacteria·wet g gut tissue plus fluids−1 throughout these regions), (3) (numbers ranged from that animals starved for five days showed a slight, but non-significant, decrease in numbers and types of bacteria in the gut, (4) that numbers and types of bacteria were somewhat greater in animals eating red algae (Spyridia sp. and Laurencia sp.) than in ones eating Ulva sp., (5) that six bacterial types appeared to be ‘resident’ forms, present in the gut regardless of diet and also present in starved animals, (6) that oral administration of antibiotics (ampicillin, neomycin, and tetracycline incorporated into artificial diets of chemical nutrients set in agar) resulted in a one order of magnitude reduction in numbers of gut bacteria over 10 days of treatment and, finally, (7) that in vitro tests of susceptibility of the bacteria to several antibiotics showed that the most effective antibiotics amongst those tested were ampicillin and chloromycetin, and the least effective, erythromycin, kanamycin, neomycin, streptomycin, penicillin, tetracycline, and polymycin B. The fact that 12 types of bacteria were identified in A. juliana (Vitalis, 1981), as opposed to 23 types in A. dactylomela, further suggests that types of bacteria may be somewhat greater in animals eating red algae than in ones eating green algae. Some of the antibiotics employed in the study (e.g. tetracycline) appeared to be distasteful to A. dactylomela in concentrations necessary to kill significant numbers of the bacteria. Also, in several instances, bacterial numbers actually increased after eight days of antibiotic treatment in the in vivo experiments, possibly caused by a secondary invasion of antibiotic-resistant forms following an early reduction in numbers of the primary populations by the antibiotics (Spence, unpubl.). Vitalis (1981) undertook a series of experiments with A. juliana, fed on Enteromorpha sp., Ulva reticulata, and U. fasciata, to assess the nutritional contribution made by gut bacteria. He showed that growth of animals exposed to antibiotics dissolved in the sea-water medium (streptomycin and penicillin each in concentrations of 10 mg·l−1) was 40– 50% less than ones in a sea-water medium without antibiotics over a 30-day period. Counts of gut bacteria in treated animals showed a 98% reduction in number compared with control animals (to bacteria·ml gizzard fluid−1). While it is tempting to conclude that the reduction in

APLYSIA

195

numbers of bacteria caused the decreased growth in these sea hares through a lessened availability of bacterial metabolites to the sea hares, the possibility exists that the antibiotic treatment itself was responsible. Vitalis tested this by exposing A. juliana to high concentrations of antibiotics. He showed that A. juliana survived well, with no obvious effects on feeding or other behaviour, in concentrations of combined streptomycin and penicillin of 1000 mg·l−1, equal to 50 times the dosage used in the experiment. While further work is needed to assess any deleterious effects of antibiotics on growth and metabolism of the host animal, the results of Vitalis’ study are none the less provocative. The possible bacterial source of required amino acids was further tested in a series of experiments with juvenile A. dactylomela (Carefoot, 1981a). Here, animals were given a chemically defined diet deficient in arginine (essential for the rat) and containing antibiotics to suppress the activities of gut bacteria. The sea hares gradually lost weight over a 20-day period on the experimental diet (Carefoot, 1981a). When the arginine-deficient diet was replaced on Day 21 with a diet complete in all nutrients, the sea hares immediately began to gain weight and continued to do so for the remaining 16 days of the experiment. These data, combined with the previous observations on A. kurodai and A. juliana, suggest that sea hares may rely on their bacterial symbionts for provision of essential amino acids and other nutrients just as has been found in other animals (e.g. insects, isopods, sea urchins, and vertebrates). This may explain how sea hares, like sea urchins (Fong & Mann, 1980), can utilize a wide variety of seaweeds as food. If bacteria are involved in this way in sea hares, further nutritional studies using traditional techniques of deletion and augmentation of specific nutrients with chemically defined artificial diets would be pointless unless a way is found to eliminate completely gut bacteria or at least to suppress their metabolic activities. Even if this is done, it must be demonstrated that this technique is affecting only the bacteria in question. Alternatively, animals could be grown axenically. Given our current knowledge of Aplysia larvae and current techniques of culturing them through metamorphosis, it should be a routine matter to grow sea hares with sterile guts. This would open a marvellous range of possibilities for their further nutritional study. FEEDING ECOLOGY The only true way to ascertain the effect of sea hares or any herbivore on an algal community is to remove the animals from an area and to keep them out for a long time. This can be done simply by hand-clearing animals or by the use of “exclosure” cages. The latter have been used in two studies on sea hares: the first on A. dactylomela and A. parvula in New Zealand (Willan, 1979); the second, on A. dactylomela in Barbados (Carefoot, 1985). In neither case were the results completely convincing. The small size of exclosure cages presented problems in both studies, as did too fine a size of mesh employed as screening, the possibility in both studies of unknown “cage effects” and, in the Barbados study, too short a period of study. If any conclusion could be drawn from the Barbados work, it would be that herbivorous fish may in fact have exerted a greater impact on the algal community than did the population of A. dactylomela (Carefoot, 1985). Willan (1979) was able to show significant differences between the biomass TABLE IX Densities of Aplysia in different geographical areas Species

Geographical area

Time of year

Density (no·m−2)

References

californica dactylomela

Ellwood, Southern California Barbados

Apr. 1975– Feb. 1977 summer

0·03 0·6

Sousa, 1979 Carefoot, 1985

196

THOMAS H.CAREFOOT

Species

Geographical area

Time of year

Density (no·m−2)

References

juliana

North Island, N.Z. Hawaii North Island, N.Z.

0·9–1·6 1·0 3·5 3–5

Willan, 1979 Sarver, 1978

parvula

Mar.–July (autumn-winter) summer/autumn winter/spring Oct.–Nov. (spring)

Willan, 1979

of Laurencia spp. within exclosure cages as opposed to without after 14 months (three of four cages ended up with a percentage cover of Laurencia about 40% greater than the uncaged control areas; the fourth cage had 25% less cover of Laurencia than its control). The author expressed a concern about possible “cage effects” in his study, such as shading and accumulation of sand within the cages, as well as how other large herbivores were excluded along with Aplysia spp. Willan did not employ a control cage or cages in his study, although there is a question whether their inclusion would be useful since it is difficult to devise adequate control cages and then correctly interpret the data obtained from them. With their large sizes and concomitantly large appetites, sea hares should exert a profound influence on the seaweed community. It is therefore surprising that so little work has been done on Aplysia in this regard in comparison to, say, sea urchins or limpets (see reviews by Lawrence, 1975 and Branch, 1981, respectively). What we find ultimately is that sea hares are never very numerous in a given area, with densities rarely exceeding 2 animals·m−2 (see Table IX). Compared with densities observed for sea urchins in areas where effects of their grazing may be pronounced (e.g. 60–1000 individuals·m−2: Paine & Vadas, 1969; Foreman, 1977; Chapman, 1981) such low numbers of sea hares, even with their greater comparative rates of feeding (Carefoot, 1981), would seem unlikely to be capable of exerting comparable large-scale effects on the seaweed community. That diet and skin colour are probably related in sea hares has been suggested by several authors (Eales, 1921, 1960; Grigg, 1949; Winkler, 1958b, 1959b, c; Kandel, 1979; Willan & Morton, 1984). The idea arose from an early and somewhat cursory observation by Garstang (1890) on colour change in A. punctata, as well as from biochemical analyses of skin pigments of various species of Aplysia. These analyses suggested that the colours in A. punctata, A. depilans, and A. californica derive from degradation products of the tetrapyrrole molecule of chlorophyll (e.g. porphyrins and bilins: MacMunn, 1899; Schreiber, 1932) or from other algal pigments (e.g. phycoerythrin and phycocyanin: Winkler, 1959b). A porphyrin has also been found in the skin of A. punctata by Kennedy & Vevers (1954) and several carotenoids in A. punctata, A. depilans, and A. fasciata by Czeczuga (1984). Since the skin colour resulting from prolonged feeding on an alga is thought to match the colour of the seaweed, a camouflage function is hypothesized. As attractive as this notion seems, there are several points that should be considered. First, the effect of diet on skin colour has not been convincingly demonstrated in experiments. Winkler (1959b) showed that brown-, dark green-, and grey-coloured A. californica changed over a period of 1–3 months to a uniform light-brown base colour when fed on diets of parsley leaves and celery tops. He further observed that if these blanched specimens were fed on a large amount of Plocamium cartilagineum, the base colour of the skin turned decidedly, but temporarily, pink. A similar pink undertone colouration has been observed also in Aplysia dactylomela in Hawaii from areas where they appear to be eating primarily red algae (Switzer-Dunlap, pers. comm.). Winkler also noticed that certain animals bearing red body streaks and purple foot colourations had been eating predominantly Plocamium cartilagineum, as determined from faecal analysis. Winkler did not attempt to re-establish the original colours in his experimental animals by feeding them subsequently on diets of different coloured seaweeds. When small (2 g live weight) sublittorally collected Aplysia punctata were fed for six weeks on Laminaria, 11 weeks on

APLYSIA

197

Enteromorpha and Ulva, and for similar lengths of time on various red algae, the animals became only a darker shade of red, similar in colour to that of freshly collected sublittoral animals of equivalent size (Carefoot, 1967a). These field animals were eating mainly the red alga Plocamium cartilagineum in their sublittoral habitat (Carefoot, 1967c). A second point with regard to diet and skin colouration is that Aplysia juliana feeds only on various species of green algae, principally Ulva spp., yet is always brown in colour, often of a rich chocolate hue. A brown coloured A. juliana cannot be described as cryptic amongst its bright green coloured food. The notion of adaptive colouration in Aplysia in response to diet arose as part of a theory of migration from deeper offshore areas to the intertidal region during an animal’s life, with a gradual colour change occurring as the animals browsed successively through red, brown, green and, finally, olive green-coloured seaweeds (see pp. 202–203). Although no evidence has been provided in support of this theory, such colour changes were none the less predicted and possibly expected by workers in the field. Despite these comments, however, it seems likely that diet does affect colouration in Aplysia. It remains to be shown convincingly in experiments, although some points support this idea. (1) It is known that eggs are coloured in response to different seaweeds being eaten and the mechanism of pigment transfer and incorporation from food to eggs has been documented (Chapman & Fox, 1969). (2) Switzer-Dunlap (1978) notes that in species of sea hares which as juveniles feed on red algae (A. brasiliana=willcoxi, A. californica, and A. parvula) the overall body colour is pink initially and grows progressively darker with continued feeding and growth. (3) Winkler (1959c) has observed that A. californica fed on Plocamium cartilagineum develop characteristically dark coloured blood and purple subcutaneous pigment deposits, and Chapman & Fox (1969) have noted that Aplysia californica fed artificially on phycocyanin pigment exhibit a distinct blueing of the inner skin. All of these points suggest that skin colour is at least partially regulated by diet in sea hares. An inverse relationship of size of animal to the depth at which it occurs was found for a sublittoral population of A. punctata in the Irish Sea (Carefoot, 1967c). As size was strongly and positively correlated with biomass of Plocamium cartilagineum, it was presumed that the animals were simply eating more, and growing larger, in shallower depths where the alga was more abundant. Willan (1979) noted also that maximum densities of Aplysia dactylomela on Echinoderm Reef, New Zealand, occurred during MarchJune when Laurencia spp., its favoured food, were dominant. The Reef, 2·76 hectares in size, hosted up to 6000 animals during peak densities in June. Willan calculated that during March-June the equivalent energy required in seaweeds by the Aplysia dactylomela population was 4·4 kJ·ha−1·month−1 while the energy available represented by Laurencia spp. at the site was never less than 59·8 kJ·ha−1·month−1. Overall, for all months of the year, the energy represented by the standing crop of Laurencia spp. never fell below six times that required by the Aplysia dactylomela population. While the author noted that other herbivores, including two aplysiid species, also grazed on the same standing crop of Laurencia spp., it would seem unlikely that the Aplysia dactylomela population in this area would ever be food-limited. An interesting aspect of feeding behaviour in A. juliana was identified by Saito & Nakamura (1961) and Frings & Frings (1965), in which animals appeared to eat only the succulent distal portions of the brown alga Undaria pinnatifida or Ulva lactuca plants, respectively, while leaving the thicker, coarser, stem or mid-rib sections. The ecological significance of this behaviour was evident, since the basal portions, at least of Ulva, were observed to produce new fronds within a week or two. Only under conditions of long starvation or if “Ulva-water” (prepared from succulent new growth) were flooded over their mouths, could the sea hares be induced to eat the tough basal sections (Frings & Frings, 1965). The authors conjectured that the phagostimulatory material present in the new growth may be absent or of insufficient concentration in the bases of the plants to stimulate the animals to eat. A similar response to food texture was shown for

198

THOMAS H.CAREFOOT

Aplysia fasciata by Susswein & Schwarz (1983), who made otherwise palatable foods (Ulva lactuca and leaves of the lily Hemerocallis fulva) too tough to swallow by enclosing them in a plastic net. These preparations became stuck in the buccal cavity and the sea hares soon learned not to eat them. In another related experiment on Aplysia oculifera, Schwarz & Susswein (1984) denervated the crop region which destroyed this ability to learn. The authors intimated that A. fasciata in their natural habitat might learn that certain seaweed species, as for example the calcium carbonate-impregnated Jania rubens, are too tough to eat (Susswein & Schwarz, 1983). Susswein, Weiss & Kupfermann (1984b) suggested that when soft food enters the crop it may activate receptors that reinforce feeding upon soft foods, whereas unsuccessful entry of tough foods may act as a negative reinforcer. Whatever the proximate causes of such behaviour, the end result for feeding sea hares would be a cropful of easily digestible soft seaweeds. In addition, if the sparing of the tough bases of Ulva plants by Aplysia juliana were a pattern followed by other Aplysia, then there exists the ultimate prospect of a renewed harvest of fronds as a result of the animals’ unique feeding ecology. THE EFFECT OF FEEDING ON OTHER BEHAVIOUR The effect of contact with food, of feeding, and of varying degrees of satiation on other behavioural activities has been extensively investigated in A. californica. Advokat (1980), for example, has shown that various defensive activities such as siphon withdrawal, locomotion, and inking are measurably attenuated following a meal (see also Advokat, Carew & Kandel, 1976). For the siphon-withdrawal reflex, actual ingestion of food is not required to produce a significant diminution of the response; this can be accomplished simply by applying food to the lips and oral veil regions. Kupfermann (1974b), Kupfermann & Carew (1974), and Susswein & Kupfermann (1974, 1975a, b) have similarly found that after a large meal, satiated A. californica stop moving and fail to bite at other foods. In some instances, animals remain in this “frozen” state for many hours. Conversely, Kupfermann & Weiss (1981) have shown that certain “aversive” stimuli, such as tail-pinching and handling, enhance feeding behaviour in A. californica. The functional significance of such behavioural modifications is unclear as it does not seem advantageous for an animal to ‘shut down’ or reduce its locomotory and defensive capabilities during or after eating. In this regard, the suggestion by Advokat (1980), that attention to food brings with it a concomitant inattention to other stimuli —that animals cannot flee and eat simultaneously—may well explain the situation, but does not clarify its biological rôle. An unusual behaviour, apparently in response to food deprivation, was observed in both the field and laboratory for A. brasiliana=willcoxi by Aspey, Cobbs & Blankenship (1977). The animals swim at the surface and exhibit a characteristic “head-bobbing” activity where the head and anterior portion of the body are stretched to extend the oral tentacles momentarily and repeatedly out of the water (Fig. 5). Observations on field animals showed head-bobbing frequency to be about 40·min−1, with bouts of activity lasting about 47 s on average. Laboratory animals were seen to increase head-bobbing frequency at the approach of their regularly scheduled feeding time, and later experiments to test this showed that the frequency increased in direct relation to the length of time they had gone without food. After 41 h without food, head-bobbing, however, stopped. Aspey et al. (1977) suggested that head-bobbing serves to increase the chances of encountering floating and/or surface food but, after several hours, becomes energetically wasteful. When touched on the oral region with food (dried laver: Porphyra sp.) or with a piece of paper towelling during swimming and/or head-bobbing, the animals dropped immediately to the bottom: a similar response to touch was noted by Hamilton & Ambrose (1975) for swimming individuals of the same species. Despite the strong correlation of frequency of bobbing with duration of food deprivation, this complex behaviour seems

APLYSIA

199

to have a function beyond that of simply finding food. In the first place, the animal’s food is benthic, not floating. Moreover, head-bobbing would be unlikely to enhance the probability of finding food: swimming on the surface would itself serve this function. Hamilton & Russell (1981a, b) and Hamilton (1986) suggest the most likely hypothesis, that while swimming allows the animal to move out of a region of food scarcity, head-bobbing aids navigation. A sexual rôle has not been intimated for this unusual behaviour as yet, even though a common response of opisthobranchs, including Aplysia, is to seek out mates, copulate, and lay eggs in times of food scarcity (see pp. 183–188). GROWTH AND ENERGETICS As might be expected in a large animal with an annual life cycle, growth rates are rapid. An early postmetamorphic animal may attain a rate of size increase of 13000% over a two-week period (Hadfield, 1975). The largest recorded sea hare was a 6·8-kg A. californica collected from Elkhorn Slough, California (MacGinitie, 1935); the largest size estimated, however, was 15·9 kg for a sublittoral A. vaccaria (Limbaugh: in Winkler & Dawson, 1963). High rates of feeding, fast growth rates, and large production of spawn were translated at one densely populated site in Hawaii to 11 kg of algae eaten per day and close to produced per day by a population of A. juliana (Sarver, 1978). one trillion eggs GROWTH The veliger larva represents the first stage of growth in sea hares and spends its life swimming and feeding in the plankton. It is separated from the second, or benthic adult phase, by a metamorphosis. Rates of growth of the veliger larvae are well known through laboratory studies (Krakauer, 1969; Kriegstein et al., 1974; Switzer-Dunlap & Hadfield, 1977). In post-metamorphic animals the pattern of growth is sinusoidal and involves a ‘lag’ phase, followed by an exponential phase, leading eventually to a levelling off (Sarver, 1978; Peretz & Adkins, 1982). Rates vary greatly in each stage and these differences must be accounted for when comparing rates in different species and especially in animals of different sizes. Growth rates of post-metamorphic animals are generally calculated from absolute measurements of live weight, or live weight converted to dry weight or energy equivalents, over time, defined as:

where W represents weight and t represents time. Because of the different rates of growth of animals at different sizes, as noted previously, such data on absolute growth rates may, however, not be useful for interspecific comparisons. Growth may be expressed better in relative terms, such as:

which gives information on growth increment per unit size per unit time, or defined according to the growth coefficient, K:

where m is the slope of a regression line relating size at time t+1 to size at time t in a Ford-Walford plot (Walford, 1946; see also Branch, 1981; for equation see Table X). The value of K is useful in growth

200

THOMAS H.CAREFOOT

Fig. 5.—“Head-bobbing” in swimming. A. brasiliana=willcoxi: figure modified from Aspey, Cobbs & Blankenship (1977).

APLYSIA

201

studies as it can be used as a comparative index of growth that is independent of body size. Moreover, when the regression line in a Ford-Walford plot is extrapolated TABLE X Growth rates of Aplysia: N=number of replicates; the Ford-Walford growth equation, is usually based on units of length L, for purposes of the comparisons below, weight was substituted instead thus, Lt=weight at time t and Lt+1=weight after a given period of time (here=2 wk); values are given for i and m; C=Chlorophyceae; i=intercept; m=slope; P=Phaeophyceae; Pr=spawn production; Ps=somatic production; R=Rhodophyceae; r=correlation coefficient Species

i

Food

N

Startin Temp. g °C weight (mean live g)

Rate of Ford-Walford growth growth equation statistics dw/dt (g live (wt·da y−1)

K growth coeffic ient K= −logem

Live g Comme Referen alga·g nts ces growth −1

(growt h=Ps +Pr)

m r

brasili ana =willc oxi

califor nica

R

califor nica

R

dactyl omela

C

Field seawe eds

9

201

13–28

1·7

16·6

1·04

0·99

−0·04

Laure ncia sp., Polysi phonia sp., Dasya sp., and Porph yra sp. Dried Porph yra sp. and fresh romain e lettuce (Lactu ca saliva longifo lia) Entero morph a sp.

9

59·3

18

6·2

49.4

1·64

0·97

−0·49

15 12 9 11 4

1·0 30 85 310 850

16 16 16 16 16

0·6 1·0 3·3 4·2 5·3

4

52·0

28

2·9

40·2

0·99

0·56

0·01

4·1

Data from a single growth curve (field study) Data from a single growth curve

Kraka uer, 1969

Numb er of days of study: variabl e up to 235

Peretz & Adkin s, 1982

Kriegs tein et al., 1974

Carefo ot, 1970

202

THOMAS H.CAREFOOT

Species

i

Food

C

R

R

Ulva fasciat a Clado phora sp. Laure ncia papillo sa Galax aura oblong ata

R C C C

juliana

C

C C juliana

Startin Temp. g °C weight (mean live g)

Rate of Ford-Walford growth growth equation statistics dw/dt (g live (wt·da y−1)

K growth coeffic ient K= −logem

Live g Comme Referen alga·g nts ces growth

8

66·2

28

1·5

19·3

1·03

0·93

−0·03

2·7

4

49·0

28

1·5

26·3

0·89

0·75

0·12

6·5

4

50·1

28

1·7

−19·2

1·85

0·84

−0·62

4·4

10

56·6

28

−1·1

−1

(growt h=Ps +Pr)

m r C

dactylom ela juliana

N

C

Laurenci a spp. Ulva fasciata Enterom orpha sp. Cladoph ora sp. Ulva fasciata

Ulva reticulata Ulva lactuca Enterom orpha sp.

10

37·4

14–17 0·8

7·12

1·11

0·98 −0·10

10

29·3

28

1·6

18·4

0·82

0·86

0·20

1·9

6

24·9

28

0·3

8·5

0·81

0·96

0·21

11·3

6

23·7

28

−0·5

0·2

0·69

0·93

0·37

10

0·01

25

0·29

0·67

343·5 0·97

−5·84

6

0·01

25

0·11

−0·07

241·7 0·91

−5·49

5

0·01

25

0·36

3·04

162·1 0·98

−5·08

10

1·0

23·5

0·32

−0·11

5·40

−1·69

0·91

Willan, 1979 Carefoot, 1970

Data from a single growth curve for each food species

Sarver, 1979

Data from a single growth

Vitalis, 1981

APLYSIA

203

curve for each food species C C punctata

R

C

C R

R

R

P P

punctata

R

Ulva reticulata Ulva fasciata Plocamiu m cartilagi neum Enterom orpha intestinal is Ulva lactuca Heterosi phonia plumosa Cryptopl eura ramosa Delesseri a sanguine a Laminari a digitata Desmare stia aculeata Plocamiu m cartilagi neum

11

1·1

23·5

0·26

−1·07

5·27

0·93

−1·66

10

1·2

23·5

0·30

−0·96

8·26

0·93

−2·11

26

9·5

15

0·20

3·4

0·94

0·99

0·06

3·0

27

8·4

15

0·19

3·5

0·90

0·96

0·11

5·4

31

6·1

15

0·09

2·1

0·85

0·97

0·16

2·9

20

4·3

15

0·08

1·8

0·84

0·99

0·17

4·4

30

4·0

15

0·03

1·4

0·77

0·97

0·26

6·0

29

3·0

15

0·04

3·3

12

1·8

15

0·02

6·2

10

1·9

15

0·0

18

6·2

15

0·2

2·95

0·99

0·99

0·01

3·2

Spawn included as “growth”

Carefoot, 1967a

Spawn included as “growth”

Carefoot, 1967b

to the ordinate axis and to its intersection with the 45° diagonal (the point at which size at time t is the same as at t+1; i.e. when there is no further growth), growth rates can be estimated for all sizes of the population, assuming that growth is constant over the period considered. Although best used for slow-growing animals (such as limpets and other snails: Ward, 1967; Hughes, 1971a, b, 1972; Balaparameswara Rao, 1976; Branch, 1981), the growth coefficient K can none the less provide a useful index for comparison of growth in other animals. Its usefulness declines, however, when sample numbers are small, when size ranges are narrow, or when animals are young and in a fast-growing phase. Data for a number of growth studies on Aplysia are presented in Table X. Growth is expressed as absolute rates in g live wt·day−1, as regression data for Ford-Walford plots of size at time t+1 against size at time t, and as values for the growth coefficient, K. Absolute growth rates of Aplysia range from 0·2–6·2 g live weight·day−1 for animals eating their optimal seaweed foods (e.g. A. punctata eating Plocamium

204

THOMAS H.CAREFOOT

cartilagineum and Aplysia californica eating Laurencia sp. and other red seaweeds), with the variability, for the most part, probably due to starting sizes of the animals. Major factors affecting growth rates in Aplysia are: (1) temperature and light—as related to season, microgeography, and the tides; (2) size and age; (3) reproductive state; (4) water movement; and (5) food, both quantity and quality. Of these, the effect of food has been studied most completely. The effect of diet quality on growth in Aplysia is shown clearly for A. punctata feeding on eight seaweed diets in the laboratory (Table X; Carefoot, 1967a). The diets represent five common species in the animal’s sublittoral habitat (the red algae, Plocamium cartilagineum, Heterosiphonia plumosa, Cryptopleura ramosa, and Delesseria sanguinea, and the brown alga, Desmarestia aculeata), including Aplysia punctata’s preferred field diet, Plocamium cartilagineum. The three other algae (the green seaweeds, Enteromorpha intestinalis and Ulva lactuca, and the brown alga, Laminaria digitata) represent abundant intertidal forms, are commonly eaten by intertidal animals, and are species thought to be principal food of Aplysia punctata (e.g. Eales, 1921). Best absolute growth occurs on a diet of Plocamium, closely matched by that on a diet of Enteromorpha. Other diets support variable, but poorer rates of growth. Values for K in this study are 0·06 and 0·11 for Aplysia punctata eating the foods which give overall best growth (Plocamium cartilagineum and Enteromorpha intestinalis, respectively). The larger values for K in the series for Aplysia punctata indicate that the regressions of size at time t+1 over size at time t flatten out as the animals increase in size; that is, the relative increment of growth becomes smaller with increasing size. This is evident for animals eating foods that are nutritionally poor (i.e. giving poor growth), such as for A. punctata eating Cryptopleura ramosa, where K=0·26 (also Aplysia dactylomela and A. juliana eating Cladophora sp., where K=0·12 and 0·37, respectively; see Table X). Similar effects of diet quality on growth of sea hares have been shown by Sarver (1979) and Vitalis (1981) in studies of Aplysia juliana. Of the several Ulva diets tested in these studies, U. fasciata proved best nutritionally for Aplysia juliana, as shown by lowest K values (−5·84 as compared with −5·49 and −5·08 for animals eating Ulva reticulata and U. lactuca, respectively; data from Sarver, 1979; see Table X). The negative values for K obtained in these studies indicate that the tiny Aplysia juliana were in their exponential phases of growth, with lowest values being associated with the steepest positive slope of the regression line in the Ford-Walford plot. The effect of diet quantity on growth of Aplysia was investigated by Sarver (1978). He reared A. juliana in laboratory culture for 137 days on “low”, “medium”, and “high” rations of Ulva lactuca (representing 5, 10, and 15 g alga·day−1·ind.−1, respectively). Sarver showed that while P (growth including spawn) as a proportion of C (food consumption) did not vary significantly on the different ration levels (values obtained were 15, 11, and 12% of C for animals eating the “low”, “medium”, and “high” rations, respectively; see Table VIII, p. 222), the absolute amount of growth which each group of Aplysia juliana attained did differ considerably. For example, animals on a “high” ration gained approximately twice as much weight as did animals on a “low” ration and produced nearly three times as much spawn; “medium”-ration animals were intermediate in both respects. In his studies on A. juliana, Sarver (1978) investigated absorption and assimilation of foods and channelling of energy into various body processes in order to compare animals of different “physiological” ages (i.e. size) with ones of different chronological ages. He found that an animal maintained on a low ration of food for several weeks and then returned to an ad libitum diet behaved physiologically like a much younger animal in terms of growth and spawn production, even though normally it would have long since stopped growing at that chronological age. Its new rate of growth corresponded to its physiological age, not to its chronological age. This line of research is highly provocative in that it provides an elegant means of investigating various processes of ageing in sea hares. It also allows comparison of sea hares from different

APLYSIA

205

geographical areas, where relative size may not be a true reflection of actual age. The availability of animals of known age through their culture in the laboratory, combined with the possibility of ascertaining age of field animals from shell size or other growth characteristics of the shell (see p. 236), would help to remove the uncertainty of not knowing an animal’s age in physiological and behavioural studies of sea hares where this knowledge is important. Few of the many environmental factors which could affect growth rates in Aplysia have been investigated, although Sarver (1978) has taken a preliminary look at the effects of light and water current on growth of A. juliana. He found that growth was about 75% faster in animals in slow-moving water than in a strong current (approximately 1 m·s−1), suggesting that currents may interfere with normal feeding or with attachment and movement. In comparison, growth was about 120% greater in animals kept continually in the dark than in those exposed to sunlight. As this nocturnal species responds to light by burying itself during the day and normally feeds only during hours of darkness, a continuous dark regime allowed almost constant feeding and greater subsequent growth (Sarver, 1978). and for K2, net growth efficiency of Values for K1, gross growth efficiency and from 15–84% sea hares are presented in Table VIII (p. 222). They range from 11–45% for Highest values of growth efficiency are generally, but not invariably, associated with foods for giving best growth. These foods, in turn, tend to be the ones favoured, and this trend is shown best for A. dactylomela and A. punctata. For A. punctata, K2 efficiencies decline from high values of 31–35% on the animal’s favoured field diet of Plocamium, to a low of 15% on the less-preferred diet of Cryptopleura (Carefoot, 1967a). Two diets eaten most poorly in the same study and giving poorest growth, the red alga Delesseria and the brown alga Laminaria, actually give the highest net growth efficiencies of 74 and 38%, respectively, suggesting that there may have been some sort of physiological compensation for the poor quality of these diets. The values for K1 and K2 efficiencies in Table VIII are calculated on the bases of either dry weight or energy. When live-weight units are used to relate growth and food consumption, a slightly different picture emerges as to the relative value of different seaweeds as food for Aplysia. This is shown in the final two columns in Table X, where growth is expressed as the number of live g of each alga required to produce a single g of either somatic or spawn tissue. On this “per mouthful” basis, Ulva spp. appear to be the optimal diet for sea hares (1·9–2·9 live g Ulva spp.·g growth−1), followed by some of the red seaweeds, including Plocamium and Delesseria. In actuality, Ulva spp. are not the best foods for most sea hares in the laboratory —only for Aplysia juliana do they promote most rapid growth. Missing from this presentation is the factor of how much of each food is actually eaten and this, in turn, depends on other factors such as palatability, texture, ability of the animal to manipulate the food, and so on. A possible compensatory type of interaction exists between growth efficiency and absorption efficiency in sea hares. Diets which are poorly absorbed are assimilated more efficiently than are ones which are highly absorbed. This negative correlation of growth efficiency and absorption efficiency is a well-known phenomenon in aquatic animals, including fish (Welch, 1968) and other gastropods (Carefoot, in press). From visual inspection of a graphical representation of these data for Aplysia (Fig. 6) there does not seem to be an effect of algal type (i.e. red, green, or brown) on the relationship between net growth efficiency and absorption efficiency. In addition to quantity and quality of diet, other factors of size, age, and reproductive state are known to affect growth efficiency in gastropods (Streit, 1976; Edwards & Huebner, 1977; Huebner & Edwards, 1981; Macé, 1981; Ansell, 1982). Of these, only age has been investigated in Aplysia and is shown by: (1) a decrease in K2 or net growth efficiency from 55 to 17% over an 80-day period for A. punctata growing from 0·9 to 18·4 g live weight (Carefoot, 1967b), and (2) a decrease in K1 or gross growth efficiency from 42 to

206

THOMAS H.CAREFOOT

Fig. 6.—Relationship of net growth efficiency (K2) and absorption efficiency in Aplysia: equation of regression, data from Carefoot (1967a, b, 1970), Sarver (1978), and Willan (1979).

12% over a 103-day period for A. juliana (including spawn as growth; Sarver, 1978). Temperature does not appear to affect either net or gross growth efficiencies in gastropods (Ansell, 1982) and, consequently, has been disregarded in the above comparisons of Aplysia where animals may have been maintained at different temperatures. Aspects of shell growth in Aplysia have been investigated by Winkler (1958a, 1959d), as have allometric relationships of growth of shell and body (Krakauer, 1974; Usuki, 1979, 1981a, b; Willan, 1979), and of growth of various organs (including shell and body) (Peretz & Adkins, 1982). By culturing A. californica to various ages, then dissecting out and measuring the maximum dimension of their shells, Peretz & Adkins (1982) have shown a precise relationship between shell size and body weight and thus, in these cultured animals, between shell size and age. Whether this relationship holds true for Aplysia under more varied conditions in the field has not yet been tested. A characteristic of growth noted for A. kurodai by Nishiwaki, Ueda & Makioka (1975) was a remarkable day-to-day fluctuation in live weight. The authors recorded daily variations of up to 40% in live weight in four individuals of A. kurodai over a two-month period in the laboratory. In some instances losses could be correlated with production of spawn, but these would have accounted for no more than about 13% of the major fluctuations noted. Similar patterns have not been observed, or at least not to the same extent, in other species of Aplysia. One possibility not considered by Nishiwaki et al. (1975) is that the fluctuations may have been a phenomenon associated with senescence, since growth measurements on these four individuals commenced at almost exactly the time of general seasonal decline in size and vigour of the laboratory population. While such fluctuations in live weight are unlikely to be reflected to the same extent in an animal’s dry weight (Carefoot, 1981b), similar periodic daily monitorings would be advisable in studies of other species that require precise and accurate measures of live weight.

APLYSIA

207

Finally, a number of studies have assessed growth of sea hares in the field. Some of these have estimated growth rates indirectly through collections at intervals and/or size-frequency analyses (Carefoot, 1967c; Audesirk, 1979); others have measured growth rates directly from tagged animals (Nishiwaki et al., 1975; Audesirk, 1979; Willan, 1979), from animals in field cages (Sarver, 1978), and from monitoring of distinctively coloured and therefore easily recognizable animals “seeded” into other field populations (Sarver, 1978). In all instances where field data have been compared with laboratory data on growth at similar temperatures, good agreement has resulted (Carefoot, 1967c; Sarver, 1979; Willan, 1979). The pattern of growth in field animals is illustrated by the results of Audesirk’s (1979) study of tagged A. californica at Santa Catalina Island. Settlement occurs in late summer or autumn with slow growth over the winter. Fastest growth is between February and April, prior to the onset of breeding (at this time a doubling in size, from mean live weights of 700 to 1400 g, was recorded by Audesirk during a single month in spring —one specimen gaining 900 g between February and March, and a further 1200 g the following month). Growth slows in late spring as breeding intensifies, and maximum weights of about 3 kg are reached in June or July. Weights decline during the most active time of breeding in August, and the population dies off in October–December. Audesirk (1979) attributes the seasonal weight loss to two main causes: to loss of foraging time due to reproductive activities, and to the large energy requirement for egg production. With some variation, similar patterns of seasonal growth have been noted in other field studies of Aplysia (e.g. A. juliana: Sarver, 1978; A. kurodai: Nishiwaki et al., 1975; A. punctata : Carefoot, 1967c). An apparent exception to this pattern of growth was noted by Krakauer (1969) for A. brasiliana=willcoxi, where the population appeared to show no seasonal decline in weight, but this may have resulted from the author terminating her field collections in late spring—possibly too early to have shown a decline in weight. ENERGY ALLOCATIONS All animals require a mechanism to control energy balance and to allocate food energy or stored energy to various metabolic pathways. In most instances regulation of energy balance is by controlling energy input; only in instances of starvation and aestivation is energy balance regulated by controlling energy output. In Aplysia, as in any animal, allocation of food energy C is to production of somatic tissue Ps and gametes Pr. Initially, all energy allocated to production is devoted to growth of somatic tissue; the first ‘decision’ comes at reproductive maturity with the shunting of some growth energy into production of eggs and sperm. Later, all energy of production save for that required for tissue maintenance is allocated to reproduction. The sum of Ps and Pr, representing an animal’s “scope for growth” (Warren & Davis, 1967), has been used by various authors as an indication of diet quality and effect of “stress” in marine invertebrates, including a few gastropods (e.g. Stickle, 1985). As a by-product of assimilation of nutrient materials into new tissue and of various maintenance costs, energy is lost in the form of urinary excretions U and as heat of respiration R. After absorption, unused food energy is egested as faeces F. The assembled energy budget is represented as the familiar equation:

Not included in energy budget models for sea hares are a number of special investments such as: (1) the shell, which represents about 0·1% of the total live weight of the animal (Peretz & Adkins, 1982; to obtain overall energy values for Ps, the shell is routinely ground up and combusted with the other body tissues, and thus actually contributes a small energy loss through endothermy: Paine, 1966); (2) opaline and ink-gland secretions, when they are lost from the body; (3) radular teeth broken off and lost during feeding, or consumed

208

THOMAS H.CAREFOOT

and lost along with the faeces; (4) mucus lost as sheathing of the faeces, and from the skin especially during locomotion (see Denny, 1980); and (5) sperm, following copulation, except in instances of simultaneous copulation when incoming sperm would ‘compensate’ to some extent for the energy loss represented by the animal’s own outgoing sperm. Of these costs, all save mucus lost during locomotion would be predicted to be small. None, however, has been reliably estimated in studies of energetics of sea hares. Energy budgets of sea hares are shown in Table XI. Several features are evident from these models. First, the energy lost in urinary secretions U, has never been determined for Aplysia. Its omission in studies of energy budgets of gastropods is usually thought to introduce only a small error. Where it has been measured in gastropods, as for example in the freshwater snail Hydrobia ventrosa by Kofoed (1975), it has, however, been shown to account for 11–19% of consumed food energy (this value also includes mucus and dissolved organic carbon losses; and it may not reflect accurately the extent of loss under the different osmotic conditions of a sea-water habitat). Secondly, where each component of the budget has been separately determined, large imbalances in the budgets are apparent (e.g. −3 to +155% for Aplysia punctata: Carefoot, 1967a, b; −46% for A. juliana: Sarver, 1978). Negative imbalances can result from failure to account for urinary and other losses such as mucus and dissolved organic molecules as previously noted, but can also result from: (1) over-estimating C during feeding by not taking into account possible loss of food materials such as plant juices or seaweed fragments, (2) under-estimating F due to loss or dissolution of, or leaching from, the faeces, or (rarely) to coprophagy (representing a double source of error), and (3) under-estimating R through “flask-effects” during respirometry measurements (Wightman, 1981). These and other errors in budgeting are discussed in more detail elsewhere (Carefoot, in press). The magnitude of one of these potential sources of error, namely the possible loss of food materials through ‘sloppy’ eating habits, was estimated in a field study of A. dactylomela using nylon-mesh bags to enclose single animals in overnight feeding experiments (Carefoot, 1985). The results showed that such where W=wastage and C=consumption, losses never exceeded 8% (calculated as expressed in Joules) in all feeding trials in the field. This was considerably less than expected in view of the rough manner of biting and tearing by which sea hares graze, which tends to fragment extensively branched or delicate algae. How much of the negative budgetary imbalances would be accounted for by such wastage in laboratory studies would depend on the degree to which uneaten remnants of food were reclaimed by the experimentor after a feeding session by Aplysia and the extent to which food bits were allowed to escape in flow-through systems. Even though such wastage probably accounts for only a small part of the missing energy in budgetary models, it does represent a possibly significant contribution to detrital food chains in areas populated by Aplysia. Allocation of ingested food energy C to somatic Ps and spawn Pr production, and to respiration R was monitored in A. juliana by Sarver TABLE XI Energy budgets for Aplysia: C=Chlorophyceae; F=faeces; P=Phaeophyceae; Pr=spawn production; Ps=somatic growth; R=Rhodophyceae; U=urine; C=consumption; R=respiration Species

Laboratory dactylomel a

R

Food

C

=Ps +Pr +R

+F +U Balance “Scope for growth” (Ps+Pr)

Remarks

References

Laurencia spp.

100

21

21

No spawn produced by these

Willan, 1979

0

41

–

−17

21

APLYSIA

Species

Food

C

=Ps +Pr +R

+F +U Balance “Scope for growth” (Ps+Pr)

juliana

C

Ulva lactuca

100

5

11

11

27

–

−46

16

punctata

R

Plocamium cartilagine um

100

14

9

16

35

−

−26

23

C

Enteromor pha intestinalis Ulva lactuca Heterosiph onia plumosa Cryptopleu ra ramosa Delesseria sanguinea Laminaria digital a Plocamium cartilagine um

100

9

9

13

41

–

−28

18

100

10

7

19

25

–

−39

17

100

10

5

16

29

–

−40

15

100

8

3

25

29

–

−35

11

100

17

10

54

55

–

+36

27

100

19

4

185

47

–

+155

23

C R

C R P punctata

R

100

22

4

14

33

–

−27

26

Field dactylomel a

Mixed red algae 100

25

6

20

49

–

0

31

Remarks

juvenile animals F estimated from a separate experiment R for animals on Plocamium diet applied to animals on other diets

209

References

Sarver, 1978

Carefoot, 1967a

Carefoot, 1967b

R estimated by difference to get perfect balance

Carefoot, 1985

(1978). These data, shown in simplified form in Figure 7, were derived from animals grown in laboratory culture from soon after metamorphosis (1 mg live weight) to almost end of life (220 g). Sarver found that at Day 5 in culture the animals were allocating about 40% of energy of C to somatic growth. This diminished to 0% after Day 60 and remained at this level to the end of the study. The first allocation of energy to spawn production was recorded at Day 17. This increased to a maximum of about 14% of C during Days 50 to 60, then diminished slowly thereafter to a level of about 10% at Day 103. Two points of major interest emerge from these data. First, when the animals initially reach reproductive maturity the two allocations to somatic

210

THOMAS H.CAREFOOT

and spawn production did not appear to ‘compete’ instantly for common energy to an extent that Ps was shut down completely. Rather, reproduction only gradually replaced somatic growth, and for about one-third of this species’ normal life span, both functions occurred together (over Days 20–50). At the same time, it should be noted that the smoothed curves in Figure 6 (simplified for the present paper) obscure an important feature noted by Sarver. This was that peaks of egg production by A. juliana were always correlated with low points for somatic growth, indicating expected short-term energy trade-offs. Secondly, diminished over the life of the animal (from 42–12%), possibly correlated with “scope for growth” an increase in percentage allocation of energy to R during the animal’s life to a maximum of about 10% of C on Day 80 (Fig. 6). The reason for this increase in R, as explained by Sarver, was that food consumption measured as a percentage of body weight actually decreased steadily and uniformly over the animal’s life, while the animal itself continued to grow larger. Only at about Days 60 to 70, when Ps became zero and allocation to Pr began to decline, did energy allocation to R begin to level off. The effect of dietary rationing on the pattern of energy allocation in sea hares was also investigated by Sarver (1978). He found that reproductively mature A. juliana maintained on low rations “favoured” egg production over weight gain, whereas on intermediate rations both somatic growth and egg production were supported about equally, and at ration levels approaching ad libitum a higher percentage of ingested energy was again allocated to reproduction. Energy flow in a coral rubble-inhabiting population of A. dactylomela in Barbados is shown in Figure 8. The data represent two groups of animals feeding in two habitats: an area dominated by Cladophora sp., and situated near the wave-break part of the sublittoral habitat; and the other, an area dominated by the red algae Gracilaria sp. and Laurencia papillosa and located a few metres offshore in slightly deeper water. The animals normally feed at night in the Cladophora area and rest during the day in the GracilariaLaurencia area, where crevices for hiding are more plentiful. These data on energy flow, representing combined field and laboratory studies (Carefoot, 1970, 1985), show that the red alga-dominated area is more energy-rich, provides more total food energy, and ultimately yields more energy represented by production than does the green alga-dominated area (by about 40%). Therefore, it is puzzling that selection favours a behaviour involving a feeding excursion each night to the shallow-water area with a return at dawn to the deeper-water resting areas. Any change in behaviour costs energy, and it would be presumed that the new activity would have to yield some advantage, either energetically, nutritionally, or whatever, to make the change worthwhile (cf. Larkin & McFarland, 1978). The ingestion of sand, as an aid to digestion, and availability of daytime hiding places, are factors that have already been considered with respect to this feeding excursion. It may be that the animals simply find Cladophora more palatable than the red seaweeds, even though Laurencia spp. are known to be preferred foods in populations of Aplysia dactylomela in other geographical areas (Morton & Miller, 1968; Switzer-Dunlap & Hadfield, 1979; Willan, 1979). In an earlier study of food choice in a field population of A. dactylomela in the same area in Barbados, Cladophora sp. was found to represent 60–70% of the foodstuffs in the dissected crops (Carefoot, 1970). LOCOMOTION Crawling is the major mode of locomotion in Aplysia, although a few species burrow and swim. The characteristic, slow movements during browsing alternate with periods of hunched immobility. Swimming has been studied most extensively in A. brasiliana=willcoxi, where it functions at least partly to enable animals to move out of shallow areas where they may be stranded by the tides, and involves a degree of navigation through celestial cueing.

APLYSIA

211

Fig. 7—Allocation of consumed food energy (C) to somatic production (Ps), spawn production (Pr), and respiration (R) in A. juliana: the animals were fed ad libitum on Ulva lactuca; “scope for growth” represents data from Sarver (1979).

CRAWLING Crawling is accomplished by a combination of muscular movements and hydraulic extensions of the anterior part of the body, the latter being involved more at high speeds. At low speeds the propulsive force is strictly by muscular waves which move from front to back along the animal’s foot (a monotaxic retrograde pattern; but see Pilsbry, 1951), one or two waves passing along it at any given time (Parker, 1917; Bebbington & Hughes, 1973). The waves lift part of the foot locally and temporarily from the substratum, enabling it to move forward while the rest anchors the animal in place (Parker, 1917). Muscles move blood into and out of lacunar spaces in the foot allowing contraction and expansion of various parts (Bebbington & Hughes, 1973). In addition to the longitudinal wave of contraction, Hening, Walters, Carew & Kandel (1979) describe a transverse contraction involving part of the body wall as well as the foot-sole which tends to constrict and narrow the foot. At higher rates of locomotion the anterior part of the body is extended by muscular antagonism against a hydrostatic skeleton, created by blood in various haemocoelic spaces. The anterior part of the foot-sole attaches to the substratum by a combination of mucous adhesion and suction, and the posterior part of the body is drawn up by muscular contraction. This produces the humping or “inchworm” style of locomotion so characteristic of a crawling sea hare. At highest speeds, as for example during escape behaviour or after full arousal following various noxious stimuli (see Wachtel & Impelman, 1973), the anterior body extends maximally and locomotion becomes a two-phase pattern of anterior extension and attachment, followed by release of the tail and pulling up of the body (Jahan-Parwar & Fredman, 1979a). In this “galloping” mode of locomotion, the body arches quite markedly. During vigorous locomotion the parapodial flaps may open and close and the mantle shelf and siphon may contract, producing a coordinated ventilatory pumping (Hening et al., 1979). Observation of such behaviour has led to the study of

212

THOMAS H.CAREFOOT

Fig. 8—Comparison of daily energy budgets for A. dactylomela feeding in two seaweed communities in Barbados: all values not otherwise indicated are in J·live g−1·day−1 for animals 60 g in live weight; the data show the ‘standing energy’ in a Cladophora- and a Gracilaria/Laurencia-dominated algal community, how much is eaten and wasted, and the allocation of consumed energy to various body processes in A. dactylomela; data from Carefoot (1970, 1985).

neural control of locomotion, initially by Jordan (1901) in A. fasciata and, more recently, by Hening et al. (1979) and Jahan-Parwar & Fredman (1978a, b, c, 1979a, b, 1980) in A. californica. Locomotory rates in some of the larger species, such as A. dactylomela, range from 10 cm·min−1 during slow crawling, to 50–150 cm·min−1 at full gallop. Strumwasser (1967) recorded a mean velocity of 1 m·h−1 in A. californica over a 12-h period of normal daytime activity, while Jacklet (1972) and Hening et al. (1979) recorded maximum velocities of 10–18 m·h−1 in the same species. Average distances moved by A. californica in laboratory studies were 35 m in 12 h (in a cycle of 12 h light: 12 h dark; Kupfermann, 1968), and 35–40 m in 24 h (in 24 h constant light; Jacklet, 1972). Kupfermann (1965) recorded movements of up to 200 m·day−1 in A. californica in the laboratory, while Willan (1979) measured normal crawling rates of 1–5 m·h−1 in A. dactylomela, with no observable correlation between size and rate. Aplysia rarely travel far in the field. Kupfermann & Carew (1974) recorded daily movements of tagged A. californica in the field of up to 100 m, although when they followed a single individual slowly grazing for one hour on its seaweed foods, it moved a distance of only 6·3 m. In studies of movements of tagged A. californica in sites at Santa Catalina Island, Audesirk (1979) noted that nine animals spent at least 4–5 months in an area not larger than 5600 m2. In comparison, tagged A. dactylomela in Barbados moved no more than 10–20 m·day−1, and many individuals remained several days in an area of about 100 m2 (Carefoot, pers. obs.). For a species like A. dactylomela which inhabits daytime hiding places, the question arises as to whether there is any fidelity to specific home-sites after nightly feeding excursions. Monitoring of movements of tagged individuals and observing known resting spots from day to day has shown none; the animals just seem to select crevices and under-rock hiding sites opportunistically as dusk nears (Carefoot, pers. obs.). Crawling can be elicited in Aplysia by several stimuli including: (1) food (Frings & Frings, 1965; Preston & Lee, 1973; Jahan-Parwar, 1972a; Audesirk, 1975a, b; Jahan-Parwar & Fredman, 1979a), (2) pheromones (Jahan-Parwar, 1976; Audesirk, 1977; Audesirk & Audesirk, 1977), and (3) various noxious stimuli (salt,

APLYSIA

213

heat, pinching, electrical shock: Wachtel & Impelman, 1973; Walters, Carew & Kandel, 1978; salt: JahanParwar & Fredman, 1979a, b). In addition, light intensity appears to affect locomotory rates. Jacklet (1972) showed in one experiment that A. californica subjected to 40 lux intensity of light over a continuous 24-h period moved a mean distance of 39 m, while those subjected to 280 lux moved a mean distance of 60m. A generous endowment of mucous glands in combination with the musculature of the foot provides Aplysia with the ability for suction attachment over the full area of the foot (Parker, 1917), but in most species development of anterior and posterior grouped mucous glands enables the animal to cling better by its front and back portions than by the remaining surface (Engel & Eales, 1957). Some species, as for example A. dura and A. juliana, even possess a sucking disc on the posterior part of the foot. This enables the animal to cling especially well by the posterior portion of the foot while raising the anterior part of the body from the substratum. Sea hares are adept at gripping with the anterior part of the foot. This is noticeable during copulation in some species (e.g. A. dactylomela) where a small pinch of the sperm-recipient’s skin may be gripped tightly by its partner. It is also evident during feeding, where an animal may use the foot to hold seaweeds, or during dislodgment, where re-attachment may be facilitated through gripping or by suction action of the anterior part of the foot. Kupfermann & Carew (1974) noted in A. californica what they described as particularly strong adhesive properties of mucous “glue” secreted by animals dislodged in strong surge conditions. Unless this species produces mucus of varying adhesive properties, as for example depending on the degree of water movement, it is likely that the observers’ fingers were stuck not be a mucous glue as described, but simply by suction generated by prehensile portions of the foot. Under normal circumstances the foot mucus in sea hares is not at all sticky. A circadian rhythm of locomotory activity is present in A. californica (Lu, Strumwasser & Gilliam, 1966; Kupfermann, 1968; Jacklet, 1972, 1974; Block & Roberts, 1981), a species that is normally active during daylight hours. When kept in constant darkness it will, however, maintain an abbreviated form of the same activity cycle for at least 48 h (Kupfermann, 1965, 1968). A remarkable pacemaker system in the eyes is thought to be involved in the circadian rhythm of locomotion. Considerable research has been done on this pacemaker system with respect to: (1) the circadian activity of the ocular pacemaker system and its entrainment by light cycles (Jacklet, 1969, 1974; Eskin, 1979; Current, Eskin & Kay, 1982), (2) its proposed relationship to locomotory activity (Jacklet, 1972, 1976), (3) the photoreceptors responsible for the entrainment to light, both within the eye itself (Eskin, 1971, 1979) and through possible activity of extraocular photoreceptors (Block, Hudson & Lickey, 1974; Jacklet, 1980), (4) the effect of removing the eyes or cutting the optic nerve on locomotion and other behavioural activities (Lickey et al., 1977; Lickey & Wozniak, 1979), (5) the possible modes of synchronization of the various pacemaker systems (Lickey, Hudson & Hiaasen, 1983; Jordan, Lickey & Hiaasen, 1985), and (6) the fine structure of the eye itself (Jacklet, Alvarez & Bernstein, 1972). There are actually three circadian rhythms in A. californica: in the abdominal ganglion (Lickey, 1969; see also Strumwasser, 1973); in the eye; and locomotory. The current view is that the circadian pacemaker in the eye is at least one of the controls for locomotory rhythm (Jacklet, 1972; Strumwasser, 1973; Lickey et al., 1977; Block & Roberts, 1981; Lickey et al., 1983), yet some extraocular regulation must be involved because eyeless animals maintain a normal diurnal pattern of activity for some time (Block, 1971; Block et al., 1974; Lickey et al., 1977; Lickey & Wozniak, 1979), although with deteriorated fidelity (Strumwasser, 1973; Lickey et al., 1977). The extraocular photoreceptors could be in the central neurons (Arvanitaki & Chalazonitis, 1961; Block & Smith, 1973; Brown & Brown, 1973), or in the siphon, rhinophore, or other areas of the skin (Lukowiak & Jacklet, 1972a; Chase, 1979b; Jacklet, 1980). Lickey et al. (1977) put

214

THOMAS H.CAREFOOT

forward the possibility, although admitted remote even by the authors, that the extraocular oscillator which drives the system in eyeless A. californica may reside in some symbiotic or parasitic organism (e.g. crabs). Assuming that the eye and optic nerve behave in the same way in vivo as in isolated preparations in the laboratory, a normal day for A. californica begins about an hour before dawn with a burst of impulses from the optic nerve, with frequency peaking just after mid-day, then attenuating through the remainder of the day (Jacklet, 1969). Locomotory activity is never completely absent during the night, but similarly shows a pronounced burst at dawn and reaches a peak in intensity by mid-afternoon, By dusk, or shortly thereafter, the animals are mostly quiescent (Kupfermann, 1968; Jacklet, 1972), and the impulse frequency from the optic nerves is diminishing at this time (Jacklet, 1969). Just as the locomotory rhythm maintains its circadian pattern in animals in total darkness, so the discharge frequency of the optic nerve retains its circadian rhythm when isolated eyes are kept in darkness. Interestingly, the free-running cycle for isolated eyes in vitro in darkness is about 26 h, with some variation, and can be maintained for a week or more in culture (Jacklet, 1974). No work has been done on a nocturnally active species, such as A. dactylomela or A. juliana, to test whether the supposed ocular or other pacemaker systems operate in these species on a 12-h phase shift. BURROWING Burrowing in sea hares may function as a mechanism to avoid light (specifically ultraviolet radiation), wave action, or intertidal exposure (see p. 173). In A. brasiliana=willcoxi, an habitual burrower, burrowing is accomplished in two phases (Aspey & Blankenship, 1975, 1976a). In the first phase, a sequence of shovelling movements with the head and oral tentacles serves to bury the front part of the body. In the second phase, following immediately after the first, a series of swelling and forward heaving movements of the entire body ensues, easing the animal into the substratum until it is fully covered save for the rhinophore tips, the opening to the mantle cavity area, and the siphon. An interesting variant on this pattern has been described for A. geographica (proposed as a new species: formerly Siphonota geographica: Willan, pers. comm.). In this animal, digging is accomplished by muscular (hydraulic?) thrusts of the head. First, the head is turned vertically and thrust into the substratum, and then is simultaneously dilated and returned to the horizontal plane, thereby parting the sand and creating a deeper hole. The oral tentacles act as wedges in the first phase and as lateral ploughs in the second, while the animal rocks back and forth throughout. Eventually, all that is visible of A. geographica are the tips of its rhinophores, the centre field of its parapodia, and its anal siphon. The animal can draw clean water for gas exchange into the mantle area through apertures created by the anterior edges of the parapodia and expel it though the siphon. Burrowing has been studied in A. brasiliana=willcoxi in relation to weight and general condition of the animal, time taken to burrow, time spent under the substratum and degree of subsequent coverage, latency of re-burrowing, degree of coverage when buried, orientation during burrowing, vigour of burrowing, propensity to ink when removed from the burrow, and behaviour following emergence (Aspey & Blankenship, 1976a). Among other things, these authors found that smallest animals burrow the fastest, are most reponsive to disturbance when they are burrowing, and are most likely to engage in reproductive activities following emergence. Animals remain under the substratum for periods ranging from one hour to several weeks. Further studies showed that buried animals induce burrowing in swimming conspecifics in aquaria, presumably through a pheromonal mediator (Aspey & Blankenship, 1976b). The emergence of one buried group member from its burrow causes other animals to follow, and mass copulation often ensues (Aspey & Blankenship, 1975). The authors proposed that burrowing in A. brasiliana=willcoxi, at least in small individuals, serves as some form of preparation for subsequent reproductive activity. In this respect, the

APLYSIA

215

situation resembles the tendency of A. dactylomela to undertake copulation following emergence from burrows and crevices in which they hide overnight (Carefoot, 1985). SWIMMING Seven species of Aplysia are known to swim (A. brasiliana=willcoxi, A. depilans, A. extraordinaria, A. fasciata, A. morio, A. pulmonica, and A. tanzanensis), and another two may (A. maculata and A. winneba; Eales, 1960). The observations by Allan (1941), MacNae (1955, 1957), Eales (1960), and Kay (1964) that A. dactylomela can swim, by Allan (1941) that A. nigra and A. parvula can swim, by Eales (1921) and Haefelfinger & Kress (1967) that A. punctata can swim, and by Kandel (1976, 1979) that A. juliana can swim, do not appear to be supported by fact. Sea hares swim by a similar parapodial flapping in all species, but with possibly different mechanisms of propulsion through the water (Neu, 1932; Pruvot-Fol, 1954; Farmer, 1970; Bebbington & Hughes, 1973; von der Porten et al., 1980, 1982). The propulsive force is a wave of muscle contraction passing from front to back along each parapodium, beginning in A. fasciata in the right parapodium slightly in advance of the left, and commencing from a “starting” position in which the flaps are folded, right over left, to cover the mantle area (Bebbington & Hughes, 1973). Parapodial flaps are fully extended momentarily at the end of this part of the propulsive stroke. The “recovery” stroke begins at the front of each parapodium and, through a wave of contraction moving from front to back, serves to fold the flaps once more over the mantle cavity. The way in which this propulsive wave translates into a swimming motion in Aplysia is unclear. Sculling and jet propulsion are the most popular theories (the latter proposed initially by Neu, 1932, and later reconsidered by Kandel, 1979), yet appear too simplistic to explain movement in A. brasiliana=willcoxi. Rather, von der Porten et al. (1982) suggest that in this species the thicker leading edge of each parapodium presents an “airfoil” which generates “lift”. This lift force is thought to be produced on both downstroke (extension) and upstroke (flexion) phases of the flapping cycle of the parapodia, and thus would explain the smooth forward progression of sea hares as they flap along, rather than the cyclical jerking which would accompany either sculling or jet propulsion. The authors propose that thrust is generated by the anterior onethird of each parapodium, while the posterior two-thirds may operate for attitude control. This theory offers an intriguing and novel approach to the study of swimming in sea hares and is bound to generate fresh interest in the subject. Neuronal control of swimming has been studied by Jahan-Parwar & Miller (1978), Pinsker et al. (1978), von der Porten et al. (1980, 1982), and others. There appears to be a neuronal oscillator in each pedal ganglion that regulates the flapping frequency. Through lesion studies, the activation “command” for swimming has been found to arise in the cerebral ganglion (Jahan-Parwar & Miller, 1978; von der Porten et al., 1980, 1982). A complete flapping cycle, including both extension and flexion, in an A. fasciata of 20-cm length takes about three seconds (temperature unspecified: Bebbington & Hughes, 1973) and carries the animal about one body length. This rate compares favourably with the rate of one cycle·2·6 s−1 recorded for A. brasiliana=willcoxi by von der Porten et al. (1982) at a temperature of 13·5°C, probably similar to the temperature used in the previous study on A. fasciata. The authors showed a marked effect of temperature on flapping frequency and swimming speed in A. brasiliana=willcoxi (0·4 beats·s−1 gives a swimming speed of 4·9 cm·s−1 at 13·5°C, while 0·6 beats·s−1, gives a speed of 6·1 cm·s−1 at 18°C). Interestingly, while temperature affects the frequency of parapodial flapping and swimming speed, it does not appear to affect rate of progression of the wave of contraction along the parapodia (von der Porten et al., 1980). Although Aplysia regulates its swimming speed (as in a current: Hamilton & Ambrose, 1975), the mechanism by

216

THOMAS H.CAREFOOT

which this is done is not known. Since the frequency of flapping does not, however, seem to vary in A. brasiliana=willcoxi when swimming at different levels of effort in currents, it may be that flap amplitude can be regulated (Hamilton & Ambrose, 1975). Most of our knowledge of swimming in sea hares comes from the work of Hamilton and colleagues on A. brasiliana=willcoxi. This species lives in intertidal and shallow subtidal areas, often in seagrass meadows (Krakauer, 1969) from the Gulf of Mexico to Martha’s Vineyard in the northeastern U.S.A. Animals often swim at the ocean surface, especially during trips of more than 4 m (Hamilton & Ambrose, 1975), where they may exhibit the peculiar “head-bobbing” behaviour described by Aspey, Cobbs & Blankenship (1977; see pp. 229–231) and, by being in the upper part of the water column, they become prone to stranding themselves on the shore (Hamilton & Ambrose, 1975). The species can swim continuously for almost two hours in a single bout and may swim for distances of up to one kilometre (Hamilton, 1985). When released in shallow water the animals generally swim in a direct line offshore. In one release of 20 sea hares in a shallow lagoon area in southwestern Florida, Hamilton (1986) recorded a median swimming duration of 9·9 min, a median distance covered of 52 m, and a median “ocean floor” speed of 5·3 m·min−1. One energetic individual swam continuously for 114 min and travelled a distance of 953 m. In comparison, another actively swimming species, A. fasciata, was observed by Susswein et al. (1984a) to swim only in calm water and then only for short periods (in the laboratory about 50% of all swimming bouts lasted for 30 s or less; one unusual animal in the field swam for 35 min.). A. brasiliana=willcoxi is able to modulate its swimming effort according to the direction it is swimming relative to current direction (Hamilton & Ambrose, 1975; Hamilton, 1984). Swimming in a current may, however, be excessively energy-demanding. For example, individuals released in an area of strong current swam for less than half the time of ones released in an area of weak current (7·8 and 21·0 min, respectively; Hamilton, 1986). The eyes of Aplysia are tiny and simple. Each has a relatively large lens and has about 7000 receptor cells in the retina (Jacklet et al., 1972). They may be sensitive to ultraviolet light (Waser, 1968). While the eyes do not appear to have image-resolving capability, Hamilton (1986a) in fact thinks that A. brasiliana=willcoxi may be able to see objects above the water’s surface. Both swimming at the surface and the “head-bobbing” behaviour described by Aspey et al. (1977) are thought by Hamilton & Russell (1982a) to aid in celestial navigation. In their studies of swimming in A. brasiliana=willcoxi, Hamilton (1979) and Hamilton & Russell (1982a, b) found that straight-line navigation is mediated through a combination of orientation to waves and celestial cueing. When animals are released from points near the shore, they swim offshore initially towards waves. This behaviour is thought to be mediated through the rhinophores, since their surgical removal brings about rapid disorientation. Since animals move in an offshore direction even in the absence of waves, however, some perception of horizon (e.g. the tree-line) may be involved in setting the initial direction (Hamilton & Russell, 1982b). For animals swimming without visual reference points on the horizon, celestial cues are thought to be used. One of the points of evidence in favour of this idea is that animals swimming under a translucent white cover become quickly disorientated (Hamilton & Russell, 1982a) suggesting that an unblocked view of the sky is important for navigation. The cues involved in celestial navigation are presumably the sun and possibly polarized light, but neither these nor the integrating mechanism involved in navigation are known. Blinded animals, however, become disorientated when swimming, so it is thought that an actual visual detection of celestial cues is responsible (Hamilton & Russell, 1982a, b). The statocyst is probably also important in the orientation of swimming Aplysia. Its structure and neural connections to the cerebral ganglion have been described by Coggeshall (1969), Dijkgraaf & Hessels (1969), Wolff (1973), Gallin & Wiederhold (1977), and Janse (1983). If the nerve from one of the paired statocysts to the cerebral ganglion is severed unilaterally there is no effect on posture or movement;

APLYSIA

217

bilateral severing of the statocyst nerves, however, markedly affects posture and causes somersaulting in swimming A. fasciata (Dijkgraaf & Hessels, 1969). The simply structured statocysts (containing only 13 cells each in A. californica and A. fasciata: Coggeshall, 1969; Dijkgraaf & Hessels, 1969) only provide information on gravity perception and not on rotation or on acceleration or deceleration (Dijkgraaf & Hessels, 1969; Wolff, 1973). Neither the function of swimming nor the need to navigate are well understood in Aplysia. In A. brasiliana=willcoxi, swimming has been implicated in reproduction (Hamilton et al., 1982), food-finding (Aspey et al., 1977), as well as in moving between eelgrass beds in search of mates, food, or more suitable habitats (Hamilton, 1985, 1986). Susswein (1984) reports that the urge to swim intensifies in A. fasciata just after dusk whether food is present or absent, but that in this noctural species swimming remains at a high level through the night only when food is absent. As A. brasiliana=willcoxi usually swim offshore from shallow inshore release points, strandings could be minimized by this behaviour (Hamilton & Russell, 1982b). A point in favour of this idea is that A. brasiliana=willcoxi are stranded more often on beaches after night-time high tides than after daytime high tides (Hamilton & Russell, 1982b; Hamilton et al., 1982; Krakauer, 1969), but whether this means that this nocturnal species just swims more at night is not known. No studies appear to have been done on the navigational abilities of A. brasiliana=willcoxi during night-time bouts of swimming. PREDATORS AND DEFENCE Aplysia has few predators. Its possible defences include ink, opaline secretions, and toxins in the digestive gland. In addition to the distinctive smell of the opaline secretions, the body often has a fruity odour, which in some way may be related to defence. As members of a molluscan group in which common or popular names are rare, sea hares are distinguished in areas of Brazil as “inkwells”, and in the Gulf of Mexico as “inkfish”, through their propensity to release clouds of reddish-coloured ink when disturbed. Interest in the chemical defences of sea hares and in their ability to sequester secondary metabolites from their foods has led to a vast outpouring of work by ‘natural products’ chemists. PREDATORS No animals are known to prey solely on sea hares, nor are ones known for which sea hares account for even a regular portion of the diet. A number of examples of predation have been reported, and are listed in Table XII, but these appear to be ‘low intensity’ interactions: examples of opportunism, or otherwise sporadic encounters leading to predation. Of these examples, predation on the eggs and juveniles predominate (e.g. MacGinitie, 1934; Winkler & Tilton, 1962; Sarver, 1979; Willan, 1979). MacGinitie (1934) notes that only after A. californica reach a size of 3–4 mm in length do they become distasteful, and it may be that their defences do not reach full operating status until a minimum size is attained. This may be equivalent to the “size refuge” attained by A. dactylomela, which protects it from predation by the starfish Coscinasterias calamaria in New Zealand (Willan, 1979), and by Aplysia juliana in response to certain carnivores in Hawaii (Sarver, 1979). Predation by the great green sea anemone Anthopleura xanthogrammica on several Aplysia californica (Winkler & Tilton, 1962) was probably a unique occurrence and one unlikely to exert a steady pressure on the population. The reasons for this are twofold. First, in a comprehensive field study on A. californica, Kupfermann & Carew (1974) never observed predation by this species of anemone on A. californica, nor has it been reported by other authors working on this species (e.g. Audesirk, 1979). Secondly, the anemone prefers a habitat of wave-swept surge channels where it feeds opportunistically on

218

THOMAS H.CAREFOOT

food materials washing by, but such channels are not favoured by A. californica (see Table I, p. 174; although it is possible that the two species may meet in tide-pools). There is also a possibility that the sea hares in Winkler & Tilton’s study were dying or already dead before being caught up in the tentacles of the anemones, as the authors did not actually witness their capture. Finally, although no data are available, it is presumed that predation TABLE XII Predators of Aplysia Species

Predator species

Aplysia spp.

Bailer shell: Melo amphora Green sea turtles: Caretta caretta Aeolid nudibranch: Favorinus japonicus Starfish: Asterina Starfish Great green sea anemone: Anthopleura xanthogrammica Cephalaspidean opisthobranch: Navanax inermis Flatworms, nemertines, annelids, crabs, isopods, hermit crabs, fish Starfish: Coscinasterias calamaria Starfish: Patiriella regularis Hermit crab: Dardanus sp., cone shell: Conus pennaceus, crab: Calappa sp., also several wrasses, two species of flatworms

brasiliana =willcoxi californica californica

dactylomela

juliana

Remarks

References

Eats the eggs

Coleman, 1975 Felger & Norris (cited in Fenical, 1975) Kay, 1979

Young individual captured Eats the eggs Only juveniles eaten

Sawaya & Leahy, 1971 MacGinitie, 1934 Winkler & Tilton, 1962

Eats sea hares in both field and laboratory

Paine, 1963

Only juveniles eaten

MacGinitie, 1934

Preys on juveniles

Willan, 1979

Eats the eggs All prey on juveniles

Sarver, 1979

on the veliger larvae of Aplysia by filter-feeding animals such as tube worms, bivalves, some anemones, and fish, is enormous. In this regard, Krakauer (1969) has observed predation on A. brasiliana=willcoxi veligers by an unidentified hydroid and the filter-feeding crab Petrolisthes armatus in the laboratory. Few animals have been seen to attack healthy adult sea hares. Susswein et al. (1984a) described an attack on Aplysia fasciata by an unidentified crab in the field, but were unable to tell if the sea hare had been previously injured. In feeding experiments with fish, Thompson (1960) noted that no live A. punctata were eaten. In tests of palatability of A. dactylomela to various fish, Russell (1966) showed that portions of the foot from otherwise healthy specimens were consistently refused, and Krakauer (1969) and Kinnel et al. (1979) showed that portions of the body of A. brasiliana=willcoxi were unpalatable to various fish and sharks (save for the buccal mass, which was consistently eaten: Kinnel et al., 1979). Similarly, Ambrose, Givens, Chen & Ambrose (1979) and DiMatteo (1981a) reported that various portions of the bodies of A. brasiliana=willcoxi and A. dactylomela were distasteful to laughing gulls, Larus atricilla, a species known as a fairly indiscriminate general scavenger.

APLYSIA

219

Defence mechanisms in sea hares can be categorized as active or passive. Active defences include: (1) escape by crawling, (2) withdrawal of the gill and siphon, and (3) inking and release of opaline secretions. Passive defences include: (1) crypsis, and (2) distastefulness. Another way to separate them, however, is by their behavioural and chemical characteristics, and it is according to these broad distinctions that the defences of sea hares will be treated. CRYPSIS While some authors would disagree (Thompson, 1960; Russell, 1966; Krakauer, 1969), sea hares are not particularly cryptic in their behaviour or in their colour. They tend to be large, bulky in shape and, to the human eye, fairly conspicuous. Their predominant colour is brown or shades of reddish-brown, often with lighter fleckings or patterns of white spots. Against their favoured foods of red and green algae, only a few species seem to be camouflaged. Yet, in opposite view, their apparent distastefulness is not reflected by any sort of flamboyant warning colouration. This led Thompson (1960) to question the concept of adaptive colouration as it applied to animals such as Aplysia which, in his view, are both cryptically coloured and distasteful. In many other opisthobranch molluscs, warning colouration is associated with various defences, such as spicules or acid secretions. Ambrose et al. (1979) have even speculated that the reddish-purple ink of sea hares functions, not as a toxin, but in warning. The function of the ink will be considered more fully in a later section. Batesian mimicry has been proposed as an explanation for the similarity in appearance of post-larval juveniles of the tropical burrfish Chilomycterus antennatus and Aplysia dactylomela (Heck & Weinstein, 1978). The two species occur sympatrically in seagrass meadows along the Caribbean coast of Panama. Heck & Weinstein point out that neither species is apparently eaten by Caribbean fishes, despite a presumed palatability of the juvenile burrfish. Quite apart from the remarkable similarity in colour patterns in the two species, Heck & Weinstein’s theory is supported indirectly by a feature of the fish’s maturation. The burrfish appears to develop unusually rapidly from the juvenile stage, where it is protected by the proposed mimicry, to the relative safety of the inflatable, spiny, adult stage, thereby possibly reducing its period of vulnerability to predators. ESCAPE Sea hares escape from predatory animals by crawling. Willan (1979) noted that A. dactylomela crawl faster on contacting the predatory starfish Coscinasterias calamaria. Kandel (1979) also cited an unpublished observation by Dieringer & Koester that when Aplysia californica was pinched by pedicellariae of the starfish Astrometis sertulifera, it brought about a galloping escape response in the sea hare. DiMatteo (1981b) recorded an increased frequency of withdrawal by Aplysia dactylomela from contact with various predators, including the gastropod Fasciolaria tulipa and the starfish Echinaster sentus, over controls (empty Fasciolaria shells and live specimens of the herbivorous gastropod Strombus gigas). The author noted that Aplysia dactylomela also seemed to flee in response to pinches by the starfish’s pedicellariae. This same species of sea hare also withdrew from the jellyfish Cassiopea xamachana (Lederhendler, Bell & Tobach, 1975; DiMatteo, 1981b), although Lederhendler et al. (1975) were unsure whether the defensive withdrawal was to a chemical released by the jellyfish or to its nematocysts, or both. DiMatteo (1981b), however, demonstrated that withdrawal was at least partly due to the nematocysts. A common inference in such investigations is the likelihood of an evolved predator-prey interaction between the animals in question. Of course, this cannot be determined from such experiments and, in this

220

THOMAS H.CAREFOOT

regard, it is unfortunate that so little is known at present about sea hares in their natural habitats. Interpretation of the results in such studies has sometimes also been confounded by apparent escape behaviour of the sea hares to non-predatory animals. This was evident in the work of DiMatteo (1981b), where Aplysia dactylomela fled as much from contact with the detritivorous holothuroid Astichopus multifidus, as it did from contact with the predatory starfish Echinaster, and fled two and one-half times more from contact with the holothuroid than from contact with the molluscivorous snail Fasciolaria. Burrowing is probably too slow to be an effective means of escape from would-be predators, and has never been observed as such in sea hares (Aspey & Blankenship, 1976a). Hamilton & Russell (1982b) and Hamilton (1986b) similarly doubt that Aplysia brasiliana=willcoxi swim to escape from predators, although Willan (1979) has observed A. extraordinaria (Willan, pers. comm.) swimming (presumably from some threat), where the first flapping strokes have been accompanied by release of ink. Finally, Dieringer, Koester & Weiss (1978) recorded an increase in rate of heartbeat of up to 67% above baseline level in A. californica actively escaping from the apparently noxious stimulus of salt being applied to its parapodia. INKING Interest in the toxicity of Aplysia dates back to ancient times. Dioscorides in the first century AD was possibly the first to observe that sea hares were poisonous (Halstead, 1965). The naturalist, Pliny, also in the first century, noted that if taken with food or drink the sea hare may be poisonous to some, while to others its very sight causes death (Bostock & Riley, 1857). Interest specifically in the toxicity of the ink of Aplysia dates from the work of Flury (1915) on A. fasciata. The ink is produced in special glands (called Blochmann’s glands after their discoverer, or ink or purple glands) opening into the mantle cavity on the underside of the free edge of the mantle shelf. All species of sea hares possess ink glands save for A. cedrosensis, A. depilans, A. dura, A. juliana, A. nigra, and A. vaccaria. Early studies on Aplysia ink which refer to A. depilans, as for example that of Schreiber (1932), presumably involve a misidentification, possibly of A. fasciata. No comparisons have been made between species with and without ink glands on any aspect of defensive behaviour. The chemical composition of the ink has been investigated by De-Negri & De-Negri (1876), Moseley (1877), MacMunn (1899), Flury (1915), Schreiber (1932), Fontaine & Raffy (1936a, b), Lederer & Huttrer (1942), Christomanos (1955), Winkler (1959a), Nishibori (1960), Rüdiger (1967a, b, 1968), and Chapman & Fox (1969). In A. fasciata it consists mostly of water and other volatile materials with about 1·6% organic substances and 4·6% minerals (Flury, 1915). The violet pigment portion is bound to a protein (Christomanos, 1955), and has been identified by Rüdiger (1967a, b) and Rüdiger, Carra & hEocha (1967) as a bile pigment (a monomethylester of a biladiene dicarboxylic acid) with the formula C34H40N4O7. Studies by Chapman, Cole & Siegelman (1967) and Chapman & Fox (1969) have shown that the ink is a monomethyl ester of phycoerythrobilin and is derived from the bilin chromophore of phycoerythrin contained in the red algae eaten by the sea hares. These authors further discovered that when A. californica are maintained on a diet of brown algae, which lack the essential phycoerythrin pigments, they become facultatively de-inked. The animals regain the ability to produce ink only when returned to a diet of red algae or to an artificial diet with phycoerythrins added. The ink of Aplysia has been referred to as aplysiopurpurin by MacMunn (1899; who also called it aplysine) and by Nishibori (1960). It has been described as consisting of two component chromoproteins aplysiovioline and aplysiorhodine by Lederer & Huttrer (1942), with the designation by Winkler (1959a) of a third component aplysioazurin (this last being

APLYSIA

221

equivalent to phycocyanobilin: Chapman & Fox, 1969). Finally, the ink has been termed aplysioviolin by Rüdiger (1967a, b) and Chapman & Fox (1969), and is the name which appears to be most commonly used. The release of ink is a high-threshold, all-or-none response (Carew & Kandel, 1977a, b, c; Shapiro, Koester & Byrne, 1979; Byrne, 1981). Rough handling in some species will often, but not always, elicit release of ink, as will separation of copulating animals. Tobach, Gold & Ziegler (1965) obtained inconclusive results in field tests of the factors causing ink to be released in A. brasiliana=willcoxi and A. dactylomela. Only sometimes did the animals release ink when they were handled or pricked with a pin, and there was no consistency in response to the different levels of stimulation. The authors did, however, observe that animals previously found in pairs or in larger groups in the field showed a slight, but significantly greater, propensity to release ink than those which were previously solitary. The significance of this observation is not known. Kupfermann & Carew (1974) never observed release of ink under natural conditions in field A. californica, but they could induce release of ink by separating copulating animals (in which case the sperm-donor frequently inked) or by vigorously hand-stimulating the mantle cavity. Interestingly, in their field studies of A. californica, Carew & Kupfermann (1974) showed a significant difference in the propensity of animals from different habitats to release ink after being stuck with a pin. Thus, sea hares from rough-water habitats were much less likely to release ink than were animals from calmwater habitats. Susswein et al. (1984a) noted that A. fasciata never released ink spontaneously, and would do so only occasionally when separated during mating. Finally, Advokat (1980) demonstrated that recently fed animals were less likely to release ink than were animals that had been starved for 24 hours. It is not known how long it takes for a sea hare to replenish its ink supply after spontaneous release, although an estimate can be made from data given by Chapman & Fox (1969) in their experiments on deinking of A. californica. The authors rubbed the ink gland and adjacent tissues with their fingers daily, and showed that an animal on a bilin-free diet (e.g. the brown alga Egregia laevigata) could be effectively deinked after five days (to about 1% of normal level), and completely de-inked after 14 days. They believe that under normal circumstances some residual ink may always be retained in the ink gland, even under conditions of severe handling. From these results it was estimated that normal animals on a diet of red algae could regenerate their ink in at least two days. Experimental animals that were kept on a bilin-free diet and manually de-inked to facultative exhaustion of the ink supply, then fed on a normal diet of Laurencia sp., required three days to begin releasing new ink (Chapman & Fox, 1969). In some species the ink is released along with considerable mucus (Schreiber, 1932; Sawaya & Leahy, 1971; Willan, 1979; DiMatteo, 1982), which may vary in amount between different animals in a given species, such as A. dactylomela, and in the same animal at different times (Carefoot, pers. obs.). If this is generally true for other species of Aplysia, it will add to the already difficult task of making quantitative assessments of the effect of ink in studies of predator-prey interactions. Release of ink in response to predation has been rarely witnessed. Willan (1979) noted that a field A. dactylomela released its ink while attempting to escape from the starfish Coscinasterias calamaria. The ink had no apparent effect on the starfish. In contrast, ink released by Aplysia parvula in response to attack by the same starfish under laboratory conditions seemed to retard the locomotory movements of the predator, thus allowing the sea hares to escape (Willan, 1979). Susswein et al. (1984a) observed a crab attacking an A. fasciata which did not release its ink, although they noted that the sea hare may have been unhealthy even before the attack. Finally, in a study on the effects of the toxin ‘holothurin’ from the sea cucumber Actinopyga agassizi on Aplysia dactylomela and of the reciprocal effects of the ink from A. dactylomela on the sea cucumber, Lederhendler et al. (1975) found that A. dactylomela released its ink in response to the presence of holothurin in the water. The sea hare’s ink did not, however, appear to affect the sea cucumber in a reciprocal manner.

222

THOMAS H.CAREFOOT

Despite the apparent defensive rôle of ink in sea hares, contradictory results have been obtained in experiments to demonstrate it. For example, Flury (1915) directly injected or added the ink of A. fasciata dropwise to the sea-water medium of a variety of animals including crabs, fishes, and frogs, and showed that it had little or no effect. Krakauer (1969) reported that pieces of shrimp soaked in the ink of A. brasiliana=willcoxi were readily eaten by blue crabs Callinectes sapidus. Similarly, Ambrose et al. (1979) soaked pieces of fish in a mixture of ink, blood, and mucus from Aplysia brasiliana= willcoxi and showed that the mixture was not unpalatable to laughing gulls Larus atricilla. On the basis of these experiments the authors concluded that Aplysia ink was possibly important as a signal of unpalatability to potential predators, rather than for its supposed toxic properties. Contrasting results were, however, obtained by DiMatteo (1981a) using the ink of A. dactylomela in similar tests with laughing gulls. He demonstrated that gulls clearly found the ink unpalatable when it was injected into pieces of fish rather than simply coated on to, or soaked into, the fish. The author suggested that Ambrose et al. (1979) may have added insufficient ink to the gull’s treated food to elicit a rejection. Further studies by DiMatteo (1982) on the possible defensive rôle of Aplysia ink showed that several species of crabs living sympatrically with A. dactylomela in their seagrass habitat (Panopeus herbstii, Mithrax sculptus, Portunus spinimanus, and Callinectes sapidus) were repelled by otherwise edible pieces of fish coated with Aplysia dactylomela ink. Willan (1979) tested the effects of A. dactylomela ink in solution on survival of a blenny Tripterygion varium, a species occupying the same habitat as the sea hare. Preliminary tests showed that ink from a single Aplysia dactylomela (80 g live weight) in three litres of sea water was toxic to the blenny, although the author could not be sure that the mucus released with the ink was not also deleterious to the fish. The blenny appeared to die of respiratory distress. It is surprising that with all the interest shown in the effect of Aplysia ink on various potential predators, no one has seriously investigated the effect on Aplysia of its own ink. In fact, Lederhendler et al. (1975) did test for this in a single A. dactylomela but found no apparent response. There is also no evidence at present that the ink of sea hares acts as a “cryptic odour” (Kittredge, Takahashi, Lindsey & Lasker, 1974) in the manner described for the ink of Octopus by MacGinitie & MacGinitie (1968). On the basis of these and other studies, the following are the possibilities for the function of ink in sea hares: (1) to rid the animal of unwanted bile pigments from biliproteins consumed in its diet (Chapman & Fox, 1969); (2) to function in defence as a “smoke-screen”, thus enabling the sea hare to escape (Eales, 1921; Halstead, 1965; Hyman, 1967), although this may only be effective in tide-pool inhabiting Aplysia (Carew & Kandel, 1977a; Kandel, 1979); (3) to function in defence through its unpalatable qualities (Beeman, 1961; DiMatteo, 1981a, 1982); and (4) to function as a warning to would-be predators of the sea hare’s other toxic properties (Ambrose et al., 1979). The only function the ink certainly fails to fulfil is that of a clothes dye: the pigment is not colour-fast (Sanford, 1922) and turns a dirty brown upon exposure to air and sunlight (Winkler, 1959a). OPALINE SECRETION In comparison with the extensive work done on the ink of Aplysia, little is known of the other secretion from the mantle cavity, that of the opaline glands. Although a defensive function is generally assumed for the opaline secretion, there is only indirect suggestion in the literature to support this. All sea hares produce the secretion in a collection of subepidermal gland cells known as the opaline glands or glands of Bohadsch, located in the mantle floor near the gonopore (Eales, 1960; Hyman, 1967). The milky fluid is described as having a musky, acrid, nauseating, odoriferous, foetid, or otherwise strong smell in various species (Flury, 1915; Ando, 1952; Eales, 1921, 1960; Kay, 1964; Hyman, 1967; Krakauer, 1969; Willan, 1979). Little is

APLYSIA

223

known of the composition of the opaline secretion. The only analysis of it seems to be that done in Flury’s (1915) early studies on A. depilans. He showed that the secretion consisted of 94·8% water and other volatile substances, with 1·6% organic matter, and 3·6% mineral material. There is little information on factors affecting release of these opaline secretions in sea hares. At the biochemical level, Tritt & Byrne (1980, 1982), and Tritt, Zigmond & Byrne (1980) investigated the effects of various neurotransmitter substances on their liberation in A. californica. The authors concluded that dopamine was most probably the substance discharged by the motor neurons of the opaline gland to effect release of the secretions. Almost no information exists on the factors leading to natural liberation of opaline secretions in field animals, although when Carew & Kupfermann (1974) placed a juvenile A. californica into the tentacles of the great green sea anemone, Anthopleura xanthogrammica, the sea hare released its opaline secretions after several minutes of entanglement in the tentacles. The sea hare later broke free of the tentacles and seemed to be unharmed by its encounter. The rôle of opaline secretion in defence is generally inferred from the study of Flury (1915). The author subjected a variety of animals, including coelenterates, annelids, molluscs, arthropods, echinoderms, fish, frogs, and rabbits, to opaline secretions from Aplysia depilans by either injecting them with it or adding it dropwise to their sea-water medium. The test animals manifested various degrees of immotility and paralysis, which often led to death. Flury prepared a distillate of the residue remaining after alcohol extraction of the opaline secretion from A. depilans and concluded that its main toxic or nerve-paralysing property was in a nitrogen-free, water-distillable, volatile oil, which behaved physically and chemically like a terpene. When injected into octopuses, fish, and frogs, the oily substance caused death. Ando (1952) similarly demonstrated a toxic effect of the opaline secretions of A. kurodai on several invertebrates and thought that the toxic component may have come from the animal’s diet of the red alga Laurencia spp. Sure enough, steam distillation of Laurencia nipponica produced an oily fraction containing terpene-like substances which were found by Ando to be toxic to certain invertebrates. These appear to be the only accounts in the literature involving tests of the toxicity of the opaline secretions and thus, indirectly, of their possible defensive rôle in sea hares. SIPHON- AND GILL-WITHDRAWAL Although not involved in protection as such, the siphon- and gill-withdrawal reflex is a characteristic response by sea hares to weak tactile stimulation, such as touch or water currents, to the siphon and mantleshelf area. The siphon is a tubular fold of mantle tissue which normally protrudes from between the parapodia. Its function is to direct exhalant water from the mantle cavity away from the body to minimize contamination of the inhalant stream. Passage over the gill alters the carbon dioxide and oxygen content of the circulating water and at various times the exhalant stream contains excretory and faecal wastes. The siphon or mantle-shelf area responds to a low-threshold of touch, such that the siphon, mantle shelf, and gill contract and withdraw into the mantle cavity according to a graded response that is proportional to the intensity of the stimulus (Carew & Kandel, 1977a). Aplysia exhibit spontaneous siphon-withdrawal (Kupfermann & Kandel, 1969; Pinsker, Kupfermann & Castellucci, 1970), an action which may not be a defensive reflex but, rather, a way that the animal uses to clear the mantle cavity of debris including faecal material. Because of ease of preparation and monitoring, most studies have involved the gill-withdrawal part of the defensive reflex. Only short-term effects can be studied from dissected preparations where the animals have been restrained and the mantle region opened and pinned out. For long-term studies, the siphon-withdrawal response must be monitored in animals that are intact and healthy. The overall response is sensitive, easy to observe and to quantify, and has featured in a number of studies of habituation and

224

THOMAS H.CAREFOOT

sensitization in Aplysia. These studies on Aplysia have contributed greatly to our understanding of the neuronal and chemical mechanisms involved in learning in sea hares and, thus, in animals in general. The first detailed study of the gill-withdrawal response was by Kupfermann & Kandel (1969) who identified in A. californica the neuronal pathways involved in the reflex response. A feature of the system later remarked on by Pinsker et al. (1970) is the relative simplicity of the neuronal circuitry. Only a few synaptic relays are involved, which increases the speed and effectiveness of the reflex and thus fulfils one of the prime requisites of a good defensive response. Later work by these and other workers on the gillwithdrawal reflex in A. californica showed: (1) the existence of short-term habituation and dishabituation responses (Pinsker et al., 1970); (2) a causal relationship of habituation to a decrease in the amplitude of excitatory synaptic potentials produced at the motor neurons of the gill, and dishabituation to an increase of these synaptic potentials (Kupfermann, Castellucci, Pinsker & Kandel, 1970); (3) that true habituation was involved and not simply fatigue of the gill-withdrawal muscles or changes in response of sensory receptors (Kupfermann et al., 1970); (4) that habituation and dishabituation were independent events using the same neuronal pathways but caused by different (opposite) synaptic events (Castellucci, Pinsker, Kupfermann & Kandel, 1970); (5) that both the purple gland and siphon provide independent afferent pathways each capable of eliciting the gill-withdrawal reflex and each capable of becoming habituated independently (Carew, Castellucci & Kandel, 1971); (6) that short-term (a few hours only) habituation and dishabituation can occur in the absence of central nervous system control, apparently involving terminations of central motor neurons in the gill itself (Peretz, 1970; Lukowiak & Jacklet, 1972a, b; Peretz & Howieson, 1973; Peretz, Jacklet & Lukowiak, 1976; Lukowiak, 1977); (7) that the central nervous system interacts with the peripheral system in the gill to mediate the adaptive behaviour of the gill-withdrawal reflex (Lukowiak & Jacklet, 1972a, b; Peretz & Howieson, 1973; Peretz et al., 1976; Lukowiak & Peretz, 1977); (8) that longterm habituation of the response can be induced, lasting for several days (Carew, Pinsker & Kandel, 1972); and (9) that through application of a second stimulus to another part of the body, a long-term sensitization of the siphon-withdrawal reflex can be induced which lasts for up to three weeks following training (Pinsker, Hening, Carew & Kandel, 1973). Pinsker et al. (1973) noted that because sensitization of the withdrawal response involves the presentation of a second stimulus, usually to a different location on the body, it resembles to some extent classical conditioning (one neuronal pathway enhancing activity in another) and, hence, may be the basic adaptive mechanism from which true associative learning evolved. Classical conditioning of the siphon- and gillwithdrawal reflex using a weak tactile stimulus to the siphon (which produces only a weak withdrawal) as the conditioned stimulus, and a strong shock to the tail (which produces a strong withdrawal) as the unconditioned stimulus, was demonstrated in A. californica by Carew, Walters & Kandel (1981) and Carew, Hawkins & Kandel (1983). A similar type of associative learning was also demonstrated in in vitro preparations of the siphon, mantle, gill, and abdominal ganglion of A. californica by Lukowiak & Sahley (1981), in which they used light as the conditioned stimulus and a tapping on the gill as the unconditioned stimulus. In addition, Walters, Carew & Kandel (1979a, b) showed for A. californica that the locomotory escape response could be learned in aversive response to a chemosensory stimulus consisting of exposure to an extract of shrimp (the conditioned stimulus), and an electrical shock to the head (the unconditioned stimulus). The authors later showed that a suite of defensive responses, including withdrawal of the head and siphon, release of ink, and escape locomotion, could be conditioned in A. californica in response to shrimp juice paired with electrical shock (Walters, Carew & Kandel, 1981). In contrast, feeding behaviour was depressed in these animals in response to conditioning. Walters and co-workers likened these combined effects in conditioned sea hares to a functional equivalent of “conditioned fear” seen in mammals and other higher animals—a choice of terminology that to some will seem inappropriate, but is none the less

APLYSIA

225

an apt analogy. In a manner similar to vertebrates, Aplysia learns to associate a stimulus with possible danger and to prepare itself for defensive action. Long-term habituation of the siphon-withdrawal reflex has been shown to last for several weeks in both laboratory (Carew et al., 1972) and field. Carew & Kupfermann (1974) directed jets of sea water from a syringe at the siphon of A. californica and measured the magnitude of the response according to how long the contracted siphon remained hidden within the parapodia following its withdrawal. The authors were able to demonstrate that animals from turbulent habitats, which were regularly exposed to water-mediated tactile stimulation, exhibited a weaker siphon-withdrawal to the water-jet stimulation and habituated significantly faster than did animals from calm-water environments. The latter animals habituated at about the same rate as control animals in the laboratory and demonstrated a ‘brisk’ withdrawal reflex. The study was especially significant in that it was the first to show that habituation learning of this sort can occur as a natural event in a sea hare’s life. A number of factors are known to influence the siphon-withdrawal response, such as food, sexual activity, and age. Contact with food (including feeding) and sexual activity generally have a modulating effect on the response (Advokat, Carew & Kandel, 1976; Advokat, 1980; Lukowiak, 1980; Lukowiak & Freedman, 1983). Similarly, age appears to affect the withdrawal reflex by: (1) increasing the rapidity of habituation (Peretz & Lukowiak, 1975), but at the same time, (2) impairing long-term retention of habituation and preventing acquisition of sensitization of the reflex (Bailey et al., 1983; see also pp. 201–202), and (3) generally suppressing the sensitivity of the response (Peretz & Lukowiak, 1975; Lukowiak, 1979; Lukowiak & Peretz, 1980; Rattan & Peretz, 1981). Overall, young Aplysia seem to be less adaptable in their response by being less able to discriminate between stimuli of varying intensities, by habituating less easily, and by having generally less suppressive control of this behaviour than older animals (Lukowiak, 1977, 1979; Lukowiak & Peretz, 1980). Lukowiak (1980) also noted that the completion of a meal did not suppress the gillwithdrawal response in young A. californica to the extent that it did in old animals. These age-related differences in the siphon-withdrawal reflex may relate to the “size-refuge” idea noted earlier in this section, in which younger animals may have less well-developed (chemical) defensive capabilities than older ones; hence, would be generally more responsive to all types of stimulation. Older animals with their diverse and well-developed armoury of alternative defences can afford to “relax” their sensitivity to various tactile stimuli (Lukowiak, 1980). CHEMICAL DEFENCES The sea hare’s main line of defence appears to reside in the toxic properties of the digestive glands (also called midgut glands, hepatopancreas, or ‘liver’). The first indication in recent times of this toxicity came with the discovery by Winkler & Tilton (1962) that the digestive glands of juvenile A. californica eaten by the sea anemone Anthopleura xanthogrammica were largely undigested. In most instances the organs were egested by the anemones before the protective membranes enclosing the glands could be perforated by the action of the digestive juices. Preliminary tests, in which water- and acetone-extractions of the digestive glands of Aplysia californica were injected intraperitoneally into mice and other small mammals, and subcutaneously into frogs and baby chicks, indicated a high level of toxicity (Winkler, 1961). The test animals entered into various states of paralysis and respiratory distress, leading to death. Winkler, Tilton & Hardinge (1962) named the toxin aplysin and noted that its effect was reminiscent of cholin esters, such as acetylcholine and succinylcholine (see also Winkler & Tilton, 1962; Langlais & Blankenship, 1972). In later studies on A. pulmonica and other related opisthobranchs in Hawaii, Watson (1973) and Watson & Rayner (1973) identified two lethal substances in digestive glands originally extracted with acetone, one

226

THOMAS H.CAREFOOT

termed water-soluble toxin, the other, ether-soluble toxin. On intraperitoneal injection into mice, the ethersoluble portion produced irritability, visciousness, and flaccid paralysis. The water-soluble portion, in contrast, caused convulsions and respiratory distress. Both toxins ultimately killed the mice, although similar extracts from A. juliana were not lethal (Watson, 1973). Watson & Rayner (1973) compared the pharmacological effects of the water-soluble toxin from A. pulmonica and other aplysiids described by Watson (1973) with the effects of A. californica aplysin described by Winkler et al. (1962) and found some similarities between the two substances. At about the same time, a number of brominated compounds were being isolated from sea hares, mainly from the digestive gland. In the first of these studies, Tanaka & Toyama (1959) identified several brominated compounds in A. kurodai. Later, two bromine-containing sesquiterpenes were isolated from the same species by Yamamura & Hirata (1963). The authors independently named the substances aplysin and aplysinol. A third brominated material aplysin-20 was later isolated from A. kurodai by Matsuda, Tomiie, Yamamura & Hirata (1967) and Yamamura & Hirata (1971). The aplysin (from A. kurodai) and a related structure debromoaplysin have since been synthesized by Yamada, Yazawa, Toda & Hirata (1968), and the optically active forms, (—)-aplysin and (—)-debromoaplysin, by Ronald, Gewali & Ronald (1980). Winkler (1969) investigated the distribution of organic bromine compounds in A. californica and concluded that over 90% of them were present in the digestive gland, with most of the remainder being located in the foot, body wall, and skin. Further studies by Stallard & Faulkner (1974a) on A. californica showed that greater than 99% of the bromine was located in the ether-soluble fraction of the digestive gland. Bromine compounds were absent from the opaline gland in A. californica (Winkler, 1969). Some confusion therefore existed concerning the toxic compounds in Aplysia, not just in the designation of the name aplysin to two different substances, but in the chemical nature of the substances and their possibly differing pharmacological effects. As noted by Blankenship, Langlais & Kittredge (1975), the cholinomimetic aplysin of Winkler (1961) was fairly well characterized in terms of its action, but its chemical identity was unknown. At the same time the halogenated terpenoids described for A. kurodai (including the “aplysin” of Yamamura & Hirata, 1963) were well known chemically, but their pharmacological action, if any, was unknown. Although Watson (1973) did not test for brominated compounds in the ether extracts he obtained from A. pulmonica and other aplysiids, he did suggest that these might be similar to the brominated substances isolated from A. kurodai. In fact, a later study by Kato & Scheuer (1974) on Steilocheilus longicauda, a related aplysiid and one included in Watson’s study, showed that the ether-soluble toxin isolated by Watson from various opisthobranchs was probably a mixture of terpenoid-like substances found also in Steilocheilus, one of which was brominated (see also Scheuer, 1975, 1977). These substances were designated by the authors as aplysiatoxin and debromoaplysiatoxin (the latter substance was later isolated from the blue-green alga Lyngbya majuscula, where it has been shown to cause dermatitis in humans similar to “swimmer’s itch” and to have antileukemic properties: Mynderse, Moore, Kashiwagi & Norton, 1977; and to be a weak tumour promoter: Fujiki et al., 1982, 1984, 1985; see also Moore, 1982, and Willey, Moser & Harris, 1984). Furthermore, the cholinomimetic toxin first described by Winkler (1961), which possessed properties similar to the water-soluble toxin described by Watson (1973), was later isolated from A. californica and identified as urocanylcholine by Blankenship et al. (1975). Urocanylcholine and related choline esters belong to a family of toxic compounds known as murexine, more familiarly known from prosobranch gastropods in the families Muricidae and Thaisidae. These compounds mimic the effects of acetylcholine, are resistant to cholinesterase, and are fairly heat stable (Blankenship et al., 1975). These authors also reviewed the literature on the effects of murexine on animals and concluded that its pharmacological effects (mainly neuromuscular blocking action) were similar to

APLYSIA

227

those described for the original aplysin by Winkler (1961) and Winkler et al. (1962). Also, the activity of Winkler’s aplysin could not be eliminated by cholinesterase, which is similar again to the response of urocanylcholine and related choline esters. It therefore appears that the cholinergic aplysin originally described by Winkler may have been urocanylcholine, a member of the murexine family of toxins. Blankenship and his colleagues suggested that since the common name “aplysin” is most frequently applied to the halogenated terpenes, it should be used in this respect only, and not in reference to the cholinomimetic compound. From the time of these first studies in the 1960s, secondary metabolites of sea hares have been intensively investigated by ‘natural-products’ chemists. This has led to a rich harvest from Aplysia spp. of mono-, sesqui-, and diterpenoid substances, as well as other compounds, most of them halogenated, and most of them from the digestive glands (e.g. Faulkner & Stallard, 1973; Faulkner, Stallard, Fayos & Clardy, 1973; Faulkner, Stallard & Ireland, 1974; Kato & Scheuer, 1974; Schmitz, Hollenbeak & Vanderah, 1978a; Schmitz, McDonald & Vanderah, 1978b; and many others). The number and variety of these substances is quite amazing. Table XIII lists them and provides information on the body part sampled, on the chemical type, molecular formula, and special names, if any, given the substances, and on their pharmacological effects, if known. Little work has been done on the biological effects of these substances, particularly with regard to their rôle in the natural biology of Aplysia. The seaweed foods of Aplysia, with their wide variety of secondary metabolites, are thought to be the source of these various compounds. In the genus Laurencia alone, over 250 natural products have been isolated, many of them representing new structural types (Erickson, 1983). Winkler (1961) was perhaps the first to suggest that the toxins may derive from Aplysia’s seaweed diet, and he later considered that various red algae known to contain brominated compounds (as Laurencia spp.; see also Augier & Mastagli, 1956; Irie, Suzuki & Masamune, 1965a; Irie et al., 1965b; Irie, Suzuki, Kurosawa & Masamune, 1966; Irie, Suzuki & Hayakawa, 1969; Craigie & Gruenig, 1967; Bhakuni & Silva, 1974; Fenical, 1975) might be the source of the brominated organic compounds (Winkler, 1969). While Darling & Cosgrove (1966) were unable to identify a possible brominated precursor for these compounds in the red alga Plocamium cartilagineum, several lines of evidence support strongly the notion of a dietary source for these materials. First of all, Yamamura & Hirata (1963) initially noted that the kind and content of brominated substances in Aplysia kurodai depended on when and where the animals were collected. Mynderse & Faulkner (1978) also recognized a possible effect of diet on the content of halogenated substances in Aplysia. These authors were able to relate variations in content of certain halogenated monoterpenes in the digestive gland of A. californica to variations in the levels of these substances in the foods eaten, especially in the red alga Plocamium cartilagineum. By isolating and identifying 12 compounds, most with new structures, from Aplysia dactylomela collected near La Parguera, Puerto Rico, Schmitz et al. (1981) also provided evidence that the digestive gland components may vary with habitat. They did not isolate any of these same compounds from A. dactylomela collected at Bimini (Schmitz et al., 1978a, b). Further indirect support for the idea of a dietary source of the compounds was provided by Watson (1973). He showed that extracts and homogenates of the digestive gland of A. juliana were not lethal. This species eats mainly the green alga Ulva which lacks brominated compounds. A number of additional studies have related substances isolated from Aplysia with similar or even identical compounds in seaweeds, often from actual foods eaten by the sea hares. For example, various sesquiterpenes isolated from Laurencia spp. were found to be structurally similar to the sesquiterpenes of Aplysia (Irie et al., 1965a, b, 1966, 1969; see Scheuer, 1971 for review). In fact, Irie et al. (1969) isolated aplysin, debromoaplysin, and aplysinol from Laurencia okamurai, which suggests that some sea hares might simply ingest their brominated terpenes intact, rather than perform chemical changes on precursor

228

THOMAS H.CAREFOOT

compounds found in the seaweeds (Scheuer, 1971). Stallard & Faulkner (1974a) have also proposed that there may be a dietary source of the halogenated substances in the digestive gland of A. californica. These authors showed that a number of halogenated terpenoids including aplysin, debromoaplysin, laurinterol, pacifenol, johnstonol, and pacifidiene, and two monoterpenes, were stored in the digestive gland. Not only are these brominated terpenes known to be found in the algal foods of A. californica (Irie et al., 1966, 1969; Sims, Fenical, Wing & Radlick, 1972; Kazlauskes, Murphy, Quinn & Wells, 1976), but the digestive glands of animals maintained solely on diets of Laurencia pacifica and Plocamium cartilagineum were shown on later analysis to contain the terpenes of each respective alga (halogenated sesquiterpenes from Laurencia and halogenated monoterpenes from Plocamium: Stallard & Faulkner, 1974a). Two of these algal metabolites, laurinterol and pacifenol, were further shown to undergo transformations to aplysin and pacifidiene, respectively, within the digestive glands of Aplysia californica (Stallard & Faulkner, 1974b). This latter finding removed some of the doubt that the algal substances were not just being identified amongst the undigested food constituents in the ducts of the digestive gland. In addition, Faulkner et al. (1974) identified a halogenated sesquiterpene prepacifenol epoxide in extracts of the digestive gland of A. californica and showed its interconversion to “johnstonol”, the latter apparently an artifact formed during isolation of the epoxide from the red alga, Laurencia johnstonii, and previously identified as being present in the alga (Sims et al., 1972). Pacifenol, in turn, is known to be present in both the digestive glands of A. californica and in at least one of its seaweed foods Laurencia pacifica (Sims, Fenical, Wing & Radlick, 1971; Stallard & Faulkner, 1974a). Imperato, Stallard & Faulkner (1977) identified several polyhalogenated monoterpenes in the digestive glands of Aplysia fasciata and showed that one of these was present in the red alga Plocamium cartilagineum. In an interesting example of investigatory chemical ecology, the authors deduced that even though the crops of the test animals were filled mainly with the red alga Gracilaria verrucosa, the sea hares must at one time have eaten Plocamium cartilagineum, because Gracilaria does not contain halogenated monoterpenes while Plocamium does. Similarly, two diterpenes, dictyol A and dictyol B, were shown by Minale & Riccio (1976) to be present in the digestive glands of Aplysia depilans which were eating mainly the brown alga Dictyota dichotoma (known to contain these substances), but not in Aplysia fasciata or A. punctata which were eating mainly red algae and no Dictyota. Several other diterpene substances, including amongst them pachydictyol A, were found by Minale and colleagues (Minale & Riccio, 1976; Danise et al., 1977) in Aplysia depilans, and by Vanderah & Faulkner (unpubl. obs., cited by Fenical et al., 1979) in A. vaccaria. These compounds are known also from various brown algae, including Pachydictyon coriaceum (the source of pachydictyol A: Hirschfeld et al., 1973) and other members of the family Dictyotaceae (e.g, Dictyota dichotoma and Dilophus ligulatus: Fattorusso et al., 1976; Danise et al., 1977). Two bromosesquiterpenes discovered in the digestive glands of Aplysia brasiliana=willcoxi, brasudol and isobrasudol, were also found in the red alga Chondria cnicophylla on which the sea hare feeds (Dieter, Kinnel, Meinwald & Eisner, 1979). Finally, a diterpene monoacetate from Aplysia dactylomela (Schmitz et al., 1981) closely resembled the diterpene diacetate isolated from the red alga Laurencia obtusa (Higgs & Faulkner, 1982). In summary, the evidence from these studies on the chemical ecology of

APLYSIA

TABLE XIII Chemical substances in the digestive glands or whole body of Aplysia Species

Part of body sampled

Chemical type

brasiliana =willcoxi

Digestive gland

Choline (possibly acetylcholine and urocanylcholi ne)

brasiliana =willcoxi brasiliana =willcoxi

Digestive gland Digestive gland

Linear C15

C15H15O2Br

Panacene

Sesquiterpene

C15H26O (two isomers: brasilenol and epibrasilenol)

Brasilenol Epibrasilenol

brasiliana =willcoxi

Digestive gland

Sesquiterpene

Brasudol Isobrasudol

brasiliana =willcoxi

Digestive gland

Linear C15

C15H25BrO (two isomers: brasudol and isobrasudol) C15H19ClO (brasilenyne), C15H20BrClO (two isomers: cisdihydrorhodo phytin and cisisodihydrorho dophytin)

californica and vaccaria

Digestive gland

Cholinomimet ic substance

Chemical formula

Name given to Pharmacologic substance al effects

References

Cholinergic action on lantern muscles of the sea urchin, Echinometra lucunter, and muscles of the sea cucumber, Ludwigothuria grisea; decreased frequency and amplitude of beating in the toad heart (Bufo ictericus) and stopped the heart in diastole

De Freitas, 1977

Brasilenyne

Aplysin

When painted on the crab, Callinectes sapidus, the extract elicited avoidance by Octopus sp. Brasudol is a potent feeding deterrent to several fish Both substances unpalatable to swordtail fish (Xiphophorus helleri)

Neuromuscula r effects on injection into frogs, baby chicks and

229

Kinnel et al., 1977 Stallard et al., 1978

Dieter et al., 1979

Kinnel et al., 1979

Winkler, 1961; Winkler et al., 1962; Winkler & Tilton, 1962

230

THOMAS H.CAREFOOT

Species

Part of body sampled

Chemical type

Chemical formula

californica

Digestive gland

Urocanylcholi ne or murexine

californica

Digestive gland Digestive gland Digestive gland and skin

Monoterpene

C10H16OBr3Cl

Monoterpene

C10H12Br3Cl3

Sesquiterpene and monoterpene

californica

Digestive gland

Monoterpene and sesquiterpene

C15H21O3Br2C l (prepacifenol epoxide), C15H21Br2ClO 3, and C15H21OBr2Cl ; also C10H13Br2Cl3 and C12H15O2Cl3 C10H16OBr3Cl and C10H12Br3Cl3 (see also Faulkner et al., 1973); also C15H19OBr (aplysin) and C15H20O (debromoaply sin), and C15H19OBr2Cl (pacifidiene), and pacifenol, johnstonol, and laurinterol

dactylomela

Whole animal (?)

dactylomela

Whole animal

californica californica

Name given to Pharmacologic substance al effects mice: caused paralysis and death Depolarizing blockage of frog rectus and rat diaphragm muscle; block of myoneural junctions

C15H24O

Langlais & Blankenship, 1972; Blankenship et al., 1975

Faulkner & Stallard, 1973 Faulkner et al., 1973 Faulkner et al., 1974; Ireland et al., 1976

Prepacifenol epoxide

Stallard & Faulkner, 1974a

Amelioration of leukemia P-388 in mice Sesquiterpene

References

DactyloxeneB

Sigel et al., 1970 Schmitz & McDonald, 1974

APLYSIA

dactylomela

Whole animal or digestive gland

Linear C15

dactylomela

Whole animal (?)

Sesquiterpene

dactylomela

Whole animal

dactylomela

Whole animal and digestive gland

C15H19OBr2Cl (two isomers: dactylyne and isodactylyne)

231

Dactylyne Isodactylyne

Intraperitoneal injection of dactylyne into mice and rats produced CNS depresssion as shown by decrease in spontaneous and locomotory activities, and decreased body temperature; also inhibits the metabolic elimination of pentobarbiton e in mice

McDonald et al., 1975; Vandereh & Schmitz, 1976; Kaul et al., 1978a, b; Kaul & Kulkarni, 1978

C15H21BrO3

Aplysistatin

Antileukemial properties against P-388 lymphocytic mouse leukemia; (−)aplysistatin has been synthesized by Shieh & Prestwich (1982) and Tanaka et al. (1984), and (+)aplysistatin by Hoye et al. (1982) and White et al. (1982)

Pettit et al., 1977; Von Dreele & Kao, 1980

Diterpene

C20H30Br2O3

Angasiol

Sesquiterpene

C15H26O

Dactylol

Pettit et al., 1978 Schmitz et al., 1978a

232

THOMAS H.CAREFOOT

Species

Part of body sampled

Chemical type Chemical formula

Name given to substance

dactylomela

Whole animal and digestive gland

Sesquiterpene

Dactyloxene Dactylenol

Schmitz et al., 1978b

dactylomela

Digestive gland

Sesquiterpene

Deodactol Isodeodactol

dactylomela

Digestive gland

Diterpene

Hollenbeak et al., 1979; Gopichand et al., 1981 Schmitz et al., 1979

dactylomela

Digestive gland

Sesquiterpene

dactylomela

Digestive gland

Linear C15

dactylomela

Digestive gland Sterol

dactylomela

Digestive gland (animals from La Parguera, Puerto Rico)

Diterpene, indole, sesquiterpene

C15H24O (dactyloxene and dactylenol) C15H25O2Br2C l (two isomers: deodactol and isodeodactol) C20H35O2Br

Some cytotoxicity shown against the National Cancer Institute’s cell cultures P-388 lymphocytic leukemia and L-1210 lymphoid leukemia

C17H27O5Br2C l C15H20OBrCl and C15H20O2BrCl 5α, 8αepidioxy sterols C22H33BrO5 (two isomers: parguerol and isoparguerol) and C22H33BrO4 (deoxypargue rol); also parguerol 16acetate and isoparguerol 16-acetate; also Nmethylindole, elatol, allolaurintol acetate, and isoobtusol acetate

Pharmacologic al effects

References

Schmitz et al., 1980 Gopichand et al., 1981 Gunatilaka et al., 1981

Parguerol Deoxyparguer ol Isoparguerol

Schmitz et al., 1981, 1982

APLYSIA

dactylomela

Digestive gland

Diterpene

C20H34O3

dactylomela

Digestive gland

Sesquiterpene

C15H22Br2O (two isomers) C9H7Br2N and C9H6Br3N C20H30O2 (dictyol A), C20H32O2 (dictyol B), and C20H32O (pachydictyol A)

dactylomela and juliana depilans

Indole Digestive gland

Diterpene

depilans

Digestive gland

Diterpene

depilans

Whole animal (?) Digestive gland

Diterpene

fasciata

kurodai

Whole animal

kurodai

Whole animal

Possesses some antimicrobial activity

C20H34O2 (dictyol C) and C20H32O2 (two isomers: dictyol D and E) C20H30O2

Monoterpene

233

Gonzalez et al., 1983a

Gonzalez et al., 1983b Kato, 1984 Pachydictyol A has mild antibiotic activity vs. Staphylococc us aureus (Hirschfeld et al., 1973)

Minale & Riccio, 1976

Danise et al., 1977

Dictyolactone

Finer et al., 1979 Imperato et al., 1977

C10H11Cl15 and C10H11Cl3Br2 C16H21OBr

kurodai

Whole animal

Diterpene

C15H19OBr Aplysin (aplysin), Aplysinol C15H20O Debromoaply (debromoaply sin sin), and C15H19O2Br (aplysinol) C20H35O2Br Aplysin-20

kurodai

Whole animal (?)

Diterpene

C20H35O2Br

Tanaka & Toyama, 1959 Yamamura & Hirata, 1963

Sesquiterpene

Isoaplysin-20

Debromoisoap lysin-20 has been synthesized by

Matsuda et al., 1967; Yamamura & Hirata, 1971 Yamamura & Terada, 1977

234

THOMAS H.CAREFOOT

Species

Part of body sampled

Chemical type Chemical formula

Name given to substance

kurodai

Digestive gland

Monoterpene

C10H15OBr3Cl2

Kurodainol

oculifera

Digestive gland

Linear C15

oculifera

Whole animal

Linear C15

C15H20Br2O2 (two isomers) C15H20BrClO

Srilankenyne

parvula

Digestive gland (?)

Diterpene

pulmonica

Digestive gland

Ether-soluble toxin

Water-soluble toxin

vaccaria

Digestive gland

Diterpene

C20H30O4 (dihydroxy crenulide), C22H32O5, C20H30O3, and C22H32O4

Pharmacologica References l effects Imamura & Ruveda, 1980 Katayama et al., 1982 Schulte et al., 1981 Dilip de Silva et al., 1983 Fenical & Howard, unpubl. obs. (cited in Fenical et al., 1979) Intraperitoneal Watson, 1973; Watson & injection into Rayner, 1973 mice causes vasoconstrictio n and irritability Intraperitoneal injection into mice causes hypotension and bradycardia; convulsions, respiratory distress, death Midland et al., 1983

Aplysia seems to offer convincing support for the theory that the diverse array of brominated compounds in their digestive glands are derived from dietary seaweeds. The majority of these compounds can be divided into four categories consisting of the monoterpenes, the sesquiterpenes, the diterpenes, and the linear Cl5 compounds, based on their carbon skeletons (see Table XIII). This simple classification is useful as a guide to the algal origin of the substances. In general, the monoterpenes come from Plocamium spp., the sesquiterpenes from Laurencia spp., the diterpenes from Laurencia spp. and various brown algae such as the Dictyotaceae, and the linear Cl5 compounds exclusively from Laurencia spp. (Andersen, pers. comm.). Data are sparse on the possible function of these materials in the seaweeds from which they are derived (e.g. possible antibacterial and antiviral activity; see Fenical, 1975; Finer et al., 1979), and almost nothing is known of their potential function in Aplysia. It is interesting to note that several sea hares which subsist primarily on green algae, such as Ulva spp., have no dietary source of brominated terpenes. Yet, these sea

APLYSIA

235

hares, of which Aplysia juliana is probably the best example, appear to be as predator-free as any other species possessing the supposedly defensive brominated compounds. Do the opaline secretions take on the major rôle of deterring predators in such species, or do these animals rely on cholinomimetic or other substances as their first line of defence? There appear to be several unexplored lines of research in the area of predatorprey interactions of sea hares, which would richly repay investigation. INTERNAL DEFENCES Research into possible internal defensive mechanisms in sea hares has focused on whether haemagglutinins are present in the haemolymph (McKay, Jenkin & Rowley, 1969; Pauley, Granger & Krassner, 1971a; Pauley, 1971; see also Bevelaqua, Kim, Kumarasiri & Schwartz, 1975), their possible function as opsonins (Pauley, Krassner & Chapman, 1971b) and, in a related way, their possible participation in macrophagemediated cytolysis of tumours in mice (Yamazaki et al., 1983a, b). Haemagglutinins appear to be present in A. californica (McKay et al., 1969; Pauley et al., 1971a, b) where they may act as opsonins (factors which enhance phagocytosis). Pauley (1971) and Pauley et al. (1971b) tested in vivo clearance rates of several species of marine bacteria and red blood cells of pigs and chickens injected into A. californica, and noted that serum agglutinin titres decreased initially after exposure to the bacteria and red blood cells. Titres later returned to normal. The authors also noted that clearance of a bacterial species was accelerated by previous exposure of the sea hare to the bacterium. The agglutinin was, however, apparently non-specific, as treatment with one bacterial species completely prevented the serum agglutinating several other bacterial species or red blood cells (Pauley, 1971; Pauley et al., 1971a). Enhanced clearance activity was retained as long as one month after the primary injection, suggesting the presence of a memory response (Pauley et al., 1971b). The authors further noted that in vitro phagocytosis of chicken red blood cells was enhanced by opsonic factors in the haemolymph of the sea hare, possibly by the same molecule or group of molecules involved in agglutination (see also Pauley, 1974). Furthermore, while the haemolymph of A. californica was found to be sterile (Pauley et al., 1971b), no bactericidal activity was evident (Johnson & Chapman, 1970; Pauley et al., 1971b); hence, elimination of marine bacteria from the haemolymph was not apparently due to a lytic effect. Sea hare eggs may possess antibiotic factors, as the egg masses appear to be free of bacteria. Kamiya & Shimizu (1981) demonstrated potent agglutinins in extracts of A. kurodai egg masses which could agglutinate mammalian erythrocytes and marine bacteria. In a later study, Kamiya, Muramoto & Ogata (1984) discovered that eggs of A. kurodai do exhibit antibacterial activity and suggested that antibacterial factors were produced in the albumen gland, such that each egg was coated with antibacterially active albumen before passing down the oviduct. Although Kamiya et al. (1984) could not actually demonstrate haemagglutinin activity in the albumen gland of A. kurodai, Gilboa-Garber, Mizrahi & Susswein (1984) and Gilboa-Garber, Susswein, Mizrahi & Avichezer (1985) found that gonadal extracts from A. californica, A. depilans, A. fasciata, and A. oculifera contained considerable haemagglutinating activity. This activity was lectin-mediated. The authors suggested a number of possible functions for the lectins (a protein that specifically binds sugars), including binding of the fertilized eggs to the capsules, and protection. Finally, Yamazaki et al. (1984, 1985) identified a cytolytic factor in the eggs of A. kurodai which lyses tumour cells in vitro and inhibits tumour growth in vivo in mice. In general, this picture of the internal defences of Aplysia conforms with that of other gastropods (see review by Bayne, 1983).

236

THOMAS H.CAREFOOT

PARASITES A number of parasites and other animal associates of Aplysia have been reported. These are: (1) unidentified microorganisms in the abdominal ganglion of A. californica (Coggeshall, 1967), possibly the same as a new species of microsporidan found also in the abdominal ganglia of A. californica and named Microsporidium aplysiae (Krauhs, Long & Baur, 1979); (2) a metacercarian larva in the left pedal ganglion of Aplysia fasciata (Vicente, 1962); (3) colonies of the hydroids Obelia and Pedicellina attached to the mantle of young Aplysia punctata (Eales, 1921); (4) the copepods Strongylopleura pruvoti, in the mantle cavity of Aplysia sydneyensis and on the ctenidium of Australian specimens of A. juliana (Eales, 1960), and Anthessius variedens var. aplysiae associated with Aplysia brasiliana=willcoxi (Cipolli & Sawaya, 1967; Sawaya & Leahy, 1971); and finally, (5) an unidentified crab in the mantle cavity of A. californica (Lickey et al., 1977), possibly the same as, or similar to, the pinnotherid crab Opisthopus transversus observed in the mantle cavity of Aplysia vaccaria (Beondé, 1968). At present, there is insufficient information to determine whether the copepod and crab associates are parasites, or some sort of commensals, or indeed whether the other parasites listed are localized or widespread in their occurrence. ACKNOWLEDGEMENTS I thank Elizabeth Carefoot, Gerri Cheng, Kathy Gorkoff, David Huggard, Vita Janusas, John McNicol, and Mark Roberts for technical help in the preparation of the manuscript; Sarah Smith for editorial comments; Anthony Barrett, Horacio de la Cueva, Tom Mommsen, and Sam Gopaul for their assistance in translating papers; Mike Hawkes and Julie Oliveira for advice on seaweeds; and Raymond Andersen, Chris Bayne, Bob Blake, Peter Hochachka, Al Lewis, Sandra Millen, and Marilyn Switzer-Dunlap for helpful scientific comments on various parts of the manuscript. The work was supported in part by an Operating Grant from the Natural Scientific and Engineering Research Council of Canada. REFERENCES Achituv, Y. & Susswein, A.J., 1985. J. exp. mar. Biol. Ecol., 85, 113–122. Advokat, C., 1980. Behav. neur. Biol., 28, 253–265. Advokat, C., Carew, T. & Kandel, E., 1976. Neurosc. Abstr., 2, 313 only. Allan, J.K., 1932. Rec. Aust. Mus., 18, 314–320. Allan, J.K., 1941. Victoria Nat., 57, 178–182. Ambrose, III, H.W., Givens, R.P., Chen, R. & Ambrose, K.P., 1979. Mar. Behav. Physiol., 6, 57–64. Ando, Y., 1952. Kagaku, 22, 87–88. Ansell, A.D., 1982. J. exp. mar. Biol. Ecol., 61, 1–29. Arch, S. & Smock, T., 1977. Behav. Biol., 19, 45–54. Arch, S., Smock, T., Gurvis, R. & McCarthy, C. 1978. J. comp. Physiol., 128, 67–70. Arnould, C. & Jeuniaux, C., 1977. Cah. Biol. mar., 18, 465–473. Arvanitaki, A. & Chalazonitis, N., 1961. In, Nervous Inhibition, edited by E.Florey, A Pergamon Press Book, The Macmillan Co., New York, pp. 194–231. Aspey, W.P. & Blankenship, J.E., 1975. Fed. Proc., 34, 418 only. Aspey, W.P. & Blankenship, J.E., 1976a. Behav. Biol., 17, 279–299. Aspey, W.P. & Blankenship, J.E., 1976b. Behav. Biol., 17, 301–312. Aspey, W.P., Cobbs, J.S. & Blankenship, J.E., 1977. Behav. Biol., 19, 300–308. Audesirk, T., 1975a. Am. Soc. Zool., 15, 796. Audesirk, T.E., 1975b. Behav. Biol., 15, 45–55.

APLYSIA

237

Audesirk, T.E., 1976. Ph.D. thesis, University of Southern California, Los Angeles. Audesirk, T.E., 1977. Behav. Biol., 20, 235–243. Audesirk, T.E., 1979. Biol. Bull. mar. biol. Lab., Woods Hole, 157, 407–421. Audesirk, T.E. & Audesirk, G.J., 1977. Comp. Biochem. Physiol., 56A, 267–270. Augier, J. & Mastagli, P., 1956. C.r. hebd. Séanc. Acad. Sci, Paris, 242, 190–192. Austin, T., Weiss, S. & Lukowiak, K., 1983. Can. J. Physiol. Pharmacol, 61, 949– 953. Baba, K., 1936. J. Dep. Agric., Kyūshū Imper. Univ., 5, 1–50. Baba, K., 1937. Dobutsu. Zasshi, 49, 57–63. Baba, K., 1970. Collect. Breed., 32, 94–96. Baba, K., Hamatani, I. & Hisai, K., 1956. Publs Seto mar. biol. Lab., 5, 209–220. Bailey, C.H., Castellucci, V.F., Koester, J. & Chen, M., 1983. Behav. new. Biol., 38, 70–81. Balaparameswara Rao, M., 1976. J. moll. Stud., 42, 136–145. Bandel, K., 1976. Stud. Neotrop. Fauna Environ., 11, 87–118. Barash, A. & Danin, Z., 1971. Israel J. Zool., 20, 151–200. Barash, A. & Zenziper, Z., 1980. Veliger, 22, 299–317. Bardach, J.E., 1975. In, Olfaction and Taste V, edited by D.A.Denton & J.P. Coghlan, Proc. Fifth int. Symp., Univ. Melbourne, Aust., Academic Press, New York, pp. 121–132. Barnes, D.M., 1986. Science, 231, 1246–1249. Bartsch, P. & Rehder, H.A., 1939. Smithson. misc. Collns, Vol. 98, 1–18. Bayne, B.L. & Scullard, C., 1978. J. exp. mar. Biol. Ecol., 32, 113–129. Bayne, C.J., 1983. In, The Mollusca. Volume 5. Physiology, Part 2, edited by A.S.M.Saleuddin & K.M.Wilbur, Academic Press, New York, pp. 407–486. Bebbington, A., 1972. Pubbl. Staz. zool. Napoli, 38, 25–46. Bebbington, A., 1974. Zool. J. Linn. Soc., 54, 63–99. Bebbington, A., 1975. Pubbl. Staz. zool. Napoli, 39, 121–128. Bebbington, A., 1977. Trans. zool. Soc. Lond., 34, 87–147. Bebbington, A. & Brown, G.H., 1975. J. Conchol, 28, 329–333. Bebbington, A. & Hughes, G.M., 1973. Proc. malac. Soc. Lond., 40, 399–405. Bebbington, A. & Thompson, T.E., 1969. Malacologia, 9, 253 only. Beeman, R.D., 1960. Bull. Sth. Calif. Acad. Sci., 59, 144–152. Beeman, R.D., 1961. Veliger, 3(suppl.), 87–102. Beeman, R.D., 1963. Veliger, 5, 145–147. Beeman, R.D., 1969. Biol. Bull. mar. biol. Lab., Woods Hole, 136, 141–146. Beondé, A.C., 1968. Veliger, 10, 375–378. Bethe, A., 1929. Pflügers Arch. ges. Physiol., 221, 344–362. Bethe, A., 1930. J. gen. Physiol. 13, 437–444. Bethe, A., 1934. Pflügers Arch. ges. Physiol., 234, 629–644. Bevelaqua, F.A., Kim, K.S., Kumarasiri, M.H. & Schwartz, J.H., 1975. J. biol. Chem., 250, 731–738. Bhakuni, D.S. & Silva, M., 1974. Botanica mar., 17, 40–51. Blankenship, J.E., 1980. In, The Role of Peptides in Neural Function, edited by J.L. Barker & T.G.Smith Jr, Marcel Dekker, Inc., New York, pp. 159–187. Blankenship, J.E., Langlais, P.J. & Kittredge, J.S., 1975. Comp. Biochem. Physiol., 51C, 129–137. Blankenship, J.E., Rock, M.K., Robbins, L.C., Livingston, C.A. & Lehman, H.K., 1983. Fed. Proc., 42, 96–100. Blankenship, J.E., Rock, M.K. & Schlesinger, D.H., 1982. Prog. clin. biol. Res., 79, 159–177. Block, G. & Smith, J.T., 1973. Comp. Biochem. Physiol. 46A, 115–121. Block, G.D., 1971. Physiol., 14, 112 only. Block, G.D., Hudson, D.J. & Lickey, M.E., 1974. J. comp. Physiol., 89, 237–249. Block, G.D. & Roberts, M.H., 1981. J. comp. Physiol., 142, 403–410. Bostock, J. & Riley, H.T., 1856. The Natural History of Pliny. Vol. V. Henry G. Bohn, London, 523 pp.

238

THOMAS H.CAREFOOT

Bostock, J. & Riley, H.T., 1857. The Natural History of Pliny. Vol. VI. Henry G. Bohn, London, 529 pp. Bottazzi, F., 1897. Arch. Ital. Biol., 28, 81–90. Braams, W.G. & Geelen, H.F.M., 1953. Arch. Néerl. Zool., 10, 241–264. Branch, G.M., 1981. Oceanogr. Mar. Biol. Ann. Rev., 19, 235–380. Breuer, J.P., 1962. Publ. Inst. mar. Sci. Univ. Texas, 8, 153–183. Bridges, C.B., 1975. Ophelia, 14, 161–184. Brown, A.M. & Brown, H.M., 1973. J. gen. Physiol., 62, 239–254. Bruce, J.R., Colman, J.S. & Jones, N.S., 1963. Marine Fauna of the Isle of Man and Its Surrounding Seas. Mem. No. 36, Liverpool University Press, Liverpool, 307 pp. Burky, A.J., 1971. Ecol. Monogr., 41, 235–251. Byrne, J.H., 1981. Brain Res., 204, 200–203. Capo, T.R., Perritt, S.E. & Berg, Jr, C.J., 1979. Biol. Bull. mar. biol. Lab., Woods Hole, 157, 360 only. Carazzi, D., 1900. Anat. Anz., 17, 77–102. Carefoot, T.H., 1967a. J. mar. biol. Ass. U.K., 47, 565–590. Carefoot, T.H., 1967b. Comp. Biochem. Physiol. 21, 627–652. Carefoot, T.H., 1967c. J. mar. biol. Ass. U.K., 47, 335–350. Carefoot, T.H., 1970. J. exp. mar. Biol. Ecol., 5, 47–62. Carefoot, T.H., 1979. Can. J. Zool., 57, 2271–2273. Carefoot, T.H., 1980. J. exp. mar. Biol. Ecol., 42, 241–252. Carefoot, T.H., 1981a. In, Proc. Fourth int. Coral Reef Symp. Vol. 2, pp. 665–669. Carefoot, T.H., 1981b. Can. J. Zool., 59, 445–454. Carefoot, T.H., 1982a. Proc. 2nd int. Conf. Aquacult. Nutr: Biochem. Physiol. Approaches Shellfish Nutr., pp. 321–337. Carefoot, T.H., 1982b. Mar. Biol., 68, 207–215. Carefoot, T.H., 1985. In, Proc. Fifth int. Coral Reef Congress, Tahiti, Vol. 4, pp. 9– 16. Carefoot, T.H., in press. In, Animal Energetics, edited by T.J.Pandian & F.J. Vernberg, Academic Press, New York (in press). Carew, T.J., Castellucci, V.F. & Kandel, E.R., 1971. Int. J. Neurosc., 2, 79–98. Carew, T.J., Hawkins, R.D. & Kandel, E.R., 1983. Science, 219, 397–400. Carew, T.J. & Kandel, E.R., 1977a. J. Neurophysiol., 40, 692–707. Carew, T.J. & Kandel, E.R., 1977b. J. Neurophysiol., 40, 708–720. Carew, T.J. & Kandel, E.R., 1977c. J. Neurophysiol., 40, 721–734. Carew, T.J. & Kupfermann, I., 1974. Behav. Biol., 12, 339–345. Carew, T.J., Pinsker, H.M. & Kandel, E.R., 1972. Science, 175, 451–454. Carew, T.J., Walters, E.T. & Kandel, E.R., 1981. Science, 211, 501–504. Carlton, J.T., 1985. Oceanogr. mar. Biol. Ann. Rev., 23, 313–371. Castellucci, V., Pinsker, H., Kupfermann, I. & Kandel, E.R., 1970. Science, 167, 1745–1748. Chapman, A.R.O., 1981. Mar. Biol., 62, 307–311. Chapman, D.J., Cole, W.J. & Siegelman, H.W., 1967. J. Am. chem. Soc., 89, 5976– 5977. Chapman, D.J. & Fox, D.L., 1969. J. exp. mar. Biol. Ecol., 4, 71–78. Chase, R., 1979a. Can. J. Zool., 57, 781–784. Chase, R., 1979b. Can. J. Zool., 57, 698–701. Cho, D.M., Pyeun, J.H., Byun, D.S. & Kim, C.Y., 1983. Bull. Korean fish. Soc., 16, 216–224. Christomanos, A., 1955. Nature, Lond., 175, 310 only. Cipolli, I.N. & Sawaya, P., 1967. Cienc. Cult. (São Paulo), 19, 432 only. Clark, K.B., 1984. Nautilus, 98, 85–97. Cobbs, J.S. & Pinsker, H.M., 1982. J. comp. Physiol. 147, 523–535. Coggeshall, R.E., 1967. J. Neurophysiol., 30, 1263–1287. Coggeshall, R.E., 1969. J. Morph., 127, 113–131. Coleman, N., 1975. Sea Front., 21, 344–346.

APLYSIA

239

Cook, D.G. & Carew, T.J. 1986. Proc. natn. Acad. Sci. U.S.A., 83, 1120–1124. Cook, E.F., 1962. Veliger, 4, 194–196. Cooper, J.G., 1863. Proc. Calif. Acad. Nat. Sci., 3, 56–60. Corrent, G., Eskin, A. & Kay, I., 1982. Am. J. Physiol., 242, R326–332. Craigie, J.S. & Gruenig, D.E., 1967. Science, 157, 1058–1059. Czeczuga, B., 1984. Comp. Biochem. Physiol., 78B, 259–264. Danise, B., Minale, L., Riccio, R., Amico, V., Oriente, G., Piattelli, M., Tringali, C., Fattorusso, E., Magno, S. & Mayol, L., 1977. Experientia, 33, 413–415. Darling, S.D. & Cosgrove, R.E., 1966. Veliger, 8, 178–180. De Freitas, J.C., 1977. Comp. Biochem. Physiol, 56C, 57–61. De-Negri, A. & De-Negri, G., 1876. Rc. Accad. Lincei, Atti (Roma), Ser. 2, 3, 394– 449. Denny, M., 1980. Science, 208, 1288–1290. Dieringer, N., Koester, J. & Weiss, K.R., 1978. J. comp. Physiol., 123, 11–21. Dieter, R.K., Kinnel, R., Meinwald, J. & Eisner, T., 1979. Tetrah. Lett., 19, 1645– 1648. Dijkgraaf, S. & Hessels, H.G.A., 1969. Z. vergl. Physiol., 62, 38–60. Dilip de Silva, E., Schwartz, R.E. & Scheuer, P.J., 1983. J. org. Chem., 48, 395–396. DiMatteo, T., 1981a. Mar. Behav. Physiol., 7, 285–290. DiMatteo, T., 1981b. Veliger, 24, 72–75. DiMatteo, T., 1982. J. exp. mar. Biol. Ecol., 57, 169–180. Downey, P. & Jahan-Parwar, B., 1972., Physiol., 15, 122 only. Dudek, F.E., Injeyan, H.S., Soutar, B., Weir, G. & Tobe, S.S., 1980a. Can. J. Zool., 58, 2220–2229. Dudek, F.E., Soutar, B. & Tobe, S.S., 1980b. Can. J. Zool., 58, 2163–2166. Eales, N.B., 1921. Liverpool Marine Biology Committee Memoirs, No. 24, edited by W.A.Herdman & J.Johnstone, Liverpool University Press, Liverpool, 84 pp. Eales, N.B., 1944. Proc. malac. Soc. Land., 26, 1–21. Eales, N.B., 1957. Bull. Mus., Sér. 2, 29, 246–255. Eales, N.B., 1960. Bull. British Mus. (Nat. Hist.), Zoology, 5, 269–404. Eales, N.B., 1970. Proc. malac. Soc. Land., 39, 217–220. Edmunds, M., Potts, G.W., Swinfen, R.C. & Waters, V.L., 1974. J. mar. biol. Ass. U.K., 54, 939–947. Edmundson, C.H., 1946. Reef and Shore Fauna of Hawaii. Bernice P.Bishop Museum Spec. Publ. 22, Honolulu, Hawaii, 381 pp. Edwards, D.C. & Huebner, J.D., 1977. Ecology, 58, 1218–1236. Emery, D.G., 1976. Am. Zool., 16, 241 only. Emery, D.G. & Audesirk, T.E., 1978. Neurobiol., 9, 173–179. Emlen, J.M., 1966. Am. Nat., 100, 611–617. Engel, H., 1928. Proc. malac. Soc. Land., 18, 147–151. Engel, H. & Eales, N.B., 1957. Beaufortia, 6, 83–114. Erickson, K.L., 1983. In, Marine Natural Products, Vol. 5, edited by P.J.Scheuer, Academic Press, New York, pp. 131–257. Eskin, A., 1971. Z. vergl. Physiol., 74, 353–371. Eskin, A., 1979. Fed. Proc., 38, 2573–2579. Farmer, W.M., 1970. Veliger, 13, 73–89. Fattorusso, E., Magno, S., Mayol, L., Santacroce, C., Sica, D., Amico, V., Oriente, G., Piattelli, M. & Tringali, C., 1976. J. Chem. Soc., No. 13, 575–576. Faulkner, D.J. & Stallard, M.O., 1973. Tetrah. Lett., 14, 1171–1174. Faulkner, D.J., Stallard, M.O., Fayos, J. & Clardy, J., 1973. J. Am. chem. Soc., 95, 3413–3414. Faulkner, D.J., Stallard, M.O. & Ireland, C., 1974. Tetrah. Lett., 40, 3571–3574. Feinstein, R., Aspey, W.P. & Schmale, M.C., 1976. Behav. Biol., 17, 271–278. Feinstein, R., Pinsker, H., Schmale, M. & Gooden, B.A., 1977. J. comp. Physiol., 122, 311–324.

240

THOMAS H.CAREFOOT

Fenical, W., 1975. J. Phycol., 11, 245–259. Fenical, W., Sleeper, H.L., Paul, V.J., Stallard, M.O. & Sun, H.H., 1979. Pure appl. Chem., 51, 1865–1874. Ferguson, G.P., Parsons, D.W., ter Maat, A. & Pinsker, H.M., 1984. Neurosc. Abstr., 10, 150 only. Finer, J., Clardy, J., Fenical, W., Minale, L., Riccio, R., Battaile, J., Kirkup, M. & Moore, R.E., 1979. J. org. Chem., 44, 2044–2047. Flury, F., 1915. Arch. exp. Pathol. Pharmakol., 79, 250–263. Fong, W. & Mann, K.H., 1980. Can. J. Fish, aquatic Sci., 37, 88–96. Fontaine, M. & Raffy, A., 1936a. C. r. Séanc. Soc. Biol., 121, 735–736. Fontaine, M. & Raffy, A., 1936b. Bull. Soc. zool. Fr., 61, 49–57. Foreman, R.E., 1977. Helgol. wiss. Meeresunters., 30, 468–484. Fox, R.S. & Ruppert, E.E., 1985. Shallow-water Marine Benthic Macroinvertebrates of South Carolina. Species Identification, Community Composition and Symbiotic Associations. Belle W.Baruch Institute for Marine Biology and Coastal Research. University South Carolina Press, Columbia, S.C., 329 pp. Frings, H. & Frings, C., 1965. Biol. Bull. mar. biol. Lab., Woods Hole, 128, 211–217. Fujiki, H., Ikegami, K., Hakii, H., Suganuma, M., Yamaizumi, Z., Yamazato, K., Moore, R.E. & Sugimura, T., 1985. Jpn J. Cancer Res. (Gann), 76, 257–259. Fujiki, H., Suganuma, M., Nakayasu, M., Hoshino, H., Moore, R.E. & Sugimura, T., 1982. Gann, 73, 495–497. Fujiki, H., Suganuma, M., Tahira, T., Yoshioka, A., Nakayasu, M., Endo, Y., Shudo, K., Takayama, S., Moore, R.E. & Sugimura, T., 1984. In, Cellular Interactions by Environmental Tumor Promoters, edited by H.Fujiki et al., Japan Sci. Soc. Press, Tokyo/VNU Science Press, Utrecht, pp. 37–45. Gallager, S.M. & Mann, R., 1980. Mar. Biol. Lett., 1, 341–349. Gallin, E.K. & Wiederhold, M.L., 1977. J. Physiol., 266, 123–137. Garstang, W., 1890. J. mar. biol. Ass. U.K., 1, 399–457. Gerencser, G.A., 1978. Comp. Biochem. Physiol., 61A, 203–208. Gerencser, G.A., 1979a. Comp. Biochem. Physiol., 63A, 519–522. Gerencser, G.A., 1979b. Comp. Biochem. Physiol., 64A, 251–255. Gerencser, G.A., 1981a. Am. J. Physiol., 240, R61-R69. Gerencser, G.A., 1981b. Comp. Biochem. Physiol., 68A, 225–230. Gerencser, G.A., 1982. Comp. Biochem. Physiol., 72A, 721–725. Gerencser, G.A., 1983. Am. J. Physiol., 244, R412-R416. Gerencser, G.A., 1984a. Comp. Biochem. Physiol., 78A, 607–611. Gerencser, G.A., 1984b. Biochim. Biophys. Acta, 775, 389–394. Gerencser, G.A. & Hong, S.K., 1977. Comp. Biochem. Physiol., 58A, 275–280. Gerencser, G.A. & Lee, S.-H., 1985. Biochim. Biophys. Acta, 816, 415–417. Gerencser, G.A. & Loughlin, G.M., 1983. Comp. Biochem. Physiol., 74A, 701–704. Gerencser, G.A. & White, J.F., 1980. Am. J. Physiol., 239, R445-R449. Gev, S., Achituv, Y. & Susswein, A.J., 1984. J. exp. mar. Biol. Ecol., 74, 67–83. Ghiretti, F., Ghiretti-Magaldi, A. & Tosi, L., 1959. J. gen. Physiol., 42, 1185–1205. Gilboa-Garber, N., Mizrahi, L. & Susswein, A.J., 1984. Mar. Biol. Lett., 5, 105– 114. Gilboa-Garber, N., Susswein, A.J., Mizrahi, L. & Avichezer, D., 1985. FEBS Lett., 181, 267–270. González, A.G., Martin, J.D., Norte, M., Pérez, R. & Weyler, V., 1983a. Tetrah. Lett., 24, 1075–1076. González, A.G., Martin, J.D., Norte, M., Pérez, R. & Weyler, V., 1983b. Tetrah. Lett., 24, 847–848. Gopichand, Y., Schmitz, F.J. & Shelly, J., 1981. J. org. Chem., 46, 5192–5197. Grahame, J., 1973. J. anim. Ecol., 42, 391–403. Grigg, U.M., 1949. J. mar. biol. Ass. U.K., 28, 795–805. Gunatilaka, A.A.L., Gopichand, Y., Schmitz, F.J. & Djerassi, C., 1981. J. org. Chem., 46, 3860–3866. Hackney, A.G., 1944. Nautilus, 58, 56–64. Hadfield, M.G., 1975. Lab. Anim., 4, 17 only.

APLYSIA

241

Hadfield, M.G. & Switzer-Dunlap, M., 1984. In, The Mollusca, Vol. 7, edited by K.M.Wilbur, Academic Press, New York, pp. 209–350. Haefelfinger, H.R. & Kress, A., 1967. Rev. Suisse Zool., 74, 547–554. Halstead, B.W., 1965. Poisonous and Venomous Marine Animals of the World. U.S. Government Printing Office, Washington, D.C., 994 pp. Hamilton, P.V., 1979. Am. Zool., 19, 992 only. Hamilton, P.V., 1984. Anim. Behav., 32, 367–373. Hamilton, P.V., 1985. In, Migration: Mechanisms and Adaptive Significance, edited by M.A.Rankin, Suppl. to Contr. Mar. Sci., Vol. 27, pp. 212–226. Hamilton, P.V., 1986. Veliger, 28, 310–313. Hamilton, P.V. & Ambrose, III, H.W., 1975. Mar. Behav. Physiol., 3, 131–144. Hamilton, P.V. & Russell, B.J., 1982a. J. exp. mar. Biol. Ecol., 56, 145–152. Hamilton, P.V. & Russell, B.J., 1982b. J. exp. mar. Biol. Ecol., 56, 123–143. Hamilton, P.V., Russell, B.J. & Ambrose, H.W., 1982. Malacol. Rev., 15, 15–19. Hardy, Sir A., 1959. The Open Sea: Its Natural History. Part II. Fish and Fisheries. Collins, London, 322 pp. Heck, Jr, K.L. & Weinstein, M.P., 1978. Biotropica, 10, 78–79. Hening, W.A., Walters, E.T., Carew, T.J. & Kandel, E.R., 1979. Brain Res., 179, 231–253. Higgs, M.D. & Faulkner, D.J., 1982. Phytochemistry, 21, 789–791. Hirsch, H.R. & Peretz, B., 1984. Mech. Ageing Dev., 27, 43–62. Hirschfeld, D.R., Fenical, W., Lin, G.H.Y., Wing, R.M., Radlick, P. & Sims, J.J., 1973. J. Am. chem. Soc., 95, 4049–4050. Hollenbeak, K.H., Schmitz, F.J., Hossain, M.B. & van der Helm, D., 1979. Tetrahedron, 35, 541–545. Howells, H.H., 1942. Q. Jl microsc. Sci., 83, 357–397. Hoye, T.R., Caruso, A.J., Dellaria, Jr, J.F. & Kurth, M.J., 1982. J. Am. chem. Soc., 104, 6704–6709. Hsü, K.J., 1972. Scient. Am., 227, 27–36. Huebner, J.D. & Edwards, D.C., 1981. Mar. Biol., 61, 221–226. Hughes, G.M. & Tauc, L., 1962. J. exp. Biol., 39, 45–69. Hughes, R.N., 1971a. J. exp. mar. Biol. Ecol., 6, 167–178. Hughes, R.N., 1971b. Mar. Biol., 11, 12–22. Hughes, R.N., 1972. Oecologia (Berl.), 8, 356–370. Hughes, R.N. & Roberts, D.J., 1980. Oecologia (Berl.), 47, 130–136. Hunt, A.R., 1878. Devonsh. Ass. Adv. Sci. Lit. Art, Rep. Trans., 10, 611–617. Hyman, L.H., 1967. The Invertebrates. Vol. VI, Mollusca I. McGraw-Hill Book Company, New York, 792 pp. Imamura, P.M. & Ruveda, E.A., 1980. J. org. Chem., 45, 510–515. Imperato, F., Minale, L. & Riccio, R., 1977. Specialia, 33, 1273–1274. Ireland, C., Stallard, M.O., Faulkner, D.J., Finer, J. & Clardy, J., 1976. J org. Chem., 41, 2461–2465. Irie, T., Suzuki, M. & Hayakawa, Y., 1969. Bull. chem. Soc. Jpn, 42, 843–844. Irie, T., Suzuki, M., Kurosawa, E. & Masamune, T., 1966. Tetrah. Lett., 17, 1837– 1840. Irie, T., Suzuki, M. & Masamune, T., 1965a. Tetrah. Lett., 16, 1091–1099. Irie, T., Yasunari, Y., Suzuki, T., Imai, N., Kurosawa, E. & Masamune, T., 1965b. Tetrah. Lett., 40, 3619–3624. Jacklet, J.W., 1969. Science, 164, 562–563. Jacklet, J.W., 1972. J. comp. Physiol., 79, 325–341. Jacklet, J.W., 1974. J. comp. Physiol., 90, 33–45. Jacklet, J.W., 1976. In, Biological Rhythms in the Marine Environment, edited by P.J.DeCoursey, The Belle W.Baruch Library in Marine Science No. 4, University S.Carolina Press, Columbia, S.C., pp. 17–31. Jacklet, J.W., 1980. J. comp. Physiol., 136, 257–262. Jacklet, J.W., Alvarez, R. & Bernstein, B., 1972. J. Ultrastr. Res. 38, 246–261. Jahan-Parvar, B., 1969. Am. Zool., 9, 1097 only. Jahan-Parvar, B., 1970. Am. Zool., 10, 287 only.

242

THOMAS H.CAREFOOT

Jahan-Parwar, B., 1972a. Am. Zool., 12, 525–537. Jahan-Parwar, B., 1972b. Physiol., 15, 180 only. Jahan-Parwar, B., 1975. In, Olfaction and Taste V, edited by D.A.Denton & J.P. Coghlan, Proc. Fifth int. Symp., Univ. Melbourne, Aust., Academic Press, New York, pp. 133–140. Jahan-Parwar, B., 1976. Physiol., 19, 240 only. Jahan-Parwar, B. & Fredman, S.M., 1978b. J. Neurophysiol., 41, 600–608. Jahan-Parwar, B. & Fredman, S.M., 1978c. Comp. Biochem. Physiol., 60A, 459–465. Jahan-Parwar, B. & Fredman, S.M., 1978a. J. Neurophysiol., 41, 609–620. Jahan-Parwar, B. & Fredman, S.M., 1979a. Behav. neur. Biol., 27, 39–58. Jahan-Parwar, B. & Fredman, S.M., 1979b. Brain Res. Bull., 4, 407–420. Jahan-Parwar, B. & Fredman, S.M., 1980. Brain Res. Bull., 5, 169–177. Jahan-Parwar, B. & Miller, D.L., 1978. Biol. Bull. mar. biol. Lab., Woods Hole, 155, 447 only. Jahan-Parwar, B., Smith, M. & von Baumgarten, R., 1969. Am. J. Physiol., 216, 1246–1257. Jahan-Parwar, B., Wells, L.J. & Fredman, S.M., 1974. Proc. int. Union physiol. Sci., 26th int. Congr. physiol. Sci., X, New Delhi, p. 230 only. Janse, C., 1983. J. comp. Physiol., 150, 359–370. Johnson, P.T. & Chapman, F.A., 1970. J. Invert. Path., 16, 259–267. Jordan, H., 1901. Z. Biol., 41, 196–238. Jordan, H., 1917. Biol. Zbl., 37, 2–9. Jordan, W.P., Lickey, M.E. & Hiaasen, S.O., 1985. J. comp. Physiol. A, 156, 293– 303. Kamiya, H., Muramoto, K. & Ogata, K., 1984. Experientia, 40, 947–949. Kamiya, H. & Shimizu, Y., 1981. Bull. Jap. Soc. scient. Fish., 47, 255–259. Kandel, E.R., 1976. Cellular Basis of Behavior. An Introduction to Behavioral Neurobiology. W.H.Freeman & Co., San Francisco, 727 pp. Kandel, E.R., 1979. Behavioral Biology of Aplysia. A Contribution to the Comparative Study of Opisthobranch Molluscs. W.H.Freeman & Co., San Francisco, 463 pp. Kandel, P. & Capo, T.R., 1979. Veliger, 22, 194–198. Katayama, A., Ina, K., Nozaki, H. & Nakayama, M., 1982. Agric. biol. Chem., 46, 859–860. Kato, Y., 1984. Ph. D. thesis, University of the Ryūkyūs, Japan. Kato, Y. & Scheuer, P.J., 1974. J. Am. chem. Soc., 96, 2245–2246. Kaul, P.N. & Kulkarni, S.K., 1978. J. pharm. Sci., 67, 1293–1296. Kaul, P.N., Kulkarni, S.K. & Kurosawa, E., 1978a. J. Pharm. Pharmac., 30, 589– 590. Kaul, P.N., Kulkarni, S.K., Schmitz, F.J. & Hollenbeak, K.H., 1978b. In, Drugs and Food from the Sea. Myth or Reality?, edited by P.N.Kaul & C.J. Sindermann, University of Oklahoma, Norman, Oklahoma, pp. 99–106. Kay, E.A., 1964. Proc. malac. Soc. Land., 36, 173–190. Kay, E.A., 1979. Hawaiian Marine Shells. Reef and Shore Fauna of Hawaii. Section 4: Mollusca. Bernice P. Bishop Museum Spec. Publ. 64(4) , Bishop Museum Press, Honolulu, Hawaii, 652 pp. Kazlauskas, R., Murphy, P.T., Quinn, R.J. & Wells, R.J., 1976. Aust. J. Chem., 29, 2533–2539. Kempf, S.C., 1981. Mar. Ecol. Prog. Ser., 6, 61–65. Kennedy, G.Y. & Vevers, H.G., 1954. J. mar. biol. Ass. U.K., 33, 663–676. Kinnel, R., Duggan, A.J., Eisner, T. & Meinwald, J., 1977. Tetrah. Lett., 44, 3913– 3916. Kinnel, R.B., Dieter, R.K., Meinwald, J., Van Engen, D., Clardy, J., Eisner, T., Stallard, M.O. & Fenical, W., 1979. Proc. natn Acad. Sci. U.S.A., 76, 3576– 3579. Kittredge, J.S., Takahashi, F.T., Lindsey, J. & Lasker, R., 1974. Fish. Bull. NOAA, 72, 1–11. Koch, U.T. & Koester, J., 1982. J. comp. Physiol., 149, 31–42. Koch, U.T., Koester, J. & Weiss, K.R., 1984. J. Neurophysiol, 51, 126–135. Kofoed, L.H., 1975. J. exp. Mar. Biol. Ecol., 19, 243–256. Kohn, A.J., 1961. Am. Zool., 1, 291–308. Koningsor, Jr, R.L. & Hunsaker, II, D., 1971. Veliger, 13, 285–289.

APLYSIA

Koningsor, Jr, R.L., McLean, N. & Hunsaker, II, D., 1972. Comp. Biochem. Physiol., 43B, 237–240. Krakauer, J.M., 1969. M. Sc. thesis, University of Florida, Gainesville, Florida, 89 pp. Krakauer, J.M., 1971. Nautilus, 85, 37–38. Krakauer, J.M., 1974. Veliger, 16, 396–398. Krauhs, J.M., Long, J.L. & Baur, Jr, P.S., 1979. J. Protozool., 26, 43–46. Kriegstein, A.R., 1977a. Proc. natn Acad Sci. U.S.A., 74, 375–378. Kriegstein, A.R., 1977b. J. exp. Zool., 199, 275–288. Kriegstein, A.R., Castellucci, V. & Kandel, E.R., 1974. Proc. natn Acad. Sci. U.S.A., 71, 3654–3658. Kupfermann, I., 1965. Physiol., 8, 214 only. Kupfermann, I., 1967. Nature, Lond., 216, 814–815. Kupfermann, I., 1968. Physiol. Behav., 3, 179–181. Kupfermann, I., 1970. J. Neurophysiol., 33, 877–881. Kupfermann, I., 1972. Am. Zool., 12, 513–519. Kupfermann, I., 1974a. Behav. Biol., 10, 89–97. Kupfermann, I., 1974b. Behav. Biol., 10, 1–26. Kupfermann, I. & Carew, T.J., 1974. Behav. Biol., 12, 317–337. Kupfermann, I., Castellucci, V., Pinsker, H. & Kandel, E., 1970. Science, 167, 1743– 1745. Kupfermann, I. & Kandel, E.R., 1969. Science, 164, 847–850. Kupfermann, I. & Pinsker, H., 1968. Commun, behav. Biol., Part A, 2, 13–17. Kupfermann, I. & Weiss, K.R., 1976. J. gen. Physiol., 67, 113–123. Kupfermann, I. & Weiss, K.R., 1981. Behav. new. Biol., 32, 126–132. Kuslansky, B., Weiss, K.R. & Kupfermann, I., 1978. Behav. Biol., 23, 230–237. Lance, J.R., 1967. Veliger, 9, 412 only. Lance, J.R., 1971. Veliger, 14, 60–63. Langlais, P.J. & Blankenship, J.E., 1972. Fed. Proc., 31, 822 only. Larkin, S. & McFarland, D., 1978. Anim. Behav., 26, 1237–1246. Laverack, M.S., 1968. Oceanogr. Mar. Biol. Ann. Rev., 6, 249–324. Lawrence, J.M., 1975. Oceanogr. Mar. Biol. Ann. Rev., 13, 213–286. Lederer, E. & Huttrer, C., 1942. Trav. Membr. Soc. Chim. Biol., 24, 1055–1061. Lederhendler, I., Bell, L. & Tobach, E., 1975. Veliger, 17, 347–353. Lederhendler, I.I., 1977. Veliger, 20, 173–178. Lederhendler, I.I., Herriges, K. & Tobach, E., 1977. Anim. Learn. Behav., 5, 355– 358. Lederhendler, I.I. & Tobach, E., 1977. Nature, Lond., 270, 238–239. Leighton, D.L., 1966. Pacif. Sci., 20, 104–113. Lickey, M.E., 1968. J. comp. physiol. Psychol., 66, 712–718. Lickey, M.E., 1969. J. comp. physiol. Psychol., 68, 9–17. Lickey, M.E. & Berry, R.W., 1966. Physiol., 9, 230 only. Lickey, M.E., Hudson, D.J. & Hiaasen, S.O., 1983. J. comp. Physiol., 153, 133– 143. Lickey, M.E. & Wozniak, J.A., 1979. J. comp. Physiol., 131, 169–177. Lickey, M.E., Wozniak, J.A., Block, G.D., Hudson, D.J. & Augter, G.K., 1977. J. comp. Physiol., 118, 121–143. Ligman, S.H. & Brownell, P.H., 1985. J. comp. Physiol. A, 157, 31–37. Littler, M.M., Littler, D.S., Blair, S.M. & Norris, J.N., 1985. Science, 277, 57–59. Lo Bianco, S., 1888. Mitt. zool. Stn Neapel, 8, 385–440. Lo Bianco, S., 1909. Mitt. zool. Stn Neapel, 19, 513–761. Lombardini, J.B., Pang, K.T. & Griffith, R.W., 1979. Calif. Acad. Sci., Occ. Papers, 134, 160–169. Longley, A.J. & Longley, R.D., 1984. Can. J. Zool., 62, 8–14. Lu, C., Strumwasser, F. & Gilliam, J.J., 1966. Calif. Inst. Tech. Biol. Div. Ann. Rep., p. 153 only. Lukowiak, K., 1977. Brain Res., 132, 553–557. Lukowiak, K., 1979. Can. J. Physiol. Pharmacol., 57, 987–997.

243

244

THOMAS H.CAREFOOT

Lukowiak, K., 1980. J. Neurobiol., 11, 591–611. Lukowiak, K. & Freedman, L., 1983. Can. J. Physiol. Pharmacol, 61, 743–748. Lukowiak, K. & Jacklet, J.W., 1972a. Science, 178, 1306–1308. Lukowiak, K. & Jacklet, J.W., 1972b. Fed. Proc., 31, 405 only. Lukowiak, K. & Peretz, B., 1977. J. comp. Physiol., 117, 219–244. Lukowiak, K. & Peretz, B., 1980. J. Neurobiol., 11, 425–433. Lukowiak, K. & Sahley, C., 1981. Science, 212, 1516–1518. Macé, A.-M., 1981. Téthys, 10, 117–120. MacFarland, F.M., 1966. Mem. Calif. Acad. Sci., No. 6, 546 pp. MacGinitie, G.E., 1934. Biol. Bull. mar. biol. Lab., Woods Hole, 67, 300–303. MacGinitie, G.E., 1935. Am. Midl. Nat., 16, 629–765. MacGinitie, G.E. & MacGinitie, N., 1968. Natural History of Marine Animals. McGraw-Hill Book Company, New York, 523 pp. MacMunn, C.A., 1899. J. Physiol., 24, 1–10. Macnae, W., 1955. Ann. Natal Mus., 13, 223–241. Macnae, W., 1957. S. Afr. J. Sci., 54, 289–292. Magnus, O., 1555. Historia de Gentibus Septentrionalibus. Rosenkilde and Bagger, Copenhagen, 1972, 815 pp. Marcus, E., 1958. Bolm Inst. oceanogr. São Paulo, 7, 31–78. Marcus, E., 1961. Veliger, 3 (suppl.), 1–84. Marcus, E. & Marcus, E., 1955. Bolm Inst. oceanogr. São Paulo, 6, 3–48. Marcus, E. & Marcus, E., 1958. Bolm Inst. oceanogr. São Paulo, 8, 3–22. Marcus, E. & Marcus, E., 1960. Bull. mar. Sci. Gulf Caribb., 10, 141–203. Marcus, E. & Marcus, E., 1962. Bull. mar. Sci. Gulf Caribb., 12, 450–488. Marcus, E.d.B.-R., 1972. Bull. mar. Sci., 22, 841–874. Marcus, E.d.B.-R., 1976. Stud, neotrop. Faun. Environ., 11, 119–150. Marcus, E.d.B.-R., 1979. Ann. Inst. Oceanogr., 55, 131–137. Marcus, E.d.B.-R. & Hughes, H.P.I., 1974. Bull. mar. Sci., 24, 498–532. Matsuda, H., Tomiie, Y., Yamamura, S. & Hirata, Y., 1967. Chem. Comm., 898– 899. McDonald, F.J., Campbell, D.C., Vanderah, D.J., Schmitz, F.J., Washecheck, D.M., Burks, J.E. & van der Helm, D., 1975. J. org. Chem., 40, 665–666. McKay, D., Jenkin, C.R. & Rowley, D., 1969. Aust. J. exp. Biol. med. Sci., 47, 125– 134. Merriman, D., 1937. Nautilus, 50, 95–97. Midland, S.L., Wing, R.M. & Sims, J.J., 1983. J. org. Chem., 48, 1906–1909. Mileikovsky, S.A., 1974. Mar. Biol., 26, 303–311. Miller, M.C., 1960. Proc. malac. Soc. Land., 34, 165–167. Minale, L. & Riccio, R., 1976. Tetrah. Lett., 31, 2711–2714. Moore, D.R., 1961. Gulf Res. Rep., 1, 1–58. Moore, R.E., 1982. Pure appl. Chem., 54, 1919–1934. Moritz, C.E., 1936. Carnegie Inst. Wash. Year Book, No. 35, 90 only. Morton, J. & Miller, M., 1968. The New Zealand Sea Shore. Collins, London-Auckland, 638 pp. Moseley, H.N., 1877. Q. Jl microsc. Sci., 17, 1–23. Mynderse, J.S. & Faulkner, D.J., 1978. Phytochem., 17, 237–240. Mynderse, J.S., Moore, R.E., Kashiwagi, M. & Norton, T.R., 1977. Science, 196, 538–540. Nagle, G.T., Painter, S.D., Kelner, K.L. & Blankenship, J.E., 1985. J. comp. Physiol. B, 156, 43–55. Neck, R.W., 1976. Veliger, 19, 107 only. Neu, W., 1932. Z. vergl. Physiol., 18, 244–254. Niell, F.X., 1977. Malacologia, 16, 207–209. Nishibori, K., 1960. Publs Seto mar. biol. Lab., 8, 327–335. Nishiwaki, S., Ueda, H. & Makioka, T., 1975. Mar. Biol., 32, 389–395.

APLYSIA

245

Oren, O.H., 1970. Underw. Sci. Technol. J., 2, 222–229. Ostergaard, J.M. 1950. Pacif. Sci., 4, 75–115. Otsuka, C., Oliver, L., Rouger, Y. & Tobach, E., 1981. Hydrobiol., 83, 239–240. Otsuka, C., Rouger, Y. & Tobach, E., 1980. Veliger, 23, 159–162. Paige, J.A., 1981. In, Advances in Invertebrate Reproduction, edited by W.H.Clark & T.S.Adams, Proc. Second int. Symp. of the int. Soc. Invert. Repr. (ISIR), Elsevier/North Holland, New York, p. 387 only. Paine, R.T., 1963. Veliger, 6, 1–9. Paine, R.T., 1966. Limnol. Oceanogr., 11, 126–129. Paine, R.T. & Vadas, R.L., 1969. Limnol. Oceanogr., 14, 710–719. Papka, R., Peretz, B., Tudor, J. & Becker, J., 1981. J. Neurobiol, 12, 455–468. Parker, G.H., 1917. J. exp. Zool., 24, 139–145. Parry, G.D., 1982. Mar. Biol., 67, 267–282. Pauley, G.B., 1971. Proc. natn Shellfish Assoc., 61, 11–12. Pauley, G.B., 1974. In, Biomedical Perspectives of Agglutinins of Invertebrate and Plant Origins, edited by E.Cohen, Ann. N.Y. Acad. Sci., 234, 145–160. Pauley, G.B., Granger, G.A. & Krassner, S.M., 1971a. J. Invert. Path., 18, 207– 218. Pauley, G.B., Krassner, S.M. & Chapman, F.A., 1971b. J. Invert. Path., 18, 227– 239. Peretz, B., 1970. Science, 169, 379–381. Peretz, B. & Adkins, L., 1982. Biol. Bull. mar. biol. Lab., Woods Hole, 162, 333–344. Peretz, B. & Howieson, D.B., 1973. J. comp. Physiol., 84, 1–18. Peretz, B., Jacklet, J.W. & Lukowiak, K., 1976. Science, 191, 396–399. Peretz, B. & Lukowiak, K.D., 1975. J. comp. Physiol., 103, 1–17. Perron, F.E., 1982. Mar. Biol., 68, 161–167. Pettit, G.R., Herald, C.L., Allen, M.S., Von Dreele, R.B., Vanell, L.D., Kao, J.P.Y. & Blake, W., 1977. J. Am. chem. Soc., 99, 262–263. Pettit, G.R., Herald, C.L., Einck, J.J., Vanell, L.D., Brown, P. & Gust, D., 1978. J. org. Chem., 43, 4685–4686. Pilsbry, H., 1951. Not. Nat. (Philadelphia), 240, 1–6. Pinsker, H., Kupfermann, I., Castellucci, V. & Kandel, E., 1970. Science, 167, 1740– 1742. Pinsker, H., Redmann, G., von der Porten, K., Heine, D. & Rothman, B., 1978. Fed. Proc., 37, 527 only. Pinsker, H.M. & Dudek, F.E., 1977. Science, 197, 490–493. Pinsker, H.M., Feinstein, R. & Gooden, B.A., 1974. Fed. Proc., 33, 361 only. Pinsker, H.M., Hening, W.A., Carew, T.J. & Kandel, E.R., 1973. Science, 182, 1039–1042. Pinsker, H.M. & Parsons, D.W., 1985. J. comp. Physiol. B, 156, 21–27. Poizat, C. & Vicente, N., 1977. Bull. Mus. natl Hist, not., Ser. 3, 34, 2–15. Preston, R.J. & Lee, R.M., 1973. J. comp. physiol. Psych., 82, 368–381. Pruvot-Fol, A., 1954. Mollusques Opisthobranches. Faune de France, Paris, No. 58, 460 pp. Rattan, K.S. & Peretz, B., 1981. J. Neurobiol., 12, 469–478. Rey, J.R. & Stoner, A.W., 1984. Estuaries, 7, 158–164. Ricketts, E.F. & Calvin, J., 1968. Between Pacific Tides. Stanford University Press, Berkeley, CA, 4th edition, 614 pp. Ronald, R.C., Gewali, M.B. & Ronald, B.P., 1980. J. org. Chem., 45, 2224–2229. Rosen, S.C., Weiss, K.R. & Kupfermann, I., 1982. Behav. new. Biol., 35, 56–63. Rothman, B.S., Weir, G. & Dudek, F.E., 1983. Gen. comp. Endocrinol., 52, 134– 141. Rüdiger, W., 1967a. Hoppe-Seyler’s Z. Physiol. Chem., 348, 129–138. Rüdiger, W., 1967b. Hoppe-Seyler’s Z. Physiol. Chem., 348, 1554 only. Rüdiger, W., 1968. Umschau, 68, 405–406. Rüdiger, W., Carra, P.O. & hEocha, C.O., 1967. Nature, Land., 215, 1477–1478. Russell, E., 1966. Pacif. Sci., 20, 452–460. Rutowski, R.L., 1983. Biol. Bull. mar. biol. Lab., Woods Hole, 165, 276–285. Ryther, J.H., 1954. Ecology, 35, 522–533.

246

THOMAS H.CAREFOOT

Saito, Y. & Nakamura, N., 1961. Bull. Jap. Soc. scient. Fish., 27, 395–400. Sakata, K., Tsuge, M. & Ina, K., 1986. Mar. Biol., 91, 509–511. Sakata, K., Tsuge, M., Kamiya, Y. & Ina, K., 1985. Agric. biol. Chem., 49, 1905– 1907. Sanford, S.N.F., 1922. Nautilus, 35, 80–82. Sarver, D.J., 1978. Ph. D. thesis, University of Hawaii, Honolulu, Hawaii, 151 pp. Sarver, D.J., 1979. Mar. Biol. 54, 353–361. Saunders, A.M.C. & Poole, M., 1910. Q. Jl microsc. Sci., 55, 497–539. Sawaya, P. & Leahy, W.M., 1971. Bol. Zool. Biol. mar., N.S., 28, 1–17. Scheller, R.H., Jackson, J.F., McAllister, L.B., Schwartz, J.H., Kandel, E.R. & Axel, R., 1982. Cell, 28, 707–719. Scheuer, P.J., 1971. Naturwissenschaften, 58, 549–554. Scheuer, P.J., 1975. Lloydia (Cincinnati), 38, 1–7. Scheuer, P.J., 1977. Accounts chem. Res., 10, 33–39. Schmekel, L., 1971. Z. Morph. Ökol. Tiers, 69, 115–183. Schmitz, F.J., Gopichand, Y., Michaud, D.P., Prasad, R.S., Remaley, S., Hossain, M.B., Rahman, A., Sengupta, P.K. & van der Helm, D., 1981. Pure appl. Chem., 51, 853–865. Schmitz, F.J., Hollenbeak, K.H., Carter, D.C., Hossain, M.B. & van der Helm, D., 1979. J. org. Chem., 44, 2445–2447. Schmitz, F.J., Hollenbeak, K.H. & Vanderah, D.J., 1978a. Tetrahedron, 34, 2719– 2722. Schmitz, F.J. & McDonald, F.J., 1974. Tetrahedron, 29, 2541–2544. Schmitz, F.J., McDonald, F.J. & Vanderah, D.J., 1978b. J. org. Chem., 43, 4220– 4225. Schmitz, F.J., Michaud, D.P. & Hollenbeak, K.H., 1980. J. org. Chem., 45, 1525– 1528. Schmitz, F.J., Michaud, D.P. & Schmidt, P.G., 1982. J. Am. chem. Soc., 104, 6415– 6423. Schreiber, G., 1932. Pubbl. Staz. zool. Napoli, 12, 291–321. Schulte, G.R., Chung, M.C.H. & Scheuer, P.J., 1981. J. org. Chem., 46, 3870– 3873 . Schwarz, M. & Susswein, A.J., 1982. In, Behavioral Models and the Analysis of Drug Action, edited by M.Y.Spiegelstein & A.Levy, Proc. 27th OHOLO Conf., pp. 479–486. Schwarz, M. & Susswein, A.J., 1984. Brain Res., 294, 363–366. Shapiro, E., Koester, J. & Byrne, J.H., 1979. J. Neurophysiol., 42, 1223–1232. Shieh, H.-M. & Prestwich, G.D., 1982. Tetrah. Lett., 23, 4643–4646. Sigel, M.M., Wellham, L.L., Lichter, W., Dudeck, L.E., Gargus, J.L. & Lucas, A.H., 1970. In, Food-drugs from the Sea, edited by H.W.Youngken, Jr, Marine Technology Society, Washington, D.C., pp. 281–294. Sims, J.J., Fenical, W., Wing, R.M. & Radlick, P., 1971. J. Am. chem. Soc., 93, 3774–3775. Sims, J.J., Fenical, W., Wing, R.M. & Radlick, P., 1972. Tetrahedron Lett., No. 3, 195–198. Smith, S.T. & Carefoot, T.H., 1967. Nature, Land., 215, 652–653. Sousa, W.P., 1979. Ecol. Monogr., 49, 227–254. Stallard, M.O. & Faulkner, D.J., 1974a. Comp. Biochem. Physiol., 49B, 25–35. Stallard, M.O. & Faulkner, D.J., 1974b. Comp. Biochem. Physiol., 49B, 37–41. Stallard, M.O., Fenical, W. & Kittredge, J.S., 1978. Tetrahedron, 34, 2077–2081. Stehouwer, H., 1952. Arch. Néerl. Zool., 10, 161–170. Stickle, W.B., 1985. In, Marine Biology of Polar Regions and Effects of Stress on Marine Organisms, Proc. 18th Europ. Mar. Biol. Symp., edited by J.S.Gray & M.E.Christiansen, John Wiley and Sons, Chichester, U.K., pp. 601–616. Stinnakre, J. & Tauc, L., 1969. J. exp. Biol., 51, 347–361. Stone, B.A. & Morton, J.E., 1958. Proc. malac. Soc. Lond., 33, 127–141. Streit, B., 1976. Oecologia (Berl.), 22, 261–273. Strenth, N.E. & Blankenship, J.E., 1976. Bull. Am. malac. Union, 42, 48 only. Strenth, N.E. & Blankenship, J.E., 1977. Veliger, 20, 98–100. Strenth, N.E. & Blankenship, J.E., 1978a. Veliger, 21, 99–103. Strenth, N.E. & Blankenship, J.E., 1978b. Bull. Mar. Sci., 28, 249–254.

APLYSIA

247

Strumwasser, F., 1967. In, The Neurosciences. A study program, edited by G.C. Quarton et al., Rockefeller University Press, New York, pp. 516–528. Strumwasser, F., 1973. Physiol, 16, 9–42. Strumwasser, F., Jacklet, J.W. & Alvarez, R.B., 1969. Comp. Biochem. Physiol., 29, 197–206. Susswein, A., Kupfermann, I. & Weiss, K.R., 1976a. Neurosci. Abstr., 2, 336 only. Susswein, A.J., 1984. Behav. new. Biol., 42, 127–133. Susswein, A.J., Gev, S., Achituv, Y. & Markovich, S., 1984a. Behav. neur. Biol., 41, 7–22. Susswein, A.J., Gev, S., Feldman, E. & Markovich, S., 1983. Behav. neur. Biol., 39, 203–220. Susswein, A.J. & Kupfermann, I., 1974. Soc. Neurosci. 4th Ann. Meet. St. Louis, Missouri, p. 445 only. Susswein, A.J. & Kupfermann, I., 1975a. Behav. Biol., 13, 203–209. Susswein, A.J. & Kupfermann, I., 1975b. J. comp. Physiol., 101, 309–328. Susswein, A.J., Kupfermann, I. & Weiss, K.R., 1976b. J. comp. Physiol., 108, 75– 96. Susswein, A.J. & Markovich, S., 1983. Israel J. Zool., 32, 103–112. Susswein, A.J. & Schwarz, M., 1983. Behav. neur. Biol., 39, 1–6. Susswein, A.J., Weiss, K.R. & Kupfermann, I., 1978. J. comp. Physiol., 123, 31–41. Susswein, A.J., Weiss, K.R. & Kupfermann, I., 1984b. Behav. neur. Biol., 41, 90– 95. Swennen, C., 1961. Zoöl. Meded., Leiden, 38, 41–75. Switzer-Dunlap, M., 1978. In, Settlement and Metamorphosis of Marine Invertebrate Larvae, edited by F.-S.Chia & M.E.Rice, Elsevier/North-Holland Press, Amsterdam, pp. 197–206. Switzer-Dunlap, M. & Hadfield, M.G., 1977. J. exp. mar. Biol. Ecol, 29, 245–261. Switzer-Dunlap, M. & Hadfield, M.G., 1979. In, Reproductive Ecology of Marine Invertebrates, edited by S.E.Stancyk, Belle W.Baruch Library in Marine Sciences, No. 9, University of South Carolina Press, Columbia, S.C., pp. 199– 210. Switzer-Dunlap, M. & Hadfield, M.G., 1981. In, Marine Invertebrates, Report of the Committee on Marine Invertebrates, Institute of Laboratory Animal Resources, National Research Council, Nat. Academy Press, Washington, D.C., pp. 199–216. Switzer-Dunlap, M., Meyers-Schulte, K. & Gardner, E.A., 1984. Int. J. Invert. Repr. Devel., 7, 217–225. Tanaka, A., Otsuka, S. & Yamashita, K., 1984. Agric. biol. Chem., 48, 2535–2540. Tanaka, T. & Toyama, Y., 1959. J. Chem. Soc. Japan, Pure Chem. Sec. (Nippon Kagakukai Zasshi), 80, 1329–1332. Thompson, T.E., 1960, J. mar. biol. Ass. U.K., 39, 123–134. Thompson, T.E., 1976. Biology of Opisthobranch Molluscs, Vol. 1. Ray Soc., Land., No. 151, 206 pp. Thompson, T.E. & Bebbington, A., 1969. Malacologia, 7, 347–380. Thorson, G., 1946. Medd. Komm. Danm. Fisk. Havunders., 4, (1), 1–523. Tobach, E., 1978. Veliger, 20, 359–360. Tobach, E., Gold, P. & Ziegler, A., 1965. Veliger, 8, 16–18. Toevs, L., 1966. Calif. Inst. Technol., Biol. Div., Ann. Rep., p. 154 only. Tritt, S.H. & Byrne, J.H., 1980. J. Neurophysiol., 43, 581–594. Tritt, S.H. & Byrne, J.H., 1982. J. Neurophysiol., 48, 1347–1361. Tritt, S.H., Zigmond, M. & Byrne, J.H., 1980. Neurosc. Abstr., 6, 703 only. Tunnell, Jr, J.W. & Chaney, A.H., 1970. Contr. mar. Sci., 15, 193–203. Usuki, I. 1970a. Mem. Sado Mus., Niigata Pref., 19, 1–10. Usuki, I. 1970b. Sci. Rep., Niigata Univ., Ser. D. (Biol.), 7, 91–105. Usuki, I. 1979. J. gen. Educ. Depart., Niigata Univ., 9, 81–84. Usuki, I. 1981a. Venus, 39, 212–223. Usuki, I. 1981b. Venus, 40, 41–49. Vanderah, D.J. & Schmitz, F.J., 1976. J. org. Chem., 41, 3480–3481. Van Weel, P.B., 1957. Z. vergl. Physiol., 39, 492–506. Vicente, N., 1962. Recl. Trav. Stn mar. Endoume, Bull. No. 26, Fasc. 41, 303–305. Vicente, N. & Poizat, C., 1977. Bull. Mus. natn. Hist, not., Paris, Sér. 3, No. 34, 19– 26.

248

THOMAS H.CAREFOOT

Vitalis, T.Z., 1981. B. Sc. thesis, University of British Columbia, Vancouver, B.C., 39pp. Von der Porten, K., Parsons, D.W., Rothman, B.S. & Pinsker, H., 1982. Behav. new. Biol., 36, 1–23. Von der Porten, K., Redmann, G., Rothman, B. & Pinsker, H., 1980. J. exp. Biol., 84, 245–257. Von Dreele, R.B. & Kao, J.P.Y., 1980. Acta Cryst., B36, 2695–2698. Wachtel, H. & Impelman, D., 1973. Fed. Proc., 32, 368 only. Walford, L.A., 1946. Biol. Bull. mar. biol. Lab., Woods Hole, 90, 141–147. Walters, E.T., Carew, T.J. & Kandel, E.R., 1978. Neurosc. Abstr., 4, 209 only. Walters, E.T., Carew, T.J. & Kandel, E.R., 1979a. Neurosc. Abstr., 5, 264 only. Walters, E.T., Carew, T.J. & Kandel, E.R., 1979b. Proc. natn Acad. Sci. U.S.A., 76, 6675–6679. Walters, E.T., Carew, T.J. & Kandel, E.R., 1981. Science, 211, 504–506. Ward, J., 1967. Bull. mar. Sci., 17, 299–318. Warmke, G.L. & Abbott, R.T., 1961. Caribbean Seashells A Guide to the Marine Mollusks of Puerto Rico and other West Indian Islands, Bermuda and the Lower Florida Keys. Livingston Publishing Company, Narberth, Penn., 348 pp. Warren, C.E. & Davis, G.E., 1967. In, The Biological Basis of Freshwater Fish Production, edited by S.D.Gerking, Blackwell Sci. Publ., Oxford, pp. 175–214. Waser, P.M., 1968. Comp. Biochem. Physiol., 27, 339–347. Watson, M., 1973. Toxicon, 11, 259–267. Watson, M. & Rayner, M.D., 1973. Toxicon, 11, 269–276. Weiss, K.R., Koch, U.T., Koester, J., Mandelbaum, D.E. & Kupfermann, I., 1981. In, Neurobiology of Invertebrates. Adv. Physiol. Sci., Vol. 23, edited by J.Salanki, Pergamon Press, Oxford, pp. 305–344. Welch, H.E., 1968. Ecology, 49, 755–759. Wells, D.W. & Hill, R.B., 1980. Comp. Biochem. Physiol., 67C, 97–106. Wells, D.W. & Hill, R.B., 1985. Comp. Biochem. Physiol., 80C, 337–345. Wells, L.J., Jahan-Parwar, B. & Fredman, S.M., 1974. Physiol., 17, 357 only. White, J.D., Nishiguchi, T. & Skeean, R.W., 1982. J. Am. chem. Soc., 104, 3923– 3928. Wightman, J.A., 1981. Oecologia (Berl.), 50, 166–169. Willan, R. & Morton, J., 1984. In, Marine Molluscs, Part 2, Leigh Marine Laboratory, University of Auckland, pp. 27–34. Willan, R.C., 1979. Ph. D. thesis, University of Auckland, Auckland, New Zealand, 198 pp. Willey, J.C., Moser, Jr, C.E. & Harris, C.C., 1984. Cell Biol. Toxicol., 1, 145–154. Winkler, L.R., 1955. Bull. Sth. Calif. Acad. Sci., 54, 5–7. Winkler, L.R., 1958a. Pacif. Sci., 12, 348–349. Winkler, L.R., 1958b. Bull. Sth. Calif. Acad. Sci., 57, 105–106. Winkler, L.R., 1958c. Bull. Sth. Calif. Acad. Sci., 57, 106–107. Winkler, L.R., 1959a. Pacif. Sci., 13, 357–361. Winkler, L.R., 1959b. Pacif. Sci., 13, 63–66. Winkler, L.R., 1959c. Bull. Sth. Calif. Acad. Sci., 58, 8–10. Winkler, L.R., 1959d. Nautilus, 72, 73–75. Winkler, L.R., 1960. Pacif. Sci., 14, 304–307. Winkler, L.R., 1961. Pacif. Sci., 15, 211–214. Winkler, L.R., 1969. Veliger, 11, 268–271. Winkler, L.R. & Dawson, E.Y., 1963. Pacif. Sci., 17, 102–105. Winkler, L.R. & Tilton, B.E., 1961. Pacif. Sci., 15, 557–560. Winkler, L.R. & Tilton, B.E., 1962. Pacif. Sci., 16, 286–290. Winkler, L.R., Tilton, B.E. & Hardinge, M.G., 1962. Arch. int. Pharmacodyn., 137, 76–83. Wolff, H.G., 1973. Mar. Behav. Physiol., 1, 361–373. Wright, D.L., 1960. Ph. D. thesis, University of California, Los Angeles, 104 pp. Wright, H.O., 1960. Veliger, 2, 63–64.

APLYSIA

249

Yamada, K., Yazawa, H., Toda, M. & Hirata, Y., 1968. Chem. Comm., No. 22, 1432 only. Yamamura, S. & Hirata, Y., 1963. Tetrahedron, 19, 1485–1496. Yamamura, S. & Hirata, Y., 1971. Bull. chem. Soc. Jap., 44, 2560–2562. Yamamura, S. & Terada, Y., 1977. Tetrah. Lett., No. 25, 2171–2172. Yamazaki, M., Esumi-Kurisu, M., Mizuno, D., Ogata, K. & Kamiya, H., 1983a. Gann, 74, 405–411. Yamazaki, M., Ikenami, M., Komano, H., Tsunawaki, S., Kamiya, H., Natori, S. & Mizuno, D., 1983b. Gann, 74, 576–583. Yamazaki, M., Kisugi, J., Ikenami, M., Kamiya, H. & Mizuno, D., 1984. Gann, 75, 269–274. Yamazaki, M., Kisugi, J., Kimura, K., Kamiya, H. & Mizuno, D., 1985. FEBS Lett., 185, 295–298. Yonge, C.M., 1947. Phil. Trans. Roy. Soc. Land., Ser. B, 232, 443–518. Yonge, C.M., 1949. The Sea Shore. Collins, London, 311 pp. Zinn. D.J., 1950. Nautilus. 64, 40–47.

Oceanogr. Mar. Biol. Ann. Rev., 1987, 25, 285–351 Margaret Barnes, Ed. Aberdeen University Press

A REVIEW OF THE COMPARATIVE ANATOMY OF THE MALES IN CIRRIPEDES WALTRAUD KLEPAL Institut für Zoologie der Universität Wien, Althanstraße 14, A-1090 Wien, Österreich

INTRODUCTION Secondary sexual dimorphism as distinct from the primary sexual dimorphism (which is expressed in the gonads, the germ cells and the copulatory organs) is a common feature in the animal kingdom. Usually there is a difference in size and in the armament in males and females. Often the females are larger as in molluscs, spiders, insects, fishes, frogs, snakes, turtles, and birds of prey. On the other hand, the males are more imposing and larger, e.g. in gallinaceous birds, ostriches, iguanas, beasts of prey, deer, horses and apes. Sexual dimorphism is also a common feature in crustaceans. It is found in the Conchostraca, Cladocera, and Anostraca and is expressed as a difference in size and shape of males and females and a difference in the formation of their appendages. In the Ostracoda it affects the carapace. In the epicaridean isopods the males are small and the females are large and sac-like without any legs. In other crustaceans the males are smaller than the females, e.g. in Gnathia. Advanced types of crustaceans like the Malacostraca show only slight differences in the appendages of males and females (Newman, Zullo & Withers, 1969). The males in the cirripedes are always considerably smaller than the females or hermaphrodites. This dwarfism is an extreme case of secondary sexual dimorphism. It is found in some parasitic crustaceans as the Rhizocephala and in cirripedes which live at the edge of some spatial horizontal or vertical distribution. The males of these cirripedes are generally referred to as dwarf males. Well-known dwarf males in other groups of animals are those of the worm Bonellia viridis and of the deep-sea fish Edriolychnus sp. In cirripedes Darwin (1851) distinguished between dwarf males, associated with a female and complemental males associated with a hermaphroditic cirripede. Crisp (1983) introduced the term “apertural male” for the males which are attached around the opercular opening of Chelonobia. These males are potential hermaphrodites which stopped in their development at the protandric stage. Males, as distinct from females and/or hermaphrodites are found in all three orders of cirripedes: Acrothoracica, Thoracica, and Rhizocephala. The Acrothoracica and the Rhizocephala are obligatory gonochoristic. Within the Thoracica both dwarf- and complemental males have been found. Whilst in the Lepadomorpha either dwarf- or complemental males are common, in the Balanomorpha so far only complemental males are known. Up to now no separate males were found in the Verrucomorpha. A number of papers have dealt with the question of sexuality in cirripedes and also with complemental males. But so far no comprehensive study has been made. This paper is an attempt to consider and to compare the main features of males in the Rhizocephala, Acrothoracica, and Thoracica as far as they are known up to now and as far as they are described in any detail. An attempt is made to show the major trends rather than to describe the male of every single species. Thus, all males will be omitted that are not relevant to the considerations in this paper.

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

251

First the descriptions of the males will be reviewed. This will be followed by an effort to elucidate the trends of evolution within the single orders and then an attempt to show the general trend of evolution of the males and of sexuality within the cirripedes in general. DESCRIPTIONS OF THE MALES ACROTHORACICA The Acrothoracica are free-living burrowing cirripedes with separate sexes in all known species. The females are about ten times as big as the males. In many species only the females have been described and the males are not known. The metamorphosis is very complex and the morphology of the males is very variable. The male may be even smaller than the cypris larva. It is motile. With the exception of Berndtia it is buried within a pocket of the female’s mantle tissue in the immediate area of the ovary. Thus its sexual development is synchronized with that of the female. It is even assumed that the male matures under some chemical mediation of the female. The embedding of the male in this way was observed in Cryptophialus melampygos, Lithoglyptes indicus, Trypetesa lateralis and Trypetesa lampas (Tomlinson, 1969). The anatomy of the males has been described by only a few authors. Three types of males may be distinguished in the Lithoglyptidae, supposedly the most primitive family of Acrothoracica (Tomlinson & Newman, 1960; this statement is based on characters of the females). (1) Pear-shaped males with a penis, homologous to the thorax of the cypris. These males are similar in most species of Lithoglyptes and in Weltneria. There are no special differentiations in the larval antennae for the attachment. (2) Males of variable shape. These are often polygonal, they have a penis and a specialized attachment organ. The latter is either a long and thin peduncle as in Kochlorine hamata or it is an orchid lobe as in K. floridana. (3) Males without a penis as in K. bocqueti or perhaps in K. ulula (Tomlinson, 1973). There is a great similarity between the adult male of K. ulula and the juvenile form of K. bocqueti. In the adult male of K. bocqueti there is no penis, but there is a specialized attachment organ which assures the attachment of the male at the side of the female. The genus Weltneria is supposed to be the most primitive one amongst the Acrothoracica. Newman (1971) described W. hessleri but he did not observe any males. He himself calls this unfortunate since Tomlinson (1969) suggested that the males in the genus have a more or less characteristic cyprid-like form with a blunt posterior end. In 1974 Newman described the male of W. exargilla; the female of the species is mm big. They have a single testis, a seminal vesicle, very similar to W. hessleri. The males are about a vas deferens and a penis. The penis seems to be in a channel, which is supposedly a rudiment of the mantle cavity. At the base of the penis there are numerous muscles. Turquier (1985b) studied the larvae of W. zibrowii very carefully as well as the metamorphosis from the cypris to the male. Three major events may be distinguished in the metamorphosis of the larvae. The beginning of the transformation is marked by some cement being secreted. One major step in the metamorphosis is the formation of a rudimentary peduncle. Then the muscles of the antennules and the cement glands degenerate. Remnants of these two organs may be seen during the whole metamorphosis. The second step is the condensation and rotation of the visceral mass. The front head region of the larva becomes the rudimentary peduncle and the visceral mass carries out a characteristic morphogenetic movement in the sagittal plane

252

WALTRAUD KLEPAL

around the tendon of the adductor muscle of the valves. Thus the mouthcone is moved backwards and the cerebral ganglion is moved posteroventrally. The lateral eyes are lost and the optic ganglion connected to the protocerebrum degenerates. The thorax withdraws into the mantle cavity and its muscles dedifferentiate. The nervous system becomes simpler and its volume decreases. The visceral mass is bent to a U-shape. The mouthcone, the cerebral ganglion and the “yellow organ” are moved towards the sternite of the cypris, thus marking the rostral face of the future male. The “yellow organ” may be a stomach rudiment with yolk-absorbing function as in Pollicipes (Batham, 1946) or derived from degenerating endodermal cells (Turquier, 1971). The third important step in the course of metamorphosis is the organogenesis of the genital apparatus. The posterior mantle cavity is now closed. The gonad on the ventral side of the larva is removed little by little from the thorax and totally occupies the head metameres. In the juvenile male the gonad takes up the central space of the body. Now there is intensive sperm production and thus the seminal vesicle is enlarged. The thorax grows vastly and re-differentiates. The thoracal muscles form a double layer of longitudinal and circular strands. The adult male of Weltneria zibrowii (Fig. 1A) has an ovoid body of three distinct regions. The long. The body proper is globular, about mm. In the rudimentary peduncle is about 50 posterior part there are the testes and at the same time it is something like a sheath for the penis. In the testes the sperms develop synchronously and they accumulate in the central part of the gonad. The pearshaped vesicle is curved to an S backwards and towards the rostral face. Then it leads into the ductus ejaculatorius and further into the penis. In the adult male the nervous system is along the ventral side of the animal. The nauplius eye and the “yellow organ” may be easily seen but the mouthcone and the first pair of cirri have disappeared. mm without antennules (Berndt, 1907). It resembles a degenerate sac In W. spinosa the male is in which a long annulated penis may be seen curling inside the body. The body is streamlined and has a mm small, deep purple nauplius eye. A “yellow organ” was not detected. In W. hirsuta the male is big. It has the usual antennae and blunt corrugated posterior projections. No cuticular structures were seen in this species. In W. reticulata the males are rare. They have paired antennules and a well-developed penis is coiled up within the body. There is a curious rounded projection from the end of the body opposite the point of attachment. Scattered clusters of cells of glandular function are seen. No size of the male is given. Tomlinson (1963) described the male of Lithoglyptes hirsutus but was not sure whether the male has a penis. In its general features the dwarf male resembles a cypris larva. Tomlinson (1969) reported that the presence of a penis in some species of Lithoglyptes was an unresolved question. According to him most species have a definite penis which can be pulled out to many times the body length. In L. spinatus Tomlinson & Newman (1960) were again not sure about the presence of a penis. The body of the male in this species is bulbous and it is connected to the antenna by a long stalk. This stalk arises from a T-shaped connection with the two normal appearing antennules and terminates in an annulated attachment to the body. In L. indicus, on the other hand, Aurivillius (1894) describes a large penis as well as a testis and a vesicula seminalis. This male is about 0·5 mm long and has the shape of a coil. The nervous system was seen by Aurivillius (1894) but the alimentary canal is missing. A penis may be present. The In L. mitis Tomlinson (1969) gave the size of the male as mm. posterior end is annulated. It does not have any heavy teeth. The antennae have very short stalks. The male of L. scamborachis was described as typical in general aspect (Tomlinson, 1969). It has a penis. The body is finely annulated. No long stalk was found between the body and the antennules. One of the males had a peculiar projection on the body. This appeared to be stuck to the body of the female. The male of L. wilsoni is a reduced bag of gametes (Fig. 2A). Only the reproductive system is well developed. There is a penis and

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

253

Fig. 1. —Anatomy of the adult males of Acrothoracica. A, Weltneria zibrowii (after Turquier, 1985b); B, Berndtia purpurea (after Utinomi, 1961); a, antenna; cg, cerebral ganglion; g, ganglion; p, penis; ps, penis sheath; rb, reserve bodies; t, testis; ta, terminal ampulla; th, thorax; vs, vesicula seminalis; yo, yellow organ; scale bars, A=150 B=200

also the antennae. The long penis sheath is faintly annulated. It has rows of very fine spines which are bifid at the tip. The cuticle of the male has rows of fine spines which are then scattered around the antennae. In L. habei the adult male is about 0·5 mm long. It has a pair of antennae but no long stalks, which are seen in some species. There is a penis. The males of Lithoglyptes are, according to Aurivillius (1894), similar to those of Alcippe (=Trypetesa). In both genera they are attached to the outer side of the female. The males do not have any appendages. There is a difference in the position of the antennae in the two genera. In Alcippe the antennae are in the middle of the body. In Lithoglyptes they are in that part of the body distal of the penis. In both genera the productive organs are well developed with a single testis, a single vesicula seminalis and a penis (penis is doubtful in some species of Lithoglyptes, see above). The nervous system consists of an elongated cerebral ganglion which is at about the height of the vesicula seminalis. From this ganglion nerves arise. From the posterior part of this ganglion a nerve goes to a second ganglion, which is nearly as long as the cerebral ganglion but narrower than the latter. From there a nerve goes to the black pigmented eye. Close to the cerebral ganglion there is a small rounded “yellow organ” whose contents are granular. This formation is reminiscent of the “gland of unknown nature” in the male of Scalpellum regium (Thoracica). mm. It has a nauplius eye, testis, vesicula seminalis The male of Kochlorine hamata is about and a penis. The penis extends into one of the “wings”, which presumably is elongated into a penis sheath.

254

WALTRAUD KLEPAL

The antennae have a stalk and together (antenna+stalk) they may be nearly 0·3 mm long. The body has two protuberances (“wings”) from the stalk; one is more pointed than the other. The male of K. floridana (Fig. 2B) attaches to the horny disk formed by the remains of the female exuviae. Young males look much like a cypris larva (Tomlinson, 1969). The mature male has an obvious cellular testis and a vesicula seminalis leading to an annulated penis. There seems to be a penis retractor muscle connected to one of these lobes projecting from the main body of the male, with the testis in another lobe; the third lobe has a few glandular-appearing cells. There is something like the “yellow organ” and the nauplius eye. The outer surface of the male is covered with rows and patterns of extremely fine dots. The mm (smaller than the cypris). This reduction in length is typical of the size of the male is about maturation of males in the Order. There are no trophic structures. It is assumed that after fertilization the male is expended. The male cypris of K. bocqueti is reminiscent of the larvae of the Trypetesidae, although that of Whilst the Kochlorine is a little bigger. The juvenile male of Kochlorine is pear-shaped and anterior region gets wider the peduncle regresses. The thorax also regresses and the posterior pallial cavity is closed. The pallial cavity is reduced to a narrow furrow and it communicates with the exterior at about the height of the distal extremity. In contrast to most males of the Acrothoracica there is no penis-Anlage developed at that stage. Whilst the visceral mass rotates the gonad grows fast and gametogenesis begins. The structure of the gonad is similar to that of other species. The globular testis is hollow. The axial cavity is full of spermatozoids or spermatids. Apart from the testis there is the vesicula seminalis whose outermost end is the remnant of the pallial cavity. Beginning with a pear-shape the body of the male becomes more and more globular. The posterior end of the mantle, which narrowed in the juvenile male disappears in the adult. A voluminous cylindrical projection differentiates arising from the antennal projection in ventral position. This organ of attachment in the oldest males. In develops at the cost of the capitulum the size of which is reduced to addition, the shape of the capitulum changes. Now it looks like a semi-cone and it is plump (Fig. 2C). The tip of the cone gets prolonged by the attachment organ which becomes long and thin, with a diameter of 30– 35 and whose wall increases in thickness. The capitulum is partly attached to the female, with the attachment organ. This consists of rows of short spines on the body wall of the male. The testis shifts into the posterior body region and the vesicula seminalis into the anterior. The ductus ejaculatorius points into the area of the attachment organ. The migration of organs also affects the nervous system and the eye spot. At the same time the trace of the posterior pallial cavity disappears. It probably gets blocked. Kochlorinopsis discoporellae has a triangular male with a single, remarkable two-jointed antenna. The mm. The posterior end is blunt; this is presumably the penis sheath. The penis size of the male is appears to swell out and to fill the posterior part of the organism. The testis is lobed, the vesicula seminalis is tubular. Utinomi (1961) gave a detailed description of the male of Berndtia purpurea (Fig. 1B). The body is 1·2– 1·5 mm long and about 0·2 mm wide in the anterior position and up to 0·06 mm wide in the posterior part. Utinomi (1961) states that he always found a remnant of the cypris sac in contact with the anterior end around the antennules. From this fact the author concludes that the male, once attached, does not moult again after the final metamorphosis. Within the cypris sac there are remnants of the cypris such as the compound eye. The body wall of the adult male consists of a thin and transparent cuticle. On the outside of this cuticle there are thorns arranged in transverse rows and usually directed posteriorly. Utinomi (1961) did not see any epithelial cell layer beneath the cuticle which is presumably due to bad preservation. He mentions strong longitudinal muscles in the middle part of the body. Up to 19 of these muscles are along the dorsal surface

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

255

and about 16 on the ventral side. The number of muscles varies with the size of the body. The length of the muscles is also variable. Utinomi did not see any cross striation in these muscles (which could also be due to the state of preservation, see above). There are no transverse muscles or fibres. In the posterior prolongation of the body there are no muscles at all. The reproductive organs are the most conspicuous internal organs of the male. They consist of a single tube which may be divided into the testis, vesicula seminalis, vas deferens and ductus ejaculatorius. The tube is coiled (because it is about twice as long as the body). The testis is large and globular. Spermatogonia are found in the basal part and spermatozoa in the portion leading into the vesicula seminalis. The spermatozoon is described as being filiform with a pinhead-like or pyriform head. The vesicula seminalis is marked by a slight swelling; it is a dilation of the proximal part of the vas deferens. The vas deferens enters into the ductus ejaculatorius. At the base of the penis there are several bundles of powerful striated muscles, one group running longitudinally and the other shorter one running obliquely. These muscles enter into the penis sheath and together form an outer circular layer of longitudinal muscles. The annulated penis is very long and greatly contractile. It runs freely, coiled up, within the widest portion of the channel-like body cavity, the penis tube. At the end of the penis there is a brush of fine setae. In the penis there are finely striated longitudinal muscles which form, together with the ringed cuticular covering, an outer sheath for the penis to enclose the ductus ejaculatorius within it. There are no circular muscles as in the penis of the males of Trypetesa and Cryptophialus. The penis tube is usually regarded as the vestige of the mantle cavity. In Berndtia sp. Utinomi (1961) never found any cuticular lining nor any epithelial cell layer which line the mantle cavity in ordinary cirripedes. Therefore Utinomi concludes that at least in B. purpurea the penis tube is a specialized lacunar channel in the body tissue. In the male of B. purpurea there is a terminal sac or ampulla which may be a kind of tactile proboscis serving in the act of fertilization. The nervous system resembles that of Trypetesa lampas (see later). Where the vas deferens communicates with the ductus ejaculatorius there is a large dark-coloured ganglion. This corresponds well with the “Hauptganglion” in Trypetesa. It has a median constriction and thus it seems to consist of two parts. Utinomi (1961) assumes that this is the fused mass of the ventral nerve ganglia. In the male of Berndtia purpurea no optic ganglion or eye has been found. There is a “large yellow peculiar body lying usually in close contact with the main ganglion” (Utinomi, 1961, p. 439). This is homologous with the “gerundete Organ” of Berndt found in the male of Trypetesa lampas. Utinomi (1961) interprets this “yellow organ” as a kind of nutritive organ which originates from the larval alimentary canal. There are ovoid globular cell masses on either side of the vas deferens and one near the proximal end of the penis. These are reminiscent of the cement glands of the males of Scalpellum velutinum and S. regium. They are not found in any other acrothoracican. There is no digestive tract nor were any excretory organs found. In larval development the male shows a close resemblance to that of Cryptophialus, whilst the female resembles Trypetesa. The male of Berndtia nodosa resembles that of B. purpurea. It is found primarily on the exuviae remains of the female; in B. purpurea the males are invariably found attached to the wall of the burrow. Information on the males of Cryptophialus sp. is very scarce and in most cases is restricted to a description of the point of attachment and on the presence of a penis. According to Tomlinson (1960) C. coronatus forms a penis late in life. Several examined males had no penis and some had a small penis only. mm. On the integument there are peg plates (polygons bearing numerous dark The animal is about spots). In C. melampygos the same author observed a long penis. Utinomi (1961) states that in C. minutus the longitudinal muscles are distinctly striated. The penis of the Cryptophialus sp. is known to have circular muscles and according to Berndt (1903a) the retractor muscles lying at its base become weaker towards the

256

WALTRAUD KLEPAL

penis and ultimately merely form an outer layer of finer longitudinal muscle fibres. Most ordinary cirripedes (supposedly the Thoracica) have an outer layer of strong longitudinal muscles in the penis (Klepal, Barnes & Munn, 1972). The circular muscles lie inside the longitudinal muscles and are formed as the outer covering of the ductus ejaculatorius. C. minutus was described by Darwin (1854) who pointed out the great similarity between the males of Cryptophialus and Alcippe inspite of the great dissimilarity of their pupae. The male of Cryptophialus minutus is about 0·3 mm long. No eye was seen, no mouth, no thorax, no cirri and no other organs except the testis, vesicula seminalis and an immensely elongated probosciformed penis. The penis can be stretched to 3–9 times its resting length. C. minutus striatus was described by Berndt (1906) as having a male which is about 0·4 mm long. The shape of the male is that of a bulged bottle with a rounded bottom and a straight neck which is compressed sideways. There is a large testis and a pear-shaped vesicula seminalis which eventually leads into a long vermiform penis. In the “neck of the bottle” there is a characteristic slit-shaped opening. There is a membranous canal which leads the penis and keeps it in its position by connective tissue fibres (and not by muscles). Around the vesicula seminalis there is a thin layer of circular muscles. The cerebral ganglion is big. The ganglion opticum and the eye are missing (but this lack may be due to bad preservation). In the mantle wall there is a system of strong cross-striated longitudinal muscles which seem to have the function of directing the penis in the stretched condition. There are also rings of small transverse muscles which may help to push out the extended penis. The penis is very long and it may be assumed that it reaches nine times its resting length. From the outside to the inside there are small circular muscles, followed by longitudinal muscles forming a tube. There are lacunae and in the centre there is the ductus ejaculatorius. At the tip of the penis there are hooks (no setae) which help to attach the penis on to the integument of the female during the ejaculation of the spermatozoa. The spermatozoa themselves resemble those of Alcippe. Batham & Tomlinson (1965) described the male of Cryptophialus melampygos [now Australophialus mm, only slightly larger than the melampygos (Berndt, 1907)]. The size of the adult male is about cyprid. It is pear-shaped, the broad end being anterior. The posterior end of the male is slightly forked, one fork carrying several multi-spiked chitinous knobs. Internally the body wall shows bands of longitudinal muscles. Young males still show some yolk globules. A conspicuous “yellow organ” is seen in the male as well as in the cyprid and the metamorphosing female. Anterior to it there is the testis and the seminal vesicle, in both of which there are elongate sperms. The long muscular penis is coiled. It is very long, several times the length of the male. The penis is annulated throughout its length and broader distally. At its end there are five small setae. In Cryptophialus heterodontus the male has a pair of antennae without any stalks. The body of the male is rounded. No penis was seen (the present author thinks that the males could have been immature specimens). No eyes, nor any “yellow organ” were noticed. C. wainwrighti has a male with a prominent penis, and paired antennae (Fig. 2D). The hyaline mantle of the male has no peg plates. The posterior end of the animal is bifid with about four small teeth. Whether eyes and “yellow organ” are present could not be made out with certainty. C. variabilis have males which, when young have a flattened cypris carapace with distinct peg plates, dorsal hairs and posterior long bristles. In addition, there is an articulated posterior plate. As usual there are the antennules; when the males are spent the antennae remain on the females. Mature males are smaller than the larvae. They have an elongated sheath for the penis. The sheath terminates in a toothed bilobed process. Other males described by Tomlinson (1969) are those of C. newmani, C. lanceolatus (which has an indication of a stalk between the body and the antennae and a cuticular single “beak” in the central line on the ventral side of the body), and C. unguiculus (whose antennae are often in a spread position). In all cases

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

257

a penis is well developed. For C. cordylacis the author gave drawings of the males but there was no description. Turquier (1985a) states that the morphology and the size of the males varies with the age of the animal; he observed this in Australophialus pecorus. Whilst soon after the metamorphosis of the cypris the males ) they become more elongated (length: width ratio=1·8 appear globular or pear-shaped (size as morphogenesis proceeds (Fig. 2E). In that particular in mature animals) and they reach a size of 350 species the testis is in the middle of the body and the seminal vesicle is orientated towards the rear end of the body. The ductus ejaculatorius turns to the ventral side before it extends into the penis. The penis itself stems from the metamorphosed larval thorax as in all males of the Acrothoracica. It is coiled in the ventral part of the body and is in a postero-dorsal position within the mantle cavity. The genital opening is in the posterior part of the body and it is armoured with teeth. The two ganglia of the nervous mass are close together, the nerves extending into the rear end of the body. Close to the nervous system there is the “yellow organ”, vesicular with a spherical opaque centre. mm. It has paired antennae and a long coiled penis. In A. turbonis the size of the male is about The end distal to the antennules has two to three teeth on the mantle surface. They look like the boring teeth of the female mantle surface. There are about 12 males per female but up to 17 were seen. Berndt (1903b) found up to 12 males in Alcippe lampas (=Trypetesa lampas). The shape of the males is that of a bulgy bottle (capitulum, Darwin, 1854) with a sideways highly compress neck (peduncle, Darwin, 1854). At the bottom of the bottle neck there are two lateral lobes of the peduncle. Down to these lobes the peduncle is buried in the disk of the female. The epithelium of the female forms a deep pocket around the peduncle of the male. The peduncle is about 0·6 mm long as is the capitulum. The whole inside of the body is freely open to the water. At the bottom of this cavity the penis arises. The basal part of the penis is the body proper of the animal. The mantle cavity forms the penis sheath. The narrow end part of the penis sheath forms a structure leading the penis during copulation. A large single testis is at the bottom of the peduncle, near the ventral side. From there the club-shaped vesicula seminalis arises and continues into the vas deferens which eventually leads into the penis. The penis may be extruded from the capitulum to 3·5 times its length. The arrangement of elements in the testis of dwarf males seems to agree with that in the testis of hermaphroditic cirripedes. There is one vesicula seminalis whose wall consists of a layer of connective tissue with bundles of circularly arranged extremely fine muscle fibres. There muscles must serve to press the spermatozoa into the penis. Thus in Alcippe (Trypetesa) the spermatozoa are ejected actively whilst in the Thoracica the ejection was described as a passive process. Within the penis there is an outer circular and an inner longitudinal layer of muscles. In the centre there is the ductus ejaculatorius. On the cuticle there are setae, and on the tip of the penis there are two long and four short setae. There are also lacunae within the penis. In a live male the penis is continuously moving. Between the ventral and the dorsal edge of the capitulum there are regularly arranged parallel muscle bundles which give this region a striated appearance. A second group of muscles is arranged as circular bars which are open on the central side. Berndt (1906) thinks that these muscles contract and thus help to press out the penis. The nervous system was first described by Aurivillius (1894). On the ventral side of the seminal vesicle there is a large elongated ganglion, with a deep indentation in its middle. The lower part of this ganglion is pear-shaped and smooth, the upper part shows a low indentation. A thin nerve arises from the lower part and reaches into the area of the eye which lies above the junction between the testis and the seminal vesicle. Here is the big ganglion opticum which is club-shaped and is about half the size of the main “ganglion”. The histology of the nervous system corresponds with that of the female.

258

WALTRAUD KLEPAL

Fig. 2.—Morphology of the mature males of various genera of Acrothoracica. A, Lithoglyptes wilsoni, reduced male with long penis sheath (after Tomlinson, 1969); B, Kochlorine floridana with an orchid lobe and a long annulated penis (after Tomlinson, 1969); C, Kochlorine bocqueti, plump male with long and thin attachment organ (after Turquier, 1977); D, Cryptophialus wainwrighti, male with prominent penis and antennae (after Tomlinson, 1969); E, Australophialus pecorus, elongated male with prominent antennae (after Turquier, 1985a); F, Trypetesa spinulosa, male with a big penis sheath, orchid lobe and lateral posterior lobes (after Turquier, 1967); key to symbols as in Fig. 1; lp, lateral posterior lobe; ol, orchid lobe; scale bars C, E and F=150 µm, in A, B and D 1 cm=140 µm, 80 µm, and 70 µm, respectively.

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

259

The antennae arise below the two lateral lobes. They are exactly like those of the larval antennae of the Lepadidae, with the cement glands and the cement ducts. The males of T. habei are essentially like those of T. lampas. They have paired antennae. The testes extend into a sac descending into the mantle pit, while the penis sheath extends forward towards the aperture. The male may be up to 0·9 mm long, although only about 0·76 mm extends beyond the mantle surface. It is about 0·12 mm wide at the penis sheath. There are several bulbous projections at the level of the entrance into the mantle pit. The testis is at the deepest portion of the lobe in the mantle pit. It is in close proximity to the assumed nauplius eye (Tomlinson, 1969). There is a penis sheath, but whether there is also a penis is doubtful. The outer surface of the male is not ornamented. Kühnert (1934) gave a good description of the development of Alcippe lampas. Good descriptions of the anatomy of Acrothoracica are given by Turquier. This author has described Trypetesa lampas, T. nassarioides, T. spinulosa, T. habei, T. lateralis, Kochlorine bocqueti etc. (Turquier, 1970a, b, 1976, 1977). He has described the cypris of Trypetesa spinulosa and states that there is no difference from the cypris of other species such as T. lampas and T. nassarioides. The adult male of T. spinulosa is also very similar to these two species. Their size is about 0·9 mm. They have a very big penis sheath which is nearly twice as long as the “lobe orchidien” (orchid lobe). There are well-marked lateral posterior lobes which are covered by strong and short spines (Fig. 2F). These cuticular structures are important for the attachment of the male on the side of the rostral tuberculum of the female. There is no peduncle and there are no antennae in the male of T. spinulosa. In T. nassarioides the males have a short penis sheath, as long as the orchid lobe and the only dorsal lobe (Turquier, 1967). The male of T. lateralis is highly reduced and there is no penis sheath. Turquier & Pochon-Masson (1969) investigated the spermatozoa of Trypetesa nassarioides. They came to the conclusion that the development of the spermatozoa in the male is a normal process. The gamete is a sperm with a flagellum. From the original condition with two centriols only one, that of the flagellum, continues to exist. There is no trace of the proximal centriol. The loss of one of the elements of the diplosome has also been noted in some other invertebrates (see Idelman, 1967). The structure of the acrosome is not original. The flagellum has the 9+2 structure, which is, according to Baccetti, Dallai & Rosati (1968) a primitive type. There is a rigid formation on the flagellum which may slow down or weaken the flagellar movement. It is also possible that this formation is the site of important metabolic activity. Thus (with this rigid formation) the gametes can compensate the reduction of their mitochondrial apparatus. It is interesting to note that such an impoverishment of the chondriome occurs also in the vesicular spermatozoa of decapods (Pochon-Masson, 1968a, b). There is also a vesicle, which is a rare formation in spermatozoa with a flagellum. Investigations of the spermatozoa of the operculate cirripedes showed that the vesicle changes in the mature spermatozoon. It is a metabolic reserve for the gamete and could in some way be a replacement for the lacking trophic cells in the testis of Trypetesa. In Trypetesa the cyprids of both sexes and the first stages of metamorphosis are identical. Later there are sexual differences. The cypris of both sexes use cement for their attachment. Then the visceral mass moves into the sagittal plane, the complex eyes histolyse and the peduncular region remains rudimentary. Now the major part of the animal develops from the capitulum and the larval cement apparatus disappears completely. The anterior part of the larval mantle cavity disappears early, so that the upper part of the posterior mantle cavity is the only one which still exists in adult animals. From this stage onwards the male and the female cyprids develop differently. When an apparently undifferentiated cypris larva of T. lateralis settles on a female it degenerates into a bag of testicular tissue (Tomlinson, 1955). After attachment the male cypris develops into a dwarf organism of very simple anatomy with only a nervous system and a genital apparatus. Its morphology differs greatly from that of the female. The male of T. lateralis loses its

260

WALTRAUD KLEPAL

carapace and becomes rounder, averaging mm. The mature male becomes progressively smaller, wrinkled, and variable in shape. This size decrease suggests that there is destructive degeneration in this form with the male discharging its gametic contents and then dying. The phenomenon of the metamorphosis of the male cypris indicates that the sexually differentiated individual is a highly evolved and specialized organism (Turquier, 1971). This could agree with the singularity of the testis and of the vesicula seminalis and the fact that the male does not have a penis. In other species of the same genus the thorax is transformed into a penis. The male moults only once during its metamorphosis whilst in the female there are two exuviations. First in the juvenile phase of the male the rotation of the visceral mass extends and the penisAnlage is there. The “Anlage” of the testis also develops further and the vesicula seminalis becomes independent. In the adult phase the animal grows, gametogenesis continues and the genital tract is completed. The male “Anlage” differentiates very early and its evolution sets in immediately after the attachment of the cypris. Thus the male is able to reproduce about two weeks after hatching from the nauplius. The female matures only several months later. The acceleration of the evolution of the genital apparatus and the lack of a second moulting in the male (as it is there in the female) may be interpreted as a process of condensation of metamorphosis. In the adult male there is only a hypertrophied genital apparatus and a reduced visceral mass which contains only the nervous system, consisting of two ganglia, and the “yellow organ”, which may be interpreted as a rudiment homologous to the digestive tract of the female. There is no trace of any metameres nor of any appendages. The larval muscles are completely de-differentiated. The penis sheath is formed by the mantle of the posterior body region. In T. lateralis Tomlinson (1969) noticed the lack of a copulatory organ. (The majority of the males have a long extensible penis.) The semen is therefore deposited in the rostral part of the pallial opening of the female. It is probable that the spermatozoids get into the pallial cavity of the water current, caused by movements of the cirri and the rhythmic retraction of the body of the female. When Tomlinson (1969) states that the “cross structure of the male seems somewhat acellular. The cell boundaries appear to have broken down and the developing spermatozoa appear to be dispersed in the central region of the organism” it seems that he worked on a deteriorating animal. Turquier (1971) did not find anything of that sort in T. nassarioides nor in T. lampas (Turquier & Pochon-Masson, 1969). It would therefore be very surprising if T. lateralis showed such an aberrant structure. The anatomical simplicity within the males of Trypetesa is relatively constant. Thus, the males of T. lampas and of the much more primitive Berndtia purpurea are very similar. In B. purpurea the male is regularly pear-shaped which is caused by the lack of the orchidien lobe, in which there is the welldeveloped genital apparatus in Trypetesa. THORACICA Balanomorpha The sedentary Thoracica are usually free-living or commensal cirripedes. Most of them are hermaphroditic with pseudo-copulation. Cross fertilization is generally the rule. In the scalpelliform cirripedes of the Lepadomorpha males have been known for a long time. As late as 1965 Henry & McLaughlin found for the first time complemental males within the genus Solidobalanus (Henry & McLaughlin, 1965, 1967), one of the most highly evolved genera of the sessile barnacles. The males are degenerate, with vestigial cirri and the typical antennules but they do not have a mouth. The reproductive organs are well developed. Since then other Balanus species with males have been found. McLaughlin & Henry (1972) compared the morphology

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

261

of four species of Balanus (Fig. 3A-C). These were B. (Solidobalanus) masignotus, B. (Conopea) galeatus, B. (C.) merilli, and B. (C.) calceolus. In all species mentioned the authors noticed a deep depression in the rostrum near the sheath. It is therefore assumed that all species with a convex rostrum may have complemental males. The external morphology of the males of the four species is similar. They have a semiglobular basis, a narrow lateral band, the opercular surface with prominent opercular valves and opercular lips. Only B. calceolus is different in having separated scuto-tergal valves and raised opercular lips. Concerning the internal morphology there are differences in the degree of degeneration in the four species. All of them have a penis-thorax. The setation of the penis varies. The more setae there are on the penis the more vestigial are the cirri. Thus in B. calceolus the penis has only six to thirteen long spines. On the thorax are six pairs of prominent cirri. In B. merilli there are 25 to 46 long spines on the penis and the six pairs of cirri are vestigial. Whilst in B. calceolus the mouthparts are relatively well developed, those of B. galeatus and B. merilli are considerably reduced. Usually the seminal vesicles are paired. In B. masignotus the vesicula seminalis may be paired or simply bilobed. In most cases the testis is a diffuse bilobed structure. All four species have a strong muscular support for the thorax and the testis. There are also muscular connections to the basis in the region of the antennules. The Musculus adductor scutorum is well developed in B. calceolus and B. merilli but it is weak or lacking in B. galeatus and B. masignotus. A progressive degeneration in cirral structure and mouthparts may be traced from B. calceolus through B. masignotus to B. galeatus and B. merilli. It is evident that close examination of other species is needed before the phylogenetic significance of complemental males in the balanids can be ascertained. Dayton, Newman & Oliver (1982) found that Bathylasma corolliforme had small individuals on the opercular valves of the hermaphrodite which were acting as males. These small individuals are not just protandric; they are definitive males settling predominantly on the terga and the exposed sheath of the carina and they remain small (see also Foster, 1980). Similar results were obtained in B. alearum (Foster, 1983) (Fig. 3D). A larva destined to become a complemental male must not only settle on an established hermaphrodite but it must also select a site from which it can effect fertilization. A larva encountering an isolated receptive hermaphrodite would have a better chance of living long enough to pass its genes on to the next generation if it settled in an appropriate position on the hermaphrodite and became a precocious male. The males of B. alearum are laterally compressed. The parietes are often cracked or deformed. The opercula are embryonic and also often deformed. Mouthparts, cirri, and penis are well developed. The penis is longer than the sixth cirrus. The reproductive organs are well developed. The alimentary canal is there but in the specimen described by Foster (1983) there were no food remains in the gut. Hui & Moyse (1984) found a complemental male in Chionelasmus darwini (Fig. 3E). In Chelonobia patula Crisp (1983) discovered what he called “apertural males” (Fig. 3F-H). These individuals settle in the opercula region of the hermaphrodites and are themselves really hermaphrodites being arrested in development. In contrast to most complemental males and dwarf males the males of Chelonobia are capable of feeding and growing and in contrast to all males described earlier they have the potential to become hermaphrodites. Crisp (1983) assumes that the hermaphrodite has the capability of reducing the growth rate of surrounding complemental males by high competition for food. Somatic growth is limited by food supply, whereas gonad and penis development proceeds more or less independently as a function of age. C. patula seems to be a pointer to the evolution of the complemental male (Crisp, 1983). It is a protandrous hermaphrodite. As long as the individuals are smaller than 2 mm in diameter they do not have a penis. The penis then develops and when the individuals are between 4 and 7 mm in diameter they are protandric males and at greater than 7 mm basal diameter they are simultaneous hermaphrodites. Individuals with a basal diameter between 2 and 4 mm have testes, vesicula seminalis and penis, but no

262

WALTRAUD KLEPAL

ovary. Their penis is abnormally well developed so that it can be assumed that these small individuals would be capable of inseminating the hermaphrodite. Lepadomorpha Within the Lepadomorpha species of the Thoracica there are a great number of dwarf or complemental males. Up to 1908, 170 species of Scalpellum had been described and intermediate types of males had been found (Pilsbry, 1908). Since then several more species of Scalpellum and other Lepadomorpha genera have been described. Repeatedly authors tried to group the various genera according to their systematic positions. Pilsbry (1908) was the first who suggested that the characters of the males should be considered just as those of the females and the hermaphrodites for the elucidation of the systematic position of the various scalpelliform species. Nilsson-Cantell (1921) follows essentially Pilsbry’s classification of the species, but he points out the great variability of the dwarf males. Therefore Nilsson-Cantell states that Pilsbry’s (1908) classification cannot be regarded as being satisfactory for all species. Where Pilsbry talks of “genera”, Nilsson-Cantell uses the term “groups”. Pilsbry (1908) distinguishes four different genera by the plates, by the degree of separation between capitulum and peduncle, by the presence of a mouth and alimentary system and by the condition of the cirri. Calantica has the most original complemental males. This genus comprises C. villosa, C. trispinosa, C. eos, C. calyculus, C. falcata, C. gemma, C. superba, and C. grimaldi. The complemental males of the genus Smilium are like those of Calantica only the females or hermaphrodites may be distinguished by the elevation of a pair of latera above the basal whorl in Smilium. This genus comprises S. peronii, S. uncus, S. pollicipedoides, S. aries, S. sexcornutum, S. scorpio, S. acutum, and S. longirostrum. The genus Euscalpellum differs from the two preceding ones by having more degenerate males. The males are saclike, not distinctly divided into capitulum and peduncle and they have only three plates, the scuta being larger than in Scalpellum. They have six pairs of articulated cirri and a mouth. The genus Euscalpellum comprises E. rostratum, E. renei, E. bengalense, E. stratum, E. squamuliferum. In the genus Scalpellum the males are very degenerate, sac-like without a peduncle or mouth, vestigial cirri or digestive tract. The plates are absent or scuta and terga are extremely small. Within the genus Scalpellum Pilsbry (1908) distinguishes three groups. (1) Group of S. scalpellum. To this group belong S. stearnsi, S. inerme, S. calcaratum, S. hamatum, S. scalpellum, S. patagonicum, and S. salartioe. (2) Group of S. californicum with S. californicum and S. osseum. (3) Group of S. stroemii. This group comprises S. stroemii, S. stroemii obesum, S. s. luridum, S. s. aduncum, S. s. septentrionale, S. s. substroemii, S. s. latirostrum, S. pressum, S. groenlandicum, S. angustum, S. nymphocola, and S. cornutum. Within the genus Scalpellum the subgenus Arcoscalpellum with two sections may be distinguished. To the stock of Arcoscalpellum belong S. velutinum, S. idioplax, and S. carinatum. The section Mesoscalpellum includes S. intermedium, S. nipponense, S. laccadivicum, S. japonicum, S. larvale, S. gruvelii, S. imperfectum, and S. sanctaebarbarae. The section Neoscalpellum comprises S. edwardsi, S. dicheloplax, S. phantasma, and S. marginatum. Both Calantica and Smilium have complemental males that look like free living juveniles (Fig. 4A, B, Table I). Capitulum and peduncle are well separated from each other. They have six large primary capitular plates. In addition to them there may be a number of small latera so that altogether there may be 15 plates

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

263

Fig. 3.—Morphology of adult males in Balanomorpha. A, generalized male of Balanus; B, Balanus calceolus, penis-thorax with cirri and mouthcone; C, Balanus masignotus, penis-thorax with vestigial cirri; (Figs A-C after McLaughlin & Henry, 1972); D, Bathylasma alearum, male in tergal niche of a larger specimen looking like a young hermaphrodite (after Foster, 1983); E, Chionelasmus darwini, male with right cirri and shell removed (after Hui & Moyse, 1984); F, Chelonobia patula, arrows indicating pits within junction of radius and paries in which cyprids settle; G, hermaphrodite with apertural males, looking like young hermaphrodites; H, apertural male with long penis, removed from shell; (Figs F, G, H, after Crisp, 1983); a, antenna; am, apertural male; ap, aperture; b, basis; c, cirri; me, mouth cone; ol, opercular lips; p, penis; pt, penis thorax; s, scutum; t, tergum; scale bars, A=0·1 mm, B=0·05 mm, C=0·1 mm, D=2 mm, E=0·4 mm; F=5 mm; G=0·5 mm, H=1 mm.

264

WALTRAUD KLEPAL

on the capitulum. Both genera have six pairs of articulated cirri and at least in Calantica there is a long extensible penis that may be beset with setae as in C. trispinosa. The penis is relatively longer than in a juvenile hermaphrodite of the same size, e.g. in C. villosa, C. studeri (Foster, 1978). The digestive tract is functioning (Nilsson-Cantell, 1931). The males are relatively big, e.g. 3 mm long and 1·8 mm wide as in Smilium sexcornutum or 1·8 mm long and 1·3 mm wide as in Calantica trispinosa. In Smilium the capitulum may be up to 5·8 mm long. Thus it exceeds in size many free-living juveniles attached to penduncles of larger specimens. In Smilium sexcornutum of the large plates the one scutum consists of two well-separated pieces (Krüger, 1911), which may point to the original condition. Krüger (1911) found up to three males attached to one female of Calantica, on its integument between the scuta, below the adductor muscle. In general the occurrence of males is rare in Calantica. Foster (1978), however, found up to 16 males between the scuta of large hermaphrodites of C. spinilatera. In the main features described above the males of some species of Scalpellum resemble those of Calantica and Smilium. The males of Scalpellum squamuliferum, S. peronii, S. villosum (now Calantica villosa), Scalpellum pilsbryi, S. scorpio, and S. longirostrum all have peduncle and capitulum separated from each other (Fig. 4C). They are relatively large (at least 1 mm long and sometimes up to 1 mm wide), they have six larger plates on the capitulum (there may be additional smaller ones as the five in S. pilsbryi). These males have six pairs of cirri, usually articulated. They may be well developed as in S. squamuliferum (Darwin, 1851) or they may be relatively short as in S. peronii, S. villosum, S. pilsbryi, S. scorpio, and S. longirostrum. In these last five species at least the first cirrus may be well separated from the following ones. All males of Scalpellum species have a penis which is well developed and extensible as in S. squamuliferum and in S. pilsbryi. In S. villosum the penis is described as being blunt and in S. longirostrum it is short. With the exception of S. villosum all males of the Scalpellum species already discussed have caudal appendages. All these species mentioned above may really belong to the genus Calantica as e.g. Scalpellum villosum (now Calantica villosa) or Scalpellum pilsbryi which is according Bocquet-Védrine (1971) identical with Calantica calyculus, or they may belong to Smilium as does Scalpellum peronii. In all cases the males are complemental males. They are attached to the integument of the hermaphrodite in a fold in the central line between the scuta, a little below the Musculus adductor scutorum. Thus, they are some way below the umbones of the scuta, and they get protected by them when they are closed. Many species of Scalpellum are known but even when males have been found they have often only been described very superficially. Some authors like Annandale (1910) and Stewart (1911) have described the males more accurately. The anatomy of the complemental male of S. squamuliferum was studied by Stewart (1911). This author concentrated on the digestive tract, the reproductive system and the tissues of the peduncle. Annandale (1910) gave an account on the “dwarf” males of several scalpellids. His main interest was the morphology. According to Stewart (1911) the adult male is 1–1·4 mm long and a maximal 0·7 mm wide. The peduncle is short, stout and well separated from the capitulum. On the capitulum there are six calcified plates; rostrum, scuta, terga, and carina. There are no peduncular plates and no latera as in the hermaphrodite. Annandale gave a precise description of the plates. The tergum is broadly triangular, the base of the triangle is rounded and the apex is pointing directly downwards. The scutum is much larger than the tergum and more narrowly triangular, with the apex pointing upwards. The carina is triangular, has a rounded base, is not quite as broad as the tergum, a little larger than the tergum, but not reaching upwards as high as the upper margin of this plate. The base is slightly lower than that of the scutum and above the apex of the rostrum. The rostrum is of about the same length as the tergum. It is rather broader than the carina and with the base produced to a point.

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

265

The outer surface of the male is covered with simple hairs. They consist of an outer cortical portion (staining intensely with iron haematoxylin) and a core which does not stain. The cirri and penis are well developed, caudal appendages are present. The mouthparts resemble those of the hermaphrodite. They are generally smaller, there is no labrum (this may be doubted—it was probably lost in preparation) and the inner teeth of the mandibles are not so distinctly separated. Thus the male belongs to the type most commonly found in the subgenus. The alimentary canal is a functioning narrow tube. Around the oesophagus there is a layer of circular muscles, at the anterior part of the narrow stomach there are two simple epithelial tubes, the coeca, which run for a short distance forward along the oesophagus. Stewart (1911) did not describe any special muscles of the rectum. The anus is dorsal to the base of the penis. The reproductive organs are tubular and consist of the testes, the seminal vesicles and the ductus ejaculatorius with the penis. The different regions are mainly distinguished by the nature of their contents. In all cases the wall consists of a fine layer of endothelium, only close to the external aperture is there a sphincter. The histology and development of the organs of the peduncle were described in more detail. There is a rostral duct in the peduncle which in the male reaches only from the root of the prosoma to the upper quarter of the peduncle. This system of spaces forms an erectile tissue by which the animal can move its peduncle in a swaying manner or by which it elongates or shortens its peduncle. The connective tissue cells within the peduncle are often large with a rounded nucleus. They may be full of reserve granules (“yolk granules”). In this case Stewart did not find it easy to distinguish them from the cement cells. Reserve material is also found in cells around the stomach and around the ventral nerve cord in the prosoma. The cement cells in the adult male are almost spherical with a spherical nucleus per cell. In their protoplasm there are large granules of irregular shape. Because the staining reaction of these granules was identical with that of yolk Stewart (1911) poses the question whether these granules may be yolk granules which are in the process of conversion to form cement. In unstained males these cells are yellow and thus they could be the “cellules jaunes de la pedoncule” of Gruvel (1905, p. 448). Stewart (1911) also gave a general outline of the development of the male. TABLE I Comparison of the males of Calantica, Smilium, Scalpellum and Pisiscalpellum: under cirri, 6=six pairs normally spaced, 1/2– 6=space between first cirrius and the rest, 1/2–6 R=space between first cirrius and the rest but all reduced; D=distinct; J.d.=just distinct; N.d.=not distinct; E=extensible; Fl=functionless; O.f.=open functional; L=large; S=small; P=plates normally developed; p=plates reduced; R=reduced; += present; –=no information; ? =doubtful. Species Calantica trispinosa Smilium sexcornutum Scalpellum squamuliferum S. peronii S. villosum S. pilsbryi

Size (mm) CapitulumPeduncle

1 1

Alimentary Canal Plates Cirri

Penis Caudal appendages

D

O.f.

6P

6

+E

–

D

O.f.

6P

6

?

–

D

O.f.

6P

6

+E

L

D D ?

O.f. O.f. O.f.

6P 6P 6P

1/2–6 + 1/2–6 + 1/2–6 +E

S − S

266

WALTRAUD KLEPAL

Species

Size (mm) CapitulumPeduncle

Alimentary Canal Plates Cirri

Penis Caudal appendages

S. scorpio S. longirostrum S. bengalense S. gigas S. hoeki S. striatum S. luteum S. intermedium S. kurchatovi S. chiliense S. galapaganum S. (S.) elongatum

? ? ? ? ? ?

O.f. R O.f. O.f. O.f. O.f. O.f. ? ? ? ? ?

+ + ? ? ? ? ? ? ? ? ? ?

S. ornatum S. rutilum S. stearnsi S. discoveryi S. gracile S. tritonis S. wood-masoni S. retrieveri S. vulgare S. rostratum S. alcockianum S. chitinosum S. compactum S. compression S. condensum S. convexum S. crinitum S. distinctum S. fissum S. gibberum S. gruvelianum S. hexagonum S. javanicum S. projection S. regium S. sessile

0·75 ? ? ? ?

J.d. J.d. N.d. N.d. N.d. N.d. N.d. N.d. N.d. N.d. N.d. J.d.

? ? ? 0·8 ? thorax 0·4 µm 1·23 1 0·67 1.3 0·67 0·8

1·17 ?

2·2

N.d. ? N.d. N.d. N.d. ? N.d. N.d. N.d. D N.d. N.d. N.d. N.d. N.d. N.d. N.d. N.d. ? N.d. N.d. N.d. N.d. N.d. ? J.d.

6P 1/2–6 6P 1/2–6 4P ? 4p 6R 4p 6R 2P2p 6R 4p 6R 4p 6 4p ? 2P2p ? 2P2p ? 4p ? – ? ? ? ? ? ? ? – O.f. ? ? ? ? ? ? ? ? ? ? ? ? ? ? Fl ?

4p 4p 4p 2P2p 2P2p 4p 4p 4p 4p 3p – – – – – – – – – – – – – – – –

4 ? ? ? ? ? ? ? 4 1/2–6R – ? ? ? ? ? ? ? ? ? ? ?

L L ? S ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? + ? ? ? ? ? ? ? ? ? ? ? ? ? ? – ?

Setae ? ? ? ? ? ? ? ? S ? ? ? ? ? ? ? ? ? S ? ? ? ? – ?

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

Pisiscalpellum withersi

0·4

N.d.

?

–

?

?

267

?

After fixation the cypris of the male is shorter than high. Its peduncular region is greater in comparison with the capitular region. Its total length is 0·75 mm. The mouth is open but there is no anus. The stomach contains excretory matter. The six valves are clearly visible through the shell. In the young adult the capitulum is much longer than in the cypris. The testes of the hermaphrodite may be seen lying at the sides or ventral of the stomach. In the cypris of the hermaphrodite the testes appear “rudimentary”. Stubbings (1936) states that the larvae of S. squamuliferum are of two sizes. The shorter one is said to be the male, whilst the longer one is that of the hermaphrodite. (This is in contrast to the cyprids of the Rhizocephala, in which the larger cypris will become the male.) Thus, as Stewart (1911) states “maleness is therefore not the result of the position of attachment” (Stewart, 1911, p. 38) as was assumed by Smith (1906). Smith considered a position on the margin of the pallial aperture to be the cause for the nondevelopment of the hermaphrodite character in the “males”. There is no description of the nervous system of the male. Stewart (1911) only states that the hairs on the outside of the capitulum are supplied with nerves. Whilst the hermaphrodites of S. squamuliferum and S. bengalense resemble each other closely, the males differ to a high degree. In the male of S. bengalense there is no marked outward boundary between the peduncle and the capitulum. There are only four valves present. The capitulum and the peduncle are covered by hairs, there is a broad band of larger hairs (having a bifid tip) on either side of the carinal midline. The mouth cavity is small in comparison with the capitulum, whose wall is relatively thick. Therefore the true body is also small. The adductor scutorum muscle persists in its usual situation, although in the majority of the specimens there are no scuta. The muscle may be able to narrow the opening of the pallial cavity. The alimentary canal has the usual V-shape. The mouth and the anus are open. The stomach is a fairly large sac, two coeca arise from the anterior end of the alimentary canal. The nervous system consists of a pair of large cerebral ganglia and a massive ventral cord. The hairs covering the outer surface fulfil a sensory function. There is also a small ganglion in the outer wall of the pallial cavity in the carinal midline. It consists of a single row of cells which are continuous with the epidermis. A thin layer of intensely staining nervous matter spreads out from the nuclei on either side under the bands of larger hairs. The cuticle is described as being either entirely absent or very much thinned over the ganglion. Stewart (1911) did not describe the reproductive system. He only states that the testes are on either side of the stomach. At the root of the prosoma there is a somewhat indefinite space which may possibly be the equivalent of the rostral duct. The cement glands consist of two groups of cells. Many of them are full of large irregular yolk-like granules. Stewart could not find any ducts. As in S. squamuliferum there is a great deal of reserve material (“yolk”) in the peduncle. It is more concentrated above the cement glands. The material is within vesicular cells with flattened nuclei. The adult male of S. gruvelii is pear-shaped. The animal Stewart (1911) described has the anterior half thicker than the posterior. The males are about 1 mm long and 0·5 mm wide. The antennae are in the same position as in the larva. The pallial cavity, lined by a fine epidermis, and the prosoma are much reduced. The alimentary tract consists of a small hollow ball of cells containing some cuticular and excretory matter (the cuticular matter could come from some prey). The nervous system is much reduced. The cerebral ganglion is on the opposite side of the stomach from the reduced ventral nerve cord. The sensory hairs are supplied with nerves in the

268

WALTRAUD KLEPAL

Fig. 4.—Morphology of adult males in Lepadomorpha. A, Calantica calyculus (after Bocquet-Védrine, 1971); B, Smilium peroni (after Krüger, 1914); C, Scalpellum squamuliferum (after Annandale, 1910); D, Scalpellum rostratum (after Darwin, 1851); E, Scalpellum wood-masoni (after Nilsson-Cantell, 1931); F, Scalpellum regium (after Thomson, 1873); in Figs A–C capitulum and peduncle may be distinguished; in Figs D–F the shap is ovoid and the plates get more and more reduced, at the same time the cuticular structures on the outer surface become denser and they are arrenged regularly; a, antenna; ca, capitulum; cs, cuticular structures; pe, peduncle; pl, plates; scale bars, A=1mm, C=0.5mm, D=0.2mm, E=0.5mm, F=1mm.

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

269

same way as the hairs of S. squamuliferum. Muscle fibres run from the anterior to the posterior pole of the animal. The testes are heart-shaped, and the vesiculae seminales were not seen by Stewart (1911). It is thought that the seminal vesicles may empty themselves by contraction of the longitudinal muscles. At the same time the prosoma carrying the ejaculatory duct is thrust through the opening of the pallial cavity. Thus the tension in the body is raised. The cement glands consist of two or three small spherical masses of cells lying posterior to the testis. As in the species described above the entire surface of the capitulum and the peduncle is covered with hairs. These are bilaterally distributed. They are parted in the carinal and rostral midlines, the hairs of each side being directed towards the rostral line. The hairs in S. gruvelii are simple and not so rigid as in S. squamuliferum and S. bengalense. Stewart (1911) also describes the post-larval development of the male of S. gruvelii. Whilst the cypris larva of this species resembles closely that of S. regium, the pupa resembles in outer form the adult male of S. velutinum (Gruvel, 1902a). Thomson (1873) described the (complemental) male of S. regium in his notes from the Challenger expedition. This author found five to nine males attached to the occludent margins of the scuta of the fully developed form. (According to Thomson the fully developed form is a female, whilst according to Hoek (1884) it is a hermaphrodite.) Thomson describes the males as being oval and sac-like whilst Hoek (1883) talks about a cylindrical shape of the males. The size is up to 2·5 mm in length and 0·9 mm in extreme width. There is a peduncular and a capitular pole, the antenna being placed on the first one, a little distant from the extremity on the ventral surface. There are no plates and the thorax is not jointed (Fig. 4F). The body wall is described as being “chitinous” (Hoek, 1884), thin and delicate. It is beset with spines in transverse rows. The spines are narrow and pointed where they attach to the wall of the body and they are broadest at the other extremity. The free margin is deeply toothed and thus the spines resemble the scales of Lepidoptera. In other places they look like combs. The epithelium consists of flat cells “with indistinct limits” (Hoek, 1884) and with conspicuous nuclei. Under the hypodermis cells a well-developed layer of muscle fibres is present everywhere. These fibres are transversely striated. The striations are in part indistinct which could have to do with the fibres being nearly functionless and rudimentary. The muscular fibres have an irregular oblique direction which in some parts approaches to a transverse, in other parts to a longitudinal position. The connective tissue consists of fibres and of delicate and finely granulated membranous plates which form the partitions between the large meshes. The occurrence of a well-developed mass of connective tissue between the different organs within the body is the rule in all cirripedes. Hoek (1884) described the mouth as functionless. According to him the oesophagus is a narrow tube which widens and passes into the stomach, whilst Thomson (1873) did not see any oesophagus or stomach. (This difference in the descriptions could be due to a possible variability in the animals.) When the stomach is present it is a pyriform pouch closed on all sides with a rudimentary intestine. In a fully grown male the stomach is almost empty, in a younger individual it is filled with a yellowish brown coloured mass of fatty nature, probably reserve material. The supraoesophageal ganglion is situated against the oesophagus. The commissure unites this ganglion with the large thoracic ganglion. The latter represents the whole ventral nerve cord and is attached before the supraoesophageal ganglion. It is elongated ovoid. The only distinct nerve coming from the thoracic ganglion arises terminally. Two stronger nerves arise from the commissures very close to the supraoesophageal ganglion. All other nerves given off from both ganglia are extremely delicate and hardly recognizable as such.

270

WALTRAUD KLEPAL

The testis is heart-shaped with the incision being directed towards the capitular extremity. This incision is the only sign of the original duplicity of the male genital gland. The vesicula seminalis is an irregular globular vesicle of a diameter of 0·3 mm. It is very closely pressed against the testis. The vesicula seminalis is only a dilation of the vas deferens at the place where it corresponds with the testis. The cylindrical terminal position of the thorax may be called a “penis”. There are six pairs of cirri. Four of them, the posterior ones are well developed and have two rami. In the first and second pair of cirri only one short branch is left. The rami of the fourth to the sixth cirrus are relatively long and narrow and they terminate in two or three very long spines. Each male has a pair of cement glands. They are of an ovoid shape and they are situated a little above the vesicula seminalis. The glands are comprised of very large cells with granular contents and a nucleus. The large granules in the cells are placed at the periphery. Cement ducts come off the glands as thread-like appendages. The male of S. regium has a highly degenerated organization. The elongated body has the shape of a bag with a slit representing the opening between the two scuta in other species. Only the antennae show their original condition, the cirri are straight and functionless, the mouthparts have disappeared. The intestine is rudimentary and has become functionless. The nervous system consists of a relatively small supraoesophageal ganglion, of a not very stout oesophageal ring and of a large thoracic ganglion. Probably the latter alone regulates the functions of the genital apparatus. The peripheral part of the nervous system is not much developed. There are no eyes or other sensory organs. The genital apparatus is, apart from the cement glands in the young males, the only well-developed organ system. The female organs are lost and the male organs show a great deal more concentration than do the same in ordinary hermaphrodite cirripedes. There is only a single testis and a single vesicula seminalis. In all these respects the males of other deep-sea species of Scalpellum, investigated by Hoek (1884) exactly correspond to the male of S. regium. So does the male of S. vulgare (a specimen from the Mediterranean) with the exception of the presence of rudimentary plates, which in that species, as in some of the deep-sea species, represent the so-called primordial valves of the young capitulum of pedunculated cirripedes. S. regium is in many respects identical with S. gruvelii. Darwin (1851) described the complemental male of S. vulgare as flask-shaped, whilst Gruvel (1898) stated that it had the shape of a “Punica granatum” and in his more precise description (Gruvel, 1899) he stated that the complemental male had the shape of a more or less rounded sac. Each male is attached to the hermaphrodite on the inside of the scuta in something like a fold which Darwin calls a “transparent spine-bearing chitine border”. There may be up to 15 males on one side, in which case they are attached so closely to each other than they have no longer their ordinary shape. The males are never found anywhere else on the hermaphrodite as Gruvel (1899) stressed. On the anterior side of the male there are eight lobes arranged in two concentric rows. There are four rudimentary, bead-shaped and calcareous plates. They are believed to be the scuta and terga, although they are placed considerably below the orifice. The eight lobes are simply evaginations of the outer cuticle of the pear-shaped body. The most backward ones are encrusted with calcium. At their base they are covered by short chitinous spines. These are homologous to the calcareous plates of the hermaphrodite. On the cuticle of the males there are small setae which form little hooks; they are united into groups of two to six, rarely more, and arranged irregularly over the whole external surface. These cuticular structures are very numerous in the anterior part of the animal, as well as on the lateral lobes. There are no cuticular structures on the antennae. In the body proper three regions may be distinguished: the head region, the thorax, and the abdomen. The head region is very much reduced. It is just a simple projection without any mouth or mouthparts. Darwin

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

271

(1851) describes the thorax as being within an inner sac or tube, about 0·4 mm long. It is transversely wrinkled and so extensible that it stretches to twice its former length. Darwin describes four pairs of short limbs on the thorax, whilst Gruvel (1899) talks about six pairs of cirri. This difference is due to the variability of the males. In the animals Darwin (1851) describes the bases of the limbs almost touch each other. The limbs have no articulations except where they unite with the thorax. The anterior smallest limb supports two or sometimes only a single spine. The second pair of cirri has three spines a little shorter than in the first pair. The two upper pairs of cirri are alike. They have two spines on their tips and a third lower down on a notch on the outer side. It is possible that the spine on the notch marks the point where in the larva there is an articulation. In the animals Darwin investigated there were 11 pairs of spines on the thoracic limbs. The spines are straight, long and not plumose. According to Darwin (1851) the four pairs of cirri must correspond to the four posterior cirri as may be inferred from their proximity to the abdominal lobe and from the three posterior pairs, closely resembling each other and differing a little from the first pair. The first pair corresponds with the third pair in the hermaphrodite. Gruvel (1899) describes the six pairs of cirri as looking all alike. There is a basal part (presumably the pedicel) on to which one or two setae-like cirri insert, which get smaller from the first to the last one. The first pair of cirri carry a single long flexible seta, absolutely smooth and ending in a fine tip. The second and third pairs have three setae, two of which are equally long and next to each other and one is a little more lateral and very short. The fourth pair of cirri does not have the two first setae of the previous pairs and the fifth and sixth pair do not have the smooth setae like the first one, but a very short one, flat and bigger at its base. At the level of the fifth and sixth pairs of cirri there are on the ventral surface three cuticular spines which are simply ornaments. The abdomen is reduced to two small cones (Gruvel, 1899). Darwin (1851) talks about one square abdominal lobe. According to the latter author there are three pairs of spines on the ventral surface, which may probably mark the three segments, which are distinct on the abdomen of the larva in the last stage of its development in Lepas and in other genera. On each of the posterior angles of the abdominal lobe there are three moderately long, very sharp spines with the tips of the outer pair bent a little inwards. Gruvel (1899) states that on each abdominal cone there is a strong short seta, similar to those of the fifth and sixth pairs of cirri. Within the thorax there are some longitudinal muscles without transverse striations which enter the short limbs but not the abdomen. At their lower ends these muscles terminate abruptly. They extend a short way beneath the lower pair of limbs and are attached to the outer integument of the animal near the base. Gruvel (1899) describes distinct muscle bundles which are orientated longitudinally. They are united by connective tissue. The arrangement of these compact muscles varies a little depending on whether they insert on the posterior or the anterior part of the sac, where some of them sometimes insert with a broad base. The contraction of these muscles causes a pulling together of the anterior and posterior part of the sac and the closure of the orifice. When the contraction ceases the orifice opens through the elasticity of the sac. There are no circular and no oblique fibres. The inner mantle surface is covered by a very thin cuticle. There is neither a mouth nor a digestive tract and no mouthparts. Darwin (1851) did not see any anus but he thinks that it may exist. The genital apparatus consists of a well-developed testis and the vesicula seminalis. Both organs are only single. (In the hermaphrodite these organs are paired.) There are no muscles in the genital tract, nor are there any muscles in the ejaculatory duct. The wall of the latter is a simple epithelium. The genital opening is between the caudal appendages on the ventral side of the body. The sperms of the male are identical with those of the hermaphrodite. The thorax serves for the emission and first direction of the spermatozoa and

272

WALTRAUD KLEPAL

“hence perhaps its singularly extensible structure” (Darwin, 1851). The longitudinal and transverse muscles lining the upper part of the outer integument of the whole animal could perhaps cause the violent expulsion of the spermatozoa, thus causing them to reach the appropriate area within the hermaphrodite. There are two ganglia in the male: a double one before the head projection and a single one in the median line of the ventral side of the thorax (Gruvel, 1898). From the latter small nerves lead into the appendages. The two ganglia are united by lateral connectives. From the cervical ganglion there lead two small connectives towards the anterior side of the body. They widen into a small nerve cell in which a nucleus may be seen and then they lead into a pigmented mass, which Gruvel (1898) believes to be the eye. Within the eye there is an anterior-posterior partitioning which indicates that originally the eye consisted of two parts. In the anterior part there are two nuclei of the pigment cells and enclosed in these cells there are two small elongated vesicles which surround a certain number of refringent rods which are orientated towards the outer opening. Gruvel (1899) states that the eyes and the setae are the only sensory organs. Darwin (1851) talks about a single eye only. According to this author the eye has a pointed oval form, consisting of an outer capsule, lined with purple pigment cells and surrounding something that looks like a lens. Gruvel (1899) describes the cement glands as being found on either side of the animal. There are two big glands of short and stout cells with big nuclei. In the histological sense they resemble the pancreatic glands described in other cirripedes. There is a big lumen in the gland from which a canal arises and which penetrates the base of the antenna. The antennae with the cement ducts are well developed. Each antenna consists of two limbs, the basal one is long and has a lateral seta, the other is very short and has a very small external appendix with very small and simple setae. The second segment of the prehensile antenna, the disk, is pointed and hoof-like (Darwin, 1851). There is a simple backward pointing spine, attached on the under side, nearly opposite the articulation of the ultimate segment. At the apex there are some excessively minute hairs or down. The ultimate segment projects rectangularly outwards as usual and has on its inner side a conspicuous notch, which bears two or three long, non-plumose spines. On the outside of the large basal segment there is a single spine curving backwards. The male of S. ornatum is very similar to that of S. vulgare. Darwin (1851) described the complemental male of S. peronii, as did Gruvel (1901, 1902a) and also Krüger (1914). Krüger talked about Smilium peronii. There are up to three complemental males within each hermaphrodite. The largest male found is 0·9 mm long and about 1·0 mm wide (from the tip of the tergum to the tip of the rostrum) and it is about 0·6 mm wide from the lowest corner of the rostrum to the corresponding one of the carina. Peduncle and capitulum may be easily distinguished. On the capitulum there are six plates: two terga, two scuta, a carina, a rostrum. All plates are well developed, and they are united by a finely-villose membrane, furnished near the orifice with much longer and thicker spines. The capitulum has the orifice not in the same line with the peduncle but almost transverse to it and therefore it is almost parallel to the surface of attachment. According to Darwin (1851) the scuta and terga are broadly oval with the primordial valves very plain at their upper ends. The carina is straight, triangular and internally slightly concave. The rostrum is shorter and internally more concave than the carina. The capitulum and peduncle are covered by fine setae. The narrow and very short peduncle commences a little below the scuta. The base is flat and truncated. There are longitudinal and oblique muscle strands in two layers. In the capitulum there is the M. adductor scutorum. The animal has a well-developed mouth and well-developed mouthparts. The labrum is highly bullate, as in the hermaphrodite. It is far removed from the M. adductor scutorum. The palpi are small and triangular. Their apices are blunt and clothed with very few scattered bristles. The mandibles have only three teeth, the lower angle is slightly pectinated. The first tooth is some distance from the second and larger than it. The maxillae bear only a few spines, furnished with a long apodeme. Beneath the upper large pair, there is a

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

273

notch under which there are two spines of considerable size and a small tuft of fine bristles. The relative size of the maxillae and mandibles are the same in the male and in the hermaphrodite. The outer maxilla is blunt, triangular with a few thinly scattered bristles on the inner face, those on the outside being longer. The digestive tract consists of the mouth, a short oesophagus, the stomach, a very narrow intestine and an anus. It is functional; diatoms were found inside the gut. The well-developed prosoma has six pairs of cirri. The first pair is far removed from the second, the rami are very short, barely exceeding the pedicel in length. They are formed of four segments, each bearing a pair of spines. On the end of the terminal segment there are three spines, the central one of these is very long. The second pair of cirri resembles the first pair and is also short. The third, fourth, fifth, and sixth pairs of cirri get gradually longer. All cirri have setae. In the sixth pair five of the six elongated segments have three pairs of long spines. The dorsal tufts are large. The cirri are furnished with transversely striated muscles just as in the hermaphrodite. The number of podomeres in the cirri of the male is variable, and considerably smaller than in the hermaphrodite. Krüger (1914) gave the number of podomeres in the cirri of the male as between three and five. The caudal appendages have no articulations, they are minute plates with a few bristles at their apices. The testes are paired and very irregular in their shape. The vesiculae seminales are long; they have muscles which help to extrude the sperms. There is a short penis, extending only up to the pedicel of the sixth pair of cirri. It has four bristles at its end. Gruvel (1902a) believes that he may have been seen immature ovarian cells, but nobody before or after him has seen such cells. The nervous system consists of a dorsal mass, situated at the base of the mouth cone. This is the cerebral ganglion which consists of two lobes. From this two small nerves arise. These go past the stomach and they are probably the two peduncular nerves. Inside these there arise two other nerves, which are certainly optical nerves. The cerebral ganglion is united with a long thoracal mass, the thoraco-abdominal chain, which is very much condensed and serving the mouthparts and the cirri. The eye in the central line between the scuta is a single and simple pigmented mass, which contains refringent bodies. It resembles very much that of Scalpellum vulgare. The cement apparatus consists of two lobed glands with large nuclei and several nucleoli. The cells are very granular and they have a free central space into which the secretory products are passed. According to Gruvel (1902a) the complemental male of S. peronii resembles closely the hermaphrodite. The hermaphrodite itself resembles closely individuals of the genus Pollicipes, an ancient type of cirripedes. Scalpellum villosum is broader and considerably higher than the male of S. peronii. The orifice of the capitulum is placed obliquely; the membrane connecting the plates is finely villose, furnished with spines, conspicuously thicker and longer than those on the male of S. peronii. The peduncle is naked, narrow and short. The capitulum is about 0·1 mm long and wide. Of the six plates the scuta and terga are much more elongated than in S. peronii. The carina descends only just below them. The rostrum is a little broader and more arched than the carina. The primordial valves are seated on the tips of the scuta, terga, and carina, but not on the rostrum. Scuta, terga, and carina of the male resemble the same plates in the hermaphrodite much more closely than in S. peronii. In both species the rostrum is of large relative size in the complemental male. This is a remarkable character which is difficult to explain (Darwin, 1851). There are six pairs of cirri. The first pair is short with only three or four podomeres in each ramus. The second cirrus has the basal segment not very thickly clothed with spines. The sixth cirrus, has seven segments, not protuberant in front, each bearing four pairs of spines without intermediate tufts. There are no caudal appendages as in the hermaphrodite. The disk of the prehensile antennae is narrower than the basal segment, only slightly pointed (different in S. peronii). At the distal end on the inner side there are two or three spines at the same place

274

WALTRAUD KLEPAL

where there are usually excessively minute hairs in some or all other species of Scalpellum and Ibla. The digestive tract is functional. The labrum is bullate with teeth on the crest, the palpi are blunt and spinose. The mandible has three teeth. Its inferior point is rather strongly pectinated. The maxilla has a considerable notch under the upper pair of large spines. The inferior part of the edge is not prominent. On the inner edge of the outer maxilla spines are arranged in two groups. The penis is thick, exceeding the length of the pedicel of the sixth cirrus. It is square at the end and furnished with some spines. The complemental males of Scalpellum villosum and S. peronii closely resemble each other, more so than the corresponding parts in the two hermaphrodites. S. pilsbryi is probably identical with Calantica calyculus (see above). The size of the male is less than 1 mm and there are 11 plates on the capitulum. These are one carina, two terga, two scuta, one rostrum well developed, one subcarina and two pairs of lateral plates. There are six pairs of cirri, the separation between the first and the subsequent pairs of cirri being less clear than in the hermaphrodite. The caudal appendages are not articulated, with one very long seta and four very short ones. The digestive tract is functional. The mouthcone is like that of the hermaphrodite. The penis is beset by setae, less numerous than in the hermaphrodite. In Scalpellum scorpio the peduncle and capitulum can be distinguished, both being beset by fine hairs. There are no real segments. The six plates are scuta, terga, rostrum and carina; the carina is the longest and is weakly bent, the rostrum is smallest, but bent more. The first pair of cirri is far away from the second. Each ramus consists of five podomeres; the rami are shorter than in the hermaphrodite. The caudal appendages reach only to the middle of the proximal segment of the protopodite and they are like those of the hermaphrodite. The digestive tract is functional. The mandible has four teeth, the next to the outermost is smallest. The mouthparts are similar to those of the hermaphrodite except that the maxilla has less spines. The penis has nearly half the length of the sixth pair of cirri. At its distal end it is much thicker than the cirri at that height. This is a difference between the male and the hermaphrodite (Aurivillius, 1894). The male of S. longirostrum is similar to S. peronii and S. villosum (Gruvel, 1902a). It has a laterally compress capitulum. The cuticle on the capitulum has short setae, arranged irregularly. The peduncle is short without any setae, or the setae are here very much reduced. Since the peduncle is very short the cement glands are entirely at the base of the capitulum. The shape of the six plates is very different from those of S. peronii or S. villosum. The cirri are very similar to those of S. peronii and S. villosum. Each ramus consists of three podomeres and there are three to four setae on the basal limbs. The caudal appendages consist of one limb only which is cylindrical and there are two terminal setae. The alimentary canal is reduced to a short oesophagus and a very globular stomach. The mouthcone is less prominent than in S. peronii and S. villosum. The labrum is big with chitinous nodules. The labial palps are cylindro-conical, very short with some setae. Mandible and maxillae are well developed. The testes are also well developed, the vesicula seminalis is reduced, leading into the ductus ejaculatorius. The penis is very short and probosciform with a few terminal setae and without any distinct annulation. The males of several species of Scalpellum are sac-like, so that capitulum and peduncle cannot be distinguished morphologically (Table I). One exception is e.g. S. (Scalpellum) elongatum which has a short peduncle. These males are smaller (<1 mm long) than the previous ones with six large plates and they have only four plates (Nilsson-Cantell, 1931; Stubbings, 1936). These may be well developed as in S. bengalense and S. hoeki, or they may be rudimentary as in S. gigas, S. striatum, S. luteum, S. intermedium, S. kurchatovi, S. elongatum, S. ornatum, S. rutilum, S. stearnsi, S. discoveryi, S. tritonis, S. wood-masoni, S. retrieveri, and S. vulgare.

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

275

The male of S. hoeki has the shape of an elongated sac. The inferior portion is globular, the antennae are very low down the ventral portion. The surface of the sac is chitinous with very short setae, arranged somewhat regularly. Beneath the cuticle there is a layer of longitudinal muscle fibres. The space of the mantle cavity is very much restricted to a simple cylindrical tube in which the cirri may be moved. The male has four well-developed plates, which are better calcified than in S. gigas. In the complemental male of S. gigas capitulum and peduncle are not distinct. There is no Musculus adductor scutorum. In the sac are a number of parallel muscle strands; one part of them inserts around the external opening, the other branching, inserts on the periphery of the sac. Apart from longitudinal muscle fibres there is a layer of oblique muscles, nearly circular. When they contract, they close the orifice (Gruvel, 1902a, b). The nervous system consists of a dorsal mass situated at the base of the mouthcone and an elongated thoracal mass. Gruvel did not find any eye. The sensory organs are very much reduced. The cement glands are below the testes. There are four rudimentary plates around the opening of the sac. These plates are probably the terga and scuta. The six pairs of cirri are more or less atrophied. The first pair has a basal limb without segmentation, it is cylindrical, ending in two setae. The second pair is better developed. The third and fourth pairs are shorter than the preceding, the fifth and sixth are very much reduced. Very much reduced appendages on the abdomen correspond to the caudal appendages of the hermaphrodite. The mouthcone is reduced to a knob, the anterior portion consists probably of a labrum and two pairs of cylindroconical mandible and maxilla. From the mouth there leads an extremely short canal which widens into a kind of sac=stomach, which is apparently without any function. The genital organs are well developed and very large. This seems to counterbalance the other organs. The testes are paired, there is only one vesicula seminalis from which a single median ductus ejaculatorius arises. The males of S. striatum are similar to those of S. hoeki. On the cuticle there are setae on the circular line and almost parallel all over the surface. No muscles were seen in the mantle (Gruvel, 1902a). The mouth is just a simple ridge of cells. The cirri are very much reduced as in S. hoeki. The genital organs do not show any particular features. The vesicula seminalis is very long as in S. gigas. The cement glands look like those in any other species. The male of S. intermedium is elongate, ovoid, about 0·75 mm long. It has chitinous hairs over the whole surface. There are four rudimentary plates. A tuft of hair-like cirri protrudes from the opening between the rudimentary plates (Hoek, 1883). According to Nilsson-Cantell (1921) the cirri are rudimentary and articulated. On the apical end of each segment is a ring of setae, the uppermost segment having longer setae. The number of cirri is difficult to make out, probably there are only two. S. (S.) elongatum has a sac-like male with a small round or oval mantle opening. On the outside it is covered with short, simple setae. The peduncle is short with longitudinal muscle fibres. The male has four small reduced plates around the mantle opening. The testis is a large spherical body in the centre of the animal. S. ornatum has a dwarf male in a transverse pouch on the under side of each scutum on the female. The general shape is elongated. The whole outer integument is thin with minute points arranged in transverse rows. The thorax is highly extensible, when stretched it exhibits five transverse folds or articulations; the abdominal lobe is rather short. The four plates are calcareous and rudimentary, thin and regularly oval. Two plates on one side are smaller than two on the other side—this is probably due to one side being pressed against the hard shelly valve of the female. The plates correspond to the scuta and terga in other cirripedes. The four pairs of cirri have short and thick spines; each limb with three spines. One of them is on a notch, low down on the outside, and it is longer than the other two. The spines on the first and second pairs of limbs are considerably shorter than those on the third pair and those on the third pair are a little shorter than those on the fourth and posterior pair. The antenna is a disk, hoof-like. The apex is pointed and clothed with

276

WALTRAUD KLEPAL

some fine down. From the lower flat surface rises a single spine pointing backwards. A single spine is on the outer edge of the basal segment. On the abdominal lobes there is a row of six moderately long and basally thick spines. The abdominal spines altogether form a brush. The males have no mouth or stomach. The male of S. rutilum has the outer integument covered with thick minute bristles. It has four calcified rudimentary plates. On the thorax are small limbs and spines. The antenna has a pointed hoof-like disk. S. stearnsi has dwarf males which, because of their number (more than a hundred were found attached to mm, the same female) give the surface of the female a crusty and grainy appearance. The males are sac-like with four rudimentary plates. Short rudimentary tentacles are attached to the surface of the mantle between small plates. These have nothing to do with the articulated antennae or other limbs of the cirripedes. On small females the same tentacles are found. These tentacles in both sexes at the tip of the capitulum attached to the chitinous surface are between the terga near the anterior extremity of the orifice— in the females large, in the males relatively much smaller. This gives entrance to the cavity in which the animal’s body is lodged. This is an example of polyandry of which probably no other one is known in animals so highly developed as the Cirripedia (Hoek, 1906, 1907; Nilsson-Cantell, 1921). The surface is hirsute with very small spines. The complemental male of S. discoveryi is sac-like and much reduced. The sac is provided with muscular bands; at the side of the pole opposite the antennae is a small lobe with the mantle opening. There are four much reduced plates. The complemental male of S. tritonis has its surface clothed with rows of very minute spines. The four rudimentaty plates are probably scutum and tergum. They are not exactly of the same size. A longer and a smaller one are placed close to one another and the two are separated by a somewhat greater distance from a quite similar couple of plates. The complemental male of S. wood-masoni is sac-like, with four rudimentary plates round the mantle opening. There are minute spines on the surface; there are muscular bands as in other reduced males. The male is about 0·8 mm long (Fig. 4E). The male of S. retrieveri is sac-like. The surface is covered by very small transversely grouped spines. The muscle fibres of the sac are distinct. Four small reduced plates are placed around the mantle-opening. The male is very much like that of S. luteum as figured by Gruvel (1902a). It is different from S. intermedium in being much more reduced. In some of the males the reduction of the plates seems to progress further: two of the four plates are smaller than the other two (Table I). This is the case in S. chiliense, S. galapaganum, and S. gracile. In all other cases of males with four plates the cirri are reduced, either in length as in S. gigas, S. hoeki, S. striatum, S. luteum, and S. vulgare (Gruvel, 1899) and/or in number as in S. ornatum and S. vulgare (Darwin, 1851) with four pairs of cirri or in S. intermedium with two pairs of cirri or hair-like projections. When caudal appendages are described they are very short as in S. gigas or there may be just spines as in S. ornatum. There are no data on the presence and/or shape and equipment of a penis. In the male of S. rostratum Darwin (1851) (Fig. 4D) described three plates (an oval carina and a pair of scuta). The peduncle is separated from the capitulum. The male has six pairs of cirri, the first pair being separated from the following ones. The pedicels of all the cirri are very long relative to the rami, thus all the cirri appear immature. The rami are slightly unequal in length, even in the sixth pair. There are six podomeres in the rami of the sixth pair, each podomere bearing two to three pairs of long spines. The caudal appendages have two to three small spines on their tips. The penis is short, blunt and has a thick apex with one or two spines on it. The capitulum and the peduncle are covered with spines, mingled with shorter ones in short rows of three and four together. The thorax is unusually elongated.

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

277

The labrum, largely bullate, is placed far from the adductor as in the hermaphrodite. The mandibles have one tooth less than in the hermaphrodite (three large, sharp teeth, with the inferior point very sharp and small). The maxillae have two or three large upper spines, the others are very thin. The outer maxillae are bilobed in front with a few short bristles on the outer side near the bottom. There is a close similarity between the mouth of the male and that of the hermaphrodite. Within the genus Scalpellum there are also species whose males are sac-like without any plates but with six pairs of cirri like S. regium. In this species the first and second pair of cirri are reduced, the third to the sixth pair are well developed. In S. distinctum the cirri are only spine-like. In S. gibberum there are three pairs of cirri. In the descriptions of all other species whose males do not have any plates the cirri are not described. Examples are S. convexum, S. compaction, Pisiscalpellum withersi, Scalpellum projectum, S. chitinosum, S. javanicum, S. sessile, S. crinitum, S. hexagonum, S. fissum, S. compressum, and S. gruvelianum. In S. alcockianum Nilsson-Cantell (1931) mentions that there are no cirri. The same is true for S. condensum and S. convexum amongst others (Nilsson-Cantell, 1921). In all these mentioned species there are no data on the presence of any penis or the caudal appendages. The digestive tract is functional in all males with at least six plates on the capitulum. It shows signs of reduction as in S. longirostrum, whose cirri are reduced and which has only a short penis. The caudal appendages are still well developed in this species. Some of the males with only four plates have a functional digestive tract (i.e. Scalpellum bengalense, S. gigas, S. hoeki, S. striatum, and S. luteum). In most species no data are given on the digestive tract. S. rostratum (Fig. 4D, p. 306) with only three plates but a separation between capitulum and peduncle, six pairs of cirri, a well-developed penis and short caudal appendages has a functional digestive tract. In S. regium the sac-like male has no plates (Fig. 4F, p. 306) and the digestive tract is functionless. This male has no penis but still six pairs of cirri (although two pairs of them are reduced in length). Of all other males without any plates there exist no data on the presence or state of any digestive tract. This could mean that no digestive tract could be traced or that only the external morphology was described. Nilsson-Cantell (1921) described the dwarf male of S. convexum as sac-like, about 0·8 mm long, with a lobe on the opposite side of the antennae. The antenna consists of three articulations. The muscles are transverse ribbons. On the cuticle are reduced comb-like scales. There are plates and no cirri. The dwarf male of S. compactum is sac-like, about 0·67 mm long, two antennae on one pole, and on the opposite pole there is the opening of the sac on a protrusion. Oblique parallel muscles are below the epidermis. The cuticle has transverse rows of comb-like scales. This male has no plates and no cirri. Pisiscalpellum withersi was described by Utinomi (1958). The dwarf male is attached on the inside of the scuta near the apex; there is a slight depression on the inside near the apex just above a deep hollow for the adductor muscle, where the mantle edge is produced inwards forming a pouch. The male is sac-like, capitulum and peduncle cannot be distinguished. It is 0·4 mm long and has “minute hairs” (probably small cuticular projections like teeth) in transverse rows on the surface. As in the previous species there are no plates and no cirri. Scalpellum projectum has dwarf males (Nilsson-Cantell, 1930a, b). The largest number of males found was four on both sides. The male is sac-like without any plates. It may be 2·2 mm long; two small lobes are at the pole where the opening for the genital organs is situated. In a younger animal these processes are missing. The male can thus have a very varying form. The dwarf males of S. chitinosum are pear-shaped. A large and a distinct opening leads into the cavity of the sac; there are short, spine-like hairs on the surface. Larger and more numerous hairs are found on the peduncle. The male is about 1 mm long.

278

WALTRAUD KLEPAL

Hoek (1907) described the dwarf male of S. javanicum as ovoid, about 1 mm long and 0·5 mm wide. There are spine-like hairs on the surface. These are longer and more numerous on the peduncular than on the capitular pole. On the capitular pole is an entrance to the sac. There are no plates, but there are slightly swollen marginal parts which may stand for rudimentary plates. Only the male reproductive organs are in a good state of development. The dwarf male of S. sessile is pear-shaped, with the peduncular pole being the narrower one. It is about mm. S. crinitum has oval dwarf males about mm. The muscle fibres of the sac are very distinct; the opening of the sac is hardly visible. The dwarf male of S. hexagonum is broad oval, about mm, extremely short and has delicate hairs on the capitular pole. mm. On the surface are very short hairs or The dwarf male of S. fissum is large, ovoid, about spines, standing alone or in groups of three or two without forming regular rows. Slightly longer hairs form a little tuft on each side of the opening giving entrance to the interior of the sac. The muscles of the sac are strongly developed. The genital apparatus is well developed. S. compression has a complemental male attached to the hermaphrodite on the interior surface of the right hand scutum, close to the occludent margin. The male is pear-shaped, about 1·3 mm long. On the outer surface are very small bristles. The antenna consists of one segment only. The testis is long and heartshaped, the vesicula seminalis is relatively small; it consists of an oval sac and communicates with the testis and with the vas deferens. In S. condensum the dwarf male is sac-like, about 0·67 mm long. The muscles are fine oblique bands; spines are directed downwards. The male has neither plates nor cirri. Iblidae Five species of Ibla are now recognized. These are Ibla quadrivalvis, I. cumingi, I. segmentata, syn. I. pygmea, I. idiotica, and I. atlantica. Only I. quadrivalvis and I. pygmea are hermaphrodites and thus have complemental males. The other species are gonochoristic and have dwarf males. Not many authors dealt with Ibla sp. Some were interested in the larval development. Anderson (1965) has described the larval development of I. quadrivalvis, and Karande (1974) that of I. cumingi. Gaonkar & Karande (1980) have described the life history of I. cumingi and Achituv & Klepal (1981) were interested in the ecology of I. cumingi in the Gulf of Elat. Annandale (1911, 1916), Stubbings (1967), For (1972), and Foster (1978) have discussed the validity and synonymy of the species. The males of I. cumingi and I. quadrivalvis were described by Darwin (1851). This author gave the best account on the anatomy of the two species, based on the methods available in his time. He made a comprehensive study of Ibla sp. Other authors who dealt with Ibla sp. are Gruvel (1902b), Stewart (1911), Broch (1922) and Hiro (1936). With the exception of Batham (1945) none of them dealt with the anatomy in any detail. Bage (1908) dealt exclusively with the excretory organs of I. quadrivalvis. Stewart (1911) described the anatomy of the male of I. cumingi. This author gave a general outline of the various organ systems of the males. He did not go into any detail, apart from a description of the sensory hairs. Klepal (1985) compared the anatomy of the female and the male of I. cumingi and compared it with the other species of Ibla described so far. Emphasis was put on the skeletal, muscle and nervous systems; these were described in great detail. The major aim of the paper was to obtain a holistic outlook on evolutionary aspects like dwarfing, reduction, and systematic implications. One important question was whether the male was simply a smaller version of the female (a “geometrically scaled down adult”; Gould, 1978) or whether it had undergone rudimentation. The vestigial organs of the male are important as one of the proofs of evolution (Osche, 1972). They are an indicator of the degree of dwarfing in the male and of

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

279

the trend of evolution of the Cirripedia s. str. Pilsbry (1908) pointed out the significance of the male cirripede for systematics in the Scalpellidae. Females whose capitulum width is about 5 mm have between one and six males attached to their mantle tissue. All males within one female are usually of different sizes and presumably also of different ages. The bigger the females the larger the possible size of the male. The males are usually attached next to each other in the median line of the scutal end. Normally the males enter immature females and they seem to develop together until they both reach sexual maturity at about the same time. The female then seems to outlive the male, which is generally the rule in cirripedes. The life span of the male of I. cumingi is not known. Experimentally isolated males lived for about one year in a Petri dish. The isolated males move with their mouthcone over the bottom of the Petri dish. Thus they seem able to sweep off small particles and/or bacteria from any surface, e.g. the body surface of the female under natural conditions. It is evident that the males have a mode of life different from that of the females. The male moults about once a month. The narrow part of the peduncle, under natural conditions embedded in the female tissue does not shed the cuticle during moulting, alternate layers of cuticle and cement are seen in this area. During moulting, but also without the stimulus of moulting, the male conducts “swaying” movements with the peduncle. The males react vigorously whenever food is added to the water in which they are kept during an experiment. They move their mouthparts and also their antennules provided the latter are not completely surrounded by cement and cuticle. Other movements the males are known to conduct are the protrusion of the thorax together with the cirri or its withdrawal into the oblique fold and also the protrusion or lowering of the mouthcone relative to the thorax. The larger cirri may be moved towards the smaller ones and the smaller cirri are moved with the whole thorax. Food is not the only stimulus for movements of the male. Another stimulus is light. Water currents do not seem to have an effect and there is no reaction to touch. The male of I. cumingi is up to 2 mm long and 0·5 mm wide. It consists of a long peduncle (about 2/3 of its total body length) terminating in the antennules, the thorax with two to three pairs of cirri, a relatively large mouthcone and a small abdomen with blunt caudal appendages. There are no opercular valve plates as in the female. The male has an oblique fold (the mantle) separating the peduncle from the mouthcone and the thorax. This fold, protecting the thorax, is homologous with the capitular plates in the female (Darwin, 1851). In the male the closure of the mantle cavity is provided by the mouthcone, especially by the labrum. Compared with the female, the male has a relatively short ventral side and an extensive dorsal side (see Klepal, 1985, her Fig. 4). The labral side of the male corresponds to the rostral side of the female and the side of the oblique fold of the male corresponds to the carinal side of the female. The cuticle is an important skeletal element in the arthropods in general. In the males of I. cumingi an exocuticle may be distinguished from an endocuticle, but when the cuticle is very thin (about 0·6 mm), e.g. around the excretory papilla, no distinction of the various cuticular zones is possible. In the female the cuticle (epi-, exo-, and endocuticle) may be twice to eight times as thick as in the male. In the male the cuticular structures are simpler than in the female. In the male there are three types of combs, whilst in the female there are five. Teeth and combs are on the peduncle as well as on the thorax. The apodemes, thickenings of the body wall on to which muscles insert are more distinct in the male than in the female of I. cumingi (Klepal, 1985). There are four endosternites in both sexes of I. cumingi. The position of the ventral endosternite varies in both sexes. It is ventral of the oesophagus and in the male on the ventral side and in the female on the dorsal side of the suboesophageal ganglion (Fig. 5A). The position of the dorsal endosternite varies in the male. Here it is either ventral or dorsal of the suboesophageal ganglion, whilst in the female it is always dorsal of the suboesophageal ganglion. In spite of the different position of the ventral endosternite relative to the suboesophageal ganglion in the male and in the female the

280

WALTRAUD KLEPAL

same muscles insert in both sexes except for the muscles which are missing in the male (inserting on the lateral part of the ventral endosternite in the female). The histology of the endosternite agrees with that of the apodemes and the connective tissue. In the fine structure there is a difference between the apodeme and the endosternite. The endosternite consists of a finely granular matrix into which fibres, about 0·01 µm wide are embedded. The structure of the endosternite is similar to that of the basement membrane but the endosternite contains a larger number of fibres. The ground substance of the endosternite is also similar to the intercellular matrix. There is no regularity in the arrangement of the fibres in the endosternite except near the surface where they are arranged parallel to the long axis of the inserting muscle fibres. In the male and in the female the junctions between the endosternite and the neighbouring muscle cell or parenchyma cell is formed by hemidesmosomes. The muscular system of the male is very complex (Klepal, 1985). There is a system of transverse and longitudinal muscles (Fig. 5A). Several muscles in the male are identical with the corresponding ones in the female, e.g. the muscles of the mouthparts. In general the muscles are smaller (less well developed) in the male than in the female. Some muscles of the female are missing in the male, e.g. the M. adductor scutorum, the M. levator primus, M. levator secundus, M. levator tertius, M. abductor capitis, M. depressor capitis, M. supinator capitis, M. dilatator ventriculi, M. longissimus inter-medius, M. attrahens pedis, M. obliquus, M. retractor pedis, M. flexor pedis, M. extensor pedis, M. arrector aortae, M. “abductor branchiae”, M. arrector pedis, the outermost layer of diagonal muscles in the peduncle and all except one transverse muscle are also missing. On the other hand, some muscles present in the male are missing in the female, e.g. M. compressor frontalis, M. compressor lateralis, M. extensor ducti, M. extensor longitudinalis. All these last four muscles are in connection with the excretory system in the male. It may be assumed that they are all descendants of muscles in close proximity to the excretory system in the female. The development of intrinsic muscles of the excretory system in the male may be due to the small body of the male and the relatively large size of the excretory system. Thus the growing together of muscles and organs which are quite distinct in an animal of bigger size may occur. The M. compressor frontalis in the male may be derived from a M. transversalis in the female, the M. compressor lateralis may be part of the M. transversalis. The M. extensor ducti may have been derived from the M. baseo-maxillulae and the M. extensor longitudinalis may have been the M. arrector pedis of the female. The male of I. cumingi may have up to three pairs of cirri, but two pairs are most frequent. The cirri next to the caudal appendages are bigger than the outer ones, which are closer to the mouthcone. Not only the number of cirri is variable, but also the number of rami per cirrus (variable between one and two) and the number of podomeres (variable between one and four—a pedicel is rarely distinct). Several features distinguish the cirri of the male from those of the female. On the tip, the dorsal and lateral sides of the cirri there are a variable number of smooth setae, on the ventral side there are none (whilst in the female most setae are on the ventral side of the cirrus). The non-sensory cuticular structures are also different in males and females (Klepal, 1983). The diameter of the base of the cirri is in the female about twice as big as that in the male. An obvious feature of the cirri of the male is that there are no muscles. The number of dendrites in the two nerves is considerably smaller (five to ten) than in the nerves of the female (about 160). Instead of the lacunae of the female there are wide irregularly shaped intercellular spaces between the parenchyma cells in the males. The antennules of the male were described by Darwin (1851), and fine structural detail has been given by Klepal (1985). Three segments may be distinguished. These are presumably segments four, three and two (or two+one) of the cypris antennule. There are up to six sensory setae on the outer distal end of segment four. On the attachment disk at the end of segment three there are cuticular villi, about 5 µm long and 1·5 µm wide. About seven sensory setae are either at the level of the disk or just above it. There is one opening of

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

281

Fig. 5.—A, anatomy of the adult male of Ibla cumingi seen from the labral side, peduncle and cirri cut off, foregut partially removed, on right side parts of muscles and organs on surface removed, on left side excretory organs removed, peduncular muscles not drawn (after Klepal, 1985); B, anatomy of the adult male of Ibla idiotica, seen from the ventral side (after Batham, 1945); a, antenna; apI, apodeme I for the insertion of head and body muscles; apII, apodeme II for the insertion of the transverse and longitudinal body muscles; apIII, apodeme III for the insertion of head muscles, longitudinal body muscles and longitudinal muscles of the peduncle; c, cirri; ca, caudal appendage; e, eye; eB, endosternite for attachment of mouthparts; eC, endosternite between efferent ducts of excretory system; eD, ventral endosternite; ex, excretory system; gt, gut; lb, labrum; lm, longitudinal body muscles; md, mandible; me, muscles of excretory system; mx1, first maxilla; mx2, second maxilla; nc, nerve cord; pg, pancreatic gland; pmd, palpus of mandible; sa, salivary gland; sg, suboesophageal ganglion; spg, supraoesophageal ganglion; st, stomach; t, testis; tm, transverse body muscles; vs, vesicula seminalis; scale bars, A and B=0·2 mm.

282

WALTRAUD KLEPAL

the cement duct in the centre of the attachment disk. The cement cells, at the origin of the duct in the peduncle are arranged in a grape-like fashion. From the cells small and short connecting ducts lead into the main duct. The male shows the simple pattern of the cement apparatus whilst the female shows the more advanced one. The cement cells of the male being so different from those of the female may be the reason why Darwin (1851) did not recognize them. In its morphology the antennule of the male resembles closely that of the cypris (of I. quadrivalvis). But there are no muscles in segments three and four of the antennule of the male, so that these segments are moved passively. The mouthparts of the male are well developed. In their shape and in their cuticular structures they are simpler than those of the female. In any case they are smaller than those of the smallest possible female. The mandible is simpler in the male in having three teeth and instead of the serrate edge of the mandible of the female there is one smaller tooth in the male. The other mouthparts (first and second maxillae and the labrum) resemble closely those of the female. In the male the labrum has the additional function of being the dorsal closing of the mantle cavity and not only the dorsal closing of the mouthcone as in all other Cirripedia Thoracica. When the animal is totally contracted the labrum and the oblique fold are pressed against each other— thus they form an effective protection of the thorax with its extremities and the rest of the mouthcone. Whilst the female has well-developed articulated caudal appendages, those of the male are only blunt projections on either side of the anus. They contain parenchyma and muscles. The alimentary tract is well developed in both sexes of Ibla. It consists of the mouth with the foregut, the midgut with the pancreatic glands and the hindgut with the anus. The salivary glands of Ibla are the labial glands. There are no suboesophageal glands. The granular inclusions of the gland cells in the male are presumably polysaccharide-protein complexes (Klepal, 1985). The genital system of the male of Ibla consists of two testes, two tube-shaped and coiled vesicula seminales whose front part is widened. From there two ducti ejaculatorii lead to one genital pore in front of the anus on the ventral side of the animal. The extension of the testes depends on the stage of maturity of the animal. Their most distal point of extension is in the peduncle on the inner side of and just above the cement cells. Around the vesiculae seminales there are strands of smooth circular muscles. The widened parts of the vesiculae may fuse into one pouch. Around the latter the muscle layers may be arranged diagonally and cross-wise against each other. The ducti ejaculatorii for about one quarter of their length are surrounded by strands of smooth, circular muscles. Amongst the epithelial cells lining the testis tube there are cells with long cellular projections as typical in Sertoli cells. These projections extend between the spermatogonia and between the spermatids. The testes and the vesiculae seminales contain cells in different stages of development. Spermatogenesis has been investigated in the males of I. cumingi (Klepal, 1985). The sperm of Ibla, ready to fertilize, has a diameter of 0·5 µm and is thus bigger than that of e.g. Balanus balanoides. It has an accessory droplet, which differs from the homogeneous appearance of B. balanoides (Barnes, Klepal & Munn, 1971) in having a reticulate structure. It differs from that of B. perforatus (Bocquet-Védrine & Pochon-Masson, 1969) by the lack of a cristalloid corpuscle, characteristic of nearly all spermatozoa at the early stages of development and in having more than one axial rod. The accessory droplet of Ibla is reminiscent of the “Nebenkern” of other crustacean spermatids (e.g. Argulus; Wingstrand, 1972) or insects (Szöllösi, 1975), although the resemblance of the structure with a mitochondrium is only superficial. So far no “Nebenkern” was found in any cirripede spermatozoon. The spermatozoa of Ibla cumingi agree in their structure with those of other cirripedes described by Pochon-Masson, Bocquet-Védrine & Turquier (1970). They have a poorly developed chondriome. This is

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

283

correlated with low oxygen consumption (Barnes, 1962). It is possible that the accessory droplet takes the place of the chondriome in providing the energy necessary for the penetration of the ova (Barnes, Klepal & Munn, 1971). The sperms of Ibla (as those of other cirripedes) represent an intermediate form in the scheme of Franzén (1970) relating the mode of fertilization to the shape of the spermatozoa. The mitochondrium is in close association with the flagellum (indicating external fertilization) and the nucleus (indicating internal fertilization). As in other cirripedes in Ibla there is pseudo-copulation: the sperms are released into the mantle cavity of the functional female. The mantle cavity communicates with the outside, thus fertilization may be called ‘external’ (Walker, 1980). The nervous system of the male consists of two major ganglia, the sub-oesophageal and the supraoesophageal ganglion, and a small optic ganglion. There is no ganglionic chain. The supraoesophageal ganglion consists in the male of two fused pear-shaped lateral lobes. Each of these lobes is smaller and together they are wider than the suboesophageal ganglion. The latter is globular and has a distinct narrow and smaller anterior portion. The oesophageal commissure connects the supraoesophageal ganglion to the ventral side of the globular suboesophageal ganglion. The optic ganglion is in close proximity to the supraoesophageal ganglion. The eye rests on the optic ganglion. Two antennary nerves (Cannon, 1947) or perpendicular nerves (Brandt, 1871; Sewell, 1926) arise from the supraoesophageal ganglion on either side of the optic ganglion. They extend into the peduncle where they branch between the muscles and the gonads, and into the antennule where they supply the muscles and the sensory organs. A labral connective arises from the ventral side of the supraoesophageal ganglion. This connective runs along the dorsal side of the oesophagus and joins on to the labral commissure at the base of the labrum. In the labrum itself it branches. Lateral nerves arise from the supraoesophageal ganglion, close to the origin of the oesophageal commissure. The lateral nerves run along the pancreatic glands into the lateral body muscles. The nerves of the mouthparts and the cirri arise on the narrow anterior portion of the suboesophageal ganglion of the male. The narrow portion of the suboesophageal ganglion of the male of Ibla resembles that of Balanus tintinnabulum (Broch, 1927). In the latter species it consists of the fused thoracic ganglia II–VI. In both cases the thoracic ganglia (one or more) are fused with the suboesophageal ganglion. In general the nervous system of the male is less concentrated than that of the female. The ganglia in the male are separated more distinctly than in any other Cirripedia Thoracica. The nervous system of the male represents clearly a more primitive condition than the female or any other cirripede. The most obvious indication of this is the separation of the buccal ganglion from the rest of the suboesophageal ganglion (Bullock & Horridge, 1965). The morphology and fine structure of the excretory system are identical in the male and in the female of Ibla cumingi (Klepal, 1985). One big difference is the presence of what seem to be intrinsic muscles in the excretory system of the male. This was never before described in any other crustacean (see muscle system, p. 323 and Klepal, 1985). Gruvel (1902b) described the anatomy of I. quadrivalvis including that of the male. The male is between 3 to 5 mm long and about 1 mm wide (so it is about four times as long as wide). As far as could be seen from Gruvel’s results, the male of I. quadrivalvis is very similar to that of I. cumingi (see also Klepal, 1983). Stubbings (1967) described I. atlantica, but not in great detail and he did not give any figure of the male. In the male a capitulum may be distinguished from the peduncle. The antennule of the male is embedded in a mass of cement in the mantle wall. There are two testes and a vesicula seminalis. There are no plates. Broch (1922) described I. pygmea. In the female there are two or three complemental males. They are only about 0·6 mm long. The eye is most obvious; it is really two eyes fused to one. The antennae originate just below the eye. They consist of a large basal segment and two smaller distal ones. They do not seem to have any prehensile function. The cirri keep the shape of those in the cirripede pupae. Behind the cirri a

284

WALTRAUD KLEPAL

short penis is seen terminating in two long setae. A more detailed description of the males of I. atlantica and I. pygmea is still not available. Batham (1945) gave a good description of I. idiotica (see Fig. 5B, p. 322). The male is about 0·4 mm long, triangular and has two large fronto-lateral horns. It has no peduncle, no mantle, and no plates. The cirri are almost entirely absent. The trophi are absent. The males are transparent, nearly motionless and their muscle fibres are almost absent. The two spinous horns give the male a bizzarre appearance. These horns are not embedded in the female’s mantle cavity. Posteriorly the body terminates in two protrusions, each with one or two setae. Anteriorly to them there are two protuberances, each with one to three setae. There are no mouthparts. Whether these horns and/or the two protrusions are modified cirri of the male is not known. The antennules are deeply embedded in the body and lie between the testes. They are completely equipped with setae, but they have no muscles. There is only a minute pore on the median anterior margin by which the antenna communicates with the exterior. The alimentary canal is reduced. There is no oesophagus. The stomach is small, globular; the intestine is thin. Batham (1945) was not sure whether there was a lumen in the gut. There is no anus. The only supply of food for the male is presumably that already stored in the body. Since the stomach is full of contents it may be assumed that it has some function in the absorption and utilization of food. When the supply of food is used up then the testes and seminal vesicle become empty, the remnants crumple up and die. It is obvious that the males are shorter lived than the females but just how much shorter is not known. The nervous system is also reduced. There is a small solid brain and a black median eye. A nerve cord runs backwards for a short distance. Associated with the lack of an oesophagus there is the absence of a circumoesophageal commissure. Apart from paired testes (sometimes there is only one testis) there is an impair seminal vesicle which ends in an opening at the body surface. Occasional strands run from the testes to the ectoderm. From the bases of the antennules to the dorsal surface there run densely staining fibres, which are probably the vestige of the muscles running from the cypris antennules. There is great variability in the appearance of the males of I. idiotica. The forms may be short or long, the antennule may protrude to a greater or a less extent. RHIZOCEPHALA The Rhizocephala are parasites, in most cases, of decapod crustaceans. Usually the sexes are separate. The females have been repeatedly described. The adult form of the Rhizocephala (Fig. 6A) has lost all trace of segmentation and appendages. “Without the evidence of embryology it would be difficult to refer the adult even to the Crustacea” (Potts, 1915, p. 3). Delage (1884) gave a complete description of the development of Sacculina, but his results were doubted until Smith (1906) published a “striking confirmation of Delage’s story”. In 1935 Day described the life history of Sacculina. It was thought that each external sac represented a cypris larva which had fixed and metamorphosed in that identical position (Giard, 1887; Coutière, 1902). The typical rhizocephalan consists of an internal root-system and a sac-like externa. The externa is enclosed in a mantle, lined by cuticle. Through the mantle aperture the developing nauplii can escape from the mantle cavity. The mantle cavity surrounds a central visceral mass consisting of a nerve ganglion, the colleteric glands and the reproductive organs. Fertilization and embryonic development take place within the mantle cavity. Usually two kinds of eggs are produced. From small eggs develop small nauplii and cyprids. The cyprids settle on host crabs and a kentrogon is formed (e.g. in Sacculinidae, Lernaeodiscidae, and Peltogastridae). Within the kentrogon a hollow cuticular stylet forms through which the parasite material is injected into the host. There the interna develops and finally the virgin externa emerges. From

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

285

large eggs large nauplii and (male) cyprids develop. These were believed to be the so-called “complemental males”, although Smith (1906) thought that they did not take part in fertilization. Now it is known (Høeg, 1986) that a trichogon is formed (in the same species that have a kentrogon). This post-cypris male larvae is then found in the receptacle where spermatogenesis begins. The cycle of Chthamalophilus delagei differs profoundly from that of Sacculina. There is no kentrogon stage. The larva hatches as the cypris is morphologically degraded. It has one pair of antennae, it can swim and infects immediately. The parasite remains external. The endoparasitic phase of Sacculina is replaced by an internal phase homologous to the endoparasitic phase. In the course of development the nucleus differentiates. The parasite is in a crypt which is formed by the tegument of the host, closed by a kind of operculum, produced by the parasite. At the time of extrusion the operculum opens and the young sac passes through; at this stage the sac is not completely developed. The root system is replaced by a lobed globular mass which remains outside. The visceral sac is very simple compared with that of Sacculina. It has only the ovary as a differentiated organ. The ovary is opposite the peduncle and is fixed to the wall by a mesentery. The mesentery is penetrated by a canal which communicates with the inner hole of the brood cavity to the exterior. There is no organ which is similar to a testis. Chthamalophilus reproduces exclusively by self-fertilization. The male reproductive cells arise at the wall of the sac shortly after its extrusion, then they are liberated into the brood-chamber as a mass which ensures the fertilization of the eggs. Bocquet-Védrine & Parent (1972) described the organogenesis in Boschmaella balani, a parasite of Balanus improvisus. In this species the gonad develops out of cells of the brood pouch coat. It is not preformed in the nodules of stock-cells injected into the cypris. Bocquet-Védrine (1968, 1972) described the morphology and the development of Boschmaella balani. Whilst Chthalamophilus has no root system, Boschmaella develops a root system late in life, after the gonad has developed. Nothing was said about any male stage. A good review on the Rhizocephala has been given by Høeg & Lützen (1985). Kentrogonida are distinguished from Akentrogonida although some authors e.g. Høeg & Lützen (1985) are careful in using these terms. This is especially so since the absence of a kentrogon (Fig. 6B) has not been proved in the Akentrogonida. On the other hand, the presence of a kentrogon has not been proved in the Sylonidae. Høeg (pers. comm.) observed that host invasion in Clistosaccidae takes place without a kentrogon. Instead the female Clistosaccus cyprids penetrate through the host cuticle using one of the antennules. Through this the primordial parasite is injected. Thus in Clistosaccus it is proved experimentally that it is an ‘akentrogonid’, although the genus is so far classified with the Kentrogonida. A subdivision (as used at present) of the Rhizocephala into Kentrogonida and Akentrogonida may be premature. In any case the Peltogastridae, Lernaeodiscidae, and Sacculinidae seem closely related to each other and more so than to the other families Sylonidae and Clistosaccidae. All of them belong to the Kentrogonida. The Chthamalophilidae and other families may belong to the Akentrogonida. The genus Duplorbis is supposedly closely related to the Chthamalophilidae. Whilst hardly anything is known about the evolution and metamorphis of the Akentrogonida, a number of Kentrogonida has been investigated. With the exception of the Chthamalophilidae, which are believed to be the most primitive rhizocephalans and self-fertilizing hermaphrodites (Bocquet-Védrine, 1961) all rhizocephalans have separate sexes. This was not always common knowledge. Originally it was believed that the externa was a self-fertilizing hermaphrodite whose receptacles were regarded as the testes. In various species the “testes” were described in different positions. Boschma (1928) described several rhizocephalan species of the Gulf of Naples. This author even distinguishes Sacculina gonoplaxae from S. carcini by the different position of the “testes”. They are in the stalk, outside the visceral mass. The same is true for S. eriphiae. In Drepanorchis neglecta the “testes” are bent. They lie alongside the mesentery and

286

WALTRAUD KLEPAL

then in the central part of the visceral mass. The stalk is directly connected to it. In Parthenopea subterranea the “testes” are in the visceral mass. In Lernaeodiscus and in Parthenopea the “testes” are on either side of the median face. The “testes” of Lernaeodiscus galatheae are not wide and they have slightly bent vasa deferentia which lead into the mantle cavity. In Sacculina inflata the visceral mass is attached to a muscular portion of the body. The “testes” are found in this muscular region (Boschma, 1931). Thus the “testes” lie outside the true visceral mass. They are globular or ovoid and they pass abruptly into the vasa deferentia. The latter are comparatively short, straight or somewhat tortuous tubes. The vasa deferentia are believed to terminate in the posterior part of the mantle cavity at the male genital opening. The proximal part of each vas deferens is lined with a layer of chitin. Boschma (1931) was under the impression that the “testis” of the left side was impeded in growth by the enormous development of that of the right side. Thus the right “testis” was more strongly developed than the left one. According to Dillon & Zwerner (1966) the male organs of the sacculinid Loxothylacus panopaei are situated in the posterior portion of the visceral mass. The two “testes” are retort-shaped and posteriorly connected to a vas deferens. Both vasa deferentia (this term has to be used cautiously, see Høeg & Lützen, 1985) extend ventrally for a short distance and then they open into a brood chamber. A single nerve ganglion is located centrally and slightly anterior to the genital organs. In the same plane there is a mantle aperture and the peduncle. The “testes” of Lernaeodiscus crenatus are described as being strongly tortuous (Boschma, 1969). The vasa deferentia run in a straight course along the margins of the comparatively short ventral mesentery. The male genital openings of this species are described on either side of the anterior end of the ventral mesentery. It is now well known that the externa is female and produces two kinds of larvae: the male cypris and the female cypris. Delage (1884) assumed that the cypris settling around the mantle aperture of the juvenile externa were males. Ichikawa & Yanagimachi (1958, 1960) showed that the virgin externa is wholly female. The emerged externa can only grow and reproduce when a male cyprid settles on it and implants cells into the male receptacles (Ichikawa & Yanagimachi, 1958, 1960; Høeg, 1982; Lützen, 1984; Høeg & Ritchie, 1985). The cyprids of the two sexes may be distinguished by their size (Yanagimachi, 1961a; Høeg, 1984) and by differences in their anatomy as well as in their attachment organs (Walker, 1985). Two size classes were also described in L. galatheae (Veillet, 1943). Male cyprids settle only on recently emerged externae and female cyprids settle only on prospective hosts (Høeg, 1984, 1985b). In Sacculina carcini and Lernaeodiscus porcellanae the two size classes of cyprids occur singly or together in the same brood depending on the time of the year (Høeg, 1982, 1985b). In L. galatheae there are two size classes of nauplii corresponding to two different egg sizes. There are big nauplii, stage I, about 0·18 mm long and small ones of the same stage, about 0·13 mm long (Veillet, 1943). A size difference in eggs and nauplii was also noted in Peltogasterella gracilis (Yanagimachi, 1960) and in Triangulus galatheae. In Peltogasterella gracilis an individual externa produces a series of broods containing only one sex. The females all look alike but they are of two genetically different kinds. Male- and female-producing females may be distinguished. The male-producing female possesses 30A chromosomes and produces large eggs with 15A chromosomes. After fertilization these eggs develop into large nauplius larvae, which then further develop into large (male) cyprids. The female-producing female has 30A+X chromosomes and produces small eggs. Half of these eggs have 15A+X chromosomes and the other half have 15A chromosomes. These eggs develop into small nauplii after fertilization and then into small cyprids, half of them having 30A chromosomes and the other half having 30A+X chromosomes (Yanagimachi, 1960). Veillet (1943) assumed that sex determination may take place before fertilization of the ova. There is no distinct structural difference between large and small eggs; the nauplius larvae are also identical in their structure. The cyprids differ in size and in structure (Yanagimachi, 1961a; Walker, 1985). The latter author found differences in

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

287

Fig. 6. —Various stages of Rhizocephala. A, section through mesentery of externa of Clistosaccus paguri without mantle aperture (after Høeg, 1982); B, horizontal section through kentrogon of Lernaeodiscus porcellanae (after Høeg, 1985b); C, section through the mantle aperture of a virgin externa, a trichogon has entered the mantle cavity (after Høeg, in press); D, contours of trichogons of Sacculina carcini (a), of Gemmosaccus sulcatus (b), and of Peltogasterella socialis (c) (after Veillet, 1985); E, cross section through the basal part of a juvenile externa, in the top receptacle there is a trichogon (tr) which has shed its cuticle (cu) in the receptacle duct (rd), this cuticle prevents another trichogon from entering the receptacle; in the bottom receptacle there is no trichogon in the terminal canal (tc); cc, cement canal; cg, cement gland; cu, cuticle; ec, epithelium of cement canal; ep, epithelial cells; hc, host cuticle; km, kentrogon muscles; m, muscles; o, ovary; pm, posterior kentrogon muscles; rc, receptacle; rw, large cells of receptacle wall; scale bars, A=2 mm B=20 µm C=1 mm D (a) and (b)=50 µm D(c)=25 µm E=0·6 mm

288

WALTRAUD KLEPAL

the antennules of the male and female cyprids. He realized that the size difference of the cyprids alone was not a good enough criterion for distinguishing the sexes. The antennule of the female is shorter than that of the male. Whilst the male has a large sac on the attachment surface and a pore on the distal tip, in the female both structures are missing. The males have a large thin-walled sac (probably aesthetascs) on the third segment and a smaller one on the fourth segment. The females have only the smaller sac on the fourth segment, but none on the third segment. Sex determination simply by size is problematic in the cyprids of Sacculina carcini because the size of the cyprids varies with the seasons (Høeg, 1984). Veillet (1945, 1960) also found differences in larval size. A size difference was thus recorded in all three major rhizocephalan families: Peltogastridae, Lernaeodiscidae, and Sacculinidae. There is no evident morphological difference between the male-producing and the female-producing female. The nauplii of S. carcini metamorphose into cyprids between four and eight days after release and the cyprids may settle two to three days after this moult. In Clistosaccus paguri the larvae are liberated as cyprids, just as in Mycetomorpha, Thompsonia, Sylon, Cyphosaccus, with the exception of C. norvegicus (Høeg & Lützen, 1985) where the larvae are released as nauplii, and the freshwater genus Sesarmaxenos (Feuerborn, 1933; Lützen, 1981a). The release as cyprids ensures an increased survival success of the cypris as the infecting stage. On the other hand, it reduces the ability of the species to spread and to explore new areas. Clistosaccus paguri has only one or very few broods within the lifetime of an externa. Thus it seems that species with nauplius stages counterbalance the greater loss of larvae by producing several broods. Species with few broods secure a high survival rate by shortening the larval life. Also in species where the larvae hatch as cyprids (Sylon, Clistosaccus, Thompsonia, etc.) there are very small larvae which may be due to a large number of larvae in a single brood. The male cyprids of Clistosaccus paguri settle both on recently emerged externae and on the host cuticle above still internal parasites (Høeg, 1985a). They penetrate into the parasites with one antennule. The function of the cyprids is to implant precursors of spermatogonia into the female parasite. This settlement of the male cyprids above the primordia is unique among the Rhizocephala yet studied. The question, therefore, arises how the male cyprids locate the still internal female parasite. Furthermore, the penetration of the cypris antennule into a hard substratum is unknown in other cirripedes. The possibility of implantation into primordia lengthens the period during which the virgin female parasite Clistosaccus can receive male cells. This is an advantage if male cyprids are rare. An adult externa on a host can supply the male cyprids necessary for growth and reproduction of the more juvenile parasites on the same host. This could be an advantage of the gregariousness seen in C. paguri (about 50% of the hosts carry more than one parasite; the parasites are usually at different stages of development). It is unknown how male rhizocephalan cyprids in the plankton locate virgin externae and select the right site on them for settlement (Lewis, 1978). They attach only within the area of host cuticle directly above the primordium. Most probably the cypris must attach with one antennule and penetrate with the other. There is a possibility that the penetration is not just mechanical but there may be also chemical dissolution of host and parasite cuticle e.g. by the cement. It is not known whether the female cypris in C. paguri uses a similar mechanism to invade the host crab. It is not known either whether this is ancestral or advanced among the Rhizocephala. To elucidate this the metamorphosis of males and females in several rhizocephalan species has to be investigated. This question is of eminent importance to our understanding of the evolution of the group. The male cypris attaches itself on to a very young female near the opening of a mantle cavity of the external sac. Later one finds in the pseudo-testicles of the young female a small number of male cells. The question remains still to be answered of what happens between the fixation of the male cypris and the presence of the cells in the pseudo-testis (see also the penetration of C. paguri, p. 332).

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

289

In the Peltogastridae, Lernaeodiscidae, Sacculinidae, Sylonidae, and Clistosaccidae the recently emerged externa is virgin. In the first three families it develops a mantle aperture soon after its emergence (within 6 to 12 days depending on the time of the year) and simultaneously it begins to attract male cyprids. These then settle in or around the aperture. The cyprids extrude cells (which are now known to be a larva, the trichogon) into the mantle cavity. They migrate into the paired male receptacles. Whilst the receptacle duct is lined by cuticle which gets shed together with that of the mantle cavity (it is in direct continuation of the latter) there is no cuticle in the centre of the receptacle. Thus the trichogon cells and the surrounding female cells are not separated from each other. This distinguishes the rhizocephalan receptacle from the receptacula seminis of other crustaceans. All authors who worked on rhizocephalans recently (with the exception of Bocquet-Védrine) are of the opinion that the implanted cells are the externa’s only source of spermatozoa (Ichikawa & Yanagimachi, 1958, 1960; Veillet, 1960; Yanagimachi, 1961b; Ritchie & Høeg, 1981; Lützen, 1984; Høeg, 1985a; Høeg, & Ritchie, 1985). Only Bocquet-Védrine (1972) maintains that the European Sacculinidae and the Chthamalophilidae are hermaphrodites. In contrast to the other groups mentioned before the mantle aperture appears late in the development of the externa in Clistosaccus, Cyphosaccus and the Chthamalophilidae. In Thompsonia no mantle aperture develops at all. Thus in these species the male cells must enter the externa differently than in other rhizocephalans. According to Yanagimachi & Fujimaki (1967) the male cypris implants cells into the young externa directly through the general body cuticle and Høeg (1985a) showed this actually to be the case in Clistosaccus paguri. The same could be true in Sylon (Lützen, 1981a). In Sacculina carcini, Peltogasterella sulcata and Peltogasterella socialis a free species-specific larva was observed in the mantle cavity. In various genera there are differences in the shape. In Sacculina carcini the males attach on the young female parasites quickly beneath a chitinous collar which forms the closure of the mantle cavity (it is the remnant of the first moult of the young female). The metamorphosis is also very fast, it takes about 15 minutes. Within the mantle cavity there are trichogons which are metamorphosed cyprids. The larvae show pigment granules. They are covered by setae which are orientated backwards. The larva migrates through the mantle cavity and ends up in the opening of the receptacle duct with setae orientated backwards. Here the setae get lost and the pigmented marks may be found in the pseudo-testis. It is worth noting, that the setae of the male larva are found in the male cypris: in the anterior half of the larva the dorsal cells have setae which are orientated towards the front of the cypris. One understands why the normal metamorphosis is so fast: the trichogon cuticle with spines is preformed and this makes the rapid metamorphosis possible. In Peltogasterella sulcata the male cypris are very big and they attach on very young sacs. The male larva is found in the mantle cavity of the sac and it is easily seen by the orange pigment. The anterior part is transparent and it contains a dozen germinal cells with big nuclei. There is no trace of any setae. The larva goes through the receptacle duct. In the pseudo-testicle one finds the remnants of pigment cells, the pigment granules, orange protein droplets, and two categories of cells: the sex cells and cells with small nuclei. Durand & Veillet (1972) described the spermatogenesis of Gemmosaccus sulcatus (now Peltogasterella sulcata) and Sacculina carcini. The larvae of Peltogasterella socialis resemble those of P. sulcata. The colourless anterior region is elongated, then there are two “shoulders” and a rounded posterior end. The front end opens and the contents of the larva go into the receptacle duct; there are lipid droplets. Veillet (1985) believes that the shape of the larvae is species specific whilst Høeg (in press) states that they are amoeboid. Looking at Veillet’s drawings this could well be the case (Fig. 6D). Høeg (pers. comm.) observed many live larvae and found that the changes of shape occur mainly when escaping through the cypris antennule and penetrating into the receptacle duct. Within the mantle cavity the shape of the trichogon is fairly constant.

290

WALTRAUD KLEPAL

The virgin externa may stay attractive to males after as long as six months’ wait as in Lernaeodiscus porcellanae (Høeg & Ritchie, 1985). Growth of the externa stops when there is no male present. But even after six months growth may be resumed. From this it may be deduced that the emergence of the externa and the acquisition of the mantle aperture bear no obligatory relation to male cypris settlement in the Rhizocephala. Reinhard (1942) believed that the male cyprids of Peltogaster paguri fertilized part of the first brood of the externa. The mechanism that attracts male cyprids to the parasite is totally unknown. It could either be a moulting hormone or some substance released from the male cell receptacles (Høeg & Ritchie, 1985) or just the physico-chemical recognition of the externa surface. The fact that virgin externa (e.g. of Lernaeodiscus porcellanae) remain attractive to male larvae and that they retain the ability for normal development after long periods of isolation is advantageous in circumstances when the larvae are rare. The density of male cyprids is variable. When the parasite is establishing itself in a new area the infection percentage is low and the abundance is also low. In Peltogasterella gracilis and P. sulcata virgin externae that fail to receive males fall away from the host within one month after emergence. The lost externae are replaced by a subsequent generation of replacement externae (Ichikawa & Yanagimachi, 1958; Perez, 1931). Regeneration of externae is common (except perhaps in solitary species like Sacculina). In most species they are lost only when damaged as in Lernaeodiscus and in Peltogaster (Day, 1935; Hartnoll, 1967; Lützen, 1981b). In Sacculina, on the other hand, the externa is always cast off after a limited number of broods. Veillet (1943, 1945) knew of two different kinds of larvae (see p. 331) but was, like other workers, aware that there was not enough information on the male cypris larvae. In 1985 Veillet published a short note on the male larvae (Fig. 6D) and in 1986 Høeg investigated the male cypris metamorphosis (Fig. 6C, E). The “trichogon” was discovered, a new post-cypris male larval form which may be considered as an extremely reduced dwarf male. The free trichogon has an irregular amoeboid shape. In the direction of its movement in the receptacle duct of the female it produces a pseudopodium-like extension. Inside the mantle cavity the trichogon has a fairly constant vermiform shape. The body surface of the trichogon is covered by spiny epicuticle with spine fields whose spines point obligatorily backwards (hence the name: trichos=hair, gonus=larva) as in Sacculina carcini, Lernaeodiscus porcellanae and in Peltogaster paguri (Veillet, 1985; Høeg, in press; Høeg, pers. comm.). In Peltogasterella there are no spines on the trichogon (Høeg, pers. comm.). In the trichogon three to four cell-types may be distinguished: the dorsolateral, the ventral epidermis, the inclusion cells, and the postganglion cells. Most of the inclusion cells and postganglion cells are in the centre of the trichogon body. Several other tissues are lacking: these are muscles, nerve cells or unicellular epidermal gland cells. Most of the trichogon body consists of the dorsolateral epidermis of the cypris i.e. the cells have large round nuclei with a single nucleolus. The formation of the trichogon begins at a moult after settlement. Before the trichogon is released it lies in the anterior end of the cypris at the bases of the two antennules, where all tissues outside the trichogon have already disintegrated. The trichogon escapes through one of the cypris antennules into the mantle aperture of the virgin externa. Then the trichogon penetrates through the mantle cavity into the receptacle duct two hours after cypris settlement. On its migration it does not follow a special path through the mantle cavity. Two or more trichogons were often seen at the entrance to the same duct. When the trichogon has entered the duct fully, it is vermiform and has shed the spiny epicuticle. The trichogon arrives at the proximal blind end of the receptacle duct. Then it contracts into an ovoid shape. It is so large so that it disrupts the terminal canal and the trichogon cells come to lie directly appressed to large female cells of the receptacle wall. The border between the male and female cells is very distinct. Only a single trichogon enters the centre of each receptacle (Fig. 6E) and thus only the first arriving trichogon at any receptacle will take part in reproduction. The later arriving trichogons will be stopped by the cuticle shed in the receptacle duct by the first arriving

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

291

trichogon. This may be the reason why the cypris metamorphoses rapidly into the trichogon and then it migrates rapidly through the mantle cavity. After implantation the trichogon cells begin to rearrange themselves and five to six days later spermiogenesis begins. Probably only the postganglion cells are actual germ cells and if this is true the postganglion cells constitute a primordial testis in the cypris. After the moult from the cypris to the trichogon the new cuticle of the trichogon consists of an epicuticular layer only. The moult only involves the cephalic part of the body just as in kentrogon formation in the female cypris. When the trichogon is inside the receptacle it is probably nourished by the female cells. The trichogon shows no movements of its own, it is probably immobile since it contains neither muscles nor contractile filaments in the epidermis. It is much more likely that it is transported within the mantle cavity of the female by peristaltic contractions of the mantle musculature (Høeg, pers. comm.). The epicuticular spines could prevent the larva from being flushed backwards in the mantle cavity. The trichogon and the kentrogon are homologous instars—both are formed from the cypris by an incomplete moult. The life time of the kentrogon is a few days whilst the trichogon lives for a few hours. The structure of the kentrogon is more complex than that of the trichogon. The kentrogon pierces the host cuticle, the trichogon does not, at least in Sacculina carcini. In Clistosaccus paguri penetration occurs in males hosts. The externae need male cyprids to proceed with their development. The infection by Rhizocephala has an effect on the sexuality of the host (Giard, 1887; Frentz, 1960). During the endoparasitic phase of Sacculina the crabs are feminized in most of their sexual qualities. The cephalothorax is affected in its general shape and also in detail. On the ventral side one sees the female genital openings on the 6th thoracal segments. Often the claws are reduced in comparison with those of normal males. The abdomen becomes rounded and thus shows a female aspect. Abdominal segments three, four, five normally grown together in the male and articulated in the female, become articulated in the parasitized males. The pleopods also become female. There are even supplementary appendages typical of the female. These males lose their aggression. The sexual instinct becomes weaker and they are incapable of fertilizing. These modifications become more emphasized at each moult; the intensity of the changes increases so that those which were caused first are those which are most highly reduced. The spermatogenesis continues on infected crabs in contrast to what happens in Inachus or Eiriocheir. Ova-testes have never been found in male Carcinus maenas. Giard (1887) stated that often male crabs get feminized without their testes becoming damaged or negatively influenced. Smith (1910, 1911, 1913) believes that the origin of the sexual modification following the infection by a rhizocephalan is a disturbance of the lipid metabolism in the crab. Sacculina prevents moulting of the host as soon as the externa is extruded. During the endoparasitic phase there is a certain influence on the growth of the crab and on its moults. The intervals between the moults get longer (Smith, 1907) and there is a reduction in growth rate from 30 to 20% or 10% during the moults which precede the extrusion of the visceral sac. The parasite develops rapidly once it is an externa and produces a succession of ova. It was thought that the Y-gland was put out of function; Veillet (1955) states that the roots of Sacculina infect exclusively the Y-gland. Another hypothesis states that Sacculina withdraws the hormone itself or a chemical intermediate stage which is necessary for the formation of the hormone. Frentz (1960) states that there is no proof for this. The cutting off of the externa in Sacculina does not cause a return of full moulting capacity in the host (Cornubert, 1954). The eye stalks which are removed are not enough to cause a moulting of the parasitized animal. When the externa and the eye stalks are cut off the host regains moulting. Peltogaster does not prevent the moulting of Pagurus (Veillet, 1945). Perez (1931) stated that in Peltogasterella sulcata the normal cyclic evolution of the Rhizocephala and the extrusion of the visceral sacs are guaranteed by the periodical moulting of the host.

292

WALTRAUD KLEPAL

DISCUSSION PALAEONTOLOGICAL CONSIDERATIONS Whilst Acrothoracica have been known since the Carbon (the Apygophora have been known since Carboniferous times and the Pygophora have been known since Triassic times), the Lepadomorpha have been known since the Silur and the Balanomorpha probably since the Jura (Withers, 1928). In the scutal plates of Arcoscalpellum from the Upper Cretaceous a depression is found. This is an indication that dwarf or complemental males have occurred in Mesozoic times. There is no fossil record of any Rhizocephala. PHYLOGENETIC CONSIDERATIONS One of the major differences between the Acrothoracica, the Thoracica in general as well as in Ibla, and the Rhizocephala is the mode of life. The Acrothoracica are burrowing and/or parasitic cirripedes. They have parasitism in common with the Rhizocephala. The Thoracica are epizoic or growing on any hard substrata. The Iblidae grow preferentially in fissures (Achituv & Klepal, 1981) and are rarely found on ‘open’ substrata. The Acrothoracica are considered to have evolved from the Lepadomorpha (Newman, Zullo & Withers, 1969; Tomlinson, 1969). In their embryology the Acrothoracica and the pedunculates agree well so that a common origin of these two groups must be assumed. Each of the two groups contains animals which are too specialized to be the origin of the other. The Acrothoracica show an original condition but many of their morphological and biological peculiarities are adaptations to their endolithic life. To find out about the origin of the Acrothoracica comparative investigations of the embryology would be necessary. Turquier (1972) states that there is no interim form known between Acrothoracica and pedunculates. In spite of the similarity of some of the species with the Iblidae the Acrothoracica are a well-defined group. The Trypetesidae are the most highly specialized family amongst the recent forms. The Iblidae are amongst the most primitive Lepadomorpha. According to Zevina (1980) they are at the same taxonomic level as the Scalpellidae. Klepal (1985) proposed that Ibla may have diverted from the cirripedian stock before the Thoracica separated from the Acrothoracica. Ibla may even be the missing link between the two orders. This presumption is based on the fact that there are several characters in common between the Acrothoracica and the male and the female of I. cumingi. This is following Pilsbry’s idea (Pilsbry, 1908) of considering male characters as well as female characters for systematic purposes. The males of Ibla and Acrothoracica have the same sperm structure and grape-like cement glands in the basal part of the antennule. In the male of Ibla and some acrothoracicans the suboesophageal ganglion narrows towards the supraoesophageal ganglion (Berndt, 1903b). There is no oblique muscle layer in the peduncle of male Ibla nor in the mantle sac of some Acrothoracica, e.g. Berndtia purpurea (Utinomi, 1960). A difference between the Acrothoracica and Ibla is the attachment of the males: in the Acrothoracica they are fixed to the outside of the female, usually close to the attachment disk, while in Ibla they are inside the mantle cavity of the female (hermaphrodite). Tomlinson (1969) considers Ibla to represent a stem-line from which the Acrothoracica evolved. This does not agree with Klepal (1985) nor with Newman (1971). Newman states that the occurrence of a rostral plate in the female of Weltneria excludes Ibla as ancestral. This plate makes it likely that the Acrothoracica had a scalpellid ancestor. Although Lithotrya is not closely related to the ancestral stock it provides an example of how the transition may have come about. Foster (1978) placed Lithotrya at the stem of scalpellid evolution. Zevina, after having first placed the Lithotryinae as the first subfamily of the Scalpellidae (Zevina, 1978a), in the second part of her revision places the

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

293

Calanticinae rather than the Lithotryinae at the stem of her phylogenetic tree (Zevina, 1978b). In doing so, she follows Broch (1922). Since then Newman (1979) had described Neolepas zevinae. This author made a mental exercise in relating the various fossils known so far to stages of development in Pollicipes. Thus he distinguishes a “Cyprilepas stage” from a “Praelepas stage”, then the “Eolepas stage” followed by the “Calantica stage”. The Neolepas stage is also seen following the Eolepas stage in the ontogeny of some species of Calantica described by Foster (1978). This ontogenetic sequence could be interpreted as progressive evolution or regressive evolution in the form of paedo-morphosis. The Neolepas level of organization could be achieved through progenesis. But there is no evidence for Neolepas being progenetic. Thus Newman (1979) came to the conclusion that Lithotrya has descended from a pre-Calantica, Neolepas-like form. Many of the dwarf males of Calantica progress no further than the Eolepas or Neolepas ontogenetic stage. Their capitular plates being apart and their peduncles being scaleless suggests that progenesis is responsible for their reduced form. Within the Calanticinae the genera Calantica and Scillaelepas may be distinguished (Zevina, 1978a, b). Foster (1978) does not agree but Newman (1980) follows up Zevina’s idea. Newman points out the differences between Calantica and Scillaelepas, one of them being that the complemental males occur in the subrostral region in Scillaelepas rather than between the scuta as in Calantica. Scillaelepas fosteri is unique in the complemental male being found attached to the peduncle between the peduncular scales, rather than on the capitulum of the hermaphrodite. In the subgenus Aurivillialepas of Scillaelepas the complemental males are in the cavity between the subrostrum and the rostrum. Newman (1980) discusses the rôle of the male in the transformation of peduncular scales into subrostral plates. The subrostra were originally peduncular scales as in Scillaelepas s. str., incorporated into the capitulum. Did the males take advantage of the presence of subrostra or were the subrostra induced by males in settling in the subrostral region? Newman (1980) shows that the subrostra were induced to form by the presence of complemental males. The male cypris must attach itself in a position to ensure the male the possibility of fertilizing the hermaphrodite. It would be an advantage for the species if the hermaphrodite provided some indication to the cypris as to where to settle. A “subrostral pheromone” must have become involved. The Acrothoracica show clear signs of reduction. They are variable in the number of segments, the number of cirri and the number of podomeres in the cirri. Highly reduced animals have no gut, no extremities and no suboesophageal ganglion. The fusion of the thoracic ganglia in the female and the incomplete development of the male may be due to parasitism. The attachment organ and the reproductive organs are the only ones that are always there in a male. Every other organ system is very variable and may be even lost. The males are very small, usually smaller than the cypris larva. They have no digestive tract, no mouthparts and no cirri. Typically the males have a single testis, a simple vesicula seminalis, a nauplius eye and paired antennae. In many males there is a “yellow organ”. Most species have males with an extensible penis. On the integument there may be teeth on the end opposite the first antennae or rectangular or polygonal plates often bearing pegs or other rugose elevations. There may be hairs, generally around the end opposite the first antennae. Tomlinson (1969) suggests that a close relationship between the Acrothoracica and the Rhizocephala is to be considered since the latter Order shows all these reductions (seen in the Acrothoracica) to an even greater degree. Tomlinson (1969) proposes that only three families be retained within the Acrothoracica. The Lithoglyptidae could incorporate the old families Berndtiidae, Balanodytidae, Kochlorinidae, Utinomiidae, and Chytraeidae. It is obvious that the Trypetesidae are more specialized than the Lithoglyptidae and the Cryptophialidae. The Lithoglyptidae is the most primitive family of the Acrothoracica. The genus Weltneria most resembles the Thoracica. It has biramous cirri, a mandibular palp, an anus, two ventral ganglia and six

294

WALTRAUD KLEPAL

pairs of cirri as in the thoracicians, and caudal appendages. The Cryptophialidae have the mouth cirri reduced to vestiges and the reduced number of cirri. They have an unusual gizzard. The Trypetesidae have uniramous cirri, no anus and only one ventral ganglion. The male of Trypetesa lateralis lacks a penis. Of the three families in the Acrothoracica, the Lithoglyptidae are the most primitive ones. The Cryptophialidae and the Trypetesidae are two divergent and more specialized families. In the female of Trypetesa the adductor of the valves of the cypris persists and it remains in the posterior and ventral position relative to the oesophagus (Turquier, 1970a, b). The same is the case in Ibla; in all other thoracicans this is an exception. In the males of the scalpelliform cirripedes the Musculus adductor scutorum may exist or it may be lost as in Scalpellum gigas which has only four rudimentary plates. In Trypetesa as well as in Ibla this muscle becomes the Musculus adductor scutorum of the adult. In all other cases the adductor of the scuta is a new formation (Walley, 1969). In the Ascothoracica the adductor muscle is in the same position as in Ibla. This may be an indication for the ascothoracid ancestry of the Cirripedia s. str. As in Ibla there is a distance between the first cirrus and the following ones in the Acrothoracica especially in Trypetesa. In the Acrothoracica in general the cirri are reduced in all three families. In Weltneria (Lithoglyptidae) (=Utinomia) the most original condition is found. There are six pairs of cirri and there are well-developed caudal appendages. Within the Lithoglyptidae there are all interim stages of the regression of cirri. The final stage of this regression is represented in the genera Kochlorine and Kochlorinopsis. In the males of the Scalpellidae there are also all stages in the regression of cirri. They get reduced in number and in length and some of the males have no cirri at all, e.g. Scalpellum alcockianum, S. condensum, S. convexum, Pisiscalpellum withersi. There may or may not be regression of the caudal appendages. The genus Weltneria is reminiscent of the pedunculates (especially of the Oxynaspidae) in its mouthparts; the mandibles have four teeth and bilobed maxilla. In the family Scalpellidae there are a number of interim stages between dwarf males which are hardly distinguishable from the hermaphrodites or females, or highly degenerate males which belong to Scalpellum s. str. (e.g. S. stroemi). In species with highly degenerate males the anatomical simplification can reach a degree similar to that of the males of the Acrothoracica. Dwarfism in the males of cirripedes is a common phenomenon. Contrary to the opinion of some earlier workers (Gruvel, 1902a, 1905) it does not seem to be correlated with trophical difficulties. The Acrothoracica live on the reserves which are in the eggs from which the males develop. The males of Ibla belong to the little reduced “peduncle-type”. Batham (1946) showed that the males of I. idiotica come into existence from the cypris after a single moult and their appearance changes considerably between the juvenile and the adult phase. Changes between the juvenile and the adult stage are also known in Acrothoracica. Here the adult male may be smaller and show a simpler organization than the juvenile. In the non-parasitic cirripedes the most deformed males are not larval forms which could be called neotenic; they are highly specialized and in them there was a condensation of morphogenetic phenomena. In all cases there is at least one moult in the males. The males of Calantica, Smilium and most of the Iblidae have mouthparts and a digestive tract which are both functional. Apart from the scalpelliform genera Scalpellum, Calantica, and Smilium, males have also been found in Pollicipes spinosus. For that reason Foster (1978) strongly suggests that Pollicipes spinosus with a complemental male should be regarded as a primitive calanticid. Up to now complemental males have not been found in any other species of Pollicipes. So far complemental or apertural males have been found in the primitive genera of the Balanomorpha within the Balanomorphoidea, Bathylasmatinae, and Chelonibiinae and within the Balanoidea (systematics after Newman & Ross, 1976). It is generally accepted that lepadomorph barnacles are more primitive than the balanomorphs (Foster, 1978). The primitive scalpellids were multi-plated and later underwent shell-plate reduction. This results in

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

295

the range of scalpellids known today. Pollicipes is a relatively unmodified descendant of early multi-plated forms. Originally the Scalpellidae are derived from non-shelled forms. The multi-plated ancestral scalpellids possibly gave rise to most of the present-day scalpellids. The calanticoid stock is probably an early derivative. In the scalpellid evolution the free-living larval stages are suppressed. This permits the colonization of greater depths. Some of the cyprids are retained in close association with the hermaphrodite. Here their sexual development into males is encouraged and chances of cross-fertilization are improved. The tendency to develop complemental males in the Scalpellidae, Iblidae, and some Balanomorpha may be interpreted as a trend away from the hermaphrodite condition and towards the unisexual dioecious condition. Thoracica are either hermaphroditic originally or they became so early in their evolution (Pilsbry, 1908; Broch, 1922; Newman, Zullo & Withers, 1969; Foster, 1978). Because of their protandric hermaphroditism they are pre-adapted to the production of complemental males (Newman, 1980). The acceleration of sexual maturity in the males (they are very small, have no peduncular and only incomplete capitular armament) indicates that the males are progenetic. A whole sequence of this development may be seen in Calantica and Scillaelepas (Newman, 1980). First the subrostral peduncular scales of the hermaphrodite get slightly enlarged as in Scillaelepas s. str., then there are specially formed scales that get incorporated into the capitulum as in Gruvelialepas (although in the species of this genus no complemental males have so far been found, then the capitular subrostrum is developed as in Aurivillialepas and finally the male is transferred from the subrostral to the suprarostral position between the scutal plates as in Calantica. This development proceeds with the male getting more and more degenerate until it is just a bag of sperms that is attached in special pockets in the scuta, as in Arcoscalpellum. Thus the hermaphrodite supports the progenetic process by providing the protective devices. The possession of complemental males seems to be more advanced than their absence (Foster, 1978). This is in contrast to Broch’s (1922) interpretation who thinks that the males are about to disappear in the genus because the males do not occur regularly. The regularity of the occurrence of the males seems to be dependent on the habitat of the hermaphrodite or female. Thus in the calanticids small complemental males seem purely facultative (Foster, 1978). Sometimes they are lacking and sometimes there are a great number of them. In Arcoscalpellida of the deep sea separate males are obligatory for sexual reproduction. From the restricted intertidal habitat of Pollicipes and Calantica spinosa Foster (1978) suggests that there may be an ancient intertidal origin of the scalpellids. C. villosa occurs in habitats protected from desiccation. This may be an indication for a shallow-water origin of Scalpellidae. Broch (1922) believes that the hermaphroditic condition is basic in the Thoracica. It seems that the males develop from normal cyprids, which are arrested in growth, except for the male reproductive system which is well developed, by some factor of the interscutal habitat. EFFECTS OF DWARFING Dwarfing in the cirripedes, so far as it is known, affects the size of the individuals; the male may be only one tenth the size of the female or hermaphrodite as in Ibla sp. or it may be even smaller than the cypris as in Kochlorine bocqueti. It causes the loss of some structures and organ systems. Thus, in the Scalpellidae a reduction of the plates may be followed. Whilst males with six plates look like small hermaphrodites (they have capitulum and peduncle well separated, they have usually six pairs of cirri, a functioning alimentary canal, a well-developed penis and often caudal appendages), individuals with only four plates also show other signs of reduction. In most cases there is no distinction between capitulum and peduncle, the cirri may be reduced in length and in number and the caudal appendages may be reduced to some setae (as in Scalpellum ornatum). There is no description of the presence or lack of a penis. The plates themselves are in

296

WALTRAUD KLEPAL

most cases reduced, often two of the four are even smaller than the other two. S. rostratum is an exception. It has only three large plates developed but capitulum and peduncle are distinct, there are six pairs of cirri, a well-developed penis and caudal appendages, and the digestive tract is functional. In most of the other species whose males have only four plates nothing was said about the condition of the alimentary canal. In the cirri a tendency may be followed from males with six plates to those with four or three towards a larger distance between the first and the following cirri and thus a formation of maxillipede. S. chitinosum has slightly swollen marginal parts which could be plates. No details are given on the other organ systems. There are also a number of sac-like males without any plates. Examples for these are S. regium, S. distinction, S. convexum, S. compaction, S. condensum, S. gibberum, S. projectum, S. chitinosum, S. javanicum, S. sessile, S. crinitum, S. hexagonum, and Pisiscalpellum withersi. In some of these species there are even six pairs of short cirri, as in Scalpellum regium, in others there are only three pairs of short cirri, as in S. gibberum. In most cases, there are no data on the presence or state of the cirri. In other cases, the cirri were counted as in S. vulgare but with differing results. This was probably due to the variability of the cirri (taking the care exercised by Darwin and Gruvel for granted). In Ibla cumingi the way of reduction was observed in the cirri. First their number is reduced, then the rami and podomeres followed by the muscles and the lacunar system, then the mitochondria and lastly the nerves and sensory organs. The reproductive system is always well developed, although there may not always be a penis as e.g. Scalpellum compressum or S. gibberum. This is deduced from the fact that nothing was said about a penis in these species. Furthermore, in S. longirostrum the penis is very short and probosciform. This also supports the idea of a reduction of the penis. In the Iblidae a penis is never developed. In the Acrothoracica, on the other hand, the penis is often extremely long so that it is curled up inside the body as in Berndtia purpurea. In Trypetesa lateralis the penis is missing. An extreme reduction of a male is seen in the Rhizocephala where the male may consist of a “bag of sperms” and perhaps a ganglion. In the males of the Thoracica and the Acrothoracica the antennae are always well developed with the exception of Klochlorinopsis discoporellae, where only a single antenna was described. In some Acrothoracica there may be a stalk between the antenna and the body as in Lithoglyptes spinatus or in Kochlorine hamata. In the males of some Scalpellidae the number of limbs in the antennae may be reduced as in Scalpellum compressum. The statement should be treated with caution as it is just possible that the author could not see the antennae well enough. In the Rhizocephala the antenna shows some variability. In the large cyprids the penultimate segment of the antenna is long and slender, in the small cyprids the corresponding part is diminished or absent (e.g. in Peltogasterella). In this Order the antennae may even have two different functions. Lewis (1978) pointed out that the cypris attaches with one antenna and it penetrates with the other. Whenever the alimentary canal is functional the mouthparts are well developed and similar to those in the female or hermaphrodite. When the gut is closed the mouthparts are reduced to a simple ridge of cells, as in Scalpellum striatum, or are lacking. Fusion may also be seen as a consequence of dwarfing. Thus ganglia or skeletal elements may fuse as in the male of Ibla. In this species fusion, of e.g. apodemes, is regarded as reduction caused by the lack of function. Apart from this fusion of elements of the same quality, Ibla provides an example for the fusion of elements of different quality. The muscles surrounding the excretory system in the female fuse in the male with the excretory system. At the same time these muscles change their function. From contracting and stretching the prosoma in the female they now act directly on the efferent duct of the excretory system of the male.

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

297

A range in the degree of the reduction may be observed in all cirripedes considered. Within the Acrothoracica the genera Weltneria and Lithoglyptes seem to be the most primitive. The males are pearshaped and homologous to the thorax of the cypris. The body of L. scamborachis is finely annulated. This could well be a remnant of metameres. In L. wilsoni, on the other hand, the male is described as a reduced bag of gametes. Thus, even within one genus there is a range in the degree of reduction. In Berndtia purpurea a number of longitudinal muscles was found, that varies with the size of the body. In contrast to the males of Trypetesa and Cryptophialus there are no circular muscles in the penis of Berndtia purpurea. This shows again the great variation in the equipment of the various organs of the male independent of the systematic position of the female. There may be two trends observed in the evolution of the male: one towards a differentiated form of attachment and the other towards a more effective way of reproduction with less material expenditure. Whilst in the more primitive species there is an ordinary antenna just like the cypris antenna, in others like Lithoglyptes spinatus and Kochlorine hamata there is a long stalk between the antenna and the body (see pp. 288– 9) (one significance of this may be to raise the male higher up towards the female) and in K. floridana or in Trypetesa spinulosa there may be a cylindrical projection of an orchid lobe. Concerning the reproductive organs the trend goes from a long annulated penis with a number of muscles at its base as in Berndtia purpurea towards no penis at all as in Trypetesa lateralis. The penis sheath is usually formed by the posterior part of the body or by the mantle cavity e.g. in Weltneria zibrowii. Either muscles or strands of connective tissue keep the penis in its position. One hypothesis put forward by Turquier (1977) is that the fertilization of the eggs of Kochlorine bocqueti could take place through the body-integument. The outermost end of the attachment peduncle and possibly also the organ for insemination is only at a distance of a few µm away from the egg sacs. If the hypothesis of the insemination through the integument were true then this could mean a special position of the species K. bocqueti within the Lithoglyptidae and the whole Order. Bisexuality is often considered as the more primitive condition from which unisexual hermaphroditism has evolved. In a population in which the males have a much shorter life expectancy and environmental requirements than females and when the males outnumber females both the genetic and reproductive efficiencies can be maximized (Scudo, 1969). The evolution of sexual dimorphism seems to be antiadaptive to the individual (like the reduction in body size or life expectancy in the males). This could also help to explain the evolution of dimorphism which is usually attributed to the action of sexual selection alone (Darwin, 1866). Male and female phenotypes cannot evolve independently (Lande, 1980). The effects of genes in the males are most probably correlated selective responses between the sexes. A comparison of closely related species often shows a trend of increasing sexual dimorphism with increasing body size, whether males or females tend to be larger due to greater ecological specialization of the sexes or stronger sexual selection in the larger species. Several authors (e.g. Clutton-Brock & Harvey, 1977; Ralls, 1977) noted that there remains considerable variance in sexual dimorphism among related species after accounting for that associated with the intensity of sexual selection. STAGES OF DEVELOPMENT OF THE MALE In all cases the males considered are at the metamorphosed level rather than at the cypris level. A detailed description of the stage of the male of Ibla cumingi was given by Klepal (1985). Judging from the skeletal elements, such as the number of endosternites and the kind and arrangement of the cuticular structures, it may be assumed that the males of Ibla sp. may have deviated in their development to the fully-developed form early after the cypris stage. This is most probably true for the males of the Acrothoracica and the

298

WALTRAUD KLEPAL

Rhizocephala, as well as for those of the balanomorph cirripedes. In the scalpelliform cirripedes the deviation need not have started very early judging from the fact that the males of some species look like miniature hermaphrodites, except that they do not have any female gonad. The males may be partial neotenic (De Beer, 1958), in showing the original condition in some of their organs e.g. the shape of the suboesophageal ganglion and the grape-like glands in Ibla cumingi. Because of the often high degree of reduction it is not always easy to detect the original characters and to separate them from reduced ones. DEVELOPMENT OF SEXUALITY AND HYPOTHESES ON THE ORIGIN OF MALES The question of whether gonochorism or hermaphroditism is original has been discussed repeatedly. Utinomi (1958) and Newman, Zullo & Withers (1969) consider that the sexes were originally separate and that hermaphroditism is a secondary phenomenon adapted to the gregarious fixed mode of life. This assumption is supported by the fact that separate sexes are found in all cirripede orders and in the Maxillopoda in general. In spite of that several authors are convinced that hermaphroditism is the original condition in cirripedes (e.g. Darwin, 1851, 1854, 1873; Hoek, 1883; Laloy, 1903; Ghiselin, 1974; Moyse, 1983). Moyse (1983) states that the plesiomorphic condition of both Ascothoracia and Cirripedia s. str. is hermaphroditism with deviation in specialized families. Moyse follows Newman, Zullo & Withers (1969) in postulating the origin of barnacles from an ascothoracic ancestor. The hypothesis of hermaphroditism being the original condition is supported by the study of chromosomes. So far no sex chromosomes have been found in any of the species investigated. Callan (1941) therefore suggests that larvae (at least in the Lepadomorpha) are potential hermaphrodites and those which settle on the appropriate place of the adult hermaphrodite develop as functional males only and that they go no further. Callan’s assumption is supported by the finding of apertural males (Crisp, 1983). These individuals are really hermaphrodites which are arrested in the protandric stage of their development. They have the potential to become hermaphrodites. Callan’s hypothesis gains further support from the work of Svane (1986) on Scalpellum scalpellum. In this species all cyprids are potential hermaphrodites (see also Svennevig, 1978). The presence of adult hermaphrodites influences sex determination. This agrees with the results of Gomez (1975) who concluded from his study on Balanus galeatus that sex is not environmentally controlled. The presence of adult hermaphrodites significantly facilitates metamorphosis to dwarf males, possibly by a chemical stimulus. As could be shown experimentally the percentage of cyprids metamorphosing into dwarf males depend on the density of available hermaphrodites. Thus Svane (1986) came to the conclusion that sex determination in Scalpellum scalpellum is controlled both genetically and environmentally. He suggests that only about 50% of the cypris larvae have the ability of metamorphosing to either dwarf males or to hermaphrodites. A similar mechanism seems to be involved in sex determination in the Rhizocephala (Høeg, pers. comm.). Two hypothetical ways for the evolution of males in cirripedes are possible. One begins with sessile protandric hermaphrodites, as in Chelonobia patula. The development of the female gonad is retarded so that pure males develop besides hermaphrodites. This condition is also represented in Ibla quadrivalvis and in all scalpelliform cirripedes with complemental males. Usually the male reproductive organs of the hermaphrodites with complemental males are reduced in comparison with those of a typical hermaphrodite. A further reduction of the male component in the hermaphrodite could then result in a pure female. Thus the gonochoristic condition could have developed in cirripedes. The other hypothetical way for the development of males could have begun with the gonochoristic condition. Males and females of equal size must be assumed (see also Foster, 1983). Once the males settled

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

299

on the females their size could be reduced. Fertilization became easier when the males moved inside the plates of the female or into their mantle cavity. The step coincides with major reductions in the male: a considerable reduction in size, loss of any plates and/or armament, loss of cirri, mouthparts and alimentary canal. In any case the reproductive organs must be well developed so that the male is sometimes just a “bag of sperms”. Examples may be seen for each step of the development discussed. After the reduction of the digestive tract a root system could be developed in the male (this kind of sequence is known from the Ascothoracica, Vagin, 1937; Vaghin, 1946). A next step could be the development of cryptogonochorism (Ghiselin, 1969) as is the case in Peltogasterella (Ichikawa & Yanagimachi, 1958, 1960). Then the function of the implanted male gonad is controlled by the female (Bresciani & Lützen, 1972). So far none of the hypotheses can be ruled out although at the present state of our knowledge the one beginning with the hermaphrodite seems to be the more likely. Within the animal kingdom hermaphroditism was probably the original form of sexuality (Schaller, 1974). Most animals are bisexual in their “Anlage”. Although free-living Ascothoracica and Acrothoracica are gonochoristic the sessile cirripedes could be originally hermaphroditic. The advantage of hermaphroditism in a sessile animal is obvious. Especially when the animals are capable of self-fertilization they can pioneer new places and they can also reproduce at the limits of species or racial distribution. Crisp (1983) argues that the presence of apertural males is advantageous to the hermaphrodite. The coevolution of apertural males and hermaphrodites represents a more primitive stage in evolution but is analogous to that put forward by Newman (1980) for the co-evolution of males and hermaphrodites with subrostral plates in Scillaelepas. Many thoracicans have a tendency to protandry. In order to increase the fitness of its own female function it might be to the advantage of the larger partner to suppress the development of the smaller protandric males. There will be competition between males and between males and hermaphrodites for the insemination of eggs. The greater the number and virility of complemental and apertural males the lower will be the fitness of male function in the hermaphrodite. Eventually this trend could lead to the dioecious condition, as evident in Ibla cumingi. Tomlinson (1966) treats self-fertilizing hermaphroditism the same as parthenogenesis. (Parthenogenesis is expected to be the next step in the evolution of sexuality in cirripedes.) In gonochoristic species the probability of contacting another individual is a function of population density. Mobility is important to gonochoristic animals, whilst hermaphroditic and parthenogenetic individuals can afford to be sessile or sluggish. Looking at the distribution of species we see that parthenogenetic forms have a much wider geographical range than the gonochoristic ones. Thus, it is evident that self-fertilizing hermaphroditism as well as parthenogenesis is extremely useful where the population densities are low i.e. at the limit of the species or racial distribution. Tomlinson (1966) argues that the production of two types of gametes in the hermaphrodites must be more troublesome than the production of a single type of gamete as in the gonochoristic species. This may explain the predominance of gonochorism in animals. The probability of parthenogenesis arising in any particular species may depend on the frequency of lethal genes, which may further depend on the relative degree of inbreeding (Cuellar, 1977). The more inbred the population, the greater the elimination of homozygous lethals and the greater the probability of parthenogenetic development. For the possible development the mode of transportation may also be important. It does not seem likely that parthenogenesis could evolve in highly vagile organisms. Asexual reproduction, as in parthenogenesis, has the advantage of a much faster population growth rate because there are no males. Furthermore, there is no critical reproductive density. This means that a single individual can colonize a new habitat, a population can persist at very low densities (Tomlinson, 1966; Ghiselin, 1974), survivors of catastrophes can readily recolonize the habitat and there is no depression of

300

WALTRAUD KLEPAL

growth rate caused by low frequency of sexual encounters. The growth rate of an asexual population is, therefore, independent of low density. Small groups of individuals colonizing a new habitat start at low densities (Gerritsen, 1980). In sparse populations the probability that mates will meet is reduced. The minimum population density for sexual reproduction, the critical density, can be a strong selective force on a species. Thus, adaptations will be favoured that increase the probability of mating and that keep the population density above the critical level. The critical density refers to the density of sexually active adults, rather than to the entire population. Hermaphroditic individuals will allocate the same amount of resources to male and female functions. The difference between bisexual and hermaphroditic species is that half the encounters in a bisexual population are between members of the same sex, while all encounters in a hermaphroditic population can result in mating. The critical density of a bisexual species is thus twice that of a similar hermaphroditic species (Gerritsen, 1980). Hermaphroditism is widespread in organisms that are sessile or sedentary and may be viewed as an adaptation to low encounter probability while retaining genetic recombination (Tomlinson, 1966; Ghiselin, 1969). An obvious advantage of hermaphroditism is that it reduces the critical density by a factor of two. In conclusion one can say that hermaphroditic reproduction requires only half the critical density of bisexual reproduction. Advantages of sexual reproduction are: a greater evolutionary rate due to recombination, a greater probability that offspring will survive intraspecific competition and unpredictable conditions induced by the evolution of their sexually reproducing competitors, predators, parasites, and prey (Maynard-Smith, 1978). ACKNOWLEDGEMENTS Thanks are due to Dr M.Barnes and Dr B.A.Foster for helping with the literature. Dr Høeg’s comments on the Rhizocephala and his help in allowing me to use some of his unpublished data are highly appreciated. Part of this work was supported by the Projekt 4323 des Fonds zur Förderung der wissenschaftlichen Forschung in Österreich. REFERENCES Achituv, Y. & Klepal, W., 1981. P.S.Z.N.I. Mar. Ecol., 2, 295–305. Anderson, D.T., 1965. Aust. J. Zool., 13, 1–15. Annandale, N., 1906. Ann. Mag. Nat. Hist. ser. 7, 18, 44–47. Annandale, N., 1910. Rec. Indian Mus. 5, 145–155. Annandale, N., 1911. Rec. Indian Mus., 6, 229–230. Annandale, N., 1916. Mem. Indian Mus., 6, 127–131. Aurivillius, C.W.S., 1894. K. svenska Vetensk Akad. Handl., 26(7), 107 pp. Baccetti, B., Dallai, R. & Rosati, F., 1968. J. Microsc. (Paris), 7(4), 20a–21a. Bage, F., 1908. Proc. R. Soc. Vict., n.s., 21, 226–232. Barnes, H., 1962. J. exp. Biol., 39, 353–358. Barnes, H., Klepal, W. & Munn, E.A., 1971. J. exp. mar. Biol. Ecol., 7, 173–196. Batham, E.J., 1945. Trans. R. Soc. N.Z., 75, 347–356. Batham, E.J., 1946. Trans. R. Soc. N.Z., 75, 405–418. Batham, E.J. & Tomlinson, J.T., 1965. Trans. R. Soc. N.Z., Zool., 7(9), 141–154. Berndt, W., 1903a. Sber. Ges. naturf. Freunde Berl., 436–444. Berndt, W., 1903b. Z. wiss. Zool., 74, 396–457. Berndt, W., 1906. Arch. Biontol., 1, 165–210.

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

301

Berndt, W., 1907. Arch. Naturgesch., 73(1), 287–289. Bocquet-Védrine, J., 1961. Cah. Biol. mar., 2, 455–593. Bocquet-Védrine, J., 1968. Archs Zool. exp. gén., 109, 257–267. Bocquet-Védrine, J., 1971. Archs Zool. exp. gén., 112, 761–770. Bocquet-Védrine, J., 1972. Cah. Biol. mar., 13, 615–626. Bocquet-Védrine, J. & Parent, J., 1972. Arch. Zool. exp. gén., 113, 109–128. Bocquet-Védrine, J. & Pochon-Masson, J., 1969. Arch. Zool. exp. gén., 110, 595– 616. Boschma, H., 1928. Proc. U.S. natn. Mus., 73, 1–10. Boschma, H., 1931. Vidensk. Meddr dansk naturh. Foren., 89, 297–380. Boschma, H., 1969. Proc. K. ned. Akad. Wet. C, 72, 413–419. Brandt, E., 1871. Bull. Acad. imp. Sci. St. Petersburg, 15, 332–340. Bresciani, J. & Lützen, J., 1972. Vidensk. Meddr dansk naturh. Foren., 135, 7–20. Broch, H., 1922. Vidensk. Meddr dansk naturh. Foren.,73, 215–358. Broch, H., 1927 In Handuch der Zoologie, edited by W.Kükenthal, De Gryter Berlin-Leipzig, 3(1), 503–552. Bullock, T.H. & Horridge, G.A., 1965. Structure and Function in the Nervous Systems of Invertebrates. Freeman & Co., San Fransisco, Vols 1, 2, xx+1719 pp. Callan, M.G. 1941. Nature Lond., 148, 258 only. Cannon, H.G., 1947. Phill. Trans. R. Soc. Ser. B., 233, 89–136. Clutton-Brock, T. & Harvey, P., 1977. J. Zool., 183, 1–39. Cornubert, G., 1954. Bull. Inst. océanogr. Monaco, No. 1039, 4 pp. Coutière, 1902. C. r. hebd. Séanc. Acad. Sci., Paris, 134, 913 only Crips, D.J. 1983. Mar. Biol. Lett., 4, 281–294. Cueller, O., 1977. Science, 197, 837–843. Darwin, C., 1851. A Monograph of the Subclass Cirripedia. Vol. 1. The Lepadidae. Ray Soc., London, 400 pp. Darwin, C., 1851. A Monograph of the Subclass Cirripedia. Vol. 2. The Lepadidae. Ray Soc., London, 684 pp. Darwin, C., 1866. The Descent of Man and Selection with Respect to Sex. Appleton & Co., New York, 2nd edition (cited in Scudo, 1969). Darwin, C., 1873. Nature, Lond., 8, 431–432. Day, J.H., 1935. Q. Jl. microsc. Sci. N. S., 77, 549–583. Dayton, P.K., Newman, W. A. & Oliver, J., 1982 J. Biogeogr., 9, 95–109. De Beer, Sir Gavin, 1958. Embryos and Ancestors. Oxford University Press, Oxford, 3rd edition, 197 pp. Delage, Y., 1884. Arcahs Zool. exp. gén., sér. 2, 2, 417–738. Dillon, W.A. & Zwerner, D.E., 1966. Trans. Am. microsc. Soc., 85, 407–414. Durand, D. & Veillet, A., 1972. Bull. Acad. Soc. Lorraines Sci., 11, 119–131. Feuerborn, H.T., 1983. Zool. Anz., 35, 127–138. Foster, B.A., 1978. N. Z. Ocean. Inst. Mem., No. 69, 160 pp. Foster, B.A., 1980. N. Z. Jl. Zool., 7, 523–531. Foster, B.A. 1983. Aust. Mus. Mem., 18, 133–139. Franzén, A., 1970. In, Comparative Spermatology, edited by B.Baccetti, Proc. Int. Symp. Rome and Siena, July 1969, pp. 29–46. Frentz, R., 1960. Contribution à l'étude biochimique du milieu interieur de Carcinus maenas Linné, Edit. Soc. Impress. Typ., Namcy, 176 pp. Gaonkar, S.N. & Karande, A.A., 1980. J. Bombay nat. Hist. Soc., 76, 305–310. Gerritsen, J., 1980. Am. Nat., 115, 718–742. Ghiselin, M.T.,1969. Q. Rev. Biol., 44, 189–208. Ghiselin, M.T.,1974. The Economy of Nature and the Evolution of Sex. University California Press, Berleley, 346 pp. Giard, A., 1887. Bull. scient. Fr. Belg., 18, 1–28. Gomez, E.D., 1975. Crustaceana, 28, 105–107. Gould, S.J., 1978. Ontogeny and Phylogeny. Harvard University Press, Cambridge, Massachusetts, 2nd edition, 501 pp.

302

WALTRAUD KLEPAL

Gruvel, A., 1898. P.v. Séanc. Sci. phys. nat. Bordeaux, Année 1897–1898, 233– 236. Gruvel, A., 1899. Archs Biol. Belg., 16, 27–47. Gruvel, A., 1901. P. v. Séanc. Soc.phys. not. Bordeaux, Année 1901, 93–94. Gruvel, A., 1902a. Expéd. sci. Travailleur et du Talisman, 1880–1883, Cirripèdes, 179pp. Gruvel, A., 1902b. P. v. Séanc. Soc. Sci. phys. nat. Bordeaux, Année 1902, 30–31. Gruvel, A., 1905. Monographie des Cirripèdes ou Thécostracés. Masson et Cie, Paris, 472 pp. Hartnoll, R.G., 1967. J. Linn. Soc. Zool., 46, 275–295. Henry, D.P. & McLaughlin, P.A., 1965. Nature, Lond., 207, 1107–1108. Henry, D.P. & McLaughlin, P.A., 1967. Crustaceana, 12, 43–58. Hiro, F., 1936. Bull, biogeogr. Soc. Japan, 6, 215–220. Høeg, J.T., 1982. J. exp. mar. Biol. Ecol., 58, 87–125. Høeg, J.T., 1984. J. exp. mar. Biol. Ecol., 76, 145–156. Høeg, J.T., 1985a. J. exp. mar. Biol. Ecol., 89, 221–235. Høeg, J.T., 1985b. Acta zoologica (Stockh.), 66, 1–45. Høeg, J.T., in press. Phil. Trans. R. Soc. Ser. B. Høeg, J.T. & Lützen, J., 1985. Marine Invertebrates of Scandinavia, Vol. 6, Universitetsforlaget, Oslo, 92 pp. Høeg, J.T. & Ritchie, L.E., 1985. J. exp. mar. Biol. Ecol., 87, 1–11. Hoek, P.P.C., 1883. Rep. scient. Res. Voyage of H.M.S.Challenger, Zoology, 8, 169pp. Hoek, P.P.C., 1884. Rep. scient. Res. Voyage of H.M.S.Challenger, Zoology, 10, No. 3, 47 pp. Hoek, P.P.C., 1906. Proc. K. ned. Akad. Wet., 8, 659–662. Hoek, P.P.C., 1907. The Cirripedia of the Siboga-Expedition. Part A.Cirripedia Pedunculata. Siboga-Expedition, 31, 1–127. Hui, E. & Moyse, J., 1984. J. mar. biol. Ass. U.K., 64, 91–97. Ichikawa, A. & Yanagimachi, R., 1958. Annotnes zool. jap., 31, 82–96. Ichikawa, A. & Yanagimachi, R., 1960. Annotnes. zool. jap., 33, 42–56. Idelman, S., 1967. Année Biol., 4, 113–190. Karande, A.A., 1974. Indian J. mar. Sci., 3, 173–177. Klepal, W., 1983. Zool. Scr., 12, 115–125. Klepal, W., 1985. P.S.Z.N.I., Mar. Ecol., 6, 47–119. Klepal, W., Barnes, H. & Munn, E.A., 1972. J. exp. mar. Biol. Ecol., 10, 243–265. Krüger, P., 1911. Abh. math.-phys. Kl. K.Bayer Akad. Wiss., II, Suppl. 6, 72 pp. Krüger, P., 1914. Die Fauna Südwest-Australiens, Vol. 4, 429–441. Kühnert, L., 1934. Z. Morph. Ökol. Tiere, 29, 45–78. Laloy, M.L., 1903. Rev. Sci. (Paris), 4th Ser., 19, 360–366. Lande, R., 1980. Evolution, 34, 292–305. Lewis, C.A., 1978. In, Settlement and Metamorphosis of Marine Invertebrate Larvae, edited by F.-S.Chia & M.E.Rice, Elsevier North-Holland, New York, pp. 207– 218. Lützen, J., 1981a. J. exp. mar. Biol. Ecol., 50, 231–254. Lützen, J., 1981b. J. exp. mar. Biol. Ecol., 53, 241–249. Lützen, J., 1984. Sarsia, 69, 91–106. Maynard-Smith, J., 1978. The Evolution of Sex. Cambridge University Press, Cambridge, 222 pp. McLaughlin, P.A. & Henry, D.P., 1972. Crustaceana, 22, 13–30. Moyse, J., 1983. J. mar. biol. Ass. U.K., 63, 161–180. Newman, W.A., 1971. J. Zool., 165, 423–429. Newman, W.A., 1974. J. mar. biol. Ass. U.K., 54, 437–456. Newman, W.A., 1979. Trans. S.Diego Soc. nat. Hist., 19(11), 153–167. Newman, W.A., 1980. Téthys, 9, 379–398. Newman, W.A. & Ross, A., 1976. Mem. S.Diego Soc. nat. Hist., No. 9, 108 pp.

COMPARATIVE ANATOMY OF MALES IN CIRRIPEDES

303

Newman, W.A., Zullo, V.A. & Withers, T.H., 1969. In, Treatise on Invertebrate Paleontology, Part R, Arthropoda, 4, Vol. 1, edited by R.C.Moore, Geol. Soc. Amer., University of Kansas Press, pp. R206–R295. Nilsson-Cantell, C.A., 1921. Zool. Bidr. Upps., 7, 75–395. Nilsson-Cantell, C.A., 1930a. J. Linn. Soc. (Zool.), 37, 61–78. Nilsson-Cantell, C.A., 1930b. Senckenbergiana, 12, 210–213. Nilsson-Cantell, C.A., 1931. Ark. Zool. 23(A), No. 18, 12 pp. Osche, G. 1972. Evolution. Grundlagen-Erkenntnisse-Entwicklungen der Abstammungslehre. Herder (Studio visuell), Basel, Wien, 116 pp. Perez, C., 1931. C. r. hebd. Séanc. Acad. Sci., Paris, 193, 195–197. Pilsbry, H.A., 1890. Proc. Acad. not. Sci. Philad., 441–443. Pilsbry, H.A., 1908. Proc. Acad. not. Sci. Philad., 60, 104–111. Pochon-Masson, J., 1968a. Annls Sci. not. Zool., Ser. 12, 10, 1–100. Pochon-Masson, J., 1968b. Annls Sci. not. Zool., Ser. 12, 10, 367–454. Pochon-Masson, J., Bocquet-Védrine, J. & Turquier, Y., 1970. In, Comparative Spermatology, edited by B.Baccetti. Proc. Int. Symp. Rome and Siena, July 1969, 206–219. Por, F.D., 1972. The Heinz Steinitz Mar. Biol. Lab. Scient. Newsl., No. 2, 3–4. Potts, F.A., 1915. Publs Carnegie Instn., 8, No. 212, 32 pp. Ralls, K. 1977. Am. Nat., 111, 917–938. Reinhard, E.C., 1942. J. Morphol., 70, 389–402. Ritchie, L.E. & Høeg, J.T., 1981. J. crust. Biol., 1, 334–347. Schaller, F., 1974. In, Grizimeks Tierleben, edited by K.Immelmann, Enzyklopädie des Tierreiches, Sonderband Verhaltensforschung. Kindler, Zurich, pp. 392–405. Scudo, F.M., 1969. Evolution, 23, 36–49. Sewell, R.B.S., 1926. Rec. Indian Mus., 28, 269–332. Smith, G., 1906. Fauna und Flora des Golfes von Neapel, 29, 123 pp. Smith, G., 1910. Q. Jl. microsc. Sci., 54, 577–604. Smith, G., 1911. Q. Jl. microsc. Sci., 57, 251–265. Smith, G., 1913. Q. Jl. microsc. Sci., 59, 267–295. Stewart, F.H., 1911. Mem. Indian Mus., 3, 33–51. Stubbings, H.G., 1936. Scient. Rep. John Murray Exped. 1933–1934, 4, No. 1, 70 pp. Stubbings, H.G., 1967. Bull. Brit. Mus. (Nat. Hist.) Zool., 15, 227–319. Svane, J., 1986. Mar. Biol., 90, 249–253. Svennevig, N., 1978. Cand. scient, thesis, University Copenhagen, 107 pp. Szöllösi, A., 1975. J. ultrastruct. Res., 50, 322–346. Thomson, W., 1873. Nature, Land., 8, 347–349. Tomlinson, J.T., 1955. J. Morph., 96, 97–114. Tomlinson, J.T., 1960. Bull. Inst. Francais d’Afrique Noire, 22A, 402–410. Tomlinson, J.T., 1963. Pacif. Sci., 17, 299–301. Tomlinson, J.T., 1966. J. theoret. Biol., 11, 54–58. Tomlinson, J.T., 1969. Bull. U.S. natn. Mus., No. 296, 162 pp. Tomlinson, J.T., 1973. Wasmann J. Biol., 31, 263–288. Tomlinson, J.T. & Newman, A., 1960. Proc. U.S. natn. Mus., 112, 517–526. Turquier, Y., 1967. Archs Zool. exp. gén., 108, 111–137. Turquier, Y., 1970a. Archs Zool. exp. gén., 111, 265–300. Turquier, Y., 1970b. C. r. Acad. Sci., Paris, 270, 960–962. Turquier, Y., 1971. Archs Zool. exp. gén., 112, 301–348. Turquier, Y., 1972. Archs Zool. exp. gén., 113, 167–196. Turquier, Y., 1976. Bull. Soc. zool. Fr., 101, 559–574. Turquier, Y., 1977. Archs Zool, exp. gén., 118, 133–154.

304

WALTRAUD KLEPAL

Turquier, Y., 1985a. Bull. Soc. zool. Fr., 110, 151–168. Turquier, Y., 1985b. Bull. Soc. zool. Fr., 110, 169–189. Turquier, Y. & Pochon-Masson, J., 1969. Archs Zool. exp. gén., 110, 453–470. Utinomi, H., 1958. Jap. J. Zool., 12, 113–122. Utinomi, H., 1960. Publs Seto mar. biol. Lab., 8, 223–279. Utinomi, H., 1961. Publs Seto mar. biol. Lab., 9, 413–446. Vaghin, V.L., 1946. C. r. Acad. Sci. Moscow N.S., 25, 273–276. Vagin, V.L., 1937. C. r. Acad. Sci. Leningrad, 15, 273–278. Veillet, A., 1943. Bull. Inst. océanogr. Monaco, No. 841, 4 pp. Veillet, A., 1945. Annls Inst. océanogr. Monaco, 22, 193–341. Veillet, A., 1955. Bull. Soc. Sci. Nancy, 14, 73–74. Veillet, A., 1960. Bull. Soc. Sci. Lorraine, 19, 90–93. Veillet, A., 1985. Bull. Acad. Soc. Sci. Lorraines, 24, 141–146. Walker, G., 1980. Phil. Trans. R. Soc. Ser. B, 233, 89–136. Walker, G., 1985. J. exp. mar. Biol. Ecol., 93, 131–145. Walley, L.J. 1969. Phil. Trans. R. Soc. Ser. B, 256, 237–280. Wingstrand, K.G., 1972. K. dansk Vidensk. Selsk. Skr., 19, No. 4, 72 pp. Withers, T.H., 1928. Catalogue of Fossil Cirripedia. British Museum (Nat. Hist.) London, 154 pp. Yanagimachi, R., 1960. Bull. mar. biol. Stn Asamushi, 10, 109–110. Yanagimachi, R., 1961a. Biol. Bull. mar. biol. Lab. Woods Hole, 120, 272–283. Yanagimachi, R., 1961b. Crustaceana, 2, 183–186. Yanagimachi, R. & Fujimaki, N., 1967. Annotnes Zool. Jap., 40, 98–104. Zevina, G.B., 1978a. Zool. Zh., 57, 998–1007, (in Russian, English abstract). Zevina, G.B., 1978b. Zool. Zh., 57, 1343–1352, (in Russian, English abstract). Zevina, G.B., 1980. Zool. Zh., 59, 689–698, (in Russian, English abstract).

Oceanogr. Mar. Biol. Ann. Rev., 1987, 25, 353–505 Margaret Barnes, Ed. Aberdeen University Press

THE BENGUELA ECOSYSTEM. PART IV. THE MAJOR FISH AND INVERTEBRATE RESOURCES R.J.M.CRAWFORD, L.V.SHANNON and D.E.POLLOCK Sea Fisheries Research Institute, Private Bag X2, Rogge Bay 8012, Cape Town, South Africa

INTRODUCTION The Benguela in the Southeast Atlantic is one of the four major eastern boundary current regions of the world ocean (Wooster & Reid, 1963), all of which are characterized by equatorward surface flow, coastal upwelling and faunal assemblages that are in many respects similar (Parrish, Bakun, Husby & Nelson, 1983). In these regions high primary productivity supports large commercial fisheries (Cushing, 1969), that for anchoveta Engraulis ringens in the Peruvian system having been the largest that the world has known (Cushing, 1982). This paper reviews aspects of the major fish and invertebrate resources of the Benguela ecosystem, and in doing so continues a series that commenced with a consideration of physical features and processes (Shannon, 1985), and continued with treatment of chemical processes (Chapman & Shannon, 1985) and plankton, including ichthyoplankton (Shannon & Pillar, 1986). Birds and mammals are not dealt with in this paper except in so far as they are of relevance to understanding the dynamics of fish resources, it being intended that these higher predators will form the subject of a later paper. The living resources of the Southeast Atlantic have been reviewed previously by Newman (1977) and De Villiers (1985), and a bibliography of marine biology in South Africa has been published by Darracott & Brown (1980) and updated by Brown (1985). Consequently, no attempt has been made in this paper to reference every work. Rather it has been essayed to synthesize important aspects of the system’s major resources. THE MAJOR FISHERIES The development of fisheries in the Benguela Current region has been sketched by De Villiers (1985), who showed that overall catches from the Southeast Atlantic climbed rapidly during the 1950s and 1960s, attaining a peak of over three million metric tons in 1968, and subsequently fluctuating around a level above two million tons (Fig. 1). De Villiers (1985) identified three major resource groups, hakes Merluccius spp., horse mackerels Trachurus spp., and epipelagic shoaling fishes notably pilchard or sardine Sardinops ocellatus, anchovy Engraulis japonicus and sardinellas Sardinella spp. He noted that since 1960 these groups had on average contributed about 87·5% of the overall catch of finfish and shellfish from the region. He demonstrated that the initial rapid rise in catches was due to expansion of purse-seine fisheries exploiting the epipelagic species, that when yields from these resources decreased catches were maintained through increased harvests by demersal trawl of the bathypelagic hakes, and that in turn a decrease in yield of hakes was offset by improved catches of horse mackerel made mainly by midwater trawl (Fig. 1).

306

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Fig. 1.—Cumulative catches of major species groups in the ICSEAF convention area, 1950–1984 (from information in De Villiers, 1985; ICSEAF Statistical Bulletins): catches of sardinellas Sardinella spp., pilchard Sardinops ocellatus, anchovy Engraulis japonicus and round herring Etrumeus whiteheadi have been made mostly with purse-seine nets; those of horse mackerels Trachurus spp. and chub mackerel Scomber japonicus up until the early 1970s mostly with purseseine nets and subsequently mostly with midwater trawls; those of snoek Thyrsites atun since the mid-1970s mostly with midwater or bottom trawls; those of hakes Merluccius spp. mostly with bottom trawls.

Many of the fish stocks are exploited by a number of nations, the International Commission for the Southeast Atlantic Fisheries (ICSEAF), which met for the first time in 1971, advising on the management of shared resources (Botha, 1980). This Commission has divided the Southeast Atlantic into a number of divisions for the reporting of catch statistics. These divisions are illustrated in Figure 2, together with major upwelling centres and some of the localities referred to in the text. PURSE-SEINE FISHERIES The purse-seine fisheries off South and South West Africa (Namibia) have been based largely on pilchard, anchovy, and Cape horse mackerel Trachurus capensis. Off South Africa chub mackerel Scomber japonicus, round herring Etrumeus whiteheadi and lanternfish, mainly Lampanyctodes hectoris, have also made important contributions to the landings. Fisheries in the two regions have generally been regarded as operating on independent stocks, although the question of stock separation remains a topic for debate (cf. Butterworth, 1983; Crawford & Kriel, 1985).

THE BENGUELA ECOSYSTEM: PART IV

307

The two main areas of exploitation, South Africa’s Western Cape and the vicinity of Walvis Bay, are separated by the principal upwelling centre of the Benguela system in the vicinity of Lüderitz and the zone of greatest negative surface temperature anomaly that occurs between 23 and 31° S (Shannon, 1985). The cool water is believed to form a major environmental barrier that effectively divides the system in two (Boyd & Cruickshank, 1983; Shannon, 1985; Schülein, 1986). Spawning of most epipelagic stocks occurs either north or south of the barrier (Shannon & Pillar, 1986), the cool, highly turbulent, weakly stratified water of the Lüderitz area not being conducive to development of early larval stages (e.g. King, 1977a; King, Robertson & Shelton, 1978). Anchovy larvae larger than 20 mm have been recorded between Lüderitz and the Orange River, and it has been suggested that they emanate from the Western Cape (Badenhorst & Boyd, 1980; Boyd & Hewitson, 1983). Tagging studies indicated no movement of pilchard from the Western Cape to Namibia and only minimal movement in the opposite direction (Newman, 1970b). Furthermore, catches of pilchard off the Western Cape fell dramatically after 1962, but those off Namibia only after 1968 (Table I). Grant (1985), however, found no genetic differences in muscle proteins between pilchards off Namibia and those off South Africa. In the Southeast Atlantic a purse-seine fishery is also conducted off Angola, in the vicinity of and to the north of the Angola-Benguela front at about 15– 17° S (Shannon, Agenbag & Buys, 1987). The main species sought are the sardinellas Sardinella aurita and S. eba and Cunene horse mackerel Trachurus trecae (Newman, 1977), although substantial quantities of pilchard have also been caught off southern Angola in the vicinity of Baia dos Tigres (16°30′ S), probably part of the northern Benguela stock (Schülein, 1971; De Campos Rosado, 1974a). Time series of catches by all gears of sardinellas, pilchard and anchovy in the Southeast Atlantic are listed in Table I. The catches were mostly made by purse-seine, and show that marked decreases in yields of pilchard from both the northern and southern Benguela were offset to some extent by TABLE I Recorded catches (thousands of metric tons) from the southeastern Atlantic of sardinellas Sardinella spp., pilchard Sardinops ocellatus and anchovy Engraulis japonicus, 1950–1984: catches of pilchard and anchovy have been allocated to northern and southern areas, corresponding approximately to the regions north and south of the Lüderitz Divide; (from De Campos Rosado, 1974a; Crawford, Shelton & Hutchings, 1983; ICSEAF Statistical Bulletins) Year 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963

Sardinella

Sardinops ocellatus

Engraulis japonicus

spp.

North

South

Total

North

South

Total

64·7 87·2 49·6 37·5 33·4 52·2 56·6 46·4

46·7 127·2 225·8 262·2 250·6 227·1 308·2 374·0 268·8 278·9 293·1 350·1 411·4 607·7

85·3 101·9 170·0 132·5 88·3 121·9 76·6 109·5 194·5 260·2 318·0 402·2 410·2 390·1

132·0 229·1 395·8 394·7 338·9 349·0 384·8 483·5 463·3 539·1 611·1 752·3 821·6 997·8

– – – – – – – – – – – – – –

– – – – – – – – 0·2 1·4 – – – 0·3

– – – – – – – – 0·2 1·4 – – – 0·3

308

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Fig. 2.—The southern African coast showing the major ICSEAF Divisions and centres of upwelling (from De Villiers, 1985; Shannon, 1985).

THE BENGUELA ECOSYSTEM: PART IV

Sardinella

Sardinops ocellatus

Year

spp.

North

South

Total

North

South

Total

1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984

85·9 42·6 54·3 44·6 78·8 156·4 67·3 86·4 144·8 166·1 166·1 21·4 21·0 135·0 240·6 181·7 182·6 160·4 181·3 153·7 179·9

717·6 780·2 775·6 969·6 1400·1 1178·4 564·6 327·5 446·9 469·2 730·4 564·8 460·6 276·0 51·0 39·1 11·8 52·4 53·3 48·5 73·3

256·1 204·5 118·0 69·7 107·8 56·1 61·8 87·6 104·2 69·0 16·0 89·2 176·4 57·8 97·0 52·9 50·4 46·2 33·5 60·5 27·2

973·7 984·7 893·6 1039·3 1507·9 1234·5 626·4 415·1 551·1 538·2 746·4 654·0 637·0 333·8 148·0 92·0 62·2 98·6 86·8 109·0 100·5

0·6 1·0 3·3 24·3 161·2 226·1 188·9 184·7 149·5 361·1 249·1 191·1 89·3 132·9 363·6 288·1 209·8 216·5 86·5 187·4 16·9

92·4 171·0 143·9 270·6 138·1 149·2 169·3 157·3 235·6 250·9 349·8 223·6 218·3 235·5 209·5 291·4 315·5 292·0 306·9 240·2 272·5

93·0 172·0 147·2 294·9 299·3 375·3 358·2 342·0 385·1 612·0 598·9 414·7 307·6 368·4 573·1 579·5 525·3 508·5 393·4 427·6 289·4

309

Engraulis japonicus

improved harvests of anchovy. Catches of sardinellas were generally higher after the mid-1960s than in earlier years. Yields of horse mackerel from the ICSEAF regions are listed in Table II and refer to the combined catches of Trachurus spp. by all gears. Catches up until the mid-1960s were made mainly with purse-seine, but the large harvests TABLE II Recorded catches (thousands of metric tons) of horse mackerel Trachurus spp. from the southeastern Atlantic, with approximate allocation to ICSEAF Divisions, 1950–1984 (from De Campos Rosado, 1974a; Babayan et al., 1985; Kinloch et al., 1986; Kinloch, in prep.; ICSEAF, 1985; ICSEAF Statistical Bulletins)

Year 1950 1951 1952

1.1

T. trecae

T. capensis

Trachurus spp. total

ICSEAF Division

ICSEAF Division

ICSEAF Division

1.2

1.3

1.4 All

1.1 1.2 1.3 1.4 1.5

1.6

2.1

– – –

50·2 100·3 102·7

0·0 0·1 0·0 0·2 0·0 0·1

– – –

– – –

– – –

– – –

2.2

Other All 50·3 100·5 102·8

All

310

Year

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

1.1

T. trecae

T. capensis

Trachurus spp. total

ICSEAF Division

ICSEAF Division

ICSEAF Division

1.2

1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965

– – – – – – – – – 0·4

1966

0·4

5·1

1967

0·2

3·0

1968

–

5·6

1969

–

2·1

1970

–

1·3

1971

–

5·5

1972

–

44·5

1973

0·8

7·5

1974

0·0

0·0

1975

0·6

2·8

1976

0·2

3·8

1977

0·1

12·7

1.3

85·7 37·0 38·4 20·0 30·1 43·6 33·5 70·4 3·2 82·5 8·1 90·6 6·2 77·5 10·3 60·2 38·9 110·3 22·4 104·6

1.4 All

1.1 1.2 1.3 1.4 1.5

1.6

2.1

2.2

– – – – – – – – – –

– – – – – – – – – – – – –

86·0 120·6 80·6 46·8 85·4 58·3 19·3 66·2 42·1 69·2 24·0 25·4 56·5

0·0 – 0·0 0·0 0·0 0·0 0·1 0·2 0·2 0·2 0·3 3·3 3·1

0·0 0·1 0·2 0·3 0·2 0·2 0·3 0·3 0·2 0·4 0·2 1·7 1·7

122·7 58·4 73·7 103·9 85·7 98·7 83·7 70·5 149·2 127·4

177· 3 163· 0 122· 5 99·9

–

163· 1 133· 2 236· 8 183· 4 132· 9 124· 8 41·8

–

238· 6

2·1

– – –

– – – – – –

– – – – – – – – – – – – –

– – – – – – – – – – – – – – – – 47·3 22·6 21·2 70·8 126·4

– – – – – – – – – – – – –

Other All 86·0 120·7 80·8 47·1 85·6 58·5 19·7 66·7 89·8 92·4 45·7 101·2 187·7

All

169·8 144·0 132·2 123·6 152·4 188·5 176·1 116·2 250·4 315·1

182· 8 166· 2 128· 1 102· 0 164· 4 138· 7 281· 3 191· 7 133· 0 128· 2 45·7

–

–

99·6

–

28·2

3·0

2·3

–

–

72·4

–

9·8

2·4

2·4

133· 1 87·0

–

–

69·2

–

1·6

5·1

3·7

79·6

–

–

46·5

–

27·2

7·0

3·8

84·5

–

–

50·4

1·0

11·1

9·7

4·5

76·7

–

–

212·4

2·1

8·4

20·2

6·8

–

–

71·7

0·7

3·6

13·5

6·1

249· 9 95·6

–

0·1

75·2

10·2

12·2

15·9

10·7

–

–

79·5

4·0

9·6

26·4

0·1

0·2

3·7

5·0

33·9

0·1

4·9

11·3

31·6

1·5

15·0

206· 2 220· 7

8·6

252· 7

202· 8 292· 9 135· 6

177· 1 101· 5 62·7

8·2

23·0

69·9

301· 4 6·1 227· 1 19·0 327· 4 7·9 0·2 563· 7 23·2 0·5 497· 6

315· 9 253· 2 207· 7 186· 5 241· 1 388· 6 376· 9 493· 1 360· 1 455· 6 609· 4 750· 3

THE BENGUELA ECOSYSTEM: PART IV

1978

17·5

42·7

–

35·9

319· 9 222· 0 57·7

1979

18·1

55·3

1980

16·1

1981

15·4

31·7

95·1

–

1982

19·5

38·1

47·5

–

1983

20·4

37·6

51·9

–

1984

1·5

6·9

46·5

–

0·00

52·3 3 8·78

0·00

0·00

0·00

10·4 1 29·6 9

47·6 7 91·1 3 87·3 2 58·7 2

59·6 4 83·0 6 92·8 6

16·3 0

81·2 0

% of 1950 s 1960 s 1970 s 1980 – 1984 1950 – 1984

0·08 2·14 11·6 0 2·44

1·8 –

0·12 0·00

0·06

380· 2 297· 2 109· 7 142· 2 105· 1 110· 0 54·9

–

0·4

–

0·4

0·0

0·1

0·0

0·0

0·0

0·1

0·1

0·1

–

–

354· 7 290· 1 440· 9 430· 9 534· 9 427· 9 353· 7

183· 5 120· 9 91·0

9·6

6·5

28·8

5·0

7·9

33·0

14·2

2·6

27·3

152· 0 118· 7 165· 7 246· 4

7·5

7·3

27·3

6·3

1·9

26·6

6·7

2·8

21·9

6·8

4·6

21·7

100· 00 36·2 6 4·78

0·00 4·10

1·91

5·23

16·3 4

7·32

76·3 4

5·2

7·1

595· 8 12·5 6·3 476· 1 9·0 0·0 585· 1 2·6 0·1 627· 7 3·6 0·1 692· 2 13·6 0·0 638· 8 11·7 – 644· 9

311

976· 0 773· 3 694· 8 769· 9 793· 3 748· 8 699· 8

12·16

in the 1970s and early 1980s were mostly taken with demersal and especially midwater trawls. The South African fishery Pilchard caught off the Western Cape were canned as early as 1935, but commercial fishing operations only commenced in 1943 in response to a demand for canned products during World War II (Du Plessis, 1959). The fishery was unrestricted prior to 1950, but subsequent management measures have included catch quotas, closed seasons and areas, minimum mesh sizes and limitations on fleet and processing capacities (Du Plessis, 1959; Stander & Le Roux, 1968; Gertenbach, 1973; Newman, Crawford & Centurier-Harris, 1979; Crawford, 1981a; Butterworth, 1983). In 1943 the combined catch of pilchard and horse mackerel was estimated as 6500 metric tons; in 1949 it was estimated as 113000 tons of which 68000 tons were pilchard. Accurate records of performance are only available from 1950, but show that pilchard and horse mackerel dominated the landings until well into the 1960s (Table III). Catches of all species rose to over 500000 tons in 1967 and from 1968 to 1985 were fairly stable with a mean of about 380000 tons. This “stability” probably resulted from quota restrictions on allowable catch. Landings of horse mackerel averaged 80000 tons between 1950 and 1958. During this period the length composition of horse mackerel in catches showed clear progression of a distinct mode (Fig. 3), which

312

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Geldenhuys (1973) identified as consisting mainly of the 1946, 1947 and 1948 year-classes. Landings from 1959 to 1966 were lower, averaging about 40000 tons, and subsequently horse mackerel was a minor contributor to the South African purse-seine catch, with landings after 1970 never exceeding 10000 tons. Annual catches of pilchard averaged 134000 tons in the 1950s but rose to in excess of 400000 tons during the early 1960s. As with horse mackerel these peak catches were based on powerful year-classes, in this instance those that were formed during the late 1950s (Centurier-Harris, 1977; Newman & Crawford, 1980). After the strong year-classes had passed through the fishery yields of pilchard fell rapidly. Only 70000 tons were harvested in 1967, and subsequently catches exceeding 100000 tons were only made on three occasions. Catches during the 20-year period 1966–1985 averaged 71000 tons, about half the mean of the 1950s. Following decreases in the catches of horse mackerel and pilchard, processors sought alternative resources to bolster diminishing returns. During 1963 1965 the minimum stretched mesh size of nets was reduced drastically to 13 mm and exploratory fishing for anchovies was initiated (Newman et al., 1979). A minimum mesh size of 38 mm had been introduced in 1950. Nets were then constructed from natural fibres, but shrinkage resulted in an effective mesh size of about 32 mm (Newman et al., 1979). This lower size had been adopted with the introduction of nets made from synthetic materials in 1956. From 1966 anchovy were consistently the major contributor to purse-seine catches off the Western Cape. A peak anchovy catch of 350000 tons was recorded in 1974. Between 1972 and 1985 yields averaged 265000 tons and never fell below 200000 tons. TABLE III Species composition of landings (thousands of metric tons) by South Africa’s Western Cape purse-seine fleet, 1950– 1985 (updated from Crawford, Shelton & Hutchings, 1983) Year

Sardinops

Trachurus

Engraulis

Scomber

Etrumeus

Lampanyctodes

Other*

Total

1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968

85·3 101·9 170·0 132·5 88·3 121·9 76·6 109·5 194·4 260·2 318·0 402·2 410·2 390·1 256·1 204·5 118·0 69·7 107·8

49·9 98·5 102·6 85·2 118·1 78·8 45·8 84·6 56·4 17·7 62·9 38·9 66·7 23·2 24·4 55·0 26·3 8·8 1·4

– – – – – – – – 0·2 1·4 – – – 0·3 92·4 171·0 143·9 270·6 138·1

– – – – 4·0 20·2 32·6 7·4 21·6 33·1 31·0 49·7 20·4 13·2 50·0 41·4 53·4 128·2 91·0

– – – – – – – – 0·8 2·6 0·1 0·1 0·1 0·2 2·7 8·2 15·4 32·0 30·3

– – – – – – – – – – – – – – – – – – 0·1

– – – – – – – – – – – – – – – – – – –

135·2 200·4 272·6 217·7 210·5 221·0 154·9 201·5 273·4 314·9 412·0 490·9 497·3 427·0 425·6 480·1 357·1 509·3 368·6

THE BENGUELA ECOSYSTEM: PART IV

Year

Sardinops

Trachurus

Engraulis

1969 56·1 26·8 149·2 1970 61·8 7·9 169·3 1971 87·6 2·2 157·3 1972 104·2 1·3 235·6 1973 69·0 1·6 250·9 1974 16·0 2·5 349·8 1975 89·2 1·6 223·6 1976 176·4 0·4 218·3 1977 57·8 1·9 235·5 1978 97·0 3·6 209·5 1979 52·9 4·3 291·4 1980 50·4 0·4 315·5 1981 46·2 6·1 292·0 1982 33·5 1·1 306·9 1983 60·5 1·4 240·2 1984 27·2 2·5 272·5 30·7 0·8 272·6 1985 * Includes Scomberesox and Thyrsites.

Scomber

Etrumeus

Lampanyctodes

Other*

Total

91·7 77·9 54·2 56·7 58·8 30·7 69·3 0·5 21·3 2·4 2·7 0·2 0·3 2·7 3·8 0·7 0·1

23·3 23·7 21·6 20·6 28·7 1·3 23·6 11·7 35·0 67·0 21·0 14·1 24·3 31·0 69·0 28·6 39·8

4·9 18·2 2·0 15·2 42·4 0·3 0·1 0·1 5·6 1·0 8·7 0·1 10·3 0·7 1·6 13·4 31·0

– – – – – – – – – – – – – – – 1·2 4·9

352·0 358·8 324·9 433·6 451·4 400·5 407·4 407·5 357·2 380·4 380·9 380·5 379·2 375·9 376·5 346·1 379·9

313

Large landings of chub mackerel were made in the late 1960s, based on the powerful 1966 and 1967 yearclasses (Newman & Crawford, 1980). By this time the size structure of the pilchard stock had been considerably reduced through introduction of the small-meshed net, leading to a greatly diminished production of canned pilchard. Therefore the improved catches of chub mackerel came at an opportune time for the industry, enabling overall production of canned fish to be maintained at a reasonable level until the mid 1970s (Crawford, 1981c). Catches of round herring were small prior to 1964, but subsequent yields have averaged about 25000 tons per annum. A peak catch of 69000 tons was recorded in 1983 during the El-Niño-Southern-Oscillation (ENSO) event of 1982–1983 (Shannon, Crawford & Duffy, 1984). The species has relatively large scales which are shed readily, resulting in clogging of the 13-mm mesh and nets being difficult to purse. Therefore, during the 1970s fishermen often avoided round herring when shoals could be identified and other fish were plentiful in the vicinity (Crawford, 1981b). Round herring has since been removed from the list of species subject to quota restrictions, technology enabling the species to be successfully canned has been developed, and the species has become a sought-after target of vessels operating from factories equipped with canning plants. Lanternfish made their first documented appearance in purse-seine catches off the Western Cape in 1966, shortly after the introduction of the smallmeshed net, but the species was only recorded separately from anchovy in 1968 and subsequently (Crawford, Centurier-Harris, Wingate & Kriedemann, 1978). Catches have been small and variable, with a peak of 42000 tons in 1973. The high oil content of the species has resulted in clogging of machinery at fish-meal plants (Centurier-Harris, 1974), and on such occasions plant managers have directed skippers to search for other species (Crawford, 1980b). Quotas were initially implemented in 1953 when the combined catch of pilchard and horse mackerel was limited to 226900 tons. This restriction, subject to the proviso that fishing would be allowed to continue

314

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Fig. 3.—Caudal length distribution of purse-seine catches of Cape horse mackerel Trachurus capensis off South Africa’s Western Cape, 1950–1961, illustrating progression of year-classes formed during the late 1940s (from information in Geldenhuys, 1973).

anyway until 31 August each year if the quota were exceeded before that date, remained in force until 1960 (Crawford, 1981a). From 1958 the measure became increasingly incapable of restricting the catch, and it was abolished in 1961. A combined-species quota was introduced in 1971 and was operative until 1982, during which period the catch ceiling ranged between 360000 and 450000 tons (Crawford, 1981a; Butterworth, 1983). In 1973 catches exceeded the limit by 15000 tons, but in other years the quota was reasonably effective although sometimes adjusted upward during the legal season. Initially processors competed for the overall quota, but from 1975 they were allocated separate quotas. Processors were then

THE BENGUELA ECOSYSTEM: PART IV

315

able to schedule their fishing activities and to wait for fish to become locally available. As a consequence rising fuel bills were contained to some extent and there was improvement in the condition of fish offloaded at processing plants, leading to higher yields of fish-meal from equivalent amounts of raw material (Crawford, 1981a). From 1983 to 1986 catches of anchovy and pilchard were controlled by separate quotas and there was no limit on the harvesting of other species. Attempts have been made to limit fishing effort through restrictions on duration of the fishing season since 1951 and on fleet size since 1953. Closed seasons have been of variable length, and in instances applicable to certain species only, but have generally fallen in the last five months of the year when many of the species spawn, render low yields of oil or are not readily available to the purse-seine fleet (Crawford, 1979). Total fleet hold capacity rose from 4400 metric tons in 1950 to about 7500 tons in 1953, over 10000 tons in 1964 and more than 13000 tons in 1973, initially through an increase in the number TABLE IV Catch and effort information for South Africa’s Western Cape purse-seine fishery, 1950–1985 (updated from Crawford, 1981a; Armstrong et al., 1985) Year

Combined catch Number of of all species catcher boats (thousands of metric tons)

Total hold capacity of fleet (metric tons)

Mean hold Effort capacity of boats (standardboat(metric tons) days)

1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973

135·2 200·4 272·6 217·7 210·5 221·0 154·9 201·5 273·4 314·9 412·0 490·9 497·3 427·0 425·6 480·1 357·1 509·3 368·6 352·0 358·0 324·9 433·6 451·4

4399 5806 6586 7547 7683 7529 7448 7448 7693 7856 7538 8445 8926 9906 10296 10641 10656 11665 12783 12609 12157 12983 12250 13176

29·32 31·73 33·43 32·96 34·15 33·02 33·55 33·47 36·63 47·61 52·71 64·47 73·16 78·00 81·71 85·13 85·25 86·41 98·33 99·28 103·03 107·30 102·08 110·72

150 183 197 229 225 228 222 210 210 165 143 131 122 127 126 125 125 135 130 127 118 121 120 119

3498 3717 3936 4482 5466 9580 8099 12058 11110 10349 11052 10126 13080 12249

Catch per unit effort (metric tons per standardboatday)

117·5 132·5 126·6 95·7 77·9 50·1 44·1 42·2 33·2 34·0 32·5 32·1 33·2 36·9

316

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Year

Combined catch Number of of all species catcher boats (thousands of metric tons)

Total hold capacity of fleet (metric tons)

Mean hold Effort capacity of boats (standardboat(metric tons) days)

Catch per unit effort (metric tons per standardboatday)

1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985

400·5 407·4 407·5 357·2 380·4 380·9 380·5 379·2 375·9 376·5 344·9 375·1

13058 12271 11055 10846 10998 10806 10055 10585 10344 10820 10320 9980

111·61 112·58 111·67 112·98 113·38 116·19 114·26 118·93 121·69 124·37 122·86 120·24

45·9 39·4 47·4 32·0 29·2 34·7 37·8 35·0 33·4 25·7 35·3 42·7

117 109 99 96 97 93 88 89 85 87 84 83

8720 10329 8597 11168 13039 10983 10051 10821 11297 14665 9764 8785

of vessels and then as a result of large increases in vessel size. Fleet hold capacity stabilized at a level of around 11000 tons from 1976 (Table IV). Therefore, in spite of restrictions, it was some two decades before growth of the fleet was curtailed. The restrictions, however, are believed to have had a beneficial influence on the industry, enabling it to cope with wide fluctuations in stock abundance, stock availability and product price, and preventing gross over-investment (Gertenbach, 1973; Newman et al., 1979). Mean hold capacity of boats rose from 30 tons in 1950 to 53 tons by 1960 and 113 tons by 1975, so increasing fishing range (Crawford, 1981a). Echo sounders were introduced between 1955 and 1959, power blocks between 1963 and 1965, sonars from 1965 and fish pumps between 1966 and 1968. An investigation of the fishing power of vessels by the technique of Robson (1966) suggested that fish pumps and sonars increased the performance of an average vessel by 36 and 18%, respectively (Newman et al., 1979). In the Peruvian anchoveta fishery installation of power blocks improved fleet efficiency by 8% (Boerema et al., 1965). For the South African fleet, storage capacity, horse-power and age were all shown to have significant impact on the fishing power of boats (Newman et al., 1979). In 1967 an aircraft was acquired by the industry to locate fish shoals and it probably facilitated exploitation of chub mackerel offshore (Crawford, 1981c). Aerial spotting was discontinued from 1981. Off Namibia, Cram (1977) reported that aerial surveys led to an improvement in catches of between 5 and 15%, but the impact of aerial spotting has not been accounted for in estimates of effort expended by the South African fleet (Newman & Crawford, 1980). Off South Africa fishing effort increased from 3500 standard-boat-days in 1960 to 13000 by 1972 and a peak of over 14000 in 1983 (Table IV). Catch rates fell by almost 60% between 1964 and 1971, then increased until 1976 before returning to, and fluctuating around, a low level (Table IV). Area of exploitation was initially confined to the environs of St Helena Bay (Davies, 1956b), but by the end of 1958 included the entire region between Lamberts Bay and Cape Hangklip (Du Plessis, 1959). In 1959, purse-seine boats entered Walker Bay for the first time. By 1964, boats were operating as far north as Port Nolloth, and by 1974 as far east as Cape Infanta. Extent of the fishing grounds at various stages in the history of the fishery are shown in Figure 4. From the mid-1960s the percentage of the overall purse-seine catch made to the east of Cape Point increased markedly (Fig. 5; Crawford & Shelton, 1981). Catch rates

THE BENGUELA ECOSYSTEM: PART IV

317

have generally been highest in St Helena Bay, off Saldanha Bay and between Walker Bay and Danger Point (Crawford, 1981a). The Namibian fishery The purse-seine fishery off Namibia commenced in 1947 (Cram, 1977) and was initially conducted from the port of Walvis Bay. For a limited period during the 1960s and the early 1970s it also operated out of Lüderitz and from sea-going processing plants (Butterworth, 1983). The sea-going plants, converted exwhaling, factory ships, were supplied by 27 catcher boats fishing the same stocks as the land-based purseseiners, but supposedly operating further than 12 nautical miles (22 km) from the coast (Troadec, Clark & Gulland, 1980). Management restrictions have been similar to those applied to the South African fishery (Troadec et al., 1980; Butterworth, 1983). Until 1968 almost all catches were of pilchard, harvests of this species rising from about 50000 metric tons in 1950 to a peak of almost 1·4 million tons by 1968 but collapsing to only 11000 tons in 1980 (Table V). Yields of pilchard averaged about 210000 tons in the 1950s, 700000 in the 1960s, 350000 in 1970s and 45000 in the early 1980s. Catches of all species exceeded 1·5 million tons in 1968 but were only 133000 tons in 1985, the lowest catch since 1951 when the fishery was at an early stage of development. In 1963 intrusion of warm water from the north (Stander & De Decker, 1969) led to southerly displacement of pilchard shoals, prompting the establishment of a processing plant at Lüderitz in 1964 (Butterworth, 1983). Following the collapse of the pilchard resource, fish in the Lüderitz vicinity, however, became scarce and the plant was closed in 1974. Area of operation of the purse-seine fishery expanded markedly in the 1960s and 1970s (Fig. 6). In particular, during the early 1970s boats were forced to search further north to find fish (Cram, 1977). In 1984 a second southward displacement of pilchard owing to occurrence of warm water off northern Namibia took place, leading to relatively high catch rates in the vicinity of Walvis Bay (Thomas & Boyd, 1985). Refrigerated sea-water boats using 28-mm nets have been employed since 1979 and have been responsible for most of the pilchard catch since 1980 (Thomas, 1986). TABLE V Species composition of landings (thousands of metric tons) by the Namibian purse-seine fleet, 1947–1985 (updated from Crawford, Shelton & Hutchings, 1983) Year

Sardinops

Trachurus

Engraulis

Other*

Total

1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960

1·0 3·0 8·0 46·7 127·2 225·8 262·2 250·6 227·1 227·9 227·5 229·1 273·5 283·3

– – – – – – – – – – – – – –

– – – – – – – – – – – – – –

– – – – – – – – – – – – – –

1·0 3·0 8·0 46·7 127·2 225·8 262·2 250·6 227·1 227·9 227·5 229·1 273·5 283·3

318

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Fig.4.—Areas of purse-seine fishing off South Africa’s Western Cape at various stages in the development of the fishery (from information in Davies, 1956b; Crawford, 1979).

THE BENGUELA ECOSYSTEM: PART IV

319

Fig. 5.—Percentage of the overall purse-seine catch off South Africa’s Western Cape recorded east of Cape Point, and population trends for Cape gannets Morus capensis and jackass penguins Spheniscus demersus located east of Cape Point, 1956–1985 (updated from information in Crawford & Shelton, 1981). Year

Sardinops

Trachurus

Engraulis

Other*

Total

1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972

343·5 397·4 555·2 635·9 666·1 718·5 926·0 1386·6 1111·0 513·7 325·1 373·5

– – – – – – – – – – 140·2 21·7

– – – 0·6 1·0 3·3 24·3 161·2 226·1 188·9 184·7 136·6

– – – – – – – – – – 5·2 2·4

343·5 397·4 555·2 636·5 667·1 721·8 950·3 1547·8 1337·1 702·6 655·2 534·2

320

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Year

Sardinops

Trachurus

Engraulis

Other*

Total

1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985

408·1 561·6 561·4 451·7 200·0 46·0 33·8 10·8 52·4 51·6 44·3 57·8 55·5

12·4 30·1 14·1 23·2 81·9 9·0 27·9 39·2 4·0 67·0 107·0 87·6 23·9

296·1 249·1 186·4 87·8 132·9 355·4 277·5 190·2 199·1 83·5 183·8 13·6 50·6

4·2 2·1 7·7 7·6 1·0 0·1 1·6 0·2 0·6 0·4 3·7 3·3 3·0

720·8 842·9 769·6 570·3 415·8 410·5 340·8 240·3 256·1 202·4 338·8 162·4 133·0

* Includes Scomber, Etrumeus, Lampanyctodes, Sufflogobius, Merluccius and Scyphozoa.

Small-meshed 11-mm nets were introduced from 1963 to 1968 but only became widely used from 1968– 1969 (Cram, 1977; Schülein, Butterworth & Cram, 1978). Anchovy were first landed in 1964, but catches remained small until 1968 when 161000 tons were harvested. Thereafter yields were relatively stable until 1983, averaging about 200000 tons per year, with a peak of 355000 tons in 1978. From 1971 to 1977, on account of the possibility that the pilchard stock might be suffering from competition with anchovy, it was thought desirable to subject the anchovy to heavy fishing pressure (Butterworth, 1983). Later it was realized that anchovy had become the mainstay of the fishery and attempts were made to protect this resource. In 1984 intrusion of warm water from the north led to greatly decreased yields of anchovy (Boyd, Hewitson, Kruger & Le Clus, 1985; Le Clus, 1985), and catches remained low in 1985. As with pilchard, a northward shift in the distribution of catches of anchovy appears to have occurred during the 1970s (Schülein, 1986). From 1971 to 1976, anchovy were seldom caught north of 20° S and some catches were made south of 25° S; from 1977 to 1982 substantial catches were recorded north of 20° S but hardly any from the southern grounds (Fig. 7). Cape horse mackerel were first recorded in the landings in 1971 when the purse-seine catch exceeded 140000 tons, the highest to the end of 1985. This species was the dominant contributor to the overall catch of 1984. Thomas (1984) recorded an unusual catch of more than 1000 tons of Sardinella aurita off Namibia in 1983, which he related to the ENSO event of that year. The time series of catch information is probably subject to inaccuracies that have arisen from illegal practices, including under-recording of landings, dumping and misidentification of species, the last two malpractices being likely to have occurred especially under conditions of individual species quotas (Cram, 1981; Butterworth, 1983). Quotas were introduced in 1954 when the pilchard catch was limited to 251000 tons. From 1955 to 1958 the limit was 227000 tons, but from 1959 quotas were increased sharply coincident with increased catches of pilchard off the Western Cape (Butterworth, 1983). Following the collapse of the Western Cape stock, quotas for pilchard off Namibia were stable between 1964 and 1966 at 653000 to 712000 tons, but were then doubled to levels of about 1·3 million tons from 1968 to 1970. The 1970 catch fell far short of the

THE BENGUELA ECOSYSTEM: PART IV

321

Fig. 6.—Areas of purse-seine fishing off Namibia at various stages in the development of the fishery (from Crawford et al., 1985).

quota. The collapse of the pilchard fishery brought a substantial reduction in the pilchard quota (452000 tons in 1971; 370000 tons in 1972) and introduction of a quota for other species, the intention of which was

322

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Fig. 7.—Percentage of overall catch of anchovy Engraulis japonicus off Namibia recorded between successive degrees of latitude, illustrating the northward shift in location of catches between 1971–1976 and 1977–1982 (from information in Schülein, 1986).

to divert fishing effort away from the pilchard resource (Butterworth, 1983). Quotas for pilchard were, however, then allowed to increase to 575000 tons in 1974. After this date a second collapse of the resource again brought rapidly reducing quotas, desire to conserve the anchovy resource, and implementation of a separate quota for anchovy of 250000 tons in 1980 (Butterworth, 1983). The number of boats engaged in the fishery increased from 66 in 1957 to 129 in 1969, but had fallen again to 52 by 1981 (Table VI). Fleet hold capacity after increasing from about 3000 tons in 1957 to over 10000 tons in 1969, however, remained above 10000 tons until 1980. Schülein (1982) used fuel consumption as a measure of fishing effort. Amount of fuel used by boats increased more than fourfold between 1965 and 1969, but in this period catch rates more than halved (Table VI). By 1972 catch rates were only 30% of the 1965 value, and despite improvement to nearly 50% in 1974, they fell to less than 20% in 1977 following the second collapse of the pilchard resource (Schülein, 1982). In 1981 effort, as measured by fuel consumption, was at a similar level to that of 1965, but the combined species catch was more than 60% less (Table VI). Since 1971 fishing has on occasion been prohibited in certain areas, especially off northern Namibia, from where a lengthy steaming time to Walvis Bay meant that fish were generally returned to processing plants in

THE BENGUELA ECOSYSTEM: PART IV

323

poor condition, in areas where pilchard contributed a high by-catch to catches of other species, or where catches were found to contain large amounts of very small anchovy (Butterworth, 1983). Closed seasons have also been in operation, usually from late spring to early autumn. As off the Western Cape they have been chosen to coincide with spawning and low yields of oil from fish (Butterworth, 1983). The Angolan fishery The early history of the Angolan purse-seine fishery has been described by De Campos Rosado (1974a). Four purse-seiners with a total gross registered tonnage of 40 metric tons and 17 fish-meal plants with a total production TABLE VI Catch and effort information for the Namibian purse-seine fishery, 1957–1985 (updated from Schülein, 1982) Year

Combined catch Number of of all species catcher boats (thousands of metric tons)

Total hold capacity of fleet (metric tons)

Mean hold Effort capacity of boats (thousands of (metric tons) litres of diesel used)

1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981

227·5 229·1 273·5 283·3 343·5 397·4 555·2 636·5 667·1 721·8 950·3 1547·8 1337·1 702·6 655·2 534·2 720·8 842·9 769·6 570·3 415·8 410·5 340·8 240·3 255·5

2996 3023 2990 3076 3203 3554 3791 5314 5542 6552 7535 8717 10475 9701 9306 10279 10270 10185 10760 11113 11904 11130 10509 9649 9709

45·4 45·8 46·0 46·6 47·8 51·5 53·4 58·4 60·9 63·0 66·1 74·5 81·2 87·4 94·0 99·8 102·7 105·0 109·8 113·4 124·0 137·4 154·5 182·1 186·7

66 66 65 66 67 69 71 91 91 104 114 117 129 111 99 103 100 97 98 98 96 81 68 53 52

4834 5595 10218 19348 22663 14949 14891 13029 13515 12393 16374 15008 15992 10263 9565 6495 4733

Catch per unit effort (metric tons per thousand litres diesel used)

138 129 93 80 59 47 44 41 54 68 47 38 26 40 36 37 54

324

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Year

Combined catch Number of of all species catcher boats (thousands of metric tons)

Total hold capacity of fleet (metric tons)

Mean hold Effort capacity of boats (thousands of (metric tons) litres of diesel used)

Catch per unit effort (metric tons per thousand litres diesel used)

1982 1983 1984 1985

202·4 338·8 162·4 133·0

9708 10493 10663 9693

194·2 201·8 209·1 210·7

25 35 35 32

50 52 51 46

8220 9723 4632 4111

capacity of 82 tons of fish-meal per hour were operative in 1945. By 1972 the fleet had increased to 361 vessels with a combined gross tonnage of 14909 tons, and the number of reduction plants to 58 with a combined hourly production capacity of 472 tons. The combined species catch fell from 349000 metric tons in 1956 to 182000 tons in 1963, the year in which a warm-water anomaly caused southerly displacement of epipelagic species off Namibia (Stander & De Decker, 1969), and possibly also off Angola (Fig. 8). Catches then climbed in a fluctuating fashion to 570000 tons in 1972 (Table VII). Between 1945 and 1972, 61% of the Angolan purse-seine catch was made in ICSEAF Division 1.3 (i.e. south of 15° S), 35% in Division 1.2 and 4% in Division 1.1 (De Campos Rosado, 1974a). In 1956 and 1957, pilchard was the main contributor to catches from Division 1.3, but from 1959 to 1972 the bulk of the catches were of horse mackerel, mainly Trachurus trecae (Fig. 8). In Division 1.2 horse mackerel dominated the catches in 1956 and were important in the landings up until 1959. Sardinellas were the most important group from 1957 to 1972 (Fig. 8) and were also the main species caught in Division 1.1 (De Campos Rosado, 1974a). Following the cessation of Portuguese rule in Angola, catches from the region of Cunene horse mackerel and sardinellas were low from 1974 to 1976 (Tables I and II), but subsequently improved largely as a result of participation in the fishery by distant-water fleets (De Villiers, 1985) using midwater and demersal trawls as well as purse-seines. Catches of T. trecae from the Southeast Atlantic peaked at 378000 tons in 1978, those of sardinellas at 241000 tons in 1978. Venediktova (1985) reported that larger T. trecae tended to predominate in purse-seine catches. TABLE VII Species composition of landings (thousands of metric tons) by the Angolan purse-seine fleet, and information on number of boats and gross registered capacity (metric tons) of the fleet, 1945–1972 (from De Campos Rosado, 1974a) Landings Year 1945 1946 1947 1948 1949 1950

Sardinella spp. Trachurus trecae Sardinops ocellatus

Number of catcher Gross registered boats capacity of fleet Other Total 4 7 13 19 31 73

40 98 219 376 694 1595

THE BENGUELA ECOSYSTEM: PART IV

325

Fig. 8.—Percentage contribution of sardinellas Sardinella spp., Cunene horse mackerel Trachurus trecae, pilchard Sardinops ocellatus and other species to Angolan purse-seine catches in ICSEAF Divisions 1.2 and 1.3, 1956–1972 (from De Campos Rosado, 1974a). Landings Year 1951 1952 1953 1954 1955 1956 1957

Number of catcher Gross registered boats capacity of fleet

Sardinella spp. Trachurus trecae Sardinops ocellatus

64·7 87·2

122·7 58·4

80·3 146·5

Other Total

81·2 38·2

98 125 142 170 207 348·9 254 330·3 294

2280 3024 3490 4355 6030 8239 10054

326

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Landings

Number of catcher Gross registered boats capacity of fleet

Year

Sardinella spp. Trachurus trecae Sardinops ocellatus

Other Total

1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972

49·6 37·5 33·4 52·2 56·6 46·4 85·9 42·6 54·3 44·6 78·8 156·4 67·3 86·4 141·8

65·5 64·7 66·8 45·9 51·1 41·3 50·7 58·2 57·0 47·2 53·9 60·9 63·9 65·9 74·3

73·7 104·0 85·7 98·7 83·7 70·5 151·4 127·4 182·8 166·2 128·1 102·0 164·4 138·7 281·3

39·7 5·4 9·8 2·6 8·0 23·5 1·7 0·1 1·1 0·6 1·5 62·4 50·9 1·6 72·8

228·5 211·6 195·7 199·4 199·4 181·7 289·7 228·3 295·2 258·6 262·3 381·7 346·5 292·6 570·2

325 347 348 349 353 354 354 360 374 376 375 372 387 387 361

11527 13156 13254 13481 13826 13624 13678 13866 14469 14552 14472 14410 15231 15540 14909

MIDWATER-TRAWL FISHERIES The main target for midwater trawlers operating in the Benguela region is horse mackerel Trachurus spp., although chub mackerel Scomber japonicus and snoek Thyrsites atun are also sought. All three genera are caught as well by demersal trawl and purse-seine, although snoek only in minimal quantities by the latter. Horse mackerel form dense concentrations in the pelagic zone at night, when they are fished successfully with midwater trawl, and congregate near the bottom by day when they are exploited with demersal trawl (Konchina, 1986). The international fisheries for horse mackerel and chub mackerel are considered in this section; that for snoek is dealt with briefly under by-catch species in the demersal-trawl fisheries (pp. 386– 387). The two horse mackerel species, Cunene horse mackerel Trachurus trecae and Cape horse mackerel T. capensis, have overlapping ranges, especially in ICSEAF Division 1.3, leading to difficulty in allocation of catches to species. The problem was most severe in the mid- and late 1970s when yields of both species from Division 1.3 were high (De Villiers, 1985). In the Southeast Atlantic three separate stocks of Cape horse mackerel are recognized, a northern stock inhabiting Divisions 1.3 and 1.4, a central stock in Divisions 1.5 and 1.6, and a southern stock in Divisions 2.1 and 2.2 (De Villiers, 1977; Draganik, 1977). For stock assessment purposes later workers (e.g. Babayan et al., 1985) have, however, lumped catches from Division 1.5 with those from Divisions 1.3 and 1.4. There is also still uncertainty as to whether the southern stock is distinct from that of Division 1.6 (Kinloch, Armstrong, Crawford & Leslie, 1986), although genetic studies have supported separation of stocks off Namibia and on the Agulhas Bank (Zenkin & Komarov, 1981). As catches from Division 1.5 have been consistently small, with a maximum of 14000 metric tons in 1980, it is likely that this Division forms a boundary between the stocks off northern Namibia and off the Western Cape, and it is therefore not of much

THE BENGUELA ECOSYSTEM: PART IV

327

consequence to which of these two stocks catches from Division 1.5 are allocated. This review assumes that catches from Divisions 1.1 through 1.5 represent the northern (Namibian) stock, from 1.6 the central (Western Cape) stock, and from 2.1 and 2.2 the southern (Agulhas) stock. As mentioned on page 360 substantial catches of Cape horse mackerel were made in Division 1.6 by the Western Cape purse-seine fleet throughout the 1950s (Table III), based largely on strong year-classes formed during the late 1940s (Geldenhuys, 1973). The fisheries off Namibia and on the Agulhas Bank commenced in the 1960s and assumed increasing importance during the 1970s and early 1980s, by which time only small yields were being taken from the central stock (Table II). From 1981 to 1984, 71 to 83% of the overall catch in the Southeast Atlantic was made with midwater trawls, 8 to 15% with bottom trawls and 1 to 17% with purse-seines. For trawl nets the minimum mesh size allowed for the capture of horse mackerel in the ICSEAF area is 60 mm (ICSEAF, 1984). The minimum size used on the Agulhas Bank by South African boats is, however, 75 mm (M.Kinloch, pers. comm.). Catches of horse mackerel from the Agulhas Bank rose from approximately 5000 tons in the mid 1960s to a peak of 93000 tons in 1977 (Tables II and VIII), the fishery using both midwater and bottom trawl (De Villiers, 1985). On 1 November 1977 South Africa declared an exclusive fisheries zone of 200 nautical miles (370 km) and introduced a quota system. As a result catches subsequently stabilized at a fairly constant level of between 29000 and 46000 tons (Kinloch et al., 1986). Effort increased by a factor of nine between 1968 and 1977, and then fell substantially (Table VIII). Catches from the northern stock climbed from under 50000 tons in the early 1960s to over 600000 tons from 1982 to 1984 (Tables II and IX). Two time series of catch rates, those for Polish and Soviet vessels, are available (Babayan et al., 1985). Both suggest a large increase in effort during the early 1970s. Catches in discrete ICSEAF Divisions are only available from 1973. Between 1973 and 1977, Division 1.4 contributed the largest catch on four occasions. In this period almost exactly half the catch came from Divisions TABLE VIII Catch and effort information for the fishery on Cape horse mackerel Trachurus capensis on the Agulhas Bank, 1968– 1983 (from Kinloch et al., 1986) Year

Catch (thousands of metric Standardized catch rates tons)

South African boats* Japanese boats

South African boats†

1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980

0·383 0·579 0·389 0·534 0·248 0·273 0·389 0·413

8·8 10·8 14·2 27·0 19·6 26·6 32·5 52·9 39·5 93·1 34·0 45·5 36·3

0·413 0·461 0·461 0·291 0·261 0·310

Effort (standardized South African units) 23005 18686 36373 50616 78948 97267 83244 128019 85672 202022 116856 174364 117194

328

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Year

Catch (thousands of metric Standardized catch rates tons)

South African boats* Japanese boats 1981 1982 1983

Effort (standardized South African units)

South African boats†

29·9 30·2 35·5

0·366 0·286

0·488 66792 0·549 55078 0·573 62003

* Not targeting for horse mackerel. † Targeting for horse mackerel. TABLE IX Catch and effort information for the fishery on Cape horse mackerel Trachurus capensis off Namibia, 1969–1984 (updated from Babayan et al., 1985) Year

Catch (’000metric tons)

1969 46·5 1970 51·4 1971 214·5 1972 72·4 1973 262·6 1974 185·0 1975 269·5 1976 512·7 1977 381·0 1978 548·2 1979 416·4 1980 546·2 1981 590·4 1982 660·0 1983 600·5 1984 606·9

Catch rate (Polish units)

42·5 48·3 27·8 37·2 37·4 30·2 39·8 40·6 42·8 46·5 47·1 44·3 43·0

Catch rate (USSR units) 6·95 7·91 9·67 6·94 8·06 8·48 5·69 4·99 4·63 2·94 2·81 2·89 7·25 4·74 5·98

Effort (Polish units) Effort (USSR units)

1704 5437 6655 7245 13708 12616 13774 10256 12762 12697 14013 13555 14114

6.69 6.50 22.18 10.43 32.58 21.82 47.36 102·74 82·29 186·46 148·19 189·00 81·43 139·24 100·42

1.1, 1.2, and 1.3 and half from Divisions 1.4 and 1.5. From 1978 to 1984 the respective percentages were 71 and 29, indicating a northward shift in distribution of catches. Overall catches of Cunene horse mackerel rose in fluctuating fashion from a mean of slightly under 100000 metric tons between 1956 and 1965 to 281000 tons in 1972. Yields then decreased following cessation of Portuguese rule in Angola, but a record catch of at least 380000 tons was recorded in 1978 (De Villiers, 1985). Thereafter catches plummeted and were only 55000 tons in 1984. Catches until 1972 were made almost entirely by purse-seine vessels (De Campos Rosado, 1974a). In the early 1980s, 47 to 76% of the catches were made with midwater trawl, 5 to 18% with bottom trawl and 15 to 47% with purse-seine. The change can be attributed to a shift from a coastal-state fishery to one dominated by distant-water fleets (De Villiers, 1985). In the late 1950s the catch of Cunene horse mackerel was evenly spread between ICSEAF Divisions 1.2 and 1.3, but in the 1960s and 1970s largely centred in Division 1.3 (Table II). During the late 1970s and

THE BENGUELA ECOSYSTEM: PART IV

329

early 1980s increased catches were recorded in the two northern ICSEAF Divisions (1.1 and 1.2). There was, therefore, a southward shift in the distribution of catches followed by a return to the north over a period of approximately 25 years. In the northern Benguela catches of Cape horse mackerel also displayed a northward shift in distribution during the late 1970s and early 1980s. In 1984, when warm equatorial water pushed south along the Namib coast (Boyd & Thomas, 1984), there was a tendency for catches of both species of horse mackerel to revert to the southern areas (Table II). All chub mackerel caught in the Southeast Atlantic are thought to represent a single stock (Crawford & De Villiers, 1984). Until 1975 most of the catches were made by the South African purse-seine fleet off the Western Cape (Table X). Thereafter purse-seining decreased in importance, and the bulk of the catch was made by long-range midwater or bottom trawlers (Crawford & De Villiers, 1984). In this latter period midwater trawl was responsible for more than half the catch in all years, except 1980 when 80% was taken by bottom trawl (Table X), possibly as a result of the powerful 1977 year-class occupying greater depths in the water column at this time (Crawford & De Villiers, 1984). Strong year-classes were responsible for mean annual catches of 104000 metric tons between 1967 and 1969 and 180000 tons from 1977 to 1978 (Crawford & De Villiers, 1984). From 1954 to 1966 the catch averaged 29000 tons, from 1970 to 1976 60000 tons, and from 1979 to 1984 44000 tons. Concomitant with the change in importance of gear types for the chubmackerel fishery there was a marked shift in the geographical area of catches. Over 90% of the catch was made in the southern Benguela (ICSEAF Divisions 1.5, 1.6, 2.1, and 2.2) from 1971 to 1975 (Table X). Subsequently the bulk of the catch was made in the nothern Benguela (Divisions 1.1, 1.2, 1.3, and 1.4), except in 1980 when half the small catch of 16200 tons came from Division 1.5. Although this trend probably resulted in large measure from increased fishing in the northern region by distant-water trawlers, the greatly decreased catches from the southern Benguela are noteworthy as aerial spotting for the Western Cape’s purse-seine fleet, used often for location of chub mackerel, continued until 1980. TABLE X Recorded catches (thousands of metric tons) of chub mackerel Scomber japonicus from the southeastern Atlantic, 1954– 1984, showing percentage of overall catch contributed by the northern (ICSEAF Divisions 1.1, 1.2, 1.3, and 1.4) and southern (Divisions 1.5, 1.6, 2.1, and 2.2) regions of the Benguela and by different types of gear (updated from Crawford & De Villiers, 1984) Percentage of overall catch contributed by Year

Catch Northern Benguela

Southern Benguela

Purse seine Midwater trawl

Bottom trawl Pound nets Hook and line

1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964

4·0 20·2 32·6 7·4 21·6 33·1 31·0 49·7 20·4 13·2 50·0

100·0 100·0 100·0 100·0 100·0 100·0 100·0 100·0 100·0 100·0 100·0

100·0 100·0 100·0 100·0 100·0 100·0 100·0 100·0 100·0 100·0 100·0

0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0

0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0

0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0

0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0

0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0

330

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Percentage of overall catch contributed by Year

Catch Northern Benguela

Southern Benguela

Purse seine Midwater trawl

Bottom trawl Pound nets Hook and line

1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984

41·4 0·0 53·4 0·0 128·2 0·0 91·0 0·0 91·7 0·0 77·9 0·0 57·5 3·8 60·9 6·2 64·8 7·1 33·2 4·2 74·4 5·3 51·7 78·6 157·8 80·4 201·5 98·2 33·5 65·4 16·2 26·0 61·3 78·9 51·1 66·7 64·5 81·2 36·6 75·1

100·0 100·0 100·0 100·0 100·0 100·0 96·2 93·8 92·9 95·8 94·7 21·4 19·6 1·8 34·6 74·0 21·1 33·3 18·8 24·9

100·0 100·0 100·0 100·0 100·0 100·0 94·4 94·4 90·9 92·3 93·2 1·3 13·6 1·2 9·9 1·1 0·7 5·4 5·9 2·1

0·0 0·0 0·0 0·0 0·0 0·0 5·0 4·2 6·2 4·7 1·9 26·0 6·3 4·0 35·0 80·5 35·3 44·1 28·7 36·1

0·0 0·0 0·0 0·0 0·0 0·0 0·6 0·5 2·7 3·1 4·9 72·7 80·0 94·8 55·1 18·5 64·0 50·5 65·4 60·7

0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·9 0·2 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0

0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 1·1

DEMERSAL-TRAWL FISHERIES Hakes Merluccius spp. contribute the bulk of the groundfish catch from the Southeast Atlantic (Newman, 1977). Soles Austroglossus spp. are a prized table food and have been sought off South Africa since the turn of the century. They are caught in shallower water than the hakes. A variety of other species contribute a bycatch to the hake fisheries, some on occasion becoming targets in their own right. Hake fisheries Three species of hake are caught in the Southeast Atlantic. The Benguela hake Merluccius polli occurs in Angolan waters and overlaps the northern

THE BENGUELA ECOSYSTEM: PART IV

331

TABLE XI Recorded catches (thousands of metric tons) of Cape hakes, Merluccius capensis and M. paradoxus, with approximate allocation to the different ICSEAF Divisions, 1955–1984 (updated from Newman, 1977; Babayan et al., 1985; ICSEAF, 1985, 1986; Leslie, 1985) ICSEAF Division Year

1.1

1.2

1.3

1.4

1.5

1.6

2.1

2.2

All

1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984

– – – – – – – – – – – – – – – – – – 0 – 1 0 0 – 0 – – – 0 –

– – – – – – – – – – – – – – – – – – 5 – 4 1 1 0 0 – – – – –

– – – – – – – – – 1 18 88 95 161 164 187 171 236 122 140 154 61 84 100 69 33 32 64 91 94

– – – – – – – – – 1 75 124 100 221 157 215 195 370 255 179 155 309 193 158 103 58 60 112 125 134

– – – – – – – – – 46 100 122 199 248 206 225 230 214 290 196 179 212 155 125 140 75 121 130 123 142

115 118 126 131 146 160 149 148 170 162 203 195 177 144 165 143 202 244 158 123 90 144 102 101 93 102 101 86 74 83

– – – – – – – – – 5 3 20 17 31 41 27 33 50 76 100 68 54 38 36 50 44 32 43 36 38

– – – – – – – – – – – – – – 1 1 2 1 1 1 6 4 2 3 4 4 3 4 5 5

115 118 126 131 146 160 149 148 170 215 399 549 588 805 734 798 833 1115 907 739 657 785 575 523 459 316 349 439 454 496

distribution of the Cape hakes M. capensis and M. paradoxus, which are found from Angola southwards around the African coast (Botha, 1980). M. paradoxus occurs in deeper water than M. capensis, although the two species co-occur at intermediate depths (Botha, 1980). The hakes are fished mainly with demersal trawl during daylight (Konchina, 1986), although off the bottom at night they may be caught as by-catch in the midwater-trawl fishery (Chlapowski, Draganik & Wyszynski, 1982).

332

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

The South African fishery for hakes has existed since the turn of the century (Jones & Van Eck, 1967), catches being of the order of 1000 metric tons per annum between 1917 and 1927 before increasing to levels of 50000 tons by 1950 (Chalmers, 1976). In the early 1960s distant-water trawlers started to arrive on the hake grounds, about 80 such trawlers operating off Namibia in 1965 (Botha, 1970) when the catch of Cape hakes was about 400000 tons (Table XI). Overall catches of Cape hakes climbed from 115000 metric tons TABLE XII Recorded catches of Benguela hake Merluccius polli (metric tons) from the different ICSEAF Divisions and of all hakes Merluccius spp. (thousands of metric tons) in the entire ICSEAF Subarea, 1973–1984 (from ICSEAF Statistical Bulletins and sources listed in Table XI) M. polli (tons)

M. capensis and M. paradoxus (’000 tons)

All Merluccius spp. (’000 tons)

ICSEAF Division

ICSEAF Division All

ICSEAF Division All

907 739 657 785 575 523 459 316 349 439 454 496

907 750 657 786 600 583 475 332 367 459 460 496

Year

1.1

1.2

1.3

1.4

All

1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 % of species catch 1973– 1984

– – 180 – 706 8047 3624 1226 2825 760 95 –

16 – 60 726 21691 12893 2542 2099 2513 1282 46 –

– 11163 156 – 2161 39245 9402 12301 12377 17818 5721 247

– – – – – – – – 11 – – –

16 11163 396 726 24558 60185 15568 15626 17726 19860 5862 247

10·2

25·5

64·3

0·0

in 1955 to a peak of 1115000 tons in 1972 and then decreased to 316000 tons in 1980, before recovering to a level of about 500000 tons in 1984 (Table XI). Yields of Benguela hake are considerably smaller, having peaked at 60000 tons in 1978 (Table XII). During the 12-year period 1973–1984 total annual catches of hakes in the Southeast Atlantic averaged 573000 tons (Table XII). Prior to 1974 some countries, notably South Africa, discarded small hake so that recorded catches in the earlier years under-estimate the actual yield (Jones & Van Eck, 1967; Newman, 1977). The Cape hakes contribute approximately 70–75% of the landings of demersal trawlers operating out of South African and Namibian ports (Payne, 1977; Botha, 1980, 1985). Off the Western Cape between 1955 and 1974 South African trawlers mostly fished at depths of 230 to 454 m (Fig. 9). At these depths in research trawls made between February 1972 and January 1973 catch rates of Cape hakes were greatest and the relative contribution of Cape hakes to catches of all species was highest—about 85% (Botha, 1985). Botha (1985) estimated that M. capensis contributed 4·8% by number and 12·9% by mass of commercial catches of Cape hakes and, in view of both the minimal by-catch of other species and the preponderance of

THE BENGUELA ECOSYSTEM: PART IV

333

Fig. 9.—Frequency distribution by depth of commercial bottom trawls off South Africa’s Western Cape for selected years combined, 1955–1974 (from Botha, 1985).

M. paradoxus in hake catches, concluded that the fishery in ICSEAF Division 1.6 could be regarded as a single-species fishery. M. capensis is the dominant hake off northern Namibia (Macpherson, Mombeck & Schülein, 1982) and on the Agulhas Bank (Hatanaka et al., 1983). For fisheries based on hake a minimum mesh size of 110 mm was adopted by ICSEAF in 1975 following assessments of the effect of mesh sizes on catches of hake (Ikeda, 1974; Newman, 1974). Earlier, mesh sizes of 90 to 102 mm had mostly been used (Newman, 1977). For mixed-species trawl fisheries a smaller

334

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

mesh is allowed, provided the percentage by mass of hake in each haul does not exceed 20% (ICSEAF, 1984). In commercial catches no distinction has been made between the two species of Cape hakes, identification not being easy. Therefore for assessment purposes the two species are considered together (De Villiers, 1985). Stock separation is not well understood but, from consideration of trends in catch and effort and geographical patterns of deployment of effort, four stocks of Cape hakes have been recognized by ICSEAF for assessment purposes: those in Divisions 1.3+1.4, in Division 1.5, in Division 1.6 and in Divisions 2.1+2.2 (Newman, 1977). Because of the international character of the fishery for Cape hakes and the large number of trawlers involved, no convenient reference exists on the size and nature of the fishing fleet (Newman, 1977). Information concerning catch rates is, however, available for certain sectors of the fishery and these have been used to provide time series of effort information for the various stocks. Total allowable catches (TACs) for each participating member country based mainly on these time series of catch and effort information are recommended annually by ICSEAF. In Divisions 1.3 and 1.4 the estimated level of fishing effort increased more than sixteenfold between 1965 and 1976. In 1976, catch rates were 24% of the 1965 level (Table XIII). They remained low through the late 1970s but improved in the early 1980s, by which stage effort had decreased to less than half its peak value. The improved catch rates are thought to have resulted from enhanced recruitment (ICSEAF, 1986). Catches peaked at 378000 tons in 1973, and the TAC recommended for 1986 was 318000 tons (ICSEAF, 1986). The stock in Division 1.5 suffered a similar decrease in abundance between 1965 and 1976, catch rates falling by 74% in this period while effort increased eightfold to its maximum level (Table XIII). Again catch rates remained low until the 1980s, when they showed signs of increasing. The highest catch of 290000 tons was taken in 1973, and the TAC recommended for 1986 was 163000 tons (ICSEAF, 1986). In Division 1.6 catch rates decreased by 47% between 1940–1947 and 1966 (Jones & Van Eck, 1967), and continued to decrease thereafter. The lowest levels, recorded in 1974 and 1975, suggested a hake density about 27% of that during 1955 (Table XIII). Effort peaked in 1972 but has subsequently decreased, concomitant with improved catch rates. The highest catch of 244000 tons was recorded 1972. South Africa declared a fishing zone of 200 nautical miles in 1977 enabling it to improve control of exploitation. For Divisions 2.1 and 2.2 catch rates fell by 71 % between 1967, when the resource was in a near-virgin state, and 1975 and remained low until 1981 before recovering slightly (Table XIII). The largest catch of 101000 tons was taken in 1974, and for Divisions 1.6, 2.1 and 2.2 combined a TAC of between 133000 and 145000 tons was suggested for 1986 (ICSEAF, 1986). By-catch species A number of species contribute a by-catch to the demersal fisheries for hakes and horse mackerel, among the more valuable of which is kingklip Genypterus capensis (Botha, 1970; Newman, 1977; De Villiers, 1985). Catches of this species made prior to 1964 were small and made primarily by South African

THE BENGUELA ECOSYSTEM: PART IV

335

TABLE XIII Catch (thousands of metric tons) and effort (relative to earliest year available for the different stocks) information for the fisheries for Cape hakes in different ICSEAF Divisions, 1955–1985 (from ICSEAF, 1986) Divisions 1.3 and 1.4 Year 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985

Division 1.5

Division 1.6

Divisions 2.1 and 2.2

Catch Catch per Effort Catch Catch per Effort Catch Catch per Effort Catch Catch per Effort unit effort unit effort unit effort unit effort 118·2 126·4 130·7 146·0 159·9 148·7 147·6 169·5 162·3 93·5 212·4 195·0 382·7 320·5 402·5 365·6 606·1 377·6 318·8 309·4 369·8 277·5 258·1 172·3 90·5 92·1 176·4 215·8 228·5 166·8

90 95 94 94 100 70 82 81 84 100 74 51 54 49 51 49 40 32 25 24 24 28 24 22 25 31 30 33 36 38

113 115 121 135 139 184 156 182 167 100 306 404 753 687 843 792 1589 1249 1336 1391 1660 1068 1132 813 379 315 628 702 674 470

99·7 122·2 199·4 247·7 206·2 224·7 229·7 214·0 290·3 195·7 178·7 211·9 154·5 125·1 140·1 74·6 120·6 130·1 123·3 141·8 201·2

100 117 66 62 51 49 64 45 45 31 37 26 31 25 33 32 38 38 40 41 45

100 105 305 403 403 459 358 481 652 628 490 821 503 502 425 236 319 348 308 346 342

115·4 100

100

203·3 195·0 176·7 143·6 165·1 142·5 202·0 243·9 157·8 123·0 89·6 143·9 102·3 101·1 92·7 101·5 100·7 86·0 73·7 82·8 60·3

281 275 265 215 287 296 427 747 476 397 288 403 317 257 227 277 260 220 169 185 135

62 62 58 58 50 42 41 28 29 27 27 31 28 34 35 32 34 34 38 39 39

17·3 31·4 41·7 27·8 34·5 51·4 77·4 100·9 73·8 57·7 40·5 38·9 53·8 47·6 35·1 46·8 41·2 43·2 44·0

100 100 100 95 89 50 44 42 29 31 33 32 36 34 31 40 38 43 43

100 181 241 168 223 593 1020 1379 1473 1065 711 700 1307 798 649 678 633 578 591

336

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

TABLE XIV Recorded catches (metric tons) of kingklip Genypterus capensis from the different ICSEAF Divisions, 1965–1984 (from Newman, 1977; ICSEAF Statistical Bulletins) ICSEAF Division Year

1.1

1.2

1.3

1.4

1.5

1.6

2.1

2.2

Not known

All

1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 % of 1976– 1984

– – – – – – – – – – – 295 45 141 1818 764 114 1503 1325 586 6·9

– – – – – – – – – – – 390 88 482 605 456 45 649 2 1 2·8

– – – – – – – – – – – 211 134 242 122 89 139 128 57 97 1·3

– – – – – – – – – – – 5384 2941 1749 1076 1454 2392 741 741 304 17·4

– – – – – – – – – – – 3447 5211 4652 3301 2172 3257 1460 2874 2557 30·0

– – – – – – – – – – – 2084 2246 1835 1979 3125 2458 1711 2770 5014 24·1

– – – – – – – – – – – 2103 916 1598 1614 1404 1125 687 807 879 11·6

– – – – – – – – – – – 530 317 827 802 879 453 553 713 651 5·9

2500 4100 4900 3300 2900 3900 5000 16300 21700 12900 10600 – 100 – – – 45 32 – –

2500 4100 4900 3300 2900 3900 5000 16300 21700 12900 10600 14444 11998 11526 11317 10343 10028 7464 9289 10089

vessels, but following arrival in the Southeast Atlantic of large numbers of distant-water trawlers exploitation increased (Newman, 1977). Recorded landings were between 2500 and 5000 tons from 1965 to 1971, but climbed to almost 22000 tons in 1973 (Table XIV). It is uncertain whether the lower values until 1971 resulted from low abundance or under-reporting by some nations because of misidentification (Terre, 1980), although the latter is thought likely (Newman, 1977; Isarev, 1983). From 1974 to 1984 catches were of the order of 10000 tons per annum. On the basis of differences in growth rate, number of vertebrae, otolith form and relationships between otolith dimensions and fish length three separate geographical stocks of kingklip are believed to exist (Isarev, 1976; Payne, 1977, 1985b). The three stocks recognized are those occurring north of Walvis Bay (23° S), between 27 and 35° S, and at the eastern edge of the Agulhas Bank. A fourth stock may exist at the western edge of the Agulhas Bank, but this is uncertain (Payne, 1985b). ICSEAF Divisions 1.5 and 1.6 contributed 54% of the catch for 1976–1984 and Divisions 1.1 through 1.4, probably most of the northern stock, 28%. The Agulhas Bank accounted for 18% of the overall catch in this period (Table XIV).

THE BENGUELA ECOSYSTEM: PART IV

337

TABLE XV Recorded catches (metric tons) of monkfish (anglerfish) Lophius spp. from the different ICSEAF Divisions, 1967–1984 (from ICSEAF Statistical Bulletins) ICSEAF Divisions Year 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 % of 1974– 1984

1.1

1.2

1.3

1.4

1.5

1.6

2.1

2.2

Not known

All 200

200 400 300 – 980 4851 – – – – 115 – – – – – – 0·1

– – 1 5 147 25 – – – – – 0·2

24 12 63 302 3092 500 493 1155 2182 2036 2271 11·5

324 510 477 3157 3127 1801 1638 8237 7500 5543 2188 32·8

– 601 349 2238 1140 1234 1072 5615 5524 4338 3392 24·2

– 72 9 3931 3070 3016 3637 3245 2759 2537 3060 24·1

– – 12 804 849 62 971 968 1056 981 1034 6·4

– – – 88 87 – 74 56 113 130 89 0·6

3716 4210 5114 209 6 30 305 20 15 – –

4064 5405 6025 10734 11633 6668 8190 19296 19149 15565 12034

Yields of monkfish or anglerfish Lophius spp. peaked at 19000 tons in 1981 and 1982. The largest catches have been made in Divisions 1.4, 1.5, and 1.6 (Table XV). There have been difficulties in identifying species in catches (Lloris & Rucabado, 1985). Limited attempts at directed fishing have been made (De Villiers, 1985). Panga (seabream) Pterogymnus laniarius has been fished on the Agulhas Bank, mainly by Japanese trawlers, which have targeted for this species and horse mackerel. Catches have also been made by South Africa, but in earlier years a large proportion of the catch by this country is reported to have been discarded (Newman, 1977). Catches were 18000 tons in 1966 but decreased to less than 2000 tons in the early 1980s (Table XVI). Catch rates of Japanese trawlers fell by 81% between 1964 and 1975 (Newman, 1977). Catches of large-eye dentex Dentex macrophthalmus were low prior to 1965 (Newman, 1977), but from 1965 to 1968 averaged 46000 metric tons (Table XVII). At this time a directed fishery was pursued by the U.S.S.R. and the species was also taken as a by-catch in the hake fisheries. Catches decreased after 1968, reportedly as a result of both reduction in the specialized Soviet fishery and a decrease in stock size (Newman, 1977). There was a resurgence in harvest during the late 1970s, but yields again fell to low levels in the 1980s. From 1972 to 1975 large catches were made in ICSEAF Divisions 1.3 and 1.4 (Newman, 1977), from 1976 to 1978 the highest harvests were made in

338

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

TABLE XVI Recorded catches (metric tons) of panga Pterogymnus laniarius from the different ICSEAF Divisions, 1965–1984 (from Newman, 1977; ICSEAF Statistical Bulletins) ICSEAF Divisions Year

1.1

1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 % of 1973– 1984

18200 9800 7300 8600 7300 8600 7300 – – – – – – – – – – – –

1.2

1.3

1.4

1.5

1.6

2.1

2.2

Not known

All 11500

– – – 1 – – – – – – – – 0·0

– 9 – 27 2 2 – – – – 2 – 0·1

– 25 – 37 16 15 – – – – 1 – 0·2

– – – 13 19 24 – – – 56 – – 0·2

– 5 5 115 436 74 3 2 2 – 10 78 1·3

5048 8429 7134 3563 4565 2380 1520 1190 1160 1336 1035 1116 67·5

2892 2104 4164 923 1492 1834 1500 406 423 394 682 744 30·8

850 622 1232 – 179 – 31 16 – – – –

8790 11194 12535 4679 6709 4329 3054 1614 1585 1786 1730 1938

Division 1.2, and in later years yields were frequently best from Division 1.1 (Table XVII). Therefore, as with many species off Angola and Namibia, there was a northward shift in the distribution of catches at the end of the 1970s. Catches of Angolan dentex D. angolensis have been much smaller and confined to the two northernmost ICSEAF Divisions (Table XVII). As a result of a ready market, South African demersal trawlers developed a semi-directed fishery for snoek off the Western Cape during the 1960s. From the early 1970s this fishery was, however, discouraged by the South African government, it being considered that snoek were the traditional catch of line fishermen. The contribution of snoek to catches by demersal vessels was closely monitored, and from the mid 1970s to 1980 was not allowed to exceed 5–10% of the catch of hakes (Table XVIII). Where it could be proved that snoek-directed trawling had been conducted convictions in a court of law were obtained (Botha, 1977b). For this reason catches of snoek by South African demersal trawlers have been relatively small, fluctuating between about 1000 and 10000 metric tons during the period 1972–1984 (Table XVIII). All snoek in the Southeast Atlantic are thought to belong to a single stock

THE BENGUELA ECOSYSTEM: PART IV

339

TABLE XVII Recorded catches (metric tons) of Dentex spp. by all nations in the different ICSEAF Divisions, 1965–1984 (from ICSEAF, 1985; Kuderskaya, 1985; ICSEAF Statistical Bulletins) D. angolensis ICSEAF Division

D. macrophthalmus ICSEAF Division

Year

1.1

1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 % of species catch 1972– 1984

51400 35100 55400 18300 18100 19400 – 9 – – – – – – 602 – – 1896 – 24·4

1.2

All

1.1

1.2

1.3

1.4

1.5

1.6

Other All 42400

– – – – 6546 963 230 – – – – 26 – 75·6

– 9 – – 6546 963 230 – 602 – – 1922 –

– 28 800 133 671 1840 5384 5250 4305 2335 5812 4425 2885 15·2

34 4 600 – 35621 13542 12828 3588 3220 1614 2333 1681 887 34·2

10593 5315 4332 7542 5313 10786 9897 11377 1051 4329 5772 3565 2639 37·1

6264 4695 1826 678 2978 658 836 739 60 92 228 117 361 8·8

1599 1423 368 256 241 72 56 44 1 2 1 – 7 1·8

810 189 66 – 55 326 – – – – – – – 0·7

– – 234 265 – 3573 297 436 20 – – – – 2·2

19300 11654 8226 8874 44879 30797 29298 21434 8657 8372 14146 9788 6779

(Crawford & De Villiers, 1985). In 1978 and subsequently large catches of the species were made by distant-water fleets fishing with demersal and midwater trawls off the Namib coast. Recorded catches from the Southeast Atlantic averaged 47000 tons between 1978 and 1984, with a peak of more than 80000 tons in 1978 (Table XIX). Between 1972 and 1984 two thirds of the catch was made in ICSEAF Divisions 1.3 and 1.4. In 1981 and 1982 more than 95% of the overall catch of snoek in the Southeast Atlantic was caught by trawl (Crawford, 1985). e Nominal (live-weight) landings of some species by the South African bottom-trawl fishery since 1917 ar listed in Table XX. It is apparent that, in addition to Cape hakes, kingklip, monkfish, panga, snoek, and soles, many other species including the white squid Loligo reynaudii and crustaceans, mainly shrimps caught off Natal, contribute to the fishery. Estimated landings of all species rose from of the order of 3000 metric tons in 1917 to a 1984 value of 182000 tons.

340

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

TABLE XVIII Recorded catches (metric tons) of snoek Thyrsites atun by South African demersal trawlers operating from South African ports and catch of snoek by these trawlers expressed as a percentage of their catch of hakes Merluccius spp., 1972–1984 (from records of the Sea Fisheries Research Institute) Year

Catch

Snoek catch as % of hake catch

1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984

1060 998 1325 1812 1979 9061 8052 7104 10503 6657 5388 3658 5443

1·3 1·1 1·5 2·5 1·8 9·2 5·7 4·7 7·3 4·9 3·7 2·9 4·1

Sole fisheries Two species of sole (Soleidae) are exploited off southern Africa: the Agulhas sole Austroglossus pectoralis and the west coast sole A. microlepis. Fishing grounds for A. pectoralis are located on the inshore reaches of the south and east coasts of South Africa at depths of less than 100 m (Botha, 1978); those of A. microlepis along the west coast between the Cunene River and Port Nolloth and within the 320 m isobath (Payne, 1985a; Fig. 10). Since 1979 large catches of A. microlepis have also been reported from Angolan waters, especially ICSEAF Division 1.1 into which the Zaire River discharges (Table XXI). For A. microlepis south of the Cunene River two distinct stocks, northern (Skeleton Coast) and southern (Orange River), are recognized (Payne, 1979, 1985a). Overall catches of A. microlepis from the Southeast Atlantic peaked at 3826 tons in 1982. Between 1973 and 1984 45% of the catch of this species was made in ICSEAF Division 1.4, 24% in Division 1.3 and 26% in Division 1.1. Of the catch of A. pectoralis 91% came from Division 2.1. The combined catch of soles reached 4643 tons in 1982, with Division 2.1 contributing 34% of the catch between 1973 and 1984, Division 1.4 29%, Division 1.3 16% and Division 1.1 14% (Table XXI). Agulhas sole have been exploited since the turn of the century, the Orange River stock of west coast sole since at least the 1930s, and the Skeleton Coast stock since 1958 (Marchand, 1933; Zoutendyk, 1973; Lucks, Payne & Maree, 1973; Chalmers, 1976; Botha, 1978; Payne, 1985a). Landings by boats operating from South African or Namibian ports for the period 1921–1984 are listed in Table XXII. These figures include landings of a sand sole Cynoglossus zanzibarensis (Cynoglossidae) which have been made by all three fisheries, particularly that for Agulhas sole (Payne, 1979), but are regarded as insig

THE BENGUELA ECOSYSTEM: PART IV

341

TABLE XIX Recorded catches (metric tons) of snoek Thyrsites atun by all nations from the different ICSEAF Divisions, 1960–1984 (from Crawford & De Villiers, 1985; ICSEAF Statistical Bulletins) ICSEAF Division Year

1.1

1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 % of 1972– 1984

18400 18300 17800 7200 12900 12300 11800 12000 8300 5400 10000 – 418 226 – – – 31 – – – 1 – – 0

1.2

1.3

1.4

1.5

1.6

2.1

2.2 Other or not known

All 17700

– 141 28 – – – 454 18 – – – – – 0

– – – – – 415 34589 10865 19084 5587 17833 49317 18310 41

1360 – 1566 1270 3221 3270 31403 7660 15388 4484 4033 11803 9115 25

– 2 – 63 145 5690 3021 1154 2580 3032 3925 2433 2495 7

2606 4976 8211 6150 12602 9156 10503 6400 6256 3019 1387 4817 5963 22

– 24 54 6 458 808 1668 3470 4557 2476 3145 1873 1805 5

– 32 4 2 41 3 15 16 – 1 – 3 1 0

1496 2965 – 2537 – 142 4 1936 8373 2498 1682 864 –

5462 8558 10089 10028 16467 19484 81688 31519 56238 21097 32006 71110 37689

nificant (Payne, 1985a). Since 1921 landings of Agulhas sole have been relatively stable, varying between 428 and 1192 metric tons. Conversely, landings of the northern stock of west coast sole have fluctuated widely, rising from a level of about 150 tons during 1963–1966 to 1895 tons in 1973. Landings decreased to under 200 tons by 1976, but climbed to 1940 tons by 1982 before again decreasing rapidly. Landings of the southern stock of west coast sole decreased in a fluctuating manner from the peak of 1404 tons in 1966 to less than 10 tons in 1975, and have remained trivial subsequently. The fishery for Agulhas sole was centred at Mossel Bay in 1977, although vessels also operated out of Hermanus, Gans Bay, Port Elizabeth and East London (Botha, 1977a, 1978; Payne, 1985a). From the late 1970s smaller vessels were replaced by larger boats capable of trawling in deeper waters and also exploiting hake (Payne, 1985a). Trawlers directed an estimated

342

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

TABLE XX Nominal live mass (thousands of metric tons) of fish, cephalopods and crustaceans landed by the South African bottomtrawl fleet, 1917–1984: catches of soles are included under other species, but are specified separately in Table XXII; —, no data available; ×, combined with other species; (from Chalmers, 1976; Annual Reports of the Chief Directorate Marine Development) Year

Merluccius spp. (Cape hakes)

Genypterus capensis Pterogymnus (Kingklip) laniarius (Panga)

Brama brama (Angel-fish)

Chelidonichthys spp. (Gurnards)

1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951

0·7 0·8 1·4 – 0·9 0·7 1·8 1·1 1·4 1·0 0·6 1·9 2·7 3·2 2·0 10·3 8·0 9·9 10·8 12·7 14·5 15·2 14·4 20·6 22·0 24·8 27·3 24·5 21·0 29·1 29·8 42·3 41·3 51·8 64·4

– – – – – – – – – – – – – – – 0·6 0·4 0·4 0·7 0·7 0·7 0·7 0·7 0·7 0·6 0·6 0·6 0·6 1·3 1·0 1·1 1·5 1·8 1·9 2·4

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – 0·1 × × × × ×

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

THE BENGUELA ECOSYSTEM: PART IV

Year

Merluccius spp. (Cape hakes)

Genypterus capensis Pterogymnus (Kingklip) laniarius (Panga)

Brama brama (Angel-fish)

Chelidonichthys spp. (Gurnards)

1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1917

63·9 67·3 75·8 80·2 82·1 87·7 90·8 100·8 111·1 103·3 102·3 98·9 102·2 99·6 124·9 115·4 115·0 101·0 91·7 105·3 112·9 128·6 130·5 107·5 107·5 98·3 141·0 151·4 143·6 135·6 142·8 125·2 134·2 –

2·8 2·7 2·0 2·3 2·0 1·5 1·7 2·0 1·5 2·1 1·7 1·8 1·4 2·5 3·9 3·1 3·1 3·1 3·2 3·9 3·8 4·9 3·5 3·6 3·6 2·7 4·3 4·6 5·3 4·1 3·1 3·4 3·6

– – – – – – – – – – – – – – – – – – – – 0·1 0·1 0·2 0·2 0·3 0·8 0·6 0·3 0·3 0·2 0·3 0·3 0·1

× × × × 0·1 0·1 × 0·3 0·3 0·4 0·4 0·4 0·2 0·2 0·2 0·2 0·5 0·5 0·5 0·5 0·5 0·4 0·4 0·3 0·4 0·6 0·5 0·6 0·5 0·6 0·6 0·7 0·6

–

– – – – – – – – – – – – – – 2·4 1·9 1·7 1·6 1·3 1·2 1·3 0·8 0·6 1·2 1·3 2·0 2·7 2·2 0·3 0·3 0·5 0·4 0·7 –

–

–

–

343

344

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Year

Helicolenus spp. Zeus spp. (John (Jacopever) dory)

Trachurus capensis (Horse mackerel)

Thyrsites atun (Snoek)

Lophius spp. (Monk-fish)

Scomber japonicus (Chub mackerel)

1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

THE BENGUELA ECOSYSTEM: PART IV

345

Year

Helicolenus spp. Zeus spp. (John (Jacopever) dory)

Trachurus capensis (Horse mackerel)

Thyrsites atun (Snoek)

Lophius spp. (Monk-fish)

Scomber japonicus (Chub mackerel)

1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984

– – – – 0·9 1·0 0·9 1·5 1·3 1·5 0·7 1·0 0·9 0·8 0·6 0·4 0·4 0·4 0·4 0·5 0·5 0·7 0·7 0·5 0·4 0·8 0·4 0·6 0·7

– – – – – 3·7 3·6 1·4 1·9 2·3 3·1 3·4 3·4 5·9 7·2 7·8 4·9 7·0 10·5 8·6 10·4 13·2 13·2 15·3 24·4 16·6 14·7 18·7 13·8

– – – – – – – – – – – – – – – – 1·5 1·4 1·9 2·6 2·6 9·1 9·1 8·1 6·8 10·5 5·4 3·7 5·5

– – – – – – – – – – – – – – – – 3·4 4·2 3·6 3·7 3·7 4·9 4·9 4·1 4·5 4·7 4·3 4·0 4·6

– – – – – – – – – – – – – – – – – – – – – – – – 7·6 4·0 5·7 5·1 4·4

– – – – 0·0 0·0 0·2 0·2 0·0 0·1 0·0 0·0 0·0 0·1 0·4 0·2 0·2 0·0 0·1 0·4 0·6 1·2 1·2 1·6 1·1 1·4 0·8 0·6 0·6

84000 hours at fishing for Agulhas sole in 1977, but this halved to levels around 40000 hours by the early 1980s (Table XXIII). Landings of Agulhas sole have been regulated since 1978 when the quota was set at 700 tons although the final catch exceeded this value. The quota increased to 930 tons by 1982 and remained at this level until 1984 (Badenhorst, 1985; Payne, 1985a). It was 950 tons in 1986 (Payne, pers. comm.). TABLE XXcontinued

346

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Year

Lepidopus caudatus (Silver scabbard-fish)

Loligo reynaudii (White squid)

Argyrosomus hololepidotus (Kob)

Crustaceans Other Total

1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

2·1 2·0 1·3 – 0·7 0·9 0·8 1·0 1·5 1·5 1·4 1·4 0·9 1·4 0·9 1·7 0·9 1·2 1·0 1·4 1.3 1·5 1·4 1·3 0·9 0·9 0·8 0·6 0·6 0·7 0·6 0·8 1·0 0·8 0·7 0·5 0·8 0·5 0·6

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

THE BENGUELA ECOSYSTEM: PART IV

347

Year

Lepidopus caudatus (Silver scabbard-fish)

Loligo reynaudii (White squid)

Argyrosomus hololepidotus (Kob)

Crustaceans Other Total

1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984

– – – – – – – – – – – – – – – – – – – – – – – – – 0·1 0·3 0·5 0·7

– – – – – 0·2 0·1 0·2 0·2 0·2 0·0 0·3 0·2 0·1 0·3 0·2 0·3 0·6 0·1 0·8 0·8 1·2 1·2 2·1 2·7 1·7 1·5 1·5 1·0

0·7 0·6 0·6 0·6 0·4 0·5 0·4 0·2 0·4 0·4 0·6 0·7 0·6 0·5 0·5 0·5 0·5 0·4 0·3 0·6 0·6 0·3 0·3 0·2 0·2 0·2 0·2 0·3 0·3

– – – – – – – – – – 0·2 0·6 0·6 0·5 0·5 0·5 0·6 0·7 1·2 1·2 1·9 1·1 1·1 1·2 1·4 2·0 1·9 1·1 0·9

– – – – – 2·4 1·7 3·2 2·4 3·6 29·5 34·9 29·1 24·9 23·4 20·7 15·8 20·0 21·9 15·2 15·6 15·4 16·3 12·1 12·0 10·8 13·0 14·3 10·6

– – – – – 113·5 111·3 107·8 110·0 110·4 165·5 161·5 155·1 139·0 129·6 141·2 146·2 169·5 175·2 146·4 149·8 151·5 197·1 204·3 211·1 193·6 195·5 175·4 182·3

From 1982 to 1984 the quota was not filled, partly because operators were subject to hake as well as sole quotas and the filling of one quota was limiting upon the other (Badenhorst, 1985). Catches of west coast sole have been unrestricted by quota. A minimum mesh size of 75 mm is used by all sole fisheries to permit adequate escape of soles less than 30 cm length, the approximate size at 50% maturity (Botha, Lucks & Chalmers, 1971; Lucks, 1972; Botha, 1978; Payne, 1985a). Certain trawlers concentrating on sole off Namibia have, however, been forced to use a net of minimum mesh size 110 mm in accordance with the ruling of the International Commission for the Southeast Atlantic Fisheries (ICSEAF) that where the catch of hake exceeds 20% by mass no smaller mesh may be used (Payne, 1985a).

348

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Fig. 10.—The fishing grounds for sole Austroglossus spp. off southern Africa, indicating locations of the different stocks (from Payne, 1985a).

Although soles contribute only a small part of the South African annual trawl-fish catch, they have the greatest value per unit mass (Botha, 1970), and landings can rarely satisfy the demand of a predominantly South African market (Payne, 1985a).

THE BENGUELA ECOSYSTEM: PART IV

349

TABLE XXI Recorded catches (metric tons) of sole Austroglossus spp. by all nations in the different ICSEAF Divisions, 1973–1984 (corrected from ICSEAF Statistical Bulletins) A. microlepis

A. pectoralis ICSEAF Division

Not known

Both species

ICSEAF Division

Year

1.1

1.2

1.3

1.4

1.5

1.6 All

1.6

2.1

2.2

All

All

1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 % of species catch 1973– 1984 % of sole catch 1973– 1984

– – – – 35 36 410 393 426 1259 926 1075 26

– – – – 181 17 56 183 148 – 16 4 3

382 279 – 83 83 322 216 356 643 1586 142 107 24

1513 663 2 169 260 675 1118 211 1097 968 685 663 45

133 43 – – 84 4 7 1 20 13 59 51 2

– – – – 6 – – – – – – – 0

2028 985 527 252 658 1054 1807 1144 2334 3826 1828 1900

– – – 150 160 – 1 1 – – – – 3

961 611 763 773 520 996 853 887 971 773 635 817 91

– – – 117 128 38 46 55 55 44 48 40 6

961 611 763 1040 808 1034 900 943 1026 817 683 857

14

2

16

29

1

0

1

34

2

– – 525 – 9 – – – – – – –

2989 1596 1290 1292 1466 2088 2707 2087 3360 4643 2511 2757

TABLE XXII Recorded catches (metric tons) of sole Austroglossus spp. by South African boats from three stocks, 1921–1985:+less than 0·5 metric tons; (from Marchand, 1933; Payne, 1979; Badenhorst, 1985; Badenhorst, pers. comm.) Year

West coast sole A. microlepis Agulhas sole A. pectoralis Southern stock A. microlepis and A. pectoralis combined

Northern stock Southern stock 1921 1922 1923 1924 1925 1926 1927 1928 1929

698 530 552 669 678 643 821 734 774

350

Year

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

West coast sole A. microlepis Agulhas sole A. pectoralis Southern stock A. microlepis and A. pectoralis combined

Northern stock Southern stock 1930 1931 1932 1933 1934 1935 1936 1937 1938 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970

734 783 676 660 808 913 1096 1039 1192 835 706 2996 2406 2216 1847 1643 1766 1430 1426 1358 1264 1373 1267 1139 1057 913 1106 2050 1050 100–150* 100–150* 100–150* 150 725 1062 953 1162

1066 608 954 1404 534 244 301 365

704 1077 808 553 595 428 620 640

THE BENGUELA ECOSYSTEM: PART IV

Year

351

West coast sole A. microlepis Agulhas sole A. pectoralis Southern stock A. microlepis and A. pectoralis combined

Northern stock Southern stock 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 * Estimated.

578 1379 1895 897 520 178 272 836 942 503 1255 1940 696 318 357

45 121 133 41 5 + 6 + + + + 1 + + 0

842 1004 963 588 733 1000 777 938 864 937 1020 824 692 863 880

TABLE XXIII Effort directed at Agulhas sole Austroglossus pectoralis and catch rates, 1973– 1983 (from Badenhorst, 1985) Year

Effort (thousands of hours)

Catch rate (kg·h−1)

1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983

73 44 32 49 84 79 49 46 41 39 42

13·2 13·9 23·7 21·4 9·6 13·1 18·2 20·3 25·3 20·8 16·2

HOOK AND LINE FISHERIES The most important fisheries with hook and line are those for kingklip and snoek along the western and southern seaboards of southern Africa, conducted mainly by the coastal states, and an international fishery for tunas. Many other species are, however, also caught.

352

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Kingklip fishery The recent development off South Africa of a line fishery targeting for kingklip has been described by Badenhorst (in prep.). The fishery commenced in 1983 when 1042 metric tons of kingklip and 230 tons of Cape hakes were caught. These values improved to 3040 and 381 tons, respectively, in 1984, and to 6970 and 1459 tons in 1985. In 1985 the line fishery for kingklip was considerably more important than the South African bottom-trawl fishery for the same species, which yielded under 4000 tons. The line fishery operates in rocky areas adjacent to the main bottom-trawling grounds, from west of Hondeklip Bay along the continental shelf and around the southern edge of the Agulhas Bank to Port Elizabeth. The best catches of kingklip have been made at depths of 300 to 450 m. Lines with lengths up to 15 km have been employed. On average lines have about 10000 hooks attached, each hook being individually baited. The lines are shot after midnight and hauled after first light, and are weighted to extend horizontally in the near-bottom zone (Badenhorst, in prep.). Snoek fishery The early history of the South African handline fishery for snoek has been summarized by Thompson (1913), Lees (1969), and Nepgen (1979b). Snoek were exploited off South Africa’s Western Cape from the 1600s (Thompson, 1913). By 1830 some 40 boats and 200 men were exclusively engaged in the fishery, and salt snoek was being exported, especially to Mauritius (Lees, 1969). This country remained a regular overseas market for more than a century, although war at times had its effect on trade. In 1881, 2000 metric tons of snoek were exported, the value being estimated at 33428 pounds sterling. In 1889 the fleet consisted of a total of 374 boats manned by 2241 fishermen, and the annual catch was between 3000 and 4500 tons (Lees, 1969). Estimates of the handline catch of snoek off South Africa’s Western Cape have been available at intermittent intervals since 1881, and off Namibia since 1924 (Table XXIV). It is evident that catches have fluctuated substantially, but the mean of catches off the Western Cape for the period 1969–1984 (3655 tons) is similar to means for the periods 1961–1962 (3090 tons) and 1921– 1938 (2958 tons) and the estimated harvest during the late 1880s (3000–4500 tons). Likewise the mean handline catch offNamibia during 1970– 1984 (1746 tons) compares well with estimates from the mid-1920s. On account of the dispersed nature of landing localities, and also the direct sale of snoek to vendors or the public, values shown in Table XXIV almost certainly under-estimate the actual handline catch of snoek. Landed mass has probably been further under-estimated by adoption of a conservative mass estimate of 2 kg per fish caught (Table XXIV). It is, however, apparent that handline catches off the Western Cape and off Namibia have averaged at least 5500 tons over the past 60 years, and probably for more than a century. The handline fishery off the Western Cape is almost entirely dependent on snoek, which in 1982 contributed 85% of the landed mass of the fishery and 80% of the overall value of its landings of about R4·1 million (Crawford & De Villiers, 1985). Approximately 2500 boats and 9000 fishermen were involved in handline fishing off the Western Cape in 1985 (A.J.Penney, pers. comm.). In South Africa, as well as Mauritius, snoek has been regarded as an important source of protein since at least the early 1800s and is used directly for human consumption (Thompson, 1913; Lees, 1969). It is sold both fresh and as a cured product, and during World War II and succeeding years was also canned (Lees, 1969). Subsequently, however, canning assumed a lower profile, primarily because the market for fresh or cured snoek was sufficient for the entire handline catch (Lees, 1969). The condition of snoek deteriorates markedly during the July-October spawning season (Nepgen, 1979b). On account of this snoek were not permitted to be caught with handlines off the Western Cape from August

THE BENGUELA ECOSYSTEM: PART IV

353

or October to November inclusive from the 1940s through 1980 (Crawford & De Villiers, 1985). Bycatches of snoek are, however, inevitable in the catches of demersal trawlers targeting for hake and, to obviate the inconsistency of allowing one sector of the fishery to exploit snoek when another may not, the closed season for handline fishermen was abolished from 1981. Multi-species fishery Many fish species are caught with handlines along the South African coast, estimated landings of the more important of which are listed in Table XXV for the period 1981 to 1984. The fishery for yellowtail, Seriola lalandi, is of particular interest. Purseseine fishing for the species commenced in 1971 and, in spite of considerable opposition from handline fishermen, who believed their livelihood was being TABLE XXIV Recorded catches (metric tons) by handline of snoek Thyrsites atun off South Africa’s Western Cape and off Namibia, 1881–1984: sources of information as follows; 1881, Lees (1969), based on exports and modified by the relationship, Catch=316·7+1·452 (Export) derived from information in Nepgen (1979) and Thompson (1913) for the period 1897– 1904 and 1906; 1889, Lees (1969); 1891–1896, as for 1881 except values for exports from Thompson (1913); 1897– 1904, Nepgen (1979b); 1905, as for 1891–1896; 1906, Nepgen (1979b); 1907–1910, as for 1891–1896; 1921–1933, Nepgen (1979b) for Western Cape and Lees (1969) for Namibia; 1935–1938 and 1961–1962, Nepgen (1979b); 1969, unpubl. records, Sea Fisheries Research Institute; 1970–1978, Nepgen (1979b) for Western Cape and Crawford (1985) for Namibia; 1979–1983, Crawford (1985) and Annual Reports of the Chief Directorate Marine Development; 1984, Annual Reports of the Chief Directorate Marine Development; in all instances when catches were reported in numbers a mass of 2 kg per snoek has been assumed, approximately equal to the mass of a fish of 700 mm fork length (Nepgen, 1979b); blanks under the Namibia column indicate no data and not zero catch Year

Western Cape

1881

3221

1889

3000–4500

1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905

1794 2088 1208 949 1034 728 873 657 810 202 259 51 604 1648 366

Namibia

Year

Western Cape

1928 3167 2437 1535 1724 1627

4542

1929 1930 1931 1932 1933 1935 1936 1937 1938

3534 3123 2643 1337

1961 1962

3885 2295

1969 1970 1971 1972

2834 3020 2654 2010

828 1556 946

Namibia

354

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Year

Western Cape

Namibia

Year

Western Cape

1906 1907 1908 1909 1910

509 823 1884 1185 998

1973 1974 1975 1976 1977

3778 4512 4254 9422 4490 1978 2336 1900 1981 1982 1983 1984

1658 1110 1334 2280 2298 5070 2070 4836 4825 4335 3655 3318

1921 1922 1923 1924 1925 1926 1927

4939 2892 2642 2237 3126 2258 6523

1979 1980 >1475 >1032

Namibia

1608

2624 1472 1226 338

affected, continued until 1982 when the purse-seine concessions were withdrawn. This followed exceptionally low catch rates and handline catches of yellowtail between 1980 and 1982 (Table XXVI). Overall effort expended on yellowtail peaked at an estimated 7000 handline boat-days in 1975–1976. The purse-seine concessions had initially been issued with the intention that they TABLE XXV Estimated landings (metric tons) of some line-fish species at South African ports, 1981–1984 (from Annual Reports of the Chief Directorate Marine Development): redfishes were Chrysoblephus spp., Cheimerius nufar, Polysteganus undulosus, and Petrus rupestris Species Atractoscion aequidens (Geelbek) Argyrosomus hololepidotus (Kob) Seriola lalandi (Yellowtail) Argyrozona argyrozona (Silverfish or carpenter) Thyrsites atun (Snoek) Scomber japonicus (Chub mackerel) Merluccius spp. (Cape hakes) Pachymetopon blochii (Hottentot) Pomatomus saltatrix

1981

1982

1983

1984

185

121

264

124

846

749

815

589

219

405

353

598

674

655

834

570

4825

4335

3655

3318

1784

16

73

389

25

10

6

35

507

424

234

203

THE BENGUELA ECOSYSTEM: PART IV

Species

1981

1982

1983

1984

(Elf) Coracinus capensis (Galjoen) Rhabdosargus globiceps (White stumpnose) Pterogymnus laniarius (Panga) Lichia amia (Leervis) Redfishes Sharks Other Total

14

20

20

25

5

2

3

0

15

11

37

53

16

11

22

78

1 409 206 385 10118

3 302 145 401 7610

3 451 62 386 7220

0 436 42 131 6593

355

should be used primarily for exploitation of tunas. From an early stage, however, yellowtail dominated the landings. The bulk of the handline catch is made by small boats operating from Gans Bay, Struis Bay, and Arniston in the vicinity of Cape Agulhas, and it was in the same region, although often further offshore, that the purse-seiners operated. Purse-seine fishing for yellowtail was initially prohibited in an area between Cape Agulhas and Cape St Blaize extending six nautical miles (11 km) seawards. In late 1978 the area was extended to include an area 15 nautical miles (28 km) seawards between Quoin Point and Struis Point. Hottentot, Pachymetopon blochii, is caught mainly in rocky areas west of Cape Agulhas where it is a common food-fish (Nepgen, 1977). Between 1898 and 1906 roman, Chrysoblephus laticeps, dominated landings of red fishes at the Strand and Hermanus, whereas red stumpnose was the main species along the eastern Cape Peninsula and seventy-four, Polysteganus undulosus, at most harbours along the southern Cape coast east of Cape Agulhas (Fig. 11). By TABLE XXVI Catch (metric tons), effort and catch per unit effort for the fishery for yellow tail Seriola lalandi off the southern coast of South Africa, 1968–1982 (from Annual Reports of the Chief Directorate Marine Development) Year

Handline catch Beach catch Purse seine catch Total catch Effort (handline boat-days)

Catch per unit effort (tons/boat-day)

1968–1969 1969–1970 1970–1971 1971–1972 1972–1973 1973–1974 1974–1975 1975–1976 1976–1977 1977–1978 1978–1979

238 198 1085 491 274 374 404 576 470 470 476

0·12 0·12 0·25 0·19 0·11 0·16 0·13 0·18 0·16 0·12 0·19

– – – 48 60 18 14 74 171 76 12

– – 8 121 238 474 394 542 228 212 157

238 198 1093 660 572 866 812 1265 869 758 645

1967 1623 4303 3492 5018 5379 6496 6989 5398 6370 3431

356

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Year

Handline catch Beach catch Purse seine catch Total catch Effort (handline boat-days)

Catch per unit effort (tons/boat-day)

1979–1980 1980–1981 1981–1982 1982–1983 1983–1984

300 154 224 378 474

0·14 0·06 0·12

18 4 20 14 15

231 163 184 – –

549 321 428 392 489

3839 4938 3598

the late 1970s roman was the dominant redfish species between Kalk Bay and Arniston and also important elsewhere, whereas seventy-four was only recorded in significant quantities from Port Alfred at the extreme east of the southern Cape (Fig. 12). At Gans Bay catch rates of red fishes decreased between 1964 and 1972, and then increased to a peak in 1977 before again decreasing. Catch rates of red fishes also increased at Struis Bay through the mid-1970s (Crawford & Crous, 1982). In the Cape Agulhas region yellowtail and geelbek, Atractoscion aequidens, are caught mainly from November to May, as is kob, Argyrosomus hololepidotus, at Gans Bay. Catch rates of kob at Struis Bay are highest from April to June. These are three of the preferred species, and consequently red fishes and silverfish (carpenter), Argyrozona argyrozona, are generally exploited in other seasons (Crawford & Crous, 1982). Off Port Elizabeth the best line catches of kob have been recorded between August and March, whereas trawl catches from further offshore have peaked in both winter and summer (Smale, 1985). Longterm trends from the trawl fishery suggest that abundance of kob off the southeastern Cape decreased after the mid-1960s (Smale, 1985). Elf, Pomatomus saltatrix, are caught off Port Elizabeth from spring to early autumn (Smale & Buxton, 1985). Adults migrate to Natal during winter, where they spawn in spring (Van der Elst, 1976). Tuna fishery Tunas were recorded in South African waters during the last century (Günther, 1860), and the existence of various species in the Benguela was documented during the first part of the twentieth century (Gilchrist, 1902; Thompson, 1918; Barnard, 1927). Tunas were generally little studied prior to the late 1940s, e.g., bigeye tuna, Thunnus obesus, were first recorded in South African waters only in 1959 (Talbot & Penrith, 1961). In an attempt to stimulate development of a South African industry for fishing and canning of tuna, research was intensified shortly after World War II (Anonymous, 1946), and Molteno (1948) published a booklet discussing five species and dealing with the economics of marketing tuna and tuna products. Tunas were caught in fair numbers by local sport fishermen during the summer months in the 1940s and 1950s, but were not exploited commercially off the Western Cape until 1960 (Talbot & Penrith, 1968) when a small longline fishery developed (the lines extending horizontally in the epipelagic zone) and stimulated further research (Talbot & Penrith, 1962; De Jager, Nepgen & van Wyk, 1963a; Talbot, 1964). The industry expanded rapidly and during the early 1960s annual landings of tunas by South African boats amounted to about 2000 metric tons (Nepgen, 1971). The South African fishery collapsed during 1964, mainly on account of a poor market (Welsh, 1968) but partly also because of low quality of landed fish (N.Bacon, pers. comm.). Subsequently it was not until the late 1970s that there was any renewed interest in commercial fishing for tunas by South African boats, although sport fishermen had continued to make reasonable catches throughout the 1960s and 1970s. By 1979 poling as a catching technique had been perfected by local fishermen and in that year a major run of yellowfin tuna, T. albacares, occurred. The fishery temporarily

THE BENGUELA ECOSYSTEM: PART IV

357

Fig. 11.—Species composition of catches of red fishes at eleven southern Cape harbours, 1898–1906 (from Crawford &Crous, 1982).

blossomed and by November 1979 some 115 vessels, excluding sport craft, were fishing for tuna (Shannon, 1980). Over 6000 tons of fish were landed that year. Equivalent availability did not occur in 1980 and again the fishery virtually collapsed. Some effort was re-directed at the longfin tuna or albacore, T. alalunga, and this species, with smaller contributions from T. albacares, T. obsesus and skipjack tuna, Katsuwonus pelamis, constituted the bulk of the South African catch of tunas of about 3000 tons per annum during the

358

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Fig. 12.—Species composition of catches of red fishes at eight southern Cape harbours, 1976–1980 (from Crawford & Crous, 1982).

early 1980s. Poling for albacore occurs primarily within the 200 nautical mile fishing zone and between 29 and 32° S. The South African fishery for tunas remains small in comparison with the combined longline fisheries of Taiwan, Japan, and Korea. During 1964–1967, 85% of Japanese longline catches off southern Africa were albacore, with best harvests being made between 25 and 35° S and 10 and 20° E during autumn, when the hooking rate was 10% (Nepgen, 1970b). A decline in catch rates of albacore by Japanese boats and improvement in those of yellowfin tuna was observed after 1971, and was attributed to a shift in deployment of effort from inshore to offshore waters (Baptista, 1979).

THE BENGUELA ECOSYSTEM: PART IV

359

Taiwan has replaced Japan as the major exploiter of tunas in the Southeast Atlantic (Yang & Sun, 1984), and hooking rates by Taiwanese vessels have remained fairly constant at around 3% since 1973. During the late 1970s and early 1980s approximately 14000–18000 tons of tuna were trans-shipped annually by Taiwanese operators in Cape Town (Yang, 1983), of which probably between 4000 and 5000 tons were caught in, or immediately west of, the Benguela region. The total annual catch of tunas from the Benguela and adjacent regions by all nations was of the order of 8000–14000 tons in the early 1980s, which at the 1986 market price for landed fish was worth 12–15 million U.S. dollars. In processed form the tuna catch would have been worth much more. Southern bluefin tuna, Thunnus maccoyii, was one of the two most abundant Benguela tunas during the 1960s, when it appeared off Cape Point in large numbers in winter and to a lesser extent in spring (De Jaeger, 1963; De Jager, Nepgen & van Wyk, 1963a, b; Talbot & Penrith, 1963; Nepgen, 1970a), but during the early 1980s was seldom caught in the area (Yang, 1983). This trend was probably due to heavy exploitation of the stock throughout its TABLE XXVII Reported catches (metric tons) of some tunas in the southeastern Atlantic by all nations, 1972–1984 (from ICSEAF Statistical Bulletins), and estimated catches by South African vessels, 1980–1984 (Sea Fisheries Research Institute, unpubl. records) Year

T. alalunga

Catches by all nations 1960 1400 1961 3300 1962 8000 1963 8600 1964 14200 1965 20800 1966 17000 1967 7300 1968 11700 1969 9900 1970 8200 1971 10300 1972 – 1973 644 1974 139 1975 154 1976 35 1977 74 1978 126 1979 344 1980 1469 1981 3713 1982 2547

T. albacares

T. maccoyii

T. obesus

K. pelamis

E. alletteratus

A. thazard

Total

6100 13900 10300 7400 6300 8600 4900 4700 3900 2000 1600 2900 663 2504 4381 73 1011 2301 2779 8811 1098 5965 1558

– – 100 – 100 300 700 200 200 600 15100 16800 – 6500 2 52 84 – – – 13 11 –

300 6300 5800 5000 4500 10000 6600 4900 5600 4400 3300 9000 – 5600 12 51 – – – 20 422 415 137

500 1400 2100 2400 1500 1400 2800 2000 4200 1900 1000 500 1691 1943 8995 689 1514 4076 3591 3640 3543 2730 2295

3000 2500 5100 4000 1700 3500 3200 4000 2700 2000 6200 3400 1646 930 – 449 10 1326 826 875 1328 1171 4202

1600 2100 1500 1200 900 1700 1400 1200 600 800 500 300 3760 1062 – 535 27 198 377 357 256 351 531

12900 29500 32900 28600 29200 46300 36600 24300 28900 21600 35900 43200 7760 19183 13525 2003 2681 7975 7699 14047 8129 14356 11270

360

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Year

T. alalunga

1983 2210 1984 2834 South African catches 1980 1469 1981 1857 1982 2547 1983 2210 1984 2834

T. albacares

T. maccoyii

T. obesus

K. pelamis

E. alletteratus

A. thazard

Total

2498 806

– –

178 36

2489 83

2101 237

531 –

10007 3996

540 186 49 457 806

13 6 – – –

422 381 137 178 36

61 110 38 104 10

– – – – –

– – – – –

2505 2540 2771 2949 3686

range. The only known spawning ground for T. maccoyii is off northwestern Australia. Ready availability of the species in the Benguela system in the 1960s may also have been related to unusual environmental conditions. Northern bluefin tuna, T. thynnus, can be regarded as extremely rare in the Benguela system. T. maccoyii and T. thynnus have in the past been mis-identified by fishermen, with resultant incorrect reporting of the catches of these two species. Albacore, yellowfin, and skipjack have all been important contributors to the tuna catch from the Southeast Atlantic (Table XXVII). Trends in the international catches of five species of tunas in the entire Atlantic Ocean for the period 1950–1984 are shown in Figure 13. Although skipjack appear to have been one of the most abundant of the Atlantic tunas since 1970, there exists no directed fishery for this species in the Benguela region. Skipjack are caught off Cape Point in summer by sport fishermen, and are an important commercial species off Angola, where between 1956 and 1969 they largely replaced T. albacares in the catches (De Campos Rosado, 1971). The geographical distribution of catches of some species during 1972 to 1984 is shown in Table XXVIII. TABLE XXVIII Percentage contribution of catches from the different ICSEAF Divisions to the overall catch of some tuna species reported from the southeastern Atlantic during 1972–1984 (from ICSEAF Statistical Bulletins) ICSEAF Division Species

1.1

1.2

1.3

1.4

1.5

1.6

2.1

2.2

T. alalunga T. albacares T. maccoyii T. obesus K. pelamis E. alletteratus A. thazard

– 0·2 – 0·2 – 21·2 1·9

– 27·3 – – 54·1 59·8 88·6

– 25·3 – – 43·7 17·3 9·1

– – – – – 0·2 0·4

– – 80·0 – 0·1 0·7 –

99·8 31·3 20·0 99·8 1·9 0·8 –

0·2 15·6 – – 0·1 – –

– 0·2 – – – – –

SET-NET, DRIFT-NET AND BEACH-SEINE FISHERIES About 25 species contribute to catches by set and drift nets and beach seines off the Western Cape, most important of which is the harder or southern

THE BENGUELA ECOSYSTEM: PART IV

361

Fig. 13.—Trends in the annual catches of five tunas (Thunnus and Katsuwonus spp.) in the Atlantic Ocean, 1950–1983 (from information in Miyake, Nordstrom & Da Rodda, 1981; Miyake, Wise, Nordstrom & Da Rodda, 1982a, b; Miyake & Da Rodda, 1984). TABLE XXIX Recorded catches (metric tons) by set nets, drift nets and beach seines off the Cape Province and of harders by purse seines, 1974– 1984 (from De Villiers, 1986, assuming an average mass of 0·25 kg per harder caught, and Annual Reports of the Chief Directorate Marine Development) Year

Set nets, drift nets and beach seines

St Joseph shark

White steenbras

Yellowtail Horse mackerel

1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984

1454 1600 1439 123 67 133 263 506 560 411

81 62 65 36 50 55 49 105 24 5

Harder

Other

Purse seines All gears Total Harder Harder

1306 1535 1662 1504 51 23 15 4 13 16 15

20 30 37 9 18 3 5

1569 1391 2122 3538 2016 1355 1555

68 31 154 188 105 34 106

1637 1422 2276 3726 2121 1389 2097

175

1481

34

1589

362

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

mullet, Mugil richardsoni, which has been exploited since the seventeenth century (De Villiers, 1986). Harders have also been caught with purse seine. Control measures applied to the harder fishery have been summarized by De Villiers (1986). A minimum stretched mesh size of 44 mm was in force prior to 1955, and from 1969 there have been restrictions on number of operators using specified gear, area of operation and net size. In particular, from 1980 only seven boats were allowed to catch harders with purse-seine nets, and in 1982 all set nets were replaced by 178 mm drift nets directed at St Joseph sharks, Callorhinchus capensis. Catches of harders off the Western Cape for the period 1974–1984 are listed in Table XXIX. Harvests were especially good in the early 1980s. Whether the inter-annual variation in catches is a result of a fluctuating stock size or of factors affecting availability is unknown (De Villiers, 1986). Although harders are caught from the Orange River to east of Knysna, 95% of the yield is from between Elands Bay and Cape Agulhas, with catches especially high in St Helena Bay and False Bay (De Villiers, 1986). To the west of Cape Point harders are caught mainly in spring and summer, possibly on account of unfavourable fishing conditions in winter, when winds from the northwest cause heavy seas along the exposed shores, or as a result of fishermen in winter directing effort towards the more lucrative snoek resource (De Villiers, 1976). Catches between Cape Point and Cape Agulhas are spread evenly throughout the year, but further east are made almost exclusively between July and November and may result from a seasonal migration (De Villiers, 1976). Catches of St Joseph sharks increased in the 1980s with the development of a fishery directed at this species (De Villiers, 1986). Other contributers to the catches of set nets, drift nets and beach seines include white steenbras, Lithognathus lithognathus, yellowtail, horse mackerel, and kob (Table XXIX). CRUSTACEAN FISHERIES Two distinct types of crustacean fisheries are conducted in the coastal waters of South Africa, namely trapfisheries (on the western and southern coasts) and a demersal trawl fishery on the eastern coast. Three main species of rock lobsters are fished for: Jasus lalandii (Western Cape and Namibia), Palinurus gilchristi (Southern Cape), and P. delagoae (eastern coast). A number of Indo-Pacific species of spiny lobster also occur on shallow inshore reefs of Natal, but these do not form the basis of any commercial fishery. The demersal trawl fishery off the eastern coast concentrates on P. delagoae, the langoustine, Nephrops andamanicus, various Indo-West Pacific penaeid prawn and caridean shrimp species, as well as a species of red crab, Geryon. Only species occurring within the Benguela system are discussed in detail below, but for the sake of completeness Table XXX shows the relative magnitudes of catches of various crustacean species or groups from the coastal waters of South Africa and Namibia. It is apparent that approximately 75% of the total crustacean catch of about 7400 tons in recent years has been comprised of Jasus lalandii, a cool-water species associated with the Benguela upwelling system. TABLE XXX Catches of crustaceans (metric tons whole weight) from the coastal waters of South Africa and Namibia, 1960–1984 Jasus lalandii Year

Jasus lalandii

South Africa

Namibia

1960 1961

11144 10391

3840 5770

Palinurus gilchristi

Palinurus delagoae

Prawns and shrimps

Total

– –

– 66

– –

14984 16227

THE BENGUELA ECOSYSTEM: PART IV

Year

Jasus lalandii

South Africa

Namibia

1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984

10090 10391 10391 10391 9186 7078 6174 5900 5600 5290 5541 5722 6288 6297 6297 6297 5352 5352 4548 3700 3700 3730 3730

7250 7277 7609 8000 8874 5343 8602 6902 3481 2396 2826 2928 1650 1506 1804 1110 1691 1373 1198 1339 1770 1865 1820

Palinurus gilchristi

Palinurus delagoae

Prawns and shrimps

Total

– – – – – – – – – – – – 744 1946 1424 1335 922 244 353 696 789 1049 968

70 86 188 279 254 32 31 143 75 44 93 105 86 53 61 19 71 106 156 143 115 66 43

– – – – – 555 605 372 422 429 490 642 1094 1137 1217 1118 882 974 1319 673 1136 729 673

17410 17754 18188 18670 18314 13008 15412 13317 9578 8159 8950 9397 9802 10939 10803 9879 8918 8049 7574 6551 7510 7439 7234

363

Western coast rock lobster J. lalandii occurs in commercially exploitable densities along a 900-km length of coastline from about 25° S in Namibia to Cape Point (34°20′ S), although the species also occurs in lower densities to the north and to the southeast of the core of distribution. Current yields of this species are approximately 5600 tons, which makes this fishery the world’s largest for any single Jasus species. This genus comprises seven species, all located in the southern hemisphere. New Zealand’s J. edwardsii takes second place with an annual yield of about 4000 tons, while Australia (including Tasmania) currently produces about 3400 tons of the species J. novaehollandiae (Winstanley, Caton, Harris & Lewis, 1983). J. lalandii is the only species of Jasus which occurs in a strongly developed upwelling regime, with its associated high primary and secondary levels of productivity. Some 2800 km west-southwest of Cape Town, in the central South Atlantic, the isolated islands of Tristan da Cunha, Inaccessible, Nightingale, and Gough support a population of J. tristani, a species which currently yields about 400 tons (Pollock, 1981). Off southern Africa rock lobsters are captured with baited traps or hoopnets. Over 260 trap-fishing vessels exist in the South African fishery, varying from about 6 to 14 m in length. These fish mainly on the deeper-water fishing grounds off the southwestern Cape coast, deploying traps at depths of up to 80 m.

364

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Dinghies, from which hoopnets are deployed, operate close inshore usually at depths not exceeding 25 m, mainly along the northern regions of the Cape coast and off Namibia. Rock lobsters generally occur in shallow waters in the region north of about 31°50′ S, where their depth distribution appears to be strongly influenced by the presence of oxygen-deficient bottom waters further offshore. As shown in Table XXX a marked decline in South African catches took place during the late 1960s. Although detailed catch statistics for individual grounds are not available for this period, yields apparently declined far more severely in the northern Cape and Namibia than elsewhere. In these northern areas, overexploitation at reduced minimum size limits had occurred during the 1960s but, in addition, it is suspected that environmental changes took place possibly resulting in gradually increased oxygen depletion and a retraction of the depth range normally occupied by the rock lobsters. Of the present total yield of about 5600 tons, approximately 84%, is exported. Frozen tails comprise about 49% of the exports, as do frozen whole raw and cooked rock lobsters. The remaining 2% is made up of live rock lobster. A minimum size limit of 89 mm carapace length has been enforced since 1933 in South Africa, with a reduction to 76 mm allowed in the far northern Cape regions only (north of about 31° S) from 1963 to 1970. Since November 1985, the minimum size limit in the latter region was again reduced, on an experimental basis, to 75 mm, to gauge whether yields from this region could be improved by harvesting at a smaller size at first capture. In Namibia, a series of changes in size limits were allowed from 1950 to 1964, culminating eventually in the complete removal of a size limit in 1968 and 1969, whereupon catches of small, mainly juvenile animals rose to alarmingly high levels. Since 1970 a size limit of 64 mm (increased to 65 mm in 1980) has been enforced in Namibia. Quotas to control the production of export tail-mass were first introduced in the Cape in 1946 and in Namibia in 1949. The South African tail-mass quota was reduced by about 50% to the equivalent of 5600 tons whole mass in 1970 when it became apparent that the declining yields of the late 1960s were not simply a shortlived phenomenon. After 1970, South African catches have varied directly in accordance with quota levels set. In Namibia on the other hand, despite a reduction in quota of about 45% to 4350 tons in 1970, total catches have remained at levels well below the quota and appear to be controlled by annual production of the stock and effort directed by the fishery. Fishing effort, in terms of numbers of vessels, is at present low off Namibia, owing to rationalization which took place before 1973 when the number of vessels was reduced from a maximum of 75 to only 21. In South Africa effort restrictions have not been as vigorously pursued, although the number of licenced vessels was ‘frozen’ in 1970 when the quota was halved. A number of additional control measures are also enforced, including closed seasons (May to October in both South Africa and Namibia), a prohibition on the landing of berried females and soft-shelled animals, and bag limits for amateurs. Red crab The red crab, Geryon maritae, occurs on the slope of the continental shelf from about 27° S off Namibia, northwards to Angola, the Congo and Ivory coast. Off Namibia, it is found at depths of about 300 to 900 m, on soft mud substrata (Beyers & Wilke, 1980; Melville-Smith, 1985). Although red crab can be caught in bottom trawls, trap-fishing is the preferred method of capture, and a trap-fishery has developed with vessels using beehive-type traps on longlines. Exploitation by three Japanese vessels commenced in 1973. A local company entered the fishery in 1976, but withdrew its two vessels in 1979. Since then fishing has only been conducted by Japanese vessels, at present numbering five. Yields in the mid-1980s were estimated at over 5000 metric tons whole mass, making this fishery the

THE BENGUELA ECOSYSTEM: PART IV

365

largest single species crustacean fishery in the northern Benguela system. No minimum size limit is observed in the fishery, nor are there any restrictions on fishing effort. MOLLUSCAN FISHERIES The molluscs are a diverse group of organisms, ranging from highly mobile groups, such as squid and to a lesser extent cuttlefish and octopus, through the more sedentary gastropods to truly sessile filter-feeding mussels or oysters. In South Africa, fisheries are directed at all three of these groupings, but vary greatly in importance. The only fishery directed at sessile molluscs is the oyster fishery of the Southern Cape, although a minor fishery for white mussels (clams), Donax serra, exists in the Western Cape. No commercial exploitation of black mussels takes place anywhere on the coast. A limited tonnage of squid is caught as a by-catch of the hake trawl-fishery off the Western Cape (see Table XX, p. 392), but most squid (mainly Loligo reynaudii) are located on the Agulhas Bank to the south and east of Cape Agulhas outside the Benguela upwelling system. Abalone, Haliotis midae, form the basis of a small yet profitable fishery conducted mainly between Cape Point and Cape Agulhas, although the species also extends north of Cape Point to the vicinity of St Helena Bay. Catches of squid and abalone since 1978 are listed in Table XXXI. Squid fishery Catches of white squid, Loligo reynaudii, have for many years been taken as a by-catch of the trawl fishery directed at ground-fish species, but only fairly recently have squid gained prominence as a sought-after food item. Owing to high prices now being offered for squid, there was a rapid growth of an inshore jig fishery directed at shoals of adult squid during 1984 and 1985. The jig fishery is conducted mainly by small boat operators in bays along the southern coast where squid aggregate for spawning, especially during the summer months. In the offshore trawl fishery, white squid are caught together with smaller quantities of red squid, Todaropsis eblanae and Todarodes angolensis, and some biomass estimates have been published for these species. Off the Western Cape, Payne, Augustyn & Leslie (1985) estimated the biomass of white squid TABLE XXXI South African catches (metric tons whole weight) of squid Loligo, Todaropsis and Todarodes spp., and abalone Haliotis midae, 1978–1985, indicating also the catch of squid by other nations Year

Squid

Trawlers South African

Other Nations

1978 1979 1980 1981 1982 1983 1984 1985

1986 2685 1876 1769 1455 1575 1168 1016

Jiggers

Total

Abalone

3003 2422 1178 1987 2149 2610 1085 742

– – – – – 500 1000 3100

4989 5107 3054 3756 3604 4685 3253 4858

795 738 739 734 727 650 655 658

366

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

to be less than 300 tons. By contrast, three joint Japanese-South African trawl surveys for squid and groundfish resources on the Agulhas Bank resulted in estimates of squid biomass (within a depth range of 80 to 200 m) of about 17000 to 36000 tons, consisting mainly of immature squid thought to be less than a year old (Hatanaka et al., 1983; Uozumi et al., 1984; Uozumi, Hatanaka, Payne & Augustyn, 1985). As yet, no restrictions have been placed on commercial squid catches, although a bag limit of 50 squid per person per day has been imposed on amateur line fishermen. Abalone fishery Three species of abalone occur on the southwestern coast of South Africa, but only Haliotis midae attains a large enough size and occurs in sufficiently high densities to attract commercial exploitation. This species has a distribution range from about St Helena Bay in the Western Cape to about as far east as 29° E on the coast of the Transkei (Newman, 1969a). Densities are greatly reduced east of Cape Agulhas, although the species appears in reasonable quantities in the vicinity of Port Elizabeth. Most commercial exploitation takes place between Cape Point and Cape Agulhas, an ecological region described by Stephenson (1944) as the “Western overlap” region (Newman, 1969a). This area is intermediate in physical characteristics between the fully developed upwelling system north of Cape Point and the region east of Cape Agulhas which is dominated by the Agulhas Current. Commercial exploitation of abalone began in 1949, but records of yields are only available from 1953 onwards (Newman, 1973). Catches rose to peak levels in the early 1960s but declined steadily after 1965 as accumulated stocks of this slow-growing mollusc were removed. A production quota was enforced from 1969 to 1982 but was replaced by a whole mass quota in 1983. Quotas in the mid-1980s were about 660 metric tons. In 1986 51 divers were licensed to collect abalone commercially. Catch rates have remained stable since the early 1980s and it is believed that the present level of catch may be sustainable (R.J.Q.Tarr, pers. comm.). A minimum size limit of 114-mm shell breadth has been enforced since 1955. Two sanctuary areas exist between Cape Point and Cape Agulhas, one at Betty’s Bay which came into operation in 1968 and the other, Dyer Island, in 1983. A closed season for fishing from August to October was introduced in 1986 in an attempt to ensure adequate supplies of abalone for the local market during the November-January holiday period. THE MAJOR RESOURCES GROWTH There have been investigations into growth with age of a variety of fish species occurring in the Southeast Atlantic, particular attention having been given to those considered of greatest commercial value as information on length at age is required for catch-based stock assessments. In certain instances the same species has been the subject of a number of studies, and often initial interpretations have later been substantially revised. Early studies were based largely on analysis of seasonal ring formation on hard structures such as otoliths and scales, although modal progressions in length frequency distributions were also used (e.g. Geldenhuys, 1973). More recently counts of daily rings on otoliths and natural tags have also been employed (e.g. Thomas, 1985). Results have most frequently been expressed in terms of the Von Bertalanffy growth formula:

THE BENGUELA ECOSYSTEM: PART IV

367

where Lt=length at age t; =asymptotic length; K=growth coefficient; t0=age at zero length. This formula appears to describe growth of most fish species in the Benguela reasonably well, although Babayan & Bulgakova (1983) preferred the relationship of Parker & Larkin (1959) for Cape horse mackerel off Namibia. Parameters derived for the Von Bertalanffy equation for some species are listed in Table XXXII. Often authors provided a single equation thought appropriate for the species under consideration, despite obtaining different parameters for the two sexes, separate geographical regions or different methods of age determination. In such instances only values of parameters pertaining to the single overall equation are shown. For studies where no unifying equation was formulated sets of parameters are listed. For species subject to more than one study, the earlier findings have not been included if age-determination techniques more recently available have suggested substantially different rates of growth. Seasonal variation in growth has been reported for both anchovy and pilchard, with growth being slower in winter (Melo, 1984; Thomas, 1985). For pilchard, Thomas (1985) showed that the standard deviation of the mean length of a recruiting group was controlled by two opposing effects: growth TABLE XXXII Values of parameters derived for the Von Bertalanffy growth formula for some fish species occurring in the southeastern Atlantic; for species that have been subject to a number of studies only selected results are shown; lengths may be assumed to represent total lengths unless otherwise indicated; * studies for T. capensis have been summarized by Babayan & Bulgakova (1983); † studies for Merluccius spp. have been summarized by Kono (1980) Species

Area

(cm)

K

t0 (years) References

S. ocellatus E. japonicus

Namibia Namibia

S. aurita T. capensis* T. trecae E. whiteheadi L. hectoris M. muelleri S. japonicus M. capensis†

Off Angola Western Cape

22·61 caudal 12·1 13·0 14·0 caudal 43·3 54·29 45·47 26·19 caudal 7·00 caudal 6·00 caudal 68·01 caudal 111·14 141·35 118·8 107·14 121·8 129·1 170 180·56 184·86

1·09 0·967 1·33 1·49 0·28 0·127 0·1823 0·334 1·66 1·15 0·207 0·12 0·0911 0·1106 0.137 0·1129 0·0989 0·05 0·0572 0·0656

−0·003 −1·319 −0·165 0·0003 −0·4 −0·3062 −1·76 −0·1664 0·06 0·06 −0·9845 −0·24 0·4731 0·0955 0·5299 −0·0034 −0.0059 −1·203 −1·8566 −0·6189

M. paradoxus†

Western Cape Western Cape Western Cape Western Cape ICSEAF Divisions 1.3 and 1.4 Divisions 1.5 and 1.6 Agulhas Bank ICSEAF Divisions 1.3 and 1.4 Divisions 1.6 and 2.1 Divisions 2.1 and 2.2

G. capensis ICSEAF Divisions 1.3 and 1.4 Division 1.5 and 1.6

Thomas (1985) Melo (1984) (using various techniques) Shcherbich (1981) Geldenhuys (1978) Venediktova (1985) Geldenhuys (1978) Prosch (1986) Prosch (1986) Baird (1977) Pozo (1976) Botha (1971) Kono (1980) Botha (1971) Kono (1980) Wrzesinski (1975) Isarev (1976)

368

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Species

P. laniarius D. macrophthalmus A. microlepis A. pectoralis H. maculatus U. canariensis P. blochii A. argyrozona

Area Division 1.5 Southeast Agulhas Bank Agulhas Bank Northern stock Agulhas Bank ICSEAF Division 1.5 ICSEAF Division 1.6 ICSEAF Division 1.6

(cm)

K

t0 (years) References

205·25 114·21 224·54

0·0624 0·06 0·0539

−0·0042 −4·08 −0·6034

Payne (1977) Morales (1980) Payne (1985b)

38·4 39·87 57·127 44·0 76·31 100·42 42·04 74·49

0·1842 0·244 0·0713 0·34 0·032 0·0553 0·1444 0·0708

−0·004 −0·525 −5·4992 0·05 −5·792 −1.658 −1·203 −1.52

Sato (1977) Kuderskaya (1985) Morales (1982) Zoutendyk (1974) Morales (1980) Venidictova (1982) Nepgen (1977) Nepgen (1977)

depensation caused by larger fish growing faster (which he presumed to result from competitive advantage) and growth compensation (caused by the general decrease in growth rate with size). He also demonstrated interannual variation in growth, and suggested that the slower growth reported by Nawratil (1960), which was derived from reading of scales, may have been a result of the larger biomass of pilchard off Namibia during 1952–1958 when the fish on which Nawratil worked were collected. FOOD, PREDATORS AND MORTALITY Pilchard and anchovy Feeding of pilchard and anchovy off Namibia during 1971 and 1972 was studied by King & Macleod (1976). They reported a low incidence of larvae (three out of 250) with food (crustacean eggs and copepod nauplii) in the digestive tract. The juveniles of both species were zooplanktophagous, feeding mainly on calanoid copepods, and the adults were phytoplanktophagous, consuming chiefly diatoms (Table XXXIII). The change in the diet of pilchard occurred at approximately 100 mm caudal length and that of anchovy at about 80 mm, and in both instances the change was attributed to development TABLE XXXIII Percentage contribution of phytoplankton and zooplankton by volume to the diet of small and large pilchard Sardinops ocellatus and anchovy Engraulis japonicus off Namibia during 1971 and 1972 (summarized from King & Macleod, 1976) Species

S. ocellatus

Caudal length (mm) Sample size Phytoplankton Zooplankton

20–100 508 12·7 86·4

E. japonicus 100–260 348 77·1 22·9

20–80 240 12·0 86·6

80–153 169 70·1 29·9

of a fine filtering mechanism. The major zooplankton forms eaten by juveniles were the calanoid copepods Calanoides carinatus, Paracalanus parvus, Centropages brachiatus and Metridia lucens, cyclopoids of the

THE BENGUELA ECOSYSTEM: PART IV

369

genus Oithona, the harpacticoid Microsetella rosea, and furcilia and calyptopis stages of euphausiids. Diatoms dominant in the diet of adults were Delphineis (=Fragilaria) karstenii and Chaetoceros spp. Rhizosolenia setigera, Stephanopyxis turris, and Coscinodiscus spp. were also recorded in high numbers. Although a higher proportion of phytoplankton was found in the diet of adult pilchard, the species were considered to be in competition with each other for food throughout their life cycles (King & Macleod, 1976). Comparison of stomach contents and plankton obtained from hauls in areas where fish were caught suggested that both pilchard and anchovy were generally non-selective filter-feeders (King & Macleod, 1976). Off the Western Cape, Davies (1957) considered pilchard to be an omnivorous filter-feeder, with phytoplankton contributing approximately twice as much as zooplankton to the diet between 1953 and 1956. Phytoplankton was also later regarded as more important for anchovy than zooplankton (Robinson, 1966). Similar to the situation off Namibia, diatoms were dominant in phytoplankton and copepods in zooplankton eaten by pilchard off the Western Cape (Davies, 1957). Copepods eaten were mainly of the genera Calanus (probably Calanoides), Centropages, and Oithona (Davies, 1957). A.G.James (pers. comm.) reviewed the feeding behaviour and diet of sardine-like fishes with special reference to species occurring off southern Africa. He observed that the larvae and juveniles of all forms described in the literature were wholly particulate-feeders, but that juveniles and adults of most species were both particulate- and filter-feeders, particulate-feeding being the dominant mode. These findings agree with those of Blaxter & Hunter (1982). Northern anchovy, Engraulis mordax, were found to feed selectively on larger particles in the Southern California Bight (Koslow, 1981). A.G.James (pers. comm.) suggested that off southern Africa a large proportion of phytoplankton in stomachs of anchovy and pilchard might represent incidental ingestion. On the basis of the diet of Sardinops spp. and Engraulis spp. from other regions of the world, he argued that the contribution of zooplankton to the diet of these genera off southern Africa had been under-estimated. In the early 1980s predators were believed to consume more than 800000 metric tons of anchovy per annum off South Africa’s Western Cape, about 656000 tons being eaten by snoek, chub mackerel and other predatory fish, about 96000 tons by marine mammals, mainly the Cape fur seal, Arctocephalus pusillus, and about 49000 tons by seabirds, with prédation by squids unknown. Natural mortality was thought to account for about 73% of all mortality (Bergh, Field & Shannon, 1985). Predators of pilchard are similar to those of anchovy (e.g. Rand, 1959a; Crawford & Shelton, 1978; Nepgen, 1979a). For pilchard an analysis of tag returns off Namibia suggested that the annual natural mortality rate (M) was unlikely to be higher than 0·5 (Newman, 1970a), and this value was later found to be consistent with aerial estimates of abundance (Butterworth, 1983). For anchovy M was estimated from the age composition of fish in catches in 1965 of the relatively unfished stock off the Western Cape to be 0·8 (Newman & Crawford, 1980). Growth, however, is now believed to be faster than initially thought (Armstrong, Shelton & Prosch, 1985), and a value for M of 1·1 may be more appropriate (Butterworth, 1983). Considerable interannual variability is likely (Crawford & De Villiers, 1985). Horse mackerel Many of the more detailed studies on the feeding of Cape horse mackerel have been conducted off Namibia (Table XXXIV). Venter (1976) found copepods to be the most important food of fish of 20-cm caudal length or less between 1972 and 1974, and in the early 1970s other researchers noted a dominance of copepods in the diet of Cape horse mackerel in ICSEAF Division 1.3 (Lipskaya, 1972; Kompowski & Slosarczyk, 1975; in Krzeptowski, 1982). In 1982 copepods contributed 74% by mass of the food of small

370

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

horse mackerel (Andronov, 1983) In the late 1970s and early 1980s the diet of larger fish consisted mainly of euphausiids (Krzeptowski, 1982; TABLE XXXIV Diet of Cape horse mackerel Trachurus trachurus off southern Africa, summarized from various sources as indicated Length (mm)

unspecifi ed

unspecifi ed

Length measured

unspecifi ed

unspecifi ed

caudal

unspecifi ed

Period of study

unspecifi ed

unspecifi ed

1972– 1974

1977

ICSEAF Division (s)

1.3

1.3 1.4

1.3 and 1. 4

1.3 and 1. 4

1.3 and 1. 4

1.3 and 1. 4

Method used

% mass

% mass

% volume

% frequency occurrenc e

% mass

% mass

Sample size

unspecifi ed

unspecifi ed

3689

all sizes 3300

2400 12

113

164

1793

Phytopla nkton Polychae ta Copepod a Amphipo da Mysidac ea Cumacea Euphausi acea Decapod a Crustace a larvae Cephalop oda Chaetogn aths Other invertebr ates Fish

–

–

–

0·6

9·3

–

–

–

–

–

–

–

–

–

2·2

14·5

–

–

–

–

–

–

63·2

59·2

0·3

58·4

11·7

–

–

73·5

5·4

71·9

1·9

–

–

–

6·9

5·5

–

–

0·1

–

–

–

9·6

–

–

0·1

0·9

–

–

–

–

–

–

– 8·9

– 20·8

– 4·8 99·7 13·8

19·0 6·8

– 65·4

– – 98·4 15·0

– 86·1

– 28·1

– 89·3

–

–

–

0·0

2·0

–

–

5·2

3·2

–

5·2

–

–

–

–

–

–

–

–

0·1

–

–

–

–

–

0·5

0·4

–

–

–

0·2

–

0·0

6·7

–

–

2·3

4·8

–

–

4·0

–

–

–

–

–

–

2·4

8·5

5·6

0·4

2·2

–

–

1·4

0·5

19·3

–

5·4

13·9

29·0

1·2

–

4·4

–

1·5

unspecifi ed

unspecifi ed

1981 1982

1983

THE BENGUELA ECOSYSTEM: PART IV

371

Length (mm)

unspecifi ed

unspecifi ed

Length measured

unspecifi ed

unspecifi ed

caudal

unspecifi ed

Period of study

unspecifi ed

unspecifi ed

1972– 1974

1977

ICSEAF Division (s)

1.3

1.3 1.4

1.3 and 1. 4

1.3 and 1. 4

1.3 and 1. 4

1.3 and 1. 4

Method used

% mass

% mass

% volume

% frequency occurrenc e

% mass

% mass

Sample size

unspecifi ed

unspecifi ed

3689

all sizes 3300

2400 12

113

164

1793

Other Referenc e

11·4 Lipskaya (1972) in Krzepto wski (1982)

0·7 Kompow ski & Slosarcz yk (1975) in Krzepto wski (1982)

2·8 Venter (1976)

2·9

–

0·6

– Androno v (1985)

–

–

– Krzepto wski (1982)

unspecifi ed

unspecifi ed

1981 1982

– Androno v (1983)

1983

Andronov, 1983). Venter (1976) reported a more broadly based diet in the early 1970s, when polychaetes, cumaceans and fish were also frequently preyed upon. Almost half of the fish eaten were gobies. Fish were again important in 1977 (Krzeptowski, 1982) and 1985, lanternfish of the genus Diaphus contributing as much as one third of the food of fish 21 to 26 cm long in the latter year (Konchina, 1986). Konchina (1986), however, regarded euphausiids to be the main food of horse mackerel. He noted that euphausiids and lanternfishes were predominant in catches made with Isaacs-Kidd trawl during 1985 and concluded that horse mackerel preyed mainly on the more numerous members of the mesopelagic fauna. Venter (1976) considered the smaller juveniles to feed in the epipelagic zone and larger fish at greater depths. Intensive feeding has been reported for the afternoon (Krzeptowski, 1982; Andronov, 1983, 1985) and in pre-dawn hours (Konchina, 1986). Off Namibia diets and feeding grounds of Cape horse mackerel and hakes are similar (e.g. Krzeptowski, 1982; Andronov, 1983; Konchina, 1986). This similarity is briefly discussed on p. 484. Off the Western Cape in 1983 and 1984 Prosch (1986) found lanternfish and lightfish, Maurolicus muelleri, to be important in the diet of horse mackerel of total length greater than 30 cm, contributing on occasion more than 50% of the mass of food ingested. On the Agulhas Bank Hecht (1976) noted that in 1974 and 1975 horse mackerel fed on euphausiids and calanoid copepods, other forms of zooplankton also being eaten. The species composition of the diet was visually similar to that in plankton tows. In the same region mysids occurred in 90% of stomachs in November and December 1981 and 81% in June 1982. Planktonic crustaceans formed 84% by mass of the food ingested in November and December 1980. Squid and fish were eaten, but less frequently (Hatanaka et

372

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

al., 1983; Uozumi et al., 1984, 1985). Off northern Namibia the daily food ration of Cape horse mackerel has been estimated to vary between 6·6 and 12·5% of body weight per day (Andronov, 1985). Off the Western Cape between 1954 and 1956 horse mackerel formed 44% by volume of the diet of Cape fur seals collected in inshore waters (Rand, 1959a). Horse mackerel are also eaten by hakes (e.g. Konchina, 1986), and juveniles by seabirds (Crawford & Shelton, 1978). Off Namibia the use of three independent methods provided estimates of M for Cape horse mackerel ranging between 0·27 and 0·52, and a middle value of 0·4 was thought appropriate (Babayan & Bulgakova, 1983). Draganik (1978) had earlier arrived at a similar conclusion for fish aged four to seven, obtaining estimates of M between 0·26 and 0·50 with a mean of 0·40. Shcherbich, Komarov & Florinskaya (1980) suggested higher values of between 0·40 and 0·68. On the basis of a comparison between estimates of abundance obtained on direct surveys over the Agulhas Bank, which were believed to under-estimate abundance, and by virtual population analysis (VPA), Kinloch et al. (1986) considered that M probably lay between 0·3 and 0·4. Pelagic goby Barber & Haedrich (1969) found Delphineis (=Fragilaria) karstenii and Coscinodiscus spp. to dominate the diet of juvenile gobies, and Ryther (1969) also drew attention to the importance of the large, chain forming Delphineis in the food of gobies. O’Toole (1978b) found that the diet of adults, juveniles, and larvae consisted predominantly of Delphineis and Chaetoceros, although he also noted the remains of copepods and euphausiids. D’Arcangues (1977), however, found mainly copepods and euphausiids in the stomachs of the juvenile and adult gobies that she studied. In 1979–1981 gobies contained numerically 93% phytoplankton (mostly diatoms) and 7% zooplankton. Zooplankton was present in 80% of the stomachs examined and phytoplankton in 90%. Delphineis karstenii was the most abundant phytoplankton species, with Chaetoceros and Coscinodiscus spp. also eaten, and euphausiids and copepods were the dominant zooplankton groups (Crawford, Cruickshank, Shelton & Kruger, 1985). Similarity between the diet of gobies and of adult pilchard was noted by Crawford et al. (1985), who considered it likely that in the intense perennial upwelling system situated between 22 and 27° S gobies partially replaced pilchards during the 1970s. During the 1970s and early 1980s pelagic gobies were a major food item for many species off Namibia, including Cape horse mackerel (Venter, 1976), Cape hakes (Chlapowski, 1977; Assorov & Kalinina, 1979; Prenski, 1980; Andronov, 1983; Konchina, 1986), kingklip (Macpherson, 1983a), monkfish (Macpherson, 1985), large-eye dentex (Kuderskaya, 1985), west coast sole (A. Badenhorst, pers. comm.), some coastalbreeding seabirds, and the Cape fur seal (Crawford et al., 1985). Although not exploited commercially to any great extent (see Table V, p. 368), the pelagic goby must therefore have been of considerable importance in the ecosystem between about 22 and 27° S. Lanternfish and lightfish The diet of lanternfish caught by purse-seine boats off the Western Cape during the 1970s was described by Prosch (1986). Food consisted entirely of zooplankton, copepods contributing 62%, amphipods 27%, and euphausiids 12%. Although feeding of lightfish in the Benguela has not been investigated, the diet is believed to be similar to that of lanternfish (Prosch, 1986). Off the South African coast tunas, snoek, chub mackerel, horse mackerel, Cape hakes, and the Cape fur seal are known to feed extensively on lanternfish or lightfish (Prosch, 1986; Sea Fisheries Research

THE BENGUELA ECOSYSTEM: PART IV

373

Institute, unpubl. records). Overall consumption of lanternfish in the southern Benguela has been estimated as about 760000 metric tons wet weight per year (Bergh et al., 1985). Off Namibia myctophids provide a major forage resource for many predators, being regularly reported in feeding studies of horse mackerel (Konchina, 1986), hakes (Table XXXVII), kingklip (Macpherson, 1983a) and large-eye dentex (Kuderskaya, 1985). The genus Diaphus appears to be abundant in the northern Benguela. From the age structure of fish in research catches made between 1982 and 1984, Prosch (1986) estimated total annual mortality rates to be 4·9 for lightfish and 2·4 for lanternfish, most of which he regarded as emanating from natural mortality. He considered that the exceptionally high rate derived for lightfish may have resulted from under-representation of the older age categories in the catches. Snoek and chub mackerel In an analysis of the stomach contents of snoek caught by handline, Nepgen (1979a, 1982) found that the species preyed largely on shoaling pelagic fishes. Pilchard were regularly eaten in the late 1950s and anchovy from the mid 1970s through the early 1980s (Table XXXV), matching shifts in the abundances of these two species. Crustaceans, especially the mantis shrimp, Squilla armata, and euphausiids, and molluscs were also eaten. Mantis shrimps were prominent in the diet in the early 1970s, at which time both pilchard and anchovy resources were probably depressed (Crawford, Shelton & Hutchings, 1983), and off the Western Cape they were again consumed by snoek in large quantities in the mid-1980s (G.K.Brill, pers. comm.). Round herring, lantern-fish, lightfish, Cape hakes, saury, and euphausiids featured prominently in the diet of snoek caught by demersal trawlers off the Western Cape in the early 1980s (Prosch, 1986; Sea Fisheries Research Institute, unpubl. records). TABLE XXXV Diet of snoek Thyrsites atun off South Africa in some selected years, 1958– 1980 (from Nepgen, 1979a, 1982) Sample size Crustaceans Stomatopoda (S. armata) Other Molluscans S. officinalis L. reynaudii Fish E. japonicus T. capensis S. japonicus S. ocellatus E. whiteheadi Fish remains Other

1958

1960

1965

1971

1974

1979– 1980

186

32

44

83

96

1068

0·4 –

– 1·3

– –

39·2 –

– –

<1·0 <1·0

– 1·2

– 1·3

– –

1·3 0·6

0·8 2·0

– –

0·8 6·1 0·4 65·7 – 2·4 0·8

11·5 – – 25·6 – 10·3 1·3

42·5 6·8 1·4 5·5 2·7 4·1 1·4

5·1 – – 1·9 3·8 3·2 1·3

26·3 – 0·4 2·4 0·8 5·6 3·6

89·0 <1·0 11·0 1·0 1·0 –

374

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Between 1969 and 1971 stomachs of chub mackerel smaller than 25 cm caudal length sampled from purse-seine catches made between Lamberts Bay and Cape Agulhas contained mainly amphipods and euphausiids (Baird, 1978b). For larger chub mackerel there was a progressive increase of fish in the diet, lanternfish being especially important and contributing almost half the food of chub mackerel larger than 46 cm (Table XXXVI). Anchovy were of little importance. By contrast anchovy were found in half the stomachs of chub mackerel sampled in False Bay in 1979 and 1980 (Nepgen, 1982). Off South Africa, chub mackerel were eaten in limited quantities by Cape hakes in the 1970s (Hecht, 1976; Botha, 1980) and they are also eaten by predatory fish such as snoek (e.g. Nepgen, 1982). Both snoek and chub TABLE XXXVI Percentage composition of the food of chub mackerel Scomber japonicus of different sizes, 1969–1971 (points method, from Baird, 1978b) Caudal length (cm) Sample size Copepoda Amphipoda Mysidacea Euphausiacea Other Crustacea Chaetognatha Tunicata Coelenterata Mollusca Unidentifiable Zooplankton Fish L. hectoris Protomyctophum spp. E. japonicus E. whiteheadi Unidentifiable remains

16 0·2 30·0 – 42·6 – – – – – 13·7

26–45 131 4·7 12·2 0·0 16·5 1·2 – – – – 30·0

110 8·1 3·6 – 30·3 0·7 – – – 0·0 12·8

– – – – 13·5

22·6 – 2·3 1·3 7·3

44·4 0·7 – 0·6 6·8

mackerel are fed on by the Cape fur seal (Rand, 1959a). Seabirds on occasion eat juveniles of the two species (e.g. Crawford & Shelton, 1978; Matthews & Berruti, 1983). The assumed age composition of the chub mackerel catch in 1954, the first year of exploitation, suggested a natural mortality value of 0·43. A value of 0·25, based on estimates for Scomber scombrus has, however, generally been used for VPA (Crawford, 1980c). Hakes Feeding by Cape hakes has been investigated in a number of studies, results of some of the more comprehensive of which are summarized in Table XXXVII. Off Namibia euphausiids were most important in the diet of Cape hakes less than 30 mm in length in 1982 (Andronov, 1983) and in many years were regularly eaten by hakes up to 600 mm. Other important food items were decapod crustaceans, gobies,

THE BENGUELA ECOSYSTEM: PART IV

375

especially for Merluccius capensis, myctophids, and hakes, especially for the larger size categories. Prenski (1978) noted, however, that in 1977 relatively few hakes were larger than 600 mm. In 1985 horse mackerel formed 32% by mass of the food of hakes between 240 and 440 mm in Division 1.3 (Konchina, 1986). Off the Western Cape euphausiids and myctophids were important in the diet of Merluccius spp. in the 1940s. Although hakes were also eaten, their contribution to the diet was not nearly as great as reported by Botha (1980) for the early 1970s when they formed 61% by frequency of occurrence of the diet of fish larger than 400 mm. In the early 1970s round herring was also an important food item, but euphausiids appear to have been overall of lesser TABLE XXXVII Diet of Cape hakes Merluccius capensis and M. paradoxus off southern Africa, 1942–1985, summarized from various sources as indicated: —indicates not encountered; blank indicates may have been encountered or was encountered and included elsewhere Species M. capens is Length of hake (mm)

150– 250

M. parado xus 260– 500

150– 250

M. capens is 260– 500

M. parado xus

M. capens is

unspec ified

150– 400

Length unspec measure ified d

1.3–1. 4

600– 900

caudal

Period 1970– of study 1972 ICSEA F Divisio n(s)

400– 600

1972– 1976 1.3–1. 5

1.3–1. 4

1.5–1. 6

1.3

1977 1.4

1.5

1.6

1.5

1.6

1.4–1. 5

Method % species index (unspecified) used weighted equally for month

% frequency of occurrence

% frequency of occurrence weighted equally for depth of fishing 1163 for all sizes

Sample size

34

662

75

370

136

Copep oda Stomat opoda Amphi poda Mysida cea Eupha usiacea Mysida cea

–

–

–

–

–

–

–

–

–

–

–

–

–

1·3

0·1

–

–

–

–

–

–

–

–

–

–

–

0·4

0·0

3·4

0·6

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

56·8

28·1

2·0

7·3

62·5

–

35·0

13·7

74·4

56·6

32·0

25·4

3·8

376

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Species M. capens is Length of hake (mm)

150– 250

M. parado xus 260– 500

150– 250

M. capens is 260– 500

M. parado xus

M. capens is

unspec ified

150– 400

Length unspec measure ified d

1.3–1. 4

600– 900

caudal

Period 1970– of study 1972 ICSEA F Divisio n(s)

400– 600

1972– 1976 1.3–1. 5

1.3–1. 4

1.5–1. 6

1.3

1977 1.4

Method % species index (unspecified) used weighted equally for month

% frequency of occurrence

Sample size and Eupha usiacea Decap oda Cephal opoda Other inverte brates Sterno ptychid s Mycto phids Sterno ptychid s and Mycto phids Gobiid ae Merluc cius spp.

1.5

1.6

1.5

1.6

1.4–1. 5

% frequency of occurrence weighted equally for depth of fishing 1163 for all sizes

34

662

75

370

136

–

2·6

0·5

–

–

10·5

44·4

–

28·7

20·0

–

–

–

–

0·8

–

9·2

–

–

–

–

–

–

–

–

–

–

0·1

–

–

–

–

–

–

–

–

–

–

–

–

–

23·0

1·6

–

–

–

–

–

–

–

–

–

–

13·4

20·5

26·5

–

25·2

23·7

–

–

–

18·4

21·7

7·4

45·1

17·8

–

0·8

–

24·8

–

–

–

–

59·2

49·9

19·5

–

1·0

–

35·5

–

–

–

22·8

–

–

4·6

26·3

88·7

THE BENGUELA ECOSYSTEM: PART IV

Species M. capens is Length of hake (mm)

150– 250

M. parado xus 260– 500

150– 250

M. capens is 260– 500

M. parado xus

M. capens is

unspec ified

150– 400

Length unspec measure ified d

1.3–1. 4

400– 600

600– 900

caudal

Period 1970– of study 1972 ICSEA F Divisio n(s)

377

1972– 1976 1.3–1. 5

1.3–1. 4

1.5–1. 6

1.3

1977 1.4

Method % species index (unspecified) used weighted equally for month

% frequency of occurrence

Sample size

34

662

75

370

136

Fish other than above Etrume us whiteh eadi Sardin ops ocellat us Engrau lis japonic us Trachu rus spp. Coelor hynchu s spp. Fish other than above

–

11·1

–

3·1

17·9

–

1.5

1.6

1.5

1.6

1.4–1. 5

% frequency of occurrence weighted equally for depth of fishing 1163 for all sizes

–

50·5

–

–

–

–

–

378

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Species M. capens is Length of hake (mm)

150– 250

M. parado xus 260– 500

150– 250

M. capens is 260– 500

M. parado xus

M. capens is

unspec ified

150– 400

Length unspec measure ified d

1.3–1. 4

1972– 1976 1.3–1. 5

1.3–1. 4

1.5–1. 6

1.3

1977 1.4

Method % species index (unspecified) used weighted equally for month

% frequency of occurrence

Sample size

136

Fish remain s Refere nce

34

600– 900

caudal

Period 1970– of study 1972 ICSEA F Divisio n(s)

400– 600

662

75

370

Assorov & Kalinina (1979)

1.5

Chlapowski (1977)

1.6

1.5

1.6

1.4–1. 5

% frequency of occurrence weighted equally for depth of fishing 1163 for all sizes

Prenski (1980)

THE BENGUELA ECOSYSTEM: PART IV

379

TABLE XXXVII-continued Specie Cape s hakes

Cape hakes

Length 28–29 150– of hake 240 (mm)

250– 350

240– 440

360– 580

M. capen sis

M. capen sis

M. capen sis

M. parad oxus

<400

<400

<400

<500

M. capens is

Length unspe measur cified ed

unspe cified

total

cauda l

total

total

Period of study

1985

1942– 1944

1946– 1947

1972– 1973

1974– 1975

ICSEA 1.3–1. F 4 Divisio n(s)

1.3

1.6

1.6

1.6

2.1–2. 2

Metho d used

% by mass

% by mass

% frequency occurrence (empty incl.)

% freque ncy occurr ence (full only)

% by mass

% by numbe r

Sampl e size

30

60

98

27

18

50

346

139

317

121

215

278

201

unspec ified

Copep oda Stoma topod a Amph ipoda Mysid acea Eupha usiace a Mysid acea and Eupha usiace a Decap oda

2·8

–

–

–

–

–

–

–

–

–

–

–

–

–

–

0·5

–

–

–

–

–

–

–

1·0

0·0

0·0

–

–

–

0·4

0·1

–

–

–

–

1·0

0·0

3·0

0·0

2·0

0·0

–

4·1

–

–

–

–

10·0

7·0

–

–

–

–

–

93·1

61·2

13·6

–

–

45·0

23·0

4·0

0·0

4·0

0·0

–

2·0

4·0

2·0

0·0

6·0

4·0

8·0

1982

–

2·9

14·4

3·6

22·6

48·0

25·0

2·0

1·0

380

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Specie Cape s hakes

Cape hakes

Length 28–29 150– of hake 240 (mm)

250– 350

240– 440

360– 580

M. capen sis

M. capen sis

M. capen sis

M. parad oxus

<400

<400

<400

<500

M. capens is

Length unspe measur cified ed

unspe cified

total

cauda l

total

total

Period of study

1985

1942– 1944

1946– 1947

1972– 1973

1974– 1975

ICSEA 1.3–1. F 4 Divisio n(s)

1.3

1.6

1.6

1.6

2.1–2. 2

Metho d used

% by mass

% by mass

% frequency occurrence (empty incl.)

% freque ncy occurr ence (full only)

% by mass

% by numbe r

Sampl e size

30

60

98

27

18

50

346

139

317

121

215

278

201

unspec ified

Cepha lopod a Other invert ebrate s Sterno ptychi ds Mycto phids Sterno ptychi ds and Mycto phids Gobii dae Merlu ccius spp.

–

–

0·7

–

6·1

2·0

4·0

1·0

6·0

0·0

1·0

30·0

1·0

32·0

–

21·2

4·8

–

–

–

1·0

–

–

–

1·0

0·0

0·0

6·0

–

–

–

–

–

3·0

6·0

6·0

0·0

9·0

2·0

–

–

–

0·4

–

54·7

19·0

15·0

7·0

0·0

25·0

4·0

–

1982

34·0

15·0

–

11·8

15·1

30·6

–

–

–

–

–

–

–

–

–

–

–

0·7

31·6

33·7

–

–

6·0

1·0

18·0

42·0

64·0

20·0

81·0

6·4

THE BENGUELA ECOSYSTEM: PART IV

Specie Cape s hakes

Cape hakes

Length 28–29 150– of hake 240 (mm)

250– 350

240– 440

360– 580

M. capen sis

M. capen sis

M. capen sis

M. parad oxus

<400

<400

<400

<500

381

M. capens is

Length unspe measur cified ed

unspe cified

total

cauda l

total

total

Period of study

1985

1942– 1944

1946– 1947

1972– 1973

1974– 1975

ICSEA 1.3–1. F 4 Divisio n(s)

1.3

1.6

1.6

1.6

2.1–2. 2

Metho d used

% by mass

% by mass

% frequency occurrence (empty incl.)

% freque ncy occurr ence (full only)

% by mass

% by numbe r

Sampl e size

30

Fish other than above Etrum eus white headi Sardi nops ocella tus Engra ulis japoni cus Trach urus spp. Coelo rhync hus spp.

–

1982

60

98

27

18

50

346

139

317

121

215

278

201

unspec ified

–

–

–

–

–

–

–

–

–

31·0

5·0

–

2·0

–

–

–

–

–

–

–

–

–

–

–

–

–

–

12·3

–

–

5·4

–

–

–

–

–

–

–

–

–

–

9·6

–

–

–

32·1

–

–

1·0

–

3·0

–

11·0

–

–

6·4

–

–

–

–

–

–

4·0

1·0

3·0

–

2·0

–

2·0

–

382

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Specie Cape s hakes

Cape hakes

Length 28–29 150– of hake 240 (mm)

250– 350

240– 440

360– 580

M. capen sis

M. capen sis

M. capen sis

M. parad oxus

<400

<400

<400

<500

M. capens is

Length unspe measur cified ed

unspe cified

total

cauda l

total

total

Period of study

1985

1942– 1944

1946– 1947

1972– 1973

1974– 1975

ICSEA 1.3–1. F 4 Divisio n(s)

1.3

1.6

1.6

1.6

2.1–2. 2

Metho d used

% by mass

% by mass

% frequency occurrence (empty incl.)

% freque ncy occurr ence (full only)

% by mass

% by numbe r

Sampl e size

30

60

98

27

18

50

346

139

317

121

215

278

201

unspec ified

Fish other than above Fish remai ns Refer ence

–

–

–

–

16·7

–

1·0

1·0

1·0

3·0

15·0

3·0

5·0

17·5

–

1·3

13·8

–

–

–

2·0

11·0

14·0

–

–

–

–

1·8

1982

Andronov (1983)

Konchina (1986)

Rattray (1947)

Davies (1949)

Botha (1980)

Hecht (1976 )

importance than in the 1940s. Large M. paradoxus fed mainly on hakes, and those smaller than 500 mm on hakes, cephalopods and myctophids (Botha, 1980). Between 1982 and 1985 lanternfish and lightfish formed more than 50% by mass of the food of M. capensis having total length between 150 and 440 mm and more than 70% of the food of M. paradoxus of total length between 150 and 340 mm (Prosch, 1986). Lightfish were especially important in the diet of the deep-water M. paradoxus. On the Agulhas Bank Hecht (1976) reported large numbers of cephalopods in the diet of M. capensis in the mid-1970s. In contrast to Namibia and the Western Cape, pilchard and anchovy were also eaten and hakes were of lesser importance. Botha (1980) concluded that at sexual maturity both species of hake change from a largely crustacean diet to an almost exclusively piscivorous one and regarded the hakes as opportunistic feeders. Feeding intensity is apparently reduced during the spawning period. Cape fur seals feed extensively on hakes (Rand,

THE BENGUELA ECOSYSTEM: PART IV

383

1959a), and from 1980 to 1982 between Cape Town and the Orange River hakes formed about half their food (Sea Fisheries Research Institute, unpubl. records). Other predators include hakes themselves, snoek, kingklip (Macpherson, 1983a) and monkfish (Macpherson, 1985). Seabirds eat netted hakes and discarded offal (Burger & Cooper, 1984), but direct predation on hakes is probably small and limited to younger fish near the surface. Early acoustic surveys indicated that the annual natural mortality rate for Cape hakes was likely to be about 0·2, and initial assessments of hake stocks in the ICSEAF region used values between 0·2 and 0·3 (Newman, 1977). Examination of the relationship between fishing mortality and effort suggested that a value of 0·25 was appropriate for fish aged four to nine, although probably low for the oldest of these age classes (ICSEAF, 1976). More recent studies have, however, suggested that the natural mortality rate may be higher, values of 0·35 for Cape hakes (Terre, 1984), 0·42 for M. capensis, 0·50 for M. paradoxus (Assorov & Shcherbich, 1979) and 0·4 for M. paradoxus in ICSEAF Division 1.6 during 1972 and 1973 (Botha, 1980) being reported. Slower growth of males may result in a higher natural mortality rate for this sex than for females and, since hakes are cannibalistic, the natural mortality rate of young is likely to be influenced by the level of exploitation of larger individuals (Botha, 1980). Off southern Namibia between 1980 and 1982 it was estimated that kingklip accounted for almost one fifth of the natural mortality of hake aged one to six, more than 40% of the hake eaten by kingklip being three-year-olds (Macpherson, 1983a). During 1984 between 23 and 28° S it was estimated that almost 8% of the natural mortality of hake aged one to three was attributable to predation by monkfish (Macpherson, 1985). Kingklip and monkfish The dragonet Paracallionymus costatus was most important in the diet of kingklip sampled off the Western Cape in terms of frequency of occurrence in both 1942–1944 and 1946–1947 (Table XXXVIII), especially of fish smaller than 90 cm (Davies, 1949). Larger kingklip ate mostly hakes Merluccius spp. (Davies, 1949). Off Namibia during 1981 and 1982 kingklip less than 40 cm ate only TABLE XXXVIII Diet of the kingklip Genypterus capensis off the western Cape in 1942–1944 (Rattray, 1947) and 1946–1947 (summarized from Davies, 1949) expressed as percentage frequency of occurrence, and for two areas off Namibia in 1980– 1982 (summarized from Macpherson, 1983a) expressed as percentage mass for different size classes of kingklip Prey

Western

Western

Cape

Cape

24–30° S

1942–

1946–

Size of kingklip (mm)

1944

1947

30–39

40–69

– 5

– –

– –

11 5

9 0

6 –

36 –

8 –

13 –

– –

– 14

– 4

– –

– 13

Cephalopoda Todarodes angolensis Unidentified Stomatopoda Squilla armata Pterygosquilla armata Decapoda

Namibia

Namibia

19–24° S 40–69

384

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Prey

Western

Western

Cape

Cape

24–30° S

1942–

1946–

Size of kingklip (mm)

1944

1947

30–39

40–69

Plesionika acanthonotus Bathynectes piperitus Other Unidentified crustaceans Pisces Helicolenus dactylopterus H. maculatus Coelorhynchus fasciatus C. flabellispinus Trachyrhynchus trachyrhynchus Nezumia aequalis Macrouridae unidentified Clorophthalmus atlanticus Diaphus sp. Myctophidae unidentified Maurolicus sp. Sufflogobius bibarbatus Paracallionymus costatus Lepidopus caudatus Merluccius capensis M. paradoxus Tripterophycis gilchristi Unidentified fish Empty stomachs Number examined

– – – 8

– – 4 –

– 100 – –

– 9 3 –

– 4 2 –

6 8 12 –

0 7 – –

– – – – – – 14 – – – 6 – 15 – 8 – – 31

– 4 13 – – – – – – – – – 46 – 13 – 8 – 47 3

– – – – – – – – – – – – – – – – – – 1 339

– – – – – 1 8 – – 0 – 21 2 – 20 – – 6 228 1656

1 – 1 – 0 – 3 – – 0 – 2 0 3 68 0 – 1 1030 141

5 – 17 9 – 12 – 2 11 – – 7 – – – – – 5 87 127

2 – 28 – – – – – – – – 0 – – 0 – – 14 103

67

Namibia

Namibia

19–24° S 40–69

decapod crustaceans, whereas gobies, myctophids, and shrimps were frequently eaten by fish up to about 70 cm (Table XXXVIII). The frequency of larger prey items such as hake, ribbonfish Lepidopus caudatus, and the squid, Todarodes angolensis increased in larger kingklip (Macpherson, 1983a). Diet of kingklip to the north and south of 24° S was similar in terms of major taxa, but differed with respect to specific composition in accordance with the geographical distribution of prey species (Macpherson, 1983a). Small hake are more abundant south than north of 24° S (Macpherson & Allue, 1980). No seasonal or sex differences were noticed in the composition of the diet (Macpherson, 1983a). Kingklip is a nocturnal, benthic-feeder and its food consumption has been estimated at between 0·35 and 0·45% of body weight per day (Macpherson, 1983a). Annual food consumption is 0·9–2·3 times the body weight, the proportion being higher in smaller sized kingklip (Macpherson, 1983a).

THE BENGUELA ECOSYSTEM: PART IV

385

Fish, mostly Merluccius capensis, were the main prey of monkfish in 1984 between 23 and 28° S, with gobies, Coelorhynchus fasciatus, and cephalopods of the genus Todarodes also important food items (Macpherson, 1985). Panga and dentex On the Agulhas Bank panga feed frequently on shrimps and crabs, with echinoderms and fish also important in the diet. Other food items include squids, shrimps, mysids, amphipods and polychaetes (Hatanaka et al., 1983; Uozumi et al., 1984, 1985). Food of large-eye dentex has been discussed by Kuderskaya (1985). Juveniles feed primarily on small crustaceans such as euphausiids and copepods, but at a length of about 15 cm fish are commonly ingested. Adults prey mainly on demersal crabs, gobies, and myctophids. An annual natural mortality rate of 0·56 has been estimated for large-eye dentex (Kuderskaya, 1985). Tunas Information on the diets of five species of Thunnus caught in or near the Benguela region is summarized in Table XXXIX. By mass or volume fish were most important in all studies, except that reported for bigeye tuna by De Jaeger (1963) in which cephalopods predominated. Saury Scomberesox saurus was a major component in the diet of yellowfin tuna, with hakes and shoaling midwater or pelagic species being regularly preyed on by longfin, southern bluefin and bigeye tuna. On account of large-scale alternation of species in commercial catches it is likely that the composition of the diets has changed since the 1960s. Pilchard has probably decreased in importance, whereas hakes, trawler offal and, in the northern Benguela, gobies and horse mackerel may well have been eaten more intensively during the 1970s. De Jager, Nepgen & Van Wyk (1963a) listed four species of sharks (Glyphis glaucus, Isurus glaucus, Carcharinus obscurus, and Alopias vulpinus) and the killer whale (Orsinus orca) as the main predators of Benguela tunas, whereas Talbot & Penrith (1963) suggested that the principal predators would probably be the mako shark (Isurus glaucus), marlins Makaira spp. and the broadbill swordfish Xiphias gladius. Talbot & Penrith (1963) commented that large, healthy specimens of Thunnus maccoyii and T. thynnus would probably have few predators. Western coast rock lobster Although spiny lobsters in general have strong jaws which are adapted for crushing a wide variety of hardshelled organisms, the population of Jasus lalandii off the Western Cape is supported to a large extent by a single species of bivalve, the ribbed mussel Aulacomya ater. This species is the dominant mollusc in the rocky subtidal environment of the Western Cape and southern Namibia and comprises by far the greatest biomass of sessile organisms on the sea bed down to about 30 m and more (Field et al., 1980; Pollock & Beyers, 1981). In localities where mussel biomass is low, stomach content analyses of rock lobster have shown that they rely also on other sources of food, including echinoderms (especially sea urchins and starfish), gastropods,

386

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

TABLE XXXIX Summary of information on the diets of five species of tunas from the Benguela and adjacent regions Species

Area

Numbers of samples analysed

Diets inferred from Reference stomach contents. (Percentages refer to contents by mass or volume and to frequency of occurrence respectively. When only one percentage is listed it refers to contents by mass or volume)

Thunnus alalunga (longfin tuna, albacore)

Southern Benguela (south of 32° S)

918

Southern Benguela (south of 32° S)

68

Southern Benguela (south of 32° S)

449

fish 40%: 50% (Maurolicus muelleri, Lepidopus caudatus juv., Helicolenus maculatus juv., Merluccius capensis) cephalopods 39%: 48% (Loligo reynaudii, Abralia gilchristi, Histioteuthis bonnellii, Argonauta spp. and others) amphipods 8%: 67% (Phrosina semilunata, Phronima sedentaria) prawns 5%: 12% (Funchalia woodwardi) fish 55%: 54% (myctophids, Sardinops ocellatus, Engraulis japonicus) cephalopods 21%: 34% crustaceans 22%: 44% (Phronima sp.) fish 56%: 42% (myctophids, Scomberesox saurus, Sardinops ocellatus) cephalopods 21%: 31% crustaceans 22%: 58% (amphipods)

Penrith (1963), Talbot & Penrith (1963)

De Jaeger (1963), De Jager, Nepgen & Van Wyk (1963a, c)

Nepgen (1970a)

THE BENGUELA ECOSYSTEM: PART IV

Species

Thunnus albacares (yellowfin tuna)

Area

Numbers of samples analysed

Diets inferred from Reference stomach contents. (Percentages refer to contents by mass or volume and to frequency of occurrence respectively. When only one percentage is listed it refers to contents by mass or volume)

Northern Benguela (21–24° S)

178

Southern Benguela (south of 32° S)

168

fish 37% (Trachurus trachurus, Scomber japonicus, Hemirihamphidae spp., Exocoetidae spp., Myctophidae spp., Brama raii, Sardinops ocellatus, Engraulis japonicus) cephalopods 32% crustaceans 28% fish 77%: 73% (Scomberesox saurus was main forage species, followed by juvenile bramids, Merluccius spp., Lepidopus caudatus and Alepisaurus ferox) cephalopods 10%: 34% (Loligo reynaudii, Argonauta spp. and others) prawns 11%: 13% (Funchalia woodwardi) megalopa larvae 1%: 20%

Southern Benguela (south of 32° S)

26

Southern Benguela (south of 32° S)

67

fish 85%: 65% cephalopods 7%: 19% crustaceans 4%: 23% (megalopa larvae) fish 90%: 77% (Scomberesox

387

Van den Berg & Matthews (1969)

Penrith (1963), Talbot & Penrith (1963)

De Jaeger (1963), De Jager, Nepgen & Van Wyk (1963a, c)

Nepgen (1970a)

388

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Thunnus maccoyii (southern bluefin tuna)

Northern Benguela (21–24° S)

21

East of Southern Benguela (34–35° S: 25–26° E)

485

Southern Benguela (south of 32° S)

263

Southern Benguela (south of 32° S)

95

saurus, myctophids, Scomber japonicus, Auxis thazard) cephalopods 5%: 23% crustaceans 5%: 39% fish 91% cephalopods 5% crustaceans 4% fish 60% (Etrumeus whiteheadi, Sardinops ocellatus, Engraulis japonicus, Scomberesox saurus) molluscs: typically 25% in fish caught offshore (Lycoteuthis diadema, Loligo reynaudii) crustaceans: megalopa larvae numerically abundant fish 64%: 28% (Merluccius spp., Lepidopus caudatus, Trachurus trachurus, Maurolicus muelleri) cephalopods 4%: 14% (Histioteuthis bonnellii, Loligo reynaudii and others) prawns 30%: 37% (Funchalia woodwardi) fish 40%: 12% (Sardinops ocellatus, Merluccius spp.) cephalopods 40%: 18% crustaceans 20%: 24% (prawns)

Van den Berg & Matthews (1969) Smale (1983)

Penrith (1963), Talbot & Penrith (1963)

47% of stomachs were empty

De Jaeger (1963), De Jager, Nepgen & Van Wyk (1963a, c)

THE BENGUELA ECOSYSTEM: PART IV

Thunnus obesus (bigeye tuna)

Southern Benguela (south of 32° S)

357

Northern Benguela (21–24° S) Southern Benguela (south of 32° S)

14 102

389

Nepgen (1970a)

fish 58%: 42% (myctophids, Scomberesox saurus, Sardinops ocellatus, Merluccius spp.) cephalopods 20%: 31% (Loligo reynaudii) crustaceans 16%: 49% (prawns) fish 86% cephalopods 7% fish 51%: 52% (Merluccius spp., Lepidopus caudatus and larger myctophids) (Maurolicus muelleri absent) cephalopods 39%: 52% (Histioteuthis bonnellii, Loligo reynaudii, Argonauta spp. and others) prawns 10%: 32% (Funchalia woodwardi)

Van den Berg & Matthews (1963) Penrith (1963), Talbot & Penrith (1963)

TABLE XXXIX-continued Species

Area

Numbers of samples analysed

Diets inferred from Reference stomach contents. (Percentages refer to contents by mass or volume and to frequency of occurrence respectively. When only one percentage is listed it refers to contents by mass or volume)

Southern Benguela (south of 32° S)

11

fish 26%: 27% (Sardinops ocellatus, Engraulis japonicus) cephalopods 56%: 73% crustaceans 18%: 18% (prawns)

De Jaeger (1963), De Jager, Nepgen & Van Wyk (1963a, c)

390

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Species

Thunnus thynnus (bluefin tuna) Katsuwonus pelamis (skipjack tuna)

Area

Numbers of samples analysed

Diets inferred from Reference stomach contents. (Percentages refer to contents by mass or volume and to frequency of occurrence respectively. When only one percentage is listed it refers to contents by mass or volume)

Souther Benguela (south of 32° S)

83

Northern Benguela (21–24° S)

54

Southern Benguela (south of 32° S) East of Southern Benguela (close inshore)

2

fish 72%: 38% (Merluccius spp., Scomberesox saurus, myctophids) cephalopods 21%: 29% crustaceans 7%: 24% fish 72% cephalopods 24% crustaceans 3% stomachs empty

(34–35° S: 25–26° E offshore)

72

288

fish, approx. 90% (mainly Sardinops ocellatus, with lesser amounts of Engraulis japonicus and Etrumeus whiteheadi) molluscs (less than 3% by mass) crustaceans (less than 4% by mass, but megalopa larvae numerically abundant) fish 65% for skipjack longer than 50 cm (mainly Etrumeus whiteheadi and Sardinops ocellatus with lesser amounts of Scomberesox saurus) molluscs (Teuthoidea, Loligo reynaudii, Lycoteuthis diadema important for fish

Nepgen (1970a)

Van den Berg & Matthews (1969) Penrith (1963) Smale (1983)

THE BENGUELA ECOSYSTEM: PART IV

Species

Area

Numbers of samples analysed

391

Diets inferred from Reference stomach contents. (Percentages refer to contents by mass or volume and to frequency of occurrence respectively. When only one percentage is listed it refers to contents by mass or volume) smaller than 50 cm (80% by mass) crustaceans: euphausiids numerically most important food item in fish smaller than 50 cm

bryozoans, and polychaetes. Even seaweeds may be ingested from time to time in some areas. Studies by Pollock (1979) and Griffiths & Seiderer (1980) have shown size preferences in feeding of rock lobsters on mussels, with both large and small rock lobsters feeding preferentially on mussels well below the maximum size which they are capable of cracking open. Benthic surveys have revealed that rock lobster growth rates tend to be high in areas where mussel biomass is high (Newman & Pollock, 1974; Pollock, 1979; Pollock & Beyers, 1981). Although a high proportion of the mussel biomass is comprised of large individuals which are relatively immune to predation by virtue of their size and strong attachment to rocks in dense clumps, it appears that areas of high mussel biomass are areas in which recruitment and production of small mussels are also relatively high. The rate of production of small mussels is probably one of the most important factors affecting rock lobster growth rates, especially in areas where rock lobster stock densities are fairly high. In areas of low mussel biomass where rock lobster growth rates are retarded (e.g. parts of the Cape Peninsula), it would appear that food sources other than mussels are either not available in densities high enough to provide an alternative optimal diet, or that other food sources are comparatively inferior in quality (calorific content). In certain shallow-water areas where rocks are interspersed with sand, a second species of mussel, Choromytilus meridionalis, sometimes occurs in high densities. These comparatively thin-shelled mussels are rarely found in the sand-free rocky subtidal areas most favoured by Jasus lalandii, possibly because they are preyed upon preferentially by a number of predators, even when still at a small size. When rock lobsters do gain access to areas in which large Choromytilus meridionalis are common, these mussels become an important food source. Even small rock lobsters are capable of cracking open fairly large individuals of this mussel species, because of its relatively thin shell (Griffiths & Seiderer, 1980). Aulacomya ater is a slow growing mussel, large individuals of which are fairly resistant to predation by virtue of their shell strength (ribbing), whereas Choromytilus meridionalis is much faster growing, thinshelled and less resistant to predation. C. meridionalis appears, however, able to tolerate much sand movement and even burial for extended periods, and therefore survives well in sandy areas, or in places which are isolated from benthic predators such as on moored surface buoys or rafts.

392

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Heydorn (1969a) found that octopus and dogsharks feed on rock lobsters, although these predators are relatively scarce in comparison with rock lobsters off the Western Cape. Hagfish, Heptatretus hexatrema, and whelks, Burnupena papyracea, often attack damaged or injured rock lobsters in the kelp beds. Although adult seals, Arctocephalus pusillus, rarely feed on rock lobsters (Rand, 1959a), small Jasus lalandii are common in the gut contents of pups under a year old (J.David, pers. comm.). Cannibalism is known to be prevalent in overcrowded situations, especially amongst juvenile rock lobsters, but predation on large juveniles and adults is considered to be low, because of the scarcity of large predators in nearshore waters. Low rates of predation are reflected in the behaviour of adult J. lalandii, which regularly feed during daylight hours in open areas, and are very easily captured by hand. This type of behaviour contrasts strongly with that of spiny lobster species from warmer waters where predators are more abundant. For example, virtually all species of Panulirus tend to seek shelter in holes, caves and crevices during the day, only leaving their shelters at night to forage in surrounding areas (Heydorn, 1969b; Herrnkind, 1980). These animals are also very adept at escaping from predators, and are correspondingly difficult to capture by hand. As Heydorn (1969a) noted, the prolific production of larvae by Jasus lalandii suggests that the mortality rate to which this species is subjected before attaining maturity must be high. By far the greatest portion of mortality is held to take place during the lengthy larval phase. The phyllosoma larvae are teleplanic, that is “long-distance” larvae (Scheltema, 1971), and are probably capable of delaying metamorphosis into the puerulus stage until a familiar set of conditions is encountered which acts as a cue for metamorphosis. Critical periods for larval mortality are likely to occur shortly after hatching, when densities are very high and a wide spectrum of coastal planktonic and nektonic predators abound, and again much later when the puerulus stage ‘runs the gauntlet’ of coastal predators as it returns across the continental shelf to settle in the kelp beds. Complete transparency of the phyllosoma and puerulus is a prerequisite for survival and, in order to remain invisible to predators, an ‘invisible’ diet is advantageous. In this connection, it is interesting to note that phyllosoma larvae of a number of scyllarids have been observed clinging to medusae of hydrozoan and scyphozoan jellyfish, and it has been suggested that food particles are not normally found in the gut contents of phyllosoma larvae caught in the plankton because their food may be in relatively liquid form such as, for example, could be obtained from soft-bodied hydromedusae (Phillips & Sastry, 1980). Red crab Known predators on red crab include skates (Macpherson, 1985) and some deep-sea fish, such as Cottuncoloides macrocephalus and Alepocephalus rostratus (Macpherson, 1983b). Squid White squid are active predators, feeding mainly on crustaceans and fish. A marked decline in catch rates of squid in bottom trawls during the night indicates that squid leave the near sea-bed zone in hours of darkness to feed at higher levels in the water column (Augustyn, in press). Hakes and myctophids are important fish species in their diets. Predators on squid include marine mammals, such as seals, dolphins and toothed whales, fish and seabirds, such as jackass penguins, cape cormorants and cape gannets (Rand, 1959b, 1960a, b).

THE BENGUELA ECOSYSTEM: PART IV

393

Abalone Abalone feed on seaweeds, the preferred diet being fronds of the kelps Ecklonia maxima and possibly Laminaria pallida (Barkai & Griffiths, 1986). Where these kelps do not occur, as for example in the eastern Cape, smaller varieties of algae are consumed, such as species of Hypnea, Plocamium, and Dictyoya (Newman, 1969a). Movement of abalone is limited (Newman, 1969b), and most animals feed on driftalgae, often detached kelp fronds, which they trap beneath their feet. The rock surfaces where abalone occur in abundance are frequently devoid of algal growth, the result of grazing activity (R.J.Q.Tarr, pers. comm.). Large abalone have virtually no predators other than man, although numbers of adults are known to be lost occasionally as a result of storm damage and the ‘sanding-over’ of reefs. Very small abalone are thought to have a high natural mortality rate, the result of predation by octopuses, predatory reef fish, rock crabs, starfish, and molluscs. Mortality rates are thus likely to decrease with an increasing size and age of individuals. DISTRIBUTIONAL ECOLOGY AND REPRODUCTION In keeping with marked seasonal patterns in the oceanic environment (Shannon, 1985; Shelton, Boyd & Armstrong, 1985) distinct seasonal cycles characterize the distribution, spawning, and recruitment of many fish species in the Benguela system and their availability to fishing fleets (e.g. Crawford, 1980a). This section summarizes patterns for the more important fish resources. The spawning grounds and seasons of many fish species in the Benguela and factors influencing development of their ichthyoplankton have been reviewed by Shannon & Pillar (1986). Therefore these aspects are only considered briefly here insofar as they bear upon the overall distributional ecology of the fish populations discussed. The environment The essential features of the Benguela environment are summarized briefly in the following paragraphs. Readers are referred to other articles in this review series (Shannon, 1985; Chapman & Shannon, 1985; Shannon & Pillar, 1986) for a more comprehensive discussion. Except where indicated the summary is based on information in these three references. The Benguela is unusual as an upwelling system, in that, because the African continent terminates at a relatively low latitude, the region is bounded at both equatorward and poleward ends by warm water regimes. The northern and southern boundaries lie around 16° S and in the vicinity of the Agulhas BankAgulhas Retroflection area, respectively. These boundaries shift seasonally, and interactions between the Benguela and the Angola system in the north and the Agulhas Current in the south extend over considerable distances, with the impact of the surface waters from these two warm currents on the Benguela being most noticeable during the first quarter of the year. The western boundary of the Benguela is best considered as fairly open ended. The prevailing winds are upwelling favourable throughout most of the year in the areas north of 31° S, and are seasonally upwelling favourable to the south of this latitude. In the south reversals in the winds with periods of between a few and about 12 days occur during the main upwelling season (September-March) because of the passage of eastward moving cyclones to the south of the continent. In the northern Benguela, winds are most strongly upwelling favourable during autumn and late winter to early spring, and least so during the first quarter, particularly in January. It is during this time of the year that water from off Angola may move farthest south (to about 20 or 21° S). Areas of strongest longshore equatorward winds exists off Lüderitz in the central Benguela and near Cape Frio (18° S) during most seasons. (At Lüderitz winds are

394

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

somewhat less strong during the second quarter.) Because of the topography and the winds, Lüderitz is the major upwelling centre in the Benguela, with a negative surface temperature anomaly of about 6°C. The Lüderitz zone effectively divides the Benguela into two and can be viewed as a major potential environmental barrier within the system, with seasons of maximum upwelling to the north and to the south being out of phase with each other. Currents over the continental shelf are variable in time and space, and satellite imagery of the Benguela suggests that the shelf circulation is dominated by a complex system of changing eddies. Surface currents are mainly determined by the wind direction and tend to flow towards the equator, but at subsurface depths the flow is on average poleward. Periodic current reversals occur which may be indicative of shelf waves. Major reversals having a longer period appear to provide a mechanism for flushing the St Helena Bay area (G.Nelson, pers. comm.). Stratification is strongest in summer in the extreme north of the Benguela, over the Agulhas Bank and west of the oceanic front, the position of which approximates to that of the shelfbreak. The shelf in parts of the Benguela is deep, and double shelf-breaks are common. Relatively shallow (upper) shelves exist in St Helena Bay and in the vicinity of Walvis Bay, and in these regions wind speeds are on average lower than those further north and south. The part of the Agulhas Bank east of 21° E is shallower than the western bank, and winter storms can mix the entire water column. As a result nutrient concentrations in the surface layer over the Agulhas Bank, excluding the nearshore region west of 20° E, are highest during winter, whereas along the west coast nutrient supply is at a maximum when upwelling is at a maximum. New nutrients are evidently imported into the southern Benguela, whereas regenerated nutrients become more important in semi-enclosed areas, such as St Helena Bay and in the northern region. The limiting nutrient is nitrogen although there is some suggestion that at times silica may also be limiting in the northern Benguela. Available literature indicates that diatoms and, to a lesser degree, dinoflagellates dominate the phytoplankton throughout the Benguela as far north as southern Angola, but it is possible that this apparent dominance may be an artefact of the sampling and examination techniques. Recent studies (largely unpublished) indicate that nanoplankton may be quite important in the system. The Benguela is floristically undistinguished, and the majority of the species also occur in the Mediterranean Sea and in the southwestern Indian Ocean. There are, however, differences in the assemblages of species between northern and southern Benguela. Skeletonema costatum is relatively uncommon off Namibia whereas Delphineis karstenii is restricted to the northern Benguela. Copepods are numerically the most abundant and diverse zooplankton group in the Benguela. Off Namibia the main zooplankton biomass occurs offshore of the peak phytoplankton biomass. Off the Cape Peninsula the zooplankton biomass is at a maximum during summer and at a minimum in winter. Paradoxically the zooplankton standing stock over the Agulhas Bank is higher than off the southern west coast in spite of lower nutrients and lower phytoplankton abundance over the Bank. This is primarily due to the contribution of pelagic tunicates. In the St Helena Bay area the relatively high standing stock of zooplankton displays little or no seasonality. Oxygen-depleted water can occur around much of the coast of South Africa and Namibia. The main areas of formation are along the Namibian shelf north of Lüderitz and the productive (offshore) region of the Angola Dome. There are also smaller localized centres of formation north of Hondeklip Bay and in St Helena Bay. In all these areas oxygen concentrations below 50 mmol/m3 (44·6 mmol/m3 =1 ml/1) are found. The areas of oxygen depletion are to the north of the major upwelling areas along the coast and result from the decomposition of material derived from the highly productive euphotic zone. Water depleted in oxygen can move southwards from the Angola Dome area at 200–400 m depth along the shelf edge and is at its most southerly position in late summer-autumn. The apparent southward limit of the low-oxygen undercurrent off Namibia is about 28° S. Further south, low-oxygen water formed near St Helena Bay may

THE BENGUELA ECOSYSTEM: PART IV

395

be advected southwards as far as Cape Point and possibly into Walker Bay. There is also local depletion on the Agulhas Bank although, because of the reduced productivity and greater turbulence here compared with the western coast, minimum values are about 75–100 mmol/m3 and the area affected is smaller. The widest extent of the low-oxygen mass is again in autumn. Several warm and cool periods in the Benguela region have been documented. A generally warm period spanned the late 1950s and early 1960s and a second warm period occurred between 1972 and 1977, but a substantial cool period began in 1978 and extended over much of the Benguela until 1983 (Shelton et al., 1985). Localized warming was experienced in the extreme southern Benguela in the summer of 1982–1983 because of reduced upwelling-favourable wind stress in the area. Only three major perturbations have been documented in the Benguela, which appear rather like an Atlantic equivalent of the El Niño although not in phase with Pacific events, viz. in 1934, 1963 and 1984. All three Benguela Niños were clearly related to changes in the equatorial Atlantic (Shannon, Boyd, Brundrit & Tanton-Clark, 1986). These local events are less intense and less frequent than their Pacific counterparts. In the Atlantic the seasonal signal is stronger than in the Pacific, but the interannual one is weaker. The major resources Pilchard and anchovy. Nought-year-old pilchard recruit along South Africa’s western coast from March onwards. In winter, maximum standing stocks of plankton occur north of Cape Columbine (Shannon, Hutchings, Bailey & Shelton, 1984), and it is in this region that catch rates of juvenile pilchard are initially highest (Crawford, 1980a). With the advent of spring and the spawning season, a southerly migration of young fish to warmer waters east of Cape Point is initiated (Fig. 14). In this region availability of mainly one-year-old pilchard to purse-seine boats is highest from January to March. Then in autumn the shoals migrate towards the east (Crawford, 1980a). One-year-olds may move as far east as the Natal coast (Armstrong, Shelton & Prosch, 1985; Thomas, 1985), where pilchard are seasonally available in winter (Baird, 1971). East of Cape Agulhas, growth of phytophlankton is frequently restricted to the region of thermocline, and pilchard often occur below the surface (Crawford, Shelton & Hutchings, 1980). Pilchard aged two or older (Thomas, 1985) tend to return to the waters west of Cape Point, where they were historically available to purse-seiners from February until May or June when they moved offshore to spawn (Crawford, 1980a). The main spawning areas have been over the Agulhas Bank and northwest of Cape Columbine, with peaks in spring and summer (Crawford, 1980a; Davies, Newman & Shelton, 1981). Spawning off Cape Columbine decreased following introduction of the small-mesh net and a consequent reduction in the age structure of the population (Crawford, 1980a). Distribution of pilchard off Namibia has been described by various authors (Matthews, 1964; Schülein, 1971, 1973; Crawford, Shelton & Hutchings, 1983; Shelton et al., 1985; Thomas, 1985) and is illustrated diagrammatically in Figure 14. South-moving fish aged six to ten months recruit to the fishery off Walvis Bay between March and August. In August a northward inshore migration of these juveniles is initiated, and one-year-old pilchard may move as far north as southern Angola (Thomas, 1985). From north to south there is a progressive increase in the length of older fish (Schülein, 1971; Cram, 1977), and the largest size classes were seasonally available to purse-seine boats near Walvis Bay between November and January immediately after spawning (Matthews, 1964). Adults move south to spawn and there have been two main centres for spawning: the Walvis Bay vicinity (23° S) where most eggs were produced between mid-spring and early summer in sea-surface temperatures ranging between 12·0 and 16·5 °C, and further north in the Dune Point region (19–21° S), where spawning attained a maximum between mid-summer and early autumn in temperatures between 16·5 and 22·8 °C (Stander, 1963; Matthews, 1964; Schülein, 1971; King,

396

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Fig. 14.—Schematic illustration of major spawning grounds and movements of pilchard Sardinops ocellatus and anchovy Engraulis japonicus off southern Africa (from information in Crawford, 1980a; Shannon, 1985; Thomas, 1985).

1977b; Le Clus & Thomas, 1982). Interannual variability in the onset of spawning has been noted (Le Clus, 1979b). Production of eggs in the Walvis Bay region declined substantially following the collapse of the pilchard stock in the early 1970s and a reduced age structure in the population, and it is likely that the heavy spawning in this region was by older adults. Spawning still occurs in the warmer waters off Dune Point. Average oocyte mass of winter and spring spawners is significantly greater than that of summer and autumn spawners, perhaps as an adaptation to the lower temperatures encountered by spawners in winter and spring (Le Clus, 1979b). Blaxter & Hunter (1982) suggest that for multiple spawners a seasonal decline in egg size can be attributed to a reduction in energy reserves over the spawning season, a change in the partitioning of energy between growth and reproduction, or a seasonal change in the age structure of the spawners. For the pilchard off Namibia oocyte mass is apparently not related to size of fish (Le Clus, 1979b). Dispersal of ichthyoplankton may be in a northerly direction (O’Toole, 1977a), although Thomas (1985) suggested that an intermittent south-moving current could cause a southward drift of larvae towards the Walvis Bay grounds where the young fish recruit to the fishery. Onshore and poleward flow within the upper 50 m in the northern spawning area is most pronounced during the first quarter of the year and may facilitate retention of eggs and larvae within the northern Benguela region (Shannon, Agenbag & Buys, 1987). The spawning grounds, distribution, and movements of anchovy are in many respects similar to those of pilchard (Crawford, 1980a; Crawford, Shelton & Hutchings, 1983) although anchovy, not being as long lived, has had only one major spawning ground in each of the northern and southern regions (Shannon &

THE BENGUELA ECOSYSTEM: PART IV

397

Pillar, 1986). Juvenile pilchard and anchovy frequently co-occur in mixed shoals together with horse mackerel and round herring (Crawford, 1980a). Young anchovy recruit to the fishery along South Africa’s western coast from March, being most readily available to the purse-seine fleet in June (Crawford, 1980a). From then until September they are caught in large quantities. The distribution of young anchovy from the stock off the Western Cape may extend to the north of the Orange River (Cruickshank, 1984). Similarly to pilchard, young anchovy move southwards at the conclusion of winter, are available to purse-seiners east of Cape Point in the first quarter of the year, and thereafter frequently migrate to the east (Fig. 14). In November 1984, the size of anchovy taken in midwater trawl on the Agulhas Bank increased eastward and with distance offshore. Highest densities of fish were recorded some distance offshore in the area west of Mossel Bay (Hampton, Shelton & Armstrong, 1985). The major spawning grounds are east of Cape Point, with dispersal of spawning products towards the north and west. Accelerated transport has been recorded in a frontal jet off Cape Town (Shelton & Hutchings, 1982). Off Namibia large shoals of juvenile anchovy recruit to the fishery in the vicinity of Walvis Bay (Fig. 14), usually from May with greatest availability occurring in August, whereas adults are caught inshore north of Cape Cross from autumn to early winter (Le Clus & Melo, 1982; Thomas, Hewitson & Böhme, 1982; Schülein, 1986). Adults spawn inshore between the Cunene River and Mercury Island from late spring to early autumn, highest concentrations of eggs being recorded off Dune Point in late summer (King, 1977b; Le Clus & Kruger, 1982; Le Clus & Melo, 1982; Shannon & Pillar, 1986). Serial spawning is believed to occur and large adults are more fecund than smaller individuals, those of caudal lengths smaller and larger than 115 mm producing, respectively, averages of 3200 and 6300 last-mode oocytes per fish (Le Clus, 1979a). Fifty per cent maturity occurs at 95 mm (Le Clus, 1979a). A major warm event was recorded in the northern Benguela region in 1963 and resulted in greatly reduced oil yield from, and gonad mass of, pilchard (Stander & De Decker, 1969). Southerly displacement of pilchard shoals is believed to have resulted (Butterworth, 1983). Temperature and salinity relationships (Stander & De Decker, 1969) and mean annual temperatures (Shelton et al., 1985) near Walvis Bay indicate a progressive warming in the region between 1959 and 1963, during which period there was also a steady decrease in measured egg production of pilchard (Fig. 15, Stander & De Decker, 1969). Pronounced warming in the northern Benguela also occurred in 1984. It concentrated pilchard close to Walvis Bay, leading to high catch rates of this species and reduced spawning by anchovy (Boyd & Thomas, 1984; Le Clus, 1985). Only two Benguela “Niños” have been documented during the past half-century, viz. 1963 and 1984 (Shannon et al., 1986). The low frequency of major warm perturbations in the Benguela could mean that fish species are less adapted to their occurrence than species in the Pacific where warm events happen more often. Similarities in the distributions of pilchard and anchovy in the northern and southern regions of the Benguela, in many respects “mirror images” of each other (Thomas, 1985), have been noted by a number of workers. For pilchard two spawning areas existed in each region: one attributable to older spawners in close proximity to upwelling, the other for younger adults in warmer water showing greater stratification and closer to the boundaries of the Benguela (Shannon & Pillar, 1986). In both systems depletion of the pilchard resources at some stage resulted in fewer older fish, a consequent contraction in the distribution of adults towards warmer waters and decreased spawning near the upwelling centres (Crawford, Shelton & Hutchings, 1983). Off both South Africa and Namibia anchovy have a tighter spawning season than pilchard and spawn in similar regions to those of the younger pilchard adults. Juveniles of the two species recruit to the fisheries in autumn and winter near upwelling centres, where feeding occurs prior to a spring migration to spawning grounds (Shelton et al., 1985), and one-year-old pilchard undertake extensive

398

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Fig. 15.—Trends in the production of pilchard Sardinops ocellatus eggs off Namibia, 1959–1963 (from Stander & De Decker, 1969).

migrations beyond the boundaries of the Benguela (Thomas, 1985). Sites of strong offshore transport are avoided as spawning areas (Shelton et al., 1985) Sardinellas. The distribution of sardinellas off Angola has been discussed by Baptista (1977), who commented that Sardinella aurita is the dominant sardinella off southern Angola but S. eba off Pointe Noire, Congo. Baptista (1977) described a seasonal migration of adult sardinellas (Fig. 16), which he related to seasonal trends in sea-surface temperatures and spawning. Spawning is initiated between the Zaire River and Pointe Noire in February and attains a maximum in March and April. Juveniles apparently remain off Pointe Noire and northern Angola, but adults migrate south and occur off central Angola from November to February. Availability to coastal purse-seiners decreases in December and January, perhaps as a result of a movement offshore (Baptista, 1977). Horse mackerel. Yields of Cape horse mackerel used to be high off South Africa’s Western Cape in the 1950s and early 1960s, but have since been small. The pattern then was for young of the year to recruit to the purse-seine fishery from May onwards, especially in St Helena Bay (Fig. 17). Larger juveniles were caught in summer and adults from late summer through winter, their scarcity in other seasons probably being linked to the reproductive cycle (Crawford, 1980a). Spawning occurs throughout the year in fairly deep water offshore, but is maximal in spring (Shannon & Pillar, 1986) after most gonads have matured during the winter months (Geldenhuys, 1973).

THE BENGUELA ECOSYSTEM: PART IV

399

Fig. 16.—Schematic illustration of the movements of sardinellas Sardinella spp. and snoek Thyrsites atun in the southeastern Atlantic: important areas of purse-seine fishing are also shown, and in the inset aspects of the movements of some species preyed on by snoek (from Crawford & De Villiers, 1985).

On the Agulhas Bank Cape horse mackerel of less than 20 cm total length occur in inshore waters between Cape Agulhas and Mossel Bay (Uozumi et al., 1985). Larger fish tend to occur further offshore and are also distributed as far east as Algoa Bay (Hatanaka et al., 1983; Uozumi et al., 1984, 1985). Catches in ICSEAF Division 2.2 have generally comprised fish larger than those of catches made in Division 2.1 (Fig. 18; Kinloch et al., 1986). The best catches on the Agulhas Bank have been made from July through

400

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Fig. 17.—Schematic illustration of major spawning and nursery grounds of Cape horse mackerel Trachurus capensis and Cape hakes Merluccius capensis and M. paradoxus in the southeastern Atlantic (from information in Macpherson et al., 1982, 1985, 1986; Payne, Augustyn & Leslie, 1985, 1986; Shannon, 1985).

October (Crawford & De Villiers, 1984) after an inshore migration of the fishable stock (Uozumi et al., 1985). The fishery is conducted with bottom trawl and only fish aged four or older are usually selected (Kinloch et al., 1986). Maturity is initiated at a total length of 20 cm, although it is only at a length of between 30 and 40 cm that most fish become sexually active (Hatanaka et al., 1983; Uozumi et al., 1984, 1985).

THE BENGUELA ECOSYSTEM: PART IV

401

Fig. 18.—Total length distribution of catches of Cape horse mackerel Trachurus capensis in two ICSEAF Divisions on the Agulhas Bank (from Kinloch et al., 1986).

Off Namibia in research catches made between March and April 1981 using midwater trawl, Cape horse mackerel of length 10–20 cm were most abundant between 18 and 19° S, but ranged as far south as Walvis Bay. The smallest individuals occurred in association with large medusae. Fish between 20 and 40 cm occurred further offshore, but mostly within the 300 m isobath. The highest densities were to the north of 19° S, but the largest fish were between 20 and 21° S (Macpherson, Mombeck & Schülein, 1982). Konchina (1986) reported that Cape horse mackerel off Namibia mature at a length of over 20 cm. Juveniles were caught by the purse-seine fleet north of 22° S in 1981 and 1982, mostly between 19 and 21° S and in water shallower than 60 m. In January of the same two years, most shoals of horse mackerel detected on research surveys were to the north of 22° S (Cruickshank, 1983). In 1985 purse-seine catches in ICSEAF Division 1. 3 were made mainly in March and April and comprised fish with a modal total length of 12–13 cm. Progressively larger fish were caught in reasonable quantities as far south as 23° S, the best catches off Walvis Bay being made in August (Schülein, in prep.). Konchina (1986) found fish of about 16 cm off Walvis Bay in April and May 1985. Slightly larger individuals were found offshore and south to 25° S, but most adults occurred at 19°30′ S. The overall pattern off Namibia therefore appears to be of juvenile horse mackerel occurring in inshore waters, with the smallest sizes being found furthest north and slightly larger individuals migrating south

402

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

towards Walvis Bay, especially in winter. Maturing fish move offshore and adults generally occur north of 21° S, with the densest concentrations in ICSEAF Division 1.3. Adults spawn between spring and autumn, mostly in summer and in water temperatures between 16 and 19 °C (O’Toole, 1977b). Spawning occurs as far south as 23° S but is maximal over the 200 m isobath between 18 and 19° S, the Cape Frio vicinity (Shannon & Pillar, 1986). Larvae occur near the thermocline, with the majority of newly hatched individuals being near to the chlorophyll maximum layer (O’Toole, 1977b). Larval abundance is greatest 50–100 km offshore (O’Toole, 1977a; Olivar & Rubies, 1983). As with pilchard and anchovy there are similarities in the distribution of Cape horse mackerel in the northern and southern regions of the Benguela system. Adults tend to occur and to spawn near the warmer boundaries of the system, but juveniles to be distributed nearer the upwelling centres. De Campos Rosado (1974b) noted that catch rates of Cunene horse mackerel by Angolan purse-seiners were highest in October and November, and believed it probable that this resulted from an inshore migration for spawning. Round herring and saury. Round herring have contributed substantial catches to the purse-seine fleet off the Western Cape but much lesser amounts off Namibia. Young of the year recruit along South Africa’s western coast (Fig. 19) from April through September when they are frequently found in mixed catches together with juvenile anchovy, pilchard, and horse mackerel (Geldenhuys, 1978). They show a tendency to move southwards at the conclusion of winter (Crawford, 1980a). In November 1983 large numbers of round herring of 10–15 cm length, sizes that are poorly represented in the commercial catch in most years and thought to be one-year-olds, were caught in mid-water over the Agulhas Bank, particularly in the vicinity of Algoa Bay (Armstrong et al., 1985). Round herring presumably occur inshore between July and December when they are seasonally eaten by jackass penguins at St Croix Island in Algoa Bay (Randall, 1983). During 1965–1971 round herring attained 50% maturity at a caudal length of 17 cm, with all fish mature at 22 cm (Geldenhuys, 1978). Adult round herring are most available to purse-seine boats in the region between Cape Columbine and Quoin Point from mid-summer through autumn, availability decreasing rapidly thereafter (Crawford, 1980a). They may extend some distance beyond the oceanic front (Crawford, 1981b) and frequently co-occur with young chub mackerel in mixed shoals (Baird, 1978a; Geldenhuys, 1978). Spawning occurs offshore along the western coast of southern Africa and the western edge of the Agulhas Bank near the 200 m isobath, being maximal from August to October (Davies et al., 1981; Shannon & Pillar, 1986). Eggs can tolerate temperatures as low as 11 °C, but development is more rapid in warmer water (O’Toole & King, 1974). Off Namibia spawning occurs over the shelf-break in the vicinity of 20° S (Shannon & Pillar, 1986). Off the Western Cape young saury tend to occupy water of 18 °C or warmer in all seasons and occur offshore of the Benguela oceanic front during inshore upwelling. Spawning occurs mainly between Cape Columbine and Cape Point in summer (Dudley, Field & Shelton, 1985). The Benguela region, however, forms only a small part of a much larger saury spawning area in the south-eastern Atlantic Ocean (Dudnik, 1975, in Dudley et al., 1985). Saury of mean fork length 35 cm were important in the diet of Cape gannets at Bird Island, Algoa Bay, between 1978 and 1981, contributing on average almost 30% of the mass of food ingested (Batchelor, 1982), but saury were not eaten in large amounts by jackass penguins at St Croix Island, Algoa Bay, during 1979– 1981 (Randall, 1983). Jackass penguins, being flightless, have a considerably smaller foraging range than Cape gannets (Frost, Siegfried & Cooper, 1976), and saury may well be distributed too far offshore along the southern coast for breeding penguins to utilize effectively. Saury have contributed to the diet of Cape gannets off the Western Cape, especially from December to March (Berruti, in prep.), and off Namibia (Crawford et al., 1985).

THE BENGUELA ECOSYSTEM: PART IV

403

Fig. 19.—Schematic illustration of major spawning grounds and movements of round herring Etrumeus whiteheadi, of the main distribution and spawning grounds of the pelagic goby Sufflogobius bibarbatus in the southeastern Atlantic, and (inset) of the spawning grounds and location of purse-seine catches of lanternfish Lampanyctodes hectoris off South Africa’s Western Cape (from information in Crawford, 1980a; Cruickshank et al., 1980; Shannon, 1985; Prosch, 1986).

Pelagic goby. The pelagic goby occurs in inshore waters around the southern African coast from the Cunene River to St Sebastian Bay, with a break in distribution between Lüderitz and the Orange River (Cruickshank, Cooper & Hampton, 1980). Small individuals have a light colouration typical of pelagic fish, whereas larger gobies have a darker colouration more characteristic of fish species living close to the sea

404

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

bottom and, although caught by demersal trawl and consumed by bottom-feeding seabirds, are rarely eaten by seabirds foraging on the ocean surface (Crawford et al., 1985). In the 1970s there were substantial stocks between Walvis Bay and Lüderitz (Fig. 19; Cruickshank et al., 1980; Cruickshank, 1982), which formed an important food item for many fish, bird and mammalian predators (e.g. see Prenski, 1980; Crawford, Cruickshank, Shelton & Kruger, 1985). On various surveys the pelagic goby has contributed between 20 and 67% of larvae collected off Namibia (Shannon & Pillar, 1986). Spawning is most intense during spring and early summer in coastal waters south of the Walvis Bay bight, but larvae have been recorded north to Cape Frio (O’Toole, 1978b). Larvae near Walvis Bay were captured in relatively cold, low-salinity water typical of coastal upwelling; those further north were caught in areas of mixing between warm, oceanic water and cold, coastal water (O’Toole, 1978b). Large catches of eggs and larvae have not been reported from the southern Benguela (Shannon & Pillar, 1986). Lanternfish and lightfish. Lanternfish and lightfish are abundant in offshore waters along the western coast of southern Africa south of about 20° S (cf. Cruickshank, 1982; Prosch, 1986; Shannon & Pillar, 1986). Off the Western Cape they are preyed on extensively by larger fish between the 100 and 500 m isobaths (Porsch, 1986). Large catches of lanternfish have been made by purse-seine boats in the vicinity of Cape Columbine at the head of the Cape Canyon (Fig. 19), especially from January to March. Availability to fishermen decreases in autumn and catch rates are minimal in winter (Crawford, 1980b). It has been suggested that sub-surface currents moving up the Cape Canyon may be instrumental in bringing about the near-surface distribution of lanternfish off Cape Columbine (Crawford, 1980b). The seasonal availability of fish to the purse-seine fleet may be influenced by changes in currents or by the reproductive cycle. In the nonspawning season, when lanternfish are most readily available to purse-seiners, the ratio of females to males in catches is much higher than at other times (Prosch, 1986). Off Namibia research catches of lanternfish have mainly been made on the outer edge of the continental shelf more than 50 km offshore, although in the Lüderitz vicinity where the shelf narrows they have been recorded nearer the coast (Cruickshank, 1982). The majority of sexually active lightfish occur at depths greater than 90 m, and older age classes may be distributed further offshore than younger ones (Prosch, 1986). In the southern Benguela Lampanyctodes hectoris and Maurolicus muelleri spawn throughout the year; lanternfish mainly between late winter and mid spring, lightfish between winter and early summer (Prosch, 1986). Between August 1977 and August 1978 most spawning by both species occurred west of Cape Agulhas and offshore in well-mixed oceanic water (Shelton, 1979; Prosch, 1986). Eggs of lightfish occur at depths in the water column that make them relatively independent of surface currents (Shelton, 1979). Off Namibia the two species have been abundant in the ichthyoplankton in offshore waters south of 19 or 20° S, lanternfish spawning in winter and spring and eggs and larvae of lightfish being abundant in July (Shannon & Pillar, 1986). Sizes at 50% maturity are approximately 42 mm caudal length for lanternfish, 30 mm for male lightfish and 38 mm for female lightfish (Prosch, 1986). Substantially more females than males have been recorded in research and commercial catches of lanternfish and in research catches of lightfish (Crawford, 1980b; Prosch, 1986). For lanternfish Prosch (1986) recorded a mean fecundity of 1126 eggs per female (range 834–1354; n=6); for lightfish a mean fecundity of 359 eggs per female (range 24–738; n=9) or 222 eggs per g wet body weight. Snoek and chub mackerel. Snoek are caught along the southern African coast from Angola in the north to Algoa Bay in the south, but mainly in cool upwelled water between 15° S and Cape Agulhas (see Table XIX, p. 389). A well-defined migratory pattern is evident, the species occurring in the northern Benguela region during the austral spring and summer, moving south in autumn and winter, and returning

THE BENGUELA ECOSYSTEM: PART IV

405

north in spring (Fig. 16). The southward movement takes place near the coast, but that towards the north further offshore, and the migrations are related to seasonal patterns of availability of prey (Crawford & De Villiers, 1985). In particular the occurrence of snoek in the northern ICSEAF Divisions coincides both with a southward shift in the distribution of sardinellas (Baptista, 1977) and with availability in the region of shoals of pilchard and anchovy (Cruickshank, 1983; Thomas, 1985; Schülein, 1986). Further south, the appearance of snoek near Port Nolloth during April and May corresponds with recruitment to the Western Cape’s purse-seine fishery in the same region of shoals of juvenile pilchard and anchovy (Crawford, 1980a). Snoek apparently follow the southward migrations of these young fish to the waters east of Cape Point (Crawford & De Villiers, 1985). In winter snoek move away from the coast of the Western Cape to spawn (De Jager, 1955), a movement that corresponds with the offshore spawning migration of round herring on which they prey (Crawford, 1980a). Spawning by snoek occurs off the Western Cape and Namibia from July to October, large concentrations of eggs having been recorded off the mouth of the Orange River (Davies, 1954; De Jager, 1955; Nepgen, 1979b; Olivar, 1982, 1984; Olivar & Rubies, 1985). During the spawning season snoek have a low content of body oil and protein, and there is marked deterioration in their condition (Van Wyk, 1944; Nepgen, 1975, 1979b). Movements of chub mackerel appear in many respects similar to those of snoek and are again related to patterns of prey availability (Crawford & De Villiers, 1984). Catch records indicate that chub mackerel are most abundant in the northern Benguela region during spring and summer, that they migrate through the central region in autumn and occur south of the Orange River from winter through mid-spring (Fig. 20). The species contributed substantial catches to the South African purse-seine fleet from the mid-1950s until 1975, but subsequently the bulk of the harvest has been made off Namibia by distant-water fleets using midwater and demersal trawl (Crawford & De Villiers, 1984). Off the Western Cape nought-year-old chub mackerel have been caught close to the coast between January and May. Older age classes consume increasingly more fish, tend to occur further offshore at greater depths, and have been most available to purse-seine and demersal trawlers during mid-year (Baird, 1978a, b; Crawford, 1981c). An onshore spawning migration during winter has been recorded by Baird (1974). Most spawning occurs between June and September, and chub mackerel mature during their fourth year of life at a caudal length of 420 mm (Baird 1977). Hakes. Botha (1980, 1985) observed that the Cape hakes occur around the southern African coast from the region of Bahia de Farto, at about 12° S off Angola, to the vicinity of Port Elizabeth, and that in ICSEAF Division 1.6 had been reported offshore to a depth of 920 m. Inada (1981) considered that Merluccius capensis may occur north to the equator and east to East London or Natal, and M. paradoxus from 15° S to the Madagascar Ridge at 33°07′ S; 44°05′ E. He listed preferred temperatures for the two species as 4–12 and 4–8°C, respectively. Botha (1980) demonstrated that off the Western Cape M. capensis occurred in shallower and M. paradoxus in deeper water, the two species being intermixed at intermediate depths, especially between 150 and 220 m (Fig. 21). Larger individuals of both species were found at greater depths than smaller fish and there was little overlap in the distribution of mature specimens of the two species (Fig. 22; Botha, 1985). Off the Western Cape males and females of M. capensis attain 50% maturity at total lengths of 36 and 48 cm, respectively, and those of M. paradoxus at 38 and 47 cm, corresponding in both species to ages of just under four (males) and five (females) years (Botha, 1980). Greater sizes at 50% maturity (43 cm for males and 50 cm for females) have been reported for Cape hakes off Lüderitz (Mombeck, 1970, in Botha, 1980). Females shed all their eggs at spawning but, on account of the extended spawning season, Botha (1980) considered it possible that fish may spawn more than once per annum. Spawning occurs mainly from spring

406

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Fig. 20.—Schematic illustration of movements of chub mackerel Scomber japonicus in the southeastern Atlantic, and (inset) of a known spawning area of chub mackerel and aspects of the movements of some species preyed on by chub mackerel (from information in Crawford & De Villiers, 1984).

to early autumn, and is generally maximal between November and December for both species, with a secondary peak in February and March for M. paradoxus (O’Toole, 1978a; Botha, 1980; Assorov & Berenbeim, 1983), although some inter-annual variability may occur. Spawning on the Agulhas Bank extends throughout the year (Kono, 1980). Spawning is believed to be heaviest over the shelf-break near St Helena Bay (33° S), along the western edge of the Agulhas Bank, and off Namibia in the vicinity of Walvis

THE BENGUELA ECOSYSTEM: PART IV

407

Fig. 21.—Numerical ratio between Merluccius capensis and M. paradoxus in relation to depth in research trawls off South Africa’s Western Cape (from Botha, 1985).

Bay (see Fig. 17; Shannon & Pillar, 1986). Off Walvis Bay M. capensis spawn most frequently where water depth is 160 to 250 m, earliest in the shallower waters (Assorov & Berenbeim, 1983). Most spawning by M. paradoxus is believed to be over the Agulhas Bank (Assorov & Berenbeim, 1983) and the species is not abundant off Namibia north of 27° S (Macpherson, Roel & Morales, 1985). Young M. paradoxus are abundant in the northern and central parts of ICSEAF Division 1.6, especially northwest of Cape Columbine (Payne, Augustyn & Leslie, 1986). To the north of 27° S Cape hakes are mostly M. capensis, with young of the year especially abundant in winter and a probable nursery area in the vicinity of 22° S (Macpherson et al., 1982, 1985). Juveniles of this species are plentiful off the Orange River (Macpherson et al., 1985) and south to 33° S (Payne et al., 1986). The same region is an important nursery region for many epipelagic and mesopelagic species (Crawford, 1980a). From this area M. capensis move to the south and west as they grow older (Payne et al., 1986). Juvenile M. paradoxus in ICSEAF Division 1.6 are found inshore in summer when they feed on euphausiids, swarms of which occur from spring to early autumn (Nepgen, 1957), and offshore in winter when they feed on deep-sea organisms (Botha, 1980; Payne et al., 1986). A similar seasonal movement has been noted in Division 1.5 (Macpherson et al., 1985). Adults are offshore during late summer and autumn at depths greater than 640 m but in early spring move inshore to depths of 330 m and then return offshore, movements that are probably related to both feeding and spawning (Botha, 1980). In the vicinity of Walvis Bay, M. capensis migrate from greater depths to shallower waters to spawn, with a return, post-spawning offshore movement from December to February (Assorov & Berenbeim, 1983). Inada (1981) reported that off Namibia M. capensis undertook seasonal migrations, moving to the south in spring and north in winter. Off northern Namibia, however, there is a tendency for catches of Cape hakes to be highest in ICSEAF Division 1.3 from November through April and in Division 1.4 from May through

408

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Fig. 22.—Numerical frequency distribution by depth of mature Merluccius capensis and M. paradoxus in research trawls off South Africa’s Western Cape (from Botha, 1985).

September (Crawford & De Villiers, 1984; ICSEAF, 1986). This would suggest that if any migrations occur they are to the south in autumn and north in spring, although in the winter of 1984 the southward displacement of hakes may have been caused by intrusion from the north of warm water (Boyd & Thomas, 1984). On the Agulhas Bank (ICSEAF Divisions 2.1 and 2.2) Cape hakes smaller than 25 cm are most abundant in inshore waters between Cape Agulhas and Cape Seal. Larger sizes occur over the entire bank, but densities decrease towards the edge of the continental shelf (Hatanaka et al., 1983; Uozumi et al., 1984, 1985). Cape hakes have been shown to undertake extensive diurnal vertical migrations away from the sea bed at night outside the spawning season, and a movement of reduced intensity during peak spawning when they are found more often in mid-water and are less available to demersal gear. The vertical migrations are mainly for feeding at night (Botha, 1973, 1980) when shoals are dispersed (Konchina, 1986). Within shoals there is no vertical stratification of sizes (Macpherson et al., 1982). Kingklip. Kingklip occur in commercial concentrations between Möwe Point off Namibia and east of Port Elizabeth on the southern coast, often close to rocky substrata (Payne, 1977, 1985b). The depth distribution of kingklip is size-dependent, smaller fish occurring inshore and larger, older fish in deeper water (Wrzesinski, 1975; Payne, 1977; Isarev, 1976, 1981). Kingklip were not encountered at depths greater than 600 m by Wrzesinski (1975). From latitudes 19 to 29° S the species is most abundant between the 200and 500-m isobaths (Macpherson, 1983a). Off Namibia kingklip are caught where water temperatures near the bottom range between 6 and 13 °C, young adults especially at temperatures between 8·5 and 9 °C and

THE BENGUELA ECOSYSTEM: PART IV

409

Fig. 23.—Relationship between monthly catch rates of kingklip Genypterus capensis and an index of upwelling in the Walvis Bay vicinity off Namibia (from Isarev, 1981).

depths of 350–400 m and older individuals at temperatures between 7 and 7·5 °C and depths of 450–500 m (Isarev, 1981). Off Namibia from 1972 to 1975, catch rates by Spanish vessels were highest between latitudes 25 and 27° S, an area of intense upwelling (Shannon, 1985), and from February to April (Macpherson, 1976; Macpherson & Lloris, 1977; Isarev, 1981). Seasonal trends in catch rates lagged about two months behind an index of upwelling based on surface wind intensity (Fig. 23), leading Isarev (1981) to speculate that aggregations of kingklip form in areas where food organisms are abundant. He attributed the lag between peak upwelling and highest catch rates to time required for a surface phenomenon to be transmitted to lower depths, and suggested that the low catch rates in August and September (Fig. 23) resulted from kingklip migrating to their spawning grounds. Spawning is believed to occur off the southern African west coast from August to October (Payne, 1977), and off Namibia mainly between 20 and 26° S in inshore waters (Isarev, 1986). During a spawning season a female, depending on size, produces on average between 124900 and 897000 eggs in three to four batches (Isarev, 1986). Most females are mature at 52 cm and males at 56 cm, in both cases corresponding to an age of about five years (Isarev, 1986). On the basis of his estimate for K in the von Bertalanffy growth equation, Wrzesinski (1975) considered that the annual natural mortality rate (M) for kingklip was likely to fall within the range 0·10–0·15. Isarev (1979) considered kingklip aged five or older to be fully selected, and for such fish estimated the total mortality rate (Z) to be of the order of 0·8 for ICSEAF Divisions 1.3 and 1.4 during 1967–1976, and about 0·45 for ICSEAF Divisions 1.5 and 1.6 during 1970–1976. Estimates of M were computed as 0·3 and 0·2, respectively (Isarev, 1979, 1980). Panga and dentex. Over the Agulhas Bank there is a tendency for panga of total length less than about 20 cm to be most abundant south of Cape Infanta, but for larger individuals to be distributed both further offshore and further east (Hatanaka et al., 1983; Uozumi et al., 1984, 1985). As with horse mackerel, panga apparently migrate inshore in winter (Uozumi et al., 1985). Sex reversal occurs (Hecht & Baird, 1977) and has been demonstrated also in two other sparids occurring on the Agulhas Bank: roman Chrysoblephus laticeps (Penrith, 1972) and dageraad C. cristiceps (Robinson, 1976). For panga the proportion of males in

410

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

research catches increased from 0% for fish of total length 20 cm or less to 50% at 35 cm and 100% at 38 cm (Uozumi et al., 1985). ICSEAF records only report catches of Angolan dentex from Divisions 1.1 and 1.2, whereas the bulk of the catch of large-eye dentex has been taken in ICSEAF Divisions 1.2 and 1.3 (see Table XVII, p. 387). Although catches of the latter species have been reported from the southern divisions of 1.6, 2.1 and 2.2, Kuderskaya (1985) considered the southern limit of the distribution of large-eye dentex to be 27°40′ S. Juvenile large-eye dentex occur inshore, but as they grow they move to near the bottom and to greater depths (Kuderskaya, 1985). Sexual maturity is reached at a length of 18–20 cm, and spawning takes place throughout the year, with a peak from October to April. The most intensive spawning occurs at depths of 75–110 m between Baia dos Tigres and the mouth of the Cunene River, but the eggs are pelagic and larvae are carried inshore to shallower waters (Kuderskaya, 1985). Dense concentrations of large-eye dentex do not occur where temperatures near the bottom are below 11°C. Fecundity ranges between 72000 and 1300000 eggs, with five to six batches being spawned (Kuderskaya, 1985). Soles. The distribution of Agulhas sole is limited to areas of benthic mud deposited by river discharge (Zoutendyk, 1973). The most extensive and productive trawling grounds lie between Cape Agulhas and Mossel Bay. Although Agulhas sole are trawled as far east as East London, rocky substratum interrupts the continuity of trawling grounds especially off the Tsitsikamma Coastal National Park (Zoutendyk, 1973). During 1970–1974 there was an offshore migration of the northern stock of west coast sole (Payne, 1985a). In 1970 and 1971 the bulk of the catch of this stock was made in inshore waters of 30–60 fm (56– 112 m) depth, in 1972 at depths of 120–150 fm (223–279 m), and in 1973 and 1974 at depths greater than 150 fm (275 m). In 1970 catch rates were high on the inshore grounds and low offshore, but by 1974 this position had been reversed (Fig. 24). Since 1974 the northern sole resource has remained in deep water (Payne, 1985a). Payne (1985a) suggests that the offshore displacement of the resource may have been triggered by excessive discarding of offal and small fish on the inshore fishing grounds, but admits the possibility of environmental changes. There may have been similar displacement of the southern stock of west coast sole, catches of which decreased rapidly after the mid-1960s (Payne, 1979, 1985a; see Table XXII, p. 395). Offshore of the traditional fishing region is a large area of untrawlable ground, so that westward displacement would have been accompanied by falling catches (Payne, 1985a). For the southern stock, fish larger than 30 cm tend to occur south of 30° S and fish of smaller size to the north of this latitude (Payne, 1979). Agulhas sole spawn throughout the year but mainly in October and November (Marchand, 1933). The spawning season of the northern stock of west coast sole is from early spring to early summer (O’Toole, 1977b). Onset of spawning by the southern stock is in late winter (Payne, 1979). Marchand (1933) recorded a mean size at maturity of 30 cm for Agulhas sole, noting that males matured at a smaller size than females. In 1971 50% maturity was at 29 cm, males again maturing at a slightly smaller size (Botha, Lucks & Chalmers, 1971). By contrast Zoutendyk (1974) reported a large difference in size at maturity between the two sexes, with maturation of males observed before they had attained 18 cm and of females at only approximately 33 cm or four years old. Length at 50% maturity of west coast sole was 29 cm for the northern stock in 1972 (Lucks, 1972), but for the southern stock 24 cm in 1978 and 21 cm from 1979 to 1981 (Payne, 1979, 1985a). Males matured at a smaller size than females (Payne, 1979). Tunas. Tunas, being scombrids, are migratory fish (Collette & Nauen, 1983). The six larger species in the Benguela system and adjacent areas all spawn in tropical waters warmer than 24°C. Species such as Thunnus albacares and Katsuwonus pelamis are essentially tropical, whereas Thunnus alalunga and T. maccoyii can be regarded as temperate tunas which migrate over large distances between spawning and feeding grounds. Although sea-surface temperature can be useful, however, in forecasting the occurrence of a tuna species in

THE BENGUELA ECOSYSTEM: PART IV

411

Fig. 24.—Catch rates and catch proportions of west coast sole Austroglossus microlepis by depth stratum, and total sole catches on the northern grounds, 1970–1974 (from Payne, 1985a).

areas where it only occurs seasonally (Nakamura, 1969), tunas are eurythermal and can tolerate a range of temperature of at least 10 °C and as much as 20 °C in the case of T. maccoyii (Nakamura, 1969). Even the tropical T. albacares when feeding may occur where temperatures are as low as 12°C (B.Rose, pers. comm.). Temperature per se is important physiologically for tunas, but it also characterizes areas of suitable visibility (Laevastu & Rosa, 1963) and of food availability such as may exist in the proximity of marked density gradients in temperature. For example, T. obesus is often associated with a thermocline, whereas T. alalunga and Katsuwonus pelamis frequently occur near fronts (Shannon, 1986). Some tunas tend to migrate within water masses rather than across boundaries, and Lima & Wise (1963) noted that the distribution of tunas and their fisheries in the Atlantic are closely related to current patterns. The cold region between about 17 and 33° S presents a barrier to the meridional migration of more tropical tunas such as Thunnus albacares and Katsuwonus pelamis. As a result occurrences of such species at the

412

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

northern and southern extremities of the Benguela probably represent origins from the Atlantic and IndoPacific, respectively. The Benguela is the only cool eastern boundary upwelling regime that is bounded at both ends by warm water. The boundaries shift seasonally (Shannon, 1985). In summer and autumn the region of upwelling is ‘pinched’ between the warm Angola and Agulhas Current systems, giving rise to intense fronts. Strong fronts also develop at the oceanic boundary during the upwelling season. These frontal regions provide suitable habitats for tunas such as Thunnus alalunga and Katsuwonus pelamis. Another factor which is important in the distributional ecology of tunas is the abrupt termination of the African landmass at about 35° S, which makes possible interchange between Atlantic and Indo-Pacific tuna populations. For temperate tunas (e.g. Thunnus alalunga) the Cape of Good Hope will not present a barrier to inter-ocean migration during winter, but the high surface temperatures associated with the Agulhas Current during summer could effectively divide the populations and minimize intermingling between them. For tropical tunas interchange between the two oceans will be facilitated during summer, whereas the presence of cooler surface water between May and September south of Africa will present a barrier. According to Alekseev & Alekseeva (1980), “Reproduction of tunas in the Atlantic Ocean remains the least investigated aspect of their biology.” Tunas tend only to spawn in water warmer than 24°C. Not surprisingly, therefore, no eggs or larvae of tunas have been recorded in ichthyoplankton from the Benguela (Shannon & Pillar, 1986). Of nearly 3000 tunas caught in the Benguela no ripe fish were encountered apart from a few T. alalunga caught off Namibia (De Jager, Nepgen & Van Wyk, 1963a, c; Penrith, 1963; Talbot & Penrith, 1963; Van den Berg & Matthews, 1969; Nepgen, 1970a), supporting the view that tunas do not spawn off southern Africa (Nepgen, 1970a). Gonads of most specimens of T. alalunga, T. albacares, T. maccoyii, and T. obesus examined by De Jager et al. (1963a) were small and inactive, but Talbot & Penrith (1963) reported that most T. maccoyii and T. albacares were mature, as were most T. alalunga longer than 85 cm and most T. obesus longer than 116 cm. It appears that no tuna populations are permanently resident in the Benguela, and those mature fish which occur seasonally must migrate into the warmer tropical waters of the Atlantic and Indian oceans to spawn. With regard to individual species, T. alalunga is the most abundant Benguela tuna constituting over 80% of the longline catch off southwestern Africa (Nepgen, 1970b; Yang, 1983). The species is widely distributed (Nepgen, 1970b) and is present in the Benguela region throughout the year. There are two areas of high abundance in the Benguela, viz. off the Cape Peninsula and south of Lüderitz (De Jager et al., 1963a, c; B.Rose & R.Lamberth, pers. comm.). These areas are immediately poleward of major upwelling centres where the shelf-break and oceanic thermal front occur close inshore. T. alalunga is not common in the southeastern Atlantic north of 20° S. Maximum abundance off the Cape is in spring and in autumn/winter (De Jager et al., 1963a; Talbot & Penrith, 1968), whereas in the area 15–30° S: 10–15° E the species is most abundant in winter (Morita, 1978). Optimum sea-surface temperature is 17–20°C (Nepgen, 1970a). Although Lebeau (1971) presented a conceptual model of the migration of albacore in the southwestern Indian Ocean, relatively little is known about the migration of T. alalunga around the Cape of Good Hope. Movement of albacore from east to west around the Cape of Good Hope during summer has, however, recently been confirmed by tagging (R. van der Elst, pers. comm.). T. alalunga spawn off Brazil south of the equator, in the central South Atlantic and in the southwestern Indian Ocean east of Madagascar between February and April (Lebeau, 1971; Bard, 1982). Seasonal movements between habitats in the South Atlantic have been illustrated diagrammatically in Le Gall (1974) and Pollock (1984). Migration after winter of mature adults from the southeastern Atlantic to the spawning area off Brazil where they arrive in summer is suggested. Adults return to southern Africa in autumn, thereby augmenting the groups of immature fish which are present in the Benguela throughout the year.

THE BENGUELA ECOSYSTEM: PART IV

413

T. albacares is seldom encountered in the cool upwelled water of the Benguela, but does occur near the boundaries of the system in warmer water. It is present in the southern Benguela between spring and autumn, optimum sea-surface temperature being about 18–22 °C, and shoals probably disperse over the Agulhas Bank during summer (De Jager et al., 1963a; Talbot & Penrith, 1968; Nepgen, 1970a). A seasonal migration of T. albacares from the Indian Ocean into the waters off southwestern Africa was shown by Hayasi (1974), and fish caught north to 33 or 32° S are probably of Indian Ocean origin. Abnormally large concentrations of yellowfin, evidently a single cohort, congregated on the western edge of the Agulhas Bank during spring 1979, when the sea-surface temperature was about 19 °C (Shannon, 1980). A major spawning area exists in the Indian Ocean, and the species also spawns off Brazil between January and March and possibly in the Gulfs of Mexico and Guinea (Albaret, 1976; Kikawa & Nishikawa, 1980). T. maccoyii used to be most abundant in the southern Benguela during late winter and to a lesser extent in spring, disappearing in summer (De Jager et al., 1963a; B.Rose, pers. comm.). It was present but not abundant off Namibia, and did not occur in the Agulhas Current (Talbot & Penrith, 1968; Van den Berg & Matthews, 1969). In the southern Benguela region it was distributed from the mid-shelf outwards (Talbot & Penrith, 1968), tending to congregate around the outer shelf-break. There is seasonal migration of adults between spawning grounds off western and northwestern Australia and cold feeding grounds off Tasmania and New Zealand (Collette & Nauen, 1983). Spawning occurs in the austral summer in warm water of 23 to 26 °C (Collette & Nauen, 1983; ICCAT, 1984). The younger fish are distributed in Australian coastal waters, but as they age they migrate at higher latitudes into the Pacific, Indian and Atlantic Oceans. Peak abundance in the Benguela used to coincide with the northward shift in atmospheric pressure systems, reduced insolation and reduced influx of Agulhas Current water into the Atlantic. T. obesus is most abundant in the Benguela during winter and spring (De Jager et al., 1963a; Van den Berg & Matthews, 1969), although Talbot & Penrith (1968) recorded maximum abundance in spring and summer in the southern part of the system. In the northern Benguela the species was the second most abundant tuna after T. alalunga (Van den Berg & Matthews, 1969). Highest concentrations tend to occur offshore in deep water over and west of the outer shelf-break, and at the northern and southeastern boundaries of the Benguela system (Nepgen, 1970b), areas where stratification is known to be seasonally intense (Shannon, 1985). Optimum range of seasurface temperature is about 18 to 21°C (De Jager et al., 1963a; Nepgen, 1970a). Widespread distribution of bigeye tuna in the southeastern Atlantic during the second and third quarters suggests that the Benguela is a feeding habitat for non-spawning fish that are part of the Atlantic stocks. The species spawns off Brazil, in the eastern Central Atlantic north of 5° N and in the Gulf of Guinea, mostly between November and April and where the seasurface temperature is 24–29 °C (Rudomiotkina, 1983). Although the diet of T. obesus is similar to that of T. maccoyii, competition between the two species is believed to be minimal on account of their different patterns of distribution (Penrith, 1963; Talbot & Penrith, 1963), T. obesus being more abundant in the northern Benguela. T. thynnus, the Atlantic bluefin, is rare but does occur in the southern Benguela. Brunenmeister (1980) reported that this species has been caught west of Cape Town between March and June and southeast of Africa during winter. According to Talbot & Penrith (1968) the species is not, however, found in the Indian Ocean. Maximum abundance of the species in the southern Benguela appears to have been in the 1940s, 1950s and early 1960s, when it was recorded during summer (B.Rose, pers. comm.), which also seems to contradict Brunenmeister (1980). It is likely that the “bluefin” was actually southern bluefin. T. thynnus spawns in the Gulf of Mexico between May and June, in the Mediterranean and sporadically in the Gulf of Guinea during February, March and August (Richards, 1976; Richards & Potthoff, 1980). Katsuwonus pelamis normally only occurs near the boundaries of the Benguela system in warmer waters. It is common off Angola during summer. Off the Cape Peninsula skipjack tuna are associated with a sea-

414

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

surface temperature of 19–20°C, arrive around December and remain until May (R. Lamberth, pers. comm.). Skipjack tuna often occur in association with yellowfin tuna. A northerly migration from southern Angola has been demonstrated by tagging and a return poleward migration during the austral summer hypothesized (ICCAT, 1984). The return migration would be consistent with a poleward migration of warm Angolan water towards the northern Benguela during the first quarter of the year (Shannon, 1985; Shannon et al., 1986). Fish caught in the southern Benguela are probably of Indo-Pacific origin. Some intermingling between Atlantic and Indo-Pacific stocks of skipjack tuna could take place offshore of the Benguela during summer. K. pelamis spawns off Brazil during summer (Kikawa & Nishikawa, 1980; Matsuura, 1982) and in the Indian Ocean (Shannon, 1986). Western coast rock lobster. Female rock lobster have a well-defined moulting and spawning cycle. These events occur sequentially each year, with most mature females moulting in about May-June, followed shortly afterwards by copulation and oviposition. The berry season extends from May-June to OctoberNovember. As in all Macrura, the incubation period is temperature-dependent. In most years incubation appears to last for about three to four months (Newman & Pollock, 1974; Beyers, 1979), suggesting a reasonably constant annual temperature regime at least during the winter and early spring months. For reasons not yet fully understood, the onset of moulting and egg-bearing in females is delayed in certain areas adjoining St Helena Bay, viz. Elands Bay (Newman & Pollock, 1971) and North Blinder (Pollock, Augustyn & Goosen, 1982). Peak hatching of eggs takes place in October-November in most areas and years. This appears to be synchronized with the onset of strong wind-induced upwelling which usually commences at this time of the year, especially towards the southern part of the Benguela system. Surface drift and Ekman transport would tend to move newly-hatched larvae in surface layers northwards and offshore. Late-stage larvae (i.e. phyllosoma stages VI to XI) are only found well offshore (Lazarus, 1967; Pollock, in prep.), so that the majority, if not all, of the larvae produced along the coast are transported into offshore waters, outside the immediate influence of the Benguela system. The oceanographic mechanisms responsible for the return of the relatively few surviving late-stage larvae, or their retention along the outside of the continental shelf, have not yet been described, although very large circulation systems, with an extensive time period, are suspected to be involved. Females reach 50% sexual maturity at about 66 mm carapace length on most grounds south of about 31° S, an exception being at one of the major upwelling sites near Olifantsbos on the Cape Peninsula. Further north, sizes at 50% maturity are smaller, at present 59 mm in the Cape just south of the Orange River, and 57 mm in Namibia. Pollock (in prep.) has suggested that the reduced sizes at maturity are the result of retarded juvenile growth rates in these northern regions. Fecundity studies conducted recently by Beyers & Goosen (in prep.) at ten sites between the Orange River and Cape Point, have shown that fecundity is low in areas where growth rates are retarded and sizes at maturity of females are smaller than elsewhere, namely in the Port Nolloth region just south of the Orange River and at Olifantsbos. The regression of egg numbers against carapace length for the females on grounds south of 31° S, except Olifantsbos, is:

where F=no. of eggs and X=carapace length (mm). Hence a female of 89-mm carapace length (the minimum size limit) produces approximately 191000 eggs, increasing to about 240000 eggs at a size of 100 mm. Red crab. Female red crabs appear to undergo a terminal puberty moult, that is they cease moulting at the attainment of sexual maturity which occurs at a carapace width varying from about 75 mm to about 110 mm,

THE BENGUELA ECOSYSTEM: PART IV

415

the maximum size attained (Melville-Smith, in press). Males, on the other hand, continue moulting after reaching sexual maturity at about 80–90 mm (Le Loeuff, Cayre & Inte’s, 1978) and attain a maximum size of about 160 mm carapace width. Adult females generally occur in shallower water than males, so that sex ratios change with depth. For example, the percentage of males in samples collected from different depths was shown by Beyers & Wilke (1980) to increase from about 47 to 53% at 400 m to more than 90% at depths greater than 800 m. Gravid females have never been found at depths greater than 600 m, so that virtually all egg production and larval release takes place towards the shallower parts of the continental slope. Mellville-Smith (in press) has pointed out that slightly higher temperatures of about 8°C in the 400– 500 m depth range would tend to accelerate incubation of eggs, compared with depths exceeding 600 m where temperatures from 4° to 6°C prevail. No seasonal cycles are evident in moulting and egg-bearing, and females are able to store sperm for an indeterminate period before oviposition takes place. Fecundity is described by the relationship:

where F=number of eggs and X=carapace width (mm) (Melville-Smith, in press). Squids. According to Augustyn (in press) Loligo reynaudii on the Agulhas Bank move inshore, mainly during the summer months, to spawn in depths shallower than about 50 m on sandy substrata that are relatively flat, especially in sheltered bays along the Southern Cape. Mating and spawning takes place in dense aggregations, after which most adults die. Females are all mature at a mantle length of about 200 mm and 50% mature at about 150–160 mm. Males approach maturity over a wider range of sizes than females, and all males are mature at mantle lengths in excess of about 250 mm. Eggs are attached in capsules to the sea bed. After hatching, early juveniles feed inshore and grow rapidly, at a rate of 10–20 mm per month. From a size of about 90 mm mantle length they begin to move to deeper waters. Maturity is thought to be reached after about a year, at which stage most animals return inshore to spawn. Some spawning activity has been noted in the winter months, and there are indications that the breeding season is longer off the Western Cape than off the Southern Cape (Augustyn, in press). Abalone. Adult abalone occur in dense ‘beds’ or aggregations where they reproduce by means of synchronous spawning. The dioecious abalone release vast quantities of eggs and sperm into the water, where fertilization takes place. The stimulus for spawning appears to be a rise in water temperature, although the presence of sperm cells in the water is also known to induce females to spawn (Newman, 1967). The fecundity of females between Cape Point and Cape Agulhas is described by the relationship:

where F=millions of eggs and X=shell breadth (cm). From this it is calculated that a female of 11·4 cm, the minimum harvestable size, produces approximately 4·3 million eggs, and this number increases exponentially with size. In the Western Cape 50% of abalone are mature at a shell breadth of 80 mm, and all animals are mature at about 105 mm. Some geographical variation is evident in size at maturity, with female abalone in the eastern Cape attaining sexual maturity at a much smaller average size than in the southwestern Cape (Newman, 1969a). Although age determination of abalone is not possible, studies by Newman (1969a) suggest that abalone in the eastern Cape also attain maturity at a younger age than elsewhere. Insufficient data were available from cooler water populations off the western coast to test whether any differences in size and age at maturity occur there.

416

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

POPULATION BIOLOGY Stock assessment Assessments of fish resources in the southeastern Atlantic have mostly been based on catch records, using especially Virtual Population Analysis (VPA) and production modelling (cf. Newman, 1977; Butterworth, 1983; De Villiers, 1985), although more recently increased attention has been devoted to direct surveys of resources supporting both demersal and purse-seine fisheries (cf. Hatanaka et al., 1983; Uozumi et al., 1984, 1985; Hampton, Shelton & Armstrong, 1985). Techniques such as mark and recapture have also been employed, especially for crustacean resources but also for pilchard (Newman, 1970a; Newman & Pollock, 1977). Up-dated assessments for many resources are published annually in the ICSEAF Collection of Scientific Papers and other ICSEAF reports, to which series the reader is referred for more detailed information. Epipelagic species, South Africa. A preliminary assessment of yield of the pelagic stocks supporting the mixed-species purse-seine fishery off South Africa’s Western Cape, based on an analysis of effort and catch rates during 1965–1972 when pilchard catches were decreasing, suggested that an equilibrium yield of approximately 360000 metric tons of all species combined could be sustained (Newman, Crawford & Centurier-Harris, 1974). This value, based on the assumption that production was independent of species composition, served as a basis for management until the late 1970s, by which time VPA assessments of biomass and year-class strength were available for the most important species (Centurier-Harris, 1977; Newman & Crawford, 1980). The assumption of constant production irrespective of the species base may not have been unreasonable (Shannon & Field, 1985). A deterministic simulation model of the fishery, taking into account the dynamics of the component species and the fishery operation, was developed in 1979 and indicated that, under conditions of average recruitment, an equilibrium yield of 375000 tons could be expected from the five major species, namely pilchard, anchovy, horse mackerel, chub mackerel, and round herring (Crawford, 1979). The model did not account for possible interaction between species, nor relationships between parent stock and recruitment. Its purpose was to explore the average situation and, because at that time it was thought unrealistic to expect average recruitment from either horse mackerel or chub mackerel, it was used to argue for a reduction in quota to a level of about 325000 tons. Stochastic modelling and risk analysis, again with the output from VPA as a basis, were introduced in the 1980s (Butterworth, 1983; Armstrong, 1984). For the South African anchovy stock it was shown that stability of yield would necessitate acceptance of lower average yields than could in the long term be attained by quota adjustment (Armstrong, 1984). VPA estimates of abundance of pilchard off South Africa are shown in Figure 25 and indicate that biomass increased from levels of about 400000 tons in the early 1950s to over one million tons in the early 1960s, largely as a result of the formation of powerful year-classes during the late 1950s. From the mid-1960s stock size fluctuated around a level of 100000 tons or less. VPA estimates of biomass of anchovy indicate a doubling of stock size during the early 1970s, with a maximum abundance during the late 1970s of somewhat less than one million tons, the actual value being dependent on the natural mortality assumed (Fig. 26). The mid-1980s saw a re-assessment of the applicability of VPA as a basis for assessment of the pilchard and anchovy stocks off the Western Cape (Armstrong, Shelton & Prosch, 1985). Surveys showed that adult anchovy occurred mostly on the Agulhas Bank, beyond the range of purse-seiners (Hampton et al., 1985), whereas growth studies indicated that catches of anchovy contained a much larger proportion of young of

THE BENGUELA ECOSYSTEM: PART IV

417

Fig. 25.—VPA estimates of the biomass of pilchard Sardinops ocellatus off South Africa’s Western Cape, 1951–1983 (from Shelton & Armstrong, 1984).

the year than previously thought, so that trends in VPA largely reflected catches of adult anchovy (Armstrong et al., 1985). It was suggested that catches of adult anchovy may have been influenced by longterm changes in sea temperature affecting availability of anchovy to the fleet (Armstrong et al., 1985). Consequently, advice for management of the South African anchovy stock is at present based on information generated by direct survey. Annual estimates of the abundance of the spawner stock derived from acoustic measurements are available since 1984 (Armstrong & Butterworth, 1986). In November 1984 the spawner biomass of anchovy off South Africa was estimated as 1 million metric tons from acoustic survey and 1·6 million tons from the egg production method. The much lower estimates previously obtained by VPA were attributed to larger fish being under-represented in commercial landings because of their concentration in offshore waters on the Agulhas Bank (Hampton et al., 1985). Armstrong & Butterworth (1986) employed the mean biomass estimate (1·32 million tons) in a modification of the Beddington & Cooke (1983) stochastic yield-per-recruit model to suggest 279000 tons as an appropriate yield for the resource. The value was obtained by first calculating an average value for year-class strength consistent with the biomass estimate, with previous catches and with estimates of population parameters, assuming that commercial catches comprised only young of the year. This information was then used to suggest that a mean spawner biomass of about 2 million tons might have been expected in the absence of fishing. To allow for uncertainties in parameters and estimates used in the model, the recommended yield was one standard deviation less than that which would on average have reduced the spawner biomass to half its unexploited level (Armstrong & Butterworth, 1986). Acoustic surveys conducted off South Africa between 1983 and 1985 indicated a low abundance of pilchard compared with anchovy and round herring. On these surveys pilchard contributed on average less than 10% by mass of the catches, and in November 1984 the estimated biomass of pilchard between St Helena Bay and Port Elizabeth was 46000 tons (Armstrong, 1986). Con-firmation of a low biomass was provided by a length-based cohort analysis, which suggested that during 1973–1985 the average biomass of pilchard aged one and older was unlikely to have exceeded 100000 tons (Armstrong, 1986). For round herring, VPA has suggested biomasses of up to 120000 tons (Crawford, Shelton & Hutchings, 1983). Because of relatively small catches of the species and a poor understanding of the pattern of

418

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Fig. 26.—VPA estimates of the biomass of anchovy Engraulis japonicus, and handline catches of snoek Thyrsites atun, off South Africa’s Western Cape, 1969–1983 (from Crawford & De Villiers, 1985).

catchability at age, assessments are, however, considered unreliable (Armstrong et al., 1985). In catches of juvenile fish with blanket nets at inshore stations off the Western Cape the contribution of round herring increased from 7% in 1955 to 81% in 1967. Average values were 15% for 1955–1957, 36% for 1958–1963 and 68% for 1964–1967, indicating a steady increase in importance of the species in the system (Sea Fisheries Research Institute, unpubl. records). During 1977 and 1978 eggs of round herring had a high abundance in the southern Benguela (Davies, Newman & Shelton, 1981). Epipelagic species, Namibia. For the pilchard off Namibia, tagging indicated a stock size of about 5·7 million tons during the late 1950s and 6·8 million tons between 1963 and 1966 (Newman, 1970a). Between 1971 and 1974 Cram (1977) estimated abundance to be between 1·7 and 3·2 million tons from egg surveys and aerial and acoustic surveys. VPA estimates suggest that biomass of the resource increased from about 5 million tons in the mid-1950s to over 11 million tons in 1964 as a result of powerful year-classes formed during the early 1960s (Thomas, 1986). The stock then collapsed to under 2 million tons in the early 1970s, before a temporary recovery resulting from a substantial 1972 year-class (Fig. 27). Since 1977 the biomass has, however, remained at a level less than 0·5 million tons. Mean ages of fish in the stock and in the catch underwent large decreases, and fishing mortality was high in the late 1960s (Fig. 27). In his analysis Thomas (1986) used catch levels reported by the Namibian fleet as input data, and made no provision for dumping of unwanted catches, under-reporting of landings, mis-identification of species in catches (Cram, 1981; Butterworth, 1983) or the harvests of pilchard made by other nations, which were especially high in the mid-1950s and mid-1960s (cf. Tables I and V, pp. 357 and 368). His trends, however, show reasonable agreement with those derived from catch per unit effort, tagging and aerial-acoustic survey (Fig. 27). Indeed, for pilchard in both northern and southern Benguela, the gross trends in abundance suggested by VPA conform in large measure with trends indicated by other estimators such as guano harvests (e.g. Fig. 28), catch rates, aerial-acoustic surveys and egg surveys (Cram, 1977; Crawford & Shelton, 1978; Troadec, Clark & Gulland, 1980; Butterworth, 1983; Crawford, Shelton & Hutchings, 1983; Thomas,

THE BENGUELA ECOSYSTEM: PART IV

419

Fig. 27.—Trends in (a) the mean age of pilchard Sardinops ocellatus in catches off Namibia and estimates of weighted fishing mortality and (b) biomass, catch rates and year-class strength (from information in Thomas, 1986).

1986). Absolute estimates of biomass may be in error, but the trends are probably mostly robust, although the possibility remains that all indices were influenced by the same environmental phenomenon. This, however, seems unlikely and certainly the collapses of the pilchard stocks are beyond dispute. In the absence of other methods VPA has been used to assess the anchovy resource off Namibia, even though a large proportion of the commercial catch is based on young of the year (Le Clus, 1985). A steady decrease in abundance since 1972 is indicated (Fig. 29). Estimates of numbers of breeding pairs of Cape cormorants Phalacrocorax capensis at both platforms and islands off Namibia closely match trends in the catch of anchovies (Fig. 30; Crawford & Williams, in prep.). Cape cormorants forage in groups and are probably dependent on large, dense shoals of fish (Duffy, Berruti, Randall & Cooper, 1984) in the nearsurface layers. Off Peru the collapse of the anchoveta resource was followed by a considerable reduction in numbers of cormorants P. bougainvillii (Brown, 1981). Epipelagic species, Angola. De Villiers (1985) reports that an assessment published by ICSEAF in 1976 and based on catch and effort information indicated that the Sardinella resources had shown no signs of

420

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

over-exploitation at catch levels of 100000 to 150000 tons. Since 1978 catches have been in excess of 150000 tons (De Villiers, 1985; see Table I, p. 357). Mesopelagic species, horse mackerel. Biomass of Cape horse mackerel in ICSEAF Division 1·6 is estimated by VPA to have decreased from nearly 700000 metric tons in the early 1950s to very low levels by the 1970s (Crawford, Shelton & Hutchings, 1983). Powerful year-classes were formed in the late 1940s (see Fig. 3, p. 362; Geldenhuys, 1973) and are believed to have been responsible for the high biomasses of the early 1950s (Fig. 31). For the Agulhas Bank, VPA indicated that biomass of the horse mackerel resource was relatively stable between 1975 and 1983, except for a noticeable drop following the large catch of 93000 tons taken in 1977 (Kinloch et al., 1986). Fish aged four and older were shown to be fully recruited to the trawl fishery and, at annual natural mortality rates of 0·25–0·30, had mean biomasses of 117000 to 200000 tons. The values for 1980–1982 matched survey estimates obtained from the aereal expansion method (Hatanaka et al., 1983; Uozumi et al., 1984, 1985). VPA suggested that year-class strength decreased between 1973 and 1981 (Fig. 31; Kinloch et al., 1986). The stochastic model in Beddington & Cooke (1983) indicated that at yields of 25000– 32000 tons the probability of stock depletion to below 20% of the unexploited biomass would be less than 10% over a 20-year period (Kinloch et al., 1986). VPA has also been used to assess the northern stock of Cape horse mackerel, although trends reported by different workers for the late 1960s are at variance, ranging between a steep decrease in biomass and little change (Babayan et al., 1986; Bergh, Butterworth & Andrew, 1986). There is also a difference in absolute estimates of biomass. All approaches, however, indicate a large increase in biomass between the late 1960s and the late 1970s followed by a decrease in the early 1980s. Powerful year-classes appear to have been formed from 1977 to 1980 (Fig. 31). Applying an F(0·1) management procedure to a yield-per-recruit analysis based on mean recruitment Bergh et al. (1986) suggested that 225000 tons would be an appropriate level of harvest. This may be compared with the mean annual catch for the period of their analysis of 270000 tons. Babayan et al. (1986) estimated mean year-class strength to be more than twice as large, and projecting forward from the 1985 stock structure suggested a yield of 485000 tons in 1986 for the same management objective, a figure that was adopted by ICSEAF (1986). Mesopelagic species, other species. Gjosaeter & Kawaguchi (1980) estimated that biomass of mesopelagic fish in the southeastern Atlantic was about 18 million tons, of which 2 million tons occurred within 160 km of the coast. If this is split between the northern and southern Benguela pro rata to the phytoplankton standing stocks for the two regions (Moloney & Field, 1985), the biomass of lanternfish and lightfish in the inshore southern Benguela region would be of the order of 570000 tons. By an input-output analysis Bergh, Field & Shannon (1985) estimated annual production of lanternfish off the Western Cape to be about 760000 metric tons wet weight. Given the suggested high annual mortality rate of about 2·4 (Prosch, 1986), this implies a mean biomass of about 320000 tons. Groundfish species, hakes. Both production models and VPA have been used to assess the hake resources, VPA being regarded as the less suitable because of problems in estimating predation on young hake, particularly by cannibalism (ICSEAF, 1986). Furthermore, Leslie (1986) has pointed out that, since the landed catch of hakes cannot be separated into the two hake species, it is necessary to consider them as a single entity in VPA. In ICSEAF Division 1.6 Merluccius paradoxus contributes a higher proportion of the catch of small fish than M. capensis, whereas the reverse is true for the catch of larger hake, so that in effect VPA attempts to relate catches of adult M. capensis with catches in earlier years of juvenile M. paradoxus (Leslie, 1986). In spite of these misgivings, ICSEAF (1986, p. 57) noted that trends determined by VPA still had value when employed in conjunction with estimates made by other means of assessment, no doubt because confidence in other methods is not absolute. Drawbacks with regard to the use of production models for hakes have been spelt out: lumping of the two species, discarding of small hake in earlier years,

THE BENGUELA ECOSYSTEM: PART IV

421

Fig. 28.—Trends in the production of seabird guano at islands and platforms off South Africa and Namibia, and in the biomass of the two pilchard Sardinops ocellatus stocks (updated from information in Crawford & Shelton, 1978; Crawford, Shelton & Hutchings, 1983).

422

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Fig. 29.—VPA estimates of the biomass of various components of the anchovy Engraulis japonicus resource off Namibia (from Le Clus, 1985).

and changes in mesh sizes leading to decreases in growth overfishing, changes in the age structure of catches and altered relationships between catch rates and effort (Newman, 1977). Growth rates for the two species are fortunately similar although, because of their depth-specific distribution, a change in fishing pattern could influence the relative contribution of the two species to the catch (Newman, 1977). VPA shows large absolute decreases in biomasses of the three western coast stocks during the 1970s, as well as a decrease in the southern coast resource (Fig. 32). A recovery of the two northernmost stocks since 1980 is indicated, and confirmed by considerably improved catch rates (see Table XIII, p. 383). Rising catch rates also suggest that abundance of the two southern stocks has increased since the early 1980s (Table XIII). For Division 1.6 this trend conforms with results obtained on direct surveys (Payne, Augustyn & Leslie, 1986), although the increases have not yet been reflected by VPA (Leslie, 1986). Year-class strength has been variable, but appears to have been good for most hake stocks during 1969–1970 and 1980– 1981 (Fig. 32). The earliest production model for Cape hakes was one developed by Jones & Van Eck (1967) for the fishery in Division 1.6. It suggested that the maximum equilibrium yield was about 100000 metric tons. More recently assessments have been undertaken annually by ICSEAF, whose publications contain substantial information relating to the topic (De Villiers, 1985). An F(0·1) management procedure is often

THE BENGUELA ECOSYSTEM: PART IV

423

Fig. 30. —Trends in the catch of anchovy Engraulis japonicus and in the breeding population of Cape cormorants Phalacrocorax capensis at islands, at platforms and overall along the Namib coast, 1956–1985 (from Crawford & Williams, in prep.).

recommended (ICSEAF, 1986). Newman (1977) listed maximum potential yields for the different hake stocks as 350000 tons for Divisions 1.3 and 1.4, 210000 tons for Division 1.5, 165000 tons for Division 1.6, and 70000 tons for Divisions 2.1 and 2.2, or a total of 795000 tons for the southeastern Atlantic. Biomass estimates obtained by the swept area method are available for some hake and other groundfish resources in recent years (Hatanaka et al., 1983; Uozumi et al., 1984, 1985; Payne et al., 1985, 1986; Macpherson, Roel & Morales, 1986). Indices obtained by research vessels from different countries are, however, not yet comparable (Payne et al., 1986). The abundances of some species relative to those of M. capensis are shown for some ICSEAF Divisions and years in Table XL. No assessment is available for M. polli, but the relatively small harvests of this species suggest that it has a limited potential yield (De Villiers, 1985). Inada (1981) reported that to the north of Angola M. polli was considered of little economic importance because of low catch rates and a small size (less than 40 cm total length) of fish.

424

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Fig. 31.—VPA estimates of trends in (a) year-class strength and (b) biomass for three stocks of Cape horse mackerel Trachurus capensis in the southeastern Atlantic (information from Crawford, Shelton & Hutchings, 1983; Babayan et al., 1986; Kinloch et al., 1986).

Groundfish species, kingklip. A preliminary cohort analysis for kingklip in ICSEAF Subarea 1 using a value of 0·3 for M was undertaken by Terre (1980). Results suggested that biomass and recruitment increased rapidly in the late 1960s and early 1970s but subsequently decreased (Fig. 33). Trends in catch rates agreed well with estimates of biomass for 1970–1975. Reported catches were low prior to 1972, in which year they climbed dramatically (Table XIV, p. 384). As it was uncertain whether the low catches up until 1971 resulted from low abundance or under-reporting by some nations owing to mis-identification (Terre, 1980), a later analysis (Isarev, 1983) used esti TABLE XL Relative abundance of some species taken by research trawls in some ICSEAF Divisions (from Hatanaka et al., 1983; Uozumi et al., 1984, 1985; Macpherson et al., 1986; Payne et al., 1985, 1986) ICSEAF Divisions

1.5*

1.5*

1.5*

1.5*

1.6

1.6

1.6

1.6

1.6

THE BENGUELA ECOSYSTEM: PART IV

425

Fig. 32.—VPA estimates of trends in (a) year-class strength and (b) biomass for four stocks of Cape hakes Merluccius capensis and M. paradoxus in the southeastern Atlantic (from information in ICSEAF, 1986; Leslie, 1986). Year Months M. capensis M. paradoxus G. capensis Lophius spp. H. dactylopterus A. microlepis T. capensis P. laniarius L. reynaudii

1983 7&8 100 14 1 7 1 1 – – –

1984 1 &2 100 8 1 3 1 0 – – –

1984 7&8 100 111 1 2 1 0 – – –

1985 7&8 100 40 1 3 1 0 – – –

1983 6&7 100 304 19 25 – – – – 5

1984 1 100 171 10 16 – – – – 19

1984 7 100 136 3 3 – – – – 0

1985 1 100 147 5 6 – – – – 5

1985 7 100 77 3 7 – – – – 6

426

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

T. angolensis ICSEAF Divisions Year Months M. capensis M. paradoxus G. capensis Lophius spp. H. dactylopterus A. microlepis T. capensis P. laniarius L. reynaudii T. angolensis

2 1 2.1 &2.2 1980 11 &12 100 – – – – – 64 32 22 –

2 1 2.1&2.2 1981 11&12 100 – – – – – 61 26 20 –

– – 2.1 &2.2 1982 6 100 – – – – – 60 18 10 –

–

–

–

* Includes the lower section (23 to 25° S) of Division 1.4.

mates of kingklip catches derived from known catches of hakes and the ratio of hakes and kingklip obtained in research trawls. This analysis distinguished between the northern (ICSEAF Divisions 1.3 and 1.4) and southern (Divisions 1.5 and 1.6) stocks on the west coast. Overall estimates of biomass were considerably higher, indicating a slight increase during the late 1960s but a decrease after 1973 (Fig. 33). Yield estimated from a production model suggested that a maximum of 10000 tons of kingklip could be harvested annually from Subarea 1 (Terre, 1980). Groundfish species, panga and large-eye dentex. Newman (1977) estimated the maximum potential yield of panga on the Agulhas Bank to be about 11000 metric tons and, from VPA, biomass of fish aged three or older to be of the order of 30000 tons or more. Sato (1980), however, estimated the yield of panga to be less than 7000 tons. Between 1980 and 1982, overall biomass was estimated by direct survey to lie between 31000 and 48000 tons (Hatanaka et al., 1983; Uozumi et al., 1984, 1985). Catch rates decreased noticeably between 1964 and 1975 (Table XLI). De Villiers (1985) stated that ICSEAF scientists recorded a decrease in the biomass of large-eye dentex from about 180000 tons in 1968 to less than 70000 tons in 1976, and that they calculated the annual sustainable yield in 1980 to be 21000 tons. Newman (1977) had earlier noted that stock size appeared to have decreased in fluctuating fashion since 1969 (Table XLI), and had suggested that an annual yield of about 9000 tons might be appropriate. Kuderskaya (1985) considered large-eye dentex to have been overexploited in the mid-1960s, when catches averaged 46000 tons, and estimated biomass to have decreased from 102000 tons in 1977 to 24000 tons in 1981. Groundfish species, soles. A preliminary assessment of the Agulhas sole fishery was undertaken by Botha (1977a, 1978), who used a production model based on catch and effort data for the period 1968–1976 to estimate a maximum sustainable yield of the order of 1100 metric tons. Catch rates appeared to have decreased between two- and threefold during the early 1970s, suggesting a state of over-fishing, and a catch restriction of 700 tons was imposed. The industry subsequently claimed that it had under-reported catches over many years, leading to revised assessments based on data for the period after 1972 TABLE XLI

THE BENGUELA ECOSYSTEM: PART IV

427

Fig. 33.—Time series for ICSEAF Subarea 1 of (a) relative proportions of kingklip Genypterus capensis and hakes Merluccius spp. in research catches (Isarev, 1983) and of commercial catch rates for kingklip, (b) for kingklip estimates from VPA for the entire Subarea of mean fishing mortality for ages four to ten and of biomass of all ages, based on reported catches (Terre, 1980), and (c) estimates for the two stocks of kingklip in Subarea 1 of biomass of all ages, based on catches of kingklip derived from relative proportions of kingklip and hakes in research catches (Isarev, 1983). Relative catch rates of panga Pterogymnus laniarius on the Agulhas Bank, 1964–1975 (from Newman, 1977), and relative stock size of large-eye dentex Dentex macrophthalmus as estimated by trawl surveys, 1969–1974 (from Komarov, 1975) Year

Panga

Dentex

1964 1965 1966 1967 1968 1969 1970 1971 1972 1973

100 78 78 55 48 50 43 27 33 29

100 32 75 82 52

428

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Year

Panga

Dentex

1974 1975

25 19

15

that predicted a maximum sustainable yield of about 930 tons (Badenhorst & Sims, 1982; Badenhorst, 1985). Catch rates showed a fluctuating trend through the 1970s and early 1980s (see Table XXIII, p. 396). From 1978 to 1981 biomass of the southern stock of west coast sole was estimated by multiplying the mean catch per unit area obtained by demersal trawl by an estimate of the area occupied by the stock, assumed equivalent to the extent of the fishing grounds in the heyday of the fishery (Payne, 1985a). Standing-stock estimates and 95% intervals showed a steady decrease over the four-year period (Table XLII). Payne (1985a) considered the otter trawls used in the surveys to have been an inefficient means of catching sole, and the estimates consequently to be conservative. TABLE XLII Estimated biomass and associated 95% confidence limits of the southern stock of west coast sole Austroglossus microlepis, 1978–1981 (from Payne, 1985a): fish younger than two years of age contributed minimally to the estimates Year

Biomass (metric tons)

95% confidence intervals

1978 1979 1980 1981

984 148 142 24

+746 +59 ±70 +10

Predatory fish, snoek and chub mackerel. A preliminary assessment of the snoek resource, based on a production model that used handline catches of snoek as an index of abundance, suggested that overall yields from the southeastern Atlantic should not exceed 34000 tons per annum. Stock size approximately halved following the large international catch of 1978 (Crawford, 1985). VPA has been conducted for chub mackerel for 1954–1975 when the bulk of the catch was taken by the South African purse-seine fleet. Biomass peaked at 250000 tons in 1967, as a result of the powerful cohorts of 1966 and 1967, but was of the order of 100000 tons in the mid-1970s (Crawford, Shelton & Hutchings, 1983). For later years, in which the bulk of the catch was made in the northern Benguela, reliable estimates of the age composition of catches are not available. There is evidence that a powerful year-class was formed in 1977, and a record catch of slightly more than 200000 tons was taken in 1978 (Crawford & De Villiers, 1984). Predatory fish, tunas. Catch-based production models have been used to assess some Atlantic stocks of tunas (ICCAT, 1984), and the perceived states of five resources are briefly sketched below. It is recognized that two stocks of Thunnus alalunga occur in the Atlantic, separated for purposes of assessment by latitude 5° N. Other stocks exist in the Indo-Pacific. South Atlantic albacore is at present exploited at the maximum sustainable yield (MSY) of around 25000 metric tons per annum and catch rates have been stable since 1973. For T. albacares there is uncertainty as to whether one or two stocks exist in the Atlantic. Catch rates declined sharply between 1956 and 1965 (Wise & Fox, 1969), and have not improved subsequently. The recent level of exploitation of 130000–150000 tons per annum is above the MSY of 118000– 131000 tons. ICCAT (1984) note that uncertainty about effort levels corresponding to an MSY is greater than uncertainty concerning the level of MSY.

THE BENGUELA ECOSYSTEM: PART IV

429

With T. obesus there is again doubt as to whether there are one or two stocks in the Atlantic. The fishery for bigeye tuna in the South Atlantic, which takes about 30000 tons per year, is operating around the estimated level of MSY. Insufficient data exist for the development of a production model for T. thynnus. Atlantic bluefin tunas were heavily exploited in the past, annual catches being of order of 25000 tons between 1950 and 1965, but the level of fishing is at present regulated and the annual yield from the Atlantic, excluding the Mediterranean Sea, is around 10000 tons. The Benguela catch is negligible. The annual Atlantic catch of Katsuwonus pelamis exceeds 100000 tons per annum, with most being caught east of 30° W. Although no reliable estimates of MSY are available, it appears that greater yields could be taken. No directed fisheries exist for skipjack and yellowfin tuna in the Benguela region. Crustaceans and molluscs. A variety of methods have been used to assess the major crustacean resources. For example, tagging is used to estimate stock sizes and mean growth rates of west coast rock lobsters, but trends in catch per unit effort and in the sizes of individuals in catches are used to monitor longterm changes in levels of exploitation. Red crabs have been assessed by means of deep-sea photographic surveys (Melville-Smith, 1983, 1985) which showed that densities increased from central Namibia towards the boundary with Angola. The abundance of abalone is at present being assessed by transect surveys conducted by SCUBA divers. Growth rates from tagging and information on the structure of the population will be used to estimate total mortality rates. Population variability In the Benguela system it appears that large, natural fluctuations in the sizes of fish populations, at least those of shoaling epipelagic species, occurred long before fisheries perturbed the stocks. Harvests of guano deposited by seabird predators are considered reliable indices of abundance of their fish prey (Crawford & Shelton, 1978) and have shown considerable variation in both the northern and southern Benguela since at least the turn of the century (see Fig. 28, p. 463). Purse-seine fisheries only commenced during World War II. Similar large fluctuations in the populations of unfished or lightly exploited clupeoids have been observed off Peru, off California, and in the North Atlantic (Blaxter & Hunter, 1982). Fossil fish scales in laminated sediment cores from waters off Namibia suggest major pre-fishery stock fluctuations. Unlike the Peruvian and Californian systems which were dominated by anchovy, the pilchard was, however, historically the dominant species off Namibia (Shackleton, 1986). Following the commencement of commercial fishing operations many of the exploited resources have undergone largescale changes in abundance, as outlined on pp. 458–472. The paragraphs below consider briefly some possible causes of variability of fish populations and their predators in the southern Atlantic. Environmental influences. Environmental factors are a frequently cited cause of variability in fish populations, for example the recovery of the Japanese sardine Sardinops melanosticta resource being attributed to favourable environmental conditions (Kondo, 1980). In the southeastern Atlantic it has been noted that a number of fish stocks may produce strong year-classes simultaneously, and that some large increases in year-class strength can be plausibly related to the environment (Crawford, Shelton & Hutchings, 1983). On account of similarity in the guano records off South Africa and Namibia (see Fig. 28, p. 463) it has been suggested that the same driving mechanism may influence trends in fish populations in both the northern and southern subsystems of the Benguela (Crawford, Shelton & Hutchings, 1983). Shannon, Crawford & Duffy (1984), furthermore, noted similarity between trends in year-class strength for pilchard and anchovy

430

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Fig. 34.—(a) The Freon (1983) trade-wind cycle for northwest Africa and (b) trends in year-class strength of pilchard Sardinops ocellatus and anchovy Engraulis japonicus in the Benguela region (from Shannon, Crawford & Duffy, 1984).

stocks off southern Africa and the trade-wind cycle (Freon, 1983) for northwest Africa (Fig. 34). They supported the view of Kawasaki (1983) that long-term environmental phenomena on a global scale may influence trends in the abundance of fish populations, and also observed that ecosystem trends were frequently similar in both the Pacific and Atlantic, but between these oceans tended to be out of phase. The hypothesis that long-term environmental factors may influence variability of fish populations in the southeastern Atlantic was supported by Shelton, Boyd & Armstrong (1985). These authors drew attention to an alternation between periods of warm and cool sea-surface temperatures (McLain, Brainard & Norton, 1985), and demonstrated that fish populations such as anchovy may be well buffered against random variability but strongly modulated by an autocorrelated environmental signal. Powerful year-classes of chub mackerel were formed in the southeastern Atlantic in 1966 and 1967 when they contributed large catches to the purseseine fleet off the Western Cape (Newman & Crawford, 1980; Crawford, 1981c). Formation of these year-classes followed poor upwelling during the spring-summer

THE BENGUELA ECOSYSTEM: PART IV

431

Fig. 35.—Size distribution of catches of chub mackerel Scomber japonicus from ICSEAF Divisions 1.3, 1.4 and 1.5: year-classes thought to be represented by modes are illustrated in parentheses and highlight the strong 1977 year-class (from Crawford & De Villiers, 1984).

season of 1965–1966 (Crawford, Shelton & Hutchings, 1983). A substantial year-class also seems to have been formed in 1977 (Fig. 35), again after a period of reduced upwelling in the Benguela (Crawford & De Villiers, 1984). Off Namibia, Boyd (1979) found a significant correlation between variability of sea-surface temperatures and year-class strength of anchovies as estimated by VPA. Year-class strength was favoured by reasonably constant temperatures, and it was argued that the feeding success of anchovy larvae was greater in stable conditions (Boyd, 1979). Intraspecific regulation. It has been asserted that clupeoid stocks tend not to have strong densitydependent regulatory mechanisms, resulting in extreme natural variability in recruitment and susceptibility to overfishing (Cushing, 1971; Blaxter & Hunter, 1982). In the Benguela system many of the clupeoid populations have undergone large-scale changes in abundance since the commencement of fishing, and probably even earlier (see Fig. 28, p. 463). There have, for example, been large decreases in the pilchard resources off both South Africa and Namibia. MacCall (1980) regarded cannibalism as indisputably a density-dependent mechanism by which clupeoid populations are regulated, and cannibalism of anchovy eggs by adults has been demonstrated on the Agulhas Bank (Szein-feld, in prep.). Off the Western Cape, size at maturity of pilchard was considerably reduced following severe depletion of the resource in the early 1960s (Crawford, Shelton & Hutchings, 1980; Shelton & Armstrong, 1983; Fig. 36a). Also, the mean condition factor of gonads of both females and males increased markedly (Shelton & Armstrong, 1983; Fig. 37). Similar compensatory increases in the mass of fish (Fig. 38) and of gonads at given lengths were observed for the pilchard stock off Namibia after its collapse in the late 1960s (Thomas,

432

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Fig. 36.—Percentage maturity at length off South Africa’s Western Cape of (a) pilchard Sardinops ocellatus during the periods prior to (1951–1963) and following (1965–1978) depletion of the resource and (b) anchovy Engraulis japonicus during the periods prior to (1965–1973) and following (1974–1979) a substantial increase in the resource as estimated by VPA (from Shelton & Armstrong, 1983).

1986). Size at maturity of anchovy off the Western Cape increased following a possible doubling of the biomass in the mid-1970s (Shelton & Armstrong, 1983; Fig. 36b). For horse mackerel off Namibia, size at 50% maturity decreased from 269 mm in 1977–1978 to 235 mm in 1983, a trend that was attributed to heavy fishing pressure on the resource (Wysokinski, 1984). With regard to groundfish species, length at 50% maturity for the southern stock of west coast sole fell from 24 cm in 1978 to 21 cm between 1979 and 1981 following a fivefold decrease in fish density (Payne, 1979, 1985a). Payne (1985a), furthermore, suggested that differences in size at maturity between the northern and southern stocks of west coast sole may have resulted from a compensatory response by the southern stock to low population levels at the time of investigation. Konchina (1986) noted that cannibalism was characteristic of various species of the genus Merluccius and considered it a mechanism for regulating population abundance. The relationship between parent stock and recruitment for fish species in the Benguela has been investigated by a number of authors mainly from data derived from VPA (e.g. Centurier-Harris, 1977; Getz, 1980; Butterworth, 1983; Shelton & Armstrong, 1983; Shelton, Boyd & Armstrong, 1985; Thomas, 1986).

THE BENGUELA ECOSYSTEM: PART IV

433

Fig. 37.—Seasonal trends in the condition factor of mature (a) female and (b) male pilchard Sardinops ocellatus off South Africa’s Western Cape at various time intervals: the resource was severely depleted in the early 1960s (from Shelton & Armstrong, 1983).

Thomas (1986) demonstrated that when gonad mass, rather than biomass, of the parent stock was plotted against year-class strength there was a marked change in the shape of the stock-recruitment plot for the Namibian pilchard resource. He concluded that a large proportion of the density-dependence exhibited by this stock could be ascribed to variations in gonad mass. Shelton & Armstrong (1983) had earlier theoretically demonstrated that a density-dependent change in fecundity would result in a plot of recruitment against parent stock having a more pronounced dome and more steeply sloping limbs and lead to higher potential yields. Influences of prey variability. There is substantial evidence that in the Benguela system the sizes of predator populations are influenced by variability of prey resources. Much of this evidence relates to coastal-breeding seabirds, for which indices of abundance are relatively easily obtained. More recently, however, similar observations have also been made for the predatory snoek. Even prior to the initiation of commercial exploitation of the smaller pelagic fishes, guano harvests at seabird islands off southern Africa were believed to have been influenced by the local availability of such

434

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Fig. 38.—Relationship between body mass and caudal length of pilchard Sardinops ocellatus off Namibia prior to (1955) and following (1970) severe depletion of the resource (from Thomas, 1986).

fish (Hutchinson, 1950). Crawford & Shelton (1978) later showed that guano harvests could be reliably used as an index of numbers of breeding seabirds, and that in many instances trends in the production of guano were closely related to estimates of the local abundance of forage fishes (cf. also Troadec, Clark & Gulland, 1980). For example, yields of guano at both islands and platforms plummeted following collapses of the pilchard fisheries off South Africa and Namibia (see Fig. 28, p. 463; Fig. 39a). At islands south of Walvis Bay and east of Cape Agulhas guano harvests were significantly related to estimates of the biomass of older pilchard, and at Lamberts Bay to those of younger pilchard (Fig. 39b, c; Fig. 40). In the early 1950s two indices of survival of juvenile Cape gannets, Morus capensis, at an island off the Western Cape were significantly related both to estimates of biomass of pilchard (Fig. 39d) in the region, and to estimates of the combined biomass of pilchard and horse mackerel (Brooke & Crawford, in prep.). These two species were at that time dominant in the diet of gannets (Davies, 1955, 1956a; Rand, 1959b). Also in the early 1950s mean numbers and flock sizes of Cape cormorants, Phalacrocorax capensis, in St Helena Bay compared well with VPA estimates of the year-class strength of pilchard (Crawford, Shelton & Berruti, 1983). Between 1956 and 1978 there were large changes in numbers of birds at the major colonies of Cape gannets, Cape cormorants and jackass penguins, Spheniscus demersus, along the southern African coast, most of which could be related to changes in the abundance of epipelagic fish (Crawford & Shelton, 1981). East of Cape Point trends in the numbers of seabird species showed similarity to the percentage of the overall purse-seine catch off the Western Cape emanating from the same region (see Fig. 5, p. 367). West of Cape Point and off Namibia sizes of the seabird populations were adversely affected by collapses of the pilchard resources, although in some instances trends appeared to have been modified by local availability of anchovy. Large increases in numbers of jackass penguins and bank cormorants Phalacrocorax neglectus at Mercury and Ichaboe Islands in the 1970s probably resulted from the presence then of a large pelagic-goby resource to the north of Lüderitz (Cooper, 1981; Crawford & Shelton, 1981; Crawford, Cruickshank, Shelton & Kruger, 1985). Three large increases in the availability of snoek to handline fishermen off the Western Cape since the 1890s have each been significantly related to estimates of the abundance of prey or of other predators of

THE BENGUELA ECOSYSTEM: PART IV

435

Fig. 39.—Relationship between indices of abundance of pilchard Sardinops ocellatus and the production of guano at (a) a seabird platform near Walvis Bay, (b) a seabird island slightly north of Lüderitz, (c) a seabird island near Port Elizabeth, and (d) between biomass of pilchard and indices of first-year survial of Cape gannets Morus capensis at a seabird island off South Africa’s Western Cape (from Crawford & Shelton, 1978; Brook & Crawford, in prep.).

shoaling epipelagic fishes (Crawford & De Villiers, 1985). Catch rates of snoek were high during the late 1890s, as were handline catches of snoek in the late 1920s. Both these periods coincided with large peaks in the amount of guano collected from Bird Island at Lamberts Bay (Fig. 40), which is situated at the centre of

436

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

the recruitment grounds for pilchard, anchovy, horse mackerel, and round herring off the Western Cape (Crawford, 1980a). As pointed out above the third known large peak in the production of guano at this island, that occurring in the 1950s, was related to the formation of powerful year-classes of pilchard (Crawford & Shelton, 1978). Unfortunately, there are no records of the performance of the snoek fishery during the 1950s, although in a questionnaire survey many fishermen selected the 1950s as the period during which snoek had been most abundant off the Western Cape (Crawford & Kriel, 1985). Handline catches of snoek off the Western Cape showed a major peak in the mid-1970s, and between 1969 and 1981 were significantly correlated with VPA estimates of anchovy abundance (Crawford, Shelton & Hutchings, 1983; see Fig. 26, p. 460). The doubling of anchovy biomass during the 1970s as suggested by VPA was caused by a substantial increase in the catch of adult fish east of Cape Point after 1973, an increase that may have resulted from greater availability of these fish to the purse-seine fleet (Armstrong, Shelton & Prosch, 1985). It is possible that the improved handline catches of snoek resulted from snoek following anchovy into inshore waters, in which case there would have been a relationship between the local abundances of predators and their prey. Off Namibia handline catches of snoek have in recent years been related to the biomass of young horse mackerel, the dominant forage species in the area (Fig. 41). Influences of predation. In the open-sea Benguela the influence of predators on variability of prey populations is considerably less obvious than the dependence of predators on adequate levels of forage resources. Crawford & De Villiers (1985) used frequency of occurrence of anchovy in the diet of snoek, abundance of snoek as indicated by handline catches off the Western Cape, and VPA estimates of anchovy biomass to calculate annual indices of the mortality of anchovy attributable to snoek. The index showed inter-annual variation with a steady increase during the 1970s (Fig. 42), but was positively related to anchovy biomass (r2=0·81) as also to the index of consumption of anchovies by snoek (r2=0·94). Of interest was the observation that the TABLE XLIII Catch rates (kg per trawl) by research vessel of kingklip Genypterus capensis, west coast sole Austroglossus microlepis and gurnard Chelidonichthys spp. off the Orange River, 1978–1981 (from Payne, 1985a) Species

1978

1979

1980

1981

Kingklip Sole Gurnard

1·2 58·6 103·6

0·2 10·7 49·8

24·2 9.4 14·0

109·0 1·2 0·9

mortality indices were relatively low during 1969–1972, immediately preceding the upswing in anchovy biomass as estimated by VPA, but high in 1979–1980, prior to the subsequent downswing. Loss of anchovies to other predators probably followed a similar trend (Crawford & De Villiers, 1984). The chub mackerel population of the southeastern Atlantic declined rapidly between 1969 and 1971 as the powerful 1966 and 1967 cohorts passed through the fishery, and in these years anchovy contributed minimal quantities (0– 2·3%) to the food of chub mackerel of all sizes collected off the Western Cape (Baird, 1978b). By contrast when VPA estimates of the biomass of anchovy off the Western Cape were high during 1979 and 1980 anchovy formed 50% of the food of chub mackerel (Nepgen, 1982). Numbers of Cape gannets, a seabird predator of anchovy, breeding at islands off the Western Cape were depressed during the late 1960s relative to the late 1970s (16000 as opposed to 26000 pairs; Crawford, Shelton, Cooper & Brooke, 1983). VPA estimates of anchovy abundance have, however, recently been called into question (Armstrong et al., 1985), so that perceived trends in the natural mortality rate of anchovy may not reflect the real situation.

THE BENGUELA ECOSYSTEM: PART IV

437

Fig. 40.—Trends in the production of seabird guano at Bird Island, Lamberts Bay, in catch rates of snoek Thyrsites atun and in year-class strength of pilchard Sardinops ocellatus off South Africa’s Western Cape (from Crawford, Shelton & Hutchings, 1983).

438

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Fig. 41.—Trends in the abundance of one-year-old Cape horse mackerel Trachurus capensis and in the handline catch of snoek Thyrsites atun off Namibia (from ICSEAF statistics and unpublished records of the Sea Fisheries Research Institute).

Fig. 42.—Relationship between the biomass of anchovy Engraulis japonicus and an index of anchovy mortality calculated as being attributable to predation by snoek Thyrsites atun off South Africa’s Western Cape, 1969–1980 (from Crawford & De Villiers, 1985).

Other examples of predators influencing prey populations are limited. Payne (1985a) refers to a “population explosion” of goby and mantis shrimp off the Orange River in 1979, which he associated with decreased catch rates of most predators of these two species in that year. In the same region kingklip were found to contain “…significant quantities of…small sole”, and during 1978 to 1981 kingklip increased in abundance in demersal trawls on research cruises while catch rates of soles and gurnards, another prey of kingklip (Payne, 1986), decreased (Table XLIII). Competition and altered trophic flow. Species replacements in catches have occurred in both the northern and southern regions of the Benguela, where severe decreases in harvests of pilchard have been followed by greatly improved yields of horse mackerel and anchovy. VPA analyses suggested that collapses of the pilchard stocks were followed by increased biomasses of horse mackerel and anchovy, and the growth of populations of these two species estimated by VPA closely matched growth that might have been expected from the Verhulst-Pearl logistic model once catches had been accounted for (Crawford, Shelton & Hutchings, 1983). The food of pilchard and anchovy is similar (King & Macleod, 1976), and off Namibia in

THE BENGUELA ECOSYSTEM: PART IV

439

Fig. 43.—VPA estimates of the biomass of pilchard Sardinops ocellatus, Cape horse mackerel Trachurus capensis and Cape hakes Merluccius capensis and M. paradoxus off Namibia, showing expansion of the horse mackerel resource following collapse of the pilchard resource and the inverse trends in biomasses of horse mackerel and Cape hakes (from information in Babayan et al., 1986; ICSEAF, 1986; Thomas, 1986).

the early 1970s the copepods Calanoides carinatus and Centropages brachiatus were important in the diets of young pilchard and young horse mackerel (King & Macleod, 1976; Venter, 1976). Young pilchard, anchovy and horse mackerel frequently co-occur in mixed catches (Crawford, 1980a). As mentioned above, there is doubt concerning reliability of the VPA estimates of anchovy biomass. A poll of fishermen, however, indicated that a large majority (37 of 45) held the opinion that there had been an increase in the anchovy resource off the Western Cape since the mid-1950s (Crawford & Kriel, 1985). Most fishermen believed the increase was during the early 1960s and no one was aware of exploitable quantities of anchovy prior to 1953. In research catches of juvenile fish with blanket nets at inshore stations off the Western Cape, anchovy contributed an annual average of 4% to the total numbers of fish caught between 1955 and 1959 and 30% between 1960 and 1965. Equivalent values for pilchard were 25% and 10%; for horse mackerel 32% and 20% (Sea Fisheries Research Institute, unpubl. records). Off Namibia there has also been considerable interest in a possible interaction between horse mackerel and Cape hakes. Wysokinski (1981) related an increase in biomass of horse mackerel in the late 1970s to a greatly decreased abundance of hakes at the same time. VPA estimates of biomass of Cape hakes, horse mackerel and pilchard off Namibia are shown in Fig. 43. Krzeptowski (1982) noted the absence of any clearly distinguishable predator-prey relationship between horse mackerel and hakes, but found marked similarity in their diet and feeding grounds off northern Namibia. He concluded that a decline in population size of one of the two species could result in improved trophic conditions for the other. Andronov (1983) noted especially similarity in the diets of horse mackerel larger than 20 cm and hake of length 15–24 cm, which he believed could lead to competition during periods of scarcity of food. Later Konchina (1986) drew attention to the fact that horse mackerel was important in the diet of hake in 1985, so that any interspecific relationship might be more complex than previously envisaged.

440

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Fig. 44.—Probable alteration in trophic flow thought to have occurred in the intense perennial upwelling region north of Lüderitz (from Crawford et al., 1985). For ocellata please read ocellatus.

Evidence has been presented for altered trophic flow in the intense, perennial upwelling area situated off Namibia between latitudes 22 and 27° S (Crawford et al., 1985). It appears likely that in this region the pelagic goby resource expanded following depletion of adult pilchard, which have a similar diet, and the change was reflected in the structure of the seabird community (Fig. 44). Between Walvis Bay and Lüderitz, colonies of Cape gannets preying largely on pilchard showed large decreases in abundance, whereas numbers of jackass penguins, Cape cormorants, and bank cormorants, birds that were able to utilize the goby resource, underwent large increases (Crawford et al., 1985). Off Namibia, in addition to probable increases in stocks of horse mackerel on the northern grounds and of gobies in the central region, Newman (1970a) and Thomas (1985) considered there to have been an increase in the anchovy population. Exploratory fishing for anchovy off Namibia commenced in 1963 but “…during this year the anchovy abundance was extremely low, and negligible catches were recorded” (Newman, 1970a, p. 12). By 1964, 16 boats were searching for anchovy, their earnings being supplemented by the remainder of the fishing fleet (Thomas, 1985). In spite of this effort, annual catches remained small

THE BENGUELA ECOSYSTEM: PART IV

441

until 1968, leading Thomas (1985) to conclude that the anchovy stock was low at a time when pilchard biomass was high (see Fig. 27, p. 462). Catches of anchovy increased dramatically from 1968 while the pilchard resource was collapsing (see Fig. 30, p. 465). Anchovy contributed less than 2% to the diet of each of Cape gannets, jackass penguins, and Cape cormorants shot near Walvis Bay between 1957 and 1959 (Matthews, 1961; Matthews & Berruti, 1983), but 53% of the food of Cape gannets sampled at islands south of Walvis Bay between 1978 and 1982 (Crawford et al., 1985). Cape cormorants generally prey heavily on anchovy when the species is plentiful (e.g. Duffy et al., 1984) and numbers of breeding Cape cormorants along the Namib coast have closely matched commercial harvests of anchovy (see Fig. 30, p. 465). Thomas (1985) furthermore asserts that fishermen reported an increase in numbers of coelenterates, particularly Chrysaora sp. (Discomedusae: Semaeostomeae), off Namibia in the 1970s. Based on 25 sets he estimated biomass of Chrysaora between latitudes 18 and 26° S and within 50 km of the coast to be 13·5 million metric tons in February 1980. Off Namibia initial policy with regard to the anchovy was to exploit the resource as heavily as possible to minimize possible competition with pilchard (Butterworth, 1983) and, although the anchovy resource may have expanded, it was primarily horse mackerel that replaced pilchard in the regional catches (Fig. 45) and the pelagic goby that became abundant south of Walvis Bay (Crawford et al., 1985). Off South Africa introduction of the small-mesh net, coupled with close proximity of most processing plants to the recruitment grounds, led to heavy exploitation of horse mackerel young of the year and early depletion of strong year-classes in 1965 and 1969 (Crawford, 1981d). In this region reduced yields of pilchard were largely offset by increased landings of anchovy (Fig. 46). The inference is that, when a resource collapses, the likelihood of a second species replacing it may be influenced by the extent to which that second species is exploited immediately following the collapse of the first. Similar replacements in catches have occurred off Japan, California and Peru-Chile. Pilchard (sardine), anchovy and horse (jack) mackerel have never been simultaneously abundant in catches from these regions, nor in those from the other two systems in which they have supported large catches, the Benguela and Canary Currents. This observed alternation of species in catches can be replicated by a simple model that assumes both competition and fishery interactions, but this finding does not invalidate other possible explanations for the replacements (Silvert & Crawford, in prep.) such as an opposite response by the species to some environmental signal. It seems reasonable to assume, however, that drastic changes in ecology have occurred when species have been reduced to low levels, and it remains of considerable importance to ascertain how the food they consumed prior to collapse is subsequently utilized (Saville, 1980). Exploitation. It is often difficult to distinguish between the influence of environmental factors and fishing, but harvests of guano at islands off both South Africa and Namibia fell following collapses of the respective pilchard stocks to lower levels than any recorded previously (see Fig. 28, p. 463). This suggests that the overall abundance of epipelagic fish in the two regions was reduced below the range of fluctuations that had occurred in the unperturbed state. For the hake stocks Newman (1974) demonstrated that mesh sizes used by bottom-trawlers prior to 1975 had resulted in growth overfishing, and there is little doubt that the introduction of small-meshed purseseine nets drastically reduced the size and age structures of a number of epipelagic fish resources in the southeastern Atlantic by increasing mortality on the younger ages (Newman & Crawford, 1980). In addition, density-dependence of the catchability coefficient (Butterworth, 1983; Shelton & Armstrong, 1983) has meant that it has been possible to maintain catches of epipelagic species when stock sizes have been decreasing (e.g. see Table VI, p. 372). With the possible exception of Thunnus alalunga, the quantity of tunas caught in the Benguela and adjacent areas is insignificant compared with the overall exploitation of Atlantic stocks. Apart from

442

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Fig. 45.—Cumulative catches of major species groups in the northern Benguela region, 1950–1984.

Katsuwonus pelamis, the other species of tunas occurring in or near the Benguela at present appear to be maximally exploited and it seems likely that the virtual disappearance of Thunnus maccoyii and T. thynnus from the region was linked with overall declines in the adult populations caused by heavy non-local exploitation. A CONCLUDING PERSPECTIVE The overall catches from the southeastern Atlantic rose rapidly through the 1950s and 1960s, but subsequently stabilized at a level of above 2 million tons (see Fig. 1, p. 354). When the region is viewed as two subsystems, the northern and southern Benguela, the same pattern of expansion followed by stabilization remains evident, although the yield from the southern Benguela has been considerably more stable as well as smaller than that from further north (Figs 45 and 46). In both instances the species composition of the overall catch was subject to considerable change. Major shifts in species dominance occurred relatively rapidly, over periods of about four to seven years, but infrequently so that, for example, anchovy has already provided substantial yields in the southern Benguela for more than 20 years and horse mackerel in the northern subsystem for more than 10. Overall stability of the finfish catch in spite of underlying changes in the species composition has been observed in other major marine systems of the world (e.g. Laevastu & Larkins, 1981; Silvert, 1982; Brown et al., 1983). The brief survey of feeding conducted in this paper suggests that many of the more important species have a fairly catholic diet, being opportunistic, rather than selective, feeders that eat whatever is available in the system at any given place and time. For example, Cape hakes have a broadly based diet that shows considerable variation from region to region and year to year. Thus, pelagic gobies have been an important food source off Namibia, myctophids of varying genera along the entire western coast of southern Africa, and clupeoids over the Agulhas Bank (see Table XXXVII, p. 421). Off the Western Cape a high degree of

THE BENGUELA ECOSYSTEM: PART IV

Fig. 46.—Cumulative catches of major species groups in the southern Benguela region, 1950–1984.

443

444

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

cannibalism has been recorded (e.g. Botha, 1980). Diet of kingklip has been shown to vary in accordance with the geographical composition of prey species (Macpherson, 1983a), whereas various workers have concluded that pilchard, anchovy, and horse mackerel are non-selective feeders (e.g. Hecht, 1976; King & Macleod, 1976; Konchina, 1986). Interannual changes in the diets of snoek and some seabirds have been related to changes in the relative abundance of prey species (e.g. Crawford & Shelton, 1978, 1981; Nepgen, 1979a). A likely consequence of non-selectivity in the diet of many species is that a particular prey species will be fed on by a number of predators. Indeed, similarity has been shown to exist in the diets of pilchard, anchovy, and smaller-sized horse mackerel (e.g. King & Macleod, 1976), as well as in those of larger horse mackerel and Cape hakes (e.g. Krzeptowski, 1982; Andronov, 1983), of snoek and chub mackerel (e.g. Nepgen, 1982), and of various seabird species (e.g. Crawford & Shelton, 1978). The pelagic goby is preyed on by a wide variety of organisms ranging from groundfish such as the west coast sole, Cape hakes, monkfish, kingklip, and large-eye dentex, through Cape horse mackerel to the Cape fur seal and seabirds. Lanternfish are eaten by many organisms, and other examples are plentiful (see pp. 414–431). In addition to similarities in feeding, the distributions of a number of species overlap, some areas appearing to be generally suitable for spawning and others as nursery grounds. In both northern and southern Benguela a number of species spawn in transitional areas between Benguela Current water and warm surface water advected into the system, but away from sites of strong offshore transport (Shelton et al., 1985). Furthermore, Shannon, Agenbag & Buys (1987) observed that onshore eastward flow could be inferred in the far northern Benguela region during the first quarter of the year, the main anchovy spawning season. Juveniles of commercially important species frequently occur nearer the major upwelling plumes where plankton standing stocks are consistently high (see Figs 14, 17, and 19, pp. 434, 439, and 442, respectively) and may shoal together (Crawford, 1980a). A further generalization is that there are often both longshore and offshore gradients in fish size, with larger individuals frequently occurring at the extremities of the system (e.g. Fig. 18, p. 440) or away from the coast. A distinct seasonal signal characterizes the physical environment of the Benguela (Shannon, 1985; Shelton et al., 1985), with the result that many of the fish species are subject to seasonal patterns of spawning, distribution, migration, condition, recruitment, and availability to the various forms of fishing gear (e.g. Crawford, 1980a). In this sense fishing operations are generally predictable, both fishermen and processors being able to schedule their activities. The similarities in diets and distributions of important species clearly provide a basis for possible changes in energy flow within the system (e.g. Fig. 44), and there is strong evidence to suggest that there have been alterations in the trophic flow (e.g. Crawford et al., 1985). Therefore, it may not be unreasonable to interpret species replacements in catches as reflecting changes in the relative abundance of species in the ecosystem. Dogmatism is precluded by the fact that some species, such as anchovy, only became important commercial targets after other resources had been severely depleted. The likely levels of abundance of such species prior to their exploitation will continue to be argued. Evidence reviewed in this paper points, however, to the fact that both scientists and fishermen active at the time regarded anchovy as scarce prior to the collapses of the pilchard resources (Newman, 1970a; Crawford & Kriel, 1985), and there is evidence, such as research catches of juvenile anchovy, that lends support to this viewpoint (see p. 484). Should production remain more or less constant (e.g. Shannon & Field, 1985), simple whole-system aproaches to resource management, such as those of Newman, Crawford & Centurier-Harris (1974), disregarding the species base, or of Moloney & Field (1985), employing a size-spectrum model (Sheldon, Sutcliffe & Paranjape, 1977), may not be inappropriate. The mechanism whereby altered trophic flow and species replacements in the Benguela may be brought about remains uncertain. A substantial decrease in an abundant resource is necessary, but it is not apparent

THE BENGUELA ECOSYSTEM: PART IV

445

whether alternation of species has resulted from exploitation decreasing the originally dominant populations and subsequent competitive replacement by other species, or from a change in environmental conditions bringing about a shift in species dominance. Debate about the relative importance of fishing and the environment in influencing fish-stock trends has a lengthy history (Clark & Marr, 1955). In the Benguela there is reason to suppose that both factors may be operative. Indeed, off Namibia intrusion of warm water was shown to increase vulnerability of pilchard to purse-seiners by temporarily enhancing local availability (Thomas & Boyd, 1985). In many instances resource depletions have followed immediately upon the taking of exceptionally large catches. The various fisheries for pilchard, horse mackerel, and hake provide clear examples, and often researchers have attributed the large decreases in abundance to excessive harvests (e.g. Cram, 1980; Troadec et al., 1980). The introduction of small-meshed nets has drastically reduced the size and age structure of a number of fish populations in the southeastern Atlantic, and off the Western Cape has led to a recent dominance in purse-seine catches of species such as anchovy having a short generation time. After a reduction in mesh size during the 1960s longer-lived species, for example pilchard and horse mackerel, have contributed substantially smaller landings, and a high fishing mortality on younger ages has kept the parent populations at low levels (Newman & Crawford, 1980; Crawford, Shelton & Hutchings, 1983). Interestingly, off Namibia horse mackerel has generally been subject to exploitation with much larger mesh sizes, 60 mm as opposed to 13 mm, and has since the mid-1970s been the dominant species in catches. Decreases in the mean age of fish in the two pilchard stocks have led to reductions in the geographical range of egg production. This could adversely affect the probability of larvae encountering favourable environmental conditions, increase chances of poor year-classes and decrease the likelihood of powerful year-classes (Newman & Crawford, 1980; Saville, 1980; Crawford, Shelton & Hutchings, 1983). With regard to environmental influences, a progressive warming off the Namib coast between 1959 and 1963 led to a reduced reproductive output by pilchard (Stander & De Decker, 1969). Later workers have speculated on the possibility of long-term environmental events influencing composition of the marine fish fauna (e.g. Crawford & Crous, 1982; Shannon, Crawford & Duffy, 1984; Shelton, Boyd & Armstrong, 1985). Long-term changes in the distribution of catches of some species have occurred and, although these could be attributed in some areas to a reduction of stocks by exploitation and a resultant change in the deployment of fishing effort, in a number of instances an environmental influence appears the more likely cause. A southward shift in the distribution of epipelagic stocks off Angola during the late 1950s and early 1960s is suggested (see Fig. 8, p. 373). During this period the relative contribution of Cunene horse mackerel to catches in ICSEAF Division 1.2 and of pilchard to those in Division 1.3 became progressively less, whereas sardinellas in Division 1.2 and Cunene horse mackerel in Division 1.3 increased in importance. Stander & De Decker (1969) noted a progressive warming off Namibia between 1959 and 1963, which led to a southerly displacement of pilchard and the opening of a processing plant at Lüderitz (Butterworth, 1983). During the late 1970s and early 1980s there was a northward shift in the distributions of catches of many species off Namibia and Angola, including anchovy (see Fig. 7, p. 370), Cunene and Cape horse mackerel (see Table II, p. 358) and large-eye dentex (see Table XVII, p. 387). Cram (1977) reported that in the early 1970s boats were forced to search further north to find fish, and there has been a clear northward shift in location of the major breeding colonies of jackass penguins off Namibia after the 1960s (Fig. 47). Also during the 1970s there were marked changes in patterns of fishing off South Africa. The percentage of the purse-seine catch (see Fig. 5, p. 367) and of the handline catch of snoek (Nepgen, 1979b) made to the east of Cape Point increased markedly. At the same time catch rates of red fishes in the vicinity of Cape Agulhas improved (Crawford & Crous, 1982) and there were increases in seabird populations breeding east

446

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

of Cape Point (Fig. 5) and decreases in those breeding further west (Crawford & Shelton, 1981). Thus, in both northern and southern regions the shifts in distribution during the 1970s were towards the extremities of the Benguela system. Whether or not these apparent ecosystem shifts were environmentally induced, e.g. by changes in the position of the South Atlantic anticyclone (high pressure cell) or changes in the intensity of the pressure gradient, is not clear. J.Taunton-Clark (pers. comm.) has, however, noted that the upwelling wind stress at a site in the southern Benguela during the 1970s was on average higher than in the preceding one and a half decades. Long-term changes in the distribution of hake larvae have been observed in the Californian upwelling system (Bailey, 1980), and MacCall (1984) has noted geographical shifts in Peruvian fish production. In the Benguela exceptionally large year-classes occurring infrequently have been known to sustain some fisheries over a number of years (Newman & Crawford, 1980; Crawford, Shelton & Hutchings, 1983), and it has been suggested that the real problem in fisheries prediction is understanding the occurrences of such very large year-classes (Brown et al., 1983). For chub mackerel in the southeastern Atlantic, formation of powerful year-classes has followed seasons of reduced upwelling (Crawford & De Villiers, 1984). Off California a negative correlation between upwelling and year-class survival of chub mackerel has been reported (Lasker & MacCall, 1983), and the strongest year-classes of Pacific hake Merluccius productus have been recruited from spawning during winters having the strongest onshore transport (Bailey, 1981). In the Benguela system strong year-classes may be formed simultaneously for a number of fish stocks (Crawford, Shelton & Hutchings, 1983). For example, all four stocks of Cape hakes gave rise to powerful year-classes in the early 1980s (see Fig. 32, p. 468). Koslow (1984) noted positive correlations in recruitment of stocks of the same species and significant positive correlation in the recruitment of certain groups of species in the northwestern Atlantic. A period of particular interest in the Benguela is that of the early 1970s. Shifts in the distribution of catches of some species and in the centres of distribution of some predators during the 1970s towards the extremities of the system have been noted above. Nearer the centre of the system, catches of the southern stock of west coast sole and production of lobster tails at Port Nolloth decreased to small, even negligible, levels between the mid-1960s and the early or mid-1970s (see Table XXII, p. 395; Pollock, 1982). Payne (1985a) documented an offshore migration of the northern stock of west coast sole in the early 1970s (see Fig. 24, p. 452). At the same time catch rates of kingklip increased (see Fig. 33, p. 470) and strong yearclasses of a number of species were formed (e.g. Figs 27, 31, 32, and 34), although for anchovy the VPA may rather have reflected increased availability to the fishing fleet (Armstrong et al., 1985). In the vicinity of Mossel Bay, Botha (1978) recorded an increase in availability of Cape hakes around 1972–1973, and after 1972 a serious decline in abundance of Agulhas sole. Off northwestern Africa, also in the early 1970s, both catches and catch rates of Sardina pilchardus decreased markedly in the northern region but increased further south (Delgado & Fernandez, 1985). In 1973 the Sardinella fisheries off Ghana and the Ivory Coast collapsed, whereas catches of sardinellas off Senegal increased (Troadec et al., 1980; Freon, 1983). Major changes occurred in other of the world’s fisheries (Shannon, Crawford & Duffy, 1984). During 1972 and 1973 there was a major positive sea surface temperature anomaly off the west coast of Africa, from about 15° S to the tip of the continent (McLain, Brainard & Norton, 1985). If the most important events operate over large scales and affect many species, it may well be that occurrences of the infrequent, exceptionally powerful year-classes will be best understood by a concentration on macro-events rather than those of smaller magnitude. Many of the most abundant fish species in the Benguela spawn over a wide area, of the order of thousands of square kilometres (e.g. Figs 14, 17, and 19), and have an extended spawning season, of the order of months (e.g. Crawford, 1980a). Some species are believed to undertake serial spawning (e.g. Le Clus, 1979a). Therefore there is a large degree of

THE BENGUELA ECOSYSTEM: PART IV

447

Fig. 47.—Percentage contribution of different islands to the overall breeding population of jackass penguins Spheniscus demersus off Namibia during the four most comprehensive censuses, 1956–1985: islands are listed from the southernmost (Sinclair) to the northernmost (Hollams Bird), and the estimated number of breeding pairs of jackass penguins during each census is indicated in parentheses; the population off Namibia decreased following collapse of the pilchard resources, and the bulk of the breeding population has shifted from the southern to the northern islands (from Crawford & Williams, in prep.).

“bet-hedging”, and as a result some fish populations appear buffered against high-frequency variability (Shelton, Boyd & Armstrong, 1985).

448

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

The most obvious species interaction in the open-sea Benguela is the dependence of predators on suitable levels of prey abundance (see pp. 478–480). Given that most resources are subject to a certain amount of variability there is obvious advantage in non-specialized feeding, and it has been shown above that many species have catholic diets. One may conclude that the greater the ability of a species to exploit a wide range of food organisms the less subject it will be to fluctuations in the abundance of any particular form of forage. This appears to be the case. For example, there has been a very large recent increase in the population of Cape fur seals (Butterworth, David, Rickett & Xulu, 1987). It may be argued that this increase has been a recovery after exploitation had reduced the population to a low level. Such a recovery, however, need not necessarily occur. Numbers of jackass penguins have continued to decrease, even after the termination of egg collections, probably as a result of food shortage (Shelton, Crawford, Cooper & Brooke, 1984). Jackass penguins, being flightless, have a limited foraging range when breeding (Frost, Siegfried & Cooper, 1976), and are not able to exploit the same diversity of food organisms as Cape fur seals (Sea Fisheries Research Institute, unpubl. records). There is limited evidence (pp. 482–483) that predators may influence the variability of prey populations, and cannibalism by Cape hakes, itself an opportunistic feeding behaviour, may have an important impact on the natural mortality rates of younger individuals (e.g. Botha, 1980). Interspecific competition for breeding space may occur at some offshore islands (Burger & Cooper, 1984), and intraspecific controls have been shown for a number of species (pp. 475–478), suggesting that energy may limit the overall biomass of fish species and that at low levels of abundance population variability is to some extent biologically controlled. Finally, it is of interest to compare variability of fish production from the northern and southern subsystems of the Benguela (Figs 45 and 46), regions that have been subject to somewhat different management approaches. After South Africa declared a fishing zone of 200 nautical miles during 1977, control over the exploitation of resources in the southern Benguela has been tighter than that in the northern subsystem where the fishery is considerably more international in nature. Furthermore, for a lengthy period the epipelagic stocks in the southern Benguela were managed as a unit by means of a combined-species quota (e.g. Newman, 1984), whereas in the northern sub-system, where the species base to the purse-seine fishery was not as broad, single-species quotas were frequently in force (e.g. Butterworth, 1983). The catch from the southern Benguela has exhibited a far greater stability than that from farther north, where twice large peaks seem to have resulted from the harvesting of powerful year-classes of pilchard (see Fig. 27, p. 462) and once from greatly improved recruitment of horse mackerel (see Fig. 31, p. 466). It appears that the fishery in the northern Benguela has been better able to exploit strong year-classes, but this has been at the expense of greater stabiity in the catch. Furthermore in the mid-1970s the pilchard collapsed to a very low level of abundance. The pilchard off South Africa also collapsed after the entry of a powerful year-class and expansion of the purse-seine fleet, during a time when there were no quota restrictions on the catch (Crawford, 1981a). The later assumption of a constant production may have limited harvests from strong cohorts, but equally there appears to have been a greater risk of resource collapse attached to more dynamic forms of management. ACKNOWLEDGEMENTS We are indebted to many of our colleagues for reading and commenting on various parts and stages of the manuscript, especially M.J.Armstrong, A. Badenhorst, H.G. van D.Boonstra, G. De Villiers, R.W.Leslie, A.I.L. Payne, R.M.Prosch, and F.H.Schülein. The artwork was prepared by A. van Dalsen and his assistants. We thank C.J.Augustyn, R.Melville-Smith, S.C.Pillar, and R.J.Q.Tarr for information.

THE BENGUELA ECOSYSTEM: PART IV

449

REFERENCES Albaret, J.J. 1976. ICCAT Coll. Vol. scient. Pap., 5(1), 94–100. Alekseev, F.E. & Alekseeva, E.I., 1980. ICCAT Coll. Vol. scient. Pap., 9(3), 695– 703. Andronov, V.N., 1983. Colln scient. Pap. int. Commn S.E. Atl. Fish., 10(1), 1–6. Andronov, V.N., 1985. Colln scient. Pap. int. Commn S.E. Atl. Fish., 12(1), 1–16. Anonymous, 1946. S. Afr. Shipp. News Fishg Ind. Rev., 1(4), 30–33. Armstrong, M.J., 1984. S. Afr. J. mar. Sci, 2, 131–144. Armstrong, M.J., 1986. Colln scient. Pap. int. Commn S.E. Atl. Fish., 13(1), 25–44. Armstrong, M.J. & Butterworth, D.S., 1986. Colln scient. Pap. int. Commn S.E. Atl. Fish., 13(1), 45–53. Armstrong, M.J., Shelton, P.A. & Prosch, R.M., 1985. Colln scient. Pap. int. Commn S.E. Atl. Fish., 12, 17–29. Assorov, V.V. & Berenbeim, D.Y., 1983. Colln scient. Pap. int. Commn S.E. Atl. Fish., 10(1), 27–30. Assorov, V.V. & Kalinina, M.I., 1979. Colln scient. Pap. int. Commn S.E. Atl. Fish., 6, 219–227. Assorov, V.V. & Shcherbich, L.V., 1979. Colln scient. Pap. int. Commn S.E. Atl. Fish., 6, 229–232. Augustyn, C.J., in press. S. Afr. J. mar. Sci. Babayan, V. & Bulgakova, T.I., 1983. Colln scient. Pap. int. Commn S.E. Atl. Fish., 10(1), 49–53. Babayan, V., Kolarov, P., Prodanov, K., Komarov, Y.A., Vaske, B. & Wysokinski, A., 1985. Colln scient. Pap. int. Commn S.E. Atl. Fish., 10(1), 55–62. Babayan, V., Kolarov, P., Prodanov, K., Vaske, B. & Wysokinski, A., 1986. Colln scient. Pap. int. Commn S.E. Atl. Fish., 13(1), 65–69. Badenhorst, A., 1985. Colln scient. Pap. int. Commn S.E. Atl Fish., 12(1), 49–52. Badenhorst, A. & Boyd, A.J., 1980. Fish. Bull. S. Afr., No. 13, 83–106. Badenhorst, A. & Sims, P.F., 1982. Colln scient. Pap. int. Commn S.E. Atl. Fish., 9(2), 53–59. Bailey, K.M., 1980. CalCOFI Rep., 21, 167–171. Bailey, K.M., 1981. Mar. Ecol. Prog. Ser., 6, 1–9. Baird, D., 1971. Investl Rep. Div. Sea Fish. S. Afr., No. 96, 19 pp. Baird, D., 1974. Ph. D. thesis, University of Stellenbosch, 241 pp. Baird, D., 1977. Zool. Afr., 12(2), 347–362. Baird, D., 1978a. Fish. Bull. S. Afr., No. 10, 50–61. Baird, D., 1978b. Fish. Bull. S. Afr., No. 10, 62–68. Baptista, S.R. da F., 1977. Colln scient. Pap. int. Commn S.E. Atl. Fish., 4, 21–24. Baptista, S.R. da F., 1979. Fish. Bull. S. Afr., No. 12, 44–48. Barber, R.T. & Haedrich, R.L., 1969. Deep-Sea Res., 16(1), 105–106. Bard, F.X., 1982. ICCAT Coll. Vol. scient. Pap., 17(2), 487–490. Barkai, R. & Griffiths, C.L., 1986. S. Afr. J. mar. Sci., 4, 37–44. Barnard, K.H., 1927. Ann. S. Afr. Mus., 21, 797–1065. Batchelor, A.L., 1982. M. Sc. thesis, University of Port Elizabeth, 53 pp. Beddington, J.R. & Cooke, J.G., 1983. FAO Fish. tech. Pap., No. 242, 47 pp. Bergh, M.O., Butterworth, D.S. & Andrew, P.A., 1986. Colln scient. Pap. int. Commn S.E. Atl. Fish., 13(1), 77–90. Bergh, M.O., Field, J.G. & Shannon, L.V., 1985. Int. Symp. Upw. W. Afr., Inst. Inv. Pesq., Barcelona, 1985, 1, 281–304. Beyers, C.J. de B., 1979. Investl Rep. Sea Fish. Brch S. Afr., No. 117, 26 pp. Beyers, C.J. de B. & Wilke, C.G., 1980. Fish. Bull. S. Afr., No. 13, 9–19. Blaxter, J.H.S. & Hunter, J.R., 1982. Adv. Mar. Biol., 20, 1–223. Boerema, L.K., Saetersdal, G., Tsukayama, I., Valdivia, J.E. & Alegre, B., 1965. FAO Fish. tech. Pap., No. 55, 44 pp. Botha, L., 1970. Investl Rep. Div. Sea Fish. S. Afr., No. 82, 9 pp. Botha, L., 1971. Investl Rep. Div. Sea Fish. S. Afr., No. 97, 32 pp. Botha, L., 1973. S. Afr. Shipp. News Fishg Ind. Rev., 28(4), 62–67. Botha, L., 1977a. S. Afr. Shipp. News Fishg Ind. Rev., 32(12), 39–41.

450

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Botha, L., 1977b. Internal Rep. Sea Fish. Res. Inst. S. Afr., 1–5. Botha, L., 1978. Colln scient. Pap. int. Commn S.E. Atl. Fish., 5, 59–64. Botha, L., 1980. Ph. D. thesis, University of Stellenbosch, 182 pp. Botha, L., 1985. S. Afr. J. mar. Sci., 3, 179–190. Botha, L., Lucks, D.K. & Chalmers, D.S., 1971. S. Afr. Shipp. News Fishg Ind. Rev., 26(10), 50–57. Boyd, A.J., 1979. Fish. Bull. S. Afr., No. 12, 80–84. Boyd, A.J. & Cruickshank, R.A., 1983. S. Afr. J. Sci., 79(4), 150–151. Boyd, A.J. & Hewitson, J.D., 1983. S. Afr. J. mar. Sci., 1, 71–75. Boyd, A.J., Hewitson, J.D., Kruger, I. & Le Clus, F., 1985. Colln scient. Pap. int. Commn S.E. Atl. Fish., 12(1), 53–58. Boyd, A.J. & Thomas, R.M., 1984. Trop. Ocean-Atmos. Newsl., No. 27, 16–17. Brown, A.C., 1985. Rep. S. Afr. national scient. Prog., No. 103, 83 pp. Brown, B.E., Anthony, V.C., Anderson, E.D., Hennemuth, R.C. & Sherman, K., 1983. In, Proceedings of the Expert Consultation to Examine Changes in Abundance and Species Composition of Neritic Fish Resources, San Jose, Costa Rica, April 1983, edited by G.D.Sharp & J.Csirke, FAO Fish. Rep., No. 291, 2, 465–505. Brown, R.G.B., 1981. Boll. Inst. Mar Peru—Callao (Extraordinario), 55–61. Brunenmeister, S., 1980. ICCAT Coll. Vol. scient. Pap., 9(2), 506–527. Burger, A.E. & Cooper, J., 1984. In, Marine Birds: their Feeding Ecology and Commercial Fisheries Relationships, edited by D.N.Nettleship et al., Canadian Wildlife Service, Dartmouth, pp. 150–159. Butterworth, D.S., 1983. In Proceedings of the Expert Consultation to Examine Changes in Abundance and Species Composition of Neritic Fish Resources, San Jose, Costa Rica, April 1983, edited by G.D.Sharp & J.Csirke, FAO Fish. Rep., No. 291, 2, 329–405. Butterworth, D.S., David, J.H.M., Rickett, L.H. & Xulu, S., 1987. In, The Status, Biology and Ecology of Fur Seals, edited by J.P.Croxall & R.L.Gentry. Centurier-Harris, O.M., 1974. S. Afr. Shipp. News Fishg Ind. Rev., 29(1), 45 only. Centurier-Harris, O.M., 1977. M. Sc. thesis, University of Cape Town, 87 pp. Chalmers, D.S., 1976. Fish. Bull. S. Afr., No. 8, 1–4. Chapman, P. & Shannon, L.V., 1985. Oceanogr. Mar. Biol. Ann. Rev., 23, 183– 251. Chlapowski, K., 1977. Colln scient. Pap. int. Commn S.E. Atl. Fish., 4, 115–120. Chlapowski, K., Draganik, B. & Wyszynski, M., 1982. Colln scient. Pap. int. Commn S.E. Atl. Fish., 9(2), 65–71. Clark, F.N. & Marr, J.C., 1955. Calif, mar. Res. Comm. Prog. Rep., 1 July 1953 to 31 March 1955. Collette, B.B. & Nauen, C.E., 1983. FAO Fish. Synop., No. 125, 2, 1–137. Cooper, J., 1981. Ostrich, 52, 208–215. Cram, D.L., 1977. CalCOFI Rep., No. 19, 33–62. Cram, D.L., 1980. In, Resource Management and Environmental Uncertainty: Lessons from Coastal Upwelling Fisheries, edited by M.H.Glantz & J.D.Thompson, Wiley, New York, pp. 137–156. Crawford, R.J.M., 1979. Ph. D. thesis, University of Cape Town, 467 pp. Crawford, R.J.M., 1980a. J. Fish. Biol., 16, 649–664. Crawford, R.J.M., 1980b. Fish. Bull. S. Afr., No. 13, 111–136. Crawford, R.J.M., 1980c. Colln scient. Pap. int. Commn S.E. Atl. Fish., 7(2), 85– 91. Crawford, R.J.M., 1981a. Investl Rep. Sea Fish. Inst. S. Afr., No. 122, 24 pp. Crawford, R.J.M., 1981b. Fish. Bull. S. Afr., No. 14, 141–181. Crawford, R.J.M., 1981c. Fish. Bull. S. Afr., No. 15, 123–166. Crawford, R.J.M., 1981d. Fish. Bull. S. Afr., No. 15, 1–41. Crawford, R.J.M., 1985. Colln scient. Pap. int. Commn S.E. Atl. Fish., 12(1), 61– 68. Crawford, R.J.M., Centurier-Harris, O.M., Wingate, G.H.L. & Kriedemann, B.D., 1978. Fish. Bull. S. Afr., No. 10, 69–88. Crawford, R.J.M. & Crous, H.B., 1982. Koedoe, 25, 13–31. Crawford, R.J.M., Cruickshank, R.A., Shelton, P.A. & Kruger, I. 1985. S. Afr. J. mar. Sci., 3, 215–228. Crawford, R.J.M. & De Villiers, G., 1984. S. Afr. J. mar. Sci., 2, 49–61.

THE BENGUELA ECOSYSTEM: PART IV

451

Crawford, R.J.M. & De Villiers, G., 1985. S. Afr. J. Sci., 81(2), 91–97. Crawford, R.J.M. & Kriel, F., 1985. S. Afr. Shipp. News Fishg Ind. Rev., 40(1), 29– 37. Crawford, R.J.M. & Shelton, P.A., 1978. Biol. Conserv., 14(2), 85–109. Crawford, R.J.M. & Shelton, P.A., 1981. In, Proceedings of the Symposium on Birds of the Sea and Shore, 1979, edited by J.Cooper, Cape Town, African Seabird Group, pp. 15–41. Crawford, R.J.M., Shelton, P.A. & Berruti, A., 1983. S. Afr. J. Sci., 79, 466–468. Crawford, R.J.M., Shelton, P.A., Cooper, J. & Brooke, R.K., 1983. S. Afr. J. mar. Sci., 1, 153–174. Crawford, R.J.M., Shelton, P.A. & Hutchings, L., 1980. Rapp. P.-v. Réun. Cons. perm. int. Explor. Mer, 177, 355–373. Crawford, R.J.M., Shelton, P.A. & Hutchings, L., 1983. In, Proceedings of the Expert Consultation to Examine Changes in Abundance and Species Composition of Neritic Fish Resources, San Jose, Costa Rica, April 1983, edited by G.D. Sharp & J.Csirke, FAO Fish. Rep., No. 291, 2, 407–448. Cruickshank, R.A., 1982. S. Afr. Shipp. News Fishg Ind. Rev., 37(6), 29–33. Cruickshank, R.A., 1983. Colln scient. Pap. int. Commn S.E. Atl. Fish., 10(2), 75– 97. Cruickshank, R.A., 1984. Colln scient. Pap. int. Commn S.E. Atl. Fish., 11(1), 67– 75. Cruickshank, R.A., Cooper, J. & Hampton, I., 1980. Fish. Bull. S. Afr., No. 13, 77–82. Cushing, D.H., 1969. Fish. Biol. Tech. Pap. FAO, No. 84, 40 pp. Cushing, D.H., 1971. J. Cons. perm. int. Explor. Mer, 33, 340–362. Cushing, D.H., 1982. Climate and Fisheries. Academic Press, London, 373 pp. D’Arcangues, C., 1977. Annls Inst. océanogr., Paris, 53(1), 87–104. Darracott, D.A. & Brown, A.C., 1980. Rep. S. Afr. national scient. Prog., No. 41, 239 pp. Davies, D.H., 1949. Investl Rep. Fish. mar. biol. Surv. Div. S. Afr., No. 11, 36 pp. Davies, D.H., 1954. Investl Rep. Div. Fish. S. Afr., No. 15, 28 pp. Davies, D.H., 1955. Investl Rep. Div. Fish. S. Afr., No. 18, 32 pp. Davies, D.H., 1956a. Investl Rep. Div. Fish. S. Afr., No. 23, 40 pp. Davies, D.H., 1956b. Investl Rep. Div. Fish. S. Afr., No. 24, 52 pp. Davies, D.H., 1957. Investl Rep. Div. Fish. Un. S. Afr., No. 30, 40 pp. Davies, S.L., Newman, G.G. & Shelton, P.A., 1981. Colln scient. Pap. int. Commn S.E. Atl. Fish., 8(2), 51–74. De Campos Rosado, J.M., 1971. J. Cons. perm. int. Explor. Mer, 34(1), 65–75. De Campos Rosado, J.M., 1974a. Colln scient. Pap. int. Commn S.E. Atl. Fish., 1, 78–101. De Campos Rosado, J.M., 1974b. Colln scient. Pap. int. Commn S.E. Atl. Fish., 1, 102–116. De Jaeger, B., 1963. FAO Fish. Rep., No. 62, 2, 588–607. De Jager, B. van D., 1955. Investl. Rep. Div. S. Afr., 19, 11 pp. De Jager, B. van D., Nepgen, C.S. de V. & Van Wyk, R.J., 1963a. Investl Rep. Div. Sea Fish. S. Afr., No. 47, 40 pp. De Jager, B. van D., Nepgen, C.S. de V. & Van Wyk, R.J., 1963b. S. Afr. Shipp. News Fishg Ind. Rev., 18(4), 62–71. De Jager, B. van D., Nepgen, C.S. de V. & Van Wyk, R.J., 1963c. S. Afr. Shipp. News Fishg Ind. Rev., 18(5), 61–69. Delgado, A. & Fernandez, M.A.R., 1985. Int. Symp. Upw. W. Afr., Inst. Inv. Pesq., Barcelona, 1985, 2, 957–968. De Villiers, G., 1976. S. Afr. Shipp. News Fishg Ind. Rev., 31(5), 57–61. De Villiers, G., 1977. Colln scient. Pap. int. Commn S.E. Atl. Fish., 4, 99–104. De Villiers, G., 1985. Int. Symp. Upw. W. Afr., Inst. Inv. Pesq., Barcelona, 1985, 2, 1005–1039. De Villiers, G., 1986. S. Afr. Shipp. News Fishg Ind. Rev., 41(11), 26–28. Draganik, B., 1977. Colln scient. Pap. int. Commn S.E. Atl. Fish., 4, 25–65. Draganik, B., 1978. Colln scient. Pap. int. Commn S.E. Atl. Fish., 5, 29–34. Dudley, S.F.J., Field, J.G. & Shelton, P.A., 1985. S. Afr. J. mar. Sci., 3, 229–237. Dudnik, Y.I., 1975. J. Ichthyol, 15(2), 182–189. Duffy, D.C., Berruti, A., Randall, R.M. & Cooper, J., 1984. S. Afr. J. Sci., 80(2), 65–69. Du Plessis, C.G., 1959. Investl Rep. Div. Fish. Un. S. Afr., No. 38, 28 pp. Field, J.G., Griffiths, C.L., Griffiths, R.J., Jarman, N.G., Zoutendyk, P., Velimirov, B. & Bowes, A., 1980. Trans. R. Soc. S. Afr., 44, 145–203.

452

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Freon, P., 1983. In, Proceedings of the Expert Consultation to Examine Changes in Abundance and Species Composition of Neritic Fish Resources, San Jose, Costa Rica, April 1983, edited by G.D.Sharp & J.Csirke, FAO Fish. Rep., No. 291, 3, 1047–1064. Frost, P.G.H., Siegfried, W.R. & Cooper, J., 1976. Biol. Conserv., 9(2), 79–99. Geldenhuys, N.D., 1973. Investl Rep. Sea Fish. Brch S. Afr., No. 101, 24 pp. Geldenhuys, N.D., 1978. Investl Rep. Sea Fish. Brch S. Afr., No. 115, 16 pp. Gertenbach, L.P.D., 1973. J. Fish. Res. Bd Can., 39(12), 2077–2084. Getz, W.M., 1980. In, Mathematical Modelling in Biology and Ecology, edited by W.M.Getz, Springer-Verlag, New York, pp. 284–303. Gilchrist, J.D.F., 1902. Mar. Invest. S. Afr., 1, 97–163. Gjosaeter, J. & Kawaguchi, K., 1980. FAO Fish. Tech. Pap., No. 193, 59–60. Grant, W.S., 1985. Int. Symp. Upw. W. Afr., Inst. Inv. Pesq., Barcelona, 1985, 1, 551–562. Griffiths, C.L. & Seiderer, J.L. 1980. J. exp. mar. Biol. Ecol, 44, 95–109. Günther, A., 1860. Catalogue of the acanthopterygian fishes in the collection of the British Museum 2, London, (as cited in Talbot & Penrith, 1968). Hampton, I., Shelton, P.A. & Armstrong, M.J., 1985. S. Afr. Shipp. News Fishg Ind. Rev., 40(4), 31–33. Hatanaka, H., Sato, T., Augustyn, C.J., Payne, A.I.L. & Leslie, R.W., 1983. Spec. Publ. mar. Fishery Resource Cent., Tokyo, Japan, 1–73. Hayasi, S., 1974. ICCAT Coll. Vol. scient. Pap., 2, 40–48. Hecht, T., 1976. Ph. D. thesis, University of Port Elizabeth, 353 pp. Hecht, T. & Baird, D., 1977. Zool. Afr., 12(2), 363–372. Herrnkind, W.F., 1980. In, The Biology and Management of Lobsters. Vol. 1. Physiology and Behaviour, edited by J.S.Cobb & B.Phillips, Academic Press, New York, pp. 1–463. Heydorn, A.E.F., 1969a. Investl Rep. Sea Fish. Brch S. Afr., No. 71, 52 pp. Heydorn, A.E.F., 1969b. Investl Rep. Div. Sea Fish. S. Afr., No. 69, 26 pp. Hutchinson, G.E., 1950. Bull. Am. Mus. Nat. Hist., 96, 1–554. ICCAT, 1984. International Commission for the Conservation of Atlantic Tunas report for the biennial period, 1982– 1983 (Part 2), 1–297. ICSEAF, 1976. Rep. Second Spec. Meet. int. Commn S.E. Atl. Fish., 1976, pp. 1–82. ICSEAF, 1984. Hdbk reg. meas. int. Commn S.E. Atl. Fish., 154 pp. ICSEAF, 1985. Doc. scient, advisory Council, No. 2, 41 pp. ICSEAF, 1986. Proc. Rep. Meet. int. Commn S.E. Atl. Fish., 1985(2), 107 pp. Ikeda, I., 1974. Colln scient. Pap. int. Commn S.E. Atl. Fish., 1, 200–208. Inada, T., 1981. Bull. Far Seas Fish. Res. Lab., No. 18, 172 pp. Isarev, A.T., 1976. Colln scient. Pap. int. Commn S.E. Atl. Fish., 3, 99–107. Isarev, A.T., 1979. Colln scient. Pap. int. Commn S.E. Atl. Fish., 6, 233–244. Isarev, A.T., 1980. Colln scient. Pap. int. Commn S.E. Atl. Fish., 7(2), 159–168. Isarev, A.T., 1981. Colln scient. Pap. int. Commn S.E. Atl. Fish., 8(2), 79–82. Isarev, A.T., 1983. Colln scient. Pap. int. Commn S.E. Atl. Fish., 10(1), 81–90. Isarev, A.T., 1986. Colln scient. Pap. int. Commn S.E. Atl. Fish., 13(1), 229–232. Jones, B.W. & van Eck, T.H., 1967. S. Afr. Shipp. News Fishg Ind. Rev., 22(11), 90–97. Kawasaki, T., 1983. In, Proceedings of the Expert Consultation to Examine Changes in Abundance and Species Composition of Neritic Fish Resources, San Jose, Costa Rica, April 1983, edited by G.D.Sharp & J.Csirke, FAO Fish. Rep., No. 291, 3, 1065–1080. Kikawa, S. & Nishikawa, Y., 1980. ICCAT Coll. Vol. scient. Pap., 9(1), 138–142. King, D.P.F., 1977a. Investl Rep. Sea Fish. Brch S. Afr., No. 114, 35 pp. King, D.P.F., 1977b. Fish. Bull. S. Afr., No. 9, 23–31. King, D.P.F. & Macleod, P.R., 1976. Investl Rep. Sea Fish. Brch S. Afr., No. 111, 29pp. King, D.P.F., Robertson, A.A. & Shelton, P.A. 1978. Fish. Bull. S. Afr., No. 10, 37–45.

THE BENGUELA ECOSYSTEM: PART IV

453

Kinloch, M.A., Armstrong, M.J., Crawford, R.J.M. & Leslie, R.W., 1986. Colln scient. Pap. int. Commn S.E. Atl. Fish., 13(1), 233–246. Komarov, Y., 1975. Colln scient. Pap. int. Commn S.E. Atl. Fish., 2, 19–24. Kompowski, A. & Slosarczyk, W., 1975. Acta Ichth. Fiscal., 2, (cited in Krzeptowski, 1982). Konchina, Y.V., 1986. Colln scient. Pap. int. Commn S.E. Atl. Fish., 13(2), 7–18. Kondo, K., 1980. Rapp. P.-v. Réun. Cons. perm. int. Explor. Mer, 177, 332–354. Kono, H., 1980. Colln scient. Pap. int. Commn S.E. Atl. Fish., 7(2), 175–209. Koslow, J.A., 1981. Fish. Bull. NOAA, 79, 131–142. Koslow, J.A., 1984. Can. J. Fish. Aquat. Sci., 41, 1722–1729. Krzeptowski, M., 1982. Colln scient. Pap. int. Commn S.E. Atl. Fish., 9(2), 111–119. Kuderskaya, R.A., 1985. Colln scient. Pap. int. Commn S.E. Atl. Fish., 12(1), 95– 97. Laevastu, T. & Larkins, H.A., 1981. Marine Fisheries Ecosystem. Its Quantitative Evaluation and Management, Fishing News Books, Farnham, 162 pp. Laevastu, T. & Rosa, H., 1963. FAO Fish. Rep., No. 6, 3, 1835–1851. Lasker, R. & MacCall, A.D., 1983. Proceedings of the Joint Oceanographic Assembly 1982—General Symposia. Canadian National Committee/Scientific Committee on Oceanic Research, Ottawa, Ontario, pp. 110–120. Lazarus, B.I., 1967. Investl Rep. Div. Sea Fish. S. Afr., No. 63, 38 pp. Lebeau, A., 1971. Science et Pêche, Bull. Inst. Pêches marit., No. 204, 1–13. Le Clus, F., 1979a. Fish. Bull. S. Afr., No. 11, 26–38. Le Clus, F., 1979b. Investl Rep. Sea Fish. Brch S. Afr., No. 119, 29 pp. Le Clus, F., 1985. Colln scient. Pap. int. Commn S.E. Atl. Fish., 12(1), 99–106. Le Clus, F. & Kruger, I., 1982. Colln scient. Pap. int. Commn S.E. Atl. Fish., 9(2), 121–146. Le Clus, F. & Melo, Y.C., 1982. Colln scient. Pap. int. Commn S.E. Atl. Fish., 9(2), 147–161. Le Clus, F. & Thomas, R.M., 1982. Colln scient. Pap. int. Commn S.E. Atl. Fish., 9(2), 163–176. Lees, R. 1969. Fishing for Fortunes—the Story of the Fishing Industry in Southern Africa and the Men who made it. Purnell & Sons S.A. (Pty.) Ltd, Cape Town, 283 pp. Le Gall, L.-Y., 1974. Synopsis FAO sur les pêches, No. 109, 70 pp. Le Loeuff, P., Cayre, P. & Inte’s, A., 1978. Docums scient. Cent. Rech. océanogr. Abidjan, 9, 17–65. Leslie, R.W., 1985. Colln scient. Pap. int. Commn S.E. Atl. Fish., 12(1), 107–117. Leslie, R.W., 1986. Colln scient. Pap. int. Commn S.E. Atl. Fish., 13(2), 27–36. Lima, F.R. & Wise, J.P., 1963. FAO Fish. Rep., No. 6, 3, 1515–1521. Lipskaya, N.J., 1972. Trudy VNIRO, 77(2) (cited in Krzeptowski, 1982). Lloris, A. & Rucabado, A., 1985. Colln scient. Pap. int. Commn S.E. Atl. Fish., 12(1), 147–155. Lucks, D.K., 1972. S. Afr. Shipp. News Fishg Ind. Rev., 27(9), 54–57. Lucks, D.K., Payne, A.I.L. & Maree, S., 1973. S. Afr. Shipp. News Fishg Ind. Rev., 28(9), 65–71. MacCall, A.D., 1980. Intergovernmental Oceanographic Commission Workshop on the Effects of Environmental Variation on the Survival of Larval Pelagic Fishes, Workshop Rep., No. 28, 201–220. MacCall, A.D., 1984. In, Marine Birds: Their Feeding Ecology and Commercial Fisheries Relationships, edited by D.N.Nettleship et al., Canadian Wildlife Service, Dartmouth, pp. 136–148. Macpherson, E., 1976. Colln scient. Pap. int. Commn S.E. Atl. Fish., 3, 119–123. Macpherson, E., 1983a. Mar. Biol, 78, 105–112. Macpherson, E., 1983b. Result. Exped. scient, (suppl. Invest, pesq.), 11, 81–137. Macpherson, E., 1985. Colln scient. Pap. int. Commn S.E. Atl. Fish., 12(1), 155–162. Macpherson, E. & Allue, C., 1980. Inf. Tec. Inst. Inv. Pesq., No. 79–80, 56 pp. Macpherson, E. & Lloris, D., 1977. Colln scient. Pap. int. Commn S.E. Atl. Fish., 4, 127–143. Macpherson, E., Mombeck, F. & Schülein, F.H., 1982. Colln scient. Pap. int. Commn S.E. Atl. Fish., 9(2), 177–194. Macpherson, E., Roel, B. & Morales, B., 1985. Colln scient. Pap. int. Commn S.E. Atl. Fish., 12(2), 1–61. Macpherson, E., Roel, B. & Morales, B., 1986. Colln scient. Pap. int. Commn S.E. Atl. Fish., 13(2), 113–136. Marchand, J.M., 1933. Rep. Fish. mar. biol. Surv. Un. S. Afr., No. 10, 110–143.

454

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Matsuura, Y., 1982. ICCAT Coll. Vol. scient. Pap., 17(1), 211–213. Matthews, J.P., 1961. Investl Rep. mar. Res. Lab. S.W. Afr., No. 3, 35 pp. Matthews, J.P., 1964. Investl Rep. mar. Res. Lab. S.W. Afr., No. 10, 96 pp. Matthews, J.P. & Berruti, A., 1983. S. Afr. J. mar. Sci., 1, 61–63. McLain, D.R., Brainard, R.E. & Norton, J.G., 1985. CalCOFI Rep., No. 26, 51–64. Melo, Y.C., 1984. S. Afr. J. mar. Sci., 2, 19–31. Melville-Smith, R., 1983. S. Afr. J. mar. Sci., 1, 123–131. Melville-Smith, R., 1985. S. Afr. J. mar. Sci., 3, 55–62. Melville-Smith, R., in press. Crustaceana. Miyake, P.M. & De Rodda, D., 1984. ICCAT Statistical Bull. (Final), No. 14, 133 pp. Miyake, P.M., Nordstrom, V. & Da Rodda, D., 1981. ICCAT Statistical Bull. (Provisional), No. 11, 121 pp. Miyake, P.M., Wise, J.P., Nordstrom, V. & Da Rodda, D., 1982a. ICCAT Historical Statistical Bull., No. 1, 79 pp. Miyake, P.M., Wise, J.P., Nordstrom, V. & Da Rodda, D., 1982b. ICCAT Historical Statistical Bull., No. 2, 109 pp. Moloney, C.L. & Field, J.G., 1985. S. Afr. J. mar. Sci., 3, 119–128. Molteno, C.J., 1948. The South Africa Tunas, S. Afr. Fish. Ind. Res. Inst., Cape Town, 34 pp. Mombeck, F., 1970. Arch. Fischwiss., 21(1), 45–61. Morales, B., 1980. Colln scient. Pap. int. Commn S.E. Atl. Fish., 7(2), 233–241. Morales, B., 1982. Colln scient. Pap. int. Commn S.E. Atl. Fish., 9(2), 195–198. Morita, S., 1978. ICCAT Coll. Vol. scient. Pap., 7(2), 232–237. Nakamura, H., 1969. Tuna Distribution and Migration. Fishing Industry (Books) Ltd, London, 76 pp. Nawratil, O., 1960. Investl Rep. mar. Res. Lab. S.W. Afr., No. 2, 43 pp. Nepgen, C.S. de V., 1957. Investl Rep. Div. Fish. S. Afr., No. 28, 30 pp. Nepgen, C.S. de V., 1970a. Investl Rep. Div. Sea Fish S. Afr., No. 87, 26 pp. Nepgen, C.S. de V., 1970b. Investl Rep. Div. Sea Fish S. Afr., No. 90, 13 pp. Nepgen, C.S. de V., 1971. In, ICCAT report for the biennial period, 1970–1971, (Part 2), 114–116. Nepgen, C.S. de V., 1975. S. Afr. Shipp. News Fishg. Ind. Rev., 30(11), 57 only. Nepgen, C.S. de V., 1977. Investl Rep. Sea Fish. Brch S. Afr., No. 105, 35 pp. Nepgen, C.S. de V., 1979a. Fish. Bull. S. Afr., No. 11, 39–42. Nepgen, C.S. de V., 1979b. Fish. Bull. S. Afr., No. 12, 35–43. Nepgen, C.S. de V., 1982. Fish. Bull. S. Afr., No. 16, 75–93. Newman, G.G., 1967. Investl Rep. Div. Sea Fish. S. Afr., No. 64, 24 pp. Newman, G.G., 1969a. Investl Rep. Div. Sea Fish. S. Afr., No. 74, 7 pp. Newman, G.G., 1969b. Investl Rep. Div. Sea Fish. S. Afr., No. 56, 20 pp. Newman, G.G., 1970a. Investl Rep. Div. Sea Fish. S. Afr., No. 85, 13 pp. Newman, G.G., 1970b. Investl Rep. Div. Sea Fish. S. Afr., No. 86, 6 pp. Newman, G.G., 1973. Ph. D. thesis, University of Stellenbosch, 268 pp. Newman, G.G., 1974. Colln scient. Pap. int. Commn S.E. Atl. Fish., 1, 212–219. Newman, G.G., 1977. FAO Fish. Tech. Pap., No. 178, 59 pp. Newman, G.G., 1984. In, Exploitation of Marine Communities, edited by R.M. May, Springer-Verlag, Berlin, pp. 313–333. Newman, G.G. & Crawford, R.J.M., 1980. Rapp. P.-v. Réun. Cons. perm. int. Explor. Mer, 177, 279–281. Newman, G.G., Crawford, R.J.M. & Centurier-Harris, O.M., 1974. Colln scient. Pap. int. Commn S.E. Atl. Fish., 1, 163–170. Newman, G.G., Crawford, R.J.M. & Centurier-Harris, O.M., 1979. Investl Rep. Sea Fish. Brch S. Afr., No. 120, 34 pp. Newman, G.G. & Pollock, D.E., 1971. Investl Rep. Div. Sea Fish. S. Afr., No. 94, 24pp. Newman, G.G. & Pollock, D.E., 1974. Investl Rep. Div. Sea Fish. Brch. S. Afr., No. 107, 16 pp. Newman, G.G. & Pollock, D.E., 1977. Colln scient. Pap. int. Commn S.E. Atl. Fish., 4, 175–187. Olivar, M.P., 1982. Colln scient. Pap. int. Commn S.E. Atl. Fish., 9(2), 221–240. Olivar, M.P., 1984. Colln scient. Pap. int. Commn S.E. Atl. Fish., 11(2), 13–21.

THE BENGUELA ECOSYSTEM: PART IV

455

Olivar, M.P. & Rubies, P., 1983. Colln scient. Pap. int. Commn S.E. Atl. Fish., 10(2), 161–172. Olivar, M.P. & Rubies, P., 1985. Colln scient. Pap. int. Commn S.E. Atl. Fish., 12(2), 85–97. O’Toole, M.J., 1977a. Ph. D. thesis, University of Cape Town, 272 pp. O’Toole, M.J., 1977b. Fish. Bull. S. Afr., No. 9, 46–47. O’Toole, M.J., 1978a. Fish. Bull. S. Afr., No. 10, 20–36. O’Toole, M.J., 1978b. Investl Rep. Sea Fish. Brch S. Afr., No. 116, 28 pp. O’Toole, M.J. & King, D.P. F., 1974. Vie Milieu (Ser. A), 24, 443–452. Parker, R.R. & Larkin, P.A., 1959. J. Fish. Res. Bd Can., 16, 721–745. Parrish, R.H., Bakun, A., Husby, D.M. & Nelson, C.S., 1983. In, Proceedings of the Expert Consultation to Examine Changes in Abundance and Species Composition of Neritic Fish Resources, San Jose, Costa Rica, April 1983, edited by G.D.Sharp & J.Csirke, FAO Fish. Rep., No. 291, 3, 731–777. Payne, A.I.L., 1977. Investl Rep. Sea Fish. Brch S. Afr., No. 113, 32 pp. Payne, A.I.L., 1979. Fish. Bull. S. Afr., No. 12, 26–34. Payne, A.I.L., 1985a. Int. Symp. Upw. W. Afr., Inst. Inv. Pesq., Barcelona, 1985, 2, 1063–1079. Payne, A.I.L., 1985b. S. Afr. J. Zool., 20, 49–56. Payne, A.I.L., 1986. Ph. D. thesis, University of Port Elizabeth, 368 pp. Payne, A.I.L., Augustyn, C.J. & Leslie, R.W., 1985. Colln scient. Pap. int. Commn S.E. Atl. Fish., 12(2), 99–123. Payne, A.I.L., Augustyn, C.J. & Leslie, R.W., 1986. Colln scient. Pap. int. Commn S.E. Atl. Fish., 13(2), 181–196. Penrith, M.J., 1963. M. Sc. thesis, University of Cape Town, 193 pp. Penrith, M.J., 1972. Koedoe, 15, 135–139. Phillips, B.F. & Sastry, A.N., 1980. In, The Biology and Management of Lobsters. Vol. II. Ecology and Management, edited by J.S.Cobb & B.Phillips, Academic Press, New York, pp. 3–390. Pollock, D.E., 1952. Investl Rep. Sea Fish. Inst. S. Afr., No. 124, 57 pp. Pollock, D.E., 1979. Mar. Biol, 52, 347–356. Pollock, D.E., 1981. Fish. Bull. S. Afr., No. 15, 49–66. Pollock, D.E., 1984. S. Afr. Marlin and Tuna Club Monthly Magazine, 1(4), 1–3. Pollock, D.E., Augustyn, C.J. & Goosen, P.C., 1982. Investl Rep. Sea Fish. Inst. S. Afr., No. 125, 30 pp. Pollock, D.E. & Beyers, C.J. de B., 1981. Trans. Roy. Soc. S. Afr., 44, 379–400. Pozo, A.E., 1976. Colln scient. Pap. int. Commn S.E. Atl. Fish., 3, 179–185. Prenski, L.B., 1978. Colln scient. Pap. int. Commn S.E. Atl. Fish., 5, 89–94. Prenski, L.B., 1980. Colln scient. Pap. int. Commn S.E. Atl. Fish., 7(2), 283–296. Prosch, R.M., 1986. M. Sc. thesis, University of Cape Town, 197 pp. Rand, R.W., 1959a. Investl Rep. Div. Fish. Un. S. Afr., No. 34, 75 pp. Rand, R.W., 1959b. Investl Rep. Div. Sea Fish. S. Afr., No. 39, 36 pp. Rand, R.W., 1960a. Investl Rep. Div. Sea Fish. S. Afr., No. 41, 27 pp. Rand, R.W., 1960b. Investl Rep. Div. Sea Fish. S. Afr., No. 42, 32 pp. Randall, R.M., 1983. Ph. D. thesis, University of Port Elizabeth, 262 pp. Rattray, J.M., 1947. Ann. S. Afr. Mus., 36(4), 315–331. Richards, W.J., 1976. ICCAT Coll. Vol. scient. Pap., 5(2), 267–278. Richards, W.J. & Potthoff, T., 1980. ICC AT Coll. Vol. scient. Pap., 9(2), 433–441. Robinson, G.A., 1966. M. Sc. thesis, University of Stellenbosch, 60 pp. Robinson, G.A., 1976. Koedoe, 19, 43–48. Robson, D.S., 1966. Res. Bull. int. Commn Northw. Atlant. Fish., 3, 5–14. Rudomiotkina, G.P., 1983. ICCAT Coll. Vol. scient. Pap., 18(2), 355–362. Ryther, J.H., 1969. Science, 166, 72–76. Sato, T., 1977. Colln scient. Pap. int. Commn S.E. Atl. Fish., 4, 151–160. Sato, T., 1980. Colln scient. Pap. int. Commn S.E. Atl. Fish., 7(2), 315–330. Saville, A., 1980. Rapp. P.-v. Réun. Cons. perm. int. Explor. Mer, 177, 513–517.

456

R.J.M.CRAWFORD, L.V.SHANNON AND D.E.POLLOCK

Scheltema, R.S., 1971. In, Fourth Europ. Mar. Biol. Symp., edited by D.J.Crisp, Cambridge University Press, Cambridge, pp. 7–28. Schülein, F.H., 1971. M.Sc. thesis, University of Stellenbosch, 63 pp. Schülein, F.H., 1973. S. Afr. natn. oceanogr. Symp. Abstr., Cape Town, p. 29 only. Schülein, F.H., 1982. Unpupl. Rep. Sea Fish. Res. Inst. S. Afr., 77 pp. Schülein, F.H., 1986. Colln scient. Pap. int. Commn S.E. Atl. Fish., 13(2), 205–224. Schülein, F.H., Butterworth, D.S. & Cram, D.L., 1978. Colln scient. Pap. int. Commn S.E. Atl. Fish., 5, 35–44. Shackleton, L.Y., 1986. M. Sc. thesis, University of Cape Town, 141 pp. Shannon, L.V., 1980. Internal Rep. Sea Fish. Inst. S. Afr., No. 37, 8 pp. Shannon, L.V., 1985. Oceanogr. Mar. Biol. Ann. Rev., 23, 105–182. Shannon, L.V., 1986. Internal Rep. Sea Fish. Res. Inst. S. Afr., No. 89, 24 pp. Shannon, L.V., Agenbag, J.J. & Buys, M.E.L., 1987. S. Afr. J. mar. Sci., 5, 11– 34. Shannon, L.V., Boyd, A.J., Brundrit, G.B. & Taunton-Clark, J. 1986. J. mar. Res., 44, 495–520. Shannon, L.V., Crawford, R.J.M. & Duffy, D.C., 1984. S. Afr. J. Sci., 80(2), 51– 60. Shannon, L.V. & Field, J.G., 1985. Mar. Ecol. Prog. Ser., 22, 7–19. Shannon, L.V., Hutchings, L., Bailey, G.W. & Shelton, P.A., 1984. S. Afr. J. mar. Sci., 2, 109–130. Shannon, L.V. & Pillar, S.C., 1986. Oceanogr. Mar. Biol. Ann. Rev., 24, 65–170. Shcherbich, L.V., 1981. Colln scient. Pap. int. Commn S.E. Atl. Fish., 8(2), 245–251. Shcherbich, L.V., Komarov, Y.A. & Florinskaya, E.S., 1980. Colln scient. Pap. int. Commn S.E. Atl. Fish., 7(2), 123–126. Sheldon, R.W., Sutcliffe, W.H. & Paranjape, M.A., 1977. J. Fish. Res. Bd Can., 34(12), 2344–2353. Shelton, P.A., 1979. M. Sc. thesis, University of Cape Town, 77 pp. Shelton, P.A. & Armstrong, M.J., 1983. In, Proceedings of the Expert Consultation to Examine Changes in Abundance and Species Composition of Neritic Fish Resources, San Jose, Costa Rica, April 1983, edited by G.D.Sharp & J.Csirke, FAO Fish. Rep., No. 291, 3, 1113–1132. Shelton, P.A. & Armstrong, M.J., 1984. S. Afr. Shipp. News Fishg Ind. Rev., 39(4), 37 only. Shelton, P.A., Boyd, A.J. & Armstrong, M.J., 1985. CalCOFI Rep., No. 26, 72–92. Shelton, P.A., Crawford, R.J.M., Cooper, J. & Brooke, R.K., 1984. S. Afr. J. mar. Sci., 2, 217–257. Shelton, P.A. & Hutchings, L., 1982. J. Cons. perm. int. Explor. Mer, 40, 185–198. Silvert, W., 1982. In, Multispecies Approaches to Fisheries Management Advice, edited by M.C.Mercer, Can. Spec. Publ. Fish. Aquat. ScL, 59, 24–27. Smale, M.J., 1983. Ph. D. thesis, Rhodes University, S. Africa, 284 pp. Smale, M.J., 1985. S. Afr. J. mar. Sci., 3, 63–75. Smale, M.J. & Buxton, C.D., 1985. S. Afr. J. mar. Sci., 3, 131–144. Stander, G.H., 1963. Investl Rep. mar. Res. Lab. S.W. Afr., No. 9, 57 pp. Stander, G.H. & De Decker, A.H.B., 1969. Investl Rep. Div. Sea Fish. S. Afr., No. 81, 46 pp. Stander, G.H. & Le Roux, P.J., 1968. Investl Rep. Div. Sea Fish. S. Afr., No. 65, 14pp. Stephenson, T.A., 1944. Ann. Natal Mus., 10, 261–358. Talbot, F.H., 1964. Mar. Biol. Assoc. India Symp. Ser., 1, 187–209. Talbot, F.H. & Penrith, M.J., 1961. S. Afr. J. Sci., 57, 240 only. Talbot, F.H. & Penrith, M.J., 1962. Nature, Land., 193, 558–559. Talbot, F.H. & Penrith, M.J., 1963. FAO Fish. Rep., No. 6, 2, 608–651. Talbot, F.H. & Penrith, M.J., 1968. Ann. S. Afr. Mus., 52(1), 1–41. Terre, J.J., 1980. Colln scient. Pap. int. Commn S.E. Atl. Fish., 7(2), 331–340. Terre, J.J., 1984. Colln scient. Pap. int. Commn S.E. Atl. Fish., 11(2), 61–73. Thomas, R.M., 1984. Colln scient. Pap. int. Commn S.E. Atl. Fish., 11(2), 87–90. Thomas, R.M., 1985. Ph. D. thesis, University of Cape Town, 289 pp. Thomas, R.M., 1986. Colln scient. Pap. int. Commn S.E. Atl. Fish., 13(2), 243–269. Thomas, R.M. & Boyd, A.J., 1985. Colln scient. Pap. int. Commn S.E. Atl. Fish., 12(2), 181–191.

THE BENGUELA ECOSYSTEM: PART IV

457

Thomas, R.M., Hewitson, J.D. & Böhme, H., 1982. Fish. Bull. S. Afr., No. 16, 95– 97. Thompson, W.W., 1913. The Sea Fisheries of the Cape Colony from van Riebeeck’s Days to the Eve of the Union. T.Maskew Miller, Cape Town, 163 pp. Thompson, W.W., 1918. Mar. Biol. Rep. S. Afr., 4, 75–177. Troadec, J.-P., Clark, W.G. & Gulland, J.A., 1980. Rapp. P.-v. Réun. Cons. perm. int. Explor. Mer, 177, 252–277. Uozumi, Y., Hatanaka, H., Payne, A.I.L. & Augustyn, C.J., 1985. Publ. Far Seas Fish. Res. Lab., S Series, No. 13, 78 pp. Uozumi, Y., Hatanaka, H., Sato, T., Augustyn, C.J., Payne, A.I.L. & Leslie, R.W., 1984. Publ. Far Seas Fish. Res. Lab., S Series, No. 11, 91 pp. Van Den Berg, R.A. & Matthews, J.P., 1969. Investl Rep. Mar. Res. Lab. S.W. Afr., No. 15, 39 pp. Van Der Elst, R., 1976. Investl Rep. oceanogr. Res. Inst. S. Afr., No. 44, 59 pp. Van Wyk, G.F., 1944. J. Soc. chem. Ind. Land., 63, 367–371. Venidictova, L.I., 1982. Colln scient. Pap. int. Commn S.E. Atl. Fish., 9(2), 337–341. Venediktova, L.I., 1985. Colln scient. Pap. int. Commn S.E. Atl. Fish., 12(2), 193– 197. Venter, J.D., 1976. M.Sc. thesis, Rand Afrikaans University, 174 pp. Welsh, J.G., 1968. Fish. Bull. S. Afr., No. 5, 32–34. Winstanley, R.H., Caton, A.E., Harris, J.G.K. & Lewis, R.K., 1983. SAFIC, 7, 7–13. Wise, J.P. & Fox, W.W., 1969. Spec, scient. Rep. U.S. Fish Wildl. Serv., No. 582, 7 pp. Wooster, W.S. & Reid, J.L., 1963. In, The Sea, Vol. 2, edited by M.N.Hill, Wiley Interscience, New York, pp. 253–276. Wrzesinski, O., 1975. Colln scient. Pap. int. Commn S.E. Atl. Fish., 2, 117–139. Wysokinski, A., 1984. Colln scient. Pap. int. Commn S.E. Atl. Fish., 11(2), 91–98. Yang, R.-T., 1983. Annual Catch Statistics of Taiwan’s Tuna Longline Fishery 1982. Tuna Research Centre, Taipei, Rep. China, 46 pp. plus figs and tables. Yang, R.-T. & Sun, C.L., 1984. ICCAT Coll. Vol. scient. Pap., 20(2), 251–269. Zenkin, V.S. & Komarov, Y.A., 1981. Colln scient. Pap. int. Commn S.E. Atl. Fish., 8(2), 291–298. Zoutendyk, P., 1973. Trans. R. Soc. S. Afr., 40(5), 349–366. Zoutendyk, P., 1974. Trans. R. Soc. S. Afr., 41(1), 33–41.

Oceanogr. Mar. Biol. Ann. Rev., 1987, 25, 507–562 Margaret Barnes, Ed. Aberdeen University Press

THE ASSOCIATION BETWEEN GOBIID FISHES AND BURROWING ALPHEID SHRIMPS ILAN KARPLUS Agricultural Research Organization, Fish and Aquaculture Research Station Dor, D.N.Hof Hacarmel 30820, Israel

INTRODUCTION The association between gobiid fishes and burrowing alpheid shrimps was described for the first time by Longley & Hildebrand (1941) from southern Florida. The further study of this partnership was delayed until the introduction of mask and snorkel and SCUBA diving as a tool for collecting material and carrying out detailed behavioural and ecological studies in the marine environment. The first aspect studied was the taxonomy of the gobies, a discipline which still leads the research of this association, probably due to the gobies’ diversity and richness in species and circumtropical distribution (Klausewitz, 1960; Lubbock & Polunin, 1977; Polunin & Lubbock, 1977; Yanagisawa, 1978; Hoese & Randall, 1982). The first ecologicalbehavioural studies were made by Luther (1958a, b) and Magnus (1967) in the Red Sea. In the northern Gulf of Elat, Red Sea, the present author has studied goby-shrimp associations for more than seven years, concentrating on communication, distribution and partner specificity (Karplus, 1970, 1976, 1979, 1981; Karplus, Szlep & Tsurnamal, 1972a, 1974, 1981; Karplus, Tsurnamal & Szlep, 1972b; Karplus & Vercheson, 1978; Karplus & Ben-Tuvia, 1979; Karplus, Tsurnamal, Szlep & Algom, 1979; Goren & Karplus, 1983). In Hawaii a detailed quantitative study was carried out on the communication between two species of shrimps, and one species of goby using both sequence and information analysis (Moehring, 1972; Preston, 1978). Polunin & Lubbock (1977) carried out a field study in the Seychelles on the distribution and partner specificity of goby-shrimp associations. The ecology, population dynamics and partner specificity of goby-shrimp associations were studied in the Great Barrier Reef by Cummins (1979). The reproduction of goby and shrimp and the initial formation of the association, areas completely unknown, were studied in southern Japan by Yanagisawa (1982, 1984) following a laboratory study by Harada (1969). Data collected by individuals in different parts of the world gradually complemented one another to allow a more comprehensive understanding of these fascinating associations. The aim of this review is to present in detail the present state of knowledge of the goby-shrimp association and to suggest areas of importance for future research. TAXONOMY OF GOBIES AND ALPHEID SHRIMPS The majority of the publications on goby-shrimp association are devoted to the taxonomy of the associated gobies which show great diversity and richness in genera and species (Fig. 1). Few associated gobies were described while traditional methods of collecting were employed. The usual indiscriminate massive poisoning of reef fishes for their collection is not effective for collecting associated gobies since these small fish are often overlooked in mass collections or hide in their holes where they either die or avoid the poison.

GOBIID FISHES AND ALPHEID SHRIMPS

459

The specific search by means of SCUBA diving for associated gobies in the last two decades in shallow and deep water has been most rewarding. Many new species are being discovered and described. The successful collection of associated gobies is practised by means of: an Hawaiian sling and a multipronged arrow (Randall, 1963), small baited hooks (Yanagisawa, 1976; Polunin & Lubbock, 1977), small amounts of poison injected in specific areas (Hoese & Randall, 1982), small dynamite cartridges (Klausewitz, pers. comm.), traps (Magnus, pers. comm.), “slurp gun” (Moehring, 1972), and hand nets after manually blocking the burrow entrance with a spade (Karplus, 1970). A special remote-operated spade was later developed for collecting live gobies for behavioural studies (Karplus & Vercheson, 1978). Several problems exist in the classification of gobies which live in association with alpheids. Many of the nominal genera involved have not been adequately diagnosed. Consequently, in many cases, species have been placed in genera with which they have little affinity (Hoese & Steene, 1978). Generic and species synonymy have often added to the confusion. A total of 70 nominal species of gobiid fishes placed within 18 nominal genera have been reported as living in association with burrowing alpheid shrimps (Table I). In some of these genera, which contain several species, all members live in association with shrimps: e.g. Amblyeleotris—20 species, Cryptocentrus—18 species, Ctenogobiops—6 species, Vanderhorstia—5 species, and Stonogobiops—4 species. In a revision of the genus Cryptocentrus at present in preparation by Hoese, about 40 nominal species are described (Hoese & Steene, 1978). Already close to 100 species of associated gobies have been discovered, so it can be assumed that their actual number is probably closer to two hundred. To date 13 species of alpheid shrimps have been reported as living in association with gobiid fishes (Table II). The majority of these species belong to the Brevirostris group. Only three species (Alpheus crassimanus, A. randalli, and A. malabaricus) belong to the Edwardsii group. Species of both these groups occupy burrows of their own construction in silty to sandy bottoms (Banner & Banner, 1982). In many of the studies which either briefly mention goby-shrimp association or deal with them at length, the shrimp taxonomy only covers the family or generic level. The two main reasons for this are the difficulty experienced in shrimp collection, as they withdraw rapidly into their burrows when approached by a diver, and the great variability of shrimp morphology and coloration (Banner & Banner, 1982). Shrimps can be collected by means of: an Hawaiian sling and a multipronged arrow (Banner & Banner, 1980), small dynamite cartridges or traps (Karplus & Vercheson, 1978), injection of an irritating liquid heavier than water (a saturated solution of NaCl with CuSO4) into the burrow, forcing the shrimp to leave it (Weiler, 1976), hand nets after blocking the burrow entrance with a spade (Karplus, 1970; Moehring, 1972), or digging up the entire burrow system of small animals (Yanagisawa, 1984). The same remote-operated spade used for collecting gobies has been utilized here, (Karplus & Vercheson, 1978). Most of the above methods required patience, skill, and practice, rendering the collecting of associated alpheids a rather difficult task. Banner & Banner (1982) discussed the colour, pattern, and structural variability of the Indo-Pacific gobyassociated alpheid shrimps of the Brevirostris group. This variability makes the determination of species from dead specimens extremely difficult. Banner & Banner (1982) suggested that the answer for species distinctions lies not in museum work with dead specimens but in careful field observations correlated with laboratory studies on living animals. Several types of associated alpheids, differing with respect to coloration, ecology and behaviour, have been described from the Seychelles (Polunin & Lubbock, 1977), the Great Barrier Reef (Cummins, 1979), and the northern Red Sea (Karplus et al., 1974). The last authors suggested that different types of Red Sea shrimp of distinct colour patterns (Fig. 2) living on different substrata, in different types of burrows and with different fish partners, may represent valid species. Species validity of these types of Red Sea alpheid shrimps was confirmed by Professor Miya in his morphological studies (Karplus et al., 1981).

460

ILAN KARPLUS

The actual number of goby-associated alpheids is probably several times larger than already reported. With the further collection of shrimps and clarification of the “type” status, more species will certainly be described. DISTRIBUTION OF ASSOCIATIONS The association between gobiid fishes and burrowing alpheid shrimps has been reported by numerous workers from many localities in tropical as well as subtropical waters. The present author has chosen one among all references for each locality as an example. These associations were found in the Red Sea (Klausewitz, 1960), the Persian Gulf (Palmer, 1963), and in the Indian Ocean, in Aldabra Atoll (Polunin & Lubbock, 1977), the Seychelle Islands (Polunin & Lubbock, 1977), South Africa (Smith, 1959), Moçambique (Macnae & Kalk, 1962), the Maldive Islands (Hoese & Randall, 1982), and Madagascar (Thomassin, 1971). These associations were also reported in the Pacific, in Palau (Bayer & Harry-Rofen, 1957), Fiji, New Caledonia, American Samoa (Lubbock & Polunin, 1977), Solomon Islands (Hoese & Randall, 1982), Hawaii (Preston, 1978), Malluca (Hoese & Steene, 1978), Indonesia (Hoese & Steene, 1978), Marquesas (Banner & Banner, 1980), Marshall Islands (Paulson, 1978), Great Barrier Reef (Cummins, 1979), New Guinea (Hoese & Randall, 1982), Japan (Yanagisawa, 1978), and the Philippine Islands (Hoese & Randall, 1982), as well as in the Atlantic in Florida and the Bahamas (Bohlke & Chaplin, 1968). Some associated gobies are very widely distributed. For example, Amblyeleotris steinitzii, was first described from the Red Sea and the Marshall Islands (Klausewitz, 1974a), and later also recorded from the Seychelles (Polunin & Lubbock, 1977), southern Japan (Yanagisawa, 1978), and the TABLE I Species of gobies reported associated with burrowing alpheid shrimps: this table is largely based on a list kindly provided by Dr D.F.Hoese; where two species names appear in one entry they are synonyms Species of goby

Reference

Acentrogobius pflaumi Gobius pflaumi Amblyeleotris aurora Polunin & Lubbock, 1977 Amblyeleotris callopareia Polunin & Lubbock, 1979 Amblyeleotris diagonalis Polunin & Lubbock, 1979 Amblyeleotris fasciata Zebreleotris fasciatus (Herre, 1953) Amblyeleotris fantanesii Gobius fontanesii (Bleeker, 1852) Amblyeleotris guttata Pteroculiops guttatus (Fowler, 1938) Amblyeleotris gymnocephala Gobius gymnocephalus (Bleeker, 1853) Amblyeleotris japonica

Harada, 1969; Yanagisawa, 1978 Polunin & Lubbock, 1977; Banner & Banner, 1980 Polunin & Lubbock, 1979 Polunin & Lubbock, 1979 Yanagisawa, 1976, 1978 Hoese & Steene, 1978 Yanagisawa, 1978 Hoese & Steene, 1978 Miya & Miyake, 1969; Harada, 1969, 1971; Yanagisawa, 1976, 1978, 1982, 1984 Polunin & Lubbock, 1979

Amblyeleotris japonicus (Takagi, 1957) Amblyeleotris latifasciata Polunin & Lubbock, 1979 Amblyeleotris macronema Polunin & Lubbock, 1979 Amblyeleotris maculata Yanagisawa, 1976

Polunin & Lubbock, 1979 Yanagisawa, 1976, 1978

GOBIID FISHES AND ALPHEID SHRIMPS

461

Fig. 1(A-E).—Species of gobies associated with shrimp in the northern Red Sea: A Cryptocentrus caeruleopunctatus; B, C. cryptocentrus; C, Amblyeleotris steinitzi; D Cryptocentrus lutheri; E, Ctenogobiops maculosus.

462

ILAN KARPLUS

Fig. 1(F-I). —F, Lotilia graciliosa; G, Eilatia latruncularia; H, Vanderhostia mertensi I, V. delagoae; from Karplus et al. (1981).

GOBIID FISHES AND ALPHEID SHRIMPS

Species of goby

Reference

Amblyeleotris ogasawarensis Yanagisawa, 1978 Amblyeleotris periopthalma Eleotris periophthalmus (Bleeker, 1853) Amblyeleotris randalli Hoese & Steene, 1978 Amblyeleotris rhyax Polunin & Lubbock, 1979 Amblyeleotris steinitzi

Yanagisawa, 1978 Hoese & Steene, 1978

Cryptocentrus steinitzi (Klausewitz, 1974) Amblyeleotris sungami Cryptocentrus sungami (Klausewitz, 1969) Amblyeleotris wheeleri Cryptocentrus wheeleri Polunin & Lubbock, 1977 Bathygobius curacao Metzelaar Butis butis Hamilton Cryptocentrus albidorsus Mars albidorsus Yanagisawa, 1978 Cryptocentrus caeruleomaculatus Mars caeruleomaculatus (Herre, 1933) Cryptocentrus caeruleopunctatus Gobius caeruleopunctatus (Ruppell, 1828) Cryptocentrus cinctus Smilogobius cinctus (Herre, 1936) Cryptocentrus cryptocentrus

Gobius cryptocentrus (Valenciennes, 1837) Cryptocentrus fasciatus Gobiosoma fasciatum (Playfair, 1866) Cryptocentrus filifer Gobius filifer (Valenciennes, 1837) Cryptocentrus inexplicatus Smilogobius inexplicatus (Herre, 1934) Cryptocentrus insignitus Batman insignitus (Whitley, 1956)

Hoese & Steene, 1978 Polunin & Lubbock, 1979 Klausewitz, 1974a; Polunin & Lubbock, 1977; Yanagisawa, 1978; Cummins, 1979; Karplus et al., 1981 Klausewitz, 1969; Banner & Banner, 1981 Polunin & Lubbock, 1977 Karplus, pers. obs. Macnae & Kalk, 1962 Yanagisawa, 1978 Yanagisawa, 1976, 1978 Klausewitz, 1960, 1964; Clark et al., 1968; Magnus, 1967; Zander, 1967; Banner & Banner, 1981 Cummins, 1979 Luther, 1958a; Smith, 1959; Abel, 1960; Klausewitz, 1960; Macnae & Kalk, 1962; Zander, 1967; Polunin & Lubbock, 1977; Banner & Banner, 1981; Karplus et al., 1981 Polunin & Lubbock, 1977; Cummins, 1979 Yanagisawa, 1978 Hoese, pers. comm. Larson & Randall, pers. comm.

463

464

ILAN KARPLUS

Species of goby

Reference

Cryptocentrus leucostictus Gobius leucostictus (Gunther, 1871) Cryptocentrus lutheri Klausewitz, 1960 Cryptocentrus malindiensis Iotogobius malindiensis Smith, 1959 Cryptocentrus maudae Fowler, 1937 Cryptocentrus nigrocellatus Mars nigrocellatus (Yanagisawa, 1978) Cryptocentrus niveatus Valenciennes, 1837 Cryptocentrus obliquus Smilogobius obliquus (Herre, 1934) Cryptocentrus shigensis Kuroda, 1956 Cryptocentrus strigilliceps Mars strigilliceps (Jordan & Snyder, 1901) Ctengobiops aurocingulus Aparrius aurocingulus (Herre, 1935) Ctenogobiops crocineus Smith, 1959

Hoese, pers. comm. Klausewitz, 1960; Palmer, 1963; Karplus et al., 1981 Polunin & Lubbock, 1977 Hoese, pers. comm. Yanagisawa, 1978 Randall, pers. comm. Yanagisawa, 1976, 1978 Hoese, pers. comm. Hoese, pers. comm. Lubbock & Polunin, 1977

Polunin & Lubbock, 1977; Lubbock & Polunin, 1977; Yanagisawa, 1978 Ctenogobiops feroculus Lubbock & Polunin, 1977 Lubbock & Polunin, 1977; Polunin & Lubbock, 1977 Ctenogobiops maculosus Fourmanoir, 1955 Klausewitz, 1960; Lubbock & Polunin, 1977; Clark et al., 1968; Paulson, 1978; Karplus et al., 1981 Ctenogobiops pomastictus Lubbock & Polunin, 1977 Lubbock & Polunin, 1977; Cummins, 1979 Ctenogobiops tangaroai Lubbock & Polunin, 1977 Lubbock & Polunin, 1977 Eilatia latruncularia Klausewitz, 1974 Klausewitz, 1974b; Karplus et al., 1981 Flabelligobius fourmanoiri Smith, 1956 Randall, pers. comm. Gobionellus saepeplallens Gilbert & Randall Bohlke & Chaplin, 1968 Gobionellus stigmalophius Mead & Bohlke Bohlke & Chaplin, 1968 Gobius nudiceps Macnae, 1957 Cattrogobius nudiceps Lotilia graciliosa Klausewitz, 1960 Klausewitz, 1960, 1970; Banner & Banner, 1981; Karplus et al., 1981 Mahidolia mystacina Yanagisawa, 1978, 1982 Gobius mystacina Valenciennes, 1837 Nes longus Nichols, 1914 Longley & Hildebrand, 1941; Bohlke & Chaplin, 1968; Weiler, 1976 Psilogobius mainlandi Baldwin, 1972 Moehring, 1972; Preston, 1978 Stonogobiops dracula Polunin & Lubbock, 1977 Polunin & Lubbock, 1977; Banner & Banner, 1980; Hoese & Randall, 1982 Hoese & Randall, 1982 Stonogobiops medon Hoese & Randall, 1982

GOBIID FISHES AND ALPHEID SHRIMPS

465

TABLE I—continued Species of goby

Reference

Stonogobiops nematodes Hoese & Randall, 1982 Stonogobiops xanthorhinica Hoese & Randall, 1982 Tomiyamichthys oni Cryptocentrus oni (Tomiyama, 1936) Tomiyamichthys randalli Goren & Karplus, 1983 Vanderhorstia ambanoro Fourmanoir, 1957 Vanderhorstia delagoae

Hoese & Randall, 1982 Yanagisawa, 1976; Hoese & Randall, 1982 Yanagisawa, 1978, 1982, 1984

Gobius delagoae (Barnard, 1937) Vanderhorstia lanceolata Yanagisawa, 1978 Vanderhorstia mertensi Klausewitz, 1974 Vanderhorstia ornatissima Smith, 1959 Vireosa hanae Jordan & Smith, 1959 Yongeichthys pavidus Smith, 1959 Amoya signatus

Goren & Karplus, 1983 Hoese, pers. comm. Macnae & Kalk, 1962; Klausewitz, 1964; Magnus, 1967; Banner & Banner, 1981; Karplus et al., 1981 Yanagisawa, 1978 Klausewitz, 1974b; Yanagisawa, 1978, 1982, 1984; Karplus et al., 1981 Polunin & Lubbock, 1977; Yanagisawa, 1978; Cummins, 1979 Harada, 1969; Yanagisawa, 1978 Polunin & Lubbock, 1977

TABLE II Species of burrowing alpheid shrimp reported associated with gobiid fishes Species of shrimp

Reference

Alpheus bellulus

Miya & Miyake, 1969; Harada, 1969, 1972; Yanagisawa, 1976, 1978, 1982, 1984; Karplus et al., 1981 Alpheus brevicrixtatus Harada, 1969; Yanagisawa, 1978 Alpheus brevirostris Karplus et al., 1981 Alpheus djiboutensis Luther, 1958a, b; Klausewitz, 1960; Harada, 1972; Karplus et al., 1972a, b, 1981; Karplus, 1970, 1981; Paulson, 1978; Banner & Banner, 1981, 1983 Alpheus floridanus Weiler, 1976 Alpheus crassimanus Macnae, 1957; Thomassin, 1971; Farrow, 1971 Alpheus malabaricus Macnae & Kalk, 1962 Alpheus ochrostriatus Karplus et al., 1981 Alpheus purpurilenticularis Karplus, 1981; Karplus et al., 1981 Alpheus randalli Banner & Banner, 1980; Hoese & Randall, 1982 Alpheus rapacida Macnae & Kalk, 1962; Magnus, 1967; Moehring, 1972; Preston, 1978; Yanagisawa, 1978, 1982, 1984; Banner & Banner, 1981, 1982 Alpheus rapax Macnae & Kalk, 1962; Magnus, 1967; Moehring, 1972; Polunin & Lubbock, 1977; Preston, 1978; Banner & Banner, 1981, 1982; Karplus et al., 1981 Karplus et al., 1981 Alpheus rubromaculatus

Great Barrier Reef (Cummins, 1979). Other species like Eilatia latruncularia and Tomiyamichthys randalli have so far only been described from the northern Red Sea (Klausewitz, 1974b; Goren & Karplus, 1983). One should be very cautious when speculating about endemism because the collection of associated

466

ILAN KARPLUS

Fig. 2–Species shrimp associated with gobies in the northern Red sea: A, Alpheus purpurilenticularis ; B A. rapax ; C, A. rubromaculatus; D, A. brevirostris ; E, A. bellulus; F, A. ochrostriatus; G, A. djiboutensis ; H, Alpheus sp.; from Karplus et al. (1981).

GOBIID FISHES AND ALPHEID SHRIMPS

467

gobies is still very sporadic and patchy, Lotilia graciliosa, first described from the Red Sea (Klausewitz, 1960) was known only in that locality until it was reported 21 years later in the Fiji Islands (Banner & Banner, 1981). L. graciliosa and Eilatia latruncularia were recently observed in the Great Barrier Reef (Hoese, pers. comm.). Some of the associated alpheid shrimps have a very wide distribution. Alpheus rapax, for example, was recorded from the Red Sea, Moçambique, Hawaii, and Australia (Macnae & Kalk, 1962; Preston, 1978; Banner & Banner, 1981; Karplus et al., 1981). Other species were only recorded from a single area, such as A. purpurilenticularis that has so far only been reported from the Red Sea (Karplus et al., 1981). In the case of the shrimp, even more so than for the goby, the collection of animals is very limited and their identification complex, so that no conclusion can be drawn from the available record on the distribution and endemism of associated shrimps. ECOLOGY OF ASSOCIATIONS The species of gobiid fishes and alpheid burrowing shrimps in the associations live in various kinds of sediments, ranging from silty mud to coral rubble, in the intertidal zone down to a depth of more than 50 m, and in a variety of habitats, e.g., mud flats and sea-grass beds (Polunin & Lubbock, 1977; Yanagisawa, 1978). The ecology of these associations is usually treated very superficially, either as an appendix to taxonomical studies, as a single element within a much larger ecological system, or as a background to behavioural studies. Most of these studies are of a general descriptive nature usually concerned with a single association, or more often, with a single species of goby. They refer to depth range, the character of the sediment and occasionally to the type of habitat. These studies have not attempted to analyse qualitatively or quantitatively differences between sympatric association. Polunin & Lubbock (1977) were the first to deal with the problem of habitat specificity of shrimp-associated gobies. In their study, carried out in the Seychelle Islands, they examined the distribution of 13 species of associated gobies in a small bay on the northwestern coast of Mahe. Seven different sandy habitats were defined in this protected bay that contained both welldeveloped coral reefs and extensive sandy habitats, extending from the intertidal zone to lower than 30 m depth. A marked degree of habitat segregation was exhibited by the different species of gobies. Five out of the 13 examined species were found in only one type of habitat while four additional species were found in only two habitats. The publication reported more generally on habitat segregation of three species of gobies at Aldabra Atoll. Most goby species showed a tendency to form local aggregations, frequently made up of one species (Polunin & Lubbock, 1977). This phenomenon could have resulted either from social interaction between gobies, from habitat segregation, or from both. In southern Japan, depth range and substratum specificity were investigated in 20 species of shrimpassociated gobies (Yanagisawa, 1978). Depth and substratum were each classified into four categories and the occurrence of gobies in these was recorded. The bottom substratum inhabited by each gobiid species is rather restricted and similar, among the localities, each species apparently having its own depth preference. The ecology of six species of gobies and their four types of shrimp partners was studied at One Tree Reef, in the Great Barrier Reef (Cummins, 1979). The specificity of the substratum defined by proportions of gravel and sand, was found to be almost entirely lacking. No single species of goby and no type of shrimp was segregated from the others, according to Pielou’s (1969) index of segregation, and to the nearest neighbour distances. These results contrast with the local aggregations and marked level of habitat segregation reported for species of gobies from Lizard Island, Great Barrier Reef (Polunin & Lubbock, 1977).

468

ILAN KARPLUS

The distribution of several sympatric species of gobies and shrimps was studied in the northern Red Sea (Karplus et al., 1981). The vertical distribution was analysed by a series of transects, parallel to the shore set at depth intervals of 2 m, down to a depth of 20 m. Different species of gobies and shrimps exhibited a different vertical distribution as regard both depth range and relative abundance. In shallow and in deep water, the species of shrimps differed whereas, as far as the gobies were concerned, the same species were found, their number decreasing with depth. In the shallow water of a sandy lagoon, four species of shrimps were found in local aggregations, with little overlap in different sub-habitats (Fig. 3). These sub-habitats were defined by the distance from the reef and the character of the sediment (mean grain size and sorting). It is difficult to evaluate the independent effect of these variables as the two are intercorrelated. The preference of different shrimp species for different types of sediment could depend on differing diets, or on a different structure of the fine hairs of the chela which are used in sediment transport. A similar segregation of burrowing alpheid shrimps in specific habitats has been described for several associated (Macnae & Kalk, 1962) and free living species (Nolan & Salmon, 1970). A. bellulus from Japan (Yanagisawa, 1982) and A. floridanus from Puerto Rico (Weiler, 1976) were both, however, reported to live on a wide range of substrata. In contrast to shallow-water shrimps, deep-water shrimps in the Red Sea do not show segregation in different sub-habitats, probably due to the more uniform character of the sediment at that depth. In contrast to the Red-Sea burrowing shrimp of shallow water, the associated gobies show less habitat specificity and occur in different habitats depending on the species of their shrimp partner. This may possibly be due to the fact that these gobies do not burrow, and unlike the shrimps, do not feed on organic material found in the sediment or on epifauna and interstitial animals (Magnus, 1967; Harada, 1969). Burrowing fish, like burrowing alpheids, show a stronger attachment to a specific type of substratum (Rao, 1939; Colin, 1972; Webb, 1974). BURROW STRUCTURE, CONSTRUCTION, AND DYNAMICS BURROW STRUCTURE The burrow structure can be divided into two components, the structure and number of its openings and its subterranean structure (Karplus et al., 1974). The typical shape of the burrow opening of a goby-associated shrimp is asymmetrical, its roof and sides being embedded by the shrimp in coral and shell fragments to prevent the collapse of its walls. It has a sandy floor leading to a shallow sloping ramp consisting of sand transported from within the burrow; the ramp is often used by the goby as a lookout post. This type of burrow opening has been described in connection with A. djiboutensis (Luther, 1958a; Karplus et al., 1972a, 1974), A. crassimanus (Farrow, 1971), A. bellulus (Yanagisawa, 1984), and A. purpurilenticularis (Karplus, 1979). The structure of the burrow opening of different species is not necessarily similar. Various types of burrow openings have been found in the northern Red Sea for different goby-associated shrimp (Karplus et al., 1974). In addition to the typical asymmetrical burrow opening, two symmetrical ones, an elevated tube-like opening and a funnel-like opening have also been described (Fig. 4). The appearance of symmetrical and asymmetrical openings depends largely on the substratum. Mainly asymmetrical openings are found in the burrows of A. djiboutensis and A. purpurilenticularis located in coarse and intermediate sand while symmetrical openings, embedded on all sides in coral fragments are more frequently found in the burrows of A. rapax (tube-like) and A. rubromaculatus (funnel-like), both located in fine sediments. A detailed analysis of the burrow opening structure of four different types of shrimps was carried out at One Tree Reef (Cummins, 1979). Each burrow entrance was classified according to four categories (e.g. angle of descent) and seven different points were measured (e.g. width of entrance). Numerical and non-

GOBIID FISHES AND ALPHEID SHRIMPS

469

Fig. 3.—Distribution of associated gobies and shrimps in the lagoon of the Elat Nature Reserve: each square is m; coarse stippling, coarse sediment; fine stippling, fine sediment; blank, intermediate sediment; black, living and dead corals; horizontal rules, stones; sloping rules, tunnels in the reef; Alpheus purpurilenticularis and Amblyeleotris steinitzi; Alpheus rubromaculatus and Lotilia graciliosa; Alpheus rapax and Ctenogobiops maculosus; Alpheus djiboutensis and Cryptocentrus lutheri; from Karplus et al. (1981).

numerical attributes of the burrows were simultaneously analysed using multivariate techniques. No single feature or combination of features was diagnostic of any one type of shrimp although two groups, each comprising two types of shrimps, could be distinguished. A shallow and narrow groove stretching from the burrow opening has been described for several shrimp species. A short (20–30 cm) and rather deep (10 cm) groove was described for A. djiboutensis (Luther, 1958a), a somewhat longer (40–50 cm) and shallow one (1–2 cm) for A. purpurilenticularis (Karplus, 1979), while a shallow (2–3 cm) and very long (up to 80 cm) groove was described for A. bellulus (Yanagisawa, 1984). These grooves facilitate the activities of the shrimp outside their burrows (Karplus, 1979). The number of openings of a single burrow can be accurately determined either by complete retrieval of a cast of the burrow or by squirting a liquid dye into a burrow entrance and noting from which holes the solution escapes (Rice & Chapman, 1971). Usually, single openings were found for four types of shrimp at One Tree Reef (Cummins, 1979). Single paired openings were described for A. floridanus (Shinn, 1968; Weiler, 1976) and multiple paired openings for A. crassimanus (Farrow, 1971). Some specificity was found in the number of burrow openings (ranging from 1–6) for several species of goby-associated shrimps in the

470

ILAN KARPLUS

Fig. 4. —The three basic structures of burrow openings of goby-associated shrimps in the northern Red Sea: Al, symmetrical tube-like opening; A2, symmetrical funnel-like opening; B, asymmetrical opening; dotted areas, sand transported by shrimp; hatched area, undisturbed sediment; black patches, coral, shell, and stone fragments; from Karplus et al., 1974).

northern Red Sea. The number of openings is not determined by the substratum because both single and multiple openings are found in fine sediments. It is rather the specific activity of the shrimp which dictates the number of its burrow openings (Karplus et al., 1974). Our knowledge of the subterranean structure of the goby-associated shrimps’ burrows, prior to the application of resin casts, was speculative, usually under-estimating their actual size. The methods for studying burrow structure were either not specified or consisted of digging up the burrow or pumping water into it. Only a 20–30 cm long burrow was suggested by Luther (1958a) for A. djiboutensis and by Palmer (1963) for an alpheid associated with Cryptocentrus lutheri. A slightly longer burrow (40 cm), usually leading under stones and other hard objects was described by Harada (1969) for Alpheus bellulus. Following unsuccessful attempts to dig out the associated goby and shrimp, Smith (1959) concluded that the burrows were deep. The burrows of two shrimps probably A. rapax and A. rapacida were described as shallow, at least 70 cm long, parallel to the surface, and occasionally branching off (Magnus, 1967). The exact study of the structure of infralittoral burrows of crustaceans only started about twenty years ago when polyester and epoxy resins became available. Resin casts are superior to those made of plaster of Paris whose use is limited to the intertidal zone. The resins have several virtues: an ability to harden underwater, a controllable viscosity, strength, impregnation of substratum and possible “freezing” of burrow producers and co-habitants (Farrow, 1971). The burrow structure of eight different species and four types of associated shrimps has been investigated by the application of epoxy and polyester resins—A. crassimanus (Farrow, 1971), A. djiboutensis, A. purpurilenticularis and A. rubromaculatus (Karplus et al., 1974), A. floridanus Weiler, 1976), A. rapax (Karplus et al., 1974; Preston, 1978), A. rapacida (Preston, 1978), A. bellulus (Yanagisawa, 1984), tapestry, pink, banded and pale shrimp types (Cummins, 1979).

GOBIID FISHES AND ALPHEID SHRIMPS

471

Goby-associated shrimps usually have shallow burrows branching off irregularly and in close contact with hard objects like coral and stone boulders within the sediment (Fig. 5). Due to the tendency of sand to collapse, these hard objects were used to support the subterranean burrow structure determining to a high degree their irregular structure and their lack of species’ specificity. The effect of the substratum on the burrow structure has been demonstrated for A. crassimanus (Farrow, 1971). When located in a coarse substratum with hard objects, it had an irregular burrow structure but, when located in muddy silts, the burrow had a regular dichotomous branching pattern. The burrow casts retrieved from sediment lacking supporting objects had an even diameter at different points, while the diameter of a burrow leading under rocks or corals was irregular with occasional chamberlike enlargements (Karplus et al., 1974; Yanagisawa, 1984). The cross section of the burrows is shaped by their position: horizontally, it was elliptical and vertically, it was more circular (Karplus et al., 1974). The floor of the burrows of A. floridanus (Shinn, 1968; Weiler, 1976) and A. bellulus (Yanagisawa, 1984) is relatively smooth due to transport of sand by the shrimp, while the burrow’s roof is more irregular. The burrow walls of four types of shrimps at One Tree Reef differed from the rest of the investigated burrows by being substantially reinforced in their upper and lower sections. In vertical burrows the walls were entirely lined by coral and shell fragments while in sloping burrows, only the roof and sides were lined (Cummins, 1979). The larger the shrimp the larger was also the diameter of its burrow, its length and depth (Karplus et al., 1974; Preston, 1978; Yanagisawa, 1984). The fact that the burrows did not penetrate deep into the substratum is remarkable, as many of the sediment-feeding organisms attempt to utilize the maximum thickness of the sediment in the construction of deep burrows (Farrow, 1971). The intensive feeding activity of the shrimp in the vicinity of the burrow’s opening and the shift of the opening in different directions probably compensates for the relatively shallow burrow (Karplus et al., 1974). BURROW CONSTRUCTION The behaviour during burrow construction has been described for several species of goby-associated shrimps in aquaria (Harada, 1969; Karplus, Szlep & Tsurnamal, 1972a) and in the sea (Luther, 1958a; Macnae & Kalk, 1962; Magnus, 1967; Farrow, 1971; Yanagisawa, 1984). Different species exhibited similar burrowing behaviour inside and outside the burrows. A. djiboutensis (Karplus et al., 1972a) and A. bellulus (Harada, 1969) used three different subterranean digging techniques: (1) digging with the first pair of chelae into a vertical sand wall and twisting until the sand collapses, (2) digging with the second pair of chelae and the third and fourth pairs of pereipods and often also the third maxiliped, and (3) digging with the pleopoda, the posterior end of the body directed towards the burrow entrance. Digging with the first pair of chelae outside the burrow has been described for A. rapax, A. rapacida, and A. purpurilenticularis in the process of accumulating sediment in the burrow from the outside upper intact layer (Magnus, 1967; Karplus, 1979). Digging with the walking legs and second chelae may be practised outside the burrow for feeding purposes (Karplus et al., 1972a). Digging with the pleopoda outside the burrow has only been described for A. djiboutensis (Luther, 1958a), while in A. bellulus (Harada, 1969) it is confined to the inside of the burrow, occasionally close to the entrance as evidenced by the turbid water streaming out of the burrow. Digging with the pleopoda, often practised inside the burrow, is usually not found outside it because during that activity the shrimp’s head is directed towards the entrance and the alpheid’s rapid withdrawal is hindered (Magnus, 1967). The transport of sand grains and small stones from the inner parts of the burrow is done by the first pair of the strongly compressed chelae. The two chelae join together to form a kind of spade broadened by rows of long hairs fringing the dorsal and ventral margins of both chelae (Magnus, 1967; Miya & Miyake, 1969).

472

ILAN KARPLUS

Fig. 5.—Resin casts of burrows of young Alpheus bellulus: A, B, and D, associated with Amblyeleotris japonica; C, associated with Stonogobius sp.; scale bars are 10 cm; from Yanagisawa (1984).

Small amounts of sediment are lifted and transported on the chelae while large amounts of sediment remain on the ground and are moved by the chelae, acting as a bulldozer. Occasionally a large flat object— like a shell—is used for a more efficient transfer of the sediment (Magnus, 1967). Twiglets of corals and shell fragments are grasped by the first pair of chelae and carried out of the burrow to the area above the entrance

GOBIID FISHES AND ALPHEID SHRIMPS

473

(Farrow, 1971; Magnus, 1967; Yanagisawa, 1984; Karplus et al., 1972a). A. bellulus seizes coral and shell fragments only with the small chela, never with the snapping chela but secures the fragments with the latter’s assistance in the burrow aperture to reinforce it. Plasticity, however, is attributed to the use of these chelipeds since individuals that have lost the small chela have been observed to handle this material with the remaining chela (Harada, 1969). A single, non-identified gobiid fish was observed at Aldabra Atoll in the process of assisting in the burrow construction of its shrimp partner, A. crassimanus. This goby enlarged the upper part of the dichotomous branching burrow in the form of a U-tube by removing mouthfuls of mud from within the burrow and ejecting them at the periphery (Farrow, 1971). No other shrimp-associated goby has been reported to take part in burrow construction. Several gobies, Nes longus (Weiler, 1976), Cryptocentrus caeruleopunctatus (Magnus, 1967), Amblyeleotris japonica (Harada, 1969), A. steinitzi and Cryptocentrus lutheri (Karplus, pres, obs.) proved their inability to burrow in aquaria when deprived of their shrimp partner, and were only able to form a depression at the bottom by splashing sand around. It is thus evident that the burrow is constructed and maintained by the shrimp in almost all cases. BURROW DYNAMICS Daily changes in the position of the burrow openings of goby-associated shrimps have been reported for several species and different types from different localities (Magnus, 1967; Karplus, Szlep & Tsurnamal, 1974; Polunin & Lubbock, 1977; Cummins, 1979; Yanagisawa, 1982, 1984). Changes in the burrow openings of a non-identified shrimp associated with Ctenogobiops feroculus at Aldabra Atoll were demonstrated by indicating on a map the burrow positions on two consecutive days, as well as the changes in frequency distribution of nearest neighbour distances (Polunin & Lubbock, 1977). Changes in the burrow openings of two species of shrimps, probably Alpheus rapax and A. rapacida, were studied in the northern Red Sea (Magnus, 1967). The positions of the burrow entrances were marked by inserting two iron rods on both sides of the opening. Changes of the position were continuous and unidirectional, usually averaging 40 cm per day and were correlated with the size of the shrimp, character of the sediment, and occurrence of obstacles within it. The changes in the position of the burrow openings of A. bellulus associated with Amblyeleotris japonica were studied in Japan (Yanagisawa, 1982, 1984). Individual burrows were identified at localities of high densities by the tagging of the associated goby that seldom exchanged its shrimp partner. The burrows were also identified by the combination of the sides of the large chela of both male and female shrimps and by their size and coloration. The distance of daily shift of the entrance reached 160 cm. There was no regular shift pattern but the positions of a burrow were confined, over several months, to a limited range usually with a horizontal extent of about a half to two square metres. The dynamics of burrow openings of several sympatric species of burrowing goby-associated shrimps have been studied in a shallow lagoon in the northern Gulf of Elat (Karplus et al., 1974). Burrows were not studied in densely populated areas to avoid confusion. The burrow openings were marked with iron rods, similarly to Magnus (1967) and their daily changes in position—distance and angle—were recorded for ten days, but several marked burrows were occasionally examined over a period of six months. The maximal daily changes of the burrow entrance position, which ranged from 30 to 80 cm were species specific, as also was the relatively small area to which the openings were confined. The daily displacement of the burrows of the different species was correlated with their type of substratum and their proximity to large coral boulders. The coarser the sediment the larger the change while the closer to the reef wall or large coral boulders the smaller the daily shift in position. The changes in the burrow structure are apparently restricted to the upper

474

ILAN KARPLUS

shallow parts of the burrow, while the deeper parts, often leading under and between large boulders, remain stable (Karplus et al., 1974). A comparative study of the changes of the burrow opening of four types of shrimps was carried out at One Tree Reef (Cummins, 1979). Maximal daily shifts were type-specific and ranged from 50 to 160 cm. Each burrow usually had three openings but usually only one was open at any time. The entrance of each burrow recurred at exactly the same position, even when the recordings were made after a period of two years. Three different mechanisms for shifting the burrow opening have been described for different species. According to Magnus (1967), the shift resulted only from the feeding activity of the shrimp which removed substratum from the area overlying the burrow, thus continuously shifting the opening backwards. The bigger the shrimp the more substratum was removed. The shift of the burrow is due to the activities of both partners according to Karplus et al. (1974). The irregular multidirectional shift results from the activity of Amblyeleotris steinitzi wedging its head through the substratum to create a new opening. Alpheus purpurilenticularis follows and enlarges the new opening while the old one rapidly collapses (Fig. 6). The fixed changes in the position of the burrow openings of the Australian types (Cummins, 1979) probably result only from the activity of the shrimp. Both the upper and lower parts of the burrow are reinforced by coral fragments, so they are stable and the shrimp is only alternately clearing or blocking them with sediment, thus reforming the openings at the same positions. The change in the position of the burrow opening is important for both shrimp and fish. The shrimp thereby gains access to additional suitable substrata outside the burrow. The territorial fish not only protects its shelter against intruders but is actively controlling the mobile system of burrow openings, spacing them out and thus controlling the density of the associations. Amblyeleotris steinitzi reacted to a camera placed on a tripod in front of its burrow by shifting the entrance 25 cm away from the camera (Karplus, pers. obs.). Similarly, instead of having to abandon its burrow because of a territorial conflict with its dominant neighbour, a goby can form a new opening further away. During the reproductive season, the fish has to reach its partner and is thus exposed to predators. The shift of the burrow opening towards the partner may reduce the danger. A detailed study of the daily changes of burrow entrance features was made at on One Tree Reef (Cummins, 1979). Some individuals of each type of shrimp showed the same structure for five days, while others showed a daily variation in the structure of the entrances. Daily fluctuations occurred in the number of burrow openings of several Red Sea shrimps. A complete blockage of all the openings was occasionally observed; this lasted for 1 to 3 days, after which a significantly greater shift occurred indicating intensive subterranean digging activity, even though no activity was observed on the surface (Karplus et al., 1974).

DIET AND FEEDING BEHAVIOUR OF GOBY AND SHRIMP Few data are available on the diet of associated gobies and shrimps studied through feeding movements and analysis of stomach content. Fishes of the genera Cryptocentrus, Ctenogobiops, Vanderhorstia, and Amblyeleotris feed by picking organisms out of the sand or by taking small mouthfuls of sand which they filter through their gill rakers thereby extracting small organisms (Magnus, 1967; Hoese & Allen, 1976; Cummins, 1979). A. steinitzi and Cryptocentrus caeruleopunctatus were occasionally observed while feeding on planktonic organisms (Karplus, pers. obs.). Amblyeleotris japonica was observed to dash 1–5 cm above the sea floor near its burrow entrance, while performing repeated biting motions. Young fish exhibited this behaviour more frequently than adult and sub-adult fish (Yanagisawa, 1982).

GOBIID FISHES AND ALPHEID SHRIMPS

475

Fig. 6.—Dynamics of the upper burrow system in the association of Amblyeleotris steinitzi with Alpheus purpurilenticularis: a, fish located in front of burrow opening (O1); b, formation of new burrow opening (O2) by fish; c, opening O1 collapsed, opening O2 functional; d, formation of additional opening (O3) by fish; e, openings O1 and O2 collapsed, opening O3 functional; from Karplus et al. (1974).

Examination of stomach and intestinal contents of A. japonica revealed that more than 90% of their food intake was corophiid amphipods and other small-sized crustacean species (Harada, 1969; Yanagisawa, 1982). The stomach content of a single Cryptocentrus lutheri consisted of small crustaceans, gastropods, and bivalves living on and in the sediment (Karplus, pers. obs.). The stomach contents of six shrimp-associated gobies at One Tree Reef were very similar, comprising invertebrates such as amphipods, copepods, bivalves, and worms as well as algae (Cummins, 1979). The stomach contents of a single non-identified shrimp-associated goby at Palau consisted entirely of shrimp larvae (Bayer & Harry-Rofen, 1957). Alpheus rapax and A. rapacida in Hawaii (Moehring, 1972), and different types of shrimp at One Tree Reef (Cummins, 1979) were observed while introducing pieces of algae in their burrows. The Hawaiian shrimp were also observed to snip off and take into their burrows pieces of worm tubes protruding around the burrow entrances (Moehring, 1972). Digging in the sediment with the first pair of chela and the subsequent introduction of the sediment into the burrow has been described for A. purpurilenticularis (Karplus, 1979), A. rapax (Karplus, 1976), A. djiboutensis (Karplus, 1976), and probably A. rapacida (Magnus, 1967). Only the upper undisturbed sediment layer, approximately 5–10 mm thick, probably rich in

476

ILAN KARPLUS

Fig. 7.—Alpheus djiboutensis cleaning Cryptocentrus cryptocentrus (from Karplus et al., 1972a).

organic materials, was introduced in this way; sediment previously removed from the burrow was never reintroduced (Magnus, 1967). The granulometric character of the undisturbed sediment, close to the burrow opening of A. floridanus did not differ significantly from that ejected from the burrow (Weiler, 1976). A comparison between the undisturbed sediment and the one removed from the burrow, as regards its organic content and composition, would be of interest but has never been determined. The stomach contents of A. bellulus consist of fairly large amounts of unidentified materials and some nematodes, copepods and amphipods, so that the diet is assumed to consist mainly of detritus, epi- and interstitial fauna (Harada, 1969). The interrelationship between the feeding behaviour of goby and shrimp is of interest. Several authors have stated that the goby seeks food in the sediment excavated by the shrimp (Abel, 1960; Farrow, 1971; Hoese & Steene, 1978) or eats small invertebrates which are disturbed by the sediment ejected from the burrow (Magnus, 1967). In both cases, the goby benefits from the digging activity of the shrimp. A mutual benefit involving feeding has been described for A. djiboutensis cleaning its fish partner Cryptocentrus cryptocentrus (Fig. 7). During the cleaning process, the shrimp’s first pair of chelae were placed on the fish

GOBIID FISHES AND ALPHEID SHRIMPS

477

while its second pair moved repeatedly from the fish to the mouth region of the shrimp (Karplus et al., 1972a). Possible predation of the goby on its shrimp partner’s larvae was suggested by Bayer & Harry-Rofen (1957) as well as by Herald (1961). This conclusion is based on a single animal and the shrimp larvae were not identified. Detailed stomach content analysis of the goby during the reproductive season of the shrimp as well as observation of the interactions between gobies and shrimp releasing larvae in artificial burrows could clarify this issue. Despite some possible overlap in the diet of gobies and shrimp (e.g. both eat polychaetes), the shrimp is basically a detritus-feeder, whereas the goby feeds on small invertebrates found both in the plankton and in the sediment close to the burrow (Karplus, 1979). THE DAILY ACTIVITY RHYTHM OF GOBY AND SHRIMP Observations on the goby-shrimp associations in the Red Sea (Magnus, 1967; Karplus et al., 1972a, 1974; Karplus, 1976, 1979), in the Seychelles (Polunin & Lubbock, 1977), and in Japan (Yanagisawa, 1982, 1984) indicated that both partners emerge from the burrow only during the day and not at night. At night, the burrow openings are usually blocked, due either to their collapse (Magnus, 1967; Karplus et al., 1974; Karplus, 1979) or to their intended closure such as by Alpheus bellulus transferring sediment to the entrance from within the burrow (Yanagisawa, 1984). Amblyeleotris japonica (Yanagisawa, 1982, 1984), A. steinitzi (Karplus et al., 1974), and Cryptocentrus caeruleopunctatus (Magnus, 1967) were all observed to renew the daily activity of their association outside the burrow, by cautiously breaking through the sediment. They are followed by their shrimp partners that immediately start to enlarge the opening. C. caeruleopunctatus were reported to resume their activity outside the burrow in the Red Sea with sunrise (Magnus, 1967). The time of emergence of Amblyeleotris japonica in southern Japan varies among the associations although most entrances are open by about one hour after sunrise (Yanagisawa, 1984). The earliest activity of goby-shrimp associations in the northern Gulf of Elat was recorded 35 minutes prior to sunrise but, on rare occasions it started as late as noon. Some differences between species were found in the light intensity at the time of emergence (Karplus, 1976; Table III). All activity outside the burrow is terminated when the goby withdraws into the burrow. A. japonica enters its burrow by sunset (Yanagisawa, 1982). Cryptocentrus caeruleopunctatus ends the activity outside the burrow at the latest 20 minutes after sunset (Magnus, 1967). Some differences of light intensity at the final retreat of the goby into the burrow were found for three sympatric Red-Sea species (Karplus, 1976; Table III). TABLE III Initiation and termination of activity of three species of gobies in the northern Red Sea in relation to light intensity: figures show number of gobies at start of activity and at final retreat into burrow in parentheses Light intensity, lux

Amblyeleotris steinitzi

Cryptocentrus lutheri

Ctenogobiops maculosus

<150 150–300 300–450 450–850 850–2500

1 4 2 1 4

1 2 7 3 4

13 2 1 2 3

(2) (5) (9) (3) –

(5) (11) (1) – –

(19) (4) (1) – –

478

ILAN KARPLUS

Light intensity, lux

Amblyeleotris steinitzi

Cryptocentrus lutheri

Ctenogobiops maculosus

2500–5000 >5000

3 3

2 2

1 1

– –

– –

– –

It was observed that several species of shrimps spent less time outside the burrow and closer to its entrance towards the end of the day. Shrimps place shell and coral fragments around the opening, probably to reinforce it and to reduce blockage at night. This activity was particularly marked in A. purpurilenticularis and A. rapax and less so in A. djiboutensis whose burrow openings are less collapsible due to their being more reinforced by shell and coral fragments (Karplus, 1976). Termination of the daily activity is more synchronous than its beginning (Table III). This phenomenon is probably due to the fact that the end of the activity is triggered by low light levels. The start of activity is more variable, being mainly guided by an endogenous rhythm, since the burrow openings usually collapse overnight and the goby within the burrow cannot perceive the light level (Karplus, 1976). The daily rhythm of the shrimp’s activity (i.e. the number of exits from the burrow, the exit’s duration and total time spent outside the burrow) have been studied for three goby-associated shrimps in the northern Red Sea on twelve consecutive calm days (Karplus, 1976, 1979). Despite some differences between species —Alpheus purpurilenticularis, A. rapax, and A. djiboutensis (Fig. 8) —all spent about a third of the time outside the burrow during early morning, reduced that time around noon and spent the majority of the time outside the burrow in the late afternoon. Exit duration and not number of exits accounted for the differences in total time spent outside the burrows, by different species at different times of the day. The duration of each exit in early morning was intermediate, low around noon and long in the late afternoon (Fig. 8). Different types of shrimps at One Tree Reef showed the reverse trend: these shrimps spent less time outside the burrow in the early morning and late afternoon and more time at noon (Cummins, 1979). Observations on the activity outside the burrow of A. bellulus associated with Amblyeleotris japonica in Japan revealed differences between the sexes. Male shrimps came out of the burrow more often than females and also spent more time outside the burrow (Yanagisawa, 1984). Magnus (1967) was the first to report on the introduction of sediment into the burrow by Red Sea gobyassociated shrimps and on the changes in this activity throughout the day, its being most pronounced in early morning and late afternoon. Changes in the sand transport and digging activity of several gobyassociated shrimps has been investigated in the northern Red Sea (Karplus, 1976). Despite specific differences between species (Fig. 9) —A. purpurilenticularis, A. djiboutensis, and A. rapax—all left their burrows in the morning and noon mostly loaded with sediment and coral fragments while, in the late afternoon, they left their burrows mainly with empty chelae. Digging with the first pair of chelae was slight in the morning and noon and pronounced in the late afternoon. The frequency of entrances into the burrow with chelae loaded with sediment was very low in the morning and increased substantially in the late afternoon. The high frequency of exits in the morning with chelae loaded with sediment is probably due to the clearance of the subterranean burrow which collapsed at night. Feeding related digging outside the burrow is mainly practised in the late afternoon and is followed by sediment introduction into the burrow. This sediment, rich in organic material will probably be consumed when activity outside the burrow is terminated. Differences between species in shrimp activity outside the burrow are probably related to the relative importance to their diet of the sediment outside the burrow. In southern Japan, during rough weather, red tide, and water temperatures below 15ºC, the burrow remained closed and their residents stayed inside it all day (Yanagisawa, 1982, 1984). In the northern Red Sea,

GOBIID FISHES AND ALPHEID SHRIMPS

479

Fig. 8.—The daily activity rhythm of three goby-associated prawns in the northern Red Sea measured during periods of 10 min: sunrise and sunset are shown; significant differences in activity at any given hour are indicated by different letters (a, b); from Karplus (1976).

when the sea is rough no activity was recorded in shallow water, while at 8–10 m or more depth the activity outside the burrow continued (Karplus, 1976).

480

ILAN KARPLUS

Fig. 9.—The daily rhythm of transporting sediment from within the burrow, digging and introducing sediment into the burrow by three goby-associated shrimps in the northern Red Sea measured during periods of 10 min: all cases, except those marked with an asterisk, showed significant deviation from random distribution of activity compared with no activity; from Karplus (1976).

Associations located in very shallow water followed the rhythm of the tides superimposed on a regular diurnal rhythm. Cryptocentrus caeruleopunctatus and Alpheus djiboutensis were still active outside the burrow when the water level was about 10 cm above the substratum at Marsah-Murach, a shallow bay 20 km south of Elat. As the water level further receded, only the goby remained positioned in the lower part of the burrow funnel, but retreated when the water level was about 3–5 cm above the substratum. Associations of the same species, several metres distant in slightly deeper water, followed a regular diurnal activity rhythm (Karplus, 1976). A similar effect of the tides on the activity of the associations found in very shallow water, was observed on the reef flat at Heron Island, Great Barrier Reef (Cummins, 1979).

GOBIID FISHES AND ALPHEID SHRIMPS

481

The main environmental factor regulating the activity of goby and shrimp outside the burrow is probably light intensity. The activity of the shrimp depending on the presence of the goby at the burrow entrance is also affected by the collapse of the subterranean burrow and the occurrence of food in the sediment outside the burrow. AGGRESSIVE BEHAVIOUR AND TERRITORIALLY OF GOBY AND SHRIMP The occurrence of aggressive interactions between gobies over burrow ownership have been reported for Ctenogobiops feroculus, C. pomastictus, and Vanderhorstia ornatissima in the Seychelles (Lubbock & Polunin, 1977; Polunin & Lubbock, 1977), a non-identified goby at Aldabra (Farrow, 1971), Vanderhorstia delagoae (Magnus, 1967), Amblyeleotris steinitzi and Cryptocentrus caeruleopunctatus in the northern Red Sea (Karplus et al., 1974; Karplus, 1979), and Amblyeleotris japonica in southern Japan (Yanagisawa, 1982, 1984). Competition for burrow ownership in shrimp-associated gobies is due to several factors. (1) The reduction of number of burrows due to the pairing of the shrimps: Alpheus bellulus starts its benthic life alone and very soon associates with a small Amblyeleotris japonica. Within several months the shrimp finds a mate and the number of burrows is thus reduced by half. Intense competition then occurs between the gobies over the remaining burrows, since the goby is only paired as an adult for short periods of time (Yanagisawa, 1982, 1984). (2) Competition for the larger burrows: A large Cryptocentrus caeruleopunctatus possessing a relatively small burrow was observed to take over a larger burrow possessed by a small goby (Karplus et al., 1974). The positive size correlation between goby and shrimp (Palmer, 1963; Klausewitz, 1964; Karplus et al., 1974; Cummins, 1979; Yanagisawa, 1984) is probably due to competition for the larger burrows. (3) Periodical desertion of burrows: Vanderhorstia delagoae and V. ornatissima are two species with a loose attachment to their shrimp partners. They often leave their burrow during feeding excursions within their wide home range which comprises the territories of several associations. They often try to take shelter in the nearest burrow when endangered (Magnus, 1967; Polunin & Lubbock, 1977). Most of the gobies form temporary pairs during the reproductive season. One of the fish in the pair has to abandon its own burrow which will rapidly be taken over by another one (Yanagisawa, 1982). After leaving its mate, this goby will have to acquire a new burrow. The burrow entrance is the centre and most protected part of the goby territory (Karplus, 1979). Usually a single or a pair of gobies occupies one burrow and protects its surroundings. In areas of high density Amblyeleotris japonica were sometimes reported to occupy several burrows at the same time and to protect their surroundings against intruders (Yanagisawa, 1982). Aggressive interactions between neighbouring gobies are usually frequent but of lower intensity than in disputes over burrow ownership. These aggressive interactions regulate the shift in the burrow opening and territory. Little information is available about the size of the goby territory defined as the area from within which it expels other shrimp-associated gobies of the same and of other species. Moehring (1972) suggested that the goby size and sex affect territory size. Several distinct types of agonistic behaviour have been described for gobies which are competing for a burrow or during a conflict between neighbours. These behaviours include aggressive acts with physical contact (e.g., mouth fighting and biting) and without physical contact (e.g., lateral display, circling, tail beat) as well as submissive acts (e.g., head down). A detailed field study on aggressive interaction between gobies has been presented by Yanagisawa (1982, 1984) for A. japonica. Prolonged fighting and circling were performed mainly by males whereas such encounters between females were relatively rare and short. Dominance was apparently determined by

482

ILAN KARPLUS

body size. Size is not always decisive in competition for burrows as, in another species, Vanderhorstia delagoae in search of a burrow avoids burrows occupied by a smaller Cryptocentrus caeruleopunctatus (Magnus, 1967). Although threat displays and body contact are common in aggressive interaction between Amblyeleotris japonica, sometimes the subordinate fish surrender their burrows to the approaching dominant ones, even without exhibiting any defensive behaviour. The size differences between the opponents in these cases probably are very large (Parker, 1974; Maynard-Smith & Parker, 1976). The reverse situation is encountered in disputes between fish of similar size, which can be severe and often terminate in the winning of the resident. A detailed laboratory study of aggressive interactions, using sequence and information analysis, was carried out on Psilogobius mainlandi (Moehring, 1972). Staged encounters between pairs of gobies were analysed from recordings on an hour-long video tape. The effect of size of interacting gobies suggested that large gobies have larger territories than small gobies, while females have larger territories than males. Agonistic behaviour occurs more frequently between large gobies and less frequently between small ones. Contrasting with Amblyeleotris japonica, interacting females perform more aggressive acts than the other two sex combinations. Dominance positions of interacting gobies are more definitely and rapidly established in interactions between large and small gobies than between gobies of similar size. Only large females dominate males, while the males dominate the females of similar size. No interactions at all outside the burrow were observed between shrimps of adjacent burrows (Karplus, pers. obs.), Yanagisawa (1984) suggested that, while digging a burrow underground Alpheus bellulus can encounter other individuals of the same sex, and as they cannot tolerate each other, they fight, attempting to expel the other individual. That may be one of the reasons for the lack of either or both chelae occasionally observed in shrimps in the sea. The agonistic interactions between shrimps were investigated in a single laboratory study, involving two species, Alpheus rapax and A. rapacida (Moehring, 1972). Staged encounters between pairs were studied from recordings on an hour-long video tape using sequence and information analysis. Several aggressive (e.g., lunge, snap, chelae spread) and submissive acts (e.g., avoid) were described. The size, sex, and species of shrimp were found to affect the frequency of aggressive interactions, the establishment of dominance and the efficiency of information transmission. REPRODUCTION OF GOBY AND SHRIMP The reproductive behaviour of shrimp-associated gobies was completely unknown up to the last few years. Palmer (1963) stated that it was not known whether the gobies utilized the shrimps’ burrows for spawning. Magnus (1967) suggested that the gobies may use burrows uninhabited by shrimps for spawning because the burrowing activity by shrimps would prevent adequate development of the goby eggs. In a recent detailed field study, many aspects of the reproductive behaviour of a shrimp-associated goby Amblyeleotris japonica were finally revealed (Yanagisawa, 1982). Males of this species were usually ready to form pairs from May until September. During this period, males moved cautiously over the substratum, rarely venturing more than 3 m from their burrows. Females were rather passive and sometimes refused to pair themselves, indicating this by nudging the male’s belly. Paired males were sometimes attacked and replaced by single males. The competition of males in this species may be accounted for by the fact that only a small percentage of females were gravid at one time, while most adult males were apparently sexually active throughout the breeding season. At most, only about 7% of all the associations were paired during July and August. All males that were presumed to have successfully fertilized eggs, were those that were larger and socially dominant. Most males, however, paired only once in several weeks, always with

GOBIID FISHES AND ALPHEID SHRIMPS

483

one female at a time, suggesting that there is no monopolization of reproduction. Established pairs were maintained for several days and no aggressive interactions were observed between the mates. Males spent more time than females inside the burrow, and retreated before their mates in case of danger, staying afterwards much longer inside than the females. These differences may be related to the stronger attachment of the male to the burrow. A 77-mm female was observed to lay an egg mass containing about 20000 ellipsoid eggs 1·1 mm long. After spawning, the females leave the burrow or position themselves at its entrance, while the males spend from four to seven days inside the burrow taking care of the eggs. Among the large individuals of several species of gobies, pairs were recorded during several months. Vanderhorstia delagoae and Amblyeleotris steinitzi paired in the Red Sea from April until November, and Ctenogobiops maculosus and Eilatia latrucularia from April until December (Karplus, pers. obs.). In Japan Vanderhorstia mertensi and Amblyeleotris japonica were reported to pair from May until September (Yanigisawa, 1982). Other species, in the same and in different localities, paired throughout the entire year: in Japan Tomiyamichthys oni (Yanagisawa, 1982); in Hawaii, Psilogobius mainlandi (Preston, 1978); and in the Great Barrier Reef, Cryptocentrus fasciatus, C. cinctus, Amblyeleotris steinitzi, and Ctenogobiops pomastictus (Cummins, 1979). The formation of pairs and their stability in the goby-associated shrimps was completely unknown due to their spending a large part of their lives underground and the difficulty in collecting and tagging them. Yanagisawa (1984) has overcome some of these difficulties in his study on the reproductive behaviour of Alpheus bellulus associated with Amblyeleotris japonica, by identifying individual prawns, by size, coloration, and the side of the large chela. Pairs of shrimps are always heterosexual. The size of the mates is positively correlated, although in adult pairs the female is slightly larger than the male. The proportion of ovigerous females was highest from mid-July to mid-August although they were sighted early in July and as late as December. The number of eggs carried by a female was positively correlated with her size and its maximal number was close to 4500. Juvenile shrimp settled on the substratum from late July to early October. The shrimps mature and participate in reproduction within one year of their settlement. At the start of their benthic lives, they are single but, with growth gradually form pairs, 50% pairing four to six months after settlement whereas the adults are mostly paired. Pairs are probably not formed on the surface, as shrimps have never been witnessed to venture far enough from their burrow entrances to reach the adjacent entrances in daytime, and at night the entrances are all plugged with sand. Yanagisawa (1984) suggested that shrimps could obtain their mates underground. Although the distance between two adjacent burrow entrances usually exceeds 50 cm, the subterranean extensions of a burrow are wide enough to come close to the adjacent burrows. A single shrimp can establish a pair when the resident of an adjacent burrow is a single one of the opposite sex. The adult shrimp’s pair-bond is usually stable; some pairs were observed for more than several months, within a maximum range of two square metres (Fig. 10). Yanagisawa (1982) suggested that the per-manency of the pair-bond and the timing of its formation depends on the degree to which movement between units of habitat is difficult and on the availability of mates. In the absence of clues pointing to the number of potential mates, as in the case of Alpheus bellulus and the difficulty of acquiring a mate subterraneously, the shrimp’s preferred method is to establish a pair-bond with a mate, even at a very early age and to maintain it for a long time. In a number of obligatory goby-shrimp associations, cases have been reported of shrimp living in heterosexual pairs; Alpheus sp. associated with Cryptocentrus caeruleopunctatus and Vanderhorstia delagoae (Magnus, 1967); Alpheus purpurilenticularis with Amblyeleotris steinitzi (Karplus, 1979), and Alpheus rapacida with Psilogobius mainlandi (Preston, 1978). The pair formation of Alpheus rapacida in large finger bowls has been studied in the laboratory (Moehring, 1972). Only shrimps of opposite sexes paired. Females communicate to males more information

484

ILAN KARPLUS

Fig. 10.—Map of the entrances of the burrows out of which a pair of adult Alpheus bellulus were observed to come: numbers indicate the month observed; a patch of entrances enclosed with a line indicate a stable association whose members did not change during the observation period; R, large chela on the right side; L, large chela on the left side; from Yanagisawa (1984).

per act and with greater efficiency than males to females. Pair formation in this species probably also takes place underground, as suggested for A. bellulus, since it was never observed to leave the burrow vicinity. A. rapax living in a more facultative type of association with Psilogibius mainlandi travel by themselves over the substratum on very hot days when the tide is low. They live in burrows in groups of three, two females and one male, or in pairs consisting of two females or of a male and a female (Preston, 1978). A stable heterosexual pair of this species may not be essential, as mates can be more easily obtained by moving to an adjacent burrow (Yanagisawa, 1984).

GOBIID FISHES AND ALPHEID SHRIMPS

485

POPULATION STRUCTURE AND DYNAMICS The study of the population structure of goby-shrimp associations and its seasonal fluctuations is complex since it involves both recruitment and pair formation of two different organisms, as well as the formation of new associations, and the changes in established ones. No data are available from field studies on the sex-ratio of shrimp-associated gobies as these fish usually lack a conspicuous sexual dimorphism. The only shrimp-associated goby known to possess marked sexual dimorphism is Cryptocentrus caeruleopunctatus. In this species, the males are larger than the females, their fins are longer and they possess a conspicuous dark caudal fin, while the females’ caudal fin is greenishgrey (Klausewitz, 1960). In a monospecific aggregation of this species in the northern Red Sea, numbering more than thirty adults the sex ratio did not deviate significantly from 1:1 (Karplus, pers. obs.). The population structure and dynamics of Alpheus bellulus and Amblyeleotris japonica were studied in southern Japan (Yanagisawa, 1982, 1984). In this area, the climate is subtropical with a rather severe winter, and consequently seasonality exists in breeding and growth. The growth pattern of Alpheus bellulus was estimated by measuring animals collected monthly. The shrimps mature and participate in reproduction within a year after settlement. Based on the largest size obtained and their growth pattern, it is suggested that the adult population is composed of one-year and two-year groups. The growth pattern of Amblyeleotris japonica was also estimated from monthly collections (Yanagisawa, 1982). Within one year of settlement, the fish participated in reproduction, and the adult population was composed of one- and two-year old individuals. The number of juvenile A. japonica associated with shrimps, compared with that of adults, changed with time. Immediately following settlement, during September and October, juvenile gobies were several times more numerous than adults. Their number decreased by about 60% during the first three months following settlement, partially due to the shrimps’ pair formation. The entire population of settled fish decreased by about 80% in a single year. Juvenile Amblyeleotris japonica settled where adult gobies were present but also as in peripheral areas where no adults were seen (Yanagisawa, 1982, 1984). Similarly, juvenile Cryptocentrus lutheri and several other species of this genus were reported as settling in areas already occupied by adults, as well as in shallow areas largely covered by stones and not inhabited by adults (Zander, 1967; Karplus, Szlep &Tsurnamal, 1981). Synchronization of the breeding seasons of Amblyeleotris japonica and Alpheus bellulus increases the likelihood of co-occurrence of non-paired juveniles of both species. The establishment of the association, as early as possible after settlement, must be essential to avoid predation: e.g., the case of a prawn with a carapace length of 1·7 mm reported associated with an Amblyeleotris japonica of 8·7 mm standard length. Yanagisawa (1982) suggested that since juvenile shrimps, whose burrows have not been occupied by fish were detected, it can be assumed that a post-larval shrimp starts digging a small burrow as soon as it settles on the bottom. A non-associated goby, exploring the bottom, may encounter such a burrow and form a partnership with its occupant. Despite the synchronization of breeding seasons, disproportional settlement of gobies and shrimps in any one area at a given time will inevitably occur, resulting in considerable mortality of the surplus animals (Yanagisawa, 1984). The population structure of goby-shrimp associations was relatively stable at One Tree Reef, Great Barrier Reef (Cummins, 1979). The frequency of the pairings did not vary seasonally, neither did the ratio of juveniles to adults. This stability of the population structure in the relatively uniform tropical climate, contrasts with its seasonal fluctuations in the subtropical climate of southern Japan (Yanagisawa, 1982, 1984). A long term study of the stability of individual partnerships was carried out at One Tree Reef (Cummins, 1979). In this analysis a large number of burrow entrances were mapped and the details of individuals within each partnership (i.e. size, species, paired or single) were recorded at a mean interval of 3·5 months.

486

ILAN KARPLUS

A large percentage of the shrimps whose individual histories were traced, were found in the second recording as being associated with a different individual of the same or of a different species of goby (Fig. 11). As these gobies were usually fully, or almost fully, grown, it can be assumed that many of them do not have life-long associations with a particular shrimp. Several mechanisms causing the turnover of gobies in individual burrows were suggested: displacement of original gobies by bigger or other species, leaving a burrow voluntarily to find a mate or a preferred shrimp partner, and death through disease or predation. Approximately half of the shrimps, which could not be traced from the first recording, were probably recruited juveniles, associated with juvenile gobies. The other half were adult shrimps, probably unrecorded in the first census due to having been subterranean at that time. It was later found that some of these shrimps had actually been recorded previously. Experimental removal of gobies from marked burrows was also carried out at One Tree Reef (Cummins, 1979). Of the 14 studied burrows, 12 were recolonized by adult gobies within a mean of two weeks, by either the species of goby preferred by each type of shrimp or by one of the generalist species. A nonpreferred species of goby recolonized a burrow for several days only, thereafter leaving it again. This experiment demonstrated the recolonization of burrows, the sequential changes of fish partners of individual shrimps, and the ability of shrimps to survive at least several weeks without gobies. At any one census, a number of burrows having no goby occupant are consequently not recorded. Magnus (1967) has suggested that there is a vast population of subterranean shrimp which is not active outside the burrows due to the lack of fish partners. The renewal of the shrimp’s activity outside the burrow, after making contact with a goby, has been documented (Magnus, 1967; Karplus, 1981; Yanagisawa, 1984), but the proportion of the subterranean prawns in the entire population is unknown. In addition to fluctuations in the population structure, occasional disasters like a typhoon (Yanagisawa, 1982) or very strong winter storms (Karplus, pers. obs.) may completely destroy entire populations. These catastrophes probably occur due to the removal of the bottom sediment inhabited by the gobies and shrimps. THE COMMUNICATION BETWEEN GOBY AND SHRIMP COMMUNICATION UNDER NATURAL CONDITIONS Most warning communication systems are acoustic or chemical and only a minority are visual or tactile (Marler, 1968; Wilson, 1975). While chemical and acoustic systems are effective both day and night, visual systems can only operate during the day and under conditions of good visibility. Tactile communication systems are even more restrictive, as they require the proximity of the communicating individuals. The goby and shrimp fulfill this condition completely. Several species of symbiotic shrimps were reported as maintaining a constant antennal contact with their fish partners when outside their burrows (Magnus, 1967; Preston, 1978; Karplus, 1979; Karplus, Szlep & Tsurnamal, 1972a; Yanagisawa, 1984). Experiments in aquaria with Alpheus djiboutensis with one partially ablated antenna, indicated that without this contact, the shrimp did not respond to the retreat of its partner Cryptocentrus cryptocentrus which would normally result in its rapid withdrawal. Constant antennal contact between goby and shrimp is thus essential to transfer information (Karplus et al., 1972a). Specialized warning signals made by gobies are rapid tail flicks, often resulting in the shrimp’s retreat into its burrow. These signals have been observed both in aquaria (Harada, 1969) and in the sea, in response to an approaching diver (Magnus, 1967; Preston, 1978). In order to understand their function the generation of these signals by Amblyeleotris steinitzi was studied in the northern Red Sea the observer hiding behind a

GOBIID FISHES AND ALPHEID SHRIMPS

487

Fig. 11.—The fate of marked associations expressed in percentages following a 3·5-month interval at One Tree Reef, Great Barrier Reef (from Cummins, 1979).

fibreglass shield (Karplus, 1979). This goby produces warning signals at the rate of 7·4 signals per hour in the late afternoon, while maintaining antennal contact with its partner, Alpheus purpurilenticularis. Signals were produced in series (i.e. spaced less than 5 s apart) their number varying from 1 to 9, with a mean of 1·7 signals per series. The warning signals are given selectively in response to the approach of certain species of fishes. The trail of the fish and its distance from the burrow entrance was estimated by laying concentric iron circles around the entrance. The size of an approaching fish and its feeding behaviour determine whether it will cause the emission by the goby of warning signals. All large fish (e.g. non-piscivorus fishes

488

ILAN KARPLUS

of the Scaridae and Labridae families) triggered the release of warning signals while no small fish had the same effect. The goby was particularly selective in its response to medium sized fishes. The majority of warning signals were triggered by approaching goatfishes. These medium sized fish that are not predators, threaten the goby or the shrimp as they can block access to the burrow entrances completely by stirring the sediment in their search for food. Medium sized piscivorous fish from different families (e.g., Parapercis hexophthalma, Pterois volitans) also triggered the release of warning signals. Medium sized fishes which were neither piscivorous nor sediment diggers (e.g., Chaetodon chrysurus, Acanthurus nigrofuscus) did not trigger the goby’s warning signals even when they moved very close to the burrow entrance. The efficiency of the goby-shrimp communication system can best be studied under natural conditions. The shrimp responded differently (i.e., either retreating or not retreating into the burrow) to a series of signals than to individual signals. The shrimp retreated into the burrow in response to approximately only 60% of the single warning signals but responded to approximately 90% of a series of the same signals. Certain series of signals which do not generate the shrimps’ retreat seem to have been made in situations of little danger, (e.g., an intruder already leaving the burrow vicinity). The emission of warning signals in the sea by Amblyeleotris japonica was reported from Southern Japan (Yanagisawa, 1984). It is hard to compare these results with data from the Red Sea (Karplus, 1979). The majority of the tail flick warning signals recorded in Japan were probably produced in response to interference by the observer. In the Red Sea, a shield was used by the observer; this could have been important since 90% of all signals were given when fishes approached the burrow. The strongest warning signal, produced by the goby and always resulting in the shrimp’s retreat into the burrow, consisted in its own rapid retreat, head first, into the burrow. This type of signal was described much earlier than the tail flick, in studies of a variety of goby-shrimp associations (Luther, 1958a; Smith, 1959; Herald, 1961). In the northern Red Sea, Amblyeleotris steinitzi retreated into its burrow at a rate of 0·3 time per hour always inducing the rapid retreat of the shrimp into the burrow (Karplus, 1979). The same fishes causing the release of warning signals, also induce, at closer range, the goby’s ‘head-first’ retreat. Similarly, Amblyeleotris japonica was reported from southern Japan to retreat into its burrow when approached by Therapon jarbua—a piscivorous fish—and to react rather indifferently to the approach of non-predatory fishes (Yanagisawa, 1984). Periods without antennal contact are very short but may occur in situations when the shrimp moves out of the burrow, in a shallow straight groove, towards its goby positioned at the groove’s end. Even without antennal contact, the shrimp is still under the goby’s protection, because in an emergency, the goby enters the burrow head first using the groove in its retreat (Karplus, pers. obs.). Some insight into the completely unknown subterranean behaviour of the goby and shrimp following this retreat of the goby may be gained by observing their behaviour in artificial burrows (Karplus et al., 1972a). Following its entry, head first, Cryptocentrus cryptocentrus moved rapidly towards the end of the burrow, turned around and slowly and cautiously moved out again. The deeper Alpheus djiboutensis had moved inside the burrow, at the moment of the goby’s retreat, the less the shrimp retreated. The goby-shrimp communication system is characterized by a high rate of warning signals emitted by the goby, and a low rate of its retreat into the burrow. Only when certain species of intruding fish cross a critical distance, and a high level of danger is thus reached, does the goby retreat. There is therefore, a zone within which the goby is aware of low danger, and transmits warning signals to the shrimp, without itself retreating into the burrow. For the alert goby, the disadvantage of being exposed to low danger, while staying outside the burrow is small and is compensated by the advantage of longer access to food. The shrimp, which has poor vision (Luther, 1958a; Magnus, 1967), and is completely dependent on the goby

GOBIID FISHES AND ALPHEID SHRIMPS

489

outside the burrow, has the advantage of being warned by tail flick signals in case of danger of low intensity, and of danger of high intensity by the goby’s retreat. A guarantee for safety signal has been described both in aquaria (Karplus et al., 1972a) and in the sea (Magnus, 1967; Yanagisawa, 1984), but has been less investigated. This signal consists in a slow undulation of large amplitude of the tail of the goby. This signal was emitted at the rate of 8·2 times per hour by Amblyeleotris japonica when touched by the antenna of Alpheus bellulus. This signal seems to have the function of eliciting the emergence of the shrimp, especially in cases of the shrimp remaining in the burrow for a long time. The goby only rarely exhibited this signal when the shrimp was engaged in constant activity in and out the burrow (Yanagisawa, 1984). WARNING SIGNAL GENERATION IN RESPONSE TO PREDATORS AND MODELS OF PREDATORS The ability of Amblyeleotris steinitzi to discriminate between predatory and non-predatory fish has been tested in a series of controlled field experiments which also took into account the level of activity of the shrimp (Karplus, 1979). A transparent box was used in this experiment to present different species of fish. A. steinitzi produced a larger number of warning signals, over 15 minutes, when exposed to Parapercis hexophthalma, a piscivore, than when exposed to Acanthurus nigrofuscus, a fish feeding mainly on algae, and when faced with the empty box (Fig. 12). No difference was found in the number of exits of the shrimps, thus the different signalling rates were probably the result of recognition of an enemy by the goby and not the different levels of activity of the shrimp. It is possible that the very young goby responds initially by emitting signals and retreating into its burrow at the approach of all types of fish. Only by a

490

ILAN KARPLUS

Fig. 12.—The number of warning signals of Amblyeleotris steinitzi given in response to Parapercis hexophthalma Acanthurus nigrofuscus and an empty box : results are means ± standard deviation; from Karplus (1979).

process of habituation does the goby cease to respond to common medium-sized, non-predatory and nondigging fishes (Karplus, 1979). The rate of signal emission by a goby in the presence of the same piscivore (i.e., Parapercis hexophthalma) is negatively correlated with the distance between goby and the piscivore, and the duration of its exposure and positively correlated with the level of movement of the piscivore. That same fish when presented to the goby in a movement restricting box, caused the release of fewer warning signals than when the fish was presented in a spacious box where it could perform frequent movements (Karplus, 1976, 1979). Various aspects of the reaction of Amblyeleotris steinitzi to a piscivore (i.e., Parapercis hexophthalma), other than the generation of warning signals, changes with the distance between the two. Exposure to the piscivore from a very short distance caused the immediate retreat, head first, of the goby. Close to the piscivore, the goby head colour became white, whereas at a greater distance from the same piscivore or following longer periods of exposure, its head colour was dark. The advantage of these colour changes is probably that, during the white phase, the goby is less conspicuous to predators as it blends well with the light-coloured sand. The conspicuous black colour phase is probably advantageous in intraspecific interactions (e.g., mate location). Because the white head coloration is positively correlated with the release of warning signals (Fig. 13), the tendency to escape is probably the motivation underlying the generation of warning signals. The release of warning signals by the goby Amblyeleotris steinitzi in response to the approach of twodimensional models of two different sizes of a piscivore of the Serranidae family has been investigated in the northern Red Sea (Karplus & Ben-Tuvia, 1979). The close approach of the model causes the retreat of the goby. The point reached by the model at the moment the goby enters its burrow was defined as the critical point. For both models, a negative correlation was established between the number of warning

GOBIID FISHES AND ALPHEID SHRIMPS

491

signals and the distance from the critical point. Thus, as in the previous study, a high level of escape tendency seems to underlie the generation of warning signals. A comparative study of the response of three shrimp-associated gobies to a small two-dimensional predator model has been carried out in the northern Red Sea (Karplus, 1976; Karplus & Ben-Tuvia, 1979). Amblyeleotris steinitzi associated with Alpheus purpurilenticularis and Ctenogobiops maculosus associated with Alpheus rapax had similar responses to the model, with respect to the large mean number of signals generated (20·8 and 14·9, respectively). Both also showed a negative correlation between the number of signals and the distance from the critical point. Both these species entered their burrows head first at similar critical distances (i.e. distance between the critical point and the burrow entrance. Cryptocentrus lutheri associated with Alpheus djiboutensis differed greatly from the previous two species by a lower mean number of warning signals (2·5) and a lack of relationship between the number of signals and the distance from the critical point. Nine out of ten of this species’ entrance into the burrows were of the tail-first type and their critical distance was approximately three times longer than in the other two species. The key to understanding the differences in signal generation of these three species lies in the way they enter their burrows. Head-first entries of fishes into the burrows or shelters appear to have been induced by stronger stimulation than tail-first entries (Magnus, 1967; Colin, 1971; Karplus et al., 1972a; Fishelson, 1975). The low frequency of signals produced by Cryptocentrus lutheri is probably the result of its habit to withdraw into its burrow, tail first, when the danger is still distant. Amblyeleotris steinitzi and Ctenogobiops maculosus remain outside their burrow even when the danger is close. They thus attain a high level of tendency to escape and generate numerous warning signals prior to entering the burrow, head first. SEQUENCE AND INFORMATION ANALYSIS The statistical analysis of the sequence of goby-shrimp acts provides an objective evaluation of bidirectional communication. This method was often applied to the study of both vertebrate and invertebrate communication (Hazlett & Bossert, 1965; Altmann, 1965; Dingle, 1969; Max-Westby, 1975) but has so far only been used twice in the study of goby-shrimp communication (Preston, 1978; Karplus, 1979). While the sequence of acts of a pair of courting or fighting individuals can be visualized as a closed system with only two actors, by definition, the warning systems are always open and include a third party—a source of disturbance. This inherent feature causes some difficulties both in obtaining sequences of goby-shrimp acts and in interpretating the results. The communication between Psilogobius mainlandi and two species of shrimp Alpheus rapax and A. rapacida has been studied in shallow areas of Kaneohe Bay, Oahu, Hawaii (Preston, 1978). Observation was carried out in shallow waters, at low tide by slowly approaching the association, the observer constituting the source of disturbance. Thirteen different acts of the goby and seven acts of the shrimp were defined and their communicatory value as inhibiting or directing a given response were analysed on the basis of two act contingency tables. Of special interest were the four warning signals of the goby which directed the retreat of the shrimp. From weakest to strongest, these signals were: “withdraw”, “tail flick”, “tail beat”, and “flee”. Both species of shrimps often responded to the “tail flick” by the “sit” act— remaining motionless at the burrow entrance, a method of concealment which requires less energy than the withdrawal. This type of response to a warning signal had not previously been described for any gobyassociated shrimp. Some differences in the response of the two species of shrimps to the same communicatory acts of the goby are probably related to the morphology of the shrimp. Because of differences in the length of the antennae of the two species of shrimps, two slightly different communication systems may have evolved. With its long antennae, A. rapax can dig at a relatively greater distance from a

492

ILAN KARPLUS

Fig. 13.—Head coloration and warning signals of Amblyeleotris steinitzi: WS, the tail-flick warning signal; W, white head; G. grey head; B, black head; from Karplus (1979).

goby and still maintain contact with it. Its antennae distinctly detect the difference between a generalized movement of a goby and a “tail flick”. On the other hand, with its relatively short antennae, A. rapacida has more body contact with the goby, and cannot usually distinguish between the generalized and specialized movements of the goby (Preston, 1978). Transfer of information, was calculated by Preston from the observed inter-phyletic two-act sequences, according to the methods used by Hazlett & Bossert (1965), Dingle (1969, 1972), and Steinberg & Conant (1974). More information was transmitted per act by A. rapax than by A. rapacida, but the latter produced more acts per encounter. The same amount of information was thus transmitted by the two species of shrimps to the goby. The fact that only one bit of information was transmitted per encounter supports the assumption that one ‘yes-no’ message was issued to the goby in each encounter. More transmission would probably be wasteful. Shrimp signals apparently informed gobies of whether or not an actively digging shrimp was present, that is, whether or not warning signals were necessary in the event of danger. Warning signals were unnecessary and therefore not made, when the shrimp was inside the burrow. The amount of

GOBIID FISHES AND ALPHEID SHRIMPS

493

information transmitted by the goby to the two species of shrimps was in both cases similar, and the goby act “tail flick” contributed the most to information transmission (Preston, 1978). The sequence of the acts of the goby Amblyeleotris steinitzi and the shrimp Alpheus purpurilenticularis have been recorded in a shallow lagoon in the northern Red Sea, when faced by a living predator— Parapercis hexo-phthalma—kept in a transparent box (Karplus, 1979). Twelve acts of the goby and five acts of the shrimp were defined. The main findings of this study were similar to those of Preston (1978). The communicatory acts of the goby— rapid body movements and special warning signals—directing the retreat of the shrimp are of interest. The first one occurs both in the absence and presence of the shrimp while the special warning signals—“tail flick” and “tail beat”—are only made in the shrimp’s presence. Some of the rapid body movements—“head-first entry”, “tail-first entry”, and “partial-tail entry”— are connected with situations of danger, while quick short movements either to collect sediment or to change position, are not. The shrimp probably responds to the last group because of the similarity of those signals to

494

ILAN KARPLUS

Fig. 14.—The head-first retreat of Cryptocentrus cryptocentrus into an artificial burrow: the black arrow indicates the area where pushing of the goby by Alpheus djiboutensis was performed; from Karplus et al. (1972a).

the initial part of the first ones. Also of interest is a rare act of the shrimp, its pinching of the goby. A similar behaviour was frequently observed in the subterranean parts of an artificial burrow. Alpheus djiboutensis pushed the goby Cryptocentrus cryptocentrus out in order to allow the shrimp’s activity outside the burrow (Fig. 14). This act was carried out using the first pair of chelae and was usually directed at the tail of the goby. In rare cases when the goby did not respond, the shrimp intensified the pushing and snapped once or twice with its big chela, thus adding the acoustic channel to the goby-shrimp communication system (Karplus et al., 1972a). Pinching outside the burrow appears to represent an extension of subterranean pushing. The validity of chi square tests, as used in the analysis of the goby-shrimp contingency tables to measure communication and the application of information theory may be questioned. Each dyade in the table should have been independent, while actually a large variable number of dyades are contributed by the same individuals (Moehring, 1972). The contingency table should be analysed as one entity thus the separate

GOBIID FISHES AND ALPHEID SHRIMPS

495

analysis of rows is probably not valid statistically. The assumption that these acts are similar is incorrect since at least the “tail flick” of Amblyeleotris steinitzi was demonstrated to be a graded signal (Karplus et al., 1979b). Changes in the goby’s response to the source of disturbance with distance and time affects the goby-shrimp interactions (Karplus, 1979). This contradicts assumption of stability when sequence and information analysis is made. The study of goby-shrimp communication using sequence analysis did not take into consideration the duration of acts which may have been crucial to their outcome (Hazlett & Bossert, 1965). Even with these violations and limitations part of which cannot be overcome, the quantitative methods applied proved to be an important tool in the objective analysis of communication systems. FILM ANALYSIS Film analysis is necessary for understanding the goby-shrimp communication, since only through it can the details of rapid and complex interactions be clarified. Regretfully, this method has only been applied once in order to study the interactions between A. steinitzi and Alpheus purpurilenticularis in a shallow lagoon in the northern Red Sea (Karplus et al., 1979). No cross-species comparisons were therefore made. Five different measurements of the tail-flick warning signal of this goby—amplitude, speed, number of components, length (i.e. total distance traversed by the caudal fin—TDT), and duration—were analysed. The conspicuous nature of this signal, compared with all other tail movements, provides it with coding characteristics (Wiley, 1973). These features increase the efficiency of the signal without making too much use of redundancy, which is harmful in any alarm system. The tail-flick warning signal belongs also to the category of graded signals which are characterized by different levels of intensity. They convey more information, than the discrete signals which are generated either in a simple on-or-off manner or display a typical intensity as a result of a wide range of stimulations (Morris, 1957; Wilson, 1975). Filming also provided the means of analysing the exact variable response of a shrimp to a warning signal (Fig. 15). The combined principal component factor analysis with a stepwise multiple regression analysis was made to define which of 18 independent variables (e.g., shrimp chelae either empty or loaded with sediment, the speed of the shrimp, and its direction) are important in determining the shrimp’s response. The structure of the preceding and actual warning signals and the area of contact of the antenna accounted for approximately a third of the differences in the shrimp’s retreat response. The most important signal variable was its length (TDT). The neurophysiologiocal mechanism, underlying the action of the mechanoreceptors on the shrimp’s antenna responding to the minute vibrations of the goby’s fins, are yet unknown but their study would be interesting. The feedback mechanism by which the goby regulates the intensity of the warning signal, according to the shrimp’s response is one of the most complex aspects of this communication system. Upon lack of response from the shrimp, the fish increases the intensity of its signal (measured by TDT), while upon the rapid retreat of the shrimp, the fish decreases the signal intensity. The goby’s regulation of the intensity of the warning signal according to the shrimp’s response is possible, since the mean time interval between two consecutive signals of a series (1·6±1·1 s) is significantly longer than the mean latency of the shrimp’s response (0·5±0·4 s). The antennal contact between shrimp and goby (Figs 16 and 17) described in field (Magnus, 1967; Karplus, 1979) and laboratory studies (Harada, 1969; Karplus, Szlep & Tsurnamal, 1972a) has been further investigated by filming. The areas of contact between shrimp and goby changed according to the position of the shrimp, relative to that of the goby. When the shrimp was behind the goby, both its antennae pointed forward and touched the goby’s caudal fin with one of them. When the shrimp was parallel to the goby, one

496

ILAN KARPLUS

antenna was bent sideways touching the second dorsal fin, while the other still pointed forward. When the shrimp was further away from the burrow than the goby, one antenna pointed backwards, touching the goby’s pectoral fin while the other was still directed forward. Warning signals are given with various fins including the caudal, second dorsal, anal and pectoral fins according to the area of antennal contact. In each case, however, the caudal fin is involved in signalling (Fig. 18). The possibility of signalling with several fins enables the shrimp to move further away from the goby while still being protected through the warning system. The goby’s head-first retreat into its burrow is so rapid that without film analysis only a cloud of sand can be perceived during its occurrence. The goby’s retreat is made in two stages. In the first short stage (0·16 s), the goby turns towards the burrow opening in a loop, its tail and head almost meeting. In the second, longer and less uniform stage (0·52±0·19 s), the goby enters head first into the burrow. This retreat causes the shrimp to withdraw very rapidly (93·7±54·9 mm/s) at a latency of only 0·16 s. The mean maximal speed of entering a burrow, in response to a series of warning signals, was only 16·5±7·2 mm/s with a latency of 0·51 ±0·36 s. The head-first retreat of the goby thus constitutes the strongest warning signal not only because it always induces the shrimp’s retreat but also due to the extreme retreat speed of the shrimp and the shrimp’s very short latency to respond. Because the goby’s head-first retreat takes place under circumstances of extreme danger as during direct attack by a predator, any delay in the shrimp’s response or any weak response on its part will cause it to be devoured and led to a direct negative selection of shrimps exhibiting this behaviour. PARTNER SPECIFICITY OF GOBY-SHRIMP ASSOCIATIONS FIELD OBSERVATIONS The joint occurrence of several different species of alpheid shrimp and gobiid fishes in the same area, poses several questions with respect to the degree of partner specificity, its regulating mechanism, and function. Specificity was initially studied by examining the occurrence or non-occur-rence of certain species of gobies and shrimp in the same burrow. Harada (1971) concluded that, in southern Japan, goby-shrimp associations are non-specific since all possible combinations between four species of gobies and two species of shrimp were found. In a more recent publication, Yanagisawa (1978) stated that some specificity was found in the combinations between gobiid fishes and snapping shrimp. Little of that is known, however, because the majority of these shrimps in southern Japan were either not collected or not identified. Polunin & Lubbock (1977) examined the composition of 170 associations in the Seychelles. They concluded, based on the number of partners, that different species of gobies and shrimps differed with respect to partner specificity. The use of the occurrence or non-occurrence of certain species of gobies and shrimps in a burrow, as the sole criterion for specificity, may be misleading, because the composition of the associations may result from a random independent distribution over burrows of both species. Partner specificity of six species of goby and four types of shrimp has been investigated in the sandy lagoon of One Tree Reef, Great Barrier Reef (Cummins, 1979). Chi square analysis was used to determine whether the distribution of goby species across types of shrimps differed significantly from a random one. Preference for a certain partner was also studied by a partial correlation of the number of gobies of each species and the number of shrimps of each type within local populations. Three species of gobies were associated both as juveniles and adults with a single “preferred” type of shrimp while, in one species of Amblyeleotris, only the adults were partner specific. One species of goby, was equally associated with two types of shrimps, and another species was associated at random with all the types of shrimps.

GOBIID FISHES AND ALPHEID SHRIMPS

497

Fig. 15.—Graphical presentation of the interaction between the tail-flick warning signal of Amblyeotris stenitzi and the response of Alpheus purpurilenticularis: a, the movement of the caudal fin of the goby, R, right and L, left; b, the movement of the Y axis (×2); c, the movement of the shrimp on the X axis (×2): from Karplus et al. (1979).

Partner specificity of goby-shrimp associations was further investigated in the northern Red Sea in a combined field and laboratory study (Karplus, Szlep & Tsurnamal, 1974, 1981; Karplus, 1981). The composition of over 750 associations was analysed following Pielou’s (1969) method for analysing associations of pairs of species found in discrete units. The following three questions were posed.

498

ILAN KARPLUS

Fig. 16.—Alpheus pupurilenticularis maintaining antennal contact with Amblyeleotris steinitzi (from Karplus, 1979)

(1) Does a species of goby occur together with a species of shrimp in the same burrow? (2) Is the number of co-occurrences significantly different (more or less) than would be expected from a random distribution of both species over burrows? (3) What is the strength of the association between two species, measured using Pielou’s (1969) correlation coefficient? Partner specificity differed greatly in deep and shallow waters. In deep water, no evidence of partner specificity was found whereas, in shallow water, three different types of specificity occurred, as follows. Type I: co-occurrences of goby and shrimp in the same burrow, with a positive correlation coefficient. Type II: co-occurrences of goby and shrimp in the same burrow, with a negative correlation coefficient. Type III: no co-occurrence of goby and shrimp in the same burrow, correlation coefficient negative. In shallow water, each of four different shrimp species occurring in a different microhabitat formed a very strong association (highly positive correlation coefficient) with a single species of goby of a different genus. Each of these shrimp species had a negative correlation coefficient with all the other gobies (Fig. 19).

GOBIID FISHES AND ALPHEID SHRIMPS

499

Fig. 17.—Alapheus randalli maintaining antennal contact with stonogobiops nematodes (from Hoese &Randall, 1982).

Species of gobies which had a wide depth range were associated with different shrimp species in shallow and deep waters (Fig. 20). The same species formed random associations with deep-water shrimps and highly specific associations with shallow-water shrimps (Karplus, 1976). The segregation of shrimp to different habitats, usually inhabited by several shrimps of the same species, increased the probability of a goby moving away from its partner to re-enter the burrow of a shrimp of the same species. The nonoccurrence of a certain species of shrimp and goby in the same burrow at Marsah-Murach, despite the spatial distribution facilitating it, suggested the involvement of a behavioural mechanism in the regulation of partner specificity (Karplus et al., 1981). LABORATORY EXPERIMENTS Laboratory experiments on goby-shrimp partner specificity have been carried out on species found in the northern Red Sea (Karplus, 1981). Although only a few species were studied, they were selected to represent the three different types of specificity (Karplus et al., 1981). Species of gobies and shrimps of the first type of specificity were mutually attracted. The goby was visually attracted to the shrimp and the shrimp was chemically attracted to the goby (Karplus, Tsurnamal &

500

ILAN KARPLUS

Fig. 18.—Amblyeleotris steinitzi: fins taking part in signalling and the areas of antennal contact: left, the area of contact (stippled); right, the frequency (%) in which various fins took part in signalling; C, caudal fin; A, anal fin; D1, first dorsal fin; D2, second dorsal fin; P, pectoral fins; from Karplus et al. (1979).

Szlep, 1972b; Karplus, 1981). The visual attraction of Amblyeleotris steinitzi was tested in a white elongated Perspex box which had two identical, water-tight transparent cells at each end. The attraction of this goby to its preferred shrimp partner, Alpheus purpurilenticularis, was demonstrated by the longer time it spent near the cell containing that shrimp, instead of near the empty cell, the longer time its head touched that cell and the longer time it attempted to enter it. In another experiment, Amblyeleotris steinitzi preferred Alpheus purpurilenticularis to A. djiboutensis with the latter of which its specificity relationship was of the second type. The chemical attraction of A. purpurilenticularis was tested in a Y-maze. This shrimp entered the arm of the Y fed with water which had flowed over Amblyeleotris steinitzi, preferring it to the arm that had received plain sea water. In still another experiment, Alpheus purpurilenticularis preferred Amblyeleotris steinitzi to Cryptocentrus lutheri, with which its specificity relationship was of the third type. Amblyeleotris steinitzi was not attracted chemically to Alpheus purpurilenticularis, its preferred shrimp partner, and neither was this shrimp attracted visually to this goby. A similar result was obtained for Cryptocentrus cryptocentrus and Alpheus djiboutensis (Karplus et al., 1972b). The mutual attraction between goby and shrimp is thus based on different sensory modalities. Species of goby and shrimp of the second and third types of specificity were not attracted to each other. While species of the second type maintained antennal contact, those of the third type did not. In aquaria,

GOBIID FISHES AND ALPHEID SHRIMPS

501

Fig. 19.—A diagrammatic representation of the strength of the association between alpheid shrimp and gobiid fish in the Elat Nature Reserve: black line, dependent occurrence of goby and shrimp in the same burrows and a significant positive correlation coefficient; white line, dependent occurrence of goby and shrimp in the same burrows and a significant negative correlation coefficient; broken line, independent occurrence of goby and shrimp in the same burrows; from Karplus et al. (1981).

several Cryptocentrus lutheri inhabited burrows excavated under a stone by several Alpheus purpurilenticularis. Over a period of 30 days, these shrimps avoided any antennal contact with the gobies which were perched at the burrow entrance. After the removal of the Cryptocentrus lutheri and the introduction of several Amblyeleotris steinitzi, antennal contact was immediately established, and the shrimp reacted to warning signals generated by these gobies (Karplus, 1981). THE MECHANISM REGULATING SPECIFICITY The major behavioural processes regulating specificity are the attraction of the goby to the shrimp’s burrow, the attraction between the partners, and the adoption of a tactile alarm system. In a sandy habitat which lacks hiding places, the shelter provided by a shrimp’s burrow plays an important rôle in the formation and maintenance of the association. Gobies isolated from their own burrows rapidly took shelter in other burrows (Karplus et al., 1974). The specific structure of the burrow openings (Karplus et al., 1974) and their size (Cummins, 1979) also possibly plays a rôle in partner selection. Part of the specificity of goby-shrimp association at One Tree Reef was controlled, according to Cummins (1979), by the size of the partners. Species of large gobies were associated with larger types of shrimp that also constructed larger burrows. Large individuals of one of the goby species, which as adults associated only with large types of shrimp, had associated, as juveniles, with all four types of shrimp, including two small types. A somewhat similar phenomenon was observed in Marsah-Murach. The very large Cryptocentrus caeruleopunctatus males only associated with Alpheus djiboutensis which attained the largest size and made

502

ILAN KARPLUS

Fig. 20.—Changes with depth in the composition of shrimp partners of two gobiid fishes in the northern Red Sea: A, Amblyeleotris steinitzi; B, Ctenogobiops maculosus; from Karplus (1976).

the largest burrows of all goby-associated shrimps in that bay. The smaller juveniles and females of this species also associated with other smaller shrimp species (Karplus, pers. obs.). The strong negative phototactic response of the goby and shrimp (Karplus et al., 1972b) may also have facilitated the maintenance of these associations. The following behavioural interactions were suggested as regulating the different types of partner specificity (Karplus, 1981). First type. In these partnerships the goby is attracted to the shrimp’s burrow, the shrimp and goby are mutually attracted and maintain antennal contacts. Examples include the association between Amblyeleotris steinitzi and Alpheus purpurilenticularis (Karplus, 1981) and between Cryptocentrus cryptocentrus and Alpheus djiboutensis (Karplus et al., 1972b). These associations are common and stable.

GOBIID FISHES AND ALPHEID SHRIMPS

503

Second type. In these partnerships, the goby is attracted to the shrimp’s burrow but the shrimp and goby were not mutually attracted although antennal contacts were maintained. For example, the association between Amblyeleotris steinitzi and Alpheus djiboutensis and between this same goby and A. rapax can be cited. These associations were rare and unstable. Amblyeleotris steinitzi and Alpheus djiboutensis were observed only once in the same burrow and for only two weeks. After this time, the goby left A. djiboutensis and moved to its preferred partner, A. purpurilenticularis. Also Amblyeleotris steinitzi and Alpheus rapax did not remain in the same burrow for more than a week. The shrimp in these partnerships did not, however, avoid antennal contact and was outside the burrow, with the goby present at the entrance. Third type. The goby in this case was attracted to the shrimp’s burrow but the goby and shrimp were not mutually attracted and did not maintain antennal contact. One example was Cryptocentrus lutheri and Alpheus purpurilenticularis. It was not possible during an entire month, to form a real association between these species in the laboratory. The goby was attracted to the burrow but the shrimp avoided any antennal contact. Observations in aquaria with artificial burrows showed that A. djiboutensis blocked the anterior part of its burrow after the removal of its partner, Cryptocentrus cryptocentrus. After the re-introduction of the goby, the opening was rapidly cleared (Karplus et al., 1972a). In the field, Ctenogobiops maculosus was observed to insert its tail into a partly blocked opening of Alpheus rapax. As soon as the shrimp touched the tail of the goby with its antennae, the opening was cleared and the shrimp was again active outside the burrow. Similarly, in the field, burrows of A. purpurilenticularis, whose partner was removed, were blocked within a short time. These openings probably remained closed when approached by gobies such as Cryptocentrus lutheri, with which Alpheus purpurilenticularis avoids antennal contact. Habitat selection of the goby cannot be a major mechanism regulating specificity, because it can only affect the initial stage of the contact between goby and shrimp. No partnership can be formed between certain species of goby and shrimp even if the goby seeks shelter in the shrimp’s burrow because the shrimp will avoid any antennal contact with these species and will not leave its burrow nor clear its entrance while the goby is positioned at the opening. The segregation of shrimps in different habitats was probably important in the evolution of partner specificity. Species of shrimp with no clear habitat segregation show a lack of partner specificity. In the initial stages of the evolution of these associations, while they were still facultative, the distribution of the shrimps probably determined to a great extent the composition of the associations. Only later, with the gradual evolution of an obligatory and mutualistic relationship, did the behavioural interactions between goby and shrimp become more important and a behavioural mechanism regulating partner specificity evolved. The specificity of goby-shrimp associations is probably of importance at the level of the species but not of the individual. The spontaneous and rapid exchange of burrows and of shrimps between Cryptocentus caeruleopunctatus of different sizes (Karplus et al., 1974), the transitional occupation of several burrows by Amblyeleotris japonica (Yanagisawa, 1982) and Vanderhorstia delagoae (Magnus, 1967), as well as the immediate formation of associations between isolated partners in aquaria (Karplus et al., 1972a; Karplus, 1981) support this assumption. Polunin & Lubbock (1977) suggested that a species’ specific communication system between goby and shrimp would enhance the efficiency of transmission but reduce the number of available hosts. A nonspecific communication system could have a lower transmission efficiency but a higher number of potential partners. The testing of this hypothesis would be most interesting as it could provide us with some understanding of the function of partner specificity. Although species’ specific differences in communication systems of several Red Sea associations were found (Karplus, 1976), the present state of knowledge on goby-shrimp communication systems does not allow us to test this hypothesis.

504

ILAN KARPLUS

THE EVOLUTION OF GOBY-SHRIMP ASSOCIATIONS The reconstruction of the evolution of a complex behavioural relationship is always complicated and speculative, but it can be aided by applying the comparative method. Comparisons between different populations of the same species, obligatory species with loose or tight relationships, and obligatory compared with facultative species may assist the reconstruction of the evolution of goby-shrimp associations. The attachment between goby and shrimp, and the local conditions were correlated as relating to Amblyeleotris japonica in two localities in southern Japan. In one locality, with a high predator pressure and a low burrow density, the gobies spent more time in close proximity to the entrances, and seldom left the burrows. In this population, fewer floaters were found, and a single or pair of gobies occupied fewer burrows than in a population of low predator pressure and high burrow density (Yanagisawa, 1982). Different species of shrimp-associated gobies differed in the degree of their attachment to their shrimp hosts (Magnus, 1967; Polunin & Lubbock, 1977). The goby Vanderhorstia ornatissima was often found in the Seychelles, far from its shrimp host, taking shelter in case of emergency in burrows of other gobies or in burrows of callianassid prawns. The loose relationship of this goby with its shrimp may be partly a result of its usual habitat, in which some protection from predators is provided by the abundant sea grasses. A similar loose relationship with its shrimp partner was described for another goby of this genus, V. delagoe, in the Red Sea (Magnus, 1967). Probably the highly cryptic colour and pattern of this fish facilitates this type of relationship. Few species of gobies form facultative relationships with shrimp. Vireosa hanae occasionally hovers at about half a metre above the burrow entrance of Alpheus bellulus, not maintaining contact with the shrimp. When Vireosa hanae retreated into the burrow, it caused the retreat of the shrimp and of Amblyeleotris japonica which often shared the same burrow (Harada, 1969; Yanagisawa, 1978). Acentrogobius pflaumi, a non-hovering goby does form facultative associations with Alpheus brevicristatus. Antennal contact was maintained by the shrimp with this goby, that often fled when approached by a diver instead of retreating into the burrow (Harada, 1969; Yanagisawa, 1978). The behaviour of an obligatory fish partner Nes longus and a facultative one, Bathygobius curacao, both associated with a facultative shrimp partner, Alpheus floridanus, has been studied at Key Biscayne, South Florida (Karplus, unpubl. data). Continuous antennal contact was maintained between the two gobies and the shrimp while it was outside the burrow. The shrimp retreated into the burrow in response to tail-flick warning signals and head-first entry of Nes longus. Bathygobius curacao warned the shrimp only by headfirst entry into the burrow, and gave no warning signals. Nes longus was continuously positioned in front of the burrow, in the shallow groove excavated by the shrimp, its tail directed to the entrance. Bathygobius curacao occasionally left the burrow entrance and moved in its vicinity, causing the shrimp to retreat when it emerged from the burrow during the goby’s absence. The goby was often positioned very close to the burrow entrance but outside the groove. The emerging shrimp was, therefore, unaware of the goby’s presence, failed to establish antennal contact and retreated. The time a shrimp spent outside its burrow differed when it was alone or when it was associated with a facultative or obligatory fish partner. Alpheus floridanus without a goby partner spent very short periods of time outside the burrow, usually only dropping the sediment very close to the entrance. When associated with Bathygobius curacao, it spent about 10% of the time outside the burrow and about 30% when associated with Nes longus. The complex obligatory associations between non-burrowing gobies and burrowing alpheid shrimps probably evolved through loose facultative partnerships. The non-associated shrimp probably spent most of its time in the subterranean burrow, occasionally dumping sediment at the entrance. Non-associated gobiid fishes under intense predatory pressure in the sandy habitat, were probably protected mainly by their cryptic

GOBIID FISHES AND ALPHEID SHRIMPS

505

coloration and a few natural shelters. Some gobies started to take occasional shelter in burrows constructed by different groups of crustaceans, polychaetes, and echiuroids (MacGinitie, 1939; Luling, 1959; MacGinitie & MacGinitie, 1968; Schembri & Jaccarini, 1978). The habit of sheltering in burrows of alpheid shrimps was the most successful, since it evolved into a highly mutually beneficial partnership. The initial phases of this process were probably similar to the loose relationship existing between Vireosa hanae and Alpheus bellulus. This hovering goby occasionally takes shelter in the shrimp’s burrow thereby warning it in case of danger, thus forming a mutually beneficial relationship. Antennal contacts were gradually established with species living on the substratum near the burrow entrance. The poor vision of the subterranean shrimp (Luther, 1958a; Magnus, 1967) probably determined the evolution of a tactile and not a visual communication system. The initial purpose of the antennal contact was probably only to inform the shrimp of the goby’s presence at the entrance, the goby’s retreat into the burrow probably being the only warning signal. Such a relationship has been described for Bathygobius curacao associated with Alpheus floridanus. Magnus (1967) and Preston (1978) have both suggested that the tail-flick warning signal evolved from intentional movements of the goby, connected with its retreat to the burrow. There are numerous examples of communicatory movements which have evolved through a mono-valent ritualization (Daanje, 1950; Andrew, 1956; Hjorth, 1966). It has been suggested that it was the escape tendency which motivated the generation of warning signals. An intermediate phase in the formation of the highly ritualized tail flick is probably the tail beat. This is still the dominant signal of Psilogobius mainlandi, and is emitted in cases of higher danger than the tail flick (Preston, 1978). The tail beat of Amblyeleotris steinitzi is only given on rare occasions, during massive body contact with the shrimp. Its dominant warning signal is the tail flick, a very effective graded and coded signal of short duration and small amplitude. In the course of the evolution of an obligatory association between gobies and shrimps, they become mutually attracted and influence one another in many ways, as their interrelationship becomes mutual and complex. The shrimp provides the goby with a burrow to use as shelter in an emergency during the day, and as a resting place at night, as well as a place to deposit and guard its eggs with safety. The goby provides the shrimp with a tactile alarm system, enabling it to clear the burrow safely and to collect food outside the burrow. Warning signals can be emitted by the goby with several fins, depending on the area of the shrimp’s antennal contact, their intensity being regulated by a feedback mechanism attuned to the shrimp’s response. The goby determines every day the beginning of the shrimp’s activity outside the burrow as well as the location of the burrow entrance by pushing its head through the substratum from within the burrow. Shrimp and goby also engage in a mutually beneficial cleaning relationship inside the burrow. In some associations the gobies were reported to feed outside the burrow on small crustaceans and polychaetes found in the sediment transported there by the shrimp. The synchronization of the breeding season of goby and shrimp facilitates the formation of the association from the very initial phases of their benthic life. The partnership between goby and shrimp provides both of them with advantages over their, free-living relatives affecting their speciation distribution and abundance. ACKNOWLEDGEMENTS I wish to thank Prof. D.F.Hoese for sending me his preliminary list of gobiid fishes associated with alpheid shrimps and for his critical comments, and Prof. L.Fishelson and Dr M.Goren of Tel Aviv University for reading and discussing the manuscript. Many thanks to Dr R.A.Cummins for sending me her Ph. D. thesis and permitting me to use her figure. I am also very grateful to Mrs Esther Sas for her superb editorial work and to Miss Lea Alfandari for most efficiently typing the manuscript.

506

ILAN KARPLUS

REFERENCES Abel, E.F., 1960. Z. Morph. Ökol. Tiers, 49, 430–503. Altmann, S.A., 1965. J. theor. Biol., 8, 490–522. Andrew, R.J., 1956. Behaviour, 10, 179–204. Banner, A.H. & Banner, D.M., 1980. Pacif. Sci., 34, 401–405. Banner, A.H. & Banner, D.M., 1983. Trav. Docum. Orstrum, No. 158, 164 pp. Banner, D.M. & Banner, A.H., 1981. Zool. Verh., Leiden, No. 190, 3–99. Banner, D.M. & Banner, A.H., 1982. Rec. Aust. Mus., 34(1), 1–357. Bayer, F.M. & Harry-Rofen, R.P., 1957. Rep. Smithson. Instn for 1956, No. 4287, 481–508. Bohlke, J.E. & Chaplin, C.C.G., 1968. Fishes of the Bahamas and Adjacent Tropical Waters. Livingston Publishing Co., Wynnewood, Philadelphia, 771 pp. Clark, E., Ben-Tuvia, A. & Steinitz, H., 1968. Sea Fish. Res. Stn Haifa, Bull. No. 49, 15–31. Colin, P.L., 1971. Copeia, No. 3, 469–479. Colin, P.L., 1972. Zoologica, 57, 137–169. Cummins, R.A., 1979. Ph. D. thesis, University of Sydney, Sydney, Australia, 252 pp. Daanje, A., 1950. Behaviour, 3, 48–49. Dingle, H., 1969. Anim. Behav., 17, 561–575. Dingle, H., 1972. In, Behaviour of Marine Animals: Current Perspectives in Research, Vol I. Invertebrates, edited by H.E.Winn & B.L.Olla, Plenum Press, New York, pp. 126–156. Farrow, G.E., 1971. Symp. zool. Soc. Land., No. 28, 455–500. Fishelson, L., 1975. Aust. J. mar.freshwat. Res., 26, 329–341. Goren, M. & Karplus, I., 1983. Senck. Biol., 63, 27–31. Harada, E., 1969. Publs Seto mar. biol. Lab., 26, 315–334. Harada, E., 1971. Biol. Mag. Okinawa, 9, 1–8 (in Japanese, English summary). Hazlett, B.A. & Bossert, W.H., 1965. Anim. Behav., 13, 357–373. Herald, E.S., 1961. Living Fishes of the World. Doubleday & Company Inc., Garden City, New York, 303 pp. Hjorth, I., 1966. In, A Discussion on Ritualization of Behaviour in Animals and Man, edited by J.S.Huxley, Phil. Trans. Roy. Soc. Ser. B, 251, 485–497. Hoese, D.F. & Allen, G.R., 1976. Jap. J. Ichthyol., 23, 199–207. Hoese, D.F. & Randall, J.E., 1982. Indo-pacific Fishes, 1, 1–18. Hoese, D.F. & Steene, R., 1978. W. Rec. Aust. Mus., 6, 379–389. Karplus, I., 1970. M. Sc. thesis. The Hebrew University of Jerusalem, Israel, 72 pp. Karplus, I., 1976. Ph. D. thesis, The Hebrew University of Jerusalem, Israel, 123 pp. Karplus, I., 1979. Z. Tierpsychol, 49, 173–196. Karplus, I., 1981. J. exp. mar. Biol. Ecol., 51, 21–35. Karplus, I. & Ben-Tuvia, S., 1979. Z. Tierpsychol., 51, 225–232. Karplus, I., Szlep, R. & Tsurnamal, M., 1972a. Mar. Biol., 15, 95–104. Karplus, I., Szlep, R. & Tsurnamal, M., 1974. Mar. Biol., 24, 259–268. Karplus, I., Szlep, R. & Tsurnamal, M., 1981. J. exp. mar. Biol. Ecol., 51, 1–19. Karplus, I., Tsurnamal, M. & Szlep, R., 1972b. Mar. Biol., 17, 275–283. Karplus, I., Tsurnamal, M., Szlep, R. & Algom, D., 1979. Z. Tierpsychol., 49, 337– 351. Karplus, I. & Vercheson, A., 1978. Crustaceana, 34, 220–222. Klausewitz, W., 1960. Senck. Biol., 41, 149–162. Klausewitz, W., 1964. Senck. Biol., 45, 123–144. Klausewitz, W., 1969. Senck. Biol., 50, 41–46. Klausewitz, W., 1970. Senck. Biol., 51, 177–179. Klausewitz, W., 1974a. Senck. Biol., 55, 69–76. Klausewitz, W., 1974b. Senck. Biol., 55, 205–212.

GOBIID FISHES AND ALPHEID SHRIMPS

507

Longley, W.H. & Hildebrand, S.F., 1941. Pap. Tortugas Lab., 34, 1–331. Lubbock, R. & Polunin, N.V.C., 1977. Revue suisse Zool., 84, 505–514. Luling, K.H., 1959. Forschn Fortschr., 28, 265–268. Luther, W., 1958a. Z. Tierpsychol., 15, 175–177. Luther, W., 1958b. Natur. Volk., 88, 141–146. MacGinitie, G.E., 1939. Am. Midl. Nat., 21, 489–505. MacGinitie, G.E. & MacGinitie, N., 1968. Natural History of Marine Animals. McGraw-Hill, New York, 2nd edition, 523 pp. Macnae, W., 1957. J. Ecol., 45, 361–387. Macnae, W. & Kalk, M., 1962. J. Anim. Ecol., 31, 93–128. Magnus, D.B.E., 1967. Helgoländer wiss. Meeresunters., 15, 506–522. Marler, P., 1968. In, Animal Communication, edited by T.A.Sebeok, Indiana University Press, Bloomington, Indiana, pp. 103–127. Maynard-Smith, J. & Parker, G.A., 1976. Anim. Behav., 24, 159–175. Max Westby, G.W., 1975. Anim. Behav., 23, 192–213. Miya, Y. & Miyake, S., 1969. Publs Seto mar. biol. Lab., 16, 307–314. Moehring, J.L., 1972. Ph. D. thesis, University of Hawaii, Honolulu, 373 pp. Morris, D., 1957. Behaviour, 11, 1–12. Nolan, B.A. & Salmon, M., 1970. Forma Functio, 2, 289–335. Palmer, C., 1963. Senck. Biol., 44, 447–450. Parker, C., 1974. J. theor. Biol., 47, 223–243. Paulson, A.C., 1978. Copeia, No. 1, 168–169. Pielou, E.C., 1969. An Introduction to Mathematical Ecology. John Wiley & Sons Inc., New York, 286 pp. Polunin, N.V.C. & Lubbock, R., 1977. J. Zool., 183, 63–101. Polunin, N.V.C. & Lubbock, R., 1979. Bull. Brit. Mus. (not. Hist.) Zool., 36, 239– 249. Preston, J.L., 1978. Anim. Behav., 26, 791–802. Randall, J.E., 1963. Underwater Nat., 1, 6–36. Rao, H.S., 1939. Proc. natn. Inst. Sci. India, 5, 275–279. Rice, A.L. & Chapman, C.J., 1971. Mar. Biol., 10, 330–342. Schembri, P.J. & Jaccarini, V., 1978. Mar. Biol., 47, 55–61. Shinn, E.A., 1968. J. Palaeont., 42, 879–894. Smith, J.L.B., 1959. Ichthyol. Bull., 13, 185–225. Steinberg, J.B. & Conant, R.C., 1974. Anim. Behav., 22, 617–627. Thomassin, B.A., 1971. Symp. zool. Soc. Land., 28, 371–386. Webb, F.J., 1974. M. Sc. thesis, University of South Florida, Tampa, Florida, 98 pp. Weiler, D.A., 1976. M. Sc. thesis, University of Puerto Rico, 65 pp. Wiley, R.H., 1973. Behaviour, 47, 129–152. Wilson, E.O., 1975. Sociobiology the New Synthesis. The Belknap Press of Harvard University Press, Cambridge, Mass., 697 pp. Yanagisawa, Y., 1976. Publs Seto mar. biol. Lab., 23, 145–168. Yanagisawa, Y., 1978. Publs Seto mar. biol. Lab., 24, 269–325. Yanagisawa, Y., 1982. Jap. J. Ichthyol., 28, 401–422. Yanagisawa, Y., 1984. Publs Seto mar. biol. Lab., 29, 93–116. Zander, D.D., 1967. Meteor Forschungergeb. Ser. D, Biol., 2, 69–84.

Oceanogr. Mar. Biol. Ann. Rev., 1987, 25, 563–575 Margaret Barnes, Ed. Aberdeen University Press

THE ECOLOGICAL IMPACT OF SALMONID FARMING IN COASTAL WATERS: A REVIEW R.J.GOWEN and N.B.BRADBURY Department of Biological Science, The University, Stirling, Scotland

INTRODUCTION During the last decade the marine cultivation of salmonids (rainbow trout, Salmo gairdnerii and Atlantic salmon, S. salar) has become a well-established industry in a number of northern European countries. For example, in 1985, production of salmon in Norway and Scotland was estimated at 27000 and 7000 tonnes (103 kg), respectively. The expansion of salmonid farming (particularly salmon) has led to an increased awareness that fish farming may have a considerable impact on the marine environment. Intensive cultivation generates large amounts of organic and inorganic waste (uneaten food, faecal and excretory material) all of which are continually produced and released at single point sources into the environment. Concern has been expressed (Anonymous, 1983) both by conservationists and by fish farmers who fear that the interaction between a farm and its environment could result in harmful feedback which may have an adverse effect on the economic viability of the farm. The purpose of this review is twofold. First, to identify waste compounds and to quantify their production. Secondly, to determine the effects of fish-farm waste on the marine environment and the scale of such effects. THE FLUX OF MATERIAL THROUGH A FISH FARM To assess the impact of cage farming on the marine environment it is necessary to know the composition and quantities of waste material released. Furthermore, for any predictions to be made prior to the establishment of a farm in a particular location, the flux of waste compounds through the farm has to be quantified. The composition and physical nature of fish-farm waste reflects the composition of the diet and the digestibility of its components. Of the three most widely used types of food (dry, moist, and wet) dry food is the most commonly used by salmon farmers (Anonymous, 1983). The main components of dry food (given by food manufacturers, for example, Fulmar, British Petroleum, and Ewos Baker) are: protein 46 to 51%, carbohydrate 18%, and lipid 14 to 17%. In addition, these diets include phosphorus and vitamins; therapeutants and pigments may also be included. All of the compounds in fish food together with the by-products of metabolism are potential waste products. Thus fish-farm waste will include the following: organic carbon and organic nitrogen (carbohydrate, lipid and protein), ammonium, urea, bicarbonate, phosphate, vitamins, therapeutants, and pigments.

ECOLOGICAL IMPACT OF FISH FARMING

509

Organic carbon and nitrogen compounds form the bulk of the waste food and faeces. Both waste food and faeces are more dense than sea water and will sink to the sediment in the immediate vicinity of a farm. Large inputs of such particulate organic material (in addition to the natural input) to the sediment can have a considerable effect on sediment chemistry and the ecology of benthic organisms (Fenchel & Riedl, 1970; Pearson & Rosenberg, 1978). Fish farms also release dissolved organic material into the sea. Any substantial and measurable increase in the concentration of dissolved nutrients has been termed hypernutrification (Anonymous, 1984). This can lead to an increase in phytoplankton growth and productivity if growth is limited by nutrients. In freshwater ecosystems this is referred to as eutrophication and we have used this term in the same context in this review. With the exception of some low salinity coastal environments (see e.g., Taft & Taylor, 1976; Schindler, 1981) dissolved inorganic nitrogen is held to be the most important growthlimiting nutrient for phytoplankton in coastal marine waters (Dugdale, 1967). The input of soluble nitrogenous compounds is known to cause hyper-nutrification of coastal waters (Ryther & Dunstan, 1971) and soluble nitrogenous waste from fish farms (ammonium and urea) may have similar effects. The detailed effects of particulate organic carbon and particulate and soluble nitrogenous waste on the marine environment are discussed below. Bicarbonate and phosphate are likely to have a minimal effect on the marine environment and in the case of vitamins, antibiotics and pigments the possible impact is poorly understood but could be significant. Bicarbonate is a by-product of respiration and is excreted via the gills. It is alkaline, but because of the natural buffering capacity of sea water it is unlikely to have any significant effect on the pH of sea water. In fresh water enhanced levels of dissolved inorganic phosphate have resulted in eutrophication (Anonymous, 1982) and phosphate released from freshwater fish farms may cause similar effects (Bergheim, Hustveit, Kittelsen & Selmer-Olsen, 1984). It has been suggested, however, that, with the exception of some low salinity coastal environments, phosphate is not important in controlling algal growth in coastal waters and is therefore not an important waste product. Vitamins such as biotin are essential for the healthy growth of salmonids (Halver, 1982). Biotin is photo-labile, with a half life of about one week (Carlucci, Silbernagel & McNally, 1969). It is thus unlikely to accumulate in the water column, but is rapidly assimilated by phytoplankton and could stimulate growth of some species. Organic compounds in fish-farm waste have been reported to stimulate the growth of particular species of phytoplankton (Nishimura, 1982) and biotin has been implicated in the toxicity of Gyrodinium aureolum, a red tide dinoflagellate (Turner, Bullock, Tett & Roberts, in press). Antibiotics are not thought to accumulate in the marine environment (Anonymous, 1983). The fate of pigments is not known. Since the effects of bicarbonate, phosphate, vitamins, antibiotics, and pigments are either minor or poorly understood they are not discussed further. WASTE FOOD The amount and composition of waste food is clearly dependent on the original composition of the food and farming practice (Anonymous, 1983). Warrer-Hansen (1982) gives the following values for unconsumed food based on data from the Danish Water Quality Institute. Dry diets (pellets) Moist diets (minced fish and binder) Wet diets (minced fish)

1–5% 5–10% 10–30%

Increased wastage is associated with higher water content and decreasing stability in water. The above data were, however, collected from tank and pond farms, and wastage at sea-cage farms is likely to be in excess

510

R.J.GOWEN AND N.B.BRADBURY

of these values. Beveridge (1984) compared food conversion ratios for trout reared in freshwater ponds and cages and found that the ratio was approximately 20% higher for cage-reared trout. Beveridge assumed that the higher ratio for cage-reared trout was due to a greater proportion of the food being wasted. Values of between 27 and 31 % for wastage of dry and moist diets can be derived from the data of Penczak et al. (1982). These values relate to cagereared rainbow trout in fresh water and are higher than the estimate of 20%, for cage-reared salmon in the sea, suggested by Braaten, Aure, Ervic & Boge (1983). As will be shown later the amount of waste food has an important effect on the scale of the impact on the benthos and good estimates of wastage are required if the impact is to be modelled accurately. There is unlikely to be any substantial loss of carbon and nitrogen by solution or microbial activity from food settling to the sea bed (Collins, 1983). Thus, uneaten food will generally arrive at the sea bed with an unaltered composition. FAECAL AND EXCRETORY WASTE The amount of waste produced by a salmonid farm will depend on the number of fish. There appears, however, to be good agreement between the amount of food consumed and the quantities of faecal material produced. Butz & Vens-Cappell (1982) calculated a faecal production of 260 g dry weight of faeces per kg of food consumed. Thus, 26% of the food eaten ends up as faeces. This estimate is in general agreement with the diet digestibility figure of 75% given by Rychly & Spannhof (1979) and by Smith (1983). The composition of the faeces and urine will obviously depend on the digestibility of the components in the food. The composition of the urine can also be influenced by the protein content of the food (Rychly, 1980). Measurements made by Penczak et al. (1982) suggest that approximately 30% of the faeces is carbon, with nitrogen and phosphorus contributing 4% and 2%, respectively. The amount of excretory nitrogen produced can be determined from the digestibility of protein and the protein content of the food. Edwards (1978) gives the nitrogen content of fish food as 7·68% and Bromley & Smart (1981) give a value of 2·72% for the nitrogen content of salmon. Thus, 1 tonne of food contains 76·8 kg of nitrogen and 1 tonne of salmon contains 27·2 kg of nitrogen. If a food conversion ratio of 2:1 and a wastage of 20% are assumed, then 122·9 kg of nitrogen are consumed but only 27·2 kg are retained. That is, 22% of the consumed nitrogen is retained and 78% lost as faecal and excretory nitrogen. The former is in good agreement with the figure of 30% for nitrogen retention measured by Penczak et al. (1982). Assuming that the nitrogen content of the faeces is 4% (Penczak et al., 1982) then between 68 and 86% of the consumed nitrogen is voided from the fish as soluble ammonium and urea. That is approximately 32 kg of ammonium are produced per tonne of food fed. These estimates are in general agreement with estimates of ammonium production by salmonids in land-based farms. For example, Liao & Mayo (1974) and Willoughby, Larson & Bowen (1972) estimated ammonium production as 28·9 and 32 kg, respectively, per tonne of food fed. Estimates of ammonium excreted per tonne of fish produced range from 45 kg (WarrerHansen, 1982) to 55·5 kg (Solbe, 1982). Ammonium forms the bulk of the excretory nitrogen, although the proportions of ammonium and urea are variable (McLean & Frazer, 1974; Rychly, 1980). Ammonium and urea are, however, both utilized by phytoplankton for growth and in this respect it is not necessary to distinguish between the two. The flux of carbon and nitrogen through a cage-farm is summarized in Figure 1.

ECOLOGICAL IMPACT OF FISH FARMING

511

Fig. 1.—The average flux of carbon and nitrogen through a salmonid farm.

THE EFFECTS OF FISH-FARM WASTE ON THE MARINE ENVIRONMENT ORGANIC ENRICHMENT The majority of coastal marine sediments have an oxic layer overlying an anoxic layer. The presence of the former depends on a balance between the consumption and supply of oxygen. One of the first effects of enrichment is an increase in oxygen consumption by heterotrophic organisms within the sediment. Enell & Löf (1983) measured oxygen consumption rates of between 45 and 55 mg O2·m−2·h−1 for the sediment beneath a freshwater cage-farm compared with 16 mg O2·m−2·h−1 for an undisturbed sediment. When the demand for oxygen exceeds the supply, the sediment becomes anoxic and at this point there are major changes in sediment chemistry and the ecology of benthic organisms. In extreme cases (low turbulence and high organic input) the water overlying the sediment can also become anoxic (Tsutsumi & Kikuchi, 1983). Even in more turbulent locations the bottom water beneath fish farms can become depleted in oxygen for long periods (Brown, Gowen & McLusky, 1987). Upwelling of oxygen-depleted water through cages could be harmful to fish. In the absence of oxygen, anaerobic processes predominate and the sediment becomes reducing. The reducing capacity of the sediment depends on the degree of enrichment, and can be assessed by measuring the redox potential (oxidation/reduction level) of the sediment. Pearson & Stanley (1979) used the redox potential of the sediment to assess the effects of wood pulp waste and found that in regions of high input the sediment became highly reducing with redox levels as low as—150 mv at 4 cm depth in the sediment. Redox potentials as low as—150 mv have been measured in the sediment beneath fish cages (Brown et al., 1987) indicating that amounts of fish-farm waste are sufficient to cause the sediment to become highly reducing. Anaerobic processes result in the production of reduced compounds (ammonium, hydrogen sulphide, and methane) which, if produced in sufficient quantities, can be released from the sediment. Blackburn (1983) suggests that there is a net production of ammonium by anaerobic sediments and Enell & Löf (1983) estimated that the rate of ammonium release beneath a freshwater cage-farm was between 2·6 and 3·3 times the rate of release from an undisturbed oxic sediment. Ammonium is soluble and will dissolve in the

512

R.J.GOWEN AND N.B.BRADBURY

overlying water. Thus, ammonium released from anoxic sediments beneath fish farms could contribute to the total soluble nitrogenous waste produced. Hydrogen sulphide is slightly soluble and methane is insoluble. It might therefore be expected that these two gases will be released from the sediment in a gaseous form. Braaten et al. (1983) observed out-gassing from the sediment beneath a cage-farm and out-gassing is commonly observed at many fish farms in Scotland (R.Gowen, unpubl. data). Hydrogen sulphide is toxic to fish (Fayette & Haines, 1980), and out-gassing of this from sediments beneath fish cages could have a detrimental effect on fish health. Such an effect was observed by Braaten et al. (1983), who related gill damage and increased mortalities to out-gassing from the sediment. It is not possible to specify at what level of enrichment out-gassing and possible damage to the fish will take place, but it is clear that in sheltered, shallow-water sites, levels of enrichment which result in highly reducing sediments and outgassing does pose a threat to the fish. In a review of the effects of organic enrichment on the ecology of macrobenthic organisms, Pearson & Rosenberg (1978) suggested that different types of organic waste (wood pulp, domestic sewage, and seaweed waste) produce similar effects. The few studies which have attempted to assess the impact of fishfarm waste (Stewart, 1984; Brown et al., 1987) indicate that the effects of fish-farm waste are similar to those described by Pearson & Rosenberg. Thus, there is a gradient of enrichment from the cages outwards. The large amounts and constant deposition of waste beneath cages can result in an azoic zone which is devoid of macrobenthic organisms (Stewart, 1984). In the immediate vicinity of the farm the macrofauna is likely to be impoverished and dominated by opportunistic species, such as Capitella capitata, which are held to be indicative of enriched sediments (Belan, 1970). Beyond this zone there is a transition zone within which enrichment is sufficient to stimulate growth. Outwith this region the biomass, species composition and abundance decline to levels which are typical for that location. EUTROPHICATION The consequences of any hypernutrification in terms of enhanced primary production and phytoplankton standing crop are complex. Enhanced levels of ammonium have been measured in the vicinity of a fish farm located in a Scottish sea loch although no increase in phytoplankton standing crop was observed (R.Gowen, unpubl. data). Enhanced primary production could take place without any increase in standing crop, if additional biomass was rapidly removed, for example by grazing. Measurements of primary production are therefore likely to be more informative than simple estimates of algal biomass, although Oviatt, Keller, Sampou & Beatty (1986) did not find a linear relationship between the level of hypernutrification and enhancement of primary production and suggested that at higher levels of hypernutrification other factors limit production. Several factors may restrict the utilization of additional nitrogen by phytoplankton. In turbid or deep, vertically mixed locations, the availability of light may be controlling growth. The flushing time of a water body can limit the accumulation of biomass within that water body (Gowen, Tett & Jones, 1983). Thus, in rapidly flushed locations (e.g., some of the small fjordic sea lochs on the western coast of Scotland) phytoplankton may not remain in the vicinity of the farm for a sufficient time to utilize any additional nitrogen. Enhanced phytoplankton growth could be a threat to fish held in cages if additional growth was in the form of toxic species. Takahashi & Fukazawa (1982) found that different species of phytoplankton responded differently to varying nutrient conditions. In particular, they found that the growth of the microflagellate Olisthodiscus luteus and a gymnodinioid dinoflagellate was stimulated by the addition of ammonia. Blooms of a small unidentified chloromonad (similar to O. luteus) have been implicated in fish kills in Scotland (Gowen, Lewis & Bullock, unpubl.) and in southern Ireland (J. Doyle, pers. comm.). It

ECOLOGICAL IMPACT OF FISH FARMING

513

Fig. 2.—The flux of carbon and nitrogen through a salmonid farm with an annual production of 50 tonnes.

should be pointed out that the experiments carried out by Takahashi & Fukazawa (1982) were on laboratory cultures of phytoplankton, and in general the effects of soluble nitrogenous fish-farm waste on the species composition of phytoplankton has not been studied in detail. Furthermore, there is no evidence to suggest that the occurrence of toxic phytoplankton blooms in Scotland (Jones et al., 1982; Gowen, in press) and in Norway (Tangen, 1977) are related to fish-farming activity. Changes in the bacterial flora of the water have been shown to occur around a freshwater farm (Korzeniewski & Korzeniewski, 1982) and eutrophication has been reported to be detrimental to salmonids in fresh water (Colby, Spangler, Hurley & McCombie, 1972). The relationships between the incidence of disease, health of the fish and eutrophication resulting from intensive salmonid farming in the sea have not been studied.

THE SCALE OF THE IMPACT RESULTING FROM FISH-FARM WASTE If the size of the farm is known it is possible to quantify the amounts of waste produced. An example based on a salmon farm with an annual production of 50 tonnes is set out in Figure 2. A farm of this size might be expected to use 100 tonnes of food per year. Thus, over a 12-month period a total of 19·4 tonnes of organic carbon waste, 2·2 tonnes of organic nitrogen waste and 4·0 tonnes of soluble nitrogenous waste will be produced.

514

R.J.GOWEN AND N.B.BRADBURY

ORGANIC ENRICHMENT It is important to distinguish between uneaten food and faeces since the settling velocities of the two are different and this will influence horizontal dispersion and amounts reaching the sea bed. Our own estimates of settling velocities of food pellets ranged from 0·09 m·s−1 to 0·15 m·s−1 (unpubl. data) which is considerably greater than settling velocities of between 0·017 and 0·06 m·s−1 for faecal pellets estimated by WarrerHansen (1979). Given the high settling velocities of uneaten food and faeces there is unlikely to be any substantial loss of carbon and nitrogen from these wastes by solution or microbial activity (Collins, 1983). Some uneaten food is likely to be consumed by wild fish. No estimates of this have been made but it is clear that this consumption could substantially reduce the amount of organic carbon and nitrogen reaching the sea bed in the form of uneaten food pellets. Faecal pellets are easily broken down into smaller particles and this could reduce the quantities of faecal waste reaching the sea bed. Petit (cited in Querellou, Faure & Faure, 1982) estimated that 86% of the suspended solids settled out of the water (14% remaining in suspension) and Querellou et al. (1982) found no intact faecal pellets in heavily stocked aerated tanks. Both of these studies were conducted at freshwater tank-farms and it is difficult to relate the behaviour of faecal pellets in tanks to their behaviour in cages. The proportion of faecal waste which eventually settles out from a cage-farm is likely to be variable being dependent on factors such as stocking density and water movement, but precise amounts are not known. Enell & Löf (1983) measured sedimentation rates of between 17 and 26 g dry material·day−1 beneath a freshwater cage-farm compared with rates of between 2·3 and 3·6 g dry material·day−1 at an undisturbed site. Similar rates were measured by Merican & Phillips (1985). Such high sedimentation rates are unlikely to result from waste food alone and suggest that much of the faecal waste settles out in the immediate vicinity of the farm. The area of sea bed over which the waste will be dispersed will depend on the following: the surface area of the farm, settling velocity of uneaten food and faeces, current speeds, and depth of water beneath the farm; it can be determined from the relationship:

where D represents horizontal distance dispersed, d is water depth, V is current speed and v is settling velocity of the waste. The waste does not, however, originate from a single point source, but from anywhere within the area occupied by the cages. Using the example of a 50-tonne farm, typical stocking densities of 20 kg of fish·m−3 and net depth of 5 m (Edwards, 1978) the area occupied by a farm of this size would be 500 m2. This area represents the minimum area of sea bed which would receive organic waste in the absence of any tidal currents. The depth of water beneath farms will obviously vary although the depth should exceed 10 m and preferably 15 m (Edwards, 1978). Current speeds at fish farms will also vary. Edwards & Edelsten (1976) suggest that current speeds of 0·035 m·s−1 are typical for the upper region of many Scottish sea lochs where fish farms are sited. Thus, if the depth of water beneath the farm is 20 m and the current speed 0·035 m·s−1, uneaten food would be dispersed 6 m beyond the cages, that is over an area of 1222 m2. Faecal waste would be dispersed 18 m beyond the cages, over an area of 3550 m2. This pattern of dispersion will give two zones. An inner zone which receives uneaten food and faeces and an outer zone which receives faecal waste only. If the losses due to consumption of uneaten food by wild fish and the proportion of faecal waste which remains in suspension are ignored, then the likely loadings to the two areas are 28 g·m−2·day−1 for the inner zone and 8 g·m−2·day−1 for the outer zone. These estimates should be regarded as approximate since the model does not take into account variation in current speed and direction. Furthermore, we have used a

ECOLOGICAL IMPACT OF FISH FARMING

515

value of 20% for food wastage and this represents 70% of the total organic carbon deposited directly beneath the cages. Food wastage is difficult to measure but good estimates are obviously necessary to derive precise estimates of enrichment beneath fish farms. The above assessment does show that the deposition of organic carbon beneath most fish farms is extremely high compared with the natural input. For example, Ansell (1974) gives a value of 250 g C·m−2 for the annual sedimentation of organic carbon in Loch Creran, a fjordic sea-loch on the western coast of Scotland. Our estimates are also considerably higher than the value of 1·43 g C·m−2·day−1 given by Pearson (1982) for organic carbon input from wood pulp waste. Although loadings will, in general, be high beneath fish farms, the scale of effect, that is the area of sea bed affected will, in most cases, be small and restricted to the immediate vicinity of the farm. At the present level of salmonid farming in coastal waters this represents a small proportion of the sea bed. In Scotland for example, the proportion of sea bed receiving fishfarm waste is probably less than 1 % of the sea bed in those sea-lochs which are used for fish farming. Some studies have suggested that the effects of fishfarm waste can be detected up to 100 m from the cages but these effects are minor (unpubl. reports by the Nature Conservancy Council of Great Britain). Only directly beneath the cages is the degree of enrichment likely to be sufficient to result in out-gassing, and it has already been suggested that this could pose a threat to fish held in cages, particularly at shallow sites. EUTROPHICATION From the preceding section on waste production it is clear that considerable amounts of soluble waste are produced by salmonid farms, and it has been suggested that the resulting hypernutrification could lead to an increase in phytoplankton growth. The final level of hypernutrification depends on not just the amount of waste generated but also the hydrography of the fish-farm site. Thus, the volume of the water body and the rate of exchange of that body with its adjacent sea area have an important bearing on the final level of enrichment. Simple methods are available for estimating the flushing time of semi-enclosed water bodies (see Gowen et al., 1983 and references cited therein) and, from this, the volume of the water body, and the amount of nitrogenous waste, the level of hypernutrification can be estimated. Such an approach assumes that the soluble waste is dispersed throughout the water body. This assumption is probably only true for small well-mixed estuaries and embayments. In larger semi-enclosed water bodies vertical stratification and limited horizontal advection (particularly during neap tides) would reduce the volume into which the nitrogenous waste is dispersed and this could lead to localized hypernutrification. In fresh water, predictions of the likely enhancement in phytoplankton biomass resulting from a particular level of phosphate enrichment are based on the relationship between chlorophyll concentration and phosphate concentration (Anonymous, 1982). The same can be done for the marine environment with nitrogen, but the relationship between chlorophyll and dissolved nitrogen is variable (P.Tett, pers. comm.). An alternative would be to use the subsistence quota; that is, the minimum cellular content of a nutrient (Droop, 1968). Tett & Droop (in press) give a mean subsistence quota for nitrogen of about 0·05 mol nitrogen: mol carbon. Assuming a carbon to chlorophyll ratio of 50:1 gives an estimate of 0·2 m mol nitrogen-mg chlorophyll−1 (Tett, Edwards & Jones, 1986). Thus, assuming all of the additional nitrogen is converted to biomass, 1 m mol of nitrogen could support a maximum chlorophyll-related phytoplankton biomass of 5 mg. It has, however, already been suggested that the additional nitrogen is not necessarily utilized by phytoplankton. Despite the assumptions which have to be made in estimating the level of hypernutrification and eutrophication, such estimates are useful as a preliminary assessment of the suitability of a site. There is a second source of ammonium to the water column and that is from the remineralization of organic nitrogen in uneaten food and faeces. The rate of ammonium release from

516

R.J.GOWEN AND N.B.BRADBURY

enriched, anoxic marine sediments is not known, but might be important in the long-term and might affect the time taken by the ecosystem within a water body to recover from the effects of fish farming. In general, widespread hypernutrification and eutrophication is unlikely to result from the farming of salmonids in marine coastal waters, at its present level. In the case of the establishment of large farms in slowly flushed locations, there is the possibility of localized eutrophication. REDUCING THE EFFECTS OF FISH-FARM EFFLUENT A number of techniques have been used to reduce the level of organic enrichment beneath fish cages but it seems that these techniques might themselves create problems. Enell, Löf & Bjorklund (1984) successfully collected most of the waste food and faecal waste by suspending a PVC funnel beneath the cages of a freshwater farm. During the use of such a system it would be necessary to ensure that the flow of water through cages was not restricted and that deploying the system did not stress the fish. Furthermore, the maintenance of any permanent fixture beneath cages would probably be difficult in more exposed marine sites. Braaten et al. (1983) assessed the use of a submersible mixer to disperse waste and suggested the use of a submersible pump to collect the sedimented waste. In the case of the former, Braaten et al. found that the submersible mixer prevented further accumulation and reduced the organic layer beneath a fish farm by one third, although they found it necessary to move the fish from the existing site before the mixer was used. In a modified system the mixer was suspended above the sea bed. This system also prevented further accumulation and could be used beneath the cages. When using such a system, it would, however, be necessary to ensure that water, low in oxygen, was not drawn into the cages and that resuspension of sedimented organic waste did not increase the biological oxygen demand in the water around the farm. Braaten et al. give a figure of N.kr 20000 (approximately £2000) for the annual operation cost of a submersible mixer and this, together with the need to move cages during a cleaning period or to monitor oxygen levels probably restricts the general applicability of this technique. We have suggested that waste food could account for 70% of the organic carbon sedimenting directly beneath fish farms. Thus, any reduction in food wastage would have a significant effect in reducing the level of enrichment. The periodic movement of cages to allow the sediment to recover might also be a means of reducing the gross effects of enrichment. At present the rate at which the sediment recovers is, however, not known. It seems likely that the only way of minimizing hypernutrification and eutrophication is to site farms in locations which are well flushed. This would also enhance the dispersal of solid waste. In this respect site assessment and the careful choice of sites before the establishment of a farm would be an important way of reducing the ecological impact of fish-farm effluent and also reduce the risk of affecting the health of the fish. CONCLUSIONS Fish-farm waste is composed of a range of organic compounds, protein, carbohydrate, urea, vitamins, antibiotics, and pigments together with the inorganic ions, ammonium, bicarbonate, and phosphate. Of these the effects of enrichment of marine sediments by organic carbon and hypernutrification of coastal waters by dissolved forms of nitrogen are best understood. The effects of vitamins, antibiotics, and pigments on marine ecosystems is poorly understood but could be significant. Bicarbonate and phosphate are probably not important waste products.

ECOLOGICAL IMPACT OF FISH FARMING

517

The deposition of organic carbon and nitrogen waste in the form of faeces and uneaten food, beneath and in the immediate vicinity of most fish farms will be substantially greater than the natural impact of organic carbon and nitrogen to the sediment. This is likely to lead to the sediment becoming anoxic and a benthic macrofauna which is impoverished or absent. Out-gassing of hydrogen sulphide from anoxic sediments could be detrimental to farmed fish, particularly in sheltered, shallow-water sites. The most severe effects of enrichment and out-gassing are probably confined to the sediment directly beneath the farm and at the present level of salmonid farming this represents a small proportion of the sea bed in a given area. The release of dissolved forms of nitrogen from fish farms together with ammonium released from anoxic, enriched sediments could result in hypernutrification and eutrophication. Localized eutrophication might occur around large farms in poorly flushed locations but at the present level of salmonid farming widespread eutrophication is unlikely. Eutrophication would pose a threat to farmed fish if the growth of toxic species of phytoplankton was stimulated. Techniques exist which can reduce the input of organic waste to the sediment and it is suggested that reducing food wastage and the periodic re-deployment of cages would reduce the effect of organic waste on the benthic ecosystem. The selection of sites with good water exchange would reduce the likelihood of hypernutrification and eutrophication and would also enhance the dispersal of solid waste. It is suggested that an understanding of the interactions between salmonid farms and the marine environment would be useful in assessing sites prior to the establishment of a farm and help to ensure the suitability of the site for salmonid growth, to minimize the ecological impact of farm waste and reduce the risk of harmful feedback affecting the fish. ACKNOWLEDGEMENTS Work on the ecological impact of salmonid farming in Scotland was supported by a research grant from the Highlands and Islands Development Board of Scotland. The authors would like to thank Drs T.Pearson and P. Tett for helpful comments during the preparation of this review and Professor R.I.Currie, Director of the Scottish Marine Biological Association for allowing the use of facilities at the Dunstaffnage Marine Research Laboratory. REFERENCES Anonymous, 1982. Eutrophication of Waters. Monitoring, Assessment and Control. OECD, 2, Rue Andre-Pascal, 75775, Paris, France, 154 pp. Anonymous, 1983. The Environmental Impact of Aquaculture. Report from the Working Group on Environmental Effects to the Steering Committee on Aquaculture, Report 83:5, Swedish Council for Planning and Co-ordination of Research in co-operation with the National Marine Resources Commission, 74pp. Anonymous, 1984. The Causes, Dynamics and Effects of Exceptional Blooms and Related Events. ICES C.M./1984/E: 42, 5 pp. Ansell, A.D., 1974. Mar. Biol., 27, 263–273. Belan, C., 1970. Mar. Pollut. Bull., 1, 59–60. Bergheim, A., Hustveit, H., Kittelsen, A. & Selmer-Olsen, A.R., 1984. Aquaculture, 36, 157–168. Beveridge, M.C.M., 1984. FAO Fish. Tech. Pap. No. 255, 131 pp. Blackburn, T.H., 1983. In, Microbial Geochemistry, edited by W.H.Krumbein, Blackwell Scientific Publications, Oxford, pp. 63–89. Braaten, B., Aure, J., Ervic, A. & Boge, E., 1983. ICES C.M. 1983/F: 26, 11 pp. Bromley, P.J. & Smart, G., 1981. Aquaculture, 23, 325–336.

518

R.J.GOWEN AND N.B.BRADBURY

Brown, J.R., Gowen, R.J. & McLusky, D.S., 1987. J. exp. mar. Biol. Ecol., in press. Butz, I. & Vens-Cappell, B., 1982. In, Report on the EIFAC Workshop on Fish Farm Effluents, edited by J.S.Alabaster, EIFAC Tech. Paper No. 41, pp. 73–82. Carlucci, A.F., Silbernagel, S.B. & McNally, P.M., 1969. J. Phycol., 5, 302–305. Colby, P.J., Spangler, G.R., Hurley, D.A. & McCombie, A.M., 1972. J. Fish. Res. Bd Can., 29, 975–983. Collins, I., 1983. B.Sc. thesis, University of Stirling, Scotland, 92 pp. Droop, M.R., 1968. J. mar. biol. Ass. U.K., 48, 689–743. Dugdale, R.C., 1967. Limnol. Oceanogr., 12, 685–695. Edwards, A. & Edelsten, D.J., 1976. Proc. R. Soc., Edinb. Sect. B, 75, 208–221. Edwards, D.J., 1978. Salmon and Trout Farming in Norway. Fishing News Books Ltd, Farnham, U.K., 195 pp. Enell, M. & Löf, J., 1983. Vatten, 39, 364–375. Enell, M., Löf, J. & Bjorklund, T.L., 1984. Institute of Limnology, Lund, Sweden (Rep. ISSN 0348–0798). Fayette, A.R. & Haines, T.A., 1980. Environ. Pollut., Ser. A, 22, 11–17. Fenchel, T.M. & Riedl, R.J., 1970. Mar. Biol., 7, 255–268. Gowen, R.J., in press. Rapp. P,-v. Réun., Cons. perm. int. Explor. Mer. Gowen, R.J., Tett, P. & Jones, K.J., 1983. J. exp. mar. Biol. Ecol., 71, 1–16. Halver, J.E., 1972. Editor, Fish Nutrition. Academic Press, New York, 713 pp. Jones, K.J., Ayres, P.A., Bullock, A.M., Roberts, R.J. & Tett, P., 1982. J. mar. Biol. Ass. U.K., 62, 771–782. Korzeniewski, K. & Korzeniewski, J., 1982. Polskie Archwm Hydrobiol., 14, 183– 187. Liao, P.B. & Mayo, R.D., 1974. Aquaculture, 3, 61–85. Merican, Z.O. & Phillips, M.J., 1985. Aquacult. Fish. Mgmt, 16, 55–69. McLean, W.E. & Frazer, F.J., 1974. Environ. Prot. Ser. Pacif. Region, Surveillance Rep., EPS5-EPR-74–5. Nishimura, A., 1982. Bull. Plankton Soc. Japan, 29, 1–7. Oviatt, C.A., Keller, A.A., Sampou, P.A. & Beatty, L.L., 1986. Mar. Ecol. Prog. Ser., 28, 69–80. Pearson, T.H., 1982. J. exp. mar. Biol. Ecol., 57, 93–124. Pearson, T.H. & Rosenberg, R., 1978. Oceanogr. Mar. Biol. Ann. Rev., 16, 229–311. Pearson, T.H. & Stanley, S.O., 1979. Mar. Biol., 53, 371–379. Penczak, T., Galicka, W., Molinski, M., Kusto, E. & Zalewski, M., 1982. J. appl. Ecol., 19, 371–393. Querellou, J., Faure, A. & Faure, C., 1982. In, Report on the EIFAC Workshop on Fish Farm Effluents, edited by J.S.Alabaster, EIFAC Tech. Paper, No. 41, pp. 87–97. Rychly, J., 1980. Aquaculture, 20, 343–350. Rychly, J. & Spannhof, L., 1979. Aquaculture, 16, 39–46. Ryther, J.H. & Dunstan, W.M., 1971. Science, 171, 1008–1013. Schindler, D.W., 1981. In, Estuaries and Nutrients, edited by B.J.Neilson & L.E. Cronin, The Humana Press Inc., Clifton, New Jersey, pp. 71–82. Smith, P., 1983. Fish Farming Int., 10, 6 only. Solbe, J.F. de L.G., 1982. In, Report on EIFAC Workshop on Fish Farm Effluents, edited by J.A.Alabaster, EIFAC Tech. Paper, No. 41, pp. 29–56. Stewart, K.I., 1984. M. Sc. thesis, University of Stirling, Scotland, 45 pp. Taft, J.L. & Taylor, W.R., 1976. In, Estuarine Processes, Volume 1, Uses, Stresses and Adaptation to the Estuary, edited by M.Wiley, Academic Press, London, pp. 79–89. Takahashi, M. & Fukazawa, N., 1982. Mar. Biol., 70, 267–273. Tangen, K., 1977. Sarsia, 63, 123–133. Tett, P. & Droop, M., in press. In, Handbook of Laboratory Model Systems for Microbial Ecosystems Research, edited by J.W.T.Winpenny, CRC Press, Baton Rouge, Florida. Tett, P., Edwards, A. & Jones, K.J., 1986. Estuar. cstl shelf Sci., 23, 641–672. Tsutsumi, H. & Kikuchi, T., 1983. Publ. Amakusa Mar. Biol. Lab. Kyushi University, 7, 17–40. Turner, M.F., Bullock, A.M., Tett, P. & Roberts, R.J., in press. Rapp. P.-v. Réun., Cons. int. Explor. Mer. Warrer-Hanson, I., 1979. Fish Farming Int., 6, 32–34.

ECOLOGICAL IMPACT OF FISH FARMING

519

Warrer-Hanson, I., 1982. In, Report on EIFAC Workshop on Fish Farm Effluents, edited by J.S.Alabaster, EIFAC Tech. Paper, No. 41, pp. 57–63. Willoughby, H., Larson, N. & Bowen, J.T., 1972. Am. Fish. U.S. Trout News, 17, 6 only.

AUTHOR INDEX

References to complete articles are given in heavy type; references to pages are given in normal type; references to bibliographical lists are given in italics. Abbott, R.T. See Warmke, G.L., 174, 176; 283 Abel, D.J., 80; 85 Abel, E.F., 513, 526; 560 Ackley, S.F., 25; 36 See Clarke, D.B., 32; 36 See Garrison, D.L., 13, 25, 32; 36 Achituv, Y., 173, 175, 178, 179, 181, 192, 193, 202, 206, 211, 320, 337; 271, 347 See Gev, S., 181, 201; 275 See Susswein, A.J., 173; 282 Addicott, J.F. See Wiens, J.A., 44; 90 Adkins, L. See Peretz, B., 202, 205, 231, 232, 236, 238; 280 Advokat, C., 205, 229, 254, 259; 271 Agenbag, J.J. See Shannon, L.V., 355, 435, 490; 503 Ainley, D.G., 13, 30, 34; 36 See Fraser, W.R., 34; 36 Albaret, J.J., 454; 495 Alden, R.W., 62; 85 Alegre, B. See Boerema, L.K., 496 Alekseev, F.E., 453; 495 Alekseeva, E.I. See Alekseev, F.E., 453; 495 Aletsee, L. See Smith Jr, W.O., 20; 38 Alexander, V., 13, 14, 16, 33; 36 See Niebauer, H.J., 14, 16; 37 See Schandelmeier, L., 13, 16, 33; 38 Algom, D. See Karplus, I., 507; 561 Allan, J.K., 175, 247; 271 Alldredge, A.L., 117, 118; 156 Allen, G.R., See Hoese, D.F., 526; 561 Allen, J.R.L., 92; 110 Allen, M.S., See Pettit, G.R., 280 Aller, R.C., See Ray, A.J., 110; 111 See Rhoads, D.C., 124; 162 Allue, C. See Macpherson, E., 424; 501

Altmann, S.A., 544; 560 Alvarez, R. See Jacklet, J.W., 245; 276 Alvarez, R.B., See Strumwasser, F., 192; 282 Ambrose, H.W. See Hamilton, P.V., 173, 203; 275 Ambrose III, H.W., 251, 252, 255, 256; 271 See Hamilton, P.V., 171, 174, 229; 275 Ambrose, K.P. See Ambrose III, H.W., 251; 271 Ambrose, Jr, W.G., 124, 154; 156 Amico, V. See Danise, B., 273 See Fattorusso, E., 274 Anderson, D.J., 77, 78; 85 Anderson, D.T., 320; 347 See McWilliam, P.S., 119; 161 Anderson, E.D. See Brown, B.E., 497 Anderson, F.E. See McLusky, D.S., 143; 161 Anderson, G.R.V. See Russell, B.C., 48; 88 Anderson, J.B. See Dunbar, R.B., 34; 36 Anderson, L., 20; 36 Anderson, M.R. See Downing, J.A., 49, 52, 53, 54, 56, 57, 58; 86 Ando, Y., 256, 257; 271 Andre, S.V. See Peterson, C.H., 124; 162 Andrew, N.L., 39–90; 48, 62, 69, 77, 78; 85 See Choat, J.H., 77; 86 Andrew, P.A. See Bergh, M.O., 465; 496 Andrew, R.J., 560; 561 Andrews, J.A., 115; 156 Andronov, V.N., 415, 416, 417, 418, 420, 422, 484, 489; 495 Angel, H.H., 140; 156 Angel, M.V. See Angel, H.H., 140; 156 See Fasham, M.J.R., 82; 86 Annandale, N., 302, 307, 315, 320; 347 Anonymous, 402, 563, 564, 565, 571; 495, 574 Ansell, A.D., 197, 236, 571; 271, 574 520

AUTHOR INDEX

Antonia, R.E., 103; 110 Anthony, V.C. See Brown, B.E., 497 Arch, S., 188, 192; 271 Archie, J.W. See Rohlf, F.J., 76, 83; 88 Ardai, J.L. See Jacobs, S.S., 13; 37 Armstrong, M.J., 364, 415, 433, 443, 458, 459, 460, 480, 483, 494; 495 See Hampton, I., 435, 458; 499 See Kinloch, M.A., 375; 500 See Shelton, P.A., 431, 459, 473, 475, 476, 477, 478, 487, 491, 494; 503, 504 Arnould, C., 215; 271 Arntz, W.E., 124; 156 Arvanitaki, A., 245; 271 Arya, S.P.S. See Businger, J.A., 102; 110 Aspey, W.P., 173, 182, 205, 229, 230, 246, 247, 248, 249, 253; 271 See Feinstein, R., 179; 274 Assorov, V.V., 418, 421, 423, 447, 448, 449; 495, 496 Atema, J. See Botero, L., 128, 130; 157 Aubin, D. See Bhaud, M., 118, 137; 157 Audesirk, G.J. See Audesirk, T.E., 182, 224;271 Audesirk, T., 208, 244; 271 Audesirk, T.E., 174, 181, 182, 183, 185, 189, 190, 193, 200, 201, 202, 203, 206, 208, 237, 238, 244, 250; 271 See Emery, D.G., 208; 274 Augier, J., 262;271 Augter, G.K. See Lickey, M.E., 278 Augustyn, C.J., 430, 457; 496 See Hatanake, H., 499 See Payne, A.I.L., 410, 439, 448, 467; 502 See Pollock, D.E., 456; 503 See Uozumi, Y., 411; 504 Aure, J. See Braaten, B., 565; 574 Aurivillius, C.W.S., 288, 295; 347 Austin, T., 216; 271 Avichezer, D. See Gilboa-Garber, N., 270; 275 Axel, R. See Scheller, R.H., 281 Ayres, P.A. See Jones, K.J., 574 Azevedo, A. See Santschi, P.H., 111 Baba, K., 174, 175, 176, 177, 191; 271 Babayan, V., 358, 375, 376, 379, 412, 413, 417, 465, 466, 484; 496 Baccetti, B., 296; 347 Badenhorst, A., 355, 391, 392, 395, 396, 471; 496 Bage, F., 320; 347 Baggerman, B., 126, 127, 132, 136, 141, 142, 143, 144, 153, 154; 156

521

Bailey, C.H., 201, 259; 271 Bailey, G.W. See Shannon, L.V., 433; 503 Bailey, K.M., 492; 496 Baird, D., 413, 419, 420, 434, 443, 445, 483; 496 See Hecht, T., 450; 499 Bakun, A. See Parrish, R.H., 353; 502 Balaparameswara Rao, M., 234; 271 Balsam, W.L. See Walsh, J.J., 38 Bambach, R.K. See Levinton, J.S., 125; 160 Bandel, K., 174, 183, 190, 193, 201; 271 Banner, A.H., 508, 509, 512, 513, 514; 561 See Banner, D.M., 508, 509, 512, 513, 514, 516; 561 Banner, D.M., 508, 509, 512, 513, 514, 516; 561 See Banner, A.H., 508, 509, 512, 513, 514; 561 Banner, F.T. See Tyler, P.A., 121, 126, 142, 153; 163 Baptista, S.R. da F., 403, 437, 445; 496 Barash, A., 175, 176, 177, 190, 191, 194; 271 Barber, R.T., 18, 417; 36, 496 Bard, F.X., 454; 496 Bardach, J.E., 209; 271 Barkai, R., 430; 496 Barker, Jr, H.R. See Dauer, D.M., 118; 157 Barnard, K.H., 402; 496 Barnes, D.M., 168; 271 Barnes, H., 71, 72, 325; 85, 347 See Klepal, W., 292; 349 Bartsch, P., 174; 272 Bartz, R., 152; 156 Bary, B.M. See Southward, A.J., 48; 89 Batac-Catalan, Z. See Porter, J.W., 117; 162 Batchelor, A.L., 443; 496 Bateman, G.I., 72; 85 Batham, E.J., 287, 292, 320, 323, 327, 340; 347 Battaile, J. See Finer, J., 274 Baumann, M. See Smith Jr, W.O., 20; 38 Baumgarten, R. von. See Jahan-Parwar, B., 208; 277 Baur Jr, P.S. See Krauhs. J.M., 270; 278 Bayer, F.M., 509, 526, 527; 561 Bayne, B.L., 117, 217; 156, 272 Bayne, C.J., 270; 272 Beatty, L.L. See Oviatt, C.A., 568; 575 Bebbington, A., 169, 171, 172, 174, 175, 176, 177, 200, 243, 247, 248; 272 See Thompson, T.E., 190, 191, 194; 282 Becker, J. See Papka, R., 201; 280 Beddington, J.R., 460, 464; 496 Beeman, R.D., 115, 174, 177, 206, 215, 256; 156, 272 Belan, C., 568; 574 Bell, J.D., 48; 85

522

OCEANOGRAPHY AND MARINE BIOLOGY

See Gray, C.A., 49; 87 Bell, L. See Lederhendler, I., 202, 252; 278 Bell, S.S., 118, 144, 154; 156 Ben-Tuvia, A. See Clark, E., 561 Ben-Tuvia, S. See Karplus, I., 507, 543; 561 Beondé, A.C., 270; 272 Berenbeim, D.Y. See Assorov, V.V., 447, 448, 449; 495 Berg Jr, C.J. See Capo, T.R., 200; 272 Bergh, M.O., 415, 418, 465, 466; 496 Bergheim, A., 564; 574 Berghuis, E.M. See Farke, H., 143, 154; 158 Bergquist, P.R., 115; 156 Berndt, W., 287, 292, 293, 295, 337; 347 Bernstein, B. See Jacklet, J.W., 245; 276 Bernstein, B.B., 46, 61, 64, 68, 74; 85 Berrie, A.D. See Mackey, A.P., 49; 88 Berrill, N.J., 115; 156 Berruti, A. See Crawford, R.J.M., 480; 498 See Duffy, D.C., 461; 499 See Matthews, J.P., 420, 486; 501 Berry, R.E. See Lickey, M.E., 214; 278 Berthet, P. See Gerard, G., 56; 86 Bethe, A., 178; 272 Betzer, P.R. See Walsh, J.J., 38 Beukema, J.J., 117, 118; 156 Bevelaqua, F.A., 269; 272 Beveridge, M.C.M., 565; 574 Beyers, C.J. de B., 410, 456, 457; 496 See Pollock, D.E., 425, 429; 503 Bhakuni, D.S., 262; 272 Bhaud, M., 118, 137; 157 Bhup, R., 133; 157 Bickell, L.R. See Chia, F.-S., 115; 157 Biggs, D.C. See El-Sayed, S.Z., 14, 30; 36 See Gilbert, P.M., 26; 36 Birkeland, C., 142; 157 Bjorklund, T.L. See Enell, M., 572; 574 Blackburn, T.H., 567; 574 Blackman, G.E., 72; 85 Blair, S.M. See Littler, M.M., 173; 278 Blake, W. See Pettit, G.R., 280 Blankenship, J.E., 183, 184, 186, 192, 261, 265; 272 See Aspey, W.P., 173, 182, 205, 229, 230, 246, 247, 248, 253; 272 See Langlais, P.J., 260, 265; 278 See Nagle, G.T., 192; 279 See Strenth, N.E., 169. 172, 174, 175, 176, 190, 199, 200, 205, 220; 282 Blaxter, J.H.S., 415, 435, 473, 475; 496

Bliss, C.I., 82; 85 Block, G., 245; 272 Block, G.D., 245; 272 See Lickey, M.E., 278 Bloom, S.A., 121; 157 See Santos, S.L., 124; 162 Boaden, P.J.S., 127; 157 Bocquet-Védrine, J., 302, 307, 325, 328, 329, 333; 347, 348 See Pochon-Masson, J., 325; 350 Boekelheide, E.F. See Ainey, D.G., 34; 36 Boerema, L.K., 365; 496 Boesch, D.F., 124; 157 Boge, E. See Braaten, B., 565; 574 Boguchwal, L.A. See Southard, J.B., 103; 111 Bohlke, J.E., 509, 513; 561 Bohme, H. See Thomas, R.M., 436; 504 Bonsdorff, E., 138; 157 Boschma, H., 329; 348 Bossert, W.H. See Hazlett, B.A., 544, 545, 547; 561 Bostock, J., 167, 253; 272 Botero, L., 128, 130; 157 Botha, L., 355, 379, 380, 382, 386, 388, 389, 393, 413, 419, 420, 422, 423, 445, 446, 447, 448, 449, 451, 470, 489, 494; 496 Botazzi, F., 216; 272 Bouchon, C., 49; 85 Bourke, R.H. See Paquette, R.G., 19; 38 Bousfield, E.L., 141; 157 Bowen, J.T. See Willoughby, H., 566; 575 Bower, P. See Santschi, P.H., 111 Bowes, A. See Field, J.G., 499 Boxhall, G.A. See Lincoln, R.J., 47; 88 Boyd, A.J., 355, 370, 377, 436, 449, 474, 475; 496 See Badenhorst, A., 355; 496 See Shannon, L.V., 433; 503 See Shelton, P.A., 431, 473, 477, 491, 494; 504 See Thomas, R.M., 367, 491; 504 Boyer, L.F. See Grant, W.D., 100, 152; 111, 158 See Rhoads, D.C., 109; 111 Braams, W.G., 211; 272 Braaten, B., 565, 567, 572; 574 Bradbury, N.B. See Gowen, R.J., 563–575 Bradley, E.F. See Wooding, R.A., 153; 164 Bradley, J.V., 46; 85 Bradshaw, P., 92; 110 See Cebeci, T., 99; 110 Brainard, R.E. See McLain, D.R., 473, 494; 501 Branch, G.M., 226, 231, 234; 272

AUTHOR INDEX

Brandt, E., 326; 348 Brandt, R.R. See Palmer, M.A., 110, 144, 154; 111, 161 Brenchley, G.A., 124; 157 Bresciani, J., 345; 348 Breuer, J.P., 174, 175, 177; 272 Bridges, C.B., 189; 272 Broch, H., 320, 326, 338, 340, 341; 348 Brock, R.E., 49; 85 Broecker, W.S. See Santschi, P.H., 111 Bromley, P.J., 565; 574 Brooke, R.K. See Crawford, R.J.M., 483; 496 See Shelton, P.A., 494; 504 Brown, A.C., 353; 497 See Darracott, D.A., 353; 498 Brown, A.M., 245; 272 Brown, B.E., 489, 492; 497 Brown, G.H. See Bebbington, A., 176, 177; 272 Brown, H.M. See Brown, A.M., 245; 272 Brown, J.R., 567; 574 Brown, P. See Pettit, G.R., 280 Brown, R.G.B., 461; 497 Brownell, P.H. See Ligman, S.H., 192; 278 Brown-Leger, L.S. See Grassle, J.F., 159 Bruce, J.R., 177, 188; 272 Brundrit, G.B. See Shannon, L.V., 433; 503 Brunenmeister, S., 455; 497 Brzezinski, M.A., 31; 36 Buchanan, J.B., 120, 125; 157 Buck, K.R. See Ackley, S.F., 25; 36 See Garrison, D.L., 25, 32; 36 Buckland-Nicks, J. See Chia, F.-S., 114, 134, 148, 155; 157 Buckley, J.R., 14; 36 Bulgakova, T.I. See Babayan, V., 412, 413, 417; 496 Bulleid, N.C. See Tranter, D.J., 163 Bullock, A.M. See Jones, K.J., 574 See Turner, M.F., 564; 575 Bullock, T.H., 326; 348 Bunt, J.S., 32; 36 See Abel, D.J., 80; 85 Burger, A, E., 423, 494; 497 Burke, R.D., 115, 131, 132, 133, 134; 157 Burkle, L.H. See Palmisano, A.C., 37 Burks, J.E. See McDonald, F.J., 279 Burky, A.J., 197; 272 Burns, B.A. See Shuchman, R.A., 38 Burns, N.M. See Hargrave, B.T., 54, 135; 87, 159 Businger, J.A., 102; 110 Butman, B., 152; 157

523

Butman, C.A., 113–165; 110, 116, 130, 133, 135, 145, 146, 147, 148, 149, 150, 151; 110, 157 Butterworth, D.S., 355, 360, 363, 365, 367, 370, 371, 415, 435, 458, 461, 477, 486, 487, 491, 494, 495; 497 See Armstrong, M.J., 459, 460; 495 See Bergh, M.O., 465; 496 See Schülein, F.H., 368; 503 Butz, I., 565; 574 Buxton, C.D. See Smale, M.J., 400; 504 Buys, M.E.L. See Shannon, L.V., 355, 435, 490; 503 Buzas, M.A. See Young, D.K., 124; 164 Byrne, J.H., 254; 272 See Shapiro, E., 254; 281 See Tritt, S.H., 256; 283 Byun, D.S. See Cho, D.M., 220; 273 Cacchione, D.A., 147; 157 Caffey, H.M., 44, 63; 85 Caldwell, D.R. See Chriss, T.M., 147; 157 Caldwell, J.W., 132; 157 Callan, M.G., 344; 348 Calvin, J. See Ricketts, E.F., 174; 280 Cameron, R.A., 133; 157 Campbell, D.C. See McDonald, F.J., 297 Campbell, J.I. See Meadows, P.S., 115, 123, 126; 161 Campbell, R. See Tranter, D.J., 163 Campbell, R.D., 115; 157 Campbell, W.J. See Johannessen, O.M., 37 See Schuchman, R.A., 38 See Zwally, H.J., 38 Cannon, H.G., 326; 348 Capo, T.R., 200; 272 See Kandel, P., 189; 277 Carazzi, D., 190; 272 Carefoot, T.H., 167–284; 172, 173, 175, 176, 177, 178, 179, 180, 181, 182, 187, 188, 191, 193, 195, 196, 197, 200, 201, 203, 204, 205, 207, 209, 210, 211, 213, 214, 215, 216, 217, 218, 219, 220, 221, 222, 223, 225, 226, 227, 228, 232, 233, 234, 236, 237, 238, 239, 240, 241, 242, 243, 247; 272, 273 See Smith, S.T., 182; 281 Carew, T. See Advokat, C., 229, 259; 271 Carew, T.J., 254, 256, 257, 258, 259; 273 See Cook, D.G., 215; 273 See Hening, W.A.V 243; 276 See Kupfermann, I., 174, 179, 181, 182, 185, 186, 187, 192, 200, 204, 205, 206, 210, 214, 219, 220, 229, 244, 245, 250, 253; 278 See Pinsker, H.M., 258; 280

524

OCEANOGRAPHY AND MARINE BIOLOGY

See Walters, E.T., 244, 259; 283 Carlton J.T., 171; 273 Carlucci, A.F., 564; 574 Carra, P.O. See Rüdiger, W., 253; 280 Carriker, M.R., 142; 157 Carsey F.D. See Zwally, H.J., 38 Carter, D.C. See Schmitz, F.J., 281 Caruso, A.J. See Hoye, T.R., 276 Case, T.J. See Wiens, J.A., 44; 90 Cassue, R.M., 70, 72, 82, 120; 86, 157 Castellucci, V., 258; 273 See Kriegstein, A.R., 173, 197, 220; 278 See Kupfermann, I., 258; 278 See Pinsker, H., 257; 280 Castellucci, V.F. See Bailey, C.H., 201; 271 See Carew, T.J., 258; 273 Caton, A.E. See Winstanley, R.H., 408; 504 Caughley, G., 41, 42, 43, 49, 50, 51, 54, 56; 86 Cayre, P. See Le Loeuff, P., 457; 500 Cebeci, T., 99; 110 Centurier-Harris, O.M., 360, 363, 458, 477; 497 See Crawford, R.J.M., 363; 497 See Newman, G.G., 360, 458, 490; 502 Chalazonitis, N. See Arvanitaki, A., 245; 271 Chalmers, D.S., 379, 388, 390; 497 See Botha, L., 393, 451; 496 Chaloupka, M.Y., 80; 86 Chaney, A.H. See Tunnell Jr, J.W., 174, 177; 283 Chang, P.K., 93; 110 Chaplin, C.C.G. See Bohlke, J.E., 509, 513; 561 Chapman, A.R.O., 226; 273 Chapman, C.J. See Rice, A.L., 520; 562 Chapman, D.G. See Leatherwood, S., 49; 88 Chapman, D.J., 174, 193, 206, 212, 227, 253, 254, 255, 256; 273 Chapman, F.A. See Johnson, P.T., 270; 277 See Pauley, G.B., 269; 280 Chapman, G., 143, 144; 157 Chapman, P., 353, 431; 497 Chase, R., 183, 245; 273 Chen, M. See Bailey, C.H., 201; 271 Chen, R. See Ambrose III, H.W., 251; 271 Chess, J.R. See Hobson, E.S., 117, 118; 159 Chia, F.-S., 114, 115, 134, 148, 154, 155; 157 See Birkeland, C., 142; 157 Chlapowski, K., 379, 418, 421; 497 Cho, D.M., 220; 273 Choat, J.H., 77; 86 Chriss, T.M., 147; 157

Christomanos, A., 253; 273 Chung, M.C.H. See Schulte, G.R., 281 Church, M. See Nowell, A.R.M., 147, 153; 161 Cipolli, I.N., 270; 273 Clapham, A.R., 72, 75; 86 Clardy, J. See Faulkner, D.J., 262; 274 See Finer, J., 274 See Ireland, C., 276 See Kinnel, R.B., 277 Clark, E., 512, 513; 561 Clark, F.N., 490, 491; 497 Clark, K.B., 171, 176; 273 Clark, P.F. See Lincoln, R.J., 47; 88 Clark, P.J., 71, 75, 76, 77, 78, 79; 86 Clark, R.B., 75; 86 Clark, W.G. See Troadec, J.-P., 365, 461, 478; 504 Clarke, D.B., 32; 36 Clarke, R.T., See Irish, A.E., 44, 52, 57, 62; 87 Clauser, F.H., 146, 147; 157 Clutton-Brock, T., 344; 348 Cobbs, J.S., 174, 179, 180, 188, 192; 273 See Aspey, W.P., 229, 230, 248; 271 Cochran, W.G., 42, 44, 46, 47, 51, 52, 53, 56, 57, 59, 60, 61, 69, 70; 86 See Snedecor, G.W., 42, 43, 44, 45, 47, 52, 53, 57, 59, 61; 89 Codispoti, L.A. See Smith, S.L., 13, 19; 38 Coggeshall, R.E., 249, 270; 273 Cohen, J., 64, 66, 67, 68, 69; 86 Coker, G. See Pot, W., 48; 88 Colby, D.R., 74; 86 Colby, P.J., 568; 574 Cole, W.J. See Chapman, D.J., 254; 273 Coleman, N., 251; 273 See Cuff, W., 60, 61; 86 Coles, J., 102; 110 Colin, P.L., 517, 543; 561 Collette, B.B., 452, 454; 497 Collins, I., 565, 569; 574 Colman, J.S. See Bruce, J.R., 188; 272 Comiso, J.C. See Zwally, H.J., 38 Conant, R.C. See Steinberg, J.B., 545; 562 Connell, J.H., 76, 78, 113, 114, 119, 134; 86,157 Cook, D.G., 215; 273 Cook, E.F., 211; 273 Cooke, J.G. See Beddington, J.R., 460, 464; 496 Cooling, D.A. See Mackey, A.P., 49; 88 Cooney, R.T. See Niebauer, H.J., 16; 37 Cooper, J., 480; 497

AUTHOR INDEX

See Burger, A.E., 423, 494; 497 See Crawford, R.J.M., 483; 498 See Cruickshank, R.A., 443; 498 See Duffy, D.C., 461; 499 See Frost, P.G.H., 443, 494; 499 See Shelton, P.A., 494; 504 Cooper, J.G., 174; 273 Copley, N.J. See Grassle, J.F., 159 Cornubert, G., 336; 348 Corrent, G., 245; 273 Cosgrove, R.E. See Darling, S.D., 262; 273 Coutière, H., 327; 348 Cox, G.M. See Cochran, W.G., 42, 44, 47, 51, 52, 56; 86 Craigie, J.S., 262; 273 Craik, G.J.S. See Bell, J.D., 48; 85 Cram, D.L., 365, 367, 368, 371, 434, 461, 491; 497 See Schülein, F.H., 368; 503 Crawford, B.J. See Chia, F.-S., 154; 157 Crawford, R.J.M., 353–505; 355, 357, 360, 361, 363, 364, 365, 366, 367, 368, 369, 377, 378, 387, 389, 397, 398, 400, 401, 402, 415, 417, 418, 419, 420, 431, 433, 434, 435, 436, 438, 440, 441, 442, 443, 444, 445, 446, 448, 449, 458, 460, 461, 462, 463, 466, 472, 473, 474, 475, 478, 479, 480, 481, 483, 484, 486, 489, 490, 491, 492, 494, 495;497, 498 See Kinloch, M.A., 375; 500 See Newman, G.G., 360, 362, 365, 415, 458, 474, 487, 490, 491, 492; 502 See Shannon, L.V., 363, 473, 474, 491, 494; 503 See Shelton, P.A., 494; 504 Creaser Jr, E.P. See Graham, J.J., 118; 158 Crisp, D.J., 114, 115, 130, 131, 132, 133, 141, 285, 299, 301, 345, 346; 157, 348 Croker, R.A., 129. 130; 157 Crous, H.B. See Crawford, R.J.M., 400, 401, 402, 491, 492; 497 Cruickshank, R.A., 435, 441, 442, 443, 444, 445; 498 See Boyd, A.J., 355; 496 See Crawford, R.J.M., 418, 443, 480; 497 Crumb, S.E., 121; 157 Cuellar, O., 346; 348 Cuff, W., 60, 61; 86 Cuhel, R.H. See Holm-Hansen, O., 30; 37 Cummins, R.A., 507, 509, 512, 513, 514, 517, 519, 520, 522, 523, 524, 525, 526, 529, 532, 535, 538, 539, 556; 561 Cuomo, M. C, 127, 128, 131, 133; 157 Cushing, D.H., 353, 475; 498 Cyr, H. See Downing, J.A., 49, 52, 53, 56, 57, 58; 86

525

Czeczuga, B., 227; 273 Daanje, A., 560; 561 Dahiya, R.C. See Alden, R.W., 62; 85 Dallai, R. See Baccetti, B., 296; 347 Danin, Z. See Barash, A., 175, 176, 177, 190, 191; 271 Danise, B., 263, 267; 273 D’Arcangues, C., 418; 498 Darling, S.D., 262; 273 Da Rodda, D. See Miyake, P.M., 405; 501 Darracott, D.A., 353; 498 Darwin, C., 285, 292, 293, 302, 307, 310, 311, 312, 313, 314, 317, 318, 320, 321, 324, 343, 344; 348 Dauer, D.M., 118, 124, 125, 154; 157, 158 David, J.H.M. See Butterworth, D.S., 494; 497 Davies, D.H., 365, 366, 414, 415, 422, 423, 424, 445, 480; 498 Davies, P.A. See Sigurdsson, J.B., 144; 163 Davies, R., 113; 158 Davies, S.L., 434, 443, 461; 498 Davis G.E. See Warren, C.E., 238; 283 Dawson, E.Y. See Winkler, L.R., 174, 177, 191, 206, 207, 210, 211, 214, 223, 231; 284 Day, J.H., 121, 131, 132, 327, 334; 158, 348 Day, R., 115; 158 Dayton, P.K., 44, 72, 122, 123, 125, 135, 137, 298; 86, 158, 348 Dean, D., 117; 158 De Beer, Sir Gavin, 344; 348 De Campos Rosado, J.M., 355, 357, 371, 372, 373, 377, 404, 441; 498 De Decker, A.H.B. See Stander, G.H., 367, 372, 436, 437, 491; 504 De Freitas, J.C., 264; 273 De Jaeger, B., 403, 425, 426, 427, 428; 498 De Jager, B. van D., 403, 425, 426, 427, 428, 445, 453, 454, 455; 498 Delage, Y., 327, 331; 348 Delgado, A., 494; 498 Dellaria, Jr, J.F. See Hoye, T.R., 276 DeMartini, E.E., 49; 86 DeMaster, D.J., 33, 34; 36 See Ledford-Hoffman, P.A. 34; 37 De-Negri, A., 253; 273 De-Negri, G. See DeNegri, A., 253; 273 Denley, E.J. See Underwood, A.J., 44, 46; 89 Denman, K.L., 44; 86 See Platt, T., 81; 88 Dennison, J.M., 56; 86

526

OCEANOGRAPHY AND MARINE BIOLOGY

Denny, M., 239; 273 Denny, M.V., 109; 110 De Villiers, G., 353, 354, 356, 358, 373, 375, 377, 382, 385, 406, 407, 458, 462, 467, 469; 498 See Crawford, R.J.M., 377, 378, 387, 389, 397, 415, 438, 440, 445, 446, 449, 460, 472, 474, 475, 480, 482, 483, 492; 497 DeVlas, J. See Beukema, J.J., 118; 156 Diamond, J. See Wiens, J.A., 44; 90 Diaz, R.J. See Boesch, D.F., 124; 157 Diegaard, R. See Sumer, B.M., 91; 111 Dieringer, N., 179, 205, 209, 253; 273 Dieter, R.K., 263, 264; 273 See Kinnel, R.B., 277 Dietrich G., 17; 36 Diggle, P.J., 72, 75, 83; 86 Dijkgraaf, S., 249; 273 Dilip de Silva, E., 268; 274 Dillon, W.A., 329; 348 DiMatteo, T., 175, 251, 252, 253, 255, 256; 274 Dingle, H., 544, 545; 561 Djerassi, C. See Gunatilaka, A.A. L., 275 Dobbs, F.C., 119, 144, 154; 158 Doherty, P.J., 69; 86 Domack, E.W. See Dunbar, R.B., 34; 36 Doochin, H., 141; 158 Dörjes, J. See Howard, J.D., 143; 160 Douglas, W.A. See Sale, P.F., 48, 49, 53; 89 Downes, B.J. See Keough, M.J., 119; 160 Downey, P., 208; 274 Downing, J.A., 40, 47, 49, 51, 52, 53, 54, 56, 57, 58, 75; 86 Doyle, R.W., 133; 158 See Todd, C.D., 154; 163 Draganik, B., 375, 417; 498 See Chlapowski, K., 379; 497 Drake, D.E. See Cacchione, D.A., 147; 157 Drewry, D.J. See Jacobs, S.S., 13; 37 Droop, M. See Tett, P., 571; 575 Droop, M.R., 571; 574 Dudek, F.E., 182, 192; 274 See Pinsker, H.M., 192; 280 See Rothman, B.S., 192; 280 Dudek, L.E. See Sigel, M.M., 281 Dudley, S.F.J., 443; 498 Dudnik, Y.I., 443; 498 Duffy, D.C., 461, 486; 499 See Shannon, L.V., 363, 473, 474, 491, 494; 503 Dugdale, R.C., 16, 31, 564; 36, 574

Duggan, A.J. See Kinnel, R.B., 277 Duhamel, G. See Bhaud, M., 118, 137; 157 Dunbar, R.B., 34; 36 Duncan, P.B. See Peterson, C, H., 143; 162 Dunstan, W.M. See Ryther, J.H., 564; 574 Du Plessis, C.G., 360, 365; 499 Durand, D., 334; 348 Dyer, K.R., 102; 110 Dyrssen, D. See Anderson, L., 20; 36 Eagle, R.A., 121; 158 Eales, N.B., 167, 168, 169, 171, 172, 174, 175, 176, 177, 180, 183, 187, 191, 202, 203, 207, 208, 210, 215, 226, 234, 247, 256, 270; 274 See Engel, H., 175, 244; 274 Eberhardt, L.L., 51; 86 Ebert, T.A., 71; 86 Eckelbarger, K.J., 115, 127, 132; 158 Eckelmann, H., 147; 158 Eckman, J.E., 103, 107, 108, 123, 137, 139, 140, 143, 153; 110, 158 See Jumars, P.A., 140; 160 See Nowell, A.R.M., 91, 105, 152; 111, 161 Edelsten, D.J. See Edwards, A., 570; 574 Edmunds, M., 211; 274 Edmundson, C.H., 175; 274 Edwards, A., 570; 574 See Tett, P., 571; 575 Edwards, D.C., 217, 236; 274 See Huebner, J.D., 236; 276 Edwards, D.J., 565, 570; 574 Einck, J.J. See Pettit, G.R., 280 Eisner, T. See Dieter, R.K., 263; 273 See Kinnel, R.B., 277 Elliott, J.M., 41, 42, 43, 46, 51, 52, 53, 56, 72, 73, 74, 82; 86 El-Sayed, S.Z., 13, 14, 24, 25, 30; 36 See Holm-Hansen, O., 30; 37 Emery, A.R., 117; 158 Emery, D.G., 208; 274 Emlen, J.M., 214; 274 Endo, Y. See Fujiki, H., 275 Enell, M., 566, 567, 570, 572; 574 Engel, H., 174, 175, 205, 244; 274 Eppley, R.W., 11, 16, 31; 36 Erickson, K.L., 262; 274 Ericksson, B. See Olsson, I., 140; 161 Errington, J.C., 81; 86 Ervic, A. See Braaten, B., 565; 574

AUTHOR INDEX

Eskin, A., 245; 274 See Current, G., 245; 273 Esumi-Kurisu, M. See Yamazaki, M., 284 Evans, F.C. See Clark, P.J., 71, 75, 76, 77, 78, 79; 85 Everson, I., 34; 36 Ewing, R.M. See Dauer, D.M., 118; 157 Fage, L., 117; 158 Fager, E.W., 126, 142; 158 Fairweather, P.G., 48; 86 Farke, H., 143, 144, 154; 158 Farmer, W.M., 247; 274 Farrelly, B.A. See Johannessen, O.M., 37 Farrow, G.E., 514, 518, 520, 522, 523, 526, 532; 561 Fasham, M.J.R., 81, 82; 86 Fattorusso, E., 263; 274 See Danise, B., 273 Fauchald, K. See Thistle, D., 123; 163 Faulkner, D.J., 261, 262, 263, 265; 274 See Higgs, M.D., 263; 276 See Ireland, C., 276 See Mynderse, J.S., 262; 279 See Stallard, M.O., 260, 262, 263, 265; 281 Faure, A. See Querellou, J., 575 Faure, C. See Querellou, J., 575 Fayette, A.R., 567; 574 Fayos, J. See Faulkner, D.J., 262; 274 Feinstein, R., 179; 274 See Pinsker, H.M., 179; 280 Feldman, E. See Susswein, A.J., 180; 282 Fell, P.E., 115; 158 Fenchel, T.M., 564; 574 Fenical, W., 251, 262, 263, 268, 269; 274 See Finer, J., 274 See Hirschfeld, D.R., 276 See Kinnel, R.B., 277 See Sims, J.J., 263; 281 See Stallard, M.O., 281 Ferguson, G.P., 182, 191; 274 Ferguson, M.M. See Pot, W., 48; 88 Fernandez, M.A.R. See Delgado, A., 494; 498 Feuerborn, H.T., 332; 348 Field, J.G., 425; 499 See Bergh, M.O., 415, 466; 496 See Day, J.H., 121; 158 See Dudley, S.F.J., 443; 498 See Moloney, C.L., 466, 490; 501 See Shannon, L.V., 458, 490; 503 Findlay, R.H. See Thistle, D., 140; 163

527

Findlay, S.E.G., 74; 86 Finer, J., 267, 269; 274 See Ireland, C., 276 Fishelson, L., 543; 561 Fisher, R.A., 72; 86 Fleming, R.H., 19; 36 See Sverdrup, H.U., 19; 38 Flint, R.W., 121; 158 Florinskaya, E.S. See Shcherbich, L.V., 417; 503 Flury, F., 253, 255, 256, 257; 274 Fong, W., 225; 274 Fonseca, M.S. See Colby, D.R., 74; 86 Fontaine, M., 253; 274 Ford, E., 120; 158 Foreman, R.E., 226; 274 Fosshagen, A., 134; 158 Foster, B.A., 299, 301, 302, 320, 338, 340, 341, 345; 348 Foster, T.D., 13; 36 Fox, D.L. See Chapman, D.J., 174, 193, 206, 212, 227, 253, 254, 255, 256; 273 Fox, R.S., 174; 274 Fox, W.W. See Wise, J.P., 472; 504 Franceschini, G.A. See Holm-Hansen, O., 30; 37 Francis, J.R.D., 92; 110 Franzén, A., 325; 348 Fraser, W.R., 34; 36 Frazer, F.J. See McLean, W.E., 566; 574 Fredman, S.M. See Jahan-Parwar, B., 209, 244; 276, 277 See Wells, L.J., 208; 283 Freedman, L. See Lukowiak, K., 259; 279 Frentz, R., 336; 348 Freon, P., 473, 474, 494; 499 Frings, C. See Frings, H., 173, 208, 209, 211, 217, 228, 244; 274 Frings, H., 173, 208, 209, 211, 217, 228, 244; 274 Frost, P.G.H., 443, 494; 499 Fuchs, N.A., 108; 110 Fujiki, H., 261; 275 Fujimaki, N. See Yanagimachi, R., 333; 351 Fukazawa, N. See Takahashi, M., 568; 575 Fukushima, H. See Meguro, 32; 37 Gaffney, J.J. See Walsh, J.J., 38 Gage, J., 72, 73, 74, 75, 121, 140; 86, 158 Gagnon, M., 75; 86 Galicka, W. See Penczak, T., 575 Gallager, S.M., 220; 275 Gallagher, E.D., 123, 124, 138, 139; 158 Gallin, E.K., 249; 275

528

OCEANOGRAPHY AND MARINE BIOLOGY

Gammelsrød, R. See Buckley, J.R., 36 Gammelsrød, T., 14; 36 Gaonkar, S.N., 320; 348 Gärdefors, D., 82, 140; 86, 158 Gardiner, F.P., 80; 86 Gardner, E.A. See Switzer-Dunlap, M., 183; 282 Gardner, W.D., 135; 158 Gargett, A.E. See Denman, K.L., 44; 86 Gargus, J.L. See Sigel, M.M., 281 Garrison, D.L., 13, 25, 32; 36 Garrison, R.L. See McIntyre, C.D., 100; 111 Garstang, W., 202, 226; 275 Gascard, J.C. See Manley, T.O., 37 Geekie, A.D. See Gage, J., 72, 73, 74, 75, 121, 140; 86, 158 Geelen, H.F.M. See Braams, W.G., 211; 272 Geldenhuys, N.D., 360, 362, 375, 412, 413, 440, 441, 443, 462; 449 Gerard, G., 56; 86 Gerencser, G.A., 221; 275 Gertenbach, L.F.D., 360, 364; 499 Gerritsen, J., 346, 347; 348 Getz, W.M., 477; 499 Gev, S., 175, 176, 181, 182, 190, 191, 201, 202, 207; 275 See Susswein, A.J., 173, 180; 282 Gewali, M.B. See Ronald, R.C., 260; 280 Ghiretti, F., 206, 207, 224; 275 Ghiretti-Magaldi, A. See Ghiretti, F., 224; 275 Ghiselin, M.T., 344, 345, 346, 347; 348 Giard, A., 327, 336; 348 Gibbs, P.E., 121; 158 Gibson, P.H., 132; 158 Gilbert, G..K., 91; 111 Gilbert, J.R. See Leatherwood, S., 49; 88 Gilboa-Garber, N., 270; 275 Gilchrist, J.D.F., 402; 499 Gilliam, J.J. See Lu, C., 245; 278 Givens, R.P. See Ambrose III, H.W., 251; 271 Gjosaeter, J., 465; 499 Glenn, S.M. See Grant, W.D., 147, 152; 158, 159 Glibert, P.M., 26; 36 Gloersen, P. See Zwally, H.J., 38 Goering, J.J. See Dugdale, R.C., 16, 31; 36 See McRoy, C.P., 13, 16; 37 See Sambrotto, R.N., 16; 38 Gold, P. See Tobach, E., 254; 283 Goldhaber, M.B. See Rhoads, D.C., 124; 162 Goldman, B. See Russell, B.C., 48; 88 Gomez, E.D., 345; 348

González, A.G., 267; 275 Goodall, D. See Ludwig, J.A., 81; 88 Goodall, D.W., 82; 87 Gooden, B.A. See Feinstein, R., 179; 274 See Pinsker, H.M., 179; 280 Goosen, P.C. See Pollock, D.E., 456; 503 Gopichand, Y., 266; 275 See Gunatilaka, A.A.L., 275 See Schmitz, F.J., 281 Gordon, A.L. See Jacobs, S.S., 13; 37 Gordon, L.I. See Nelson, D.M., 37 Goren, M., 507, 514; 561 Goudsmit, E.M. See Sanders, H.L., 120; 162 Gould, S.J., 320; 348 Gove, P.B., 47; 87 Gowen, R.J., 563–575; 568, 571; 574 See Brown, J.R., 567; 574 Graham, J.J., 118; 158 Grahame, J., 197; 275 Granger, G.A. See Pauley, G.B., 269; 280 Grant, J., 144, 153, 154; 158 See Muschenheim, D.K., 92, 105; 111 Grant W.D., 100, 104, 114, 145, 147, 152; 111,158, 159 See Butman, C.A., 110, 135; 110, 157 Grant, W.S., 355; 499 Grass, A.J., 91; 111 Grassle, J.F., 82, 123, 124, 136, 140; 87, 159 See Grassle, J.P., 135; 159 See Maciolek, N.J., 121, 123; 160 Grassle J.P., 132, 135; 159 See Grassle, J.F., 124, 136; 159 Gray, C.A., 49; 87 Gray, J.S., 53, 56, 70, 115, 121, 122, 123, 125, 127, 128, 129, 130, 132; 87, 159 See Hulings, N.C., 53; 87 Green, R.H., 41, 42, 44, 45, 46, 47, 51, 52, 53, 54, 56, 60, 68, 69, 73, 74; 87 Greenblatt, P.R. See Levin, L.A., 118, 132; 160 Greig-Smith, P., 53, 72, 74, 81; 87 Grenfell, T.C. See Maykut, G.A., 14; 37 Griffith, R.W. See Lombardini, J.B., 178; 278 Griffiths, C.L., 429; 499 See Barkai, R., 430; 496 See Field, J.G., 499 Griffiths, R.J. See Field, J.G., 499 Grigg, U.M., 175, 177, 201, 203, 226; 275 Gross, T.F., 147, 152; 159 Gruenig, D.E. See Craigie, J.S., 262; 273

AUTHOR INDEX

Gruvel, A., 303, 308, 310, 311, 312, 313, 314, 315, 316, 317, 320, 326, 340; 348, 349 Guérin, J.-P., 115, 136; 159 See Massé, H., 136; 160 Gulland, J.A. See Troadec, J.-P., 365, 461, 478; 504 Gunatilaka, A.A.L., 266; 275 Günther, A., 400; 499 Gurvis, R. See Arch, S., 192; 271 Gust, D. See Pettit, G.R., 280 Gust, G. See Palmer, M.A., 119, 144, 154; 161 See Paola, C., 103; 111 Guy, H.P., 91; 111 Hackney, A.G., 177; 275 Hadfield, M.G., 115, 155, 187, 188, 189, 192, 198, 199, 200, 231; 159, 275 See Kempf, S.C., 132; 160 See Miller, S.E., 155; 161 See Switzer-Dunlap, M., 173, 180, 181, 188, 189, 190, 191, 192, 193, 194, 197, 198, 200,205, 207, 208, 211, 213, 220, 231, 243; 282 Hadl, G., 129, 130; 159 Haedrich, R.L. See Barber, R.T., 417; 496 See Gardiner, F.P., 80; 86 Haefelfinger, H.R., 247; 275 Haegele, C.W. See Schweigert, J.F., 62; 89 Hagerman Jr, G.M., 118, 144, 154; 159 Haines, T.A. See Fayette, A.R., 567; 574 Hairston, N.G., 73; 87 Hakii, H. See Fujiki, H., 275 Hakkinen, S., 15, 21; 36 Hall, D.N. See Chaloupka, M.Y., 80; 86 Halstead, B.W., 253, 256; 275 Halver, J.E., 564; 574 Hamatani, I. See Baba, K., 271 Hameedi, M.J., 19; 36 Hamilton, P.V., 171, 173, 174, 203, 229, 248, 249, 250, 253; 275 Hammer, R.M., 118; 159 Hamner, W.M. See Omori, M., 48, 70, 83; 88 Hampson, G.E. See Sanders, H.L., 120; 162 Hampton, I., 435, 458, 459, 460; 499 See Cruickshank, R.A., 443; 498 Hankin, D.G., 62, 63; 87 Hannan, C.A., 115, 119, 123, 134, 135, 137, 138, 141, 144, 150, 153; 159 Harada, E., 507, 512, 514, 517, 520, 522, 523, 526, 540, 548, 552, 559; 561 Harada, K. See Noriki, S., 33; 37

529

Harbottle, G. See Walsh, J.J., 38 Hardinge. M.G. See Winkler, L.R., 260; 284 Hardy, Sir A., 202; 276 Hargrave, B.T., 54, 135; 87, 159 Harker, J. See Keast, A., 50; 87 Harris, C.C. See Willey, J.C, 261; 283 Harris, G.P., 44; 87 Harris, J.C.K. See Winstanley, R.H., 408; 504 Harry-Rofen, R.P. See Bayer, F.M., 509, 526, 527; 561 Hart, T.J., 13, 24; 36 Hartnoll, R.G., 334; 349 Harvey, P. See Clutton-Brock, T., 344; 348 Harvey, P.H., 78; 87 Hatanaka, H., 381, 411, 417, 425, 440, 441, 449, 450, 458, 464, 467, 469; 499 See Uozumi, Y., 411; 504 Haury, L.R., 48, 50; 87 Hawkins, R.D. See Carew, T.J., 258; 273 Hay, W.W. See Dennison, J.M., 56; 86 Hayakawa, Y. See Irie, T., 262; 276 Hayasi, S., 454; 499 Hayward, P.J. See Harvey, P.H., 78; 87 Hazlett, B.A., 544, 545, 547; 561 Heck Jr, K.L., 252; 276 Hecht, T., 417, 419, 422, 423, 450, 489; 499 Heggarty, D. See Fleming, R.H., 19; 36 Heimdal, B.R., 20; 37 Heine, D. See Pinsker, H., 280 Heisig, D.M., 60; 87 Helshe, J.F., 53, 72, 75; 87 Henderson, F.M., 94; 111 Henderson, J.A., 132; 159 Hening, W.A., 243, 244; 276 See Pinsker, H.M., 253; 280 Hennemuth, R.C. See Brown, B.E., 497 Henry, D.P., 298; 349 See McLaughlin, P.A., 298, 301; 349 hEocha, C.O. See Rüdiger, W., 253; 280 Herald, C.L. See Pettit, G.R., 280 Herald, E.S., 527, 541; 561 Hermans, C.O., 136; 159 See Schroeder, P.C., 115; 163 Herriges, K. See Lederhendler, I.I., 182; 278 Herrman, K., 151, 155; 159 Herrnkind, W.F., 430; 499 Hess, F.R. See Bartz, R., 156 Hessels, H.G.A. See Dijkgraaf, S., 249; 273 Hessler, R.R. See Grassle, J.F., 159 Hewitson J.D. See Boyd, A.J., 355, 370; 496

530

OCEANOGRAPHY AND MARINE BIOLOGY

See Thomas, R.M., 436; 504 Heydorn, A.E.F., 429, 430; 499 Hiaasen, S.O. See Jordan, W.P., 245; 277 See Lickey, M.E., 245; 278 Hiegel, M.H. See Holland, A.F., 159 Higgins, H.W. See Tranter, D.J., 163 Higgs, M.D., 263; 276 Highsmith, R.C., 131, 133; 159 Hildebrand, S.F. See Longley, W.H., 507, 513; 561 Hill, H.B. See Webb, J.E., 129, 130; 164 Hill, O., 80; 87 Hill, R.B. See Wells, D.W., 175, 216; 283 Hill, R.W. See Hairston, N.G., 73; 87 Hinze, J.O., 101; 111 Hirata, Y. See Matsuda, H., 260; 279 See Yamada, K., 260; 284 See Yamamura, S., 260, 261, 262, 267, 268; 284 Hiro, F., 320; 349 Hirsch, H.R., 201, 202; 276 Hirshfeld, D.R., 263, 266, 267; 276 Hisai, K. See Baba, K., 271 Hjorth, I, 560; 561 Hobson, E.S., 117, 118; 159 Høeg, J.T., 328, 329, 331, 332, 333, 334, 335; 349 See Ritchie, L.E., 333; 350 Hoek, P.P.C., 308, 309, 310, 316, 317, 319, 344; 349 Hoenig, J.M. See Heisig, D.M., 60; 87 Hoese, D.F., 507, 508, 509, 512, 513, 514, 526, 551; 561 Hogg, J.J. See Bergquist, P.R., 115; 156 Hogue, E.W., 74, 139, 143, 144, 153, 154; 87, 159 Holdsworth, G. See Jacobs, S.S., 13; 37 Holland, A.F., 124, 125; 159 See Mountford, N.K., 121; 161 Holland, J.S. See Flint, R.W., 121; 158 Holland, W.R. See Wiebe, P.H., 48, 50, 51, 53, 55, 56; 90 Hollenbeak, K.H., 266; 276 See Kaul, P.N., 277 See Schmitz, F.J., 262; 281 Hollister, C.D. See Nowell, A.R.M., 109; 111 Holm, T.R. See Statzner, B., 92; 111 Holme, N.A., 43, 52, 53, 71, 72, 120; 87, 160 Holm-Hansen, O., 30; 37 See El-Sayed, S.Z., 14, 30; 36 See Neori, A., 19; 37 See Rönner, U., 26; 37 See Sakshaug, E., 31; 37 Hong, S.K. See Gerencser, G.A., 221; 275 Honjo, S. See Wefer, G., 33; 38 Hopkins, T.S. See Smith, J.D., 151; 163

Horner, R. See Matheke, G.E.M., 32; 37 Horner, R.A., 32; 37 Horridge, G.A. See Bullock, T.H., 326; 348 Hoshino, H. See Fujiki, H., 275 Hossain, M.B. See Hollenbeak, K.H., 276 See Schmitz, F.J., 281 Howard, J.D., 143; 160 Howells H.H., 215, 220; 276 Howieson, D.B. See Peretz, B., 258; 280 Hoye, T.R., 266; 276 Hsii, K.J., 172; 276 Hubbe, M.A., 105; 111 Huber, B. See Nelson, D.M., 37 Hudson, D.J. See Block, G.D., 245; 272 See Lickey, M.E., 245; 278 Huebner, J.D., 236; 276 See Edwards, D.C., 217, 236; 274 Hughes, G.M., 206; 276 See Bebbington, A., 175, 243, 247, 248; 272 Hughes, H.P.I. See Marcus, E.d.B.-R., 175, 176; 279 Hughes, R.N., 121, 197, 234; 160, 276 Hui, E., 299, 301; 349 Hulberg, L.W., 124; 160 See Oliver, J.S., 135; 161 Hulings, N.C., 53; 87 Hunkins, K.L. See Manley, T.O., 37 Hunsaker II, D. See Koningsor Jr, R.L., 221; 277 Hunt, A.R., 202; 276 Hunter, J.R. See Blaxter, J.H.S., 415, 435, 473, 475; 496 Hunter, V.D. See Bloom, S.A., 121; 157 Huppert, H.E. See Jacobs, S.S., 13; 37 Hurlbert, S.H., 44, 46, 54, 61, 68, 83; 87 Hurley, D.A. See Colby, P.J., 568; 574 Husby, D.M. See Parrish, R.H., 353; 502 Hustveit, H. See Bergheim, A., 564; 574 Hutchings, L. See Crawford, R.J.M., 357, 361, 368, 419, 434, 435, 436, 460, 461, 463, 466, 472, 473, 474, 475, 480, 481, 483, 491, 492; 498 See Shannon, L.V., 433; 503 See Shelton, P.A., 436; 504 Hutchinson, G.E., 478; 499 Huttrer, C. See Lederer, E., 253, 254; 278 Hyman, L.H., 256; 276 ICCAT, 454, 455, 472; 499 Ichikawa, A., 331, 333, 334, 345; 349 ICSEAF, 358, 375, 379, 382, 383, 387, 423, 449, 465, 467, 468, 484; 499 Idelman, S., 296; 349

AUTHOR INDEX

Ikeda, I., 381; 499 Ikegami, K. See Fujiki, H., 275 Ikenami, M. See Yamazaki, M., 284 Imai, N. See Irie, T., 276 Imamura, P.M., 268; 276 Impelman, D. See Wachtel, H., 243, 244; 283 Imperato, F., 263, 267; 276 Ina, K. See Katayama, A. 277 See Sakata, K., 281 Inada, T., 445, 449, 468; 499 Injeyan, H.S. See Dudek, F.E., 274 Inte’s, A. See Le Loeuff, P., 457; 500 Ireland, C., 265; 276 See Faulkner, D.J., 262; 274 Irie, T., 262, 263; 276 Irish, A.E., 44, 52, 57, 62; 87 Isarev, A.T., 384, 413, 449, 450, 468, 470; 499 Ito, K. See Meguro, H., 32; 37 Jaccarini, V. See Schembri, P.J., 559; 562 Jacklet, J.W., 244, 245, 246, 249; 276 See Lukowiak, K., 245, 258; 279 See Peretz, B., 258; 280 See Strumwasser, F., 192; 282 Jackson, G.A., 132; 160 Jackson, J.C.B. See Woodin, S.A., 114, 125; 164 Jackson, J.F. See Scheller, R.H., 281 Jacobs, S.S., 13; 37 See Ainley, D.G., 13, 30; 36 Jahan-Parvar, B., 208, 215; 276 Jahan-Parwar, B., 182, 205, 208, 209, 244, 248; 276, 277 See Downey, P., 208; 274 See Wells, L.J., 208; 283 Janse, C., 249; 277 Jansson, B.-O., 128, 130; 160 Jarman, N.G. See Field, J.G., 499 Jeannin, P.F. See Manley, T.O., 37 Jelley, E. See Thomas, M.L.H., 117; 163 Jenkin, C.R. See McKay, D., 269; 279 Jensen, P., 129, 130; 160 Jeuniaux, C. See Arnould, C., 215; 271 Johannessen, J.A. See Buckley, J.R., 36 See Johannessen, O.M., 37 See Smith, D.C., 15; 38 Johannessen, O.M., 14, 15, 20, 21; 37 See Buckley, J.R., 36 See Schuchman, R.A., 38 Johnson, M.W. See Sverdrup, H.U., 19; 38 Johnson, P.T., 270; 277

531

Johnson, R.B., 71, 77; 87 Johnson, R.G., 121; 160 See Whitlatch, R.B., 140; 164 Johnson, R.M. See Gray, J.S., 129, 130; 159 Jones, B.W., 379, 380, 382, 467; 499 Jones, D.A., 129, 130; 160 Jones, J.G., 49; 87 Jones, K.J., 568; 574 See Gowen, R.J., 568; 574 See Tett, P., 571; 575 Jones, M.L., 140; 160 See Jumars, P.A., 80; 87 Jones, N.S. See Bruce, J.R., 188; 272 Jonsson, I.G., 104, 105; 111 Jordan, H., 206, 207, 211, 244; 277 Jordan, W.P., 245; 277 Jørgensen, B.B., 92; 111 Josberger, E. See Johannessen, O.M., 37 Josberger, E.G. See Schuchman, R.A., 38 Jumars, P.A., 71, 72, 74, 80, 81, 92, 102, 103, 122, 123, 127, 140; 87, 111, 160 See Gallagher, E.D., 123, 124, 138, 139; 158 See Miller, D.C., 106, 110; 111 See Nowell, A.R.M., 91–112; 91, 92, 105, 114, 139, 141, 144, 146, 152; 111, 161 See Taghon, G.L., 105, 110; 112 Kalinina, M.I. See Assorov, V.V., 418, 421; 496 Kalk, M. See Macnae, W., 509, 512, 513, 514, 516, 517, 522; 562 Kalle, K. See Dietrich, G., 17; 36 Kamiya, H., 270; 277 See Yamazaki, M., 284 Kamiya, Y. See Sakata, K., 281 Kamykowski, D., 31; 37 Kandel, E. See Advokat, C., 229, 259; 271 See Kupfermann, I., 258; 278 See Pinsker, H., 257; 280 Kandel, E.R., 168, 226, 247, 252, 256; 277 See Carew, T.J., 254, 256, 257, 258; 273 See Castellucci, V., 258; 273 See Hening, W.A., 243; 276 See Kriegstein, A.R., 173, 197, 220; 278 See Kupfermann, I., 257, 258; 278 See Pinsker, H.M., 258; 280 See Scheller, R.H., 281 See Walters, E.T., 244, 259; 283 Kandel, P., 189; 277 Kao, J.P.Y. See Pettit, G.R., 280

532

OCEANOGRAPHY AND MARINE BIOLOGY

See Van Dreele, R.B., 266; 283 Kapuscinski, R. See Kirchman, D., 57; 88 Karande, A.A., 320; 349 See Gaonkar, S.N., 320; 348 Karplus, I., 507–562; 507, 508, 509, 511, 512, 513, 514, 515, 516, 517, 518, 519, 520, 522, 523, 524, 525, 526, 527, 528, 529, 530, 531, 532, 533, 536, 537, 538, 540, 541, 542, 543, 544, 546, 547, 548, 549, 550, 552, 553, 554, 555, 556, 557, 558; 561 See Goren, M., 507, 514; 561 Kashiwagi, M. See Mynderse, J.S., 261; 279 Katayama, A., 268; 277 Kathirgamatamby, N., 72; 87 Kato, Y., 261, 262, 267; 277 Kaul, P.N., 265; 277 Kaumeyer, K.R. See Holland, A.F., 159 Kawaguchi, K. See Gjosaeter, J., 465; 499 Kawasaki, T., 473; 499 Kay, E.A., 171, 172, 173, 174, 176, 177, 190, 191, 203, 247, 251, 256; 277 Kay, I. See Current, G., 245; 273 Kazlauskas, R., 263; 277 Keast, A., 50; 87 Keck, R., 128; 160 Keller, A.A. See Oviatt, C.A., 568; 575 Kellogg, T.B. See Truesdale, R.S., 34; 38 Kelner, K.L. See Nagle, G.T., 192; 279 Kempf, S.C., 132, 198, 199, 205; 160, 277 Kenchington, R.A., 56; 87 Kendziorek, M. See Anderson, D.J., 77, 78; 85 Kennedy, G.Y., 227; 277 Kennedy, V.S., 116; 160 Kennelly, S.J., 48, 49, 50, 62, 63, 70; 87 Keough, M.J., 119; 160 Kershaw, K.A., 81; 87 Kikuchi, T. See Tsutsumi, H., 567; 575 Kikawa, S., 454, 455; 500 Kim, C.Y. See Cho, D.M., 220; 273 Kim, K.S. See Bevelaqua, F.A., 269; 272 Kimmel, J.J., 49; 87 Kimura, K. See Yamazaki, M., 282 King, D.P.F., 355, 414, 435, 436, 483, 484, 489; 500 See O’Toole, M.J., 443; 502 King, J.M. See Alldredge, A.L., 117, 118; 156 Kingston, P.F. See Buchanan, J.B., 125; 157 Kinloch, M.A., 358, 375, 376, 417, 440, 464, 466; 500 Kinnel, R., 264; 277 See Dieter, R.K., 263; 273 Kinnel, R.B., 251, 264; 277

Kinner, P.C. See Watling, L., 53; 89 Kinzie, R.A., 48, 50, 78; 87 Kirchman, D., 57, 62; 88 Kirkup, M. See Finer, J., 274 Kisker, D.S. See Lie, U., 121; 160 Kisugi, J. See Yamazaki, M., 284 Kittelsen, A. See Bergheim, A., 564; 574 Kitting, C.L. See Seapy, R.R., 71; 89 Kittredge, J.S., 256; 277 See Blankenship, J.E., 261; 272 See Stallard, M.O., 281 Klauser, M.D., 129, 130; 160 Klausewitz, W., 507, 509, 512, 513, 514, 516, 532, 537; 561 Klepal, W., 285–351; 292, 320, 321, 323, 324, 325, 326, 337, 338, 344; 349 See Achituv, Y., 320, 337; 347 See Barnes, H., 325; 347 Koch, U.T., 209; 277 See Weiss, K.R., 283 Koester, J. See Bailey, C.H., 201; 271 See Dieringer, N., 179, 205, 253; 273 See Koch, U.T., 209; 277 See Shapiro, E., 254; 281 See Weiss, K.R., 283 Kofoed, L.H., 239; 277 Kohn, A.J., 208, 209; 277 Kolarov, P. See Babayan, V., 496 Komano, H. See Yamazaki, M., 284 Komar, P.D. See Miller, M.C., 152; 161 Komarov, Y., 471; 500 Komarov, Y.A. See Babayan, V., 496 See Shcherbich, L.V., 417; 503 See Zenkin, V.S., 375; 505 Kompowski, A., 415, 416; 500 Konchina, Y.V., 374, 379, 417, 418, 420, 422, 441, 449, 477, 485, 489; 500 Kondo, K., 473; 500 Koningsor Jr, R.L., 221; 277 Konishi, R., 15, 16; 37 Kono, H., 413, 447; 500 Konstantinova, M.I., 155; 160 Korzeniewski, J. See Korzeniewski, K., 568; 574 Korzeniewski, K., 568; 574 Kosler, A., 72, 75; 88 Koslow, J.A., 415, 492; 500 Kothbauer, H. See Hadl, G., 130; 159

AUTHOR INDEX

Krakauer, J.M., 169, 173, 174, 179, 180, 181, 190, 192, 194, 200, 201, 202, 203, 205, 206, 231, 232, 236, 238, 248, 250, 251, 252, 255, 256; 277, 278 Krassner, S.M. See Pauley, G.B., 269; 280 Krauhs, J.M., 270; 278 Krauss, W. See Dietrich, G., 17; 36 Kreger, D., 141; 160 Kress, A. See Haefelfinger, H.R. 247; 275 Kriedemann, B.D. See Crawford, R.J.M., 363; 497 Kriegstein, A.R., 173, 188, 190, 197, 198, 199, 200, 205, 220, 231, 232; 278 Kriel, F. See Crawford, R.J.M., 355, 480, 484, 490; 497 Kristensen, I., 142, 144; 160 Kruger, I. See Boyd, A.J., 370; 496 See Crawford, R.J.M., 418, 443, 480; 497 See Le Clus, F., 436; 500 Krüger, P., 302, 307, 312, 313; 349 Krzeptowski, M., 415, 416, 417, 484, 489; 500 Kuderskaya, R.A., 387, 413, 418, 425, 451, 470; 500 Kühnert, L., 296; 349 Kulkarni, S.K. See Kaul, P.N., 265; 277 Kumarasiri, M.H. See Bevelaqua, F.A., 269; 272 Kupfermann, I., 174, 178, 179, 181, 182, 185, 186, 187, 189, 190, 192, 200, 204, 205, 206, 209, 210, 214, 215, 219, 220, 229, 244, 245, 246, 250, 254, 257, 258; 278 See Carew, T.J., 254, 256, 259; 273 See Castellucci, V., 258; 273 See Kuslansky, B., 219, 220; 278 See Pinsker, H., 257; 280 See Rosen, S.C., 209; 280 See Susswein, A., 209; 282 See Susswein, A.J., 205, 209, 214, 215, 219, 220, 228, 229; 282 See Weiss, K.R., 283 Kurosawa, E. See Irie, T., 262; 276 See Kaul, P.N., 277 Kurth, M.J. See Hoye, T.R., 276 Kuslansky, B., 219, 220; 278 Kusto, E. See Penczak, T., 575 LaBarbara, M. See Vogel, S., 91, 105; 112 Lacalli, T.C., 132; 160 Lacroix, G. See Gagnon, M., 75; 86 Laevastu, T., 453, 489; 500 Lafedes, R.N., 47; 88 Laloy, M.L., 344; 349 Lance, J.R., 172, 176, 177, 189, 191; 278 Lande, R., 343; 349 Langlais, P.J., 260, 265; 278

533

See Blankenship, J.E., 261; 272 Lannelongue, N. See Shuchman, R.A., 38 Larkin, P.A. See Parker, R.R., 412; 502 Larkin, S., 243; 278 Larkins, H.A. See Laevastu, T., 489; 500 Larsen P.F., 121; 160 Larson, N. See Willoughby, H., 566; 575 Laskaridou, P. See Rees, E.I.S., 124; 162 Lasker, R., 492; 500 See Kittredge, J.S., 256; 277 Laverack, M.S., 209; 278 Lawrence, J.M., 226; 278 Laws, E.M., 94; 111 Lazarus, B.I., 456; 500 Leahy, W.M. See Sawaya, P., 188, 193, 204, 205, 206, 211, 251, 255, 270; 281 Leatherwood, S., 49; 88 Leberau, A., 454; 500 Le Clus, F., 370, 435, 436, 461, 464, 494; 500 See Boyd, A.J., 370; 496 Lederer, E., 253, 254; 278 Lederhendler, I., 202, 252, 255, 256; 278 Lederhendler, I.I., 175, 182, 184; 278 Ledford-Hoffman, P.A., 34; 37 Lee, C.C. See Bunt, J.S., 32; 36 Lee, H., 151, 155; 160 Lee, R.M. See Preston, R.J., 205, 208, 209, 244; 280 Lee, S.-H. See Gerencser, G.A., 221; 275 Lee, W.Y., 62; 88 Lees, R., 396, 397, 398; 500 Le Gall, L.-Y., 454; 500 Legendre, R. See Page, L., 117; 158 Lehman, H.K. See Blankenship, J.E., 272 Leighton, D.L., 204, 206, 212, 214; 278 Le Loeuff, P., 457; 500 Le Roux, P.J. See Stander, G.H., 360; 504 Leslie, R.W., 379, 467, 468; 500 See Hatanaka, H., 499 See Kinloch, M.A., 375; 500 See Payne, A.I.L., 410, 439, 448, 467; 502 See Uozumi, Y., 504 Lettenmaier, D.P. See Millard, S.P., 44, 46, 52, 63, 64, 68; 88 Levin, L.A., 118, 119, 124, 132, 135, 137, 138, 143; 160 Levinton, J.S., 124, 125; 160 Lewis, C.A., 115, 332, 342; 160, 349 Lewis, D.B., 129, 130; 160 Lewis, J.R., 52; 88 Lewis, R.K. See Winstanley, R.H., 408; 504

534

OCEANOGRAPHY AND MARINE BIOLOGY

Li, W.K.W., 31; 37 Liao, P.B., 566; 574 Lichter, W. See Sigel, M.M., 281 Lick, W., 152; 160 Lickey, M.E., 214, 245, 270; 278 See Block, G.D., 245; 272 See Jordan, W.P., 245; 277 Lie, U., 121, 125; 160 Ligman, S.H., 192; 278 Lima, F.R., 453; 500 Lima, G.M., 132; 160 Lin, G.H.Y. See Hirschfeld, D.R., 276 Lincoln, R.J., 47; 88 Lindsey, J. See Kittredge, J.S., 256; 277 Lipskaya, N.J., 415, 416; 500 Littler, D.S. See Littler, M.M., 173; 278 Littler, M.M., 173; 278 Livesy, J.L. See Laws, E.M., 94; 111 Livingston, C.A. See Blankenship, J.E., 272 Livingston, R.J. See Mahoney, B.M.S., 124; 160 Lloris, A., 385; 500 Lloris, D. See Macpherson, E., 449; 501 Lo Bianco, S., 190, 191, 202; 278 Loeb, V.J. See Shulenberger, E., 34; 38 Loeng, H. See Rey, F., 19, 20; 38 Löf, J. See Enell, M., 566, 567, 570, 572; 574 Lombardini, J.B., 178; 278 Long, C.E., 102, 104; 111 Long, J.L. See Krauhs, J.M., 270; 278 Longley, A.J., 187; 278 Longley, R.D. See Longley, A.J., 187; 278 Longley, W.H., 507, 513; 561 Loughlin, G.M. See Gerencser, G.A., 221; 275 Loya, Y., 70; 88 Lu, C., 245; 278 Lu, P-C., 97; 111 Lubbock, R., 507, 509, 513, 532; 561 See Polunin, N.V.C., 507, 508, 509, 512, 513, 514, 516, 517, 523, 528, 532, 533, 553, 558; 562 Lucas, A.H. See Sigel, M.M., 281 Lucas, J.S. See Henderson, J.A., 132; 159 Luckenbach, M.W., 125, 134, 138; 160 Lucks, D.K., 388, 393, 451; 500 See Botha, L., 393, 451; 496 Ludwig, J.A., 81; 88 Lukowiak, K., 245, 258, 259, 260; 278, 279 See Austin, T., 216; 271 See Peretz, B., 258; 280 Lukowiak, K.D. See Peretz, B., 259; 280

Luling, K.H., 559; 561 Lumley, J.L. See Tennekes, H., 101; 112 Lussenhop, J., 84; 88 Luther, W., 507, 513, 514, 518, 519, 520, 522, 541, 559; 561 Lützen, J., 331, 332, 333, 334; 349 See Bresciani, J., 345; 348 See Høeg. J.T., 328, 329, 332; 349 Luxton, J.A. See Antonia, R.E., 103; 110 MacCall, A.D., 475, 492; 500, 501 See Lasker, R., 492; 500 Macé, A.-M., 236; 279 MacFarland, F.M., 174, 204; 279 MacGinitie, G.E., 174, 181, 184, 189, 190, 192, 193, 194, 206, 256, 559; 279, 562 MacGinitie, N. See MacGinitie, G.E., 184, 256, 559; 279, 562 Maciolek, N.J., 121, 123; 160 Mackay, A.P., 49; 88 Mackintosh, N.A., 13; 37 Macko, S.A. See Walsh, J.J., 38 Macloed, P.R. See King, D.P.F., 414, 483, 484, 489; 500 MacMunn, C.A., 226, 253, 254; 279 Macnae, W., 169, 174, 175, 176, 247, 509, 512, 513, 514, 516, 517, 522; 279, 562 Macpherson, E., 381, 413, 423, 424, 425, 430, 439, 441, 448, 449, 467, 469, 489; 501 Madsen, O.S. See Grant, W.D., 104, 114, 145, 147, 152; 111, 159 Magno, S. See Danise, B., 273 See Fattorusso, E., 274 Magnus, D.B.E., 507, 514, 517, 520, 522, 523, 524, 526, 528, 529, 532, 533, 534, 535, 538, 540, 541, 543, 548, 558, 559, 560; 562 Magnus, O., 167; 279 Mahoney, B.M.S., 124; 160 Makioka, T. See Nishiwaki, S., 203, 237; 279 Malouf, R. See Keck, R., 128; 160 Mandelbaum, D.E. See Weiss, K.R., 283 Manley, T. See Shuchman, R.A., 38 Manley, T.O., 15, 21; 37 Mann, K.H. See Fong, W., 225; 274 See Hughes, R.N., 121; 160 Mann, R., 134, 155; 160 See Gallager, S.M., 220; 275 Mapstone, B.D. See Andrew, N.L., 39–90 Marchand, J.M., 388, 395, 451; 501 Marcus, E., 172, 174, 175, 176, 190, 206; 279

AUTHOR INDEX

See Marcus, E., 174, 175, 190, 206; 279 Marcus, E.d.B.-R., 173, 174, 175, 176, 203; 279 Mare, M.F., 123; 160 Maree, S. See Lucks, D.K., 388; 500 Margalef, R., 35; 37 Markovich, S. See Susswein, A.J., 173, 180, 205, 209; 282 Marler, P., 540; 562 Marr, J.C. See Clark, F.N., 490, 491; 497 Marr, J.W.S., 11, 34; 37 Marsden, J.R. See Bhup, R., 133; 157 Marshall, P.T., 13, 14; 37 Marshall, S.M. See Barnes, H., 71, 72; 85 Marshall, R.K. See Wooding R.A., 153; 164 Martin, J.D. See González, A.G., 275 Masamune, T. See Irie, T., 262; 276 Massé, H., 136; 160 See Guérin, J.-P., 136; 159 Mastagli, P. See Augier, J., 262; 271 Matheke, G.E.M., 32; 37 Matsuda, H., 260, 268; 279 Matsuura, Y., 455; 501 Matthews, J.B., 13; 37 Matthews, J.P., 420, 434, 435, 486; 501 See Van Den Berg, R.A., 426, 427, 428, 453, 454, 455; 504 Maurer, D. See Keck, R., 128; 160 See Watling, L., 53; 89 Max Westby, G.W., 544; 562 Maykut, G.A., 14; 37 Maynard, N.G., 35; 37 Maynard-Smith, J., 347, 533; 349, 562 Mayo, R.D. See Liao, P.B., 566; 574 Mayol, L. See Danise, B., 273 See Fattorusso, E., 274 McAlice, B.J. See Lee, W.Y., 62; 88 McAllister, L.B. See Scheller, R.H., 281 McCall, P.L., 124, 137; 160 McCann, L.D., 128, 130; 160 McCarthy, C. See Arch, S., 192; 271 McCarthy, J.J. See Glibert, P.M., 26; 36 McCave, I.N., 152; 160 See Bartz, R., 156 See Miller, M.C., 152; 161 See Nowell, A.R.M., 109; 111 McCombie, A.M. See Colby, P.J., 568; 574 McDonald, F.J., 265; 279 See Schmitz, F.J., 262, 265; 281 McDougall, K.D., 141; 161 McEdward, L. See Day, R., 115; 158

McFarland, D. See Larkin, S., 243; 278 McGrorty, S., 124; 161 McIntire, C.D., 100; 111 McIntyre, A.D., 48, 53; 88 See Holme, N.A., 43, 52, 53; 87 See Wigley, R.L., 120; 164 McKay, D., 269; 279 McLain, D.R., 473, 494; 501 McLaughlin, P.A., 298, 301; 349 See Henry, D.P., 298; 349 McLean, N. See Koningsor Jr, R.L., 221; 277 McLean, S.R., 152; 161 See Smith, J.D., 114, 147, 152; 163 McLean, W.E., 566; 574 McLellan, T. See Grassle, J.F., 159 McLusky, D.S., 143; 161 See Brown, J.R., 567; 574 McMaster, G.S. See Ebert, T.A., 71; 86 McNally, P.M. See Carlucci, A.F., 564; 574 McNulty, J.K., 119, 120; 161 McRoy, C.P., 13, 16; 37 See Sambrotto, R.N., 16; 38 McWilliam, P.S., 119; 161 Mead, R., 81, 82; 88 See Perry, J.N., 72, 75; 88 Meador, J.P. See Bernstein, B.B., 74; 85 Meadows, P.S., 115, 123, 126, 129, 130; 161 See Crisp, D.J., 132, 133; 157 Meguro, H., 32; 37 Meinwald, J. See Dieter, R.K., 263; 273 See Kinnel, R.B., 277 Melo, Y.C., 412, 413; 501 See Le Clus, F., 436; 500 Melville-Smith, R., 410, 457, 472; 501 Mendenhall, W. See Scheaffer, R.L., 42, 52, 56; 89 Merican, Z.O., 570; 574 Merriman, D., 169, 177; 279 Merz, R.A., 110; 111 Meyers-Schulte, K. See Switzer-Dunlap, M., 183; 282 Michael, A.D. See Cassie, R.M., 120; 157 Michaud, D.P. See Schmitz, F.J., 281 Middleton, G.V., 91, 105, 108; 111 Midland, S.L., 268; 279 Mihursky, J.A. See Holland, A.F., 125; 159 See Mountford, N.K., 121; 161 Mileikovsky, S.A., 114, 116, 134, 155, 199; 161, 279 Millard, S.P., 44, 46, 52, 63, 64, 68; 88 Miller, C.B. See Hogue, E.W., 139, 143, 153; 159 Miller, D.C., 106; 111

535

536

OCEANOGRAPHY AND MARINE BIOLOGY

Miller, D.L. See Jahan-Parwar, B., 248; 277 Miller, M. See Morton, J., 207, 243; 279 Miller, M.C, 152, 177, 181, 188, 191, 201, 203; 161, 279 Miller, S.E., 155; 161 Mills, E.L. See Muschenheim, D.K., 92, 105; 111 See Sanders, D.L., 120; 162 Milne, A. See Clark, R.B., 75; 86 Minale, L., 263, 267; 279 See Danise, B., 273 See Finer, J., 274 See Imperato, F., 263; 276 Mitchell, K.A. See Meadows, J.S., 126; 161 Mitchell, R. See Kirchman, D., 57; 88 Miya, Y., 512, 514, 522, 523; 562 Miyake, P.M., 405; 501 Miyake, S. See Miya, Y., 512, 514, 522, 523; 562 Mizrahi, L. See Gilboa-Garber, N., 270; 275 Mizuno, D. See Yamazaki, M., 284 Moehring, J.L., 507, 508, 509, 513, 514, 526, 533, 534, 536, 547; 562 Molinski, M. See Penczak, T., 575 Moloney, C.L., 466, 490; 501 Molteno, C.J., 402; 501 Mombeck, F., 447; 501 See Macpherson, E., 381, 441; 501 Montagna, P.A., 62; 88 Montgomery, M.P. See Day, J.H., 121; 158 Moore, D.R., 177; 279 Moore, H.B. See McNulty, J.K., 119, 120; 161 Moore, P.G., 126; 161 Moore, R.E., 261; 279 See Finer, J., 274 See Fujiki, H., 275 See Mynderse, J.S., 261; 279 Morales, B., 413; 501 See Macpherson, E., 448, 467; 501 Morgan, E., 129, 130; 161 Morin, A., 47, 52, 53, 56, 57, 58; 88 Morison, J. See Johannessen, O.M., 37 Morison, J.H. See Smith, D.C., 15; 38 Morisita, M., 71, 73, 74; 88 Morita, S., 454; 501 Moritz, C.E., 190; 279 Mork, M. See Gammelsrød, T., 14; 36 Morris, D., 547; 562 Mortensen, T., 131; 161 Morton, J., 207, 243; 279 See Willan, R., 169, 172, 176, 177, 226; 283 Morton, J.E. See Stone, B.A., 221; 282

Moseley, H.N., 253; 279 Moser Jr, C.E. See Willey, J.C., 261; 283 Mountford, M.D., 71, 76,; 88 Mountford, N.K., 121; 161 See Holland, A.F., 125; 159 Moyse, J., 344; 349 See Hui, E., 299, 301; 349 Munn, E.A. See Barnes, H., 325; 347 See Klepal, W., 292; 349 Muramoto, K. See Kamiya, H., 270; 277 Murphy, P.T. See Kazlauskas, R., 263; 277 Muschenheim, D.K., 92, 105; 111 Muus, B.J., 125; 161 Muus, K., 123, 125, 134, 135; 161 Myers, A.C., 124; 161 Mynderse, J.S., 261, 262; 279 Nagle, G.T., 192; 279 Nakamura, H., 452; 501 Nakamura, N. See Saito, Y., 176, 211, 213, 217, 218, 219, 228; 281 Nakayama, M. See Katayama, A., 277 Nakayasu, M. See Fujiki, H., 275 Natori, S. See Yamazaki, M., 284 Nauen, C.E. See Collette, B.B., 452, 454; 497 Nawratil, O., 414; 501 Neck, R.W., 172, 175, 177; 279 Nelson, C.S. See Parrish, R.H., 353; 502 Nelson, D.M., 15, 26; 37 See Smith Jr, W.O., 13, 14, 16, 18, 26, 30, 31; 38 See Wilson, D.L., 31, 34; 38 Neori, A., 19; 37 Nepgen, C.S. de V., 396, 397, 398, 399, 403, 413, 415, 419, 426, 427, 428, 445, 453, 454, 455, 483, 489, 492; 501 See De Jager, B. van D., 403, 425, 426, 427, 428, 453; 498 Neu, W., 247; 279 Newell, G.E. See Chapman, G., 143, 144; 157 Newman, G.G., 353, 355, 360, 362, 364, 365, 378, 379, 380, 381, 382, 383, 384, 385, 386, 411, 415, 423, 429, 431, 456, 457, 458, 461, 467, 469, 471, 474, 486, 487, 490, 491, 492, 495; 502 See Davies, S.L., 434, 461; 498 Newman, W.A., 285, 287, 337, 338, 340, 341, 344, 346; 349 See Dayton, P.K., 298; 348 See Tomlinson, J.T., 286, 288; 350 Newton, J.F. See Paquette, R.G., 19; 38

AUTHOR INDEX

Nicholaidou, A. See Rees, E.I.S., 124; 162 Nichols, F.H., 121; 161 Nie, H.W. de, 49; 88 Niebauer, H.J., 11, 14, 15, 16, 18; 37 See Alexander, V., 13, 14, 16, 33; 36 Niell, F.X., 204, 207, 211; 279 Nilsson-Cantell, C.A., 299, 302, 307, 315, 316, 317, 318, 319; 350 Nishibori, K., 193, 253, 254; 279 Nishiguchi, T. See White, J.D., 283 Nishikawa, Y. See Kikawa, S., 454, 455; 500 Nishimura, A., 564; 575 Nishiwaki, S., 190, 203, 237, 238; 279 Nittrouer, C.A. See Ledford-Hoffman, P.A., 34; 37 Nixon, S.W. See Oviatt, C.A., 82; 88 Noakes, D.L.G. See Pot, W., 48; 88 Nolan, B.A., 517; 562 Nordstrom, V. See Miyake, P.M., 405; 501 Noriki, S., 33; 37 Norris, J.N. See Littler, M.M., 173; 278 Norte, M. See González, A.G., 275 Norton, J.G. See McLain, D.R., 473, 494; 501 Norton, T.R. See Mynderse, J.S., 261; 279 Nott, J.A. See Gibson, P.H., 132; 158 Nowell, A.R.M., 91–112; 91, 92, 105, 107, 109, 114, 139, 141, 144, 145, 146, 147, 152, 153; 111, 161 See Bartz, R., 156 See Eckman, J.E., 103, 107, 108, 153; 110, 158 See Gross, T.F., 147, 152; 159 See Jumars, P.A., 92, 102, 103, 122, 123, 127; 111, 160 See Miller, D.C, 110; 111 See Taghon, G.L., 105, 110; 112 Nozaki, H. See Katayama, A., 277 Nybakken, J.W. See Oliver, J.S., 135; 161 Nyffleler, U.P. See Santschi, P.H., 111 O’Brien, D.P., 93; 111 O’Brien, J.J. See Røed, L.P., 14; 38 O’Connor, E.F. See Ainley, D.G., 34; 36 Ogata, K. See Kamiya, H., 270; 277 See Yamazaki, M., 284 Ohlhorst, S.L., 119; 161 Olivar, M.P., 441, 445; 502 Oliver, J. See Dayton, P.K., 298; 348 Oliver, J.S., 124, 125, 134, 135, 137, 139; 161 See Dayton, P.K., 122, 123, 125, 135, 137; 158 See Hulberg, L.W., 124; 160 Oliver, L. See Otsuka, C., 200; 280

537

Olsson, I., 140; 161 Omori, M., 48, 70, 83; 88 Oren, O.H., 171; 279 Oriente, G. See Danise, B., 273 See Fattorusso, E., 274 Orrhage, L. See Gärdefors, D., 82, 140; 86, 158 Orth, R.J., 124, 143; 161 Orton, J.H., 141, 142; 161 Osche, G., 320; 350 Ostergaard, J.M., 189, 191, 199; 280 Österman, C.-S. See Bonsdorff, E., 138; 157 O’Toole, M.J., 418, 435, 441, 443, 444, 447, 451; 502 Otsuka, C., 177, 184, 191, 200, 205, 220; 280 Otsuka, S. See Tanaka, A., 282 Ott, L. See Scheaffer, R.L., 42, 52, 56; 89 Oviatt, C.A., 82, 568; 88, 575 Paige, J.A., 190, 199, 200; 280 Paine, R.T., 226, 239, 251; 280 Painter, S.D. See Nagle, G.T., 192; 279 Palmer, C., 509, 513, 520, 532, 534; 562 Palmer, M.A., 110, 119, 144, 154; 111, 161 Palmisano, A.C., 14, 32; 37 See Sullivan, C.W., 14; 38 Pamatmat, M.M., 143; 161 Pang, K.T. See Lombardini, J.B., 178; 278 Paola, C., 103; 111 Papka, R., 201; 280 Paquette, R.G., 19, 20; 38 Paranjape, M.A. See Sheldon, R.W., 490; 503 Parent, J. See Bocquet-Védrine, J., 328; 348 Parker, C., 533; 562 Parker, G.A. See Maynard-Smith, J., 533; 562 Parker, G.H., 243, 244; 280 Parker, R.R., 412; 502 Parkinson, C.L. See Zwally, H.J., 38 Parrish, R.H., 353; 502 Parry, G.D., 197; 280 Parsons, D.W. See Ferguson, G.P., 182; 274 See Pinsker, H.M., 182, 190, 193; 280 See Von der Porten, K., 283 Partheneides, E., 152; 161 Paul, V.J. See Fenical, W., 274 Pauley, G.B., 269, 270; 280 Paulson, A.C., 509, 513, 514; 562 Payandeh, B., 75; 88 Payne, A.I.L., 380, 381, 384, 388, 389, 391, 393, 395, 410, 413, 439, 448, 449, 451, 467, 469, 471, 476, 483, 492; 502

538

OCEANOGRAPHY AND MARINE BIOLOGY

See Hatanaka, H., 499 See Lucks, D.K., 388; 500 See Uozumi, Y., 411; 504 Pearse, J.S., 115; 161 Pearson, T.H., 121, 564, 567, 571; 162, 575 Pechanec, J.F., 53; 88 Pechenik, J.A., 132; 162 See Lima, G.M., 132; 160 Peer, D.L. See Hughes, R.N., 121; 160 Penczak, T., 565, 566; 575 Penrith, M.J., 426, 427, 428, 450, 453, 455; 502 See Talbot, F.H., 402, 403, 425, 426, 427, 453, 454, 455; 504 Perdue, W.F. See Paquette, R.G., 19; 38 Peretz, B., 202, 205, 231, 232, 236, 238, 258, 259; 280 See Hirsch, H.R., 201, 202; 276 See Ludowiak, K., 258, 259; 279 See Papka, R., 201; 280 See Rattan, K.S., 201, 259; 280 Perez, C., 334; 350 Pérez, R. See González, A.G., 275 Perritt, S.E. See Capo, T.R., 200; 272 Perron, F.E., 197; 280 Perry, J.N., 72, 75; 88 See Taylor, L.R., 75, 82; 89 Peter, R. See Hadl, G., 130; 159 Petersen, C.G.J., 119; 162 Peterson, B.J. See Eppley, R.W., 16, 31; 36 Peterson, C.H., 124, 125, 143; 162 Peterson, E.H., 103; 111 Petrecca, R.F. See Grassle, J.F., 159 Pettit, G.R., 266; 280 Phillips, B.F., 430; 502 Phillips, D.W. See Suer, A.L., 131, 132; 163 Phillips, M.J. See Merican Z.O., 570; 574 Phillips, P.J., 129, 130; 162 Phinney, H.K. See McIntire, C.D., 100; 111 Piattelli, M. See Danise, B., 273 See Fattorusso, E., 274 Pielou, E.C., 51, 71, 72, 74, 76, 79, 80, 81, 82, 83, 517, 553; 88, 562 Pihl, L., 48, 50, 52, 55; 88 Pikanowski, R.A. See Saila, S.B., 52, 63; 88 Pillar, S.C. See Shannon, L.V., 353, 355, 431, 435, 436, 440, 441, 443, 444, 447, 453; 503 Pilsbry, H., 175, 243; 280 Pilsbry, H.A., 299, 301, 320, 337, 340; 350 Pinsker, H., 248, 257, 258; 280 See Castellucci, V., 258; 273

See Feinstein, R., 179; 274 See Kupfermann, I., 205, 214, 258; 278 See Von der Porten, K., 205; 283 Pinsker, H.M., 179, 182, 190, 192, 193, 258; 280 See Carew, T.J., 258; 273 See Cobbs, J.S., 174, 179, 180, 188, 192; 273 See Ferguson, G.P., 182; 274 Platt, T., 81; 88 Pochon-Masson, J., 296, 325; 350 See Bocquet-Védrine, J., 325; 348 See Turquier, Y., 296, 297; 350 Poizat, C., 177; 280 See Vicente, N., 191, 200; 283 Polgar, T.T. See Holland, A.F., 125; 159 Pollard, D.A. See Bell, J.D., 48; 85 Pollock, D.E., 408, 425, 429, 454, 456, 492; 502, 503 See Crawford, R.J.M., 353–505 See Newman, G.G., 429, 456, 458; 502 Polunin, N.V.C., 507, 508, 509, 512, 513, 514, 516, 517, 523, 528, 532, 533, 553, 558; 562 See Lubbock, R., 507, 509, 513, 532; 561 Poole, M. See Saunders, A.M.C., 190, 191; 281 Por, F.D., 320; 350 Porter, H.J. See Williams, A.B., 117; 164 Porter, J.W., 117; 162 Porter, K.G. See Porter, J.W., 117; 162 Pot, W., 48, 49; 88 Potthoff, T. See Richards, W.J., 455; 503 Potts, F.A., 327; 350 Potts, G.W. See Edmunds, M., 211; 274 Pozo, A.E., 413; 503 Prasad, R.S. See Schmitz, F.J., 281 Pratt, D.M., 126, 141, 142, 143, 153; 162 Premuzic, E.T. See Walsh, J.J., 38 Prenski, L.B., 418, 420, 421, 443; 503 Preston, J.L., 507, 509, 513, 514, 516, 520, 522, 535, 536, 537, 540, 544, 545, 546, 560; 562 Preston, R.J., 205, 208, 209, 244; 280 Prestwich, G.D. See Shieh, H.-M., 266; 281 Pringle, J.D., 52, 53, 56, 57; 88 Prodanov, K. See Babayan, V., 496 Prosch, R.M., 413, 417, 418, 419, 423, 442, 444, 466; 503 See Armstrong, M.J., 415, 433, 459, 480; 495 Pruvot-Fol, A., 175, 247; 280 Pyeun, J.H. See Cho, D.M., 220; 273 Querellou, J., 570; 575 Quinn, R.J. See Kazalauskas, R., 263; 277

AUTHOR INDEX

Radlick, P. See Hirschfeld, D.R., 276 See Sims, J.J., 263; 281 Raffy, A. See Fontaine, M., 253; 274 Rahman, A. See Schmitz, F.T., 281 Ralls, K., 344; 350 Rand, R.W., 415, 417, 420, 423, 429, 430, 480; 503 Randall, J.E., 508; 562 See Hoese, D.F., 507, 508, 509, 513, 514, 551; 561 Randall, R.M., 443; 503 See Duffy, D.C., 461; 499 Rao, H.S., 517; 562 Rasmussen, E., 132; 162 Rattan, K.S., 201, 259; 280 Rattray, J.M., 422, 424; 503 Ray, A.J., 110; 111 Rayner, M.D. See Watson, M., 260, 268; 283 Reading, C.J. See McGrorty, S., 124; 161 Redman, G. See Pinsker, H., 280 See Von der Porten, K., 205; 283 Rees, E.I.S., 124; 162 Render, H.A. See Bartsch, P., 174; 272 Reid, J.L. See Wooster, W.S., 353;505 Reidenauer, J.A. See Thistle, D., 140; 163 Reinhard, E.C., 334; 350 Reise, K., 124, 140; 162 Reish, D.J., 136; 162 Remaley, S. See Schmitz, F.J., 281 Resh, V.H., 40, 44, 48, 51, 52, 53, 54; 88 Revsbach, N.P. See Jørgensen, B.B., 92; 111 Rey, F., 19, 20; 38 Rey, J.R., 188, 190, 193, 194, 203; 280 Rhoads, D.C., 109, 122, 124, 140; 111, 162 See Young, D.K., 124; 164 Riccio, R. See Danise, B., 273 See Finer, J., 274 See Imperato, F., 263; 276 See Minale, L., 263, 267; 279 Rice, A.L., 520; 562 Rice, M.E., 132; 162 Richards, W.J., 455; 503 Richardson, E.V. See Guy, H.P., 111 Richmond, R.H., 132; 162 Richter, W., 136; 162 See Sarnthein, M., 136; 162 Rickett, L.H. See Butterworth, D.S., 494; 497 Ricketts, E.F., 174; 280 Riedl, R.J. See Fenchel, T.M., 564; 574 Rieger, R.M. See Hagerman, G.M., 118, 144, 154; 159 Riley, G.A., 14; 38

Riley, H.T. See Bostock, J., 167, 253; 272 Ripley, B.D., 81, 83; 88 Ritchley, T.A. See Helshe, J.F., 53, 72, 75; 87 Ritchie, L.E., 333; 350 See Høeg, J.T., 331, 333, 334; 349 Ritte, U. See Hairston, N.G., 73; 87 Robbins, L.C. See Blankenship, J.E., 272 Roberts, D. See DeMartini, E.E., 49; 86 Roberts, D.J. See Hughes, R.N., 197; 276 Roberts, M.H. See Block, G.D., 245; 272 Roberts, R.J. See Jones, K.J., 574 See Turner, M.F., 564; 575 Robertson, A.A. See King, D.P.F., 355; 500 Robinson, G.A., 415, 450; 503 Robson, D.S., 365; 503 Rock, M.K. See Blankenship, J.E., 183, 192; 272 Roe, H.S. See Fasham, M.J.R., 82; 86 Røed, L.P., 14; 38 See Buckley, J.R., 36 See Gammelsrød, T., 14; 36 Roel, B. See Macpherson, E., 501 Rohlf, F.J., 76, 83; 88 See Sokal, R.R., 40, 47, 61, 64, 69, 73; 89 Romia, R.D. See Southard, J.B., 103; 111 Ronald, B.P. See Ronald, R.C., 260; 280 Ronald, R.C., 260; 280 Rönner, U., 26, 31; 38 Rosa, H. See Laevastu, T., 453; 500 Rosati, F. See Baccetti, B., 296; 347 Rosen, S.C., 209; 280 Rosenberg. R., 72; 88 See Pearson, T.H., 564, 567; 575 See Pihl, L., 48, 50, 52, 55; 88 Ross, A. See Newman, W.A., 340; 349 Rotenberry, J.T., 64, 65, 66, 68; 88 Rothman, B. See Pinsker, H., 280 See Von der Porten, K., 205; 283 Rothman, B.S., 192; 280 See Von der Porten, K., 283 Rouger, Y. See Otsuka, C., 184, 200; 280 Routledge, R.D. See Pielou, E.C., 80; 88 Rowe, F. See Tranter, D.J., 163 Rowe, G.T. See Grassle, J.F., 159 See Walsh, J.J., 38 Rowley, D. See McKay, D., 269; 279 Rubies, P. See Olivar, M.P., 441, 445; 502 Rucabado, A. See Lloris, A., 385; 500 Rüdiger, W., 253, 254; 280 Rudomiotkina, G.P., 455; 503

539

540

OCEANOGRAPHY AND MARINE BIOLOGY

Rumohr, H. See Arntz, W.E., 124; 156 Rumrill, S.S. See Cameron, R.A., 133; 157 Ruppert, E.E. See Fox, R.S., 174; 274 Russell, B.C., 48, 49; 88 See Bell, J.D., 48; 85 Russell, B.J. See Hamilton, P.V., 173, 203, 229, 249, 250, 253; 275 Russell, E., 251, 252; 281 Rutowski, R.L., 187; 281 Ruveda, E.A. See Imamura, P.M., 268; 276 Rychly, J., 565, 566; 575 Ryland, J.S. See Harvey, P.H., 78; 87 Ryther, J.H., 223, 417, 564; 281, 503, 575 Saetersdal, G. See Boerema, L.K., 496 Sahley, C. See Lukowiak, K., 258; 279 Saila, S.B., 52, 57, 61, 63; 88 Saito, M. See Konishi, R., 15, 16; 37 Saito, Y., 176, 211, 213, 217, 218, 219, 228; 281 Sakata, K., 217; 281 Sakshaug, E., 31; 38 Sale, P.F., 48, 49, 53, 54, 55; 89 See Doherty, P.J., 69; 86 See McWilliam, P.S., 119; 161 Salmon, M. See Nolan, B.A., 517; 562 Sambrotto, R.N., 16, 18; 38 Sameoto, D.D., 129, 130, 143; 162 Sammarco, P.W. See Abel, D.J., 80; 85 Sampou, P.A. See Oviatt, C.A., 568; 575 Sanders, H.L., 120, 122; 162 See Grassle, J.F., 159 Sanford, L.P. See Grant, W.D., 110, 152; 111, 158 Sanford, S.N.F., 169, 177, 256; 281 Santacroce, C. See Fattorusso, E., 274 Santos, S.L., 118,, 119, 120, 123, 124, 137, 154; 162 Santschi, P.H., 110; 111 Sarnthein, M., 136; 162 See Richter, W., 136; 162 Sarvala, J., 132; 162 Sarver, D.J., 173, 175, 176, 178, 180, 181, 191, 194, 196, 197, 199, 201, 202, 207, 211, 218, 219, 221, 222, 223, 226, 231, 233, 234, 235, 237, 238, 239, 240, 241, 242, 250, 251; 281 Sastry, A.N., 115; 162 See Phillips, B.F., 430; 502 Sato, T., 413, 469; 503 See Hatanaka, H., 499 See Uozumi, Y., 504 Saunders, A.M.C., 190, 191; 281

Saville, A., 487, 491; 503 Sawaya, P., 188, 193, 204, 205, 206, 211, 251, 255, 270; 281 See Cipolli, I.N., 270; 273 Schaller, F., 345; 350 Schandelmeier, L., 13, 16, 33; 38 Scheaffer, R.L., 42, 43, 44, 52, 56, 59; 89 Scheffe, H., 56; 89 Scheibling, R.E., 72, 143; 89, 162 Scheller, R.H., 192; 281 Scheltema, R.S., 115, 116, 132, 133, 430; 162, 163, 503 Schembri, P.J., 559; 562 Scheuer, P.J., 261, 262; 281 See Dilip de Silva, E., 274 See Kato, Y., 261, 262; 277 Schulte, G.R., 281 Schindler, D.W., 564; 575 Schlesinger, D.H. See Blankenship, J.E., 183, 192; 272 Schlichting, H., 146; 163 Schmale, M. See Feinstein, R., 179; 274 Schmale, M.C. See Feinstein, R., 179; 274 Schmekel, L., 187; 281 Schmidt, P.G. See Schmitz, F.J., 281 Schmitz, F.J., 262, 263, 265, 266, 267; 281 See Gopichand, Y., 275 See Gunatilaka, A.A.L., 275 See Hollenbeak, K.H., 276 See Kaul, P.N., 277 See McDonald, F.J., 279 See Vanderah, D.J., 265; 283 Schrader, G..C. See Horner, R.A., 32; 37 Schreiber, G., 226, 253, 255; 281 Schroeder, P.C., 115; 163 Schülein, F.H., 355, 368, 370, 371, 372, 434, 435, 436, 445; 503 See Macpherson, E., 381, 441; 501 Schulte, G.R., 268; 281 Schwartz, J.H. See Bevelaqua, F.A., 269; 272 See Scheller, R.H., 281 Schwartz, R.E. See Dilip de Silva, E., 274 Schwarz, M., 205, 215, 228; 281 See Susswein, A.J., 215, 228; 282 Schweigert, J.F., 62, 63, 64, 70; 89 Scudo, F.M., 343; 350 Scullard, C. See Bayne, B.L., 217; 272 Seapy, R.R., 71; 89 Seber, G.A.F., 42, 44, 46, 47; 89 Segerstråle, S.G., 142; 163 Seiderer, J.L. See Griffiths, C.L., 429; 499

AUTHOR INDEX

Self, R.F.L. See Jumars, P.A., 122; 160 See Nowell, A.R.M., 105; 111 Selmer-Olsen, A.R. See Bergheim, A., 564; 574 Sengupta, P.K. See Schmitz, F.J., 281 Sewell, R.B.S., 326; 350 Seymour, M.K., 117; 163 Shackleton, L.Y., 473; 503 Shames, I.H., 99; 111 Shannon, L.V., 353, 355, 356, 363, 403, 431, 433, 434, 435, 436, 439, 440, 441, 442, 443, 444, 447, 449, 453, 454, 455, 458, 473, 474, 490, 491, 494; 503 See Bergh, M.O., 415, 466; 496 See Chapman, P., 353, 431; 497 See Crawford, R.J.M., 353–505 Shapiro, E., 254; 281 Sharp, B.J. See Sale, P.F., 48, 49, 53, 54, 55; 89 Shcherbich, L.V., 413, 417; 503 See Assorov, V.V., 423; 496 Shea, P.J. See Toft, C.A., 64, 65, 66, 84; 89 Sheader, M. See Buchanan, J.B., 125; 157 Sheldon, R.W., 490; 503 Shelley, J. See Gopichand, Y., 275 Shelton, P.A., 431, 433, 436, 444, 459, 473, 475, 476, 477, 478, 487, 490, 491, 494; 503, 504 See Armstrong, M.J., 415, 433, 459, 480; 495 See Crawford, R.J.M., 357, 361, 365, 367, 368, 415, 417, 418, 419, 420, 434, 435, 436, 443, 460, 461, 462, 463, 466, 472, 473, 474, 475, 478, 479, 480, 481, 483, 489, 491, 492; 497, 498 See Davies, S.L., 434, 461; 498 See Dudley, S.F.J., 443; 498 See Hampton, I., 435, 458; 499 See King, D.P.F., 355; 500 See Shannon, L.V., 433; 503 Sherman, K. See Brown, B.E., 497 Sherman, K.M. See Bell, S.S., 118, 144, 154; 156 Shieh, H.-M. 266; 281 Shields, A., 91, 152; 111, 163 Shimizu, Y. See Kamiya, H., 270; 277 Shin, P.K.S., 121; 163 Shinn, E.A., 520, 522; 562 Shuchman, R.A., 15, 21; 38 See Johannessen, O.M., 37 Shudo, K. See Fujiki, H., 275 Shulenberger, E., 34; 38 Sibert, J.R. See Schweigert, J.F., 62, 70; 89 Sica, D. See Fattorusso, E., 274 Siedler, G. See Dietrich, G., 17; 36 Siegelman, H.W. See Chapman, D.J., 254; 273

541

Siegfried, W.R. See Frost, P.G.H., 443, 494; 499 Sigda, J. See Kirchman, D., 57; 88 Sigel, M.M., 265; 281 Sigurdsson, J.B., 144; 163 Silbernagel, S.B. See Carlucci, A.F., 564; 574 Silva, M. See Bhakuni, D.S., 262; 272 Silvert, W., 489; 504 Simberloff, D., 78; 89 Simon, J.L. See Bloom, S.A., 121; 157 See Dauer, D.M., 124, 125, 154; 157, 158 See Santos, S.L., 118, 119, 120, 123, 124, 137, 154; 162 Simons, D.B. See Guy, H.P., 111 Sims, J.J., 263; 281 See Hirschfeld, D.R., 276 See Midland, S.L., 279 Sims, P.F. See Badenhorst, A., 471; 496 Sinclair, D.F., 77; 89 Sinclair, M. See Bergquist, P.R., 115; 156 Skeean, R.W. See White, J.D., 283 Skellam, J.G., 75; 89 Slattery, P.N. See Oliver, J.S., 135; 161 Sleeper, H.L. See Fenical, W., 274 Slosarczyk, W. See Kompowski, A., 415, 416; 500 Smale, M.J., 400, 427, 428; 504 Smart, G. See Bromley, P.J., 565; 574 Smidt, E.L.B., 123; 163 Smith, D.C., 15, 19, 21; 38 Smith, D.F. See Tranter, D.J., 119; 163 Smith, F.G.W., 141; 163 See Doochin, H., 141; 158 Smith, G., 307, 327, 328, 336; 350 Smith, G.A. See Palmisano, A.C., 37 Smith, H.F., 53; 89 Smith, J.D., 114, 147, 151, 152; 163 Smith, J.L.B., 509, 513, 520, 541; 562 Smith, J.T. See Block, G., 245; 272 Smith, M. See Jahan-Parwar, B., 208; 277 Smith, P., 565; 575 Smith, R.L. See Barber, R.T., 18; 36 Smith, S.J., 62, 69; 89 Smith, S.L., 13, 19, 20, 21; 38 Smith, S.T., 182; 281 Smith Jr, W.O., 11–38; 13, 14, 16, 18, 20, 21, 30, 31; 38 See Nelson, D.M., 15; 37 See Smith, S.L., 13, 19; 38 See Wilson, D.L., 31, 34; 38 Smock, T. See Arch, S., 188, 192; 271 Snedecor, G.W., 42, 43, 44, 45, 47, 52, 53, 57, 59, 61; 89

542

OCEANOGRAPHY AND MARINE BIOLOGY

Snider, R.H. See Kinzie, R.A., 48, 50, 78; 87 Sokal, R.R., 40, 47, 61, 64, 69, 73; 89 Solbe, J.F. de L.G., 566; 575 Solov’yeva, A.A. See Vedernikov, V.I., 20; 38 SooHoo, J.B. See Palmisano, A.C., 37 See Sullivan, C.W., 14; 38 Sörensson, F. See Rönner, U., 26; 38 Sousa, W.P., 226; 281 Soutar, B. See Dudek, F.E., 182; 274 Southard, J.B., 103; 111 See Middleton, G.V., 91, 105, 108; 111 See Nowell, A.R.M., 105; 111 See Paola, C., 103; 111 Southward, A.J., 48; 89 Southwood, T.R.E., 42, 51, 52, 53, 54, 72, 73, 74, 75; 89 Spangler, G.R. See Colby, P.J., 568; 574 Spannhof, L. See Rychly, J., 565; 575 Spärck, R., 120; 163 Squire, V.A. See Wadhams, P., 19, 21; 38 Stallard, M.O., 260, 262, 263, 264, 265; 281 See Faulkner, D.J., 261, 262, 265; 274 See Fenical, W., 274 See Ireland, C., 276 See Kinnel, R.B., 277 Stander, G.H., 360, 367, 372, 435, 436, 437, 491; 504 Stanley, S.O. See Pearson, T.H., 567; 575 Stanton, G.R. See Palmisano, A.C., 37 Statzner, B., 92; 111 Steene, R. See Hoese, D.F., 508, 509, 512, 526; 561 Stehouwer, H., 211; 281 Steinberg, J.B., 545; 562 Steinitz, H. See Clark, E., 561 Stephen, A.C., 120; 163 Stephenson, T.A., 411; 504 Sternberg, R.W., 107; 111 Stewart, F.H., 302, 303, 307, 308, 320; 350 Stewart, G. See Pechanec, J.F., 53; 88 Stewart, K.I., 567; 575 Stickle, W.B., 238; 281 Stimson, J., 71, 76, 78; 89 Stinnakre, J., 178, 208; 282 Stocker, L.J. See Andrew, N.L., 48, 77, 78; 85 Stocker, M. See Schweigert, J.F., 62; 89 Stoddart, D.R., 53; 89 Stoenner, R. See Walsh, J.J., 38 Stolzenbach, K.D. See Butman, C.A., 110, 135; 110, 157 Stone, B.A., 221; 282 Stoner, A.W. See Rey, J.R., 188, 190, 193, 194, 203; 280 Strathmann, R.R., 115, 131; 163

See Jackson, G.A., 132; 160 Streit, B., 236; 282 Strenth, N.E., 169, 172, 174, 175, 176, 190, 199, 200, 205, 220; 282 Stretch, J.J., 49; 89 Strumwasser, F., 192, 244, 245; 282 See Lu, C., 245; 278 Stubbings, H.G., 307, 315, 320, 326; 350 Suer, A.L., 131, 132; 163 Suganuma, M. See Fujiki, H., 275 Sugimura, T. See Fujiki, H., 275 Sulkin, S.D., 116, 155; 163 Sullivan, C.W., 14; 38 See Garrison, D.L., 13, 32; 36 See Palmisano, A.C., 32; 37 Sumer, B.M., 91; 111 Summersom, H.C. See Peterson, C.H., 143; 162 Sun, C.L. See Yang, R.-T., 403; 505 Sun, H.H. See Fenical, W., 274 Susswein, A., 209, 219; 282 Susswein, A.J., 173, 175, 178, 179, 180, 181, 183, 184, 185, 186, 187, 189, 192, 193, 205, 207, 209, 211, 214, 215, 217, 219, 220, 228, 229, 248, 249, 251, 254, 255; 282 See Achituv, Y., 173, 175, 178, 179, 181, 192, 193, 202, 206, 211; 271 See Gev, S., 181, 201; 275 See Gilboa-Garber, N., 270; 275 See Schwarz, M., 205, 215, 228; 281 Sutcliffe, W.H. See Sheldon, R.W., 490; 503 Suzuki, M. See Irie, T., 262; 276 Suzuki, T. See Irie, T., 276 Svane, J., 345; 350 Svendsen, E.A.S. See Johannessen, O.M., 37 Svennevig, N., 345; 350 Sverdrup, H.U., 14, 19; 38 Sweatman, H.P.A., 68, 69; 89 Swennen, C., 172; 282 Swinfen, R.C. See Edmunds, M., 211; 274 Switzer-Dunlap, M., 115, 173, 180, 181, 183, 184, 185, 188, 189, 190, 191, 192, 193, 194, 197, 198, 200, 205, 207, 208, 211, 213, 220, 227, 231, 243; 163, 282 See Hadfield, M.G., 187, 188, 189, 192, 198, 199, 200; 275 Sykes, J.B., 47; 89 Szelp, R. See Karplus, I., 507, 522, 523, 537, 548, 553, 555; 561 Szöllösi, A., 325; 350

AUTHOR INDEX

Taft, J.L., 564; 575 Taghon, G.L., 105, 110; 112 Taguchi, S. See Ackley, S.F., 25; 36 See El-Sayed, S.Z., 13, 25; 36 Tahira, T. See Fujiki, H., 275 Takahashi, F.T. See Kittredge, J.S., 256; 277 Takahashi, M., 568; 575 Takayama, S. See Fujiki, H., 275 Talbot, F.H., 402, 403, 425, 426, 427, 453, 454, 455; 504 See Russell, B.C., 48; 88 Tanaka, A., 266; 282 Tanaka, T., 260, 267; 282 Tangen, K., 568; 575 Tauc, L. See Hughes, G.M., 206; 276 See Stinnakre, J., 178, 208; 282 Taunton-Clarke, J. See Shannon, L.V., 433; 503 Tay, D. See O’Brien, D.P., 93; 111 Taylor, L.R., 47, 58, 75, 82, 83; 89 Taylor, W.D., 47, 58; 89 Taylor, W.R. See Taft, J.L., 564; 575 Teal, J.M., 129 130; 163 Tegner, M. See Dayton, P.K., 44; 86 Tennekes, H., 101; 112 Terada, T. See Yamamura, S., 268; 284 ter Maat, A. See Ferguson, G.P., 182; 274 Terre, J.J., 384, 423, 468, 469, 470; 504 Tett, P., 571; 575 See Gowan, R.J., 568; 574 See Jones, K.J., 574 See Turner, M.F., 564; 575 Thistle, D., 123, 140; 163 See Jumars, P.A., 80; 87 Thomas M.L.H., 117; 163 See Hughes, R.N., 121; 160 Thomas, R.M., 367, 370, 412, 413, 433, 434, 435, 436, 445, 461, 462, 476, 477, 478, 484, 486, 491; 504 See Boyd, A.J., 377, 436, 449; 496 See Le Clus, F., 435; 500 Thomassin, B.A., 509, 514; 562 Thompson, G.B. See Shin, P.K.S., 121; 163 Thompson, H.R., 81; 89 Thompson, T.E., 175, 190, 191, 194, 251, 252; 282 See Bebbington, A., 200; 272 Thompson, W.W., 396, 397, 398, 402; 504 Thomson, W., 307, 308, 309; 350 Thorson, G., 115, 119, 120, 123, 125, 126, 132, 136, 141, 142, 177, 194, 199; 163, 282 Tilton, B.E. See Winkler, L.R., 205, 216, 250, 251, 260, 264; 284

543

Titman, C.W. See Sigurdsson, J.B., 144; 163 Tobach, E., 175, 181, 254; 282, 283 See Lederhendler, I., 202, 252; 278 See Lederhendler, I.I., 175, 182, 184; 278 See Otsuka, C., 184, 200; 250 Tobe, S.S. See Dudek F.E., 182; 274 Toda, M. See Yamada, K., 260; 284 Todd, C.D., 82, 154; 89, 163 Toevs, L., 189; 283 Toft, C.A., 64, 65, 66, 84; 89 Tomiie, Y. See Matsuda, H., 260; 279 Tomlinson, J.T., 286, 287, 288, 289, 292, 293, 295, 296, 297, 337, 338, 339, 346, 347; 350 See Batham, E.J., 292; 347 Tosi, L. See Ghiretti, F., 224; 275 Tourtellotte, G.H. See Dauer, D.M., 118; 157 Townsend, A.A., 103; 112 Toyama, Y. See Tanaka, T., 260, 267; 282 Tranter, D.J., 119; 163 Tranter, H.A. See Tranter, D.J., 163 Tringali, C. See Danise, B., 273 See Fattorusso, E., 274 Tritton, D.J., 97; 112 Tritt, S.H., 256; 283 Troadec, J.-P., 365, 461, 478, 491, 494; 504 Trueblood, D.D. See Gallagher, E.D., 123, 124, 138, 139; 158 Trueman, E.R., 154; 163 Truesdale, R.S., 34; 38 Tsuge, M. See Sakata, K., 281 Tsukayama, I. See Boerema, L.K., 496 Tsunawaki, S. See Yamazaki, M., 254 Tsunogai, S. See Noriki, S., 33; 37 Tsurnamal, M. See Karplus, I., 507, 522, 523, 537, 548, 553, 555; 561 Tsutsumi, H., 567; 575 Tudor, J. See Papka, R., 201; 280 Tunnell Jr, J.W., 174, 177; 253 Turner, J.T. See El-Sayed, S.Z., 30; 36 Turner, M.F., 564; 575 Turquier, Y., 287, 289, 293, 295, 296, 297, 337, 339, 343; 350 See Pochon-Masson, J., 325; 350 Tyler, P.A., 121, 126, 142, 153; 163 Ueda, H. See Nishiwaki, S., 203, 237; 279 Underwood, A.J., 44, 45, 46, 47, 53, 57, 61, 63, 64, 65, 66, 68, 77, 78, 80, 83, 84; 89 See Fairweather, P.G., 48; 86

544

OCEANOGRAPHY AND MARINE BIOLOGY

See Kennelly, S.J., 48, 49, 50, 52, 62, 63, 70; 87 Untersteiner, N. See Smith, D.C., 15; 38 Uozumi, Y., 411, 417, 425, 440, 441, 449, 450, 458, 464, 467, 469; 504 Upton, G.J.G., 81, 82; 89 Usher, M.B., 81; 89 Ussing, H. See Thorson, G., 120; 163 Usuki, I., 172, 174, 176, 179, 181, 182, 188, 189, 191, 194, 201, 205, 207, 211, 236; 253 Utinomi, H., 289, 290, 291, 292, 319, 337, 344; 350, 351 Vadas, R.L. See Paine, R.T., 226; 280 Vaghin, V.L., 345; 351 Vagin, V.L., 345; 351 Valdivia, J.E. See Boerema, L.K., 496 VanBlaricom, G.R., 124, 137, 143, 153; 163 Van Den Berg, R.A., 426, 427, 428, 453, 454, 455; 504 Vanderah, D.J., 265; 283 See McDonald, F.J., 279 See Schmitz, F.J., 262; 281 Van Der Elst, R., 400; 504 Van der Helm, D. See Hollenbeak, K.H., 276 See McDonald, F.J., 279 See Schmitz, F.J., 281 Van Dyke, M., 92; 112 Van Eck, T.K. See Jones B.W., 379, 380, 382, 467; 499 Vanell, L.D. See Pettit, G.R., 280 Van Engen, D. See Kinnel, R.B., 277 van Leer, J., 35; 38 See Manley, T.O., 37 Van Fleet, E.S., 50; 89 Van Weel, P.B., 175, 178, 179; 283 Van Wyk, G.F., 445; 504 See De Jager, B. van D., 403, 425, 426, 427, 428, 453; 498 Vaske, B. See Babayan, V., 496 Vaughan, D.S. See Saila, S.B., 52, 63; 88 Vedernikov, V.I., 20; 38 Veillet, A., 331, 332, 333, 334, 335, 336; 351 See Durand, D., 334; 348 Velimirov, B. See Field. J.G., 449 Venediktova, L.I., 373, 413; 504 Venidictova, L.I., 413; 504 Venrick E.L., 62; 89 Vens-Cappell, B. See Butz, I., 565; 574 Venter, J.D., 415, 416, 417, 418, 484; 504 Vercheson A. See Karplus, I., 507, 508, 509; 561 Verwey, J., 141, 142; 163 Vevers, H.G. See Kennedy, G.Y., 227; 277

Vicente, N., 191, 200, 270; 283 See Poizat. C, 177; 280 Vijverberg, J. See Nie, H.W. de, 49; 88 Villanueva, J.Z. See Manley, T.O., 37 Vinje, T.E., 20; 38 Virnstein, R.W., 124; 163 See Boesch, D.F., 124; 157 Vitalis, T.Z., 205, 207, 211, 213, 224, 233; 283 Vogel, S., 91, 105, 145; 112, 163 Von der Porten, K., 205, 247, 248; 283 See Pinsker, H., 280 Von Dreele, R.B., 266; 283 See Pettit, G.R., 280 Vozarik, J.M. See Dobbs, F.C., 119, 144, 154; 158 Wachtel, H., 243, 244; 283 Wodhams, P., 19, 21; 38 Waldo, R. See Thistle, D., 140; 163 Walford, L.A., 231; 283 Walker, G., 325, 331, 332; 351 Walley, L.J., 339; 351 Walsh, J.J., 33; 38 Walters, E.T., 244, 259; 283 See Carew, T.J., 258; 273 See Hening, W.A., 243; 276 Ward, J., 234; 283 Warmke, G.L., 174, 176; 283 Warren, C.E., 238; 283 See McIntire, C.D., 100; 111 Warrer-Hanson, I., 565, 566, 569; 575 Waser, P.M., 249; 283 Washecheck, D.M. See McDonald, F.J., 279 Waters, V.L. See Edmunds, M., 211; 274 Watling, L., 53; 89 Watson, M., 260, 261, 262, 268; 283 Watzon, M.C., 124, 125, 135, 138; 163, 164 Wawra, E. See Hadl, G., 130; 159 Webb, C., 134; 164 Webb, F.J., 517; 562 Webb, J.E., 129, 130; 164 Wefer, G., 33; 38 Weiler, D.A., 509, 513, 514, 517, 520, 522, 523, 526; 562 Weinberg, J.R., 124; 164 Weinberg, S., 48, 50, 83; 90 Weinstein, M.P. See Heck Jr, K.L., 252; 276 Weir, G. See Dudek, F.E., 274 See Rothman, B.S., 192; 280 Weiss, K.R., 209; 283 See Dieringer, N., 179, 205, 253; 273

AUTHOR INDEX

See Koch, U.T., 209; 277 See Kupfermann, I., 178, 229; 278 See Kuslansky, B., 219, 220; 278 See Rosen, S.C., 209; 280 See Susswein, A., 209; 282 See Susswein, A.J., 205, 209, 214, 228; 282 Weiss, S. See Austin, T., 216; 271 Welch, H.E., 236; 283 Wellham, L.L. See Sigel, M.M., 281 Wells, D.W., 175, 216; 283 Wells, L.J., 208; 283 See Jahan-Parwar, B., 209; 277 Wells, R.J. See Kazlauskas, R., 263; 277 Welsh, J.G., 403; 504 Wethey, D.S., 114; 164 Weyler, V. See González, A.G., 275 Whitaker, T.M., 13, 32; 38 White, D.C. See Palmisano, A.C., 37 White, J.D., 266; 283 White, J.F. See Gerencser, G.A., 221; 275 Whitelegge, T., 127; 164 Whitlatch, R.B., 121, 125, 138, 140; 164 See Zajac, R.N., 124, 137; 165 Wiebe, P.H., 48, 50, 51, 53, 55, 56, 57; 90 Wiederhold, M.L. See Gallin, E.K., 249; 275 Wiens, J.A., 44; 90 See Rotenberry, J.T., 64, 65, 66, 68; 88 Wieser, W., 120, 128, 130; 164 Wightman, J.A., 239; 283 Wigley, R.L., 120; 164 Wilde, P.A.W.J. de. See Farke, H., 143, 154; 158 Wiley, R.H., 547; 562 Wilke, C.G. See Beyers, C.J. de B., 410, 457; 496 Willan, R. 169, 172, 176, 177, 226; 283 Willan, R.C., 172, 175, 176, 177, 180, 188, 190, 191, 192, 193, 200, 201, 207, 213, 218, 219, 222, 225, 226, 228, 233, 236, 237, 238, 240, 243, 244, 246, 250, 251, 252, 253, 255, 256; 283 Willey, J.C., 261; 283 Williams, A.B., 117, 129, 130, 131; 164 Williams III, A.J. See Grant, W.D., 147; 159 Williams, B., 83; 90 Williams, G.P., 91; 112 Williams, J.G., 123, 137, 139; 164 Williams, P.M. See Van Fleet, E.S., 50; 89 Williams, W.T. See Abel, D.T., 80; 85 Willoughby, H., 566; 575 Wilson, D.L., 31, 32, 33, 34; 38 See Smith Jr, W.O., 20; 38

545

See Smith, S.L., 13, 19; 38 Wilson, D.P., 115, 123, 125, 127, 128, 130, 131, 132, 155; 164 See Day, J.H., 131; 158 Wilson, E.O., 540, 547; 562 Wilson, J.G., 76, 78; 90 Wilson Jr, W.H., 124; 164 Wimbush, M., 145; 164 Winer, B.J., 43, 44, 45, 47, 48, 61, 64, 66, 68; 90 Wing, R.M. See Hirschfeld, D.R., 276 See Midland, S.L., 279 See Sims, J.J., 263; 281 Wingate, G.H.L. See Crawford, R.J.M., 363; 497 Wingstrand, K.G., 325; 351 Winkler, L.R., 174, 177, 191, 205, 206, 207, 210, 211, 214, 215, 216, 223, 226, 227, 231, 236, 250, 251, 253, 254, 256, 260, 261, 262, 264; 283, 284 Winstanley, R.H., 408; 504 Wise, J.P., 472; 504 See Lima, F.R., 453; 500 See Miyake, P.M., 405; 501 Withers, T.H., 337; 351 See Newman, W.A., 285, 337, 340, 344; 349 Woiwod, I.P. See Taylor, L.R., 75, 82; 89 Wolf, C.C. See Mann, R., 134, 155; 160 Wolf, P. de, 141; 164 Wolfe-Murphy, S. See McLusky, D.S., 143; 161 Wolff, H.G., 249; 284 Wood, E.J.F., 141; 164 Woodin, S.A., 114, 124, 125, 138; 164 Wooding, R.A., 153; 164 Wooster, W.S., 353; 505 Work, R.C. See McNulty, J.K., 119, 120; 161 Wormuth, J.H. See Shulenberger, E., 34; 38 Wozniak, J.A. See Lickey, M.E., 245; 278 Wright, D.L., 173, 177; 284 Wright, H.O., 174, 181, 190; 284 Wrzesinski, O., 413, 449, 450; 505 Wysokinski, A., 476, 484; 505 See Babayan, V., 496 Wyszynski, M. See Chlapowski, K., 379; 479 Xulu, S. See Butterworth, D.S., 494; 497 Yaglom, A.M., 146, 147; 164 Yalin, M.S., 94, 152; 112, 164 Yamada, K., 260; 284 Yamaguchi, M., 132; 164 Yamaizumi, Z. See Fujiki, H., 275

546

OCEANOGRAPHY AND MARINE BIOLOGY

Yamamura, S., 260, 261, 262, 267, 268; 284 See Matsuda, H., 260; 279 Yamane, T., 64; 90 Yamashita, K. See Tanaka, A., 282 Yamazaki, M., 269, 270; 284 Yamazato, K. See Fujiki, H., 275 Yanagimachi, R., 331, 332, 333; 351 See Ichikawa, A., 331, 333, 334, 345; 349 Yanagisawa, Y., 507, 508, 509, 512, 513, 514, 516, 517, 518, 519, 520, 521, 522, 523, 526, 528, 529, 531, 532, 533, 534, 535, 536, 537, 538, 539, 540, 541, 542, 552, 558, 559; 562 Yang, R.-T., 403; 505 Yasunari, Y. See Irie, T., 276 Yazawa, H. See Yamada, K., 260; 284 Yearsley, J.R. See Millard, S.P., 44; 88 Yingst, J.Y. See Thistle, D., 123; 163 Yonge, C.M., 202, 208; 284 Yoshioka, A. See Fujiki, H., 275 Young, C.M. See Chia, F.-S., 114, 134, 148, 155; 157 Young, D.K., 124; 164 See Rhoads, D.C., 122, 124, 140; 162 Young, M.W. See Young, D.K., 124; 164 Young, R.A., 109; 112 Young, R.J. See Alden, R.W., 62; 85 Youngbluth, M.J., 49; 90 Zajac, R.N., 124, 137; 165 See Whitlatch, R.B., 138; 164 Zalewski, M. See Penczak, T., 575 Zalinski, J. See Bernstein, B.B., 46, 61, 64, 68; 85 Zander, C.D., 512, 513, 514, 537; 562 Zanevald, J.R.V. See Bartz, R., 156 Zenkin, V.S., 375; 505 Zentara, S.-J. See Kamykowski, D., 31; 37 Zenziper, Z. See Barash, A., 175, 190, 191, 194; 271 Zevina, G.B., 337, 338; 351 Ziegelmeier, E., 125; 165 Ziegler, A. See Tobach, E., 254; 283 Zigmand, M. See Tritt, S.H., 256; 283 Zimmer, W.J. See Johnson, R.B., 71, 77; 87 Zimmerman, R.C. See Hammer, R.M., 118; 159 Zinn, D.J., 177; 284 Zoutendyk, P., 388, 413, 451; 505 See Field, J.G., 499 Zullo, V.A. See Newman, W.A., 285, 337, 340, 344; 349 Zwally, H.J., 11, 17, 21, 24, 30; 38 Zwart, P.R. See O’Brien, D.P., 93; 111 Zwerner, D.E. See Dillon, W.A., 329; 348

SYSTEMATIC INDEX

References to complete articles are given in heavy type; references to pages are given in normal type. Abralia gilchristi, 426 Acanthohaustorius, 129 millsi, 129 Acanthophora, 175 spicifera, 200, 207 Acanthurus nigrofuscus, 540, 542 Acentrogobius pflaumi, 512, 559 Acrochaetium, 173 Acroscalpellida, 341 Acroscalpellum, 310, 337, 341 Acrosorium uncinatum, 206 Acrothoracica, 285, 286, 287, 289, 290, 293, 295, 296, 337, 338, 339, 340, 342, 343, 344, 345 Actinopyga agassizi, 255 Akentrogonida, 328, 329 Aktedrilus monospermatecus, 128 Alciopidae, 118 Alcippe, 288, 292 lampas, 296 lampas (=Trypetesa lampas), 293 (=Trypetesa), 288, 295 Alepisaurus ferox, 426 Alepocephalus rostratus, 430 Alopias vulpinus, 425 Alpheus, 515, 554, 555 bellulus, 514, 515, 517, 518, 519, 520, 521, 522, 523, 526, 528, 529, 532, 534, 535, 536, 537, 542, 554, 555, 559 brevicristatus, 514, 559 brevirostris, 514, 515 crassimanus, 508, 514, 518, 520, 523 djiboutensis, 514, 515, 518, 519, 520, 522, 526, 527, 529, 532, 540, 541, 543, 546, 547, 554, 555, 556, 557 floridanus, 514, 517, 520, 522, 526, 559, 560 malabaricus, 508, 514 ochrostriatus, 514, 515, 554, 555

purpurilenticularis, 514, 515, 516, 518, 519, 520, 522, 524, 525, 526, 529, 535, 540, 543, 545, 547, 549, 550, 554, 555, 556, 557 randalli, 508, 514, 551 rapacida, 514, 520, 522, 523, 526, 534, 536, 545 rapax, 514, 515, 516, 518, 519, 520, 522, 523, 526, 529, 534, 536, 543, 545, 554, 555, 557 rubromaculatus, 514, 515, 518, 519, 520, 554, 555 Amblyeleotris, 508, 526, 553 aurora, 512 callopareia, 512 delicatulus, 512 diagonalis, 512 exilis, 512 fasciata, 512 fontanesii, 512 guttata, 512 gymnocephala, 512 japonica, 512, 521, 523, 526, 528, 529, 532, 533, 534, 535, 537, 538, 541, 542, 558 japonicus, 512 latifasciata, 512 macronema, 512 maculata, 512 ogasawarensis, 512 periopthalma, 512 randalli, 512 ryhax, 512 steinitzi, 509, 510, 512, 518, 523, 524, 525, 526, 528, 532, 535, 540, 541, 543, 544, 545, 547, 549, 550, 552, 554, 555, 556, 557, 560 sungami, 512 wheeleri, 512 Amphictenidae, 118 Amphipoda, 416, 420, 421, 422 Anostraca, 285 547

548

OCEANOGRAPHY AND MARINE BIOLOGY

Anthessius variedens var. aplysiae, 270 Anthopleura xanthogrammica, 250, 251, 257, 260 Aparrius aurocingulus, 513 Aplysia, 167–284; 167, 168, 169, 171, 172, 174, 178, 179, 181, 182, 183, 185, 186, 188, 189, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 208, 209, 210, 211, 214, 215, 217, 219, 220, 221, 223, 225, 226, 227, 228, 231, 232, 234, 235, 236, 237, 238, 239, 240, 244, 247, 248, 249, 250, 251, 252, 253, 254, 255, 256, 257, 258, 259, 261, 262, 264, 269, 270 brasiliana, 174 brasiliana=willcoxi, 169, 173, 179, 180, 181, 182, 183, 184, 186, 188, 190, 192, 193, 194, 199, 200, 201, 203, 205, 206, 211, 220, 227, 229, 230, 232, 238, 243, 246, 247, 248, 249, 250, 251, 253, 254, 255, 263, 264, 270 californica, 172, 173, 174, 179, 181, 182, 183, 185, 186, 187, 188, 189, 190, 192, 193, 194, 197, 199, 200, 201, 202, 203, 204, 205, 206, 208, 209, 210, 211, 212, 214, 215, 216, 217, 219, 220, 221, 226, 227, 229, 231, 232, 234, 236, 238, 244, 245, 246, 249, 250, 251, 252, 253, 254, 256, 257, 258, 259, 260, 261, 262, 263, 264, 265, 269, 270 cedrosensis, 173, 174, 253 cervina, 173, 174, 190, 204, 206 cornigera, 174 cronullae, 174 dactylomela, 169, 171, 172, 173, 174, 178, 180, 181, 182, 183, 184, 185, 186, 187, 188, 190, 192, 194, 195, 196, 197, 200, 201, 203, 204, 205, 207, 208, 211, 213, 214, 215, 216, 217, 218, 219, 220, 222, 223, 224, 225, 226, 227, 228, 232, 233, 234, 235, 239, 240, 241, 242, 243, 244, 245, 246, 247, 250, 251, 252, 253, 254, 255, 256, 262, 263, 265, 266, 267 denisoni, 175 depilans, 173, 175, 179, 182, 190, 192, 193, 194, 198, 201, 202, 205, 206, 211, 215, 224, 226, 227, 247, 253, 256, 257, 263, 267, 270 dura, 169, 173, 175, 244, 253 euchlora, 171, 172, 175 extraordinaria, 173, 175, 247, 253 fasciata, 171, 173, 175, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 190, 192, 193, 194, 198, 201, 202, 205, 207, 211, 215, 224, 227, 228, 244, 247, 248, 249, 251, 253, 254, 255, 263, 267, 270 geographica, 173, 246 gigantea, 175 gracilis, 169, 175 inca, 175

Juliana, 169, 171, 172, 173, 175, 178, 179, 180, 182, 185, 188, 191, 192, 194, 195, 196, 197, 199, 200, 201, 204, 205, 207, 208, 211, 213, 214, 217, 218, 221, 222, 223, 224, 225, 226, 227, 228, 231, 232, 234, 235, 236, 237, 239, 240, 241, 242, 244, 246, 247, 250, 251, 253, 260, 262, 269, 270 keraudreni, 176, 207 kurodai, 176, 179, 182, 188, 189, 191, 193, 194, 201, 207, 211, 213, 218, 219, 223, 225, 237, 238, 257, 260, 261, 262, 267, 268, 270 maculata, 173, 176, 247 morio, 170, 173, 176, 247 nigra, 173, 176, 247, 253 oculifera, 176, 181, 189, 191, 198, 205, 215, 228, 268, 270 parvula, 169, 171, 172, 176, 182, 188, 189, 191, 194, 198, 200, 207, 208, 225, 226, 227, 247, 255, 268 pulmonica, 177, 178, 191, 208, 247, 260, 261, 268 punctata, 169, 172, 177, 178, 179, 184, 188, 191, 193, 194, 195, 196, 197, 200, 201, 202, 203, 204, 205, 207, 210, 211, 213, 214, 218, 220, 221, 222, 223, 226, 227, 228, 233, 234, 235, 236, 238, 239, 240, 247, 251, 263, 270 rehderi, 177 reticulata, 177 reticulopoda, 177 robertsi, 177 rosea, 202 rudmani, 177 sagamiana, 177 sowerbyi, 177 sydneyensis, 177, 270 tanzanensis, 173, 177, 247 vaccaria, 169, 173, 177, 180, 191, 194, 204, 207, 221, 231, 253, 263, 264, 268, 270 willcoxi, 174, 177 winneba, 177, 247 Aplysiidae, 169 Apygophora, 337 Arctocephalus pusillus, 415, 429 Arenicola marina, 117, 118 Argonauta, 426, 427 Argulus, 325 Argyrosomus hololepidotus, 392, 399, 400 Argyrozona argyrozona, 399, 400, 413 Aricidea, 118 Armandia, 119, 135 brevis, 118, 135, 136, 137 Ascothoracia, 344

SYSTEMATIC INDEX

Ascothoracica, 339, 345 Asterina, 251 Astichopus multifidus, 253 Astrometis sertulifera, 252 Atractoscion aequidens, 399, 400 Aulacomya ater, 425, 429 Aurivillialepas, 338, 341 Australophialus melampygos, 292 pecorus, 293, 295 turbonis, 293 Austroglossus, 378, 393, 394, 395 microlepis, 388, 394, 395, 413, 452, 469, 471, 483 pectoralis, 388, 394, 395, 396, 413 Autolytus, 118 Auxis thazard, 404, 405, 427 Balanodytidae 339 Balanoidea, 340 Balanomorpha, 286, 298, 301, 337, 340 Balanomorphoidea, 340 Balanus, 298, 301 balanoides, 325 calceolus, 298, 301 (Conopea) calceolus, 298 (Conopea) galeatus, 298 (Conopea) merilli, 298 galeatus, 298, 345 improvisus, 328 masignotus, 298, 301 merilli, 298 perforatus, 325 (Solidobalanus) masignotus, 298 tintinnabulum, 326 Bathygobius curacao, 512, 559, 560 Bathylasma alearum, 299, 301 corolliforme, 298 Bathylasmatinae, 340 Bathynectes piperitus, 424 Batman insignitus, 513 Berndtia, 286, 291 nodosa, 291, 292 purpurea, 289, 290, 291, 292, 298, 337, 342, 343 Berndtiidae, 339 Bonellia viridis, 285 Boschmaella, 328 balani, 328 Brama brama, 390 raii, 426 Branchiostoma nigeriense, 129

Brania, 118 Bryopsis adriatica, 207 Bufo ictericus, 264 Burnupena, papyracea, 429 Bursatella, 169 Butis butis, 512 Calanoides, 415 carcinatus, 414, 483 Calantica, 301, 302, 304, 338, 340, 341 calyculus, 301, 302, 307, 314 eos, 301 falcata, 301 gemma, 301 grimaldi, 301 spinilatera, 302 spinosa, 341 studeri, 302 superba, 301 trispinosa, 301, 302, 304 villosa, 301, 302, 341 Calanticinae, 338 Calanus, 415 finmarchicus, 33 hyperboreus, 33 Calappa, 251 Callianassa islandgrande, 129 jamaicense louisianesis, 129 Callinectes sapidus, 255, 264 Callithamnion byssoides, 200 Callorhinchus capensis, 407 Capitella, 128, 131, 135, 137 capitata, 135, 568 Carcharinus obscurus, 425 Carcinus maenas, 336 Caretta caretta, 251 Cassiopea xamachana, 252 Caulerpa racemosa, 200 Centroceras clavulatum, 206, 207 Centropages, 415 brachiatus, 414, 483 Cephalopoda, 416, 421, 422, 424 Ceramiales, 200, 209 Ceramium eatonianum, 206, 210 Chaetoceros, 205, 414, 418, 485 Chaetodon chrysurus, 540 Chaetognatha, 420 Champia laingii, 213 parvula, 175, 206

549

550

OCEANOGRAPHY AND MARINE BIOLOGY

Cheimerius nufar, 399 Chelidonichthys, 390, 483 Chelonibiinae, 340 Chelonobia, 285, 299 patula, 299, 301, 345 Chilomycterus antennatus, 252 Chionelasmus darwini, 299, 301 Chlorophyceae, 196, 218, 222, 232, 240 Chondria californica, 206 cnicophylla, 263 leptacremon, 206 Chondrococcus, 200 hornemanni, 200 Chondrus, 200 Choromytilus meridionalis, 429 Chromadorita tenuis, 129 Chrysaora, 486 Chrysoblephus, 399 cristiceps, 450 laticeps, 399, 450 Chthamalophilidae, 329, 333 Chthamalophilus, 328 delagei, 328 Chytraeidae, 339 Cirripedia, 317, 320, 324, 326, 339, 344 Cladocera, 285 Cladophora, 196, 207, 211, 213, 216, 217, 218, 222, 232, 233, 234, 241, 243 Cladophorales, 209 Clistosaccidae, 328, 329, 333 Clistosaccus, 328, 329, 332, 333 paguri, 331, 332, 333, 336 Clorophthalmus atlanticus, 424 Codium, 206 fragile, 212, 213 Coelenterata, 420 Coelogynopora schulzii, 128 Coelorhynchus, 421, 422 fasciatus, 424, 425 flabellispinus, 424 Colpomenia, 206 Conchostraca, 285 Conus pennaceus, 251 Convoluta, 129 Copepoda, 416, 420, 421, 422 Coracinus capensis, 399 Corallina, 207 gracilis, 206 pinnatifolia, 206

vancouveriensis, 206 Corophium arenarium, 129 volutator, 129 Coscinasterias calamaria, 250, 251, 252, 255 Coscinodiscus, 414, 417, 418 Cottuncoloides macrocephalus, 430 Crustacea, 115, 327, 416, 420 Cryptocentrus, 508, 526 albidorsus, 512 caeruleomaculatus, 512 caeruleopunctatus, 510, 512, 523, 526, 528, 532, 533, 535, 537, 556, 558 cinctus, 512, 535 cryptocentrus, 510, 513, 527, 540, 541, 546, 547, 554, 556, 557 fasciatus, 513, 535 filifer, 513 inexplicatus, 513 insignitus, 513 leucostictus, 513 lutheri, 510, 513, 518, 520, 523, 526, 528, 537, 543, 554, 556, 557 malindiensis, 513 maudae, 513 nigrocellatus, 513 niveatus, 513 obliquus, 513 oni, 514 shigensis, 513 steinitzi, 512 strigilliceps, 513 sungami, 512 wheeleri, 512 Cryptophialidae, 339 Cryptophialus, 291, 292, 343 cordylacis, 293 coronatus, 292 heterodontus, 293 lanceolatus, 293 melampygos, 286, 292 minutus, 292 minutus striatus, 292 newmani, 293 unguiculus, 293 variabilis, 293 wainwrighti, 293, 295 Cryptopleura, 195, 210, 236 ramosa, 188, 196, 207, 210, 213, 222, 233, 234, 240 Ctenogobiops, 508, 526

SYSTEMATIC INDEX

aurocingulus, 513 crocineus, 513 feroculus, 513, 523, 532 maculosus, 510, 513, 518, 528, 535, 543, 544, 554, 555, 557 pomastictus, 513, 532, 535 tangaroai, 513 Cumacea, 416 Cumella vulgaris, 128 Cynoglossidae, 388 Cynoglossus zanzibarensis, 388 Cyphosaccus, 322, 333 norvegicus, 332 Cyprilepas, 338 Cystoseira osmundacea, 212 Dardanus, 251 Dasya, 173, 200, 232 Decapoda, 416, 421, 422, 424 Delesseria, 195, 207, 210, 236 sanguinea, 188, 196, 207, 210, 211, 213, 221, 222, 233, 234, 240 Delphineis, 418 (=Fragilaria) karstenii, 414, 417 karstenii, 418, 432, 485 Dendraster excentricus, 133 Dentex, 387 angolensis, 386, 387 macrophthalmus, 385, 387, 413, 471 Desmarestia, 177, 188 aculeata, 188, 233, 234 Diaphus, 417, 418, 424 Dictyopteris zonarioides, 212 Dictyota, 173, 177, 188, 263, 430 crenulata, 200 dichotoma, 263 Dictyotaceae, 263, 269 Dilophus ligulatus, 263 Diopatra ornata, 118 Diplanthera wrightii, 205, 206 Discomedusae, 486 Dolabella, 169 Donax serra, 410 variabilis, 117 Drepanorchis neglecta, 329 Dunaliella tertiolecta, 205 Duplorbis, 329 Echinaster, 253

551

sentus, 252 Echinometra lincunter, 264 Ecklonia cava, 213 maxima, 430 Ectocarpus, 206 Edriolychnus, 285 Egregia, 177, 204, 206, 207 laevigata, 254 menziesii, 212 Eilatia latruncularia, 511, 513, 514, 535 Eiriocheir, 336 Eisenia, 206 arborea, 212 Eleotris periophthalmus, 512 Endarachne binghamiae, 213 Engraulis, 361, 368, 415 japonicus, 354, 357, 370, 413, 414, 419, 420, 421, 422, 426, 427, 428, 434, 460, 464, 465, 474, 476, 482 mordax, 415 ringens, 353 Ensis directus, 117 Enteromorpha, 176, 177, 195, 196, 200, 206, 207, 209, 210, 211, 213, 218, 222, 223, 224, 227, 232, 233, 234 compressa, 176, 207 intestinalis, 196, 207, 210, 212, 213, 218, 221, 222, 233, 234, 240 linza, 207 Eolepas, 338 Erichthonius braziliensis, 76 Eteone lactea, 117 longa, 117 Etrumeus, 361, 368 whiteheadi, 354, 355, 413, 419, 420, 421, 422, 427, 428, 442 Eunoe, 118 Euphausia superba, 11, 34 Euphausiacea, 416, 420, 421, 422 Eurydice affinis, 129 pulchra, 129 Euscalpellum, 301 bengalense, 301 renei, 301 rostratum, 301 squamuliferum, 301 stratum, 301 Eurhynnus alletteratus, 404, 405 Evechinus chloroticus, 180 Exocoetidae, 426 Exogone, 118, 119

552

OCEANOGRAPHY AND MARINE BIOLOGY

lourei, 119 Fabricia sabella, 129 Fasciolaria, 252, 253 tulipa, 252 Favorinus japonicus, 251 Ferrissia rivularis, 197 Flabelligobius fourmanoiri, 513 Fucus, 177, 203, 207 serratus, 177 Funchalia woodwardi, 426, 427 Galaxaura oblongata, 213, 222, 232 Gelidium, 200, 206 coulteri, 206 pulchellum, 207 purpurascens, 213 pusillum, 207 Gemmosaccus sulcatus, 331, 334 Genypterus capensis, 382, 384, 390, 413, 424, 450, 469, 470, 483 Geryon, 407 maritae, 410 Gigartina, 206, 210 acicularis, 207 armata, 212 canaliculata, 206 Gigartinales, 209 Glycera, 117 capitata, 117 dibranchiata, 117, 118 Glyphis glaucus, 425 Gnathia, 285 Gobiidae, 421, 422 Gobionellus saepeplallens, 513 stigmalophius, 513 Gobiosoma fasciatum, 513 Gobius caeruleopunctatus, 512 cryptocentrus, 513 delagoae, 514 filifer, 513 fontanesii, 512 gymnocephalus, 512 leucostictus, 513 mystacina, 513 nudiceps, 513 pflaumi, 512 Gracilaria, 175, 200, 206, 207, 216, 241, 242, 263 verrucosa, 263

Grateloupia filicina, 213 okamurai, 213 tsurutsuru, 213 Gruvelialepas, 341 Gyptis brevipalpa, 118 vittata, 118 Gyrodinium aureolum, 564 Halimeda opuntia, 200 Haliotis midae, 410, 411 Halopteris scoparia, 207 Haustorius, 129 canadensis, 129 Helicolenus, 391 dactylopterus, 424, 469 maculatus, 413, 424, 426 Hemerocallis fulva, 215, 228 Hemirihamphidae, 426 Heptatretus hexatrema, 429 Herposiphonia, 213 Heterosiphonia, 195, 210 plumosa, 196, 207, 210, 213, 221, 222, 233, 234, 240 Histioteuthis bonnellii 426, 427 Homarus americanus, 128 Hydrobia ventrosa, 239 Hypnea, 176, 430 musciformis, 206 valentiae, 206, 210 Hypoglossum, 173 Ibla, 314, 320, 325, 326, 337, 338, 339, 340, 341, 343, 344 atlantica, 320, 326, 327 cumingi, 320, 321, 323, 324, 325, 326, 337, 342, 344, 346 idiotica, 320, 323, 327, 340 pygmea, 320, 326, 327 quadrivalvis, 320, 324, 326, 345 segmentata, 320 Iblidae, 320, 337, 340, 342 Inachus, 336 Iotogobius malindiensis, 513 Isochrysis galbana, 205, 208, 220 Isurus glaucus, 425 Jania rubens, 207, 228 tenella, 206 Jasus, 408 edwardsii, 408

SYSTEMATIC INDEX

lalandii, 407, 408, 425, 429, 430 novaehollandiae, 408 tristani, 408 Katsuwonus pelamis, 403, 404, 405, 428, 452, 453, 455, 472, 489 Kentrogonida, 328, 329 Kochlorine, 290, 339 bocqueti, 286, 290, 295, 296, 341, 343 floridana 286, 289, 295, 343 hamata, 286, 289, 342, 343 ulula, 286 Kochlorinidae, 339 Kochlorinopsis, 339 discoporellae, 290, 342 Labridae, 540 Lactuca saliva longifolia, 232 Laminaria, 197, 206, 207, 210, 212, 227, 236 digitata, 176, 196, 210, 211, 213, 222, 233, 234, 240 farlowii, 212 pallida, 430 Lampanyctodes, 361, 368 hectoris, 355, 413, 420, 442, 444 Larus atricilla, 251, 255 Laurencia, 172, 177, 200, 206, 207, 209, 210, 213, 214, 218, 222, 224, 226, 228, 232, 233, 234, 240, 242, 243, 255, 257, 262, 263, 269 distichophylla, 200 gemmifera, 206 hybrida, 207 johnstonii, 263 nipponica, 257 obtusa, 263 okamurai, 262 pacifica, 199, 200, 206, 210, 212, 214, 263 papillosa, 175, 196, 207, 213, 216, 217, 222, 232, 241 pinnatifida, 207 Lepadidae, 295 Lepadomorpha, 285, 298, 299, 307, 337, 344 Lepas, 311 Lepidactylus dytiscus, 129 Lepidoptera, 309 Lepidopus caudatus, 392, 424, 426, 427 Leptastacus constrictus, 129 Lepus marinus, 167 Lernaeodiscidae, 328, 329, 332, 333 Lernaeodiscus, 329, 334 crenatus, 329

553

galatheae, 329, 331 porcellanae, 331, 334, 335 Lichia amia, 399 Lithoglyptes, 286, 288, 343 habei, 288 hirsutus, 288 indicus, 286, 288 mitis, 288 scamborachis, 288, 343 spinatus, 288, 342, 343 wilsoni, 288, 295, 343 Lithoglyptidae, 286, 339, 343 Lithognathus lithognathus, 407 Lithothrix aspergillum, 206 Lithotyra, 338 Lithotryinae, 338 Littorina, 197 Loimia, 118 Loligo, 411 reynaudii, 387, 392, 410, 419, 426, 427, 428, 457, 469 Lomentaria articulata, 200 Lophius, 385, 391, 469 Lophosiphonia, 206 Lotilia graciliosa, 511, 513, 516, 518, 554 Loxothylacus panopaei, 329 Ludwigothuria grisea, 264 Lumbrineris, 118 Lycoteuthis diadema, 427, 428 Lygdamis muratus, 128 Lyngbya majuscula, 261 Macoma balthica, 117 Macrocystis, 204, 206 pyrifera, 212, 213 Macrouridae, 424 Macrura, 456 Mahidolia mystacina, 513 Makaira, 425 Malacostraca, 285 Mars albidorsus, 512 caeruleomaculatus, 512 nigrocellatus, 513 strigilliceps, 513 Martensia, 200 Maurolicus, 424 muelleri, 413, 416, 426, 427, 444 Maxillopoda, 344 Mediomastus ambiseta, 119 Melo amphora, 251

554

OCEANOGRAPHY AND MARINE BIOLOGY

Mercenaria mercenaria, 128, 143 Merluccius, 354, 368, 378, 380, 388, 390, 399, 413, 420, 421, 422, 423, 426, 427, 428, 470, 477 capensis, 379, 380, 413, 420, 421, 422, 423, 424, 425, 426, 439, 445, 446, 447, 448, 449, 467, 468, 469, 484 paradoxus, 379, 380, 413, 421, 422, 423, 424, 439, 445, 446, 447, 448, 467, 468, 469, 484 polli, 378, 380, 467, 468 productus, 492 Mesoscalpellum, 301 Metridia lucens, 414 Microhedyle milaschewitchii, 129 Microsetella rosea, 414 Microsporidium aplysiae, 270 Minuspio, 118 Mithrax sculptus, 255 Mollusca, 420 Monochrysis, 205 lutheri, 205 Monostroma, 207 angicava, 176 Morus capensis, 367, 478, 479, 485 Mugil richardsoni, 407 Mulinia lateralis, 138 Muricidae, 261 Mycetomorpha, 332 Myctophidae, 424, 426 Mysidacea, 416, 420, 421, 422 Mytilus edulis, 117 Nannochloris, 208 Nassarius obsoletus, 116 Navanax inermis, 251 Nematonereis unicornis, 118 Neoagardheilla baileyi, 200 Neohaustorius biarliculatus, 129 schmitzi, 129 Neolepas, 338 zevinae, 338 Neoscalpellum, 301 Nephrops andamanicus, 407 Nephtys discors, 117 Nereis succinea, 117, 118 virens, 117 Nes longus, 513, 523, 559 Nezumia aequalis, 424 Nitzschia curta, 32, 34 Nothria elegans, 137

Obelia, 270 Octopus, 256, 264 Odontosyllis, 118 phosphorea, 119 Oithona, 414, 415 Olisthodiscus luteus, 568 Ophelia bicornis, 127, 128, 130 Opheliidae, 117 Ophryotrocha, 118 Opisthopus transversus, 270 Orbiniidae, 117 Orsinus area, 425 Ostracoda, 285 Owenia fusiformis, 142 Oxynaspidae, 340 Pachydictyon coriaceum, 263 Pachymetopon blochii, 339, 413 Padina, 174 durvillaei, 176 Pagodroma nivea, 13 Pagurus, 336 Palinurus delagoae, 407, 408 gilchristi, 407, 408 Panopeus herbstii, 255 Panulirus, 430 Paracalanus parvus, 414 Paracallionymus costatus, 423, 424 Parahaustorius longinerus, 129 Parahesione luteola, 118 Parapercis hexophthalma, 540, 542, 543, 546 Parastenocaris vicesima, 128 Parthenopea, 329 subterranea, 329 Patiriella regularis, 251 Pavlova lutheri, 205, 208 Pectenogammarus plancrurus, 129 Pedicellina, 270 Peltogaster, 334, 336 paguri, 334, 335 Peltogasterella, 335, 342, 345 gracilis, 331 socialis, 331, 333, 334 sulcata, 333, 334, 336 Peltogastridae, 328, 329, 332, 333 Pelvetia fastigiata, 212 Penaeus aztecus, 129 duorarum, 129 setiferus, 129

SYSTEMATIC INDEX

Petalonia debilis, 212 Petricola pholadiformis, 117 Petrolisthes armatus, 251 Petrus rupestris, 399 Phaeodactylum, 205 Phaeophyceae, 196, 218, 222, 232, 240 Phalacrocorax bougainvillii, 461 capensis, 461, 465, 480, 485 neglectus, 480, 485 Pherusa affinis, 117 Phronima, 426 sedentaria, 426 Phrosina semilunata, 426 Phyllodoce, 118 Phyllospadix, 206 torreyi, 174, 182, 213 Pisces, 424 Pisiscalpellum, 304 withersi, 305, 318, 319, 339, 342 Platynereis bicanaliculata, 118 Plesionika acanthonotus, 424 Plocamium, 195, 200, 206, 209, 210, 214, 221, 234, 236, 240, 263, 269, 430 cartilagineum, 177, 188, 196, 197, 206, 207, 210, 212, 213, 214, 218, 221, 222, 227, 228, 233, 234, 240, 262, 263 costatum, 172, 177, 188, 207, 213 Polinices alderi, 197 Pollicipes, 287, 314, 338, 340, 341 spinosus, 340 Polychaeta, 115, 416 Polydora antennata, 118 ciliata, 127 Polyphthalmus pictus, 118 Polysiphonia, 173, 200, 232 Polysteganus undulosus, 399 Pomatomus saltatrix, 399, 400 Porphyra, 193, 205, 208, 217, 220, 229, 232 Portunus spinimanus, 255 Postelsia palmaeformis, 72 Praelepas, 338 Prionospio heterobranchiata newportensis, 118 pygmaea, 137 Protodorvillea gracilis, 118 Protodrilus hypoleucus, 128 rubropharyngeus, 128 symbioticus, 128 Protohaustorius deichmannae, 129 Protomyctophum, 420

555

Pseudoeurythoe, 118 Pseudopolydora kempi japonica, 106 paucibranchiata, 137 Psilogobius mainlandi, 513, 533, 535, 536, 545, 560 Pterocladia, 206, 210 capillacea, 207, 213 pyramidale, 212 Pteroculiops guttatus, 512 Pterogymnus laniarius, 385, 386, 390, 399, 413, 469, 471 Pterois volitans, 540 Pterosiphonia, 206 Pterygophora californica, 212 Pterygosquilla armata, 424 Pygophora, 337 Pyraminonas, 205 Rhabdosargus globiceps, 399 Rhizocephala, 285, 286, 307, 327, 328, 329, 331, 332, 333, 334, 336, 337, 342, 344, 345 Rhizosolenia setigera, 414 Rhodophyceae, 196, 218, 222, 232, 240 Rhodymenia palmata, 205, 208 Rhodymeniales, 209 Rhynchospio arenincola, 119 Sabellaria alveolata, 128 spinulosa, 128 Sacculina, 327, 328, 334, 336 carcini, 329, 331, 332, 333, 334, 335, 336 eriphiae, 329 gonoplaxae, 329 inflata, 329 Sacculinidae, 328, 329, 332, 333 Salmo gairdnerii, 563 salar, 563 Sordina pilchardus, 494 Sardinella, 354, 357, 373, 374, 438, 462, 494 aurita, 355, 371, 413, 437 eba, 355, 437 Sardinops, 361, 368, 415 melanosticta, 473 ocellatus, 354, 357, 373, 374, 413, 414, 419, 421, 422, 426, 427, 428, 434, 437, 459, 462, 463, 474, 476, 477, 478, 479, 481, 484, 485 Sargassum, 175, 176 vulgare, 207 Scalpellidae, 320, 337, 338, 339, 340, 341, 342 Scalpellum, 299, 301, 302, 304, 310, 314, 315, 318, 340 alcockianum, 305, 318, 339

556

OCEANOGRAPHY AND MARINE BIOLOGY

angustum, 301 bengalense, 304, 307, 308, 315, 318 calcaratum, 301 californicum, 301 carinatum, 301 chiliense, 304, 317 chitinosum, 305, 318, 319, 342 compactum, 305, 318, 319, 342 compressum, 305, 318, 319, 342 condensum, 305, 318, 320, 339, 342 convexum, 305, 318, 339, 342 cornutum, 301 crinitum, 305, 318, 319, 342 dicheloplax, 301 discoveryi, 305, 315, 317 distinctum, 305, 318, 342 edwardsi, 301 elongatum, 315 fissum, 305, 318, 319 galapaganum, 304, 317 gibberum, 305, 318, 342 gigas, 304, 315, 316, 317, 318, 339 gracile, 305, 317 groenlandicum, 301 gruvelianum, 305, 318 gruvelii, 301, 308, 310 hamatum, 301 hexagonum, 305, 318, 319, 342 hoeki, 304, 315, 316, 317, 318 idioplax, 301 imperfectum, 301 inerme, 301 intermedium, 301, 304, 315, 316, 317 japonicum, 301 javanicum, 305, 318, 319, 342 kurchatovi, 304, 315 laccadivicum, 301 larvale, 301 longirostrum, 302, 304, 315, 318, 342 luteum, 304, 315, 317, 318 marginatum, 301 nipponense, 301 nymphocola, 301 ornatum, 305, 312, 316, 317, 318, 341 osseum, 301 patagonicum, 301 peronii, 301, 302, 304, 312, 314, 315 phantasma, 301 pilsbryi, 302, 304, 314

pressum, 301 projectum, 305, 318, 319, 342 regium, 289, 291, 305, 307, 308, 309, 310, 318, 342 retrieveri, 305, 315, 317 rostratum, 305, 307, 318, 341 rutilum, 305, 315, 317 salartioe, 301 sanctabarbarae, 301 scalpellum, 301, 345 (Scalpellum) elongatum, 304, 315, 316 scorpio, 302, 304, 315 sessile, 305, 318, 319, 342 squamuliferum, 302, 304, 307, 308 stearnsi, 301, 305, 315, 317 striatum, 304, 315, 316, 317, 318, 342 stroemii, 301, 340 stroemii aduncum, 301 stroemii latirostrum, 301 stroemii luridum, 301 stroemii obesum, 301 stroemii septentrionale, 301 stroemii substroemii, 301 tritonis, 305, 315, 317 velutinum, 291, 301, 308 villosum, 302, 304, 314, 315 vulgare, 305, 310, 312, 314, 315, 317, 342 wood-masoni, 305, 307, 315, 317 Scaridae, 540 Schistomeringos longicornis, 118 Scillaelepas, 338, 341, 346 fosteri, 338 Scolecolepides viridis, 118 Scolelepis fuliginosa, 129 Scoloplos fragilis, 117 Scomber, 361, 368 japonicus, 354, 355, 374, 378, 391, 399, 413, 419, 420, 426, 427 scombrus, 420 Scomberesox, 361 sauras, 425, 426, 427, 428 Scyphozoa, 368 Semaeostomeae, 486 Sepia officinalis, 419 Seriola lalandi, 397, 399, 400 Serranidae, 543 Sesarmaxenos, 332 Siphonota geographica, 246 Skeletonema costatum, 432 Smilium, 301, 302, 304, 340

SYSTEMATIC INDEX

557

acutum, 301 aries, 301 longirostrum, 301 peroni, 301, 307, 310 pollicipedoides, 301 scorpio, 301 sexcornutum, 301, 302, 304 uncus, 301 Smilogobius cinctus, 512 inexplicatus, 513 obliquus, 513 Soleidae, 388 Solemya (=Solenomya) velum, 117 Solen viridis, 117 Solidobalanus, 298 Sphacelaria, 206 Sphaerosyllis hystrix, 118 Spheniscidae, 13 Spheniscus demersus, 367, 480, 485, 493 Spio martinensis, 118 Spisula raveneli, 117 Spyridia, 174, 176, 200, 224 filamentosa, 200, 207 Squilla armata, 419, 424 Staphylococcus aureus, 267 Steilocheilus, 261 longicauda, 261 Stephanopyxis turris, 414 Stomatopoda, 419, 424 Stonogobiops, 508 dracula, 513 medon, 513 nematodes, 514, 551 xanthorhinica, 514 Stonogobius, 521 Streblospio benedicti, 119, 128, 137 Strombus gigas, 252 Strongylopleura pruvoti, 270 Stylocheilus, 169 Stylochus, 118 Sufflogobius, 368 bibarbatus, 424, 442, 485 Syllidae, 117, 118 Sylon, 332 Sylonidae, 328, 329, 333 Syringodium filiformis, 205, 206

Tellina tennis, 76 Teuthoidea, 428 Thaisidae, 261 Thalassia, 175 testudinum, 175 Thallasiosira tumida, 24 Therapon jarbua, 541 Thompsonia, 332, 333 Thoracica, 285, 286, 289, 292, 298, 299, 324, 326, 337, 340, 341, 342 Thunnus, 405, 425 alalunga, 403, 404, 405, 426, 452, 453, 454, 455, 472, 487 albacares, 403, 404, 405, 426, 452, 453, 454, 472 maccoyii, 403, 404, 405, 425, 427, 452, 453, 454, 455, 489 obesus, 402, 403, 404, 405, 427, 453, 455, 472 thynnus, 404, 405, 425, 428, 455, 472, 489 Thyrsites, 361 atun, 354, 374, 388, 389, 392, 398, 399, 419, 438, 460, 481, 482 Tiffaniella snyderae, 206 Todarodes, 411, 425 angolensis, 410, 424, 469 Todaropsis, 411 eblanae, 410 Tomiyamichthys oni, 514, 535 randalli, 514 Trachurus, 354, 357, 358, 361, 368, 374, 421, 422 capensis, 355, 358, 362, 375, 376, 391, 413, 419, 439, 440, 466, 469, 482, 484 trachurus, 416, 426, 427 trecae, 355, 358, 373, 374, 375, 413 Trachyrhynchus trachyrhynchus, 424 Triangulus galatheae, 331 Tripterophycis gilchristi, 424 Tripterygion varium, 256 Trypetesa, 291, 296, 298, 339, 343 habei, 296 lampas, 286, 291, 296, 297, 298 lateralis, 286, 296, 297, 339, 342, 343 nassarioides, 296, 297 spinulosa, 295, 296, 343 Trypetesidae, 290, 337, 339 Tunicata, 420 Turbanella hayalina, 129

Tagelus divisus, 117 Tapes japonica, 137

Uca minax, 129 pugilator, 129

558

OCEANOGRAPHY AND MARINE BIOLOGY

pugnax, 129 Ulva, 172, 174, 175, 176, 177, 178, 182, 195, 200, 201, 204, 206, 207, 208, 209, 210, 211, 221, 223, 224, 227, 228, 234, 236, 262, 269 fasciata, 196, 207, 213, 217, 218, 222, 224, 232, 233, 234 gigantea, 207 lactuca, 182, 193, 196, 197, 205, 206, 207, 208, 210, 213, 215, 217, 218, 221, 222, 228, 233, 234, 235, 240, 242 pertusa, 207, 213, 217, 218 reticulata, 207, 213, 224, 233, 234 Ulvales, 209 Umbrina canariensis, 413 Undaria pinnatifida, 211, 213, 217, 218, 228 Urechis caupo, 133 Utinomia, 339 Utinomiidae, 339 Vanderhorstia, 508, 526 ambanoro, 514 delagoae, 511, 514, 532, 533, 535, 558 lanceolata, 514 mertensi, 511, 514, 535 ornatissima, 514, 532, 558 Verrucomorpha, 286 Vireosa hanae, 514, 558, 559 Weltneria, 286, 287, 338, 339, 340, 343 exargilla, 287 hessleri, 287 hirsuta, 288 reticulata, 288 spinosa, 287 zibrowii, 287, 289, 343 Xiphias gladius, 425 Xiphophorus helleri, 264 Yongeichthys pavidus, 514 Zebreleotris fasciatus, 512 Zeus, 391 Zonaria farlowii, 212 Zostera, 172, 175, 176, 205, 206 marina, 206

SUBJECT INDEX

References to complete articles are given in heavy type; references to sections of articles are given in italics; references to pages are given in normal type. Abalone, 410, 411, 473 eggs, 457, 458 Abundance, estimation of in marine ecology, 43–70 in marine ecology, abundance and methods, 48–51 accuracy and precision, 47–51 optimization and pilot studies, 52–64 precision, 51–52 questions and scale, 44–47 Accuracy, definition of, 47 Acoustic surveys and hake mortality, 423 fisheries, 459, 460 Acrothoracica, males of, 286–298 Activity rhythms of gobies and shrimps, 528–532 Aerial-acoustic surveys, fisheries, 461 Aerial surveys for dolphins, 49 Aggressive behaviour and territoriality of gobies and shrimps, 532–534 Agulhas Bank, 375, 376, 381, 384, 385, 396, 411, 413, 417, 423, 425, 432, 433, 434, 435, 440, 443, 447, 448, 449, 450, 454, 459, 464, 471, 475, 489 —Agulhas Retroflection area, 431 squid, 410, 411, 457 Agulhas Current, 411, 431, 453, 454, 455 system, 453 Albacore, 403, 404, 426, 472 Aldabra Atoll, goby-shrimp association, 523, 532 Aleutian low pressure system, 18 Algoa Bay, South Africa, 402, 440, 443, 444 America, Aplysia, 188 American Samoa, goby-shrimp associations, 509 Analysis of variance, 45 Anchovy, 354, 355, 357, 360, 368, 370, 371, 412, 414, 415, 419, 435, 436, 441, 445, 458, 459, 460, 461, 464, 465, 473, 474, 476, 480, 482, 483, 484, 486, 487, 488, 489, 491, 492

egg production, 436 Anglerfish, 385, 390 Angola, 371, 372, 373, 374, 377, 386, 404, 413, 432, 434, 444, 445, 491 Benguela front, 355 Current system, 453 Dome, 433 red crab, 410 sardinellas, 437 upwelling system, 431, 432 Angolan fishery, 371–374 Antarctic, Aplysia, 169 Circumpolar Current, 169 Convergence, 24, 34 krill, 11, 13 Marine Ecosystem Research, at the Ice Edge Zone (AMERIEZ), 25 Peninsula, 24 upwelling, 15 Aplysia, Batesian mimicry, 252 biology and ecology, 167–284 breeding aggregations, 181–183 breeding season, 190, 191 burrowing, 246–247 cannibalism, 205 Cell R15, 178 chemical defences, 260–269 chemical substances in, 260, 261, 262, 263, 264, 265, 266, 267, 268, 269 chemosensory response to food, 208, 209 circadian activity, 245 competition with sea urchins, 180 coprophagy, 223 copulation and copulatory rôles, 183–188 crawling, 243–246 559

560

OCEANOGRAPHY AND MARINE BIOLOGY

crypsis, 252 daily activity, 180 defence mechanisms, 251 digestion and absorption, 220–223 distribution, 169–181 dried algal food, laver “nori”, 205 effect of feeding on other behaviour, 229–231 egg-laying, 188–194 egg-laying hormone (ELH), 192 eggs and larvae, 190, 191 antibiotic factors, 270 xanthophyllic carotenoid, 193 energy, allocations, 238–243 budgets, 240, 242 escape, 252–253 feeding and movement of food through gut, 215–220 ecology, 225–228 field diets, 204–205 food and feeding, 204–231 choice and past dietary history, 214, 215 for larvae, 205–208 foods of, 206, 207 “galloping”, 244 growth, 231–238 and energetics, 231–243 efficiencies, 222 rates, 232, 233 habitat and depth distribution, 174, 175, 176, 177 “head-bobbing”, 188, 229, 230, 248, 249 ink, chemical composition of, 253, 254 effect on blennies, 256 crabs, 255 sea cucumber, 255 inking, 253–256 internal defences, 269–270 laboratory foods, 205 larval feeding rates, 220 life, 197–199 transport in ballast water, 171 learning in, 258, 259 life cycle, 181–204 span, 190, 191 local habitats, 172–181 locomotory rates, 244 locomotion, 243–250 methods of navigation, 249 neurophysiological studies, 183 nutrition, 223–225 opaline secretion, 256–257

parasites, 270 perception of food, 208–215 periodicity of occurrence, 202–204 pheromones, 183, 244 photoreceptors, 245, 246 phycoerthrins from algae, 193 population densities, 226 predators, 250–252 and defence, 250–270 reproductive effort, 195–197 sea hare, origin of common name, 167 senescence and death, 201–202 settlement and metamorphosis, 199–201 siphon- and gill-withdrawal, 257–260 skin colour, 226, 227 swimming, 247–250 tagging of, 203 time spent on reproductive activities, 186 used by neurobiologists, 167, 168 world-wide, distribution, 169–172 Y-maze tests, 208, 209 Aplysine, 254 Aplysioazurin, 254 Aplysiopurpurin, 254 Aplysiorhodine, 254 Aplysiovioline, 254 Arctic circle, Aplysia, 169 marginal ice zone, 15–24; 14 Arniston, 399, 400, 402 Aswan Dam, 172 Alantic Ocean, Aplysia, 169 salmon, 563 tuna, 455 Australia, Aplysia, 169 goby-shrimp associations, 516, 524 lobster, 408 tuna, 454 Bahamas, Aplysia, 171 goby-shrimp associations, 509 Baia dos Tigres, Angola, 355 Baja California, Aplysia, 169 Balanomorpha, males of, 298–299 Baleen whales, 31 Baltic Sea, 72 Barbados, Aplysia, 180, 181, 186, 211, 217, 220, 225, 226, 242 Barents Sea, marginal ice zone, 19–24; 15 Barnacles, New South Wales coast, 44

SUBJECT INDEX

Benguela Current, 353, 486, 490 Benguela ecosystem, abalone, 430–431, 457–458 abalone fishery, 411–412 crustacean fisheries, 407–410 crustaceans and molluscs, 472–473 demersal-trawl fisheries, 378–396 environment, 431–433 epipelagic species, Angola, 462 Namibia, 461–462 South Africa, 458–461 fish and invertebrate resources, perspective, 489–495 fisheries, by-catch species, 382–388 tagging, 355 food, predators, mortality, 414–431 groundfish species, kingklip, 468–469 panga and dentex, 469–470 soles, 470–471 growth of resources, 412–414 hake, 420–423, 445–449 hook and line fisheries, 396–405 horse mackerel, 415–417, 437, 439–441 kingklip, 449–450 and monkfish, 423–425 fishery, 396 lanternfish and lightfish, 418, 444 major fish and invertebrate resources, 353–505 fisheries, 353–412 resources, 412–489 mesopelagic species, horse mackerel, 462–465 other than horse mackerel, 465–468 midwater trawl fisheries, 374–378 molluscan fisheries, 410–412 multi-species fishery, 397–400 panga and dentex, 425, 450–451 pelagic gobies, 417–418, 443–444 phytoplankton, 432, 433, 434 pilchard and anchovy, 414–415, 433–437 predatory fish, snoek and mackerel, 471–472 tunas, 472 purse-seine fisheries, 355–374 red crab, 410, 430, 456–457 resources, 433–458 competitor and trophic flow, 483–487 distributional ecology and reproduction, 431–458 environmental influences, 473–475 exploitation, 487–489 influence of prey variability, 478–482 intraspecific regulation, 475–478 population biology, 458–489

561

variability, 473–489 predation, 482–483 round herring and saury, 441–443 sardinellas, 437, 438 set-net, drift-net and beach-seine fisheries, 405–407 snoek and chub mackerel, 419–420, 444–445 snoek fishery, 396–397 sole fishery, 388–396, 451–452 squid fishery, 410–411, 430, 457 stock assessment, 458–473 tuna fishery, 400–405, 425, 426–428, 452–456 western coast rock lobster, 408–409, 425, 429–430 zooplankton, 432, 433 Benguela El Niño, 436 upwelling centres, 454 system, 407, 410, 431 Benthic organisms, factors affecting or inducing settlement or metamorphosis, 115 studies involving grain size of sediments, 119, 120, 121 surveys, locations, 120, 121 Bering Sea, 23, 32, 33, 34 marginal ice zone, 15–19; 14 Bermuda, Aplysia, 171 Bigeye tuna, 402, 425, 427 Bird Island, Lamberts Bay, 443, 480, 481 Black mussels, 410 Bluefin tuna, 403, 425, 427, 428, 472 Brazil, Aplysia, 171, 173, 188, 203, 211 “inkwells”, 250 spawning of albacore, 454 skipjack tuna, 455 Breeding aggregations, Aplysia, 181–183 Burrowing alpheid shrimps and gobiid fishes, association of, 507–562 Aplysia, 246–247 Burrow structure, construction and dynamics, gobies and shrimps, 517–525 Buzzards Bay, U.S.A., boundary layers, 146, 149 velocity profiles, 149, 150 California, Aplysia, 169, 172, 179, 186, 202, 214, 226, 231, 238 catch replacements, fish, 486 clupeoids, 473 Californian upwelling system, 492 Canary Current, 486 Cannibalism in hake, 423, 494 in rock lobster, 429

562

OCEANOGRAPHY AND MARINE BIOLOGY

of anchovy eggs by adults, 475 Western Cape, 489 Cape Agulhas, 399, 400, 407, 410, 411, 412, 419, 434, 440, 445, 449, 451, 478, 492 Canyon, 444 Columbine, 433, 443, 444, 448 Cross, 436 Frio, 432, 443 fur seal, 415, 417, 418, 420, 423, 490, 494 Hangklip, 365 Infanta, 365, 450 of Good Hope, 453 albacore, 454 Peninsula, 399, 432, 454 rock lobster, 456 skipjack tuna, 455 Point, 365, 367, 403, 404, 407, 408, 410, 412, 429, 433, 435, 436, 443, 445, 480, 492 rock lobster, 456 Province, 406 St. Blaize, 399 Seal, 449 Cape Town, 403, 408, 423 bluefin, 455 frontal jet, 436 Caribbean fishes, 252 Sea, Aplysia, 169, 171, 252 corals, 80 Cetaceans, aerial surveys of, 40 Chemical defences, Aplysia, 260–269 Chub mackerel, 354, 355, 362, 374, 377, 378, 391, 399, 415, 418, 419, 445, 446, 475, 483, 487, 488, 489, 492 Chuckchi Sea, 33 marginal ice zone, 15–19 Cirripedes, apertural males, 285 complemental males, definition of, 285 dwarf males, definition of, 285 males, Acrothoracica, 286–298 a “subrostral pheromone”, 338 Balanomorpha, 298–299 comparative anatomy of, 285–351 descriptions of 286–336 development of sexuality, 344–347 effects of dwarfing, 341–344 hypotheses on origin of, 344–347 Iblidae, 320–327 Lepadomorpha, 299–320 palaenontological considerations, 337 phylogenetic considerations, 337–341

Rhizocephala, 327–336 stages of development, 344 Thoracica, 298–327 parthenogenesis, 346 Coastal Zone Color Scanner (CZCS), 35 Coefficient of dispersion, 71–73 variation, 51, 52 Communication between gobies and shrimps, 540–552 Comparative anatomy of cirripede males, 285–351 Congo, red crab, 410 sardinellas, 437 Contiguous quadrat analysis, 81 Copulation and copulatory rôles, Aplysia, 183–188 Coriolis force, 147 Cormorants, 430, 461, 465, 480, 486 Cunene River, 388, 436, 443, 451 Dageraad, 450 Danish Water Quality Institute, 565 Dentex, 386, 418, 471, 489, 491 spawning, 451 Diet and feeding behaviour, gobies and shrimps, 526–528 Digestion and absorption, Aplysia, 220–223 Discovery expeditions, 24 Distribution, Aplysia, 169–181 Dogsharks, 429 Dragonet, 423 Drake Passage, 24 Dune Point, South Africa, 435, 436 Dwarfing, cirripede males, 341–344 East Greenland Current, 19, 20 Sea, marginal ice zone, 14 East London, 389, 445, 451 Echo sounders and South African fisheries, 365 Ecological impact of salmonid farming in coastal waters, 563–575 Egg-laying, Aplysia, 188–194 Ekman transport, 14 Elands Bay, 407 Elat Nature Reserve, goby-shrimp associations, 518, 554 Elephant Island, 34 Elf, 399, 400 El Niño, 433 Southern Oscillation (ENSO), 18, 363, 371 Energy budgets, Aplysia, 240, 242 False Bay, 407, 419

SUBJECT INDEX

Feeding ecology, Aplysia, 225–228 Fiji, goby-shrimp associations, 509, 516 Fish and invertebrate resources, Benguela ecosystem, 353– 505 Fish-farm effluent, reducing effects of, 572–573 flux of material through, 563–566 wastes and eutrophication of marine environment, 568–569, 571–572 composition of, 564, 565, 573 effect on marine environment, 566–569 organic enrichment of marine environment, 566– 568, 569–571 scale of impact on marine environment, 569–573 Fish-farming, faecal and excretory waste, 565– 566 impact on coastal waters, 563–575 waste food, 565 Fish foods, composition of artificial ones, 563, 565 Florida, Aplysia, 193, 201, 203, 248 Current, 169 Flumes and marine ecosystem, 110 driving the flow, 94–95 entrance conditions, 93–94 Euler number, 98 examples, 105–109 exit conditions, 94 experimental considerations, 104–105 for simulation of benthic environments, 91–112 Froude numbers, 94, 97, 100, 103, 105, 107 Karman’s constant, 101 mass and momentum balance, 95–102 Navier-Stokes equations, 92, 98, 101 necessary scalings, 102–104 Reynolds number, 100, 103 Rouse number, 105 SeaDuct, 109 SeaFlume, 109 Shields parameter, 105, 106, 107 Strouhal number, 97, 98, 105, 106 test sections, 95–105 types of flow channels, 105, 106 Flux of material through a fish farm, 563–566 Food and feeding, Aplysia, 204–231 Fram Strait, 33 marginal ice zone, 15–24; 12 Fulmars, 32 Galjoen, 399 Gannets, 367, 430, 443, 478, 480, 483, 485, 486 Gans Bay, 389, 399, 400, 402

563

Geelbek, 399, 400 Georges Bank, sediment transport, 152 Ghana, sardinella fishery, 494 Gobies, 417, 420, 425 Gobies and shrimps, aggressive behaviour and territoriality, 532–534 burrow construction, 522–523 dynamics, 523–525 structure, 517–522 communication between, 540–552 under natural conditions, 540–542 daily activity rhythms, 528–532 diet and feed behaviour, 526–528 distribution of associations, 509–516 ecology of associations, 516–517 field observations of partner specificity, 552–554 laboratory experiments on partner specificity, 554–556 mechanism regulating partner specificity, 556–558 partner specificity, 552–558 population structure and dynamics, 537– 540 reproduction, 534–537 synchronization of breeding seasons, 537, 538 warning signals in response to predators, 542–544 film analysis of, 546–552 sequence and information analyses, 544– 546 Gobies, area of shrimp antennal contact, 552 associated with shrimps, 512, 513, 514 eggs of, 534, 535 methods of retreat into burrow, 548 movements made to warn shrimps, 545, 547 spawning, 443, 444 techniques of catching, 508 Gobiid fishes and burrowing alpheid shrimps, association of, 507–562 Goby and shrimp associations, effects of catastrophes, 539 evolution of, 558–560 head coloration and warning signals, 544 number of warning signals, 542, 543 partner specificity, 554, 555, 556 warning signals, observation of, 540 Goby-pushing shrimp, 546 resource in Benguela ecosystem, 486 Goodness-of-fit tests, 73–74 Gordon’s Bay, 402 Gough Island lobsters, 408 Great Barrier Reef, 44, 45 goby-shrimp associations, 507, 509, 514, 517, 519, 520, 524, 525, 526, 529, 535, 538, 539, 553, 556 Great Bitter Lake, 171

564

OCEANOGRAPHY AND MARINE BIOLOGY

Groundfish species, hakes, 467–468 Growth and energetics, Aplysia, 231–243 rates, Aplysia, 232, 233 Guano deposits, 463, 473, 478, 479, 480, 481 harvests as indication of biomass of fisheries, 461 Gulf of Aden, Aplysia, 171 Elat, goby-shrimp associations, 507, 524, 528, 529 Guinea, albacore, 454 bluefin, 455 tuna, 455 Mexico, albacore, 454 Aplysia, 169, 171, 248, 250 bluefin, 455 Gulf Stream Current, 169 Gurnard, 390, 483 Habitat selection, hydrodynamical constraints, 148–151 laboratory studies, 130–131 Hagfish, 429 Hake, 354, 374, 378, 379, 380, 382, 383, 385, 387, 388, 390, 392, 396, 417, 418, 419, 424, 425, 470, 484, 487, 488, 489, 492, 494 fisheries, 378–382 nursery grounds, 439 Harder, 406, 407 Hawaii, Aplysia, 172, 194, 201, 226, 231, 250, 260 goby-shrimp associations, 507, 516, 526, 535, 545 Hermanus, 389, 399 Herring, 62, 63 High Energy Benthic Boundary Layer Experiment, 109 Hondeklip Bay, 396, 433 Horse mackerel, 354, 355, 357, 358, 360, 362, 363, 370, 373, 374, 375, 382, 385, 391, 406, 407, 415, 416, 417, 418, 435, 440, 441, 466, 476, 480, 482, 483, 484, 485, 486, 487, 488, 489, 490, 491 nursery grounds, 439 Hottentot, 399 Iblidae, males of, 320–327 Ice, algal community, impact on ice edge phytoplankton, 32–33 edge eddies, 15 effect on light attenuation, 14 effect on water column, 14 Inaccessible Island, lobsters, 408 Index of dispersion, 72, 74, 77, 78 Indian Ocean, 432 albacore, 454 Aplysia, 169, 171

bluefin, 455 goby-shrimp associations, 509 spawning of skipjack tuna, 455 tuna, 455 Indo-Pacific, goby-shrimp associations, 509 tuna, 453 Indo-West Pacific, 407 Inking, Alpysia, 253–256 International Commission for the Southeast Atlantic Fisheries (ICSEAF), 354, 355, 356, 393 scientists, 469 Irish Sea, Aplysia, 188, 203, 221 Isaacs-Kidd trawl, 417 Israel, Aplysia, 179, 180, 181, 186, 202 Ivory coast, red crab, 410 sardinella fishery, 494 Jackass penguins, 367, 430, 443, 480, 486, 492, 493, 494 Jacopever, 391 Japan, Aplysia, 173, 194 catch replacements in fisheries, 486 goby-shrimp associations, 507, 516, 523, 528, 531, 535, 537, 541 longline fisheries, 403 red tide 531 Japanese sardine resource, 473 trawlers, 385 John dory, 391 Johnson and Zimmer’s index, 77 Kalk Bay, 400, 402 Killer whale, 425 Kingklip, 382, 384, 386, 390, 396, 418, 423, 424, 470, 483, 489 catch rates by Spanish vessels, 449 Kob, 392, 399, 400, 407 Korea, longline fisheries, 403 Krill, 31 Lamberts Bay, 365 Lanternfish, 355, 363, 417, 423, 442 spawning, 444 Larvae, chemoreception, 132 Larval life, Aplysia, 197–199 settlement, active habitat selection hypothesis, 127– 139 bottle collectors, 125 chemotaxis hypothesis, 133 choice experiments, 127, 128, 129, 130, 131

SUBJECT INDEX

compatibility of the alternative hypotheses, 154–155 cue detection, 132–134 field experiments, 134–139 hydrodynamical constraints on active habitat selection, 148–151 in the bottom boundary layer, 145–154 laboratory studies of habitat selection, 130–131 methods of observation, 117, 118, 119 of soft-sediment invertebrates, 113–165 passive accumulation, 142–143 passive deposition hypothesis 141–145 passive particle transport and deposition, 151–154 passive resuspension and transport, 144 passive sinking, 144 pattern, distribution and active habitat selection, 139– 141 rôle of hydrodynamical processes, 113– 165 rôle of, on soft substratum, 123–127 site perusal, 132–134 site selection, 132–134 spatial scales of pattern explained by active habitatselection, 113–165 studies, definition of terms used, 115, 116, 119 Leervis, 399 Lepadomorpha, males of, 299–320 Life cycle, Aplysia, 181–204 Lightfish, 423 spawning, 444 Loch Creran, Scotland, organic carbon, 571 Locomotory rates, Aplysia, 244 Locomotion, Aplysia, 243–250 Longhurst-Hardy plankton recorder, 50 Longline tuna, 403, 425, 426 Lüderitz, 355, 365, 367, 433, 443, 444, 447, 454, 479, 480, 485, 491 Divide, 357 zone, 432 Macrofauna beneath fish-farm cages, 567, 568 Madagascar, albacore, 454 goby-shrimp associations, 509 Ridge, 445 Mako shark, 425 Maldive Islands, goby-shrimp associations, 509 Malloy Deep, 19, 21, 22 Mantis shrimp, 419 Marginal ice zones, Arctic, 15–24; 14 Barents Sea, 19–24; 15 Bering Sea, 15–19; 14

565

Chuckchi Sea, 15–19 East Greenland Sea, 14 Fram Strait, 15–24; 12 partitioning of primary productivity, 33–34 physical-biological interactions in, 13–15 phytoplankton dynamics in, 11–38 phytoplankton growth in, 15–32 Ross Sea, 26–32; 15 Southern Ocean, 24–32; 14 upwelling, 14 Weddell Sea, 24–26; 15 Marine Biological Association, Plymouth, Aplysia, 202 Marine ecology, describing non-random pattern, 80–83 detecting non-random pattern, 71–80 estimating abundance, 43–70 fitting contagious distributions, 82–83 power analysis, 64–69 spatial patterns, general conclusions, 83–85 spatial patterns in, 39–90 Marlin, 425 Marshall Islands, goby-shrimp associations, 509 Mauritius, 396, 397 Mediterranean Sea, Aplysia, 169, 171, 172, 179, 202 bluefin, 455 Mesopelagic species, others, 465–466 Metamorphosis, Aplysia, 199–201 definition of, 116, 119 laboratory studies, 131–132 Migratory routes between Red Sea and Mediterranean Sea, 171, 172 Moçambique, goby-shrimp associations, 509, 516 Monkfish, 385, 386, 391, 418, 423, 425, 489 Mossel Bay, 389, 435, 440, 451 Multi-stage sampling, 60–64 Namibia, 365; 355, 367, 368, 369, 370, 371, 372, 375, 376, 380, 381, 386, 388, 393, 397, 398, 407, 408, 413, 414, 415, 418, 420, 423, 425, 432, 433, 434, 441, 443, 448, 449, 454, 461, 462, 463, 464, 465, 473, 475, 476, 477, 478, 480, 483, 484, 485, 486, 487, 492, 493 intrusion of warm water, 491 red crab, 410 rock lobster, 408, 409, 456 Namibian fishery, 365–371 Natal, 386, 400, 407, 433, 445 Nature Conservancy Council of Great Britain, 571 Negative bionomial distribution, 77 New Guinea, goby-shrimp associations, 509 New South Wales coast, barnacles, 44

566

OCEANOGRAPHY AND MARINE BIOLOGY

New York Bight, benthic invertebrates, 63 New Zealand, Aplysia, 169, 188, 201, 225, 226, 228, 250 tuna, 454 Nightingale Island, lobsters, 408 North America, Aplysia, 169, 248 North Atlantic clupeoids, 473 Northern anchovy, 415 Hemisphere, extent of pack ice, 17 Norway, Aplysia, 169 fish farming, 568 salmon production, 563 Nutrition, Aplysia, 223–225 Octopus, 429 Olifantsbos, rock lobster, 456 Opaline secretion, Aplysia, 256–257 Orange River, 355, 388, 407, 423, 435, 443, 445, 448, 483 rock lobster, 456 Out-gassing under fish-farm cages, 571, 573 Pacific Ocean, Aplysia, 169, 171 El Niño, 433 goby-shrimp associations, 509 tuna, 455 Panama, Aplysia, 252 Panga, 385, 386, 390, 399, 471 Parameters, definition of, 43 Parasites, Aplysia, 270 Parthenogenesis, cirripedes, 346 Partner specificity in gobies and shrimps, 552– 558 Pattern Analysis, 81, 82 Patterns in abundance, definition of, 40 of juvenile and adult distribution, 119–123 Pelagic-goby resource, Benguela ecosystem, 480 Penguins, 32, 34 marginal ice zone, 13 Persian Gulf, goby-shrimp associations, 509 Peru, anchoveta fishery, 365, 461 Peru-Chile, catch replacements in fishery, 486 Peru, clupeoids, 473 Peruvian fish production, 492 Petrels, 32 Philippine Islands, goby-shrimp associations, 509 Phytoplankton dynamics in marginal ice zones, 11–38 growth in marginal ice zones, 15–32 ice edge, impact of ice algal community, 32– 33 recent techniques in studying distribution of, 35 Pielou and Mountford’s α, 77

Pilchard, 354, 355, 357, 362, 365, 367, 371, 372, 373, 412, 414, 415, 418, 419, 425, 433, 434, 435, 436, 441, 445, 458, 459, 460, 462, 474, 475, 476, 477, 478, 479, 480, 481, 483, 484, 485, 486, 487, 488, 489, 491, 495 egg production, 436, 437 tagging of, 461 Plankton, Friedinger sampler, 49 Schindler sampler, 49 Victor Hensen, 84 Pointe Noire, sardinellas, 437 Polar Front, 19, 20 Port Alfred, 400, 402 Port Elizabeth, 389, 396, 400, 411, 445, 449, 460, 479 Port Nolloth, lobsters, 492 Power analysis for planning, 66–68 in marine ecology, 64–69 Poisson distribution, 71, 72 Post hoc power analysis, 68–69 Precision, definition of, 47 Predators and defence, Aplysia, 250–270 Primary productivity at ice edge and in open ocean, 18 marginal ice zone, 33–34 Proportional Stratified Sampling, 60 Quoin Point, 399, 443 Rainbow trout, 563 Red crab, 407, 472 eggs, 457 Red fishes, 399, 400, 401, 402, 492 Red Sea, Aplysia, 169, 171, 172 goby-shrimp associations, 507, 509, 514, 515, 516, 517, 523, 529, 530, 531, 532, 535, 540, 541, 546, 547, 554, 558 Refrigerated sea-water boats, 367 Reproduction of gobies and shrimps, 534–537 Reproductive effort, Aplysia, 195–197 Reynolds number, 95, 97, 100, 101, 103, 104, 105, 106, 107, 145, 146 Rhizocephala, males of, 327–336 Ribbed mussel, 425 Ribbonfish, 424 River Nile, 172 Rock lobsters, 472 eggs, 456 Roman, 450 Ross Ice Shelf, 30 Ross Sea, 32, 33, 34 marginal ice zone, 26–32

SUBJECT INDEX

Round herring, 354, 355, 363, 435, 442, 460, 461, 480, 488 St. Croix Island, 443 St. Helena Bay, 365, 407, 410, 411, 432, 433, 437, 447, 460, 480 rock lobster, 456 St. Joseph shark, 406, 407 Saldanha Bay, 365 Salmonid farm, flux of carbon and nitrogen, 566, 569 Salps, 31 Sardine, 354, 486 Sardinella, 354, 355, 357, 373, 438, 445, 487 Saury, 419, 425, 443 Scotia Sea, 24 Scotland, fish farming, 568 salmon production, 563 SCUBA diving, 507, 508 abalone surveys, 473 Sea bream, 385 Sea Fisheries Research Institute, South Africa, 398, 419, 423, 461, 482, 494 Seals, 32, 429 Sediments beneath fish-farm cages, 566, 567 Senegal, sardinella fishery, 494 Settlement and metamorphosis, Aplysia, 199– 201 Sexual dimorphism, 285 Seychelle Island, goby-shrimp associations, 507, 509, 516, 532, 558 Sharks, 399, 425 Shrimp antennae, mechanoreceptors, 548 cleaning of gobies, 527 Shrimps associated with gobies, 514 Silverfish, 400 Silver scabbard-fish, 392 Simple Random Sampling, 60, 63 Simulation of benthic environments, flumes, 91–112 Size and number of sampling units, 53–59 Skate, 430 Skeleton coast, 388 Skipjack tuna, 403, 404, 428 Snoek, 374, 386, 388, 389, 391, 396, 397, 398, 399, 415, 418, 445, 460, 478, 480, 481, 482, 487, 488, 489 Snow petrels, 13 Soft sediments, larval settlement, 113–165 substrata, general features of boundary layer flows, 145–148 larvae studied, 128, 129

567

Sole, 386, 388, 389, 391, 392, 393, 394, 395, 396, 418, 483, 492, 494 catch rates, 452 spawning, 451 Solomon Islands, goby-shrimp associations, 509 South Africa, goby-shrimp associations, 509 African fishery, 360–365 Atlantic albacore, 454 anticyclone, 492 Southeast Atlantic, 353, 355, 358, 373, 375, 377, 378, 380, 384, 386, 387, 388, 393, 403, 404, 405, 438, 443, 446 Southern California Bight, anchovy, 415 Cape, South Africa, 407 oysters, 410 squid, 457 Hemisphere, extent of pack ice, 17 Ocean, 34 marginal ice zone, 24–32; 14 Spain, Aplysia, 204, 211 Spatial pattern, advantages of various indices, 79 Clark and Evans’ R, 75–76 definition of, 40 detection in heterogeneous areas, 80 indices based on distances between organisms, 75–79 indices based on sample units, 74–75 in marine ecology, 39–90 Johnson and Zimmer’s I, 77 measures from sample units, 71–75 Pielou and Mountford’s α, 76–77 terms and concepts, 41–43 the term population, 41 Spectral Analysis, 81 Steenbras, 402 Stochastic modelling for estimation of fish resources, 458, 460 Strait of Gibraltar, 172 Stratified sampling, 59–60 Stratified Sampling with Optimal Allocation, 59, 60 Stratified Simple Random Sampling, 59, 61 Struis Bay, 399, 400, 402 Point, 399 Stumpnose, 402 Submersible mixer used beneath fish-farm cages, 572 Suez Canal, Aplysia, 171, 172 Swimming, Aplysia, 247–250 Swordfish, 425 Tagging surveys, fisheries, 461 Taiwan longline fisheries, 403

568

OCEANOGRAPHY AND MARINE BIOLOGY

Taylor’s power law, 75 Tethys Sea, 172 Texas, Aplysia, 186 Thoracica, males of, 298–327 Total allowable catches (TAC), 382 Toxic phytoplankton blooms and fish farms, 568 Trade-wind cycle, 473 Transkei, 411 Tristan da Cunha, Aplysia, 169 lobsters, 408 Trophic flow, Lüderitz, 485 Tsitsikamma, 402 Coastal National Park, 451 Tuna, 400, 487, 488 migration and current patterns, 453 spawning, 453 Underwater microscope, 49 Upwelling, Antarctic, 15 areas, perennial, 485 centres, South Africa, 356 marginal ice zones, 14, 24 Variance/mean ratio, 71–73; 75 Verhulst-Pearl logistic model, 483 Victoria Land, 30, 34 Virtual Population Analysis (VPA) for assessment of fish resources, 458, 459, 460, 461, 464, 467, 468, 470, 472, 476, 477, 480, 482, 483, 492 Von Karman’s constant, 151 Walker Bay, 365 Walvis Bay, 355, 365, 384, 432, 434, 435, 436, 441, 443, 444, 447, 448, 450, 478, 479, 485 Weddell Sea, 34 marginal ice zone, 24–26; 15 Western Cape, South Africa, 355, 360, 361, 362, 364, 366, 367, 371, 375, 377, 380, 381, 386, 396, 397, 398, 403, 405, 407, 411, 413, 415, 417, 418, 419, 420, 424, 425, 429, 435, 437, 442, 445, 446, 447, 458, 459, 460, 461, 466, 475, 476, 477, 479, 480, 482, 483, 484, 489, 491 abalone, 457 white mussels (clams), 410 squid, 457 Western Port Bay, Australia, 60 West Indies, Aplysia, 169 Whales, 32

marginal ice zone, 13 White squid, 386, 392, 410 steenbras, 406, 407 stumpnose, 399 Wood pulp waste, 571 World War II, 360, 397, 402, 473 Yellowfin tuna, 403, 404, 425, 426 Yellowtail, 397, 399, 400, 406, 407 Zaire River, 388 sardinella, 437