Advances in Ecological Research, Volume 2

Advances in

ECOLOGICAL RESEARCH Edited by

J. B. CRAGG The Nature Conservancy, Merlewood Research Station, Grange-over-Sands, Lancashire, England

VOLUME 2

1964

ACADEMIC PRESS London and New York

ACADEMIC PRESS INC. (LONDON) LTD. BERKELEY SQUARE HOUSE LONDON,W.1

U.S. Edition published by ACADEMIC PRESS INC. 111 FreTH A v E ~ NEW , YORK10003, NEWYORX Copyright @ 1964 by Academic Press Inc. (London) Ltd.

Second Printing 1968 All rights reserved N O PART O W THIS BOOK MAY BE REPRODUCED IN ANY FORM BY PHOTOSTAT, MICROFILM

OR ANY OTHER MEANS, WITHOUT WRITTEN PERMISSION FROM THE PUBLISHERS

Library of Congress Catalog Card Number: 62-21479

Contributors to Volume 2 J. R. BRAY,Botany Division, D.S.I.R. Zealund.

Palmerston North, New

M. B.DALE,Botany Department, University of Southampton, England. E. GORHAM, Botany Department, University of Minnesota, Minneapolis, Minnesota, U.S.A.

J. HESLOP-HARRISON, Department of Botany, University of Birmingham, England. J. M. LAMBERT, Botany Department, University of Southampton, England.

M. E. SOLOMON, Agricultural Research Council, Pest Infestation Laboratory, Slough, England.

Preface The main aim of Advances in Ecological Research, as was pointed out in the preface to Volume 1,is “. . .to present comprehensive accounts of selected topics of ecological research in such a way that biologists with a general interest in ecology as well as specialists in ecology, can obtain a balanced picture of what is taking place”. Mr. M. E. Solomon’s review of processes involved in the natural control of insects will certainly not be the last word in,this very controversial field of study. In taking the insects as his basic material and utilizing information from other groups of animals, he has presented a personal viewpoint of this branch of population dynamics. This should help the general ecologist who cannot hope to keep up with the vast literature and provide workers in population dynamics with many points for discussion and development. I n the first number of Advances, Professor M. E. D. Poore gave an account of his approach to the analysis and description of plant communities. His discussion of classification was of value to animal as well as to plant ecologists. I n this number Dr. Joyce Lambert and Mr. M. B. Dale have looked at the classification of plant communities in a different way and they challenge some of the views expressed by Professor Poore. Their paper, in discussing methods of analysing phytosociological data, gives readers a chance of assessing the value of computers in this branch of ecology. Now that the International Biological Programme is taking shape, the information and discussion in Dr. Gorham’s and Dr. Bray’s paper will provide a valuable starting point for those who will soon be engaged in studying the production of terrestrial communities as part of an international effort. Finally Professor J. Heslop-Harrison’s extensive review of genecology provides the ecologist not familiar with the extensive links between genetics and plant ecology, with a broad perspective of the subject and presents a challenge to the animal ecologist. It was originally planned that Advances in Ecological Research should appear every two years. However, sufficient contributions of high quality are coming forward to justify annual volumes.

J. B.CRAUU

September, 1964

vii

Analysis of Processes Involved in the Natural Control of Insects

. .

M E SOLOMON

Agricultural Research Council. Pest Infestation Laboratory. Slough. England I . Introduction .......................................................... I1. ConceptsandTerms ................................................... I11 Three Types of Processes influencing Abundance........................... A Regulation by Density-Dependent Processes ........................... B . Modification of the Regulatory Processes .............................. C . Imposition of Changes in Abundance Independently of Density ............ IV Estimating the Roles of different Factors and Processes in Natural Control ..... A . Variation and its Causes ............................................. B Detecting and Assessing Regulation ................................... C. Assessing the Role of Particular Factors ............................... D . Interpreting Effects of Successive Mortalities ........................... ,E . The Status of Methods of Estimation .................................. V Density Relationships in the Action of Predators and Paraaites A. Functional and Numerical Responses of Natural Enemies ................ B. A Laboratory Model of Functional Response ........................... C Functional Responses of Some Insect Parasites ......................... D . Mammalian Predators of the Pine Sawfly .............................. E . Functional Responses of Other Vertebrate Predators of Insects ........... F ANoteonTerms ................................................... V I . Questions bearing upon a Theory of Parasite-Host Interaction ................ A Is Parasite Fecundity Not a Limiting Factor? .......................... B . Do Parasite-Host Oscillationstend to Increase in Amplitude? C. How might Expanding Oscillations be Damped? ........................ Acknowledgments ......................................................... References................................................................

.

.

.

.

.

...............

.

. .

..............

1

3 8 8

9 9 10

10

16 26

36

41 42 42 42 44 44 47 49 49 49 61 61 64 64

I. INTRODUCTION A short title may cover a wide field. The title of this article is by no means long enough to show precisely what I propose to deal with and which topics will be omitted . The word insects does duty for terrestrial . insects and mites. Birds and small mammals appear only in the role of predators upon insects. The emphasis will be on the results of practical studies of population dynamics. especially in the field. on the sorts of data that are needed for the study of natural control. and on their elementary analysis. I shall not deal with statistical methods. nor

2

M. E. SOLOMON

developments in mathematical modelling, nor the methods of sampling and counting animals. I shall touch only incidentally on philosophical questions such as whether the numbers of animals are regulated or not, and consider instead how to assess the degree of regulation. However, this cannot be done effectively unless there is understanding between writer and reader as to what is meant by this term and certain others. To make these matters clear, I have included explanations of terms and a simple framework of ideas about natural control, which I hope will show the bearing of the topics discussed upon the central problem of how the numbers of animals are regulated. I n the last few years a good deal of new information has become available through studies of insect and mite populations in the field and in small-scale experiments. At the same time, new or newly adapted methods of analysing population dynamics have been introduced and put to work on the results of these investigations. The most notable body of new data and methods comes from studies of forest insects in England (by Varley and Gradwell), the Netherlands (by Klomp and his colleagues) and particularly in Canada. I shall make frequent reference to the recently published account, by Dr R. F. Morris and his colleagues of the Green River Project, on the spruce budworm in the fir and spruce forests of New Brunswick. This project is remarkable for the concentration of manpower over an extended period, for the broad approach to what is a major economic problem, and for the amount of attention devoted to problems of measurement, analysis and mathematical formulation. Other data I have found very instructive include those'of Richards and Waloff (1961) on the broom beetle and those of Holling on predation. The new data are particularly welcome to students of insect population dynamics, since their thinking has generally suffered from an insufficient basis of ascertained fact. The information that has been available has been mainly derived from laboratory experiments, from biological and chemical control undertakings in partly or completely unnatural circumstances, from the simpler examples of regulation in the field, or from investigations that did not go far enough, or not in the right directions, to uncover the regulatory processes. There has been a special shortage of facts about the more difficult, but widely typical, populations that are members of complex communities and subject to mani.fold influences. The work on forest insects is now providing more data of this sort. It would be valuable, from this point of view, if more of the original observations on the spruce budworm were published. The report (Morris, ed. 1963) presents the relationships found, in impressive completeness, but includes very little of the observational data. At this stage the problems of how to set about analysing the dynamics

ANALYSIS O F PROCESSES I N CONTROL O F INSECTS

3

of populations, and the elementary biological thinking that should guide us in these matters, are of prime importance. The questions involved include the following, which I propose to discuss in later pages. How can the roles played by different factors and sorts of factors in natural control be estimated? How can the presence of regulation be detected, and how measured? How is the effect of a mortality factor changed when it is preceded or followed by other mortalities of various types? I n what ways is the action of predators and insect parasites related to the density of the prey? Can different aspects of this action be considered separately? Methods of attacking these questions will be illustrated as far as possible by use of published data from field investigations, but in some cases by means of hypothetical examples. My aim will be to deal with the methods and examples in their simplest forms. Simple procedures based on elementary ideas are not only easily assimilated; their implications are relatively clear, and they are amenable to development in various directions to meet the needs of particular investigations. I shall not deal with some of the more sophisticated methods and models which forego some or all of these advantages in the interests of specialization for a particular set of circumstances. This does not imply any depreciation of the making of mathematical models, an important aspect of population dynamics which has recently undergone vigorous development, as may be seen from the papers of Watt (1961, 1962), Holling (1962) and others. I agree with Watt’s view that in the study of insect populations, as already in fisheries research, this sort of theory is likely soon to become a major means of advance, the more so to the extent that the models are kept in touch with field data, and field investigations are organized in such a way as to use and test the models. But the present article will deal rather with the elements from which complex models may be constructed. Theories of natural control involve assumptions that must be tested by observation or experiment if the theories are to be seriously employed. This aspect of the relationship between theory and practice emerges explicitly in Section VI, in connection with the influence of predators and parasites upon insect populations. Thus the article concentrates on a few aspects of a wide subject, and refers to other aspects only briefly or not at all. Within the chosen topics, selection has often been necessary, and sometimes inescapably arbitrary .

11. CONCEPTSAND TERMS We must begin with the truism that the numbers of animals in natural populations are limited - strikingly so in view of the high rates

M. E. SOLOMON

4

of repPoduction &hat many species can achieve under favourable conditions. Whatever processes are responsible for this restriction are referred to collectively as natural control. Franz (1962) has suggested the alternative term limitation. If a more formal statement is necessary, natural control can be d e h e d as the process(es) keeping the numbers of animals, in a population not controlled by man, within the limits of fluctuation observed over a sufficiently representative period (cf. Solomon, 1957, p. 132, also Stern et al., 1959,p. 87).

80

b

L

40

1

120

I

I

I

160 200 2 4 0 Large larvae per IOsq. ft. of folioge 80

I

280

FIG.1. A density-dependentrelationship in the spruce budworm, after Miller (1963a). Each point is the mean of ten values. The effect can be explained in terms of food supply or starvation.

Among the processes involved in natural control some can be distinguished as density-dependent ; their action (measured proportionately, as percent mortality or as mean effect per individual of the population) becomes increasingly adverse when density rises, and , 15, 16). decreasingly so when density falls (Fig. 1, and cf. Figs. 9 ~ 14, This relationship between adverse action and density may show itself promptly, as in some forms of competition, or in a lagging reaction, as in the case of an increase of parasites or predators following an increase in the hosts or prey (Fig. 3). Because the proportionate adverse action of density-dependent processes declines when density falls, as well as intensifying when density rises, such processes tend to curtail fluctuations, whether upwards or downwards, that go beyond the average or normal levels of abundance. Nicholson (1933) regarded them as acting in a compensatory way against any departure from an equilibrium level which continually changes. The restriction of increases in density in this way is an active process ; the curtailment of downward fluctuations is not -

ANALYSIS O F PROCESSES I N CONTROL O F INSECTS

5

it is simply an effect of the relaxation of the active psocess. The principle of this action is the same as that of the governor on an engine; it is the cybernetic principle of negative feed-back. Thus it can be said that

I

I

I

I

0-

100

200

I

I

300

500

400

Larvae per 1OOsq. ft. of foliage

FIG.2. Graph of data tabulated by Morris (1959) for larval population density and yo parasitism of the black-headed budworm, AcZeris variana (Fern.) (Tortricidcte) by a complex of Ichneumoid and Tachinid parasites, in successive generations in a stand of conifers in northern New Brunswick.

0

I

I

I

I

I

!

100 200 300 400 500 Lorvol populotion of previous generation, per IOOsq. ft. of foliage

FIG.3. The same data as Fig. 2, but current yo parasitism graphed against host population density in the previous generation, showing (delayed) density-dependent relationship.

6

M . E. SOLOMON

density-dependent processes tend to regulate abundance or population density, and I restrict the use of the term regulation to this process (cf. Nicholson, 1954b, density regulating factors). If an increase of population is stopped by some process other than a density-dependent one, this may for the time being constitute an aspect of natural control as defined above, but I do not call it regulation. I n an earlier paper (Solomon, 1949) I equated natural control and regulation, but the general tendency has been away from this strict interpretation of natural control, and I have since used the more inclusive definition. I n practice, regulation cannot be studied adequately without reference to the wider aspects of population dynamics, for anything that happens to a population, and anything that it does, may have an effect on its regulation. Regulation can be imposed by all types of density-dependent processes : by the action of predators, parasites, or pathogens, by intra-specific competition for various requisites including food, shelters and nesting-sites, and by mutual interference or agression which can also be interpreted as an aspect of competition for space or resources. Competition may lead to losses by emigration. In animals that have a social organization, regulation through competition may be mediated by restrictions imposed by the population upon its members. Density-dependent processes are distinguished from inverse processes, which operate in the opposite sense, i.e. their adverse action becomes proportionately weaker as density rises, or intensifies as density falls. For example, in a sparse population reproduction may be hindered by the infrequency of encounters between the sexes; or, as density increases the proportion parasitized may decline (Fig. 4, and 'cf. Figs. 9~ and 17). Many natural enemies behave as inverse factors under certain environmental conditions, or when the ratio of enemies to prey is low. This is a consequence of their limited capacity for attack. The significance of this feature was first emphasized by Thompson (1939, and earlier). Examples will be cited in Section V. The action of density-independent processes is not significantly dependent upon population density. A little more should be said about the differences between prompt and lagging density-dependence. Intra-specific competition generally seems to be promptly density-dependent, and so, at times, does the influence of predators. But in a common type of parasite-host interaction, part of the response of the parasites to an increase in host density is to increase in abundance, which cannot be done promptly. The parasites commonly fail t o increase for a time even after the host increase from a low density has been resumed. The result of this delay, as Varley (1953) has pointed out, is that in parts of the parasite-host

A N A L Y S I S O F P R O C E S S E S I N CONTROL O F I N S E C T S

7

Pupal density (no.per IOsq. ft. of foliage)

FIG. 4. A predominantly inverse density relationship in the influence of parasites upon the spruce budworm. (From Miller, 196310.)

oscillation the parasite acts like an inverse factor, at other times like a promptly density-dependent one. While the differences between lagging and prompt density-dependence are often important, they both tend in practice towards regulation, oscillatory in the one case, plain in the other ; and both are likely to be rather irregular in most natural environments. The more rapidly the parasite can develop and reproduce compared with the host, the more closely is its action likely to approximate to that of a promptly density-dependent factor. Also some predators and insect parasites can react at once to a rise in prey density by attacking at a proportionately higher rate, in which case this part of their response is promptly density-dependent (Solomon, 1949). Some lagging density-dependent processes do not arise from natural enemies, but from damage to the environment, or, as in one recorded instance (Wallace, 1962), from a toxic effect of feeding on dead bodies. Taken as a group, natural enemies cannot be classified as all promptly density-dependent, all lagging density-dependent, nor as inverse factors, nor as density-independent. Their action is varied, and affords examples of all of these relationships except perhaps the last. A classification of types of action or types of density relationship is one thing, a descriptive classification of factors (weather factors, parasites, etc.) is another, and the two cannot be fitted neatly together. Certain correspondences occur, not in accordance with firm rules but rather as tendencies, subject to various qualifications and exceptions. I n discussing the influences acting upon a population we may refer to any element in the situation, e.g. predators, competitors, predation or competition, as a factor. At the same time, predation and competition are processes, and may be studied as such. If we are concerned with the way in which the effect of a process or factor varies with the population density, we are studying a relationship, of a type that may

8

M. E. SOLOMON

be distinguished as a density relationship. Thus, if we find that a population suffers a density-dependent mortality, and identify a predatory population as the cause of this, the predators or their predation constitute a density-dependent factor, the predation is a density-dependent process, and the process can be regarded as the expression of a density-dependent relationship. The use of these different terms need imply no more than choosing the words appropriate to the context. The term population, as I use it in this paper, simply means any group of animals, usually of one species, that can conveniently be considered as a unit. For the present purposes we may assume that all individuals of the same species and of a particular stage of development are equivalent, although in practice it is desirable, when possible, to take account of differences in age and genetics. In practice, males and females may sometimes be differently involved in density relationships. When different developmental stages occur together, one should take account of the numbers of each stage separately.

111. THREE TYPES OF PROCESSES INFLUENCING ABUNDANCE On an elementary and fundamental level one can make a three-fold classification of the processes involved in natural control. The three categories are (a) regulation by density-dependent processes, ( 6 ) modification of the regulatory processes, and (c) imposition of changes in abundance independently of population density. The use of these simple distinctions is convenient in thinking about population dynamics on the theoretical level and in analysing field observations. The following paragraphs are intended to establish this classification for the purposes of the succeeding discussions.

A. REGULATION BY DENSITY-DEPENDENT PROCESSES The simpler aspects of regulatory action are the immediate effects of increases or decreases in density, e.g. in intensifying or alleviating competition. When there are significant lagging effects the picture is more complicated, for lagging or persistent effects tend to give rise to oscillations. For example, it sometimes happens that predators, parasites and phytophagous animals do not immediately relax their attack on the supply of food if this becomes over-taxed, but rather exploit it more intensively; then their numbers are belatedly reduced by shortage of food; the consequent relaxation of attack allows the food to increase again; the consumers, owing to such delays as the time required for gfrowth and reproduction, at first increase rather slowly, and only later come again to dominate the food supply.

ANALYSIS O F PROCESSES I N CONTROL O F INSECTS

9

Regulation of a population may be influenced by more than one density-dependent process simultaneously. It may be exerted by different processes at different times and places. It may be continuous or intermittent, or, presumably, even absent in the case of marginal populations which are ephemeral offshoots of more permanent ones.

B. MODIFICATION

OF THE REGULATORY PROCESSES

Density-dependent regulation is of course subject to many modifying influences bearing upon the population and its environment. Influents such as weather may greatly modify a population’s reproductive and survival rates, its food supply, and its competitors and natural enemies. Such influences often play a prominent part in determining which regulatory process(es) shall, for the time being, be the decisive one(s), and at what level of density further increase would be prevented. Nicholson (1954b)refers to modifying influences as “legislative”. The legislative or modifying factors include many and perhaps most genetic changes. Only when such changes are geared to the level of density in some way can they play a direct part in regulation. While examples are available of genetic change which seems to act as a modifying influence, leading to increases in abundance, the density-related action of genetic change is still almost entirely a matter of theory and speculation (cf. Franz, 1949 ;Pimentel, 196lb).

c.

IMPOSITION OF CHANGES I N ABUNDANCE INDEPENDENTLY OF DENSITY

It is generally agreed that influents such as weather, not responsive to population density, commonly have a density-independent action on a population. (Of course, severe weather may give rise to a situation in which competition occurs for a small residue of favourable sites; then the situation may be expressed as “density-independent influent + limited space+density-dependent competition +regulation” ; this is the same class of situation as that described under B - weather determining which regulatory process is to be prominent and also modifying its intensity.) It has occasionally been argued that the effects of influents that are apparently density-independent may in fact often depend in some small degree on population density (Andrewartha and Birch, 1954; Chitty, 1960). However that may be, if their action cannot be shown to be substantially dependent on density, it is realistic to treat them as density-independent. There is little point in discussing in a general way the relative contributions of regulatory factors and density-independent factors to natural control. A more constructive approach to this question is to find means of identifying and measuring the action of both types of

10

M . E . SOLOMON

factors in actual situations, and of assessing their impacts on populations. This matter is taken up in the following Section. IV.

ESTIMATINQ THE ROLESOF DIFFERENT FACTORS AND PROCESSES IN NATURAL CONTROL

I n Section 111, natural control was represented as a three-fold process

- regulation, modification of this, and variation not related to density.

Density-independent variations in abundance may be imposed by many Werent types of factors, but, particularly with invertebrates that are exposed to its influence, weather is the predominant cause of such variations. Weather influences the state of affairs in the environment in many ways, and to a great extent “sets the stage” for the processes of population dynamics, including regulation ; it partly determines the capacity of the environment with respect to the animals that live in it (Solomon, 1949). It often influences abundance directly by periodically killing a fraction of the population, and often determines the favourable conditions that enable a population to increase rapidly in abundance. We may examine the correlations between weather conditions and population increase, decrease and abundance; this has been a major preoccupation of field entomologists, and is obviously necessary and important. Yet, from the point of view of an ecologist wishing to understand natural control, it is only a preliminary stage. He must go on to discover how abundance is regulated, If he succeeds in this, he will be able to explain how it is that, in spite of great powers of increase most animals can display under favourable conditions, and in spite of the fact that the action of the weather often bears no significant relationship to the level of abundance at the time, yet variation remains within definite limits, and populations are maintained in a loose state of equilibrium: most of them neither die out nor maintain a net upward trend of abundance. Since I have set out this argument at sufficient length elsewhere (Solomon, 1957, 1962a),and others have done likewise, it seems unnecessary to dwell upon it here. A more elaborate study of the matter, using namerical examples, has been published by Klomp (1962). A.

VARIATION A N D I T S C A U S E S

1. Amplitude of Fluctuation Violent fluctuations in abundance are a characteristic of many insect populations. The red locust, Nomadacris septemfasciata (Serville), in a self-contained, controlled outbreak in Africa has been shown t o vary in annual abundance by a factor of over 750 times (Gunn and Symmons, 1959); the increase over two breeding seasons, 1929 and 1930, in the outbreak areas has been estimated at x 140; this was followed by an

ANALYSIS O F PROCESSES IN CONTROL O F INSECTS

11

increase of x 45 over the succeeding 4 years, averaging x 2.9 per annum (Gunn, 1960). Gunn and Symmons (Zoc. cit.) commented: “if . . . a population exists in generally favourable conditions, such as outbreak areas are now supposed to provide, and unfavourable factors are few and powerful, then large fluctuations may be expected.” Glasgow and Welch (1962) presented estimates of the annual abundance of ‘an insect of unusually low fecundity, the tsetse fly Glossina swynnertoni Aust., in an area of 40-50 sq. miles of thorn bush in the Shinyanga District of Tanganyika. In the first 5 years the range of the annual estimates was approximately 4.5-fold, and in the following 5-year periods it was about 2.1-, 4.2- and 3.2-fold. Taking 10- instead of 5-year periods, the fluctuation was about 18-fold in the first period and 5.4-fold in the second. Taking all together as a single 20-year period, the range is just over 18-fold. I have divided the run of years as above in order t o illustrate the point that the amplitude of fluctuation depends partly on the number of years or generations covered, and to provide a basis for comparing the values for Glossina with those quoted below, for some other insects. Richards and Waloff (1961) estimated the numbers of the beetle Phytodecta olivacea on a more or less isolated patch of about two acres of the leguminous shrub, broom (Sarothumnus scoparius), in southern England. Over a period of 5 years, the estimated numbers of adults in spring (some of them a year older than the others) ranged from 17 027 in the second year to 4 000 in the fifth, i.e. a 4.3-fold range of fluctuation. For a number of species of oak-feeding caterpillars in Wytham Wood, near Oxford, Varley and Gradwell (1963) have illustrated annual estimates of population density for the years 1949-62. Over the first 10 of these years, their record is complete for 9 species. I have measured the approximate ranges of fluctuation from their graph, and converted the values from their logarithmic to a geometric scale. I n the first 5 years, the values range from nearly x 3 (for Eucosma isertana) to x 25 (for the winter moth Operophtera brumata). I n the second 5 years, they range from about x 2 (for Erannis defoliaria) to x 10 for Cosrnia trapezina). Taking the two together as one 10-year period, the values range from about x 3-4 (for Eucosma) to x 25 (for Operophtera). Klomp (1962) shows graphically the annual density of caterpillars of six species of Lepidoptera living in the foliage in a forest of Scots pine in the Netherlands (Fig. 5). I n the first 5 years of his 10-yearperiod, the approximate range of fluctuation varied from x2.9 (Panolis) to x 28 (Thera), and in the second 5 years from x 5.3 (Thera) to x 33 (Ellopia). Taking the 10 years as one period, the range of fluctuation varied from x 16 (PanoZis) to x 95 (Eupithecia).The similarity to the

12

M. E. SOLOMON

-

\-------

10 -

/ Bupolus oe%

./.

O

I-

O

00

Ponolis Ellopio ~

Eupithecia Thera

j 0.1 -

5

z

Semiothiso

1

1

1950

1

,

1952

,

,

1954

,

,

1956

,

,

1958

I

1960

FIG.5. Population density of phytophagous caterpillars in a forest of Scots pine. (From Klomp, 1962.) The horizontal line has been added to indicate Klomp’s estimated mean density for Bupalus.

figures of Varley and Gradwell (above) is rather close, in spite of the great differences between pine and oak. Another set of data for Lepidoptera feeding on pine foliage was published by Schwerdtfeger (1935, 1941).These were based on censuses of overwintering pupae or larvae on the ground in forests at Letzlinger Heide, Germany. They were conveniently re-graphed on a logarithmic scale by Varley (1949);I have read off approximate values of the ranges of fluctuation on this logarithmic scale and converted them to a geometric scale. Two of the species were among those dealt with by Klomp (loc. cit.), namely the Noctuid Panolis griseovariegata (Goeze) and the Geometrid Bupalus piniarius (L.) -the pine beauty and the bordered white. A third species, the pine hawk, Hyloicus (or Sphinx) pinastri (L.) was also dealt with by both authors, but Klomp’s data were too incomplete for the present purpose. A fourth species studied by Schwerdtfeger was the Lasiocampid Dendrolimus pini L. The graphs show marked fluctuations of all four species, rather irregular but with a tendency to oscillations of 7 or 8 years duration (9 or 10 years in the case of Hyloicus). Some of the peaks in these fluctuations represent severe outbreaks. Normally, one should take account of these facts in selecting the length of the periods in which to measure the ranges of fluctuation, but I have simply started at the beginning of the records and used successive 10-year periods for comparison with the other examples. The figures cover the period 1881-1940, except the last 10 years for Hyloicus.

13 Taking 10-year periods, the ranges of fluctuation for Panolis are approximately x 86, x 81, x 7.4, x 209, x 14.5 and x 77. The corresponding value from Klomp’s data (above) was x 16. The diEerence between this relatively low value and the three higher ones from Schwerdtfeger’s data is an expression of the fact remarked upon by Klomp (loc. cit., p. 97) that in the pine forests of the Netherlands real outbreaks of pine caterpillars do not occur (except, rarely and locally, for Panolis). Klomp interprets this as evidence of regulation in the Netherlands forests. Returning to Schwerdtfeger’s data, the ranges of fluctuation of Panolis over three 20-year periods are approximately x 188, x 205 and x 124; for the whole 60 years it is x 438. The corresponding values for Bapalus are: 10-year periods, x 753, x 393, x 3 236, > x 3 236, x 14 190, x 102; 20-year periods, x 1021, x 3 236, x 14 190; 60 years, x 27 160. Klomp’s data give the value x 30 over 10 years. The values for Hyloicus are: 10-year periods, x 53, x 6.6, x 6.0, > x 86, x 72; 20-year periods, x 53, > x 86; 50 years, x 330. I n the Netherlands, this species remains at a low level (Klomp, loc. cit.). The values for Dendrolimus are: 10-year periods, x 841, x 7-9, x 16-9, x 19.6, x 124, x 1 611; 20-year periods, x 2 489, x 19.6, > x 3 381; 60 years, > x 17 580. Equally striking fluctuations in abundance were demonstrated by Nicholson (1954b) in laboratory experiments with the blowfly Lucilia cuprinu Wed. I n this case, the external conditions were constant, and the fluctuations were due to delayed responses to a limiting food supply. For strictly valid comparisons of the ranges of fluctuation one should take account of which stages in the life cycle are the more subject to disturbing influences, and which are the more closely regulated. When this is not done, we may inadvertently compare one population in its most variable phase with another at its least variable. To avoid this sort of error, we need censuses of the different stages, or life tables for the successive generations of the populations concerned, to be examined as on pages 31-35. We should certainly not be justified in assuming, without some such observations, that the adults will fluctuate less, in proportion to their numbers, than the larvae do, or vice versa. The range of fluctuation is no doubt influenced by many different factors, including the variability of the climate, the reproductive capacity of the species concerned, the degree of regulation, and the number of different regulatory factors. I n connection with the last of these, it has been argued that the ffuctuations are likely to be less violent in a population that is part of a complex community, which provides a variety of inter-compensatory influences buffering the effect of any ANALYSIS OF PROCESSES I N CONTROL O F INSECTS

14

M. E. SOLOMON

major disturbing factor (Voilte, 1946; Solomon, 1949; MacArthur, 1955; Pimentel, 1961a; Elton, 1958). Insect populations tend on the whole to fluctuate much more violently than those of birds and mammals, which have a relatively small reproductive potential, exercise parental care of the young, and are less influenced by variations in weather. These vertebrates are more socially organized than most invertebrates, and have developed patterns of behaviour which tend to regulate their abundance in accordance with the resources of the environment, a subject recently discussed a t length by Wynne-Edwards (1962). Fish are to a great extent sheltered from weather effects, and are long-livedenough for the impact of a heavy mortality or high survival of one year’s fry to be buffered by the older and the succeeding generations.

2. Types of Fluctuation Although, superficially, fluctuation is perhaps the most definite and easily measured of population phenomena, there are so many different causes of variation in abundance that comparisons between different populations, or over different intervals of time, must be made with caution. Abundance in a particular habitat may be influenced by an annual breeding season, with annually varying reproductive success, followed by a period of variable decline in numbers due to factors such as weather conditions, food shortage, natural enemies and disease, until the next breeding season. Clearly the time of year and the age distribution should be taken into account when these fluctuations are assessed. In comparisons of the abundance from year to year these complications are generally avoided by counting at a particular stage of the life cycle. Other causes of fluctuations are migratory, dispersive or aggregative movements, as in locusts, or, less strikingly, in the spruce budworm (Morris, ed., 1963). Another complication is that fluctuations between generations are often caused by lagging density effects, notably in the parasite-host oscillations: thus, while a parasite may be regulating its host, it may be regulating it through an oscillatory course of marked systematic increases and decreases in abundance, instead of steering it towards a constant (or randomly varying) level of abundance. Other types of lagging density-dependent factors can have a similar effect. The striking fluctuations in Nicholson’s laboratory populations of blowflies, as already mentioned, were caused by the delayed effects of intense competition for limited resources. A field exam9le of this type is provided by the work of Wallace (1957) on the lucerne flea, Xminthurus viridis (L.), in W. Australia. .This collembolan is a pest of pastures containing clovers. It is very

ANALYSIS O F PROCESSES I N CONTROL O F INSECTS

15

dependent on favourable conditions of rainfall and temperature, and the population fluctuations caused by changes in these factors have been demonstrated by earlier workers. But when Wallace made a close study of small areas of field, he found local variations in abundance bearing no relation to the prevailing moisture and temperature. There was a changing mosaic of densities that could not be explained by migratory movement. When he plotted densities early in the season (May) with densities on the same patches late in the season (September),the points were fitted quite well by a hyperbolic curve (Fig. 6). The coefficient of

10ol 0

0

1

50

,

I

100 150 Population density in Mow

1

200

FIG.6. Relation between densities of lucerne flea populations at given p0htS early and late in the 1953 season. (FromWallace, 1957.)

correlation between the estimates of initial and final densities in the year concerned (1953) was - 0-426 (significant at P < O - O O l ) . Wallace remarked that “where densities exceeded approximately 20 per aquare link in May they had decreased by September, and where they were below this figure in May they had increased, there being only a few minor departures from this relation. Also, in general, the greater the departure from 20 in May, the greater was the departure, of opposite sign, in September.’’I n observations over 4.successive years a significant negative correlation was maintained between May and September values, and the density mosaic was different each year, so that it could not be related to soil differences. Later observations provided an

16 M. E. SOLOMON explanation of the phenomenon in terms of the lucerne fleas’ habit of feeding upon the dead bodies of their own species, which exert a toxic effect (Wallace, 1962). The marked fluctuations observed by Wallace were out of phase with each other; when the density on some patches was high, on some of the nearby patches it was low or intermediate in level. This might easily have been regarded as local variation of little general significance and sampling replicated until the effects were ((ironed out)’. Wallace’s findings show that the fluctuations of very localized sub-populations can be most instructive, and that vital information may be lost if they are regarded simply as a nuisance. B. DETECTING A N D ASSESSING REGULATION One could consider, on the one hand, means of detecting only the presence of regulation in the dynamics of populations, and, on the other hand, means of assessing its extent or effectiveness. But since the first operation nearly always involves something of the second, it is convenient to deal with both together. 1. B y the Tendency of Density to Return towards a Mean This is the most direct way of detecting and measuring regulation, the nearest approach to observing the process itself. It is often made difficult by the fact that the actual “target” level of density, towards which regulation tends to direct a population, itself changes continually under the direct and indirect influence of various environmental conditions. Nevertheless, as we shall see, it is sometimes practicable. Obviously, if we are to observe the return of density towards a recognizable (‘target” level or a mean density, there must first be a displacement of density from this level. Such a displacement may be either natural or artificial. a. After naturally imposed changes. A simple example of this method is given by Klomp (1962), using the annual estimates of the density of pine looper larvae (Bupalus piniariw) shown in Fig. 5. Klomp gives a value of 20 as his estimate of a mean density for the period, and I have added SL horizontal line to the diagram at this level. The argument is that in a highly regulated population the trend from one point to the next should be upwards ( + ) if the former point is below the mean, and downwards ( - ) if it is above the mean. From the graph it can be seen that the expected and actual trend from each point is as follows: Point

Expected trend Observed trend

1

2

-

-

3

4

5

6

7

8

9 1 0

+ + - - - + + + - + + +

+ - + + + -

ANALYSIS O F PROCESSES I N CONTROL O F INSECTS

17

Only in two cases in ten do observation and expectation disagree. Klomp writes: “Under the hypothesis that no tendency to return to the constant level is operating, we can expect that in half of the cases there is no agreement. The probability of the above result under this hypothesis is 0-03 (tested one-sided). Consequently, this hypothesis can be rejected.” This gratifyingly simple method can succeed only with highly regulated populations, where the regulatory action is dominant over the non-regulatory fluctuations. Where the opposite is the case, the method does not detect the regulatory element. This can be demon. represtrated by means of the population graph J in Fig. 1 0 ~ This sents an imaginary population in which the main fluctuations are density-independent, but in which the rate of change upwards or downwards is density-dependent. The scale of density used is logarithmic, and the mean is the geometric mean, 2.26. The graph provides thirteen steps in which the expected and the observed trends can be compared. There are only six agreements, and seven disagreements. b. After arti$cially imposed changes. Nicholson (1957) and Hairston (1957) have remarked that it should be possible to demonstrate the existence of regulation artificially by imposing a change in density upon a field population and observing its returns towards the normal level. Such experiments would have to be done with care to avoid interference with natural enemies, where these are important ; for this reason, many pest control operations would not provide a suitable test;

I

I

FIG.7. Simplified model in which a population tending to assume a constant density is reduced to lower densities by exploitation (vertical arrows), and its subsequent increase observed. (FromSolomon, 1962b.)

18

M. E. SOLOMON

however, many of them might serve well for the purpose, if the aftereffects were adequately observed and reported from this point of view. Another field in which evidence of this sort could be collected is in the exploitation of natural populations. Men concerned with the preservation of fisheries and of game stocks have a good deal of knowledge about the speed of recovery of populations in the face of reduction by hunting. I n a discussion about the light this sort of information might be expected to throw upon regulation (Solomon, 1962b), I suggested that the systematic exploitation of natural populations might be expected to give measures of the degree of regulation in the following ways, illustrated in Fig. 7. (i) Assuming a population can recover from low density as illustrated in Fig. 7a, after a measured reduction in density the speed of its recovery to normal density can provide an index of the degree of regulation (Fig. 7b). (ii) If a population is reduced by a set amount at intervals, and the intervals are shortened until the population can barely reach its normal density (Fig. 7c), this can provide an index of the degree of regulation. Fig. 7d relates to the following sub-section.

2. B y the Way in which Mortality, Reproduction, or Net Increase are Related to Density Aspects of regulation, sometimes even the entire basis of the process, may be represented by an increase in mortality, or by a decline in the reproductive rate, as density rises. Examples of such relationships have been referred to in Section 11, to illustrate density-dependence. Alternatively, we may study the relationship between density and net increase or decrease. Of these criteria, net increase or decrease come nearer to the basic idea of regulation. Referring to Fig. 7d, if we observe the speed of recovery from reductions to different density levels, we can assess to what extent increase is greater a t the lower densities, and use this as an index of the degree of regulation. I n the context of exploitation, Fig. 7d implies an increased productivity following reduction of density. Living models of this sort of thing may be seen in certain laboratory experiments with insects. Those of Nicholson (1954b)with laboratory populations of the blow-fly Lucilia cuprina have already been referred to. I n some of these experiments (Nicholson, 1954a), the main limiting factor was the supply of food for the larvae. When the experimenter systematically removed 99% of all the emergent adults, the consequent alleviation of adult crowding allowed a 6-fold increase in the numbers reaching the adult stage. Watt (1955) used populations of the flow beetle Tribolium confusum Duv. for a study of the optimum yield

19

ANALYSIS O F PROCESSES I N CONTROL O F INSECTS

problem. As with Lucilia, the productivity of the beetle cultures increased with the rate of exploitation, sometimes up to an exploitation rate of about 90%, depending partly on the age-distribution of the residual population. 3. From the Form of the Population Curve at Different Levels of Density If we are to go beyond measuring the range of fluctuation of a population, and try to interpret the fluctuations, we need to be able to distinguish, for example, between reverses in increase that are due to the more or less direct effects of weather, and reverses or decelerations that are imposed through density-dependent processes. Whether or not the action of these latter may have been modified by the weather may also be an important consideration, but it is a secondary one. Some evidence about these alternatives can often be gained by inspecting the population curves. If the reversal is sudden, it is likely to be caused by a change in some environmental factor, such as weather. If it is approached by a gradual deceleration of increase, it may be the result of regulation, although one cannot be sure, without other evidence, that it is not due to a gradual change in weather, for example. Evidently, methods of inspecting population curves need to be formalized and made objective. One way of doing this has been described and used by Morris (19638, 1963b). He finds in his work with forest insect populations that if the logarithm of population density in

. 4-

+

P m

-

(B)

(A1 I

I

I

.. ..

I

I

I

I

I

I

\-I

FIG.8a. Annual population density of spruce budworm larvae in one plot at Green

River, New Brunswick, 1944-60. The two points below the curve are values corrected for immigration. B. Graph of log density in each generation against log density in the following generation;values from two plots. (FromMorris, 1963a.)

20

M . E. SOLOMON

one year (log N,) is plotted against that of the following year (log and the points treated as a scatter diagram, a straight line can be fitted (Fig. 8). The logarithmic scale has the effect of stabilizing the variance over the range of densities involved, so that valid regression analysis can be carried out. With reference to an artificial example based on work with the fall webworm, Hyphntria cunea (Dru.), Morris (1963b) writes ; “the slope, b = -5, which is a reasonably average value for forest insects, provides an index .of the degree of density dependence in the system. If the rate of increase in population did not decrease with density, the slope would of course be 1.0.” Morris goes on to examine how regression can be improved, and how the slope of the line is altered, when the estimated effects of certain factors are eliminated, but I do not propose to go into this. I do, however, wish to look more closely at the implications of the slope in the graph of log N , against log N,,,. Tt is not easy to do this with reference to either of the examples given by Morris. I shall use simple artificial examples to show what may be expected of the procedure with different sorts of density relationships. I n real examples, complicating factors would cause more or less scatter ; this would make the fitting of a line a matter for statistical procedure, but &s Morris (1963a) has shown, quite practicable in at least some instances. Fig. 9 will serve to show the difference between slopes for densityindependent, density-dependent and inversely density related population trends. Lines C and D in Fig. 9~ represent sustained unvarying 5-

4-

*+

c

z

m

0

3-

3

2-

1-

0

’

I

1

I

2

I

3

I

4

I

5

I

6

Time, or generotions (A)

I

7

I

8

4

L

0

I

1

I

I

3

2

I

4

I

5

log N”

(B)

FIQ. 9 ~ Curves . ilustrating various density relationships: ‘2, D, constant, densityindependent geometric increase;E , F,density-dependent;a, inverse. B. C‘ to a’, corresponding slopes of log N,,, against log N,.

21

ANALYSIS O F PROCESSES IN CONTROL OF INSECTS

geometric increase at two different rates, i.e. there is no densitydependent effect. Plotting log N , versus log N,+l for these, after reading off the values from the graph of C and I), gives the two lines C’ and D’ in Fig. 9 ~ both , with a slope of 1. Curves E and P represent increase which decelerates, on the geometric or logarithmic scale, as density increases; such a pattern, if it is established as not being due to some progressive change in weather, etc., would be interpreted as a density-dependent phenomenon. The corresponding lines E’ and F’ each have a slope less than unity, and the more strongly densitydependent F’ has a lower slope than E‘. Curve G represents the effect of an inverse relationship ; the rate of geometric increase rises as density rises. The slope of the corresponding line G is greater than 1, approximately 2.0. It may seem unrealistic to introduce this example, since a population of this sort would be inherently unstable. However, such relationships do occur, often as a temporary phase. I n practice, we often have to deal with populations fluctuating under

1 1 , 2 , 3.,4 , 5 , 6 , 7 ,8 , 9 , l o , I I , 1 2 , 1 3 , 1 2 3 4 5 6 7 8 9 10 I I 12 13 Time, or generotions

0

0

2 1

2 log Nn

3 1

2 log N”

L

3

-

2

3 log Nn

FIG.10~.Curves for three hypothetical populations which increase and decrease in unison, in response to environmental changes; but the mode of rise or fall is denaityindependent in H . density-dependentin J,and inverse in K . B. H’, J‘ and K’, corresponding slopes of log N,,, against log N,. B

C.E.R.

22

M. 1. SOLOMON

the influence of weather, etc. Can the method be used to detect densitydependence in such populations? Figure 10 shows three curves, H , J and K , representing the density of three hypothetical populations increasing and decreasing in unison as if under the influence of common environmental changes. They differ however in that H increases at the same geometric rate at all the levels of density involved, whereas J increases at a slower rate as density becomes greater, and K increases at a faster rate as density becomes greater. Correspondingly, H declines at a constant proportionate rate as density falls, whereas J declines more slowly as density falls, and K more rapidly. In H and J the rate of decrease is greater than the rate of increase (there is of course no reason to expect rates of increase and decrease to be equal, and sustained increase followed by a relatively rapid decrease is characteristic of many insect populations, and others). The graphs of log N,,, against log N , again show a slope of 1 in the case of the density-independent lines H‘. There are two lines because of the difference between the rates of increase and decrease in H , but both have the same slope. Similarly, there are two lines J’,representing the density-independent increases and decreases in J . Their slope is 0.625 and 0.60 respectively; the similarity is accidental, since the rules governing the density-dependent increases and decreases in J were separately and arbitrarily invented. The lines K’ representing the increase and decrease of K have slopes of 1.33 and 1.44 respectively, the values above unity indicating the inverse density relationship embodied in K . I n brief, Fig. 10 demonstrates that the method can be used to assess the degree of density-dependence in populations that are fluctuating mainly under the influence of densityindependent factors ; also, it shows that phases of increase and decrease tend to require separate treatment, a point not dealt with by Morris (1963a,b). As already mentioned, Nicholson (1933) envisaged populations as being regulated towards an equilibrium level which continually changed. While acknowledging the value of this idea on a theoretical level, I have objected that it would be difficult in practice to distinguish movement towards an equilibrium, on the one hand, from changes in the equilibrium level, on the other (Solomon, 1949). Another difficulty that might often arise is the imposition of sudden reductions by densityindependent influences that have no reference to any equilibrium level. I believe the method outlined above might in certain circumstances be used to distinguish between (i) movement towards an equilibrium, (ii) a change in the equilibrium level, and (iii) density-independent changes. Fig. 11 illustrates a simple hypothetical example embodying these features. As before, the graphs of logN,,, against logN have separate lines for the rising and falling approaches to equilibrium. The

23

ANALYSIS O F PROCESSES I N CONTROL O F INSECTS

4

3

‘*FaFirst equilibrium

\*--level-____--\*--level--

5

E2 level

1

0

1

1

1

2

1

3

1

4

1

5

1

6

1

7

1

8

Time, or generations

(6)

0

1

1

9

10

1

II

1

1

12 13

1

FIG. l l ~Curve . of a hypothetical regulated population which tends towards an equilibrium level (steps 1, 2, 3), then suffers a density-independent reduction (4), rises again towards the equilibrium level (6, 6), then falls towards a modified equilibrium level (7, 8, 9), suffers another density-independent reduction (lo), then rises again towards equilibrium (11, 12, 13). B. Slopes of log N,,, against log N,.

interesting point is that the change in the equilibrium level is clearly apparent; if we had not known about it in advance, the arrangement of the points along two different linear tracks would have revealed it. Also, the points for the two density-independent reductions fall well away from the others. One can readily envisage conditions in which these differences would be less clear, even indistinguishable. One of the conditions necessary for the positive success of the method is that the changes in equilibrium

24

M. E. S O L O M O N

level should not be too frequent. (Failing this condition, the method could still provide the negative information that the population density was not being uninterruptedly regulated towards any equilibrium level that remained substantially the same over three or more census intervals.) Nevertheless, under some conditions it might provide a somewhat penetrating analysis of the dynamics of a population. So far, only the prompt type of density-dependence has been considered. How does the method function if a population is involved in a parasite-host cycle, so that the density-dependence operates with a lag

FIG. 12. Numbers in successive generations graphed as log N,,, against log N,, for populations fluctuating under the influence of parasites. A. A numerical example of the Nicholson-Bailey model. B. Data for the black-headed budworm in conifer forest in New Brunswick. From Morris (1959),with addition of lines of slope 1.0.

between generations rather than within each generation? As pointed out by Morris (1959), the Nicholson-Bailey theoretical model for parasite-host interaction gives a spiral when log N,+l is plotted against . slope changes stepwise through 360"as one follows log N , (Fig. 1 2 ~ )The the successive generations around the spiral. Morris (Zoc. cit.) also illustrated a real example of parasite-host oscillation, using the estimates of population density of the black-headed budworm that have already beengraphed here in Fig. 2. The graph of log N,+l against log N , (Fig. 1 2 ~forms ) a closed spiral or ellipse. One way of dealing with a figure of this sort is to ignore the linkage of successive points and treat it as a scatter diagram. When this is done, one can calculate the slope of the straight line that best fits the points. It seems that the points for examples of the Nicholson-Bailey model require a line of slope 1.0. Morris calculated the slope for the blackheaded budworm data as 0.78. If the slope of 1.0 is assumed to represent an inherent tendency of actual parasite-host oscillations, the lower value may indicate that other, more immediate density effects are influencing the population.

ANALYSIS O F PROCESSES IN CONTROL O F INSECTS

25

But generally, when parasite-host oscillations and elliptical or polygonal patterns are involved, the method should be regarded as only a first step in analysis. Some later steps are illustrated by Morris (ZOC.

cit.).

The main requirement of the method is a series of estimates over an adequate run of generations; also, as Morris (Zoc. cit.) pointed out, it is desirable to have figures for a wide range of densities to work upon, so that density-effects may show up clearly. Where adequate data are available, the method can be expected to provide a simple and useful index of the degree of density-dependence, or regulation. The slope of the line shows whether the density-dependent relationship comes into play strongly or only feebly as density rises and falls ; and the scatter of the points about the fitted line may show the degree of influence of minor disturbing or modifying factors ; but if consistent densitydependence continues to operate during the course of population changes imposed by density-independent influences, no such scatter arises (Fig. 10). Marked changes representing breaks in the densityrelationships tend to give points widely separated from the densitydependent slopes (Fig. 11). A marked change in the “target” equilibrium level gives rise to a new set of points to which a separate line must be fitted (Fig. 11).

4. By Correlation between Density and Weather Indices This method can be illustrated by reference to a much discussed example. Entomologists at the Waite Institute, Adelaide, S. Australia, collected twenty rose blooms from the garden of the Institute daily over a number of years and counted the numbers of the apple blossom thrips, Thrips imaginis Bagnall, in each sample (Davidson and Andrewartha, 1948a,b). The insect bred in various flowers, but not in these roses, which served merely as traps. The numbers of thrips in the samples depended on the weather at the time and on the abundance of roses and of other flowers, as well as on the abundance of thrips. After declining from a minor autumn peak, their abundance greatly increased during the main flowering season in spring (September onwaras), reaching a peak in November or December, then rapidly fell as the hot dry season set in. The relationships between the density of the population in the roses and the weather factors were set out by Davidson and Andrewartha (194813). The degree of association between the density of thrips and the weather during the preceding months was measured by partial regression. Four components of the physical environment were selected and 78% of the variance of the population was attributed to their influence. The maximum density attained in the spring was considered

26

M . E . SOLOMON

to be largely determined by the weather of the preceeding autumn, but sometimes modified by conditions during early spring. Andrewartha and Birch (1954) discussed the work in full and concluded that since they had shown most of the variation was due to weather, there was no room for density-dependent regulation. They claimed that the abundance each year was determined by the previous density and the influence of density-independent factors, favourable and unfavourable. The conclusions of Andrewartha and Birch (Zoc. cit.) have been subjected to a series of adverse criticisms on the grounds that they are extremely improbable or that density-dependence can be demonstrated in the data, or both, by Solomon (1957), Kuenen (1958), Nicholson (1958), Smith (1961; see also Andrewartha, 1963 and Smith, 1963), Klomp (1962) and Varley (1963). The question I raised was how, without regulation, did the population fluctuate about the same level over the years, instead of drifting upwards or downwards persistently? Nicholson (Zoc. cit.) went into this matter more fully and argued that if the “peak numbers” were highly correlated with weather conditions, then the implication was “that the weather-induced changes in numbers of thrips took place each year from some approximately constant number”. Therefore, some factor “must have adjusted the intensity of its action in relation to the numbers coming under its influence, for these varied greatly, from year to year. I n short, a density-governing factor is required to operate in this situation”. Klomp (1962) elaborated the same point to good effect with the aid of diagrams. I n simple general terms, the argument may be summarized as follows: if a persistently good correlation is found between a seasonal weather index and density, this implies that the weather effects operated each year on a population starting from some fairly constant base level of density, a feature that can be maintained only by some form of regulation.

c.

ASSESSING’ THE ROLE OF PARTICULAR FACTORS

1. By Direct Inspection of the Relationship between Effect and Density I have already referred to the detection of density-relationships in rates of mortality, reproduction and net increase (e.g. as in Fig. 1). The rate of loss by dispersal or gain by immigration can be examined in the same way when such movements occur, since it cannot be taken for granted that they are density-independent. The next step may be to examine the density-relationship of the effect of particular factors on one or more of the above-mentioned processes. For example, we may enquire into the effect of a particular predator, or of several species together, upon the species we are studying, to

ANALYSIS O F PROCESSES I N CONTROL O F INSECTS

27

determine whether this predation causes a higher or lower percent mortality when the prey is (or has recently been) sparser or more abundant (cf. Fig. 4).There is no need to dwell on this, for examples of density-relationships or of density-dependent (etc.) factors are often presented in such terms.

2. By K e y Fmtor Analy& One of the first phases of the study of a field population is often an attempt to discover which factors are mainly responsible for the variations in abundance. Whether or not their action is dependent on the density, we need to know about them in order to carry the analysis to the stage of understanding regulation, to say nothing of the needs of economic entomology and the importance of being able to predict increases in abundance. Morris (1959, 1963a, b) argues the practical advantages of this approach, and shows how it has been developed in the study of Canadian forest insects. Life-table data for the spruce budworm (to quote from the former paper) “suggested that the factors affecting this species in any one place are of two types - those that cause a relatively constant mortality from year to year and contribute little to population variation, and those that cause a variable, though perhaps smaller, mortality and appear to be largely responsible for the observed changes in population (Morris, 1957). A factor of the latter type will here be called a ‘key factor’, meaning simply that changes in population density from generation to generation are closely related to the degree of mortality caused by this factor, which therefore has predictive value. . . . If a key factor is suspected in a population, can its existence be demonstrated effectively by a very limited ‘single-factor’approach in which only population density and the mortality caused by this factor are measured in each generation?” He gives two examples of this single factor analysis (Morris, 1959). We need refer to only one of them, based on the observations over twelve generations of the black-headed budworm referred to earlier (Figs. 2 and 1 2 ~ )Parasitism . in the larval stage was the suspected key factor. Population density was estimated at the appropriate stage of larval development, and parasitism was measured by dissecting or rearing the larvae from this sample. The test was made by comparing the numbers surviving after this factor had acted (8,) with the total numbers of larvae in the next generation (N,+,). The correlation had to be significantly better than in a straight-forward comparison of N , and N,+l, if the parasitism were really acting as a key factor. I n fact, it was found to be so ; r = 0.93 for log N,+l and log S,,, r =0.67 for log N,,, and log N,. Squaring these

28

M . E. SOLOMON

values of r gives 0.86 and 0.45, suggesting that 45% of the variance in larval density is accounted for when log N , is used for prediction, as compared to 86% when log S, is used. The regression formula is log N,,,

= 0.53

+ 0.92 log s,,

in which the estimated slope of b =0.92 is not significantly different from the value b = 1 to be expected for a perfect key factor accounting for all the variation. It is concluded that larval parasitism was a key factor and that the insect was remarkably unresponsive to variations that must have occurred in other factors. This is in marked contrast to what was found for the spruce budworm, in which effects of weather accounted for most of the variability (Morris, 1963a). This short account omits many aspects of the paper, but should make at least brief reference to Morris’ insistence that “the single-factor approach should be recognized as a useful lead to more complete studies but certainly not as a substitute for them”. A final point that should be noted is that the black-headed budworm population was involved in a parasite-host oscillation, i.e. in a delayed type of densitydependent relationship ;inspection of the data shows that the percentage parasitism was correlated with the host population density of the previous generation rather than with current density. Varley and Gradwell (1960), in a note on key factors as defined by Morns (Zoc. cit.), described a graphical method of demonstrating them. They compared killing powers of a series of successive mortality factors acting on a population of the winter moth, Operophtera brumata (L.), on oaks in Wytham Wood near Oxford. They estimated the numbers at two stages in each life cycle of this insect, which has one generation per year. For each generation they took the difference between the logarithms of the two estimates. This value (k-value) is equivalent to the logarithm of the factor by which the population has been reduced; e.g. if the density is 50 on the first occasion and 10 on the second, the k-value is log 50 -log 10 = 0.699, or log (50/10). Further, with the aid of additional information, the mortality was subdivided into 6 parts according to the developmental stages and causal agents, and each part was treated as above, giving values of k,, k, ... k6 which in sum are equal to K . When the process was repeated with the data for each of 10 successive years, and the values plotted against time (Fig. 13), two different sorts of k-values were distinguishable. That for “winter disappearance’’ (kl) followed the same fluctuating course as K , only the fluctuations tended to be greater, But the k-value for pupal predation (k5)ran counter to the fluctuations of k,, and so partly offset them. Clearly, k, represents the “key factor”. The second type (kS)apparently tends to compensate

29

ANALYSIS O F PROCESSES I N CONTROL O F INSECTS

0.1

.-.-.-./.-.-.

r k~.-./m-.-.-.,

-*

0 .o

1950

1955

1960

FIG. 13. Mortality affecting winter moth at Wytham, Berkshire, on logarithmic scale. From Varley and Gradwell (1963), with addition of population density (crosses), converted to same scale after reading from their Fig. 2. K , total mortality=k, +k,...k,; k,, winter disappearance; k,, pupal predators; k,, k,, k,, k,, mortalities due to various parasites, and disease.

for variations in mortality and so to reduce the variability of total mortality from year to year; this was thought to be an aspect of a density-dependent relationship : “pupal predation in any one year was higher under trees with a high pupal population than under adjacent trees with low populations.” It is interesting to compare these mortalities with the levels of larval density in the same generations. This can be done by superimposing a graph of population density (crosses and broken line), as in Fig. 13. Winter disappearance (k,) then seems to be inversely related to the abundance of the surviving larvae, at least over the period 1951-60 inclusive. “Winter disappearance includes all the mortality from the time the adult moths are sampled in November and December to the B2

C.E.R.

30

M . E. S O L O M O N

subsequent count of fully fed larvae in May, and is calculated on the basis of an assumed constant egg production.” They attribute it chiefly to mortality of first-stage larvae, which they know to be both great and variable. It is understandable that the higher values of k, are usually followed in the same generation by relatively low values for larval density, while the lower values of k, allow greater larval abundance. The addition of the oensus data to the diagram also seems to confirm emphatically that pupal predation is a promptly density-dependent process, causing a proportionate mortality (k,)that rises and falls with population density. Provided the data are good enough, the method described by Varley and Gradwell can be used to distinguish inverse, lagging and prompt density relationships. However, it calls for more detailed observations than the single factor method described by Morris (Zoc. cit.). The latter requires only one annual census and an estimate of the mortality due to the suspected key factor. Varley and Gradwell need mortality data for each factor or group of factors to be represented by a k-curve. I n return for these extra demands, their method reveals the importance and operational nature of as many factors as are represented by adequate data. Thus, although they introduce their method as a type of key factor analysis, its scope is wider than this. It has a good deal in common with the analysis of survival rates undertaken by Morris and his colleagues in their work on the spruce budworm, in which survival ratios from successive mortalities in the life-cycle are multiplied together and related to the change in numbers from each generation to the next. The essence of Morris’ key factor method is to concentrate on a single factor, irrespective of the way in which its action may turn out to be related to density, consideration of other factors being postponed until the time comes for more extended analysis by other methods. Some of the ways forward from this stage are indicated by Morris (1963) and Morris (ed., 1963). The method of Varley and Gradwell, if reduced to its simplest form, would be the same as that of Morris. One census each year would be required, together with observations to provide a value for the mortality (k,) due to a suspected key factor. The total mortality ( K )would be estimated from the census figures and a value for the egg output. Then all the mortality except k, would be called k, ( = K - kl),and a graph would show what proportion of the variation in K was due to k,.

3. By Life-table Analysis I n studying an insect population in the field one normally begins without knowing which is the key factor (causing the main variation in numbers from one generation to the next), and which is the regula-

ANALYSIS O F PROCESSES I N CONTROL O F INSECTS

31

tory factor. (There may be more than one of each, but for verbal convenience I write here as if there were not.) A natural early step is to try to identify these factors by looking for their effects on abundance. To discern these effects by direct inspection we must look not only at the adult or any other single stage, but at the whole life-cycle. When we know at what stage the variation is imposed or the regulation exerted, it may be possible to identify the factors responsible. The appropriate data for this approach are tables of the successive mortalities through the life-cycle. For this purpose, what we need are not average or standard life-tables, but a series of tables, each of them based on censuses in a different generation or year. They should preferably be life- and fertility-tables, showing the variation in reproduction as well as in survival. To demonstrate these points, I shall examine the tables given by Richards and Waloff (1961) for a population of the broom beetle Phytodecta olivacea-in the grounds of the Imperial College Field Station in Berkshire. The species disperses relatively little in this area, and the population is considered to be nearly self-contained. It is associated with a plantation of about two acres of broom, Xarothamnus scoparius. The beetles emerge from the soil from late April to June and lay their eggs on the plants. The eggs and the four larval instars are found on the broom simultaneously for most of the time between May and the end of August. The fully fed larvae descend and pupate in litter and grass roots under the bushes. The adults emerge in autumn and return to the plants to feed, but again descend to hibernate, remaining immature until the spring. A fifth to a third of them survive another year and hibernate a second time. The thoroughness of the sampling and the cross-checking of estimates by different methods were impressive, and the paper gives a full account which repays study, but these matters can be passed over for the present purpose. Table I shows the annual changes in population density in five successive phases of the life-cycle. During the &year period the amount of broom, estimated in “armfuls”, was halved in the winter of 1955-6, and almost halved again in the winter of 1957-8. Because of these changes, it is particularly important to deal here with density rather than with total numbers. I have therefore converted all the population estimates to numbers per armful, even for insects on the ground, where 2-70 m2 corresponds on average to one armful of foliage above. I n the right-hand column of Table I, I have calculated the coefficient of variation (100 x S.D./arith. mean) of the five annual values for each phase of the life-cycle. Comparisons of these expressions of the degrees of fluctuation suggest the following conclusions about the population of Phytodecta over the period concerned,

32

M . E. SOLOMON

TABLEI Data from Richards and Waloff (1961, Table 34),on a Population of the Broom Beetle, Phytodecta, recalculated in Terms of Numbers per Armful of B’oliage. The Figures for Numbers of Armfuls are from their Table I and from Richards (1963)* 1954

1955

Armfuls 4769 Eggs 96.6 4th instar larvae and pupae in soil 13.4 Adults, autumn 7.1 Autumn adults surviving to spring 2.8 Spring adults surviving to following spring 0.56

4615 142.9

1957

1958

Coeff. of variation

2100* 2576 334.1 78.4

1429 87.5

72.4

1956

29.8 8.2

3.4 2.5

17.4 2.9

7.3 3.7

71.8 53.1

2.6

1,35

1.9

1.9

27.0

1.08

1.69

0.73

0.88

44.2

Firstly, the “key factor(s)” responsible for the main variation in abundance comefs)after the counting of the adnlts, at which stage the coefficient of variation (C.V.)was 27-0 and 44.2, and before the counting of the eggs, by which stage the C.V. had risen to 72.4. This suggests that some density-independent influence such as a weather factor affected the numbers of eggs laid. Richards and Waloff estimated the mean fecundity per female (in the field) in the five successive years as 77.3, 71-4, 78.3, 58.7, and 32.9. This may largely account for the low densities of 1957 and 1958, but does not explain the very high density of 1956. This latter can be attributed mainly to the greatly reduced amount of host plant in that year, leading to greater concentration of all stages. (It should be mentioned that the population of eggs was not as dense as the figures in Table I suggest, because they were laid throughout the summer, and were seen in the field overperiods ranging from 86 days in 1954 to 112 days in 1958. This would have to be taken into account if density-effects on the eggs were being considered; but for the present purpose we can ignore it, noting only the long periods during which the egg-laying females may have been exposed to disturbing influences.) Whatever may be the correct explanation, the C.V. values strongly suggest that the chief cause of fluctuation, i.e. the key factor, operated at or about the time of the laying of the eggs. Table I shows there was no appreciable change in the C.V. from the egg stage to the counting of the 4th stage larvae and pupae in the soil. This suggests that neither regulation nor any persistent source of fliictuation operated during this period. It could be of course that both

33 these phenomena were active and their effects cancelled. This seems unlikely, but since the fluctuations of larvae and pupae followed a somewhat different pattern from those of the eggs (Table I), there is an element of accidental coincidence in the similarity of the C.V. values. While this result suggests there was no major key factor nor regulatory process in action during juvenile development, there was a good deal of mortality, as is obvious from the values in Table I. Most of this was due to Mirid bugs and other predators which together caused mortalities of larvae and eggs estimated at values from 77.8% in 1957 to 98.9% in 1956. Richards and Waloff wrote: “The population of mirids is probably in no way dependent on Phytodecta and is controlled by its own complex of parasites and by fratricidal predation.” But: “If the quantity of broom is reduced, Phytodecta and its predators become more concentrated and predation becomes more intense.” Richards (1963) extended this argument and tabulated figures showing that percent predation increased when the quantity of broom was reduced (1956, 1958). When percent predation is graphed against the density of adults in spring of the same years, there appears to be a well-marked density-dependent relationship. But when the estimates of egg density are used instead of adults this relationship does not hold, except for 1956 compared with the other years as a group. It seems clear that a heavy reduction of Phytodecta, due to contraction of its host plants and concentration of its predators, occurred a t least in 1956. Table I shows the reduced numbers of late larvae and pupae in this year. It also shows that this reduction overshot the mark, taking the numbers to an appreciably lower level than in the other four years. This overshooting (resembling the action of a lagging rather than a prompt density-dependent factor) contributed to the high value of the C.V. for late larvae and pupae. The second hypothesis to be drawn from the ranges of variation shown in Table I arises from the fact that the coefficient of variation was 71.8 at the time of sampling the larvae and pupae in the soil and only 53.1 at the sampling of adults in the autumn. This suggests that a regulatory effect came between these two events. Evidence in support of this conclusion can be found by taking from the life-tables the numbers of 4th stage larvae and pupae in the soil, converting them to density values, and plotting them against the percent mortality occurring between that stage and the autumn adult stage; these mortality values, as percentages of the larvae and pupae in the soil, are read from the life-tables. The resultant graph (Fig. 14) suggests that the mortality between the two stages was strongly density-dependent. The relationship shown in Fig. 14 is such as to compensate to a considerable extent ANALYSIS O F PROCESSES IN CONTROL OF INSECTS

34

M . E. SOLOMON

19.57

1925

1924

L

0

I

10

20

I

30

No. of 4Ih lorvoe and pupoe in soil, per armful of plant

FIQ.14. Estimated mortality from pupal or 4th larval stage of the broom beetle, in the soil, t o adults in autumn, in relation t o population density of the former. An armful of plant ~ 0 . 3 m2 7 on the ground. Data from Richards and Waloff (1961).

for the effects of heavy predation of the earlier stages while they were on the plants : in years when the density had been most heavily reduced in this way the percentage mortality a t this latter stage was least hence the reduced coefficient of variation in Table I for the density of adults in the autumn. Whatever factor was partly or wholly responsible for this apparent regulation may have acted either while the insects were on the plants, presumably after the main impact of predation, or in the soil before the census was taken. Table I shows a further reduction in the coefficient of variation from 53-1for the adults in autumn to 27-0 for the survivors of these adults in the following spring. This suggests that a further regulatory process was a t work between these two samplings. By applying the same procedure as above, Fig. 15 was drawn. This graph supports the conclusion that mortality between these two samplings was density-dependent. Table I further shows that the density of adults which survived from one spring to the next was rather more variable than the numbers in the first spring (44.2 compared with 27.0). I shall not speculate on the possible significance of this.

ANALYSIS O F PROCESSES I N CONTROL O F INSECTS

35

1955

'Or

e54 c

c

-0 3

1927 I

I

I

I

I

1

Adults per armful of plant, in autumn

FIG.15. Estimated mortality of adult broom beetles from autumn to following spring, in relation to their autumn population density. Data from Richarda and Waloff (1961).

4 . By Artijicial Reduction or Exclusion of a Regulatory Factor If a particular factor is thought to play a part in regulating the abundance of a population, this can often be tested by eliminating the factor or reducing it, and assessing the consequent change in numbers. Entomologists at Riverside, California, have developed this approach for evaluating the effect of natural enemies on scale insects and mites infesting fruit trees (DeBach et al., 1949, 1950, 1951, DeBach 1955, 1958). The methods used have included (i) enveloping small branches in fine cloth sleeves, some open at the end, others closed to exclude parasites, (ii) local treatment with a selective insecticide that will kill the natural enemies and leave the host population little affected (see also Huffaker and Spitzer, 1951), and (iii) allowing ants, artificially fostered if necessary, to have access to certain trees where they greatly reduce the effectiveness of natural enemies. The third method has recently been used by Banks (1962) in experiments with Aphis fabae Scop. on potted bean plants in cages. These were placed near nests of the ant Lasius niger in a garden, but the ants were excluded from some of the plants. He demonstrated that the ants drove away most predators, while the colonies of aphids not protected in this way were eliminated or had their numbers persistently restricted by the predators. Lack (1954) gives an account of some analagous experiments with vertebrates.

D. I N T E R P R E T I N G EFFECTS O F SUCCESSIVE MORTALITIES A comparative study of the numbers in a population killed by different factors is only a beginning in the investigation of its population dynamics, and, by itself, readily misleading. The building up of

36

M . E. SOLOMON

mortality tables is of course valuable, indeed indispensable for a thorough investigation. The crux of the matter, however, is how to assess the real significance of the different mortalities, since the ultimate effect of each depends on the contributions of the others, and the degree of responsiveness of each mortality factor to difference%in population density is also an essential consideration. One line of thought, with various branches, is represented by the papers of Thompson (1928, 1955), Bess (1945), Morris (1957), and some others. The earlier ones were reviewed by Morris (Zoc. cit.) who discussed the significance and limitations of the proposals of Thompson and Bess, and indicated some general principles governing the effects of mortalities in combination. Another line of thought on the subject was developed by Nicholson (1933) and Nicholson and Bailey (1935). They were primarily concerned with the elaboration of their model representing parasite-host interactions. They gave some consideration to the effects of several species of parasites attacking a common host, and to the effects of densityindependent mortality impinging on a parasite-host system, in the terms of the model. Varley (1947) applied their theory to the interpretation of field observations on the knapweed gall-fly, including a consideration of the effects on its abundance when density-independent mortality acted before or after a major parasite. Neither of these lines can be easily summarized, and I do not propose t o attempt it. It seems more useful to state clearly some simple principles concerning the combination of mortalities. These rules or relationships are self-evident when clearly understood, and do not need mathematical proof. The numerical examples are for illustration, not proof. I n calculations of this sort, it is generally more convenient to combine survival rates than mortality rates. For example, ifa 70% mortality is followed by a 90% mortality of the survivors, the simplest way of calculating the result is to consider the survival rate, 10% of 30%, i.e. 3% of the original numbers. The fact that successive survival ratios in a life-cycle can be combined by multiplying together has been used as the basis of models and calculations by the investigators of the spruce budworm (Morris, ed., 1963). They use ratios instead of percentages, so that if N , animals are reduced by one mortality to N , and by a later one to N , , one writes N , / N , x N J N , =N , / N , , which is convenient for dealing with population counts. It makes no difference to the calculation, of course, what numbers one writes in one of these fractions, e.g. N , / N , , so long as the ratio is preserved. It is often convenipnt to convert a percentage reduction to this ratio form, e,g. 3/10 for the survival ratio from a 70% mortality.

ANALYSIS O F PROCESSES I N CONTROL O F INSECTS

37

Another necessary preliminary is to distinguish the following types of mortality. (i) Density-independent : killing a percentage of the total population, irrespective of density. (ii) Promptly density-dependent : killing a higher percentage at high density, a lower percentage at low density. Fig. 16 shows three relationships of this sort, which will be used in the numerical examples. (iii) Killing a more or less constant number of animals, irrespective of density. In the context of one generation, this is an inverse density-relationship (the higher the density, the lower the proportion killed). But it is introduced here as a simplified representation of what may happen in certain parasite-host interactions (delayed density-dependent), in the early upswing of each cycle. It is assumed that the parasites have a more or less constant capacity each to attack a certain number of hosts over the range of densities involved (cf. Morris, 1957). Whether or not it is widely representative, the implications of this “constant-number” type of mortality are worth consideration. For the numerical examples, a hypothetical population of 200 young insects is taken, and the mortalities are considered to operate in succession on the juvenile stages. This figure could also represent the egg output per female under favourable conditions, and reductions in this output could be treated like mortalities.

1. Density-independent Mortalities in Combination Rule i : The calculated final effect of a succession of density-independent mortalities is not affected by the order in which they operate. This is generally recognized (e.g. Morris, 1957). For example, if 70% mortality is followed by 90% mortality of those remaining, or vice versa, in either case we have 200 x 3/10 x 1/10 = 6 survivors. 2. Density-dependent Mortalities in Combination Rule ii : The calculated final survival from two density-dependent mortalities acting in succession is lower when the more powerful mortality operates first. This follows because the greatest effect is achieved if the more potent factor has the advantage of high density to act upon. Consider, for example, mortality A in Fig. 16 followed by mortality C. The percentage reduction in current density inflicted by each will depend on the level of that density at the time ; the appropriate values are read off from Fig. 16 : 200 x 2.5/100 = 5 ; then 5 x l O O / l O O = 5 survive.

I n reverse order (CA) : 200 x 60/100 = 120; 120 x 22.5/100 = 27 survive.

38

M. E. S O L O M O N

Although it is not so obvious, worked examples confirm that if three density-dependent mortalities follow in succession,the greatest reduction is achieved when the most potent comes first, and the least when it comes last. If two lines like those in Fig. 16 intersect, the change in the order of potency beyond the crossing-pointmust be taken into account.

3. Denszty-dependent and Density-independent Mortalities in Combination Rule iii: I n a calculation of the final effect of a density-dependent mortality and one or more density-independent mortalities, the greatest

Population density FIG.16. Three density-dependentmortality relationships.

reduction is achieved when the density-dependent mortality acts first. This follows from the fact that density-dependent mortality is the more increased at high density, and the more reduced at low density. For example consider a density-independent mortality M of 20% in succession with B (Fig. 16) :

BM : 200 x 25/100 =50 ; 50 x SO/lOO =40 survive MB : 200 x 80/lOO = 160; 160 x 40/100 = 64 survive.

4. Constant-number and Density-independent Mortalities in Cornbination Rule iv: I n a calculation of the final effect of one or more densityindependent mortalities and one killing an approximately constant number of individuals, the greater reduction is achieved when the density-independent mortality comes first. This follows because the density-independent mortality kills a set percentage, hence a greater number when the density is higher. (It is perhaps worth repeating here

ANALYSIS O F PROCESSES IN CONTROL O F INSECTS

39

that it is by definition the percentage, killed by a density-independent factor, that is independent of density -not the number killed.) For example consider a density-independent mortality M of 50%, and a constant-number mortality K killing 50 individaals :

KM: 200 - 50 = 150; 150 x 50/100 = 75 survive MK : 200 x 50/100 = 100; 100 - 50 = 50 survive. 5 . Constant-nurhber and Density-dependent Mortalities in Combination Rule v: I n a calculation of the final effect of one or more densitydependent mortalities and one killing an approximately constant number of individuals, the greater reduction is achieved when the density-dependent mortality comes first. This follows because densitydependent mortality kills a greater number of individuals when the density is higher (to an even greater extent than in the case of densityindependent mortality). For example, consider the density-dependent mortality B (Fig. 16) and a constant-number mortality K killing 50 individuals :

KB : 200 - 50 = 150; 150 x 43-5/100= 65 survive BK: 200 x 25/100= 50; 50 - 50 = 0 survive. Obviously, much more complicated models could be built upon these simple foundations, with mathematics taking over from intuition. This would defeat the purpose of the present exercise, which is to show that the effects of a succession of mortalities of different known types can be calculated very easily -provided, of course, we know or can assume how many, or what proportion, of the population will be killed by each mortality at the level of density prevailing when it operates. It is assumed in all this that variations in environmental conditions, if they cannot be ignored, can be adequately allowed for in the calculations.

6. Assesszng the Signi$cance of Regulatory Mortalities One reason for studying the effects of different mortalities acting in combination is to learn something of the contributions they make to regulation. Nicholson (1933) and many others since have emphasized that a knowledge of the numbers or percentage killed by a factor cannot by itself tell us whether the factor is making a great or small contribution to regulation; it is the ability to offset a rise in density that is primarily important in this connection. Obviously, if a markedly density-dependent factor (prompt or lagging) accounts for most of the mortality in each generation, the case is clear - unusually so ! But a factor that causes only a small fraction of the total mortality per generation can play a dominant part in regulating abundance. Many

40

M . E. SOLOMON

animals produce large numbers of young most of which die before maturity. I n many such cases, the greater part of this mortality seems to be density-independent ; a t least it is difficult to prove otherwise. Suppose in an insect population of 10,000 adults, half of them females, spread Over a particular habitat, each female produces on average 200 young, and that density-independent mortality always kills at least goyo, and sometimes as many as 98%. If there were no subsequent mortality, the increase would be tenfold at the lower level of mortality and still twofold a t the higher level. Suppose however there is a sensitively and promptly density-dependent factor regulating the survivors of this mortality to approximately the same density as the parent generation. This regulating factor kills a relatively small percentage of the original number of young, yet it is entirely responsible for regulation. This is one way of looking a t the matter. But it may well be that this regulating factor can maintain control only over a limited range of abundance, so that the stability of the system is dependent on the heavy preliminary mortality as well as on the “finishing touches” by the regulating factor. Of course, these “finishing touches” seem to be small because we have assessed mortality in terms of the original numbers of young. This is a somewhat unrealistic way of looking at the action of a densitydependent factor, even in an over-tidy example of the sort we are considering. For its action (when not of the lagging type) is dependent on the density of population at the time, not on the initial density. If there has been a preliminary mortality of go%, the regulatory factor will operate on a population of 100,000, which for perfect stability, and assuming no other mortalities in the life-cycle, it will have to reduce to 10,000 - a reduction of 90%. If, on the other hand, the preliminary mortality is as much as 98%, the regulatory factor will have to reduce 20,000 individuals to 10,000 by a 50% mortality. I n brief, if a regulatory factor operates by “finishing off” after most of the young have already been killed, it will kill only a small percentage of the original numbers, but this may be a high percentage of the numbers present a t the time when it acts. Certainly it must have the capacity to kill a high proportion of those present when the residual density is higher than usual, if it is to maintain effective regulation. A real example of these relationships can be selected from the data of Richards and Waloff (1961) on the broom beetle (cf. Table I). The original high numbers of eggs were reduced to low numbers of adults in the autumn. These were reduced to even lower numbers by the following spring, as also shown in Table I. The halving of the value of the coefficient of variation from autumn to spring strongly suggests the operation of a regulatory process, and this is supported by Fig. 15.

41

ANALYSIS O F PROCESSES I N CONTROL O F INSECTS

TABLEI1 Late-stage Mortality of Broom Beetle Phytodecta olivacea, f r m the Life-Tables of Richards and Waloff (1961) ~~

Mortality as yo initial numbers (eggs) up to autumn adult stage up to spring adult stage difference Mortality of adults, autumn to spring, aa % of autumn adults

1954

1955

1956

1957

1958

92.6 97.0 4.4

94.3 98.2 3.9

99.2 99.6 0.4

96.4 97.6 1.2

95.7 97.8 2.1

60.3

68.2

46.4

33.6

48.8

Table I1 shows that this regulatory effect was achieved by a small increment in the high mortality that had already taken place when the autumn adults were counted. This mortality ranged from 92.6 to 99.2% of the original numbers (eggs) in different years, and the incremental mortalities represented from 0.4% to 4.4% of the eggs. I n spite of the smallness of the incremental mortalities they considerably improved the regulation of the population. This was possible because of the preceding high mortality (cf. Morris, 1957, on this point). Also, although the incremental mortalities represented small percentages of the eggs, they represented substantial percentages of the population at the times when they occurred - 33.6 to 68.2% of the autumn adults.

E.

THE STATUS O F METHODS O F ESTIMATION

This Section began with the quantitative description of observed degrees of population fluctuation or of constancy. This was followed (in B and C) by the discussion of various methods involving calculations (or their graphical equivalent) from observations; the aim of the calculations is always to provide evidence for or against some particular hypothesis, and, if it is supported, to suggest an analysis of the observations. However correct the hypothesis may be, the analysis may be faulty, because of circumstances unknown or not understood. ,Nevertheless, the way forward is along such uncertain pathways, and in the exploration and identification of blind alleys. To put the matter in another way, although correlation does not establish causality, it may correctly suggest it, and it does provide evidence. Several of the methods that have been used or suggested depend on experiment. This can provide a transition from hypothesis to established fact, decisive in proportion to the degree of support given t o the experimental procedure by the use of controls, repetitions and statistical tests. Although it does not go beyond words and simple arithmetic, the

42

M. 1. S O L O M O N

function of the sub-section on successive mortalities is the same as that of mathematical studies. It points out consequences of the different ways in which mortality may be related to population density. While a number of actual examples have been studied, some of the methods have been illustrated with hypothetical examples. This is partly but not primarily a matter of convenience. I n spite of the increase in the data available from studies in population dynamics, there are still too few realistic demonstrations and convincing tests of many of the methods discussed. More detailed information is needed, and more extended studies through long series of generations. It is also desirable that field investigators should bear in mind the various ways in which the phenomena of population dynamics can be assessed, so that these methods and the investigations can be developed together. Although simple ideas make good starting points, they often have to be modified and elaborated before they can be successfully applied to real examples.

V. DENSITYRELATIONSHIPS IN THE ACTIONOF PREDATORS A N D PARASITES A. F U N C T I O N A L A N D N U M E R I C A L RESPONSES O F NATURAL E N E M I E S Predators and insect parasites (parasitoids) are so often involved in natural control that it is appropriate to pay some attention to recent advances in our understanding of their mode of action. Since my prime concern is with the role that natural enemies may play in regulating the populations of their prey, I shall concentrate attention on the ways in which their action is related to the prey population density. Some time ago, in a discussion of the theory of natural control (Solomon, 1949), I pointed out that natural enemies could respond to a change in the density of their prey (e.g., an increase) in two distinguishable ways: (i) by a functional response, in which each predator takes more of the prey, ‘or takes them sooner, and commonly also (ii) by increasing in numbers (numerical response) through increased survival or. reproduction or through immigration. Recently, C. S. Holling has greatly developed this theme, both with reference t o his own field studies of predation by shrews and deer-mice on pupae of the pine sawfly Neodiprion sertijer (Geoff.) (Holling, 1959a),and also in the course of a more general study of the relationships involved in predation (Holling, 195910, 1961).

B. A LABORATORY M O D E L O F F U N C T I O N A L R E S P O N S E Unless a predator can reproduce fairly promptly when the prey increases (or unless there is a reduced mortality of the predator or

ANALYSIS OF P R O C E S S E S IN CONTROL OF INSECTS

43

a gain in numbers by some adjustment of its migratory movements), predatory responses in the short term must be purely functional. I t is therefore interesting to enquire whether or not functional responses tend in general to be density-dependent. Holling (1959b) investigated the basic relationships involved in the functional response by setting up a simple laboratory model of predation. A blindfolded human “predator” detected and removed discs of sandpaper scattered on a table. The greater the initial number of discs, the more were picked up in a set period. But the rising curve of “number of discs picked up” became less steep as the initial number was increased (Fig. 17). This was

0

I

50

I

100

I

I50

I

200

1

250

No. discs per 9 sq. f t. FIG. 17. Functional response of a human subject searching for sandpaper discs on an area of 9 ftaby touch. (From Holling, 1959b.)(Averages752 S.E., 8 replicates.)

evidently a result of the time spent in handling and removing discs after having discovered them. The shape of the curve illustrates what is presumably an inherent tendency of the functional response ; although it increases as prey density rises (at least up to a certain level), on a proportional scale the increase is not as great as the rise in density. The relationships involved in this and similar experiments were represented in simple algebraic equations. When allowance was made for the time spent in dealing with discs that had been found, and only the active searching was considered, the rate of discovery was found, as expected, to be in rectilinear proportion to the initial number of discs present. It was the time spent in dealing with “finds” that caused the discovery curve to turn away from this density-independent course and become inversely density-related.

44

M. E . SOLOMON

c.

FUNCTIONAL RESPONSES OF SOME INSECT PARASITES

Holling (1959b)went on to examine some published examples of the functional responses of insect parasites to rises in host density. These were from experiments with various Hymenopterous parasites, namely Dahlbominus fuscipennis (Zett.) searching for cocoons of the pine sawfly in laboratory cages (Burnett, 1951), on a lawn (Burnett, 1954), and in a plot of woodland (Burnett, 1958d);Chelonw texanus Cress. searching for eggs of the flour moth Anagtgasta kuehniella (Zell.) (Ullyett, 1949a); Cryptus inornatus Pratt searching for cocoons of the beet webworm Loxostege sticticalis (L.) (Ullyett, 1949b); and Mormoniella vitripennis (Walk.) searching for puparia of the housefly (DeBach and Smith, 1941b); this last experiment, and Ullyett’s, were in the laboratory. I n every case, the graph of the number of affected hosts against host density, illustrated in Holling’s paper, was of the same general form as Fig. 17. I n a later paper, Holling (1961) cited two further examples of the same sort, one unpublished, the other from the work of Miller (1959, 1960) on two parasites of the spruce budworm, and one could now add the experimental findings of Chant (1961) with the phytophagous mite Tetranychus tehrius and a predatory mite (a Typhlodromus). Varley and Edwards (1957) reinterpreted the highly artificial experiments of DeBach and Smith (1941a,b, 1947) taking account of the way in which the parasite’s behaviour and physiology influence its response to availability of hosts. It can be inferred from their conclusions that under more natural conditions the functional response curve would probably have an early density-dependent phase before assuming the form shown in Holling’s diagram. On the whole, however, the adherence of insect parasites to the basic curve of functional response, concave below, is impressive. We may conclude that there is a general tendency for the number of hosts affected to increase in less than linear proportion to the total number or density of hosts. This is an inverse density relationship, one that cannot by itself lead to regulation. When host density increases, percentage parasitism will generally decline, until perhaps the functional response is followed by a delayed numerical response. The same principles apply to the action of predators: up to a certain limit, they can be expected to take more prey as prey density rises, nevertheless, other things being equal (including the numbers of predators), the percentage taken will be smaller. D.

MAMMALIAN PREDATORS OF THE P I N E SAWFLY

It might be expected that vertebrate predators would show more

complex responses to changes in the density of their prey than insects

ANALYSIS O F PROCESSES I N CONTROL O F INSECTS

46

do. The study of the functional and numerical responses of mammalian predators of the pine sawfly by Holling (1959a) gives a most instructive account of the responses of these predators. The investigation was carried out in a sand-plain area of S.W. Ontario. The plantations he used were 15-20 years old, with a closed canopy reducing ground vegetation to a trace. The sawflies spend the egg and feeding stages on the foliage, but in June the larvae drop from the trees and penetrate the pine needles on the forest floor, where they spin cocoons at the surface of the soil. Here they remain until the latter part of September, when most emerge as adults. A certain proportion, however, overwinter in cocoons, and emerge the following autumn. Only the cocooned stage is susceptible to attack by the small mammals. These are two species of shrew (a Sorex and a Blarina) and a deer-mouse (a Peromyscus). Each species makes a characteristically marked opening in the cocoon which provides a relatively permanent record of the act of predation and of the taxonomic identity of the predator. The effects of cocoon density on predation by Peromyscus were studied in two plots over three years, and the effects on that of Sorex and Blarina in one plot in one year. Owing to the treatment of parts of the area with a spray containing virus pathogenic to the sawfly larvae, a range of different levels of prey densities was available. The densities of Peromyscus and Blarina were estimated by trapping ; that of Sorex had to be calculated from the results of laboratory experiments on the rate of predation per individual.

1. Functional and Numerical Responses of Sawfly Predators Holling showed that the number of cocoons opened per mammal per day increased as the prey density rose, up to a certain value of prey density above which the rising curve of predation rapidly turned over and levelled off (Fig. 1 8 ~ )The . slope and the maximum was different for each species of predator, but the curves were of the same general form. Curves for the individual Peromyscus studied in the laboratory were also of this form. Turning to the numerical responses of the predators, Rolling’s observation period of four summers was too short for more than tentative conclusions. His graphs (Fig. 1 8 ~ suggest ) that Peromyscus and Sorex, but not Blarina, were more abundant at higher prey densities, up to a certain level of prey abundance that was different for the two. 2. Combined Raponses of Sawfly Predators When Holling compounded the estimated numerical and functional responses (Fig. 18c),the calculated combined response of each predator species was strongly density-dependent, i.e. percent predation greatly

M. E. S O L O M O N

46 U

Blortno

00 200 400 600 800 1000 No. of cocoons per ocre, thousands

400 600 800 1000 1200 1400 No. of cocoons per acre, thousands

200

200

600 1000 1400 1800 No. of cocoons per ocre, thousands

FIG.18. Functional (A) numerical (B) and combined ( c ) responses of small mammal predators to density of prey (sawfly cocoons). (After Holling, 1959e.)

increased as prey density rose - up to a certain level of prey density. Beyond this level, the ascending curve of percentage predation turned over and gradually declined as prey abundance continued to increase (inverse density relationship). The implications of this general result are that any of the predators might be able to regulate the abundance of the sawfly under certain environmental conditions, but if, in spite of predation, the prey exceeded a certain abundance, predation would no longer tend to regulate it. As Holling points out, several species of predators with different critical levels of prey density might be expected to provide a securer control of the sawfly than a single species.

ANALYSIS O F PROCESSES I N CONTROL OF INSECTS

47

3. Density Relations of the Responses of SawJEyPredators A point worth examining is the density-dependence or otherwise of the separate functional and combined responses. When a functional response is graphed as predation rate per individual predatoq per day , the response is densityagainst prey density, as in Fig. 1 8 ~ whether dependent or inversely density related depends on whether the rising curve is concave or convex. Only if the increase in individual predation rate is proportionately greater than the increase in prey density is the effect density-dependent ; for density-dependence is measured as an effect per individual of the prey population, such as percentage mor) tality. The rising curves drawn to the data by Holling (Fig. 1 8 ~ show slight density-dependence in the responses of Blarina and Sorex, and a more marked density-dependence in that of Peromyscus. Holling wrote : “Unfortunately the data for any one functional response curve are not complete enough to establish a sigmoid relation, but the six curves presented thus far and the several curves to be presented in the following section all suggest a point of inflection.” The numerical responses, as , be tested for density-dependencein the same way. graphedin Fig. 1 8 ~can There is a suggestion’of some density-dependence in the rising curve for Sorex, but the curve for Peromyscus indicates an inverse density relationship throughout (Idiffer from Holling’s statement on this point), and this is the general tenor also of the roughly horizontal curve for Blarina. I n the graph of combined responses (Fig. 18c) where the ordinate is in proportionate terms (percent predation), the criterion of densitydependence is simply a rising curve, which each of the three species achieves up to a certain level of prey density. It is appropriate here to make the general point that when both functional and numerical responses are involved, the combined response may be density-dependent even when the separate components are not. For example, if prey density rises from 1 000 to 3 000 the functional response of a predator may be an increase in the daily individual predation rate from 2 to 4, and its numerical response an increase in density from 10 to 20. Each of these responses amounts only to an inverse density relationship. Yet their combined effect, which is what matters most in practice, is an increase from 2 x 10 to 4 x 20, that is a-fold, compared with the %fold increase in prey density - a clearly density-dependent relationship.

E.

FUNCTIONAL RESPONSES O F OTHER VERTEBRATE PREDATORS O F INSECTS

As we have seen (Fig. 18), the functional responses of the mammalian predators of the pine sawfly were not quite of the simple type of the

48

M . E. S O L O M O N

laboratory model illustrated in Fig. 17. The curves tended to be sigmoid in the rising phase (Fig. 18). In a subsequent review, Holling (1961) expressed the opinion that “the functional response curves of vertebrate predators in general seem to have an S-shaped rise to a plateau”. I n support of this contention he quoted Leopold (1933) as suggesting that vertebrate predators attack scarce prey by chance but develop the ability to find a greater proportion where the prey become abundant. He referred to the experiments of De Ruiter (1952) in which tame jays (Garrulus glandarius) were shown to develop a lively interest in twig-like objects after stick-like caterpillars had been placed among sticks in which the birds had previously soon lost interest. He also referred to the proposal of L. Tinbergen that woodland birds preying on a variety of insects developed a “searching image” of any acceptable species that became abundant. An account of the work of Tinbergen and his colleagues on this subject has now appeared in English (Tinbergen, 1960; Mook et al., 1960) and deserves close study by ecologists. The following is part of Tinbergen’s summary. “When a new species appears in the environment, its risk is low at first, and then increases suddenly. A detailed examination of this phenomenon leads to the hypothesis that tits (Parus spp.) when searching for prey concentrate on one or a few species at a time, and that, by a kind of learning process, they adopt ‘specificsearching images’ for these species. The main factors that determine whether or not this process will take place are discussed. . . . Among these, density of the prey species is important. It was found that the relation between the density of a prey species and its percentage in the food cannot be explained from probability of encounters alone. At low densities, consumption is lower than would be expected on that basis. At moderate densities it is unexpectedly high, and at high densities it falls again below expectation. The difference in risk between low and high densities is explained by assuming that the birds do not adopt a specific searching image for a species of prey that is scarce. The decrease in risk at high densities is supposed to be due to the fact that, in order to obtain a sufficiently varied diet, the birds stop using a searching image when the species concerned forms more than a certain critical percentage in the total food.” Another study of the density-relationships between birds and their insect prey is summarized by Mook (1963). He gives curves for the responses of the bay-breasted warbler to changes in the density of larvae of the spruce budworm. The curve for the functional response is identical in general form to the sigmoid plateau curves for Holling’s , the concave-upward aspect of the small mammals (Fig. 1 8 ~ ) although slope is again not firmly established. The curve of the numerical re-

ANALYSIS OF PROCESSES I N CONTROL OF INSECTS

49

sponse is based only on years when the budworm numbers were higher than in the previous year; as far as it goes, it rises at a decreasing rate, i.e. less than density-dependently. The curve of combined response is similar to Holling’s curve for BZurina in Fig. 1 8 ~ .

F.

A NOTE ON TERMS

The terms functional and numerical response were originally intended to apply to the responses of natural enemies to changes in the density of their prey. I n developing his analysis of predation, Holling (1961) introduces an extension of the terminology, namely, functional response of predators to their own density. This is simply a new name for the effects of competition and interference between predators, and group stimulation, as measured by the effects upon their prey. (Holling reviews a number of studies on this aspect of predation.) I regret only that this may make the terms more cumbersome, since the user, to make his meaning clear, would often have to write, e.g. ‘(numerical response to prey density’’ or “functional response to predator density”. I am more dubious about the distinction made by Holling (1959a, 1961) between “direct” numerical responses to prey density, in which the numbers of predators increase as prey numbers rise, and “inverse” responses, in which predator numbers decline as prey numbers increase. His illustrations of direct responses show predator numbers rising, but sometimes not at as great a proportionate rate as prey numbers. I n the latter cases we have an inverse density relationship; if it is at the same time regarded as a direct numerical response, there may be some confusion. Moreover, the need for the term “inverse numerical response” to describe a decline in predator numbers in response to a rise in prey numbers must be rather slight, since although predators often do decline at a time when prey density is rising, such a fall in predators must very seldom be a response to the increase in prey.

VI. QUESTIONSBEARING O N A THEORY OF INTERACTION

PARASITE-HOST

A . IS PARASITE FECUNDITY NOT A LIMITING FACTOR? Nicholson (1933) and Nicholson and Bailey (1935) put forward a model of parasite-host interaction in which one of the assumptions was that the capacity of parasites to attack hosts is great enough to match the numbers of hosts that they can find, so that it is the supply of hosts and their ability to find them that limits the parasites’ increase, not their own capacity to deal with hosts found. The model generates oscillations of the host and parasite populations, the oscillations becoming increasingly violent and ending in extinction of the populations.

60

M . E. SOLOMON

Among the main lines of thought surrounding this model, one has been to challenge or investigate the assumptions, including the assumption of insatiable parasites. While it is obvious that parasites do not have unlimited fecundity, it is not necessary, for the purposes of the model, to assume that they do. One is required to assume only that the limit of their fecundity is not reached, and that fecundity does not decline at the levels of host and parasite abundance involved (cf, Varley, 1947). This assumption receives no support from the above-mentioned experimental data on the functional response of insect parasites reviewed by Holling (1959b).The diagrams, similar in form to Fig. 17, show that the number of hosts parasitized rises proportionately less than host density does, from the lowest levels of density studied. While these results are not simply an index of parasite fecundity, they do remind us that parasites, besides seeking and finding hosts, spend some time in dealing with the discovered hosts and depositing their eggs; the greater the density of hosts, the greater is the proportion of the parasites’ time that is spent in this way instead of seeking more hosts (cf. Tinbergen and Klomp, 1960). They also show that, however high is the hosts’ density, the parasites are unable to attack more than a definite number of them. (To explain how such parasites could ever, without the help of additional factors, put a stop to the increase of their hosts, we must invoke also the numerical response, but this is a separate matter from the assumption under discussion.) One recent example shows an opposite effect to that of parasite satiation. I n experiments.with the whitefly Trialeurodes vaporariorum (Westw.) and its Hymenopterous parasite Encarsia formosa Gahan, Burnett (1958b)c)demonstrated that with low but increasing densities of the host, the percentage of hosts found also increased (with further host increase this percentage became approximately constant). The reduced percentage parasitism at low host density evidently had something to do with the parasites’ searching eaciency (it did not occur when the experimental area was reduced). The maintenance of the same percentage parasitization at higher host densities showed that fecundity of the parasites was not a limiting factor within the range of densities considered. Burnett (1959), in the course of a very useful review of parasite-host experiments, pointed out that we do not know how much departure from the initial assumptions is required to upset the conclusions from the Nicholson-Bailey model. This is a question that could well be investigated theoretically. Nicholson (1933, 1954b) and Nicholson and Bailey (1935) have elaborated the original model rather than tested the effects of relaxing the basic assumptions. However, in the present instance it seems that any falling off of effective parasite

ANALYSIS O F PROCESSES I N CONTROL O F INSECTS 51 fecundity at the higher host densities would inevitably lead to greater increases of the host (of. Tinbergen and Klomp, 1960).

B.

DO PARASITE-HOST OSCILLATIONS T E N D TO INCREASE I N AMPLITUDE?

Experimental studies have provided a little support for the conclusion of Nicholson and Bailey that parasite-host interactions tend towards increasingly violent oscillations. In the early experiments of Gause (1934)with the predatory mite Cheyletus eruditus (Schr.) and an Acarid mite as prey, the predators killed all the prey and then died out. This has happened in more recent laboratory or greenhouse experiments with mite predators and prey (e.g. Bravenboer and Dosse, 1962, with mites on peach in a greenhouse). Burnett (1958a), in experiments with Trileurodes and Encarsia in a small cage, produced two cycles with’an increase in amplitude in the second. However, this experiment was so highly manipulated (somewhat in the style of the DeBach and Smith experiments), that it seems closer to the theoretical model than to any set of natural conditions. I n a brief account of experiments with the same animals on tomato plants in a greenhouse, Burnett (1959)reported that “in general, the growth forms of the host and parasite populations exhibited fluctuations of increasing amplitude”. However, the experiments stopped after three oscillations, and these did not always increase in amplitude. At least one experimenter (Utida, 1953, 1955a,b,c) has reported parasite-host oscillations that did not increase in size. He used pulse beetles Callosobruchus spp. and Hymenopterous parasites in small dishes of bean seeds, in which the host increase was restricted by crowding; Burnett (1959) suggested there must also have been some form of protection of the hosts. Burnett (1959)has.expressed the opinion that in some of the parasitehost or predator-prey experiments in the laboratory too little space has been allowed for the populations to become properly established. Nicholson (1954b) expressed a similar opinion about the well-known experiments of Gause (1934) with Paramecium and Didinium. Varley and Edwards (1957) raised a series of objections to the procedure used in the experiments of DeBach and Smith (1941a,b, 1947) on the housefly and Mormoniella. There is no doubt that the conditions imposed by experimental design need scrutinizing as closely as the assumptions of a, theoretical model.

C. HOW MIGHT EXPANDING OSCILLATIONS B E DAMPED? Another approach to the Nicholson-Bailey model and its prediction of increasingly violent oscillations has been to assume that it may apply

52

M. E. SOLOMON

in nature if there are also mechanisms at work that damp the oscillations, and to suggest or investigate what these mechanisms may be. There has been no lack of suggestions. Nicholson (1933) himself argued that host populations become fragmented into sub-populations some of which, at any given time, have not yet been discovered by the parasite, while others are in different phases of interaction with it. Although increasing oscillations lead to the extinction of some of the subpopulations, others will survive for a time, and new ones will be formed. Some ecologists have protested that they do not see this pattern in nature (e.g. Tinbergen and Klomp, 1960). Evidence that it could have the kind of effect Nicholson sought is provided by the experiments of Huffaker (1958),who set up an experimental equivalent of Nicholson’s idea, using a phytophagous and a predatory mite on groups of oranges in trays. Nicholson (1933) also argued that oscillations could be damped if the parasite had one or more other hosts which were not regulated by it. This possibility seems to have been generally neglected. Tinbergen and Klomp (1960)argue against it on general grounds. Varley (1947) claimed that “it can be shown with Nicholson and Bailey’s theory that if a proportion of hosts is not available to parasitism, oscillations will be damped instead of increasing in amplitude”. Nicholson (1954b) replied that this was so only with low powers of host increase combined with the protection of the greater part of the hosts that must survive to ensure population maintenance. Retreating slightly, Varley and Gradwell (1958)listed host-protection as having an effect “in the direction which leads to quenching”. Recently, Bailey et al. (1962) returned to the theoretical model to examine the effects “when some host individuals are more difficult to find than others”. They deduced that “systems of damped or growing oscillations are produced according to the circumstances, which are defined”. The set of circumstances required to produce damped oscillations and a stable system was very restricted. DeBach and Smith (1941a) pointed out that oscillations could be damped if the reproductive rate of the host were strongly densitydependent, but considered such a thing likely only at very high densities, a view endorsed by Tinbergen and Klomp (1960).Possibly the amplitude of fluctuation in Utida’s experiments with pulse beetles and their parasites was prevented from increasing by the effect of crowding in restricting the increase of the host. Varley and Gradwell (loc. cit.), continuing their list of processes tending t o damp oscillations, mentioned the protection of some hosts by the odour of parasites that had previously examined and rejected them ; imperfect synchronization of the susceptible host stage with the peak of parasite activity; hindering of the parasites by unfavourable

53

ANALYSIS OF PROCESSES IN CONTROL O F INSECTS

weather at the time the hosts are passing through the susceptible stage; and differences in the dispersion of hosts and parasites, so that some hosts occur in areas with few parasites. L. Tinbergen (see Klomp, 1958) claimed that in theory an additional density-dependent mortality of the host could be effective in damping oscillations if it occurred at medium levels of density; he went on to relate this idea to his studies of mortality in the caterpillars of woodland insects due to predation by tits (Parus spp.). Tinbergen and Klomp (1960) combined a Nicholson parasite-host model with empirical data on the insects, on their parasites, and on the mortality due to bird predation; this last was density-dependent so long as host density did not rise too high. They concluded; “When the birds eliminate a considerable part of the population at intermediate densities of the prey (more than 25%) this density dependent predation has a damping effect on the oscillations of a host-parasite model. . . . Consequently, under certain conditions, the system host-predator-parasite is self-regulating.” Varley and Gradwell (1963) illustrated a simple model showing how pupal predation might restrict the population oscillations of the winter moth and a parasite. They used the Nicholson-Bailey model and substituted constants based on their field observations. These combinations of theory with empirical data are a welcome and valuable development. When so many “damping” processes are suggested, one may well doubt whether any expression of a tendency towards increasingparasitehost oscillations under natural conditions would survive long enough to be detectable. For that makter, even more “damping” agents could be listed, for almost any density-dependent processes acting on the hosts would tend to have this effect. We are concerned here with the limiting aspects of natural control at high densities, and with the relaxation of controls, and the operation of protective influences, at low densities (Solomon, 1949, section on Phases of Control). The question, whether these processes would be adequate to damp increasing oscillations before they reached dangerously high and low extremes, cannot be settled in general, but must be determined according to the circumstances of particular cases. It is obvious that hosts do at times overcome any controlling influence their parasites may have upon them, and reach high levels of density. But unless there is a history of systematic oscillation with the population of a specific parasite, there is no reason to expect that the parasites will overtake the host while it is abundant, and cause it to “crash”. Often some other influence reduces the host to low density, a t which the effect of parasitism again becomes prominent. This is quite a different matter from the damping of a host-parasite oscillation. C

C.E.R.

54

M. 1. SOLOMON

ACKNOWLED QMENTS For permission to reproduce Figs. 1, 4 to 8, 12, 13, 17 and 18 I am indebted t o the authors mentioned in the captions, t o Blackwell Scientific Publications Ltd. (Fig. 7), and to the editors or publishers of the following journals : Memoirs of the Entomological Society of Canada (Figs. 1, 4 and 8), The Canadian Entomologist (Figs. 17 and 18), Ecology (Fig. 12), Archives Nderlandaises de Zoologie (Fig. 5 ) , Nature (Fig. 6), and Proceedings of the Ceylon Associationfor the Advancement of Science (Fig. 13). REFERENCES Andrewartha, H. G. (1963). Ecology 44, 218-220. Density-dependence in the Australian thrips. Andrewartha, H. G. and Birch, L. C. (1954). “The Distribution and Abundance of Animals”, 782 pp. Illinois: Univ. of Chicago Press. Bailey, V. A., Nicholson, A. J. and Williams, E. J. (1962). J . theor. Bio.?.3, 1-18. Interaction between hosts and parasites when some host individuals are more M c u l t to find than others. Banks, C. J. (1962). Ann. appl. Bioz. 50, 669-679. Effects of the ant L&ua niger (L.) on insects preying on small populations of Aphisfabae Scop. on bean plants. Bess, H. A. (1945). Ann. ent. SOC.Amer. 38, 472-481. A measure of the influence of natural mortality factors on insect survival. Bravenboer, L. and Dosse, G. (1962). Entomologia ezp. app.?.5, 291-304. Phytoseiulua riegeli Dome als Priidator einiger Schadmilben au8 der Tetranychua urticae-Gruppe. Burnett, T. (1951). Amer. Nat. 85, 337-352. Effects of temperature and host density on the rate of increase of an insect parasite. Burnett, T. (1964). Physiol. Zool. 27,239-248. Influences of natural temperatures and controlled host densities on oviposition of an insect parasite. Burnett, T. (19584. Proc. 10th Int. Congr. Entom. (Montreal, 1956) 2, 679-686. A model of host-parasite interaction. Burnett, T. (1958b). Canad. Ent. 90, 179-191. Effect of host distribution on the reproduction of Encarsiaformosa Gahan (Hymenoptera: Chalcidoidea). Burnett, T. ( 1 9 5 8 ~ )C. a d . Ent. 90, 225-229. Effect of area of search on reproduction of Encarsiaformosa Gahan (Hymenoptera: Chalcidoidea). Burnett, T. (1958d).Canad. Ent. 90, 279-283. Dispersal of an insect parmite over a small plot. Burnett, T. (1959). Ann. Rev. Ent. 4, 235-250. Experimental host-parasite populations. Chant, D. A. (1961).Canad. J . Zoo.?.39,311-315. The effect of prey density on prey consumption and oviposition by adults of Typhlodromua (T.) occidentalia Nesbitt (Acarina:Phytoseiidae)inthe laboratory. Chitty, D. (1960). Canad. J . Zool. 38, 99-113. Population processes in the vole and their relevance to general theory. Davidson, J. and Andrewartha, H. G. (194%). J . a n h . Ecol. 17,193-199. Annual trends in a natural population of Thrips imaginis (Thysanoptera). Davidson, J. and Andrewartha, H. G. (1948b). J . anim. Ecol. 17, 200-222. The influence of rainfall, evaporation and atmospheric temperature on fluctuations in the size of a natural population of Thrips imagini.9 (Thysanoptera).

ANALYSIS O F PROCESSES I N CONTROL O F INSECTS

55

DeBach, P. (1955). J . econ. Ent. 48, 584-588. Validity of the insecticidal check method as a measure of the effectiveness of natural enemies of Diaspine scale insects. DeBach, P. (1958). J . econ. Ent. 51, 474-484. The role of weather and entomophagous species in the natural control of insect populations. DeBach, P. and Smith, H. S. (1941a). Ecology 22, 363-369. Are population oscillations inherent in the host-parasite relation? DeBach, P. and Smith, H. S. (1941b). J . econ. Ent. 34, 741-745. The effect of host density on the rate of reproduction of entomophagousparasites. DeBach, P. and Smith, H. S. (1947). Ecology 28, 290-298. Effects of parasite population density on rate of change of host and parasite populations. DeBach, P., Dietrick, E. J. and Fleschner, C. A. (1949).J . econ. Ent. 42, 546-547. A new technique for evaluating the efficiency of entomophagous insects in the field. DeBach, P., Fleschner, C. A. and Dietrick, E. J. (1950). J . econ. Ent. 43, 807-819. Studies of the efficacy of natural enemies of citrus red mite in southern California. DeBach, P., Fleschner, C. A. and Dietrick, E. J. (1951). J . econ.Ent. 44, 763-766. A biological check method for evaluating the effectiveness of entomophagous insects. DeRuiter, L. (1952). Behaviour 4,222-232. (From Holling, 1961.) Elton, C. E. (1958). “The Ecology of Invasions by Animals and Plants”, 181 pp. London : Methuen. Franz, J. (1949). 2.angew. Ent. 31, 228-260. ffber die genetbchen Grundiagen des Zusammenbruches einer Massenvermehrung am inneren Ursachen. Franz, J. M. (1962). Verh. X I Intern. Kongr. Entom. (wien 1960) 2, 670-674. Definitions in biological control. Gause, G. F. (1934). “The Struggle for Existence”, pp. 163. Baltimore: Williams and Wilkins. Glasgow, J. P. and Welch, J. R. (1962). Bull. ent. Res. 53, 129-137. Long-term fluctuations in numbers of the tsetse fly Glossina swynnertoni Austen. G w , D. L.(1960). A . Rev. Ent. 5, 279-300. The biological background of locust control. G m , D. L. and Symmons, P. M. (1959). Nature, Lon& 184, 1425. Forecasting locust outbreaks. Hairston, N. G. (1957). In Reynoldson, T. B. (1957). Cold Spring Harb. Symp. quant. Biol. 22, 313-327. Discussion. Holling, C. S. (1959a). Canad. Ent. 91, 293-320. The components of predation as revealed by a study of small-mammal predation of the European pine sawfly. Rolling, C. S. (195913). Canad. Ent. 91, 385-398. some characteristics of simple types of predation and parasitism. Holling, C. S. (1961).A. Rev. Ent. 6, 163-182. Principles of insect predation. Holling, C. S. (1962). Mem. ent. Soc. can. 32, 22-32. An experimental component analysis of population processes. Huffaker, C. B. (1958). Hilgardia 27,343-383. Experimental studies on predation: dispersion factors and predator-prey oscillations. Hufhker, C. B. and Spitzer, C. H., Jr. (1951). J . econ. Ent. 44, 519-522. Data on the natural control of the cyclamen m i t e on strawberries. momp, H. (1958). Arch. n6erZ. 2001.13, 134-145. On the theories of host-parasite interactions.

66

M. E. SOLOMON

Klomp, H. (1962). Arch. nderl. Zool. 15, 68-109. The influence of climate and weather on the mean density level, the fluctuations and the regulation of animal populations. Kuenen, D. J. (1958). Arch. rder.!. Zool. 13, 335-341. Some sources of misunderstanding in the theories of regulation of animal numbers. Lack, D. (1954). “The Natural Regulation of Animal Numbers”, 343 pp. London: Oxford University Press. Leopold, A. (1933). “Game Management.” New York: Charles Scribner’s Sons. MacArthur, R. (1955). Ecology 36, 533-536. Fluctuations of animal populations, and a measure of community stability. Miller, C. A. (1959). Canad. Ent. 91, 475-477. The interaction of the Spruce Budworm, Choristoneura fumiferana (Clem.), and the parasite Apanteles fumiferame Vier. Miller, C. A. (1960). Canad. Ent. 92, 839-850. The interaction of the Spruce Budworm, Choristoneura fumifera?.a (Clem.) and the parasite cflypta fumiferanae (Vier.). Miller, C. A. (1963a).I n Morris, ed. (1963),pp. 75-87. Miller, C. A. (1963b).I n Morris, ed. (1963),pp. 228-244. Mook, L. J. (1963).InMorris, ed. (1963),pp. 268-271. Mook, J. H., Mook, L. J. and Heikens, H. S. (1960).Arch. nderl. 2002.13, 448465. Further evidence for the role of “searching images” in the hunting behaviour of titmice. Morris, R. F. (1957). Canad. Ent. 89, 49-69. The interpretation of mortality data in studies on population dynamics. Morris, R. F. (1959). Ecology 40, 580-588. Single-factor analysis in population dynamics. Morris, R. F., ed. (1963).Mem. ent. SOC. Can. 31,332 pp. The dynamics of epidemic spruce budworm populations. Morris, R. F. (1963a).InMorris, ed. (1963),pp. 116-129. Morris, R. F. (1963b). Mem. ent. SOC. Can. 32, 16-21. Predictive population equations based on key factors. Mott, D. G. (1963).InMorris, ed. (1963),pp. 42-52. Nicholson,A. J. (1933).J . anim. Ecol.2,132-178.Thebalanceofanimalpopulations. Nicholson, A. J. (1954a). Aust. J . ZooZ. 2, 1-8. Compensatory reactions of populations to stresses, and their evolutionary sign5cance. Nicholson, A. J. (195413). Awt. J . Zool. 2, 9-65. An outline of the dynamics of animal populations. Nicholson, A. J. (1957). I n Reynoldson, T. B. (1957). Cold Sp&g Harb. Symp. quant. Biol. 22,313-327. Discussion. Nicholson, A. J. (1958).Ann. Rev. Ent. 3,107-136. Dynamics of insect populations. Nicholson, A. J. and Bailey, V. A. (1935). Proc. zool. SOC. Lond., 551-598. T h e balance of animal populations, Part I. Pimentel, D. (1961a). Ann. ent. SOC. A m r . 54, 76-86. Species diversity and insect population outbreaks. Pimentel, D. (1961b). Amer. Nat. 95, 65-79. Animal population regulation by the genetic feed-back mechanism. Richards, 0. W. (1963). Proc. 16th Int. Congr. Zool. (Washington1963) 3, 353-356. Some factors controlling insect populations living on Scotch broom. Richards, 0. W. and Waloff, N. (1961). Phil. Trans. 244, 205-257. A study of a natural population of Phytodecta olivucea (Forster) (Coleoptma, Chrysomeloidea).

ANALYSIS O F PROCESSES I N CONTROL O F INSECTS

57

Schwerdtfeger, F. (1935). Z . For8t- u. Jagdw. 67, 449-482, 513-540. Studien iiber den Massenwechsel einiger Fortschiidlinge, 11. Schwerdtfeger, F. (1941). 2. angew. Ent. 28, 254-303. Uber die Ursachen des Massenwechsels der Insekten. Smith, F. E. (1961). Ecology 42, 403-407. Density dependence in the Australian thrips. Smith, F. E. (1963). Ecology 44,220. Density-dependence. Solomon, M. E. (1949). J. anirn. Ecol. 18, 1-35. The natural control of animal populations. Solomon, M. E. (1957). A . Rev. Ent. 2, 121-142. Dynamics of insect populations. Solomon, M. E. (1962a). Verh. 11.int. Kongr. Ent. ( Wien 1960) 2, 126-130. Status of the idea that weather can control insect populations. Solomon, M. E. (1962b). In “The Exploitation of Natural Animal Populations” (E. D. Le Cren and M. W. Holdgate, eda.), pp. 373-375. Oxford: Blackwell. Stern, V. M., Smith, R. F., van den Bosch, R. and Hagen, K. S. (1959). Hilgardia 29,81-101. The integrated control concept. Thompson, W. R. (1928). Par&tology 20, 90-112. A contribution to the study of biological control and parasitic introduction in continental areas. Thompson, W. R. (1939). Parasitology 31, 299-388. Biological control and the theories of the interactions ofpopulations. Thompson, W. R. (1955). C a d . Ent. 87, 26&275. Mortality factors acting in a sequence. Tinbergen, L. (1960). Arch. nkerl. 2002.13, 265-343. The natural control of insects in pinewoods I. Factors influencing the intensity of predation by songbirds. With Appendix by L. de Ruiter. Tinbergen, L. and Klomp, H. (1960). Arch. nkerl. Zool. IS, 344-379. The natural control of insects in pinewoods 11. Conditions for damping of Nicholson oscillations in parasite-host systems. Ullyett, G. C. (1949a). Canad. Ent. 81,25-44. Distribution of progeny by Chelonus tezanus Cress. (Hymenoptera: Braconidae). Ullyett, G. C. (1949b) Canad. Ent. 81, 285-299. Distribution of progeny by Cryptus inornatua Pratt (Hymenoptera:Ichneumonidae). Utida, S. (1953). Ecology 34, 301-307. Interspecific competition between two species of bean weevil. Utida, S. (1955a). Mem. Coll. Agric. Kyoto 71, 1-34. Population fluctuation in the system of host-parasite interaction. Utida, S. (195513).Ecology 36,202-206. Fluctuations in the interacting populations of host and parasite in relation to the biotic potential of the host. Oyo-kontyu 11, 43-48. Population fluctuation in the system of Utida, S. (1955~). interaction between a host and its two species of parasite. Varley, G. C. (1947). J . anim. Ecol. 16, 139-187. The natural control of population balance in the knapweed gall-fly (Urophorajaceana). Varley, G. C. (1949). J . anim. Ecol. 18, 117-122. Population changes in German forest pests (Review). Varley, G. C. (1953). Tram. 9th Intern. Congr. Entomol. (Amterdam 1951) 2, 210-214. Ecological aspects of population regulation. Varley, G. C. (1963). Proc. R. ent. SOC.Lond. (Ser.C) 27, 52-57. The interpretation of change and stability in insect populations. Varley, G. C. and Edwards, R. L. (1957). J . anim. Ecol. 26, 471-477. The bearing of parasitic behaviour on the dynamics of insect host and parasite populations.

68

M. E. SOLOMON

Vmley, G. C. and Gradwell, G. R. (1958). Proc. 10th Intern. Congr. Entornol. (Montreal 1956) 2, 619-624. Balance in insect populations. Varley, G. C. and Gradwell, G. R. (1960).J . unim. Ecol. 29, 399-401. Key factors in population studies. Varley, G. C. and Gradwell, G. R. (1963). Proc. 18th Ann. Session Ceylon Asm. Adv. Sci., Part II, 142-156. The interpretation of insect population changes. VoQte,A. D. (1946). Arch. nderl. Zool. 7, 435-470. Regulation of the density of the insect-populationsin virgin forests and cultivated woods. Wallace, M. M. H. (1957).Nature, Lo&. 180, 388-390. Field evidence of densitygoverningreaction in Sminthurus viridis (L.). Wallace, M. M. H. (1962). In (Aust.) Rep. Div. Ent. C.S.I.R.O., 1961-62, p. 38. Ecology of the lucerne flea Sminthurus wiridis. Watt, K. E. F. (1955). Eool. Monogr. 24, 269-290. Studies on population productivity, I. Three approaches to the optimum yield problemin populations of Tribolium confuaum. Watt, K. E. F. (1961). Canad. Ent. 93, Suppl. 19, 62 pp. Mathematical models for use in insect pest control. Watt, K. E. F. (1962).A . Rev. Ent. 7,243-260. Use of mathematics in population ecology. Wynne-Edwards, V. C. (1962). “Animal Dispersion in Relation to Social Behaviour”, 653 pp. Edinburgh and London: Oliver and Boyd.

The Use of Statistics in Phytosociology

. .

. .

J M LAMBERT and M B DALE

Botany Department. University of Southampton. England

.

I Introduction .......................................................... A The Aim of the Contribution ......................................... B DefinitionofTerms ................................................. C. The Function of Statistics in Phytosociology ........................... I1. The Nature of Phytosociological Data .................................... A . The Nature of Vegetation ............................................ B . The Nature of the Variables Concerned ................................ C The Nature of the Sites ............................................. D The Nature of the Measures .......................................... I11 Methods of Analysis ................................................... A The Statistical Properties of Phytosociological Data ..................... B Ordination and Classification......................................... C. MethodsofClassification ............................................ D Inverse Analysis ................................................... E . Nodal Analysis..................................................... F. Relationship with the Environment .................................. IV Comparison of Phytosociological Concepts ................................ A TheBasic Assumptions ............................................. B . TheBctsic Approach ................................................ C. The Basic "Vegetation-Unit" ........................................ V The Future of Statistics in Phytosociology ................................ Acknowledgments ........................................................ References ................................................................

. .

. . . . . .

.

.

.

59 59

60 62 64 64 66 67 68 70

70 72 77 81 83 85 87 87 90 93 95 97 98

I . INTRODUCTION 8.

T H E A I M O F THE CONTRIBUTION

The title of this paper may at first sight seem misleading. We shall not concern ourselves with an enumeration of the various applications of mathematics which have progressively invaded the study of vegetation within the last 30 years; nor shall we consider every facet of ecological work which could conceivably be included under the heading of phytosociology in its widest sense; and. even within the restricted field with which we shall be dealing. we shall not attempt a detailed comparison of all the various techniques and parameters which have been proposed a t one time or another for the assessment of a given situation . Certain of these aspects have already been covered excellently in recent publications by Goodall (1952. 1954a. 1962) and Greig-Smith

60

J . M. LAMBERT AND M. B . D A L E

(1964), whose latest contributions between them provide a comprehensive bibliography of the relevant literature t o date. With such modern surveys available, we see no point in repetition of similar material here. Instead, we shall attempt to examine the present situation more particularly in terms of current concepts which may be involved in efforts to put the study of vegetation oq an increasingly objective basis; and this in turn will require the close examination of certain basic assumptions, with regard both to the nature of the data to be analysed and to the statistical methods which can be employed.

B.

D E F I N I T I O N OF TERMS

1. The DeJinition of “Statistics” The modern usage of the term “statistics” has tended to restrict its meaning to probabilistic studies, i.e. to studies of problems involving, a degree of uncertainty and, more specifically, to those involving estimation of parameters and testing of hypotheses previously erected. However, the word has an older -though now dubiously respectable meaning, derived from its original use for the description of “state” data (i.e. political and economic data) and thereby including nonprobabilistic methods for data simplification and generation of hypotheses. Although both types of method have a part to play in ecological work, the non-probabilistic techniques are in fact more generally relevant to the kind of empirical situation usually met with in the field; and although this aspect has been largely rejected from modern statistical terminology, there is as yet no convenient alternative word available t o cover it. I n the present contribution, therefore, we shall deliberately revert to the older and wider definition of statistics covering both aspects, rather than use still more ambiguousterms like “numerical” and “quantitative” methods. This wider definition thus allows us to include two important primary functions beyond those appertaining to more orthodox techniques. The first concerns the reduction, subject to some efficiency criterion, of a large and complex mass of data into a more accessible and convenient form; such simplification aims only at allowing the investigator to reduce the amount of information to be handled for future examination. The second - and generally more useful - function is that of hypothesis-generation. Hypotheses normally postulate underlying causal factors which may be thought of as having generated the original data in all their complexity, and there is no a priori reason why they should not be developed from the original data. I n complex situations, there may be so many variables that the whole pattern cannot be intuitively grasped; if, however, the data can be so simplified that their internal

THE USE O F STATISTICS IN PHYTOSOCIOLOQY

61

interrelationships can be economically displayed, the investigator may with greater confidence suggest an hypothesis concerning the causal factors involved. Such an hypothesis may be regarded as a model system which would serve to generate the system under examination. However, if hypothesis-generation is the ultimate aim, it may be more useful to begin with an a priori hypothesis as to the nature of the model; for example, it may be postulated that a given number of causal factors is involved. It may then be possible, first, to test this primary hypothesis ;and, second, to simplify the data in such a way as to display the simplest set of interrelationships which would be consistent with the postulated degree of complexity of the model. The two approaches are more different than may appear at first sight, and require quite different statistical techniques. In the one case the data are unimpaired, the manipulation is directed solely to simplification, and the model arises subsequent to all numerical manipulation. I n the second case, the nature of the model required itself determines the nature of the manipulation. We find it convenient to distinguish these two different activities as “primary” and “derivative” model-making. Both types of method may be used to generate hypotheses; but the mathematical routes by which the end-products are sought are different, and the distinction cannot be avoided. It is, moreover, very important to realize that any system of organizing information to generate rather than to test hypotheses is not self-sufficientin a scientific context : some further observations outside the immediate system are required before such hypotheses can be accepted or rejected.

2. The DeJnition of “Phytosociology” Whereas plant ecology is usually defined as the study of plants in relation to their habitat, phytosociology is based on a study of the interrelationships of the plants themselves within an area of vegetation. Taken at face value, the term “phytosociology” is self-explanatory, in that it refers merely to the study of plants as gregarious entities. I n this sense, it could equally well apply to the study of the interrelationships of the vegetative offshoots of a single parent, or to a set of individuals of a single species growing together in the field. More usually, however, the term is taken to mean the study of sets of species forming communities under natural or semi-natural conditions. The use of the word “community” in turn imparts the concept of separate units of vegetation, comprised of sets of plants with at least a degree of internal organization in their interrelationships. Moreover, the current identification of phytosociological studies with certain schools of thought has tended to restrict the meaning of the term still further, so that it is now CZ

C.E.R.

62

J . M . LAMBERT A N D M. B . DALE

frequently equated solely with attempts to classify such communities into a generalized scheme. For the purpose of this communication, we shall again revert to the wider meaning of the term; for reasons which will later be apparent (see IIA), we prefer this to the use of “vegetation studies” to cover the field with which we shall be dealing. Used in this unrestricted sense, phytosociology now carries no a prior; assumptions as to the nature of the interrelationships expected. Moreover, it allows us to include other analytical as well as purely classificatory techniques, without any primary assumption that the latter are those required.

c. THE

FUNCTION OF STATISTICS I N PHYTOSOCIOLOGY

A criticism which has often been levelled at phytosociology is that

it is far too frequently concerned with a mere reordering of the data into a more simplified form, irrespective of whether the results are interpretable in a wider context. If such an ordering is all that is required, i.e. to display the information or docket it for future reference, then the only advantage of using statistical techniques is to enable the phytosociologist to reject peripheral information on a uniform but predetermined basis. However, w_herephytosociological studies are regarded primarily as a means of entry into more complex vegetational relationships involving the habitat as well, the hypothesis-generating function of certain statistical methods becomes important. For instance, a statistical system constructed on a plant/plant basis can provide some insight into the internal structure of the population in terms of associations between the plants concerned; and if the assumption is made that some relationship exists between the distribution of the plants and features of the physical environment, the structure revealed may then be used to postulate the nature of possible plantihabitat relationships. The prime criterion for the use of statistics in any branch of study is often believed to be that it makes the study more “objective” and hence scientifically more respectable. It is true that the use of such methods eliminates personal bias and produces reproducible results for a given set of data analysed by a given method. However, in very complex situations like those usually encountered in the field, the subjective element must enter, in one guise or another, at almost every stage. For instance, the nature of the data to be collected, the form of the analysis, and the exact parameters to be used, are all subjective decisions frequently made either on past experience or with some foreknowledge of the type of phenomena expected. Where practical considerations are overriding, however, the study may be objective in the sense that decisions are often based on factors largely irrelevant to the actual system under study. This introduces the concept of efficiency,

63 which we ourselves believe to be the more important reason for the use of statistical techniques in much phytosociological and ecological work. Some difficulties arise in the definition of “efficiency” in the sense we intend it here. In the modern, restricted, statistical literature, the efficiency of a statistic is related to the precision of an estimate from a given size of sample in an unknown population. I n the present context, however, this definition is largely irrelevant except in so far as similar underlying concepts are involved. For present purposes, we shall define the term as the optimization of the amount of information extracted from a given situation for a given quantity of work; and this includes the wider concepts of time and effort spent in data-collecting, in processing the data, and in interpreting the results. As regards efficiency of analytical methods, the most relevant property of most phytosociological (and ecological) data is its complexity. With modern computer facilities, however, a number of quite elaborate nonprobabilistic techniques appropriate to complex data are now coming within reach; and the saving of personal time and effort by using mechanical methods of analysis is often so considerable that no modern ecologist can afford to ignore the possibilities which exist. A t the same time, computer-time costs money, and any statistical technique proposed should be rigorously scrutinized on the efficiency criterion suggested before it is adopted. For instance, it is often necessary tobalance the efficiency of a given parameter for a given estimation against a disproportionate increase in total computing time ; and methods which give the maximum information are not necessarily to be preferred against more approximate methods which are faster to operate. Similarly, the method of data-collection should be pruned as far as possible consistent with an acceptable reduction in information-content of the samples to be analysed: it is often more efficient to work with a larger number of samples of low individual information-content than with a smaller number of much more elaborate and time-consuming records. In any use of statistical techniques as investigational tools, therefore, there are three basic requirements to be met. First, the tool must be appropriate to the material to be worked, i.e. the nature and form of the data, and the scale of the problem, must be considered in selecting the most efficient method; secondly, the tool itself must be of sound construction (i.e. the underlying mathematics must be sound), and any area of weakness must be clearly understood so that an undue strain is not imposed; and thirdly, the work to be done must itself be clearly defined, i.e. the questions to be asked must be precisely formulated, so that a misleading or nonsensical answer is not inadvertently obtained. Although such general requirements are obviously not unique to THE USE O F STATISTICS IN. PHYTOSOCIOLOOY

64

J . M. LAMBERT AND M. B . DALE

phytosociological studies, the last of them is especially relevant. The more complex a problem, the more difficult it may be to disentangle assumptions, concepts, postulates and the like from the situation actually requiring statistical examination. This is not to imply that a mixture of concepts may not occasionally be used when speed of operation is the limiting factor; but exceptional care is then required in interpreting the results and the pitfalls must be f d y realized (wide, for example, the discussion on the use of frequency data in Greig-Smith, 1964, pp. 9-19). Similarly, the more elaborate the method of analysis itself, the more it is necessary to understand the underlying assumptions and logical consequences of each mathematical step. The use of statistics in phytosociological work has thus an additional hidden function in that it forces the investigator to clarify his mind as to the precise nature of his current problem: the demands of exacting methods can prove a powerful antidote to any tendency to loose or superficial thinking.

11. THE NATURE OF PHYTOSOCIOLOGICAL DATA A. THE NATURE O F VEGETATION Since phytosociology is concerned with vegetational phenomena, the concept of vegetation itself must first be examined in a statistical context. Briefly, vegetation may be defined as plant material growing on the earth’s crust. There are thus three elements to be distinguished: (1) the plants themselves; (2) the sites (defined by position in space) in which they occur; and (3) the environmental features associated with these sites. This threefold plant/site/habitat distinction seems only vaguely recognized in most vegetation studies ; it nevertheless provides a useful mechanism for differentiating between strictly phytosociological phenomena (i.e. plant/plant relationships in a number of different sites) and wider ecological phenomena concerning the whole of the vegetational system. It may at first seem inappropriate t o include environmental features in any general concept of vegetation. However, we rarely think of vegetation as a set of plants completely isolated from their normal habitat. For instance, we do not envisage saltmarsh vegetation divorced from a saline soil and represented merely by a collection of halophytes laid out on a laboratory bench, or even growing in a garden: although the plants occupy spatial positions and a plant/site relationship thereby exists, a further site/habitat relationship is necessary before the system fulfils our usual concept of such vegetation. For statistical purposes, the sites may be regarded as a number of individuals with certain attributes or properties potentially in common. The particular plants attributes which concern us here are, first, the

THE USE OF STATISTICS I N PHYTOSOCIOLOQY

65

pertaining to each site and, secondly, all other properties of the sites which could conceivably affect the growth of plants. For the general ecological picture, where plantlhabitat relationships are ultimately sought, the choice before us now is either to examine both sets of properties simultaneously in a single operation, or first to analyse one set alone and use it as a reference system to which the other can then be related. From mere considerations of economy in the number of analyses involved, the first would seem the more efficient operation; but, apart from the fact that computing-time often increases disproportionately with any increase in the number of attributes handled together, there are both theoretical and practical objections to this course. I n the f i s t place, the features of the habitat which could be recorded are almost infinite; to select only those features regarded on a priori grounds as relevant immediately introduces a subjective bias based on preconceptions which may not be applicable to the particular system under study. Secondly, whereas the plants are visible entities which can usually be recorded directly in the field with reasonable ease, environmental features tend to be far more difficult and time-consuming to assess with any degree of confidence beyond a very superficial level. Thirdly, it is statistically inconvenient - and therefore inefficient - to mix two sets of attributes of very diverse nature in a single analysis if this can be avoided; although this may be necessary in certain statistical fields (as in a numerical taxonomy, where only diverse information may be available), a vegetational system offers the opportunity of dealing separately with a set of attributes measurable in comparable units for the primary analysis. Lastly - and most important - by definition our main concern in plant ecology is with plants in relation to habitats; and this in itself suggests that the plant component of the plantlsitelhabitat system should be considered first. But even if habitat features are excluded from the primary analysis, they still must be eventually considered in a full ecological investigation. The usual practice in ecological survey is to attempt to record a number of environmental features at the same time as the plants. However, apart from the question of bias considered above, this practice is not necessarily the- most efficient. It has already been indicated earlier (p. 62) that an analysis of plantlplant relationships by hypothesisgenerating methods can be used to give a lead as to the particular environmental features most likely to be involved. In certain circumstances, therefore, a preliminary phytosociological survey may be more rewarding, in terms of total effort, than a more elaborate programme of recording plants and habitat together. Investigations into the most appropriate statistical techniques for

66

J. M. LAMBERT AND M. B. DALE

eventually relating plant/site and site/habitat systems are still at a very early stage, and such methods as exist will only be dealt with briefly in a subsequent section (IIIF). As already implied, the immediate purpose of the present contribution is to examine methods of phytosociological analysis : it is the plant/site system, therefore, which must be now more fully explored.

B.

THE NATURE O F THE VARIABLES C O N C E R N E D

In any investigation of plant/site relationships within a given region, the first concern is to define and identify the entities to be recorded. The sites have already been defined as individual positions on the surface of the earth: they are thus conveniently identified for record purposes by any suitable spatial coordinates, such as individual map references. For independent identification, therefore, the units of plant material must now be separately defined in terms of characteristics other than spatial position. For certain types of vegetation study, it may be appropriate t b record the variation in plant material from site to site in terms of a single measurable property. For instance, the variation in total plant mass from one part of an area to another is often used in studies primarily concerned with general vegetational productivity ; or differences in density of an individual species may be required in autecological work. These are essentially univariate systems, however, and do not concern us here. I n contrast, in phytosociological work, the plant cover is assumed to be composed of a number of different units, separately defined, and each capable of variation across the sites. A phytosociological system is therefore multivariate in nature, and any statistical analysis of such a system must thus make use of multivariate methods. Theoretically, plant units could be defined, identified and recorded by reference to any property it is possible to conceive. There is in principle no reason, for instance, why properties like chemical composition or geometrical form should not be used, though clearly the more bizarre the unit, the more restricted will be the application of results. It is important to realize, however, that, whatever criteria are used for definition, the choice itself is a subjective step, and depends on the ultimate purpose for which the information is required. For particular purposes, therefore, criteria may be selected relevant to that particular investigation alone. Thus tree-size classes might be required for the study of tree-regeneration in a given wood; or life-form records might be appropriate for a study of vegetation on an inter-continental scale. Assuming that the ultimate aim of the investigation is to reveal some structure in the system in order to interpret it in terms of other

67 phenomena, there are, however, four basic requirements to be borne in mind. First, there should be sufficient variables for it to be possible for significant differences, if any, to appear; secondly, for practical convenience, the variables selected should be easily categorized, and the categories easily identified in the field; thirdly, the definition of the categories should be unambiguous and as independent as possible of the system under study; and lastly, to be useful for interpretation, the categories must be referable to other phenomena outside the immediate system. For ((general-puqose))survey, the most useful criteria for categorizing plant material are those which have fairly wide relevance outside and combine external information from a number of different sources. The erection of such categories must itself involve prior abstraction from other independent variables, and the level of abstraction to be used is again a, subjective decision. Thus the widespread use of species in traditional phytosociological work stems from a general acceptance of the species concept as an economical method of characterizing plant material by a number of properties simultaneously at a generally convenient level of abstraction. The use of species composition for the plant component in fact largely fulfils the four requirements outlined above; and for convenience -without prejudice as to the possible use of other plant units in other circumstances -we shall similarly use species for categorizing our plants in the rest of this contribution. THE USE O F STATISTICS IN PHYTOSOCIOLOGY

c. THE NATURE

O F THE SITES

We must now consider the criteria to be used in determining the individual sites for record. I n any given region, there is an infinite number of possible sites, and decisions must be made as to their size, shape and distribution. I n a large area, complete coverage of the surface is of course impracticable, and some method of sampling must be adopted. Since the ultimate aim of the investigation is to reveal a structure in the plant cover, and not to impose it, the sampling must be as unbiased as possible. The scale of any pattern which emerges will obviously depend on size and spacing of the sample sites, while the precision of the pattern will depend on the number of samples. Although a great deal of work on the (‘correct)’size and shape of sample adequate to represent an area of vegetation has been published, most of this is irrelevant in the present context, where considerations of precision of estimate are not involved. There are, however, other considerations to be borne in mind arising from the nature of phytosociological investigations, which are essentially concerned with the grouping of species, with the interrelations of such groups, and in some sense with their spatial distribu-

68

J . M. LAMBERT A N D M . B . D A L E

tion. Obviously the samples must be sufficiently large for each to accommodate more than one species: otherwise there can be no interrelationships for analysis. The minimal size of sample site necessary to give an unbiased sample of interrelations can only be determined if the interrelations and distribution of the species are already known ; but the bias will be limited if an arbitrary, rather than subjective, sample size is fixed. I n the absence of complete coverage, the spatial relationships will be limited by the intersample distance. Again, without foreknowledge of the situation, an arbitrary spacing must be fixed. Taking the area as a whole, it is usually preferable to have ‘equal information about spatial relations, rather than varying amounts, so the intersample diKtance should remain constant. This leads to the use of systematic sampling, usually on some form of grid, rather than random sampling. Such .considerations clearly preclude the subjective assessment of species-group area and distribution implicit in any use of “stand” samples as “representative”, and which will inevitably result in the introduction of an unknown subjective bias. This should not, however, be taken to imply that all sites which lie on the grid should necessarily be recorded, provided some external criterion is used to reject unsuitable sites. For instance, in an investigation specifically concerned with variations in forest vegetation, newly felled sites lying within the region under study might well be excluded as irrelevant to the immediate problem in hand; or, in a mountainous area, some sites might have to be omitted because they were completely inaccessible.

D.

THE NATURE O F THE MEASURES

The final decision to be made concerns the nature of the information about each species to be collected at each site. A wide variety of measures have been proposed, ranging from a simple presence-orabsence system to the complex (‘indicesof importance” used by Curtis and McIntosh (1951). For instance, measures of density, dry weight, leaf area, percentage cover, vitality and so on have all been adopted at one time or another in various phytosociological studies. Given that the prime requirement is to obtain unbiased information as efficiently as possible, the different types of measure must now be looked at in this light. In the first place, there is little to be said on theoretical grounds in favour of composite, mixed quantitative measures. Examples of these are Curtis’s “Importance Value” (a sum of non-additive numbers, i.e. relative density, relative frequency and relative cover) and the various “cover-abundance” scales much used in continental work. The latter, in fact, are little more than mere subjective estimations, given respectability by being displayed in a pseudo-quantitative form.

THE USE O F STATISTICS I N PHYTOSOCIOLOQY

69

The choice between using simpler quantitative measures and purely qualitative records rests on more subtle considerations. It seems rarely recognized in phytosociological work that, whereas presence-orabsence records form a self-contained logical system, all quantitative data are truncated in that no record is possible except zero for the absence of a species: although the amount by which a species is present can be recorded, there is no corresponding measure of the extent to which it is absent. The quantitative pattern is thus superimposed on an underlying qualitative pattern and, even with sites from roughly similar vegetation, the qualitative differences may easily override the quantitative element in information-content. We would therefore dissent from the views of Greig-Smith (1964, p. 160) when he states: “Even if the arbitrary simplification is made of considering presence and absence of species only, the important difference between stands lies in the amount of different species. . . . With only slight exaggeration we may, within the limits of a set of broadly similar stands, regard absence as simply the extreme value of a continuous variable.” I n fact, a method now exists for assessing the relative importance of qualitative and quantitative information in a given set of data; this is the partition correlation analysis of Williams and Dale (1962). So far, only one set of quantitative data, based on a crude frequency measure, has been tested by this method (Dale, unpublished), but the results are very suggestive and worth quoting briefly here. For twenty-eight species in fifteen rather similar heathland sites, the information-content (as measured by the sums of squares of the partitioned correlations) of the purely qualitative element was more than double that of the purely quantitative element, and also substantially exceeded that of both qualitative/quantitative and purely quantitative elements added together. Although quantitative differences may be important in detailed studies of pattern involving relatively few species, in large-scale phytosociological work any increase in information gained by using quantitative instead of qualitative methods is thus likely to be extremely small. It is common experience that the recording of quantitative measures is vastly more time-consuming than mere presence-or-absence recording ; moreover, the use of quantitative instead of qualitative data in any subsequent statistical analysis may well increase computing-time by as much as a hundredfold. I n general, therefore, there is much to recommend the adoption of simple floristic recording as the “standard” method of data-collecting on grounds of general efficiency.

70

J. M. LAMBERT A N D M. B . DALE

111. METHODSOF ANALYSIS

A.

THE STATISTICAL PROPERTIES O F PHYTOSOCIOLOOICAL DATA

So far, the only assumption which we have made about the nature of the vegetation to be analysed is that the plant component of the plant/ site system is composed of entities identifiable as species. We have further decided that, for the purpose in hand, the species present at each site shall be the properties we measure, either as a quantitative value or, as recommended, on a purely qualitative presence-or-absence system. On either basis, the raw data can now be displayed as a twodimensional matrix, with the individual sites along the top and the species down the side: each column will show the species present in each site, and each row will show the sites in which each species is represented. From this, two other sets of matrices can be constructed, one indicating relationships between all possible pairs of species (calculated over all sites), and the other indicating relationships between all possible pairs of sites (calculated over all species). We must now make a decision as to whether we are primarily interested in a comparison of the sites on the basis of the species they contain, or in a comparison of the species on the basis of the sites in which they occur. At present, we will confine our attention to the traditional phytosociological approach of comparing the sites in terms of their floristic composition: we shall return to the other approacbin a subsequent section (IIID). From a statistical point of view, the raw data to be analysed are already ordered to some extent, in that the sites can be represented geometrically on a number of species-axes. Completely unordered data, where relationships alone are known, are extremely rare in phytosociological work; the sole example we have encountered (G. A . Yarranton, unpublished) is one in which the number of times two species concur is known, but the number of separate occurrences of the species across the sites is not recorded. Such data require special treatment for analysis, which need not concern us here; instead, the more typical partially ordered data, capable of immediate display on axes, will form the starting point for subsequent discussion. Theoretically, since any-species can be present in any site, it is possible to conceive of a situation in which all sites contain all species; this might occur, for instance, with a set of sites in poor heathland containing only three species all of which are intimately intermingled. I n quantitative data, if the quantities ofthe species are also equivalent, the sites are clearly indistinguishable and the data completely homogeneous.

THE USE O F STATISTICS I N PHYTOSOCIOLOGY

71

However, such data may also be considered homogeneous if the quantitative measures of the separate species all show a unimodal (or amodal) distribution across the sites : the sites are then statistically part of a single population. A t the other extreme, each site might contain an entirely different set of species from all others; for instance, if one site lies in a field, one in a wood, and a third at the margin of a lake, there could easily be no overlap of species. I n such a situation, the sites have no single property in common, and the data are completely heterogeneous. I n quantitative data with, for the moment, no qualifative differences between the sites, heterogeneity can occur if a similar set of values obtains within any sub-set of the sites against a different pattern elsewhere. I n qualitative data, if sub-sets of sites exist such that within a sub-set some species are everywhere absent, then at least two sets of species-axes are necessary to define the sites and the data can again be considered heterogeneous. If the data contain both quantitative and qualitative elements, as is usual with “quantitative” measures in p4ytosociology, heterogeneity can derive from either or both elements. Thus, although it is not impossible to obtain homogeneous data in the presence of qualitative differences, it seems reasonable to assume that overall homogeneity is unlikely to exist in practice. This assumption of inherent heterogeneity in most phytosociological data in fact provides their most important statistical property in the present context, since it largely determines the nature of the most efficient methods available for their analysis : it is the reduction of this heterogeneity t o an acceptable level, rather than the establishment of homogeneity, which normally forms the basis of such methods. The widespread occurrence of heterogeneity in most phytosociological situations occasionally leads investigators to suggest, either that only the “more important” species should be used for analysis, or that such species should be weighted in some way to compensate for the qualitative differences between the sites. However, although there may occasionally be grounds for weighting on some objective criterion internal to the analysis (see p. 79), there appears no justification at all for a priori weighting or selection. The occurrence together of certain rare or inconspicuous species in certain of the sites could well prove a powerful indication of phytosociological similarities of ecological importance, which could easily be obscured by concentration on the more prominent species. For present purposes, therefore, the individual species will be treated as initially all equally relevant to the situation to be analysed, with no arbitrary assumptions as to their relative importance. Apart from questions of external weighting, however, there are

72

J . M . LAMBERT AND M . B . DALE

decisions to be made as to the form in which the data shall be subjected to analysis. If the values are not standardized in some way, the more variable species will tend to dominate the,anaIysis and information of value may be lost. I n contrast, if the data are adjusted by reduction to zero mean and unit variance, the balance will be shifted in favour of the commoner and rarer species. There is much to be said for distortion of the original data in this way: it in fact provides for the more equitable disposal of the available information and is thus to be recommended as likely to be in general more efficient.

B.

ORDINATION A N D CLASSIFICATION

Unless the data are assumed to be homogeneousin the strict statistical sense, classical statistical techniques are not available for their analysis, since these are mostly dependent on a particular distribution --the multivariate normal -for the variables and are sensitive to departures from this distribution. With basically heterogeneous data usually to be accommodated, ecologists have reacted to the situation in one of two ways. One school, accepting the concept of vegetation as composed of discrete communities, has proposed systems of cZussiJcation ; the other, leaning towards the idea of vegetation as essentially continuous, has welcomed schemes of ordination. In classification, the individuals (i.e. the sites) are arranged in groups, the members of which have certain properties in common ;in ordination, the individuals are arranged on axes, with their properties determining their positions. Both types of method are essentially “structuring” techniques, in that both are aimed at seeking a simpler structure than that of the original raw data. Before the two groups of methods can be-compared, the concept of continuity in vegetation itself needs further examination; it is rarely defined in any objective way, and it is by no means clear from the literature exactly what is intended by the use of terms like “vegetational continuum”. The position is further confused by the fact that some authors appear to use the expression in a strictly phytosociological sense (i.e. independent of the actual position of the sites), while others equate it with spatial continuity on the gr0und.t Statistically, continuous variation means that the variables concerned can take any value within the limits of truncation adopted; strictly, therefore, the concept of continuity can only apply to vegetation with no qualitative differences. I n vegetation differing floristically. therefore, the term needs a wider connotation before it can be useful. f Since this paper was written, a useful further contribution by Goodall (1963) has appeared, in which (pp. 308-310) the distinct concepts of continuity in the “vegetational space” of a geometrical model, and continuity of vegetation in real space, are fuuy appreciated and discussed.

THE USE O F STATISTICS IN PHYTOSOCIOLOUY 73 We are indebted to Professor W. T. Williams, of this Department, for the following logical definition, which we shall here adopt : “If we have a set of sites such that for every site in the group there is at least one other site with which it has one or more species in common, and if this group cannot be divided into two or more groups such that all the sites of any one group have no species in common with those of another, then the vegetation may be said to be continuous. MoreoGer, if in any group of sites which is continuous in this sense, a linear or cyclic order of species can be found such that every site is defined by an uninterrupted segment of this order, the vegetation can be said to be progressively continuous; if not, it will be said to be diffusely continuous, in which the diffuseness itself can take various forms.” With such a definition, most phytosociological data are likely to be continuous to a greater or lesser degree, except in such cases as the sites to be compared have been taken only from widely disparate ecological situations with no transitions between : such a situation can easily be tested, and needs no elaborate statistical analysis. A more generally useful concept than continuity and discontinuity, therefore, is that of the degree of heterogeneity between the different sites. I n fact, the equation of heterogeneity with discontinuity we believe to have been the cause of much of the present confusion among ecologists ; as we have here defined the terms, though a discontinuous set of sites must be heterogeneous, a heterogeneous set of sites need not be discontinuous. Further, there seems to be a common misconception that classification is only properly applicable to “discontinuous” data, while ordination techniques are more appropriate to continuous systems. I n contrast, it cannot be too strongly emphasized that there is no a priori reason why the use of either method should be restricted in this way: continuous systems can be efficiently classified if classification is desired, while “discontinuous” (i.e. markedly heterogeneous) systems can be ordinated if ordination is thought more useful for the immediate purpose in hand. Moreover, there is in principle no reason why classification and ordination techniques should be mutually exclusive : classified units can be ordinated, and ordinated units classified. Which method to adopt at a given stage of the investigation is entirely a matter for the user, irrespective of any subjective concept of the “real” nature of vegetation. The relative efficiency of the two approaches can most easily be compared by considering methods of ordination fist. Here we must first distinguish between “ordination” of the species/site information on external axes directly related to the environment (as in Whittaker’s “gradient analysis”, to be mentioned in Section IIIF), and structuring

74

J. M . LAMBERT A N D M. B . DALE

of the species/site data on purely internal axes. To avoid confusion, we shall here restrict the term to the latter situation. As indicated earlier (IIIA), the raw data themselves are already crudely ordered, in that the sites can be displayed as points along axes representing the species. The assumption now is that the existence of relationships between the species may allow reduction, on some predetermined efficiency criterion, in the number of species-axes required to accommodate the sites. The results of ordination are thus a set of new and simpler axes, on which the sites can then be replotted in a more convenient form. If enough new axes are used, the data may be completely redescribed; and the sites specified, in terms of these new axes; but since each axis is now associated with a value indicating the variability in the total population which it represents, those axes with little information-value can be rejected and the situation correspondingly simplified. This form of ordination, known as “principal component analysis”, is thus mainly a method for efficient description and display: it requires no assumption of common structure in the population analysed, and any suggestion as to the meaning of the axes is purely a matter of subjective hypothesization. A more sophisticated approach, however, is to assume that new axes can be found which will reiate to some more fundamental structure. The sites are now thought of as being informed by underlying factors-such as differences in the soil-varying over the whole area; the species will be responding to these in groups, and if groups can be found they can be used to generate hypotheses concerning the ecological properties of the sites. But the species are also responding to-aspects of the environment to which they alone are sensitive individually; variation of this sort will give no information about the area as a whole and is better eliminated. As the result of such elimination, the sites can be eventually ordinated on new axes describing only-the common variation. Methods of this sort, which seek to remove the variation due to individual interests, are known under the general title of “factor analysis”. The term involves a variety of techniques, and the computation involved is usually formidable. At the start, either the number of axes, or the common variance of each species, must be estimated, a difficulty which is resolved only by iterative procedures for approximating the required values. Moreover, if the underlying structure is assumed t o contain correlated factors, rotation of axes is permissible subject to further constraints, and this involves additional computation. Unlike component analysis, which uses all the information about the species and makes no assumption of any “common” variation, factor analysis is primarily concerned with extracting common information

THE USE O F STATISTICS IN PHYTOSOCIOLOGY

75

and thus assumes that it exists; it will therefore tend to fail, in that the results may be uninterpretable, in markedly heterogeneous situations. However, in both cases, the object of performing the analysis is not only to place the sites in relation to one another, since the primary data themselves contain such relative information if this is all that is required. Both methods aim to simplify the situation by ordering the axes in terms of information-content. I n component analysis, the new axes are mainly simplification constructs to show the nature of the population; in factor analysis, the axes themselves are of primary importance since it is they which form the basis for hypothesis-generation. It is these two interrelated sets of techniques which provide the most fundamental mathematical approach to ordination so far available. Their exponents have beeq few, but two such studies deserve some mention here. One is that of Goodall (1954b), who used a form of principal component analysis (under the misnomer of factor analysis) in a study of random samples from the Victorian mallee; however, his attempts to interpret the ‘(factors” he obtained were unsuccessful owing to marked heterogeneity in his data. Secondly, there is the work of Dagnelie (1960),who used factor analysis for a variety of studies of vegetation and environment, and obtained interpretable results ; but his data consisted of selected stands, where at least some common variation could be expected. The massive computation involved in these formal methods has probably been partly responsible for the development of more approximate methods by some workers. Thus Smensen (1948) employed measures of similarity between sites to erect a single, grossly simplified, ordination axis; and Bray and Curtis (1957) invented a method suggestive of component analysis, in which, however, the axes extracted had no common origin nor any rigorous mathematical relation to each other. I n contrast to these crude approximations to a genuine mathematical system, the literature is also beset with a number of empirical attempts at ordination, most of which are so subjective in approach that they do little more than extract an answer already built in by the use of weighted data, biased selection of “important” species, and uncritical use of inappropriate parameters. Moreover, it is far too frequently assumed that the achievement of ordination by such methods provides a demonstration that vegetation is essentially (‘continuous”. The further assumption is then often made that classification is inappropriate, with statements such as “differentiation is not sufficiently discrete to allow classification on anything but an arbitrary basis” (Anderson, 1963, p. 409). All classification is arbitrary in that the limits to the classes are set by the investigator; but all ordination is equally arbitrary in

76

J. M . LAMBERT AND M . B . DALE

that the number of axes to be used is also determined by the investigator. One of the more revealing features of much ordination work is that many of the proponents of this approach use data selected on the basis of an earlier subjective classification before ordination is attempted (cf., for instance, the use of stands of the same vegetation type by Ssrensen (1948) and Curtis (1959)). We have already been at some pains to emphasize that the more powerful ordination methods (factor analysis) are only reasonably successful when they are applied to data which have a number of species in common between the sites: although component analysis can be used in any situation to explore the degree of heterogeneity of the data, the usefulness of factor analysis is limited when applied to qualitatively different sites. I n investigations concerned essentially with the overall picture of the vegetation of an area, therefore, the relative efficiency of ordination and classification must now be looked at in this light. It is unfortunate that much of the discussion concerning the applicability of the one or the other method has frequently been confounded by the erection of false antitheses. For instance, Goodall (1954a) discusses the “inadequacies of classification” and considers that “ordination can hardly fail to be more precise than a classificatory system”. I n fact, both ordination and classificationare equally precise as methods. I n either case, as many constructs -axes or classes - are needed to accommodate the data completely as there are non-identical sites; in either case, the primary aim is to simplify a complex situation by reduction in the number of axes or classes to be assessed; and, in either case, the degree of simplification eventually adopted depends on criteria imposed by the user and not on the method itself.? A general assessment of the two approaches has recently been made by Greig-Smith (1964, Chap. 7), who comes to the conclusion (p. 161) that “unless . . . there are strong reasons for preferring classification in a particular case, ordination is the sounder initial procedure even though the results may conceivably indicate that a classification is the best means of summarizing the information finally”. Bearing in mind the requirement for overall efficiency, we shall here take the opposite point of view, at least as far as primary survey is concerned: if most vegetation is inherently heterogeneous (as distinct from discontinuous) in fforistic composition, if ordination techniques are computationally

t In his most recent paper (cf. footnote, p. 72), Goodall (1963), p. 303, further claims that ordination is “more informative” than classification, and contrasts a classification of four towns in t e r n of their land-mass relationships with an ordination of the towns on latitudinal and longitudinal axes. However, the example is not valid, since different criteria are used in the two caaes, so that the two processes give different information and are not comparable.

THE USE OF STATISTICS I N PHYTOSOCIOLOGY

77

cumbersome, and if - as Greig-Smith implies - classification is the most generally useful method of eventually docketing the information, there is much to be said for proceeding directly to a system of classification. Moreover, it need not be assumed that any advantages of the more informative methods of ordination will automatically be lost if classification is used, since similar statistical operations are frequently involved in the better classificatory techniques; in fact, the choice between ordination and classification as a primary approach rests mainly on convenience, both in the performance of the analyses and in the use of the results.

c. METHODS

O F CLASSIFICATION

The process of classification is used in so many fields that it is not surprising that the study of statistical classificatory techniques “numerical taxonomy’’ -is considerably in advance of that of pure ordination methods. The basic problems associated with the manufacture of such techniques have been currently reviewed by Williams and Dale (1964)in a complementary paper to this contribution: we shall here only outline the main points a t issue as far as the user is concerned. The first distinction to be made is that between techniques which are aimed at allotting individuals to existing classes and those concerned with the actual erection of the classes themselves from information in the data. The former type of method, known as “discriminant analysis”, is much used in situations where the systematic framework is already widely recognized and accepted : the sole function of statistics then is to facilitate decisions as to the best placing of new individuals. At present, however, we are more concerned with methods which aim to extract the maximum information from the system actually under study, rather than with attempts to accommodate the data within classes already erected on external criteria; the justification for this as a primary approach to phytosociological situations will be dealt with more fully in a subsequent section (IVA). Methods of discriminant analysis will therefore not be considered further in the present discussion, which deals essentially with possible methods of primary analysis. Among those ecologists who favour a classificatory approach, the common concept of vegetation as “a multidimensional network” has led many to suggest that some form of non-hierarchical classification provides the most effective means of representing such variation. For instance, one finds in the literature statements such as: “It became clear in the course of field study that, although communities could be forced into hierarchical groups, the true relationships of communities were multidimensional” (Poore, 1955~).Such views, however, stem largely

78

J . M. LAMBERT AND M. B. DALE

from attempts to describe vegetation as it occurs in the field rather than to subject it to efficient and searching analysis. With objective analytical techniques, the question of “forcing” units into one type of grouping or another does not in fact arise: the data are made to reveal their own internal relationships, and the classification eventually constructed depends on the requirements of the user rather than on any preconception as to “true” relationships. Much of the controversy concerning the appropriateness or otherwise of hierarchical classification in fact seems to be due t o a confusion of concepts : “multidimensional” is frequently regarded as synonymous with “reticulate”, and either or both terms are often used as the direct antithesis of “hierarchical”. We shall attempt to clarify the situation here by defining the terms as we understand them in a statistical context. The term “hierarchical” normally applies to methods of obtaining groups at several levels which can be ranked in importance, in which a unique relationship is specified between any group and the entire population at any given level ; the hierarchy need not be dichotomous, but the route by which the groups arise, as distinct from intergroup relationships, can be simply displayed in not more than two dimensions. The term “reticulate” concerns intergroup distances irrespective of the method by which the groups are extracted, and methods of “reticulate classification” could in fact be regarded as hierarchical methods with a single polychotomous division. The term “multidimensional” concerns attempts to display intergroup relationships ; it can be applied equally as well to a reticulum provided by intergroup distances defined at any level in an hierarchy as to the groups derived directly from a reticulate classification. I n hierarchical classification, therefore, the emphasis is on the extraction of groups at successive levels of relationship, while reticulate methods are concerned with a single level only: as soon as any reticulately-formed groups are combined into “supergroups” or split into r( subgroups’’, the classification immediately becomes hierarchical. Since both methods of analysis are aimed at simplifying the data to a greater or lesser degree, some information must be disregarded in either form of classification; and the more powerful the method used for either type, the more the final results of the one are likely to diverge from those of the other. The decision between the use of the one or the other method must therefore again rest on both the ease of computation and the general convenience of the user. There have so far been very few statistical attempts to proceed directly to a reticulate system, and the fact that in practice most classificatory systems in other fields are essentially hierarchical in nature is itself a pointer to their generally greater usefulness and efficiency. We shall accordingly follow the

79 general consensus of opinion and confine our attention to methods of hierarchical classification as the primary approach. A second decision to be made concerns the use of “subdivisive” or “agglomerative” methods. The former begin with the whole population of sites and divide it successively into smaller groups, each group being examined independently for possible further subdivision as it is extracted; the latter begin at the bottom and combine the individual sites which are most alike until all individuals are eventually united in a single population. Subdivisive methods thus concentrate essentially on differences,while agglomerative methods seek similarities. Two points are of particular interest here. First, it should be noted that subdivisive methods start from maximal information obtained over the whole population, while agglomerative techniques start from single units of minimal information. Secondly, subdivisive methods can be terminated at any convenient level, while agglomerative methods require the whole analysis to be completed before the large-scale divisions at the top of the hierarchy can be obtained. I n general, therefore, subdivisive methods are to be preferred on theoretical grounds, although the actual calculations involved frequently require more computing time. A further choice lies between the use of “polythetic” or ‘‘monothetic” methods. Polythetic methods employ a combination of characters to form the groups, while monothetic methods use only a single character for each division; monothetic methods can only be subdivisive, but polythetic methods can be either subdivisive or agglomerative. The theoretical advantage of polythetic systems is that the classification obtained is usually more stable and, by its nature, more informative; against this, monothetic methods usually involve much less computation. I n addition to the above, there are yet further decisions to be made on more subtle statistical considerations. These have been set out fully by Williams and Dale in the parallel paper already mentioned, and need only very brief consideration here. The &st concerns the question of “internal weighting”. We have already (p. 71) condemned the practice of external weighting on subjective estimations of the “most important” species, but the problem of internal weighting of those species found to be most informative during the course of the analysis itself is less easily resolved. Given that the primary requirement is to find maximum similarities between closely related sites, and maximum dissimilarities between the groups, the analysis can frequently be made more powerful by internal weighting. The principle involved is roughly as follows. While the population of sites as a whole may possess a multiplicity of attributes, any single site may in fact possess very few, so that the site could contain very little information in its own right. By considering THE USE O F STATISTICS IN PHYTOSOCIOLOQY

80

J . M . LAMBERT AND M. B . D.ALE

the relationships between species, however, a number of measures of “importance” of each species, in either the whole population or in a particular part of the population, can be calculated. This information can then be imported into the sites as weighting coefficients, thereby increasing the informativeness of the individual sites. I n subdivisive methods, the weights can be recalculated for each subgroup as it is obtained; agglomerative methods can only use weights calculated over the whole population throughout the analysis, and this provides a further reason for preferring subdivisive method as a general rule. Again, there is the choice between using “self-structuring” and “transposed-structuring” techniques. The former examine site/site relationships directly in terms of their species composition ; the latter employ species/speciesrelationships to generate the groups of sites. I n addition to the decisions to be made as to the form of the method, there are also choices within the techniques themselves as to the actual parameter to be used in calculation. For instance, in self-structuring methods, where the parameter commonly chosen is some measure of “similarity”, some parameters involve correction for “double-zero” matches between species while others accept such matches as informative. Again, in transposed-structuring, which usually employs a “correlation” measure, the relative weight to be given to small values can be altered by appropriate transformation. With so many choices available, it is not surprising to find that a variety of different combinations have been used in practice. Thus Goodall (1953) devised a basically valuable subdivisive, monothetic, transposed method but used an inefficient parameter for subdivision ; the underlying idea was later adopted by Williams and Lambert (1959, 1960), who recast and refined the technique and used it with success on several test-communities. Examples of the use of other methods in ecology can be seen in the work of Bellamy (1962), who used a simple self-structuring agglomerative method for analysing the vegetation of a number of Polish bogs; and of Harberd (1962), who again used a selfstructuring technique but employed a distance measure derived by a needlessly circuitous route. The ultimate decision as to the specific method to be used in a given situation must depend partly on the relative power of the methods themselves, and partly on practical considerations. The most efficient method is that which uses the greatest amount of relevant information to maximize the intergroup differences (or minimize the intragroup differences) subject to computational practicability. Thus, although a polythetic subdivisive method with internal weighting might be theoretically most desirable, the time involved in actual computation could easily put the method out of reach for any but the simplest

THE USE O F STATISTICS I N PHYTOSOCIOLOBY

81

situation. I n large-scale survey, it is the overall pattern of the vegetation, rather than the exact relationships of every individual site, which provides the major interest; in such a situation, a polythetic method might well be rejected in favour of a cruder, but quicker, monothetic technique, since “misclassification” of a few isolated sites is unlikely seriously to disturb the general ecological picture. Once the groups of sites have been eventually extracted, they may either be used directly for the generation of hypotheses as to the ecological nature of their interrelationships, or they could be manipulated further. For instance, the groups at a given hierarchical level could be explored for reticulate relationships, although, if the classification has been efficient, such exploration may only confuse instead of clarify the issue. Similarly, the groups in theory could be ordinated, but, again, an efficient primary classification is likely to render subsequent ordination unprofitable. However, although the use of the more powerful classificatory techniques may thus make subsequent structuring on other systems inadvisable, the potential value of the primary results for direct hypothesization is itself increased : by maximizing the fundamentally important similarities or differences between the sites, the central relationships are exposed more clearly by greater elimination of peripheral information. I n short, assuming that the most important function of such statistical analysis is to facilitate hypothesization rather than merely to describe, the use of the most powerful classificatory method available seems justified; once the data have been efficiently structured in this way, the primary statistical operation may be terminated as far as the site relationships alone are concerned.

D.

INVERSE ANALYSIS

So far, we have dealt exclusively with ultimate sitelsite relationship, defined in terms of their floristic composition; such classification of sites, in fact, is comparable in effect to the classification of “stands” in traditional phytosociology , since the so-called stands are basically individual sites with their whole complement of species. However, the interest of the data should not be regarded as lying entirely in the site relationships, since the extraction of separate species-groups may also provide information of value to the phytosociologist: a complementary classification of the species, in terms of the sites in which they occur,/ may in fact reveal similarities and differences in their ecological behaviour which might otherwise be obscured. This we may call “inverse” analysis, to contrast it with the “normal” process of siteanalysis. It is obvioupthat the type of information obtained by “inverse” analysis will be different from that provided by “normal” methods. I n

82

J . M . LAMBERT AND M. B . DALE

site-classification, sites with comparable sets of species are grouped together, from which it can be postulated that roughly comparable habitat conditions will obtain over the sites of any given site-group; in “inverse” analysis, however, groups of species which occupy roughly the same range of sites will be extracted, and the hypothesis now is that the constituent species of each species-group will have certain physiological/ecologicalproperties in common. Although there have been a number of contributions dealing with the detection of associations between individual pairs of species (see Greig-Smith, 1964, Chap. IV), few phytosociological analyses directed at the formation of species-groups as the primary aim of the operation have appeared in the literature so far. The somewhat empirical method of Hopkins (1957) extracted “basic units” of reticulately-linked species by using positive associations between species to erect the groups, but concentrated more on the number of associations than on their relative values; while Kershaw (1961) again extracted reticulately-linked groups by using both positive and negative associations at different block sizes of quadrats to calculate the overall co-variance of species. There is, however, in principle no reason why any of the more‘powerful hierarchical classificatory techniques already discpssed should not be applied to the species as well as to the sites. The only example we know of the independent use of the same hierarchical method for both site- and species-classification is that of “association-analysis” used normally for the sites and inversely for the’ species (Williams and Lambert, 1961a). This method requires the use of a correlation-matrix for the variables, which are necessarily brought to zero mean and unit variance. Differences in “abundance” of the species are thus eliminated from the normal analysis but remain in the inverse analysis, while differences in “richness” of the sites are eliminated from the inverse analysis but remain in the normal analysis ;but since both “abundance” and “richness” are ecologically meaningful, this does not diminish the interpretability of the results. When methods of equal power are used for both normal and inverse analyses, however, there is a further feature to be considered. Whereas each site is an individual in its own right and floristically identical sites could easily occur, each species is already an abstraction from taxonomic data and it is unlikely that any one species will have an exactly comparable ecological range to any other. On a priori grounds, therefore, the degree of heterogeneity between the species could well be considerably greater than that between the sites, and this must be recognized in any cross-comparison of the results from the two independent analyses. ’

THE USE O F STATISTICS IN PHYTOSOCIOLOQY

E.

83

NODAL ANALYSIS

Whereas the normal analysis alone gives site-groups and the inverse gives species-groups, the central information concerning specieslsite relationships must clearly lie in a correlation between the two. I n any set of data from a natural community, it is hardly to be expected that species/species and site/site relationships will correlate perfectly : this situation would only occur if all the sites except identicals were completely discrete from one another with no overlap of species. Instead, it is common experience that most sites contain “accidental” species whose main phytosociological relationships lie elsewhere in the population; conversely, species frequently occur aberrantly in sites outside their normal ecological range. On the other hand, there may well be groups of ecologically similar sites in which one or other of their more typical species happens to be absent from certain of the sites ;conversely, a mqmber of a group of phytosociologically related species may occasionally be missing from among its usual companions. These chance presences and absences in fact represent peripheral information largely irrelevant to the central situation: a method is therefore needed to reject these individual variations and to synthesize the central information. Since both sets of results derive from the same set of primary data, some coincidence between the two is only to be expected. Theoretically, i t would appear more satisfactory to extract this central information by means of a single analysis rather than by attempting to correlate two independent sets of results. However, as Williams and Dale (1964) point out, there is a fundamental dificulty in such an approach: since the original data-matrix is not symmetrical, simultaneous manipulation of the sites and species is impossible with any statistical method involving the geometrical concept of alternative site- and species-axes in Euclidean space. Methods of extracting the central information by other statistical means are under consideration, but the difficulties appear formidable. At present, the only statistical method which exists for approximating to this central information is one in which coincidences are sought by dividing the sites and species with reference to each axis in turn at each successive subdivision in “association-analysis” (Williams and Lambert, 1961b; Lambert and Williams, 1962). The concentrates of information eventually obtained by such a method may be called “noda”, each of which is now defined by two sets of parameters, one relating to the site- and the other to the speciessubdivisions. Dense, highly ordered concentrations are clearly of greater importance than more diffuse aggregations, but the level of concentration at which aggregates are to be rejected or retained has still to be

84

J. M. LAMBERT A N D M. B. DALE

decided. The ranking method originally suggested by Williams and Lambert is only appropriate to monothetic classifications, and Dale (1964) has suggested an alternative numerical method based on information statistics which, though computationally more cumbersome, might prove to be more applicable in certain circumstances. I n any situation, there is in principle no reasoh why coincidences should not occur between a given group of species and more than one group of sites, or between a given group of sites and more than one group of species. However, although such aggregates may. be indistinguishable on one axis, they are distinguishable on the other and may therefore be separately defined by reference to that axis. For instance, an area of grassland might well contain one group of species exclusive to grazed sites, another which is graze-tolerant but not codned to pastures, and a third which is normally found in meadows but can just survi(P-eunder grazing ; conversely, any of these groups of species could occur over a variety of Werent soils, but associated in each case with a different subsidiary flora. Any one group of sites may thus bear anumber of different noda, while any one group of species may form different noda across a range of sites. Once the noda have been obtained, they represent abstractions from the data which can then be used as phytosociological units in their own right; by whatever method they are eventually obtained, their value clearly lies in their double definition in terms of both species and sites. However, to be useful as independent “vegetation-units”, they still require to be characterized for identification purposes as economically as possible. The simplest characterization is to define and identify each nodum by reference to the single species and site carrying most information concerning the unit in question, so that each is uniquely represented by a single species/site coincidence. But the decision must then be made as to the nature of the information actually required by the user. For instance, the extraction of the “characteristic” species and site could be made by reference solely to the information contained within each nodum itself; this will give maximum information about the internal composition of any particular nodum, but none about its interrelationships with other units. Alternatively, the information could be derived from the sets of species and sites from which each nodum was directly obtained; this method will incorporate information about the lateral relationships of any one nodum with others of similar site- but different species-groups, and of similar species- but different site-groups, respectively, but some information concerning the composition of the nodum may correspondingly be lost. Again, another method (and the one actually used so far) is to derive the information from the immediate parent populations of species and sites fiom which

85

THE USE OX STATISTICS I N PRYToSOCIOLOOY

the coincident units were generated; and this-will give most weight to those species and sites likely to differentiate most powerfully between closely related noda. There is clearly more work to be done on determining the most efficient methods of both abstracting and characterizing noda by statistical methods: all that can be said at present is that the concept seems valuable and might help to resolve existing phytosociological difficulties as to the nature and relationships of abstract vegetation-units.

F.

RELATIONSHIP WITH THE ENVIRONMENT

Although the erection of abstract units defined by both species and site could be legitimately regarded as the end of the strictly phytosociological operations on the data from a given area, it must be remembered that, for analytical purposes, the sites have so far been regarded as purely spatial entities with no external reference except their position on the surface of the earth. For ecological purposes, however, some knowledge of the particular habitat conditions relating to particular groups of species is required: we must therefore now return to the original concept (p. 64) of vegetation as a threefold system of plantlsitelhabitat relationships. The sites form the link between the species and the general environment ; and, whereas at present we have dealt exclusively with specieslsite relationships, we must now examine sitelhabitat relationships and use the sites to establish connection between the two. The question of what habitat features to record in a given situation must always be to some extent subjective, based on the investigator’s intuition as to the range of features most likely to be involved; within this range, however, the question of which of the recorded features show the best correlations with variations in the plant cover needs to be objectively resoIved. We have already indicated (p. 65) that there are three basic approaches to the general problem. We may examine specieslsite relationships in the first instance and use these as the reference system for site/habitat relationships ; we may make sitelhabitat relationships the focus of our interest and relegate specieslsite relationships to a subsidiary position ; or we may examine both systems independently and attempt to correlate the two sets of results. Although there are a priori reasons for preferring the &st approach, some mention should nevertheless be made of the other two. Once site-groups have been obtained from specieslsite data by any method of statistical analysis, these groups can actually be mapped and this in itself may give some guide to habitat va3iations of ecological importance. Agglomerative methods of classification will have concentrated on floristic similarities between the sites, while subdivisive D

C.E.R.

c

86 J. M. LAMBERT AND M. B. DALE metho& will have been mainly concerned with differences: mapping of agglomerated groups will therefore tend to reveal foci of habitat resemblance, while mapping of subdivided groups will pinpoint attention on boundaries between different environmental conditions. Nevertheless, although such mapping may serve to define areas of overall similarity or difference, the operative habitat factors have still to be identified; and this requires separate examination of such other properties of the sites as could be relevant. So far, all correlations on this basis have been subjective only, but there’is in principle no reason why more objective assessments should not be attempted. An example of the opposite approach, i.e. the examination of site/ habitat relationships as the primary system, is seen in Whittaker’s “gradient analysis” (Whittaker, 1956). Although his sites were ordinated instead of classified, the comparison of sites primarily by means of habitat factors instead of species composition still remains the central principle; “important” habitat factors were selected to provide axes on which to ordinate the sites, and the latter were then inspected subjectively for species relationships. Other examples can be found in Greig-Smith (1964, pp. 191-7); in fact, this type of approach, through habitat factors first, now seems to be becoming increasingly common in some American and Continental work (cf. Gounot, 1961), but so far much of it is essentially empirical in nature. An interesting preliminary study in the independent assessment of species/site and sitelhabitat relationships was made by Hughes and Lindley (1955), who analysed separately the species composition of two subjectively selected types of plant community and the chemical properties of six subjectively selected soil series bjr’means of a distance function ;but, although-they suggest the possibility of collation between environmental factors and related vegetational characteristics as a practicable exercise, no collation was actually attempted. The most appropriate method of direct collation would appear to be by some form of canonical analysis (vide e.g. Kendall, 1957); such methods deal with the interrelationships of interdependent sets of variables but, with the exception of a few preliminary studies in the social sciences, canonical techniques have so far been little used in practice. Unfortunately, there is still a general lack of empirical knowledge as to the relative appropriateness and efficacy of different methods, a lack which is due partly to the scale of computation and partly to the extreme frailty of some of the available techniques. However, on statistical as well as general grounds, at present it seems that the most efficient means of objective collation are likely to reside in further attempts to correlate habitat information directly with the resdts of an initial species/site analysis. The question then arises as to whether

THE USE O F STATISTICS I N PHYTOSOCIOLOGY

87

such collation should be attempted over the whole range of sites, and whole range of species, in the original data, or whether the nodal abstractions only should be used. Although the former will provide more information for analysis, the practical difticulties are correspondingly greater. For instance, for a complete analysis of a given area, habitat observations would be needed for all the sites; but if the vegetational noda alone are used, a “characteristic” site for each will have been defined, detailed habitat observations can be concentrated on these sites, and a great deal of field labour (as well as computation) will thereby be eliminated. Even the use of noda, however, still requires other decisions to be made. For instance, it must be decided whether the nodal aggregates of species are to be represented by their “characteristic” species alone, or whether the other species pertaining to each nodum are also to be included; further, if the noda are to be used as separate entities, a decision is needed as to whether they are to be regarded as of equivalent value on a qualitative basis, or whether they are to be represented by some form of quantitative measure; and again, if quantitative measures are required, there is the question of whether such measures should be based on the status of the noda as regards the diffuseness or concentration of their composition, on the relative number of sites occupied by individual noda, or on the number of species concerned in each case. As in any statistical work, these questions must be settled by balancing the optimum requirements of the user against the availability of the appropriate mathematical techniques and the amount of computing time which would be involved; in short, the criterion of overall efficiency again is all that can be used in arriving at the ultimate decisions.

IV. COMPARISONOF PHYTOSOCIOLOGICAL CONCEPTS A. THE BASIC ASSUMPTIONS We have deliberately restricted ourselves in the foregoing pages to the application of statistics to primary survey work, i.e. to situations where a general assessment of the overall pattern is required before detailed work can begin on more specific aspects. Here, there are two apparently irreconcilable views among ecologists as to the best procedure to adopt : one school of thought would apparently wish to try to fit the vegetation of new areas into a pre-existing classificatory framework and gradually build up a world-wide classification of vegetation comparable to that of individual plants in orthodox taxonomy; the other, realizing the magnitude of this task, is content to deal with each area under investigation as a separate entity and attempt to extract the maximum ecological information from it. We ourselves believe the

88

J. M. LAMBERT A N D M. B. DALE

latter course to be both more realistic and more generally useful; we shall therefore concern ourselves with the application of phytosociological principles in general to situations where the vegetation is comparatively unknown, but shall not assume a necessary relationship with pre-existing schemes of classification. I n any scientific investigation, certain assumptions have to be made at one point or another before the work can proceed; the initial assumption$ may be very general ones -for instance, the assumption that there is some pattern to be found in the material under study -but without such assumptions the work would not even begin. Moreover, unless the investigation is to be completely aimless, the purpose of the study has to be clearly defined and kept in mind throughout; this constantly calls for decisions by the investigator as to what information is useful for this purpose and what can be regarded as irrelevant. The actual course of the investigation therefore depends on both assumptions and decisions, both of which are essentially subjective ; however, decisions are largely based upon assumptions and, unIess the assumptions are recognized and constantly examined for their validity, the answer expected by the investigator may be inadvertently built into the results. It is probably true to say that, in any investigation, the fewer the assumptions the wider the applicability of the results, irrespective of whether the results themselves have been obtained by subjective assessment or by statistical means. We shall therefore attempt to examine some of the basic assumptions underlying much phytosociological work in an effort to eliminate those which are unnecessary; and this will involve us in a comparison of concepts as to the nature of vegetation as material for study. Our key references for this section are the works of Poore (1955a,b,c; 1956; 1962) and of Whittaker (1962), whose extensive researches into phytosociological literature provide an admirable basis for such a comparison. The first concept to be examined is that of vegetation itself. This seems to be a fluctuating concept among workers in that some would restrict it solely to plant material, while others would allow the concept to include an element of the natural habitat. We have ourselves defined the term as a threefold plantfsitelhabitat system, which allows us to include both plant/site and site/habitat relationships and yet dispense with any assumption that plants and environment are causally related and should be assessed together. We may later wish to mamine plant/ habitat relationships for ecological purposes ; but this is a secondary step and does not affect the strictly phytosociological study of plant/ site relationships. Although the plantfsite and sitefhabitat distinction may seem unnecessarily precise to ecologists concerned essentially with vegetation in its widest sense, we have nevertheless tried to show

THE USE O F STATISTICS I N PHYTOSOCIOLOCIY

89

that decisions as to the initial data to be collected, and the analytical methods to be used, ultimately depend on.the recognition of this distinction. The second basic concept is that the plant component of most vegetation consists of separately identifiable units, differing in properties other than mere spatial position. Without this assumption, the science of phytosociologywould have no meaning, and we must therefore accept it as a necessary assumption for any phytosociological work. However, the categorization of the units to be recorded is a matter for personal decision, and must be determined by the purpose of the investigation. We have already indicated (p. 67) that definition of the units by reference to generally accepted taxonomic categories (i.e. species) is likely to be most generally useful, but this does not deny the possibility of using other categories of units for specific purposes. The decision to use species as the categories of plants to be recorded leads us inevitably to a further assumption based on common experience. Our existing knowledge of vegetational systems all points to the fact that, even in very limited areas, there will be some variation in floristic composition from site to site, with some species absent from many of the sites. Since the existence of qualitative differences of this sort immediately imposes a degree of heterogeneity (p. 71) on the vegetation, we can make the assumption that we are dealing with a basically heterogeneous system; and this will help us to decide on the most effective methods of analysis. The question of whether vegetation is fundamentally “continuous” or “discontinuous” is less easy to resolve from common observation; as Poore (1962, p. 53) has pointed out, “workers who have long famitiarity with vegetation and who have examined the question critically can come to quite opposite conclusions.” However, we ourselves believe (p. 72) that much of the confusion has arisen from a general reluctance among ecologists to define precisely the terms which they are using, so that “discontinuity” and “heterogeneity” are frequently confounded. Since the assumption of heterogeneity is itself sufficient to swing the balance in favour of classification rather than ordination techniques for primary analysis, we do not feel that either of the opposing concepts of discrete communities or “vegetational continua” is particularly useful as a basis for investigation; we prefer instead to make no assumption either way, but to use only the concept of heterogeneity as a working tool. If no initial assumption of the existence of discrete communities is made, then any sampling method used must be independent of this assumption. This in itself destroys the basis for the traditional phytosociological practice of subjectively selecting “stands” of “homo-

90

J . M. LAMBERT A N D M. B . DALE

geneous” vegetation for comparison : the selection of such stands can only be made on the assumption that they exist, and the fact that workers believe that they can recognize such stands in a given situation is itself no proof of the general applicability of the method. Moreover, if the concept of the stand as the basic unit is abandoned in favour of more objective sampling, then the contentious concept of “minimal area” can also disappear, together with a priori “tests of homogeneity” for uniformity within a given stand. Apart from the saving in time, it is surely better for ‘areas of relative homogeneity to emerge from an analysis, rather than being imposed at the outset by the preconceptions of the individual investigator. If the foregoing argument is accepted, we are left with a deiinition of vegetation which includes habitat relationships as well as the plants themselves, but with no assumptions about the nature of the relationship; further, the only assumptions about the nature of the plant component is that it consists of separately identifiable units, that these can be recognized as species on accepted taxonomic criteria, and that the use of species for categorizing the plant material is likely to confer a basic heterogeneity on the system, On the assumption of such heterogeneity, together with practical considerations (see pp. 76-77), the decision is taken to classify rather than ordinate as the primary approach. From this point onwards, therefore, we may now look at other phytosociologicalconcepts related to classification.

B. THE

BASIC APPROACH

The place of phytosociology in vegetation studies is defined by Poore (1962, p. 51) as follows: “The proper province of plant phytosociological studies should be to describe vegetation and to discover and define problems for solution by more exact methods.” Elsewhere (p. 36), he states: “Every description is an abstraction from the available data” and (p. 51) “current statistical methods are inappropriate for [the description of stands for classification]”. Such statements, which are representative of many others in phytosociological literature, show a misunderstanding of the main function of a statistical approach. The place of classification lies between description and abstraction, in that it represents an ordering of the described material in a form suitable for abstraction. There are thus four distinct stages in the operation : the irlitial selection of the phenomena to be described; the description based on selection; the classijcation based on description; and the ultimate abstraction of common material for further use, such as hypothesization. The criteria to be used for the selection, the method of description, the form of classification, and the amount of abstraction, are d subjective decisions; but, whereas “description” seeks to record

THE USE O F STATISTICS IN PHYTOSOCIOLOCIY

91

the available information in as much detail as possible, “abstraction” in contrast seeks only to retain that which is important for the purpose in hand and to eliminate the rest. The function of statistics is not to describe, but to assist the investigator in the process of abstraction; moreover, by indicating the more significant features in the data, it may also assist in the process of hypothesization. If the function of phytosociology is only “to describe vegetation and to discover and define problems”, then it is scarcely worthy to be recognized as a science at all. We prefer to think of phytosociology in a wider context, i.e. as scientific investigation rather than a description of vegetational relationships. The essence of scientific work lies in the erection and testing of hypotheses, from which predictions can be made. The function of primary survey is to establish an initial pattern; this is then used as the basis for the erection of hypotheses, the testing of which requires the collection of more information. The hypothesis to be tested must not inadvertently be incorporated into the pattern, and if the object of the exercise is to determine plant/habitat relationships, then the one or the other must be excluded from the pattern used for hypothesization. However, once correlations have been established and independently tested, then abstractions can be made from the joint information for use in other contexts. Where phytosociological (i.e. plantlsite) relationships are used as a means of entry into ecological problems, rather than just as a means of discovering and defining them, then we have already suggested (p. 65) that the most practical approach is to establish the phytosociological pattern first. Given that no assumption of discrete communities is made (p. 89) and that the sampling system is predetermined on a systematic basis, the question now arises as to whether it is more efficient to determine the overall pattern by “wholistic” methods (i.e. simultaneous assessment of the total information) or by “sequential” means. If we understand Poore aright (1962, p. 38), his “method of successive approximation” involves the sequential assessment of information. However, it is by no means clear from his description of the method whether it refers only to the process of progressively adjusting hypotheses in the light of sequential observation, or whether in fact he is also concerned with methods of testing hypotheses or deriving estimates by the use of samples in sequence so that a decision can be taken as soon as the samples provide sufficient information according to some criterion : the latter approach has some parallel in “sequential” statistical methods and might be justifiable in certain circumstances. I n his advocacy of the method, Poore writes (Zoc. cit., p. 39): “The method of successive approximation would appear to be the most economical way

92

J. M. LAMBERT AND M. B. DALE

of obtaining a comprehensive understanding of vegetational variation. It is par excellence a reconnaissance method but is not limited to

reconnaisance. When the main framework is established, more detailed investigations either of the community or of the habitat will reveal finer relationships. This, not earlier, is the proper stage for the entry of more strictly quantitative methods; for only at this stage will they produce results which are commensurate to the work they involve.” It is clear from this statement that Poore has primary survey particularly in mind and, ignoring for the moment the question of the use of “quantitative methods”, we may thus look at the “method of successive approximation” itself in connection with primary survey : although Poore is specifically dealing with subjectively recognized “communities” as his basic units, presumably his argument is applicable to any type of sample. Truly sequential methods suffer from one major defect in that, strictly applied, the fist few assessments w i l l be either very vague or liable to gross error; in “wholistic” methods, however, the range of variability over the population as a whole is used to indicate the major points of interest as the basis for hypothesization. We maintain that the function of primary survey is to establish an overall pattern in the area from which hypotheses can be made as to relationships between plants and habitat. To be properly objective in approach, the investigator should keep an open mind until the data have been actually assembled and analysed. It is at this stage that incidental observations may be used, but for interpretation rather than assessment ; if such incidental observations are included haphazardly in the analysis itself, they may well bias the results and obscure the more fundamental features. It is true that the method of successive approximation in a very general sense forms the basis of all scientific work; but this only applies to the adjustment of hypotheses once they have been erected, not to the establishment of hypotheses in the first instance. The construction of any hypothesis needs a certain amount of initial information; and the greater this initial information, the more stable any hypothesis is likely to be. To form an initial hypothesig from the limited information in a single site - or even a small number of sites -in an essentially complex situation will lead a t best to vagueness and at worst to complete error; and if subsequent observations are to be based in any sense upon such an initial assessment, much effort may be spent in the correction of such error. If, on the other hand, a certain amount of information is obtained from sites distributed over the whole area under study before any hypothesization is attempted, a firmer and more objective basis can be established for decisions as to what is worth investigating further.

THE USE O F STATISTICS I N PHYTOSOCIOLOQY

93

The question as to the most appropriate stage at which to introduce “quantitative methods” depends on what is meant by such a term. Poore’s later text in fact suggests a lack of distinction between the use of quantitative methods and the use of quantitative measures. Since we have already indicated (p. 69) that more information resides in qualitative differences than in quantitative differences in the floristic composition of different sites, we would agree that quantitative memures should be reserved for detailed work rather than used in primary survey. However, if the term “quantitative methods” is equated with the use of statistical techniques, then - provided the method selected is both appropriate and efficient -we believe that such methods have as much virtue in primary survey as in subsequent more detailed work. We shall return to the question of the practicability of using such techniques in large-scale survey in our final section. Meanwhile, since we are ultimately concerned with the abstraction of information as efficiently as possible, we must examine phytosociological concepts involved in the erection of abstract units and try to assess the general situation regarding schemes of vegetational classification.

C.

THE BASIC “VEQETATION-UNIT”

The value of classification as an investigational tool is that it brings together common information to enable abstractions to be made efficiently. If vegetation is regarded as a tripartite system of species/ site/habitat, but with no assumption that species and habitat factors are coincidentally related, then a common abstraction must have due regard to information from each side. Much of the argument which constantly arises as to whether environmental features should be used in vegetational classification seems in fact to stem from the fluctuating concept of vegetation we have already mentioned (p. 88);it is frequently not realized that different workers are quite legitimately using different concepts. This point is recognized by Poore (1962, p. 61), who distingGshes “vegetation” from “vegetation-habitat complex”, and suggests that no features of the habitat should be used in the definition of vegetation-units. However, although we have suggested that in general it is most efficient to deal initially with plant/site relationships alone, we believe it is more informative to include also site/habitat relationships in the ultimate definition of vegetation-units. Such doublydefined units we have elsewhere (p. 83) called “noda”, and this concept now needs to be matched against pre-existing ideas. The original concept of a nodum as an abstract unit in phytosociological work can be attributed to Poore himself, whose “noda”, however, were subjectively extracted and defined solely by reference to floristic composition. The noda we are here concerned with possess a D2

C.E.R.

94

J . M. LAMBERT AND M . B . DALE

site-definition as well as a species-definition, so that the physical properties of the relevant sites can also be invoked for purposes of recognition of similar units elsewhere; although the habitat factors are thus involved as a secondary step, their inclusion adds to the pool of common information. The idea of simultaneous definition of vegetationunits on both species and habitat together is by no means new, since both criteria have long appeared in common terms like “reedswamp”. However, most systems of vegetational classification which have hitherto made use of environmental features have tended to use the two kinds of criteria separately at different stages in a single hierarchical system ; in contrast, the present proposal is to erect both definitions simultaneously by independent analyses. The essence of nodal analysis as we conceive the situation is that the noda finally abstracted are separate units separately related to the sites on which they occur, so that any one sample of vegetation can contain a number of noda. Perhaps the chief stumbling-block to any earlier emergence of conceptions of this type has lain in the common phytosociological practice of regarding the whole complement of species at a given site as an inviolable unit for purposes of classification. However, the idea of “noda” in our sense seems to have something in common with the long-established concept of “synusiae”, which have been variously defined but which appear to have the common property of representing tightly-knit groups of species of similar ecological requirements. Nevertheless, the direct comparison fails in that the usual concept of synusiae does not appear to embody the idea of physical intermingling of different synusiae ; the closest approximation is seen when the term is applied to different layers in stratified vegetation, since here there is at least some coincidence of site in a horizontal sense. Since the noda are also separately related to the groups of ecologically similar species from which they are derived, their erection as abstract units also allows us to accommodate the concept of physical overlap of groups of species of different ranges of ecological tolerance. The idea of these doubly-defined noda thus to some extent effects a compromise between the opposing concepts of “vegetational continua” and discrete communities ; but whether the new concept will prove as useful in practice as it appears in theory still remains to be seen. We have already mentioned (p. 87) possible practical advantages accruing from the concentration of the properties of each nodum into a single characteristic species and a single characteristic site. Although the idea of a “characteristic site” is new, the concept of a “characteristic species” is long-established and has led to widespread argument in phytosociological circles: “dominance”, “constancy” and “fidelity”

THE USE O F STATISTICS IN PHYTOSOCIOLOGY

95

all have their adherents, and there is still no general agreement. If only qualitative data are available, “dominance” can scarcely have a meaning, but “constancy” and “fidelity” can still be accommodated. If the “characteristic species” is extracted statistically by a monothetic method, it will necessarily be constant in the noda to which it applies; and if the extraction is made with reference to the parent population of sites from which each nodum is derived (p. 85), its power of differentiation between noda of closely similar habitats seems likely to invest it with a reasonable degree of local fidelity. Again, therefore, there seems to be the possibility of some compromise between opposing schools of thought. Since the noda are erected and defined by reference only to the data from a given situation, they will not necessarily have counterparts in other areas of vegetation. If the groupings are strong enough, they may in fact recur in various independent analyses, and we have ourselves found certain well-marked noda emerging consistently from different areas analysed statistically; but, although noda may serve as referencepoints for further work on ecological problems within the region under study, it is doubtful if any very useful purpose would be served by attempting to use such noda extensively beyond the boundaries of the general area from which they were derived. However, since the search for an acceptable method of abstracting, defining and characterizing “vegetation-units’’ is usually associated in ecologists’ minds with the phytosociological pipe-dream of eventually erecting a series of such units as the basis for a world-wide classification comparable to that of orthodox taxonomy, we cannot end this section without at least some reference to this concept. Whether it is appropriate or possible to use statistical methods as an aid to such classification, however, seems to us less important than whether any further extension of existing phytosociological schemes would remain useful for long enough to warrant the tremendous effort which would be involved. Vegetation patterns are constantly altering under the increasing influence of man’s activities, and any general classification would have to be constantly adjusted. Instead, we should prefer to see the wider use of more restricted classifications produced quickly and efficiently in relation to particular regional problems, and discarded without compunction once their purpose had been served.

V. THE FUTURE OF STATISTICS IN PHYTOSOCIOLOGY Although probably few modern ecologists would question the value on theoretical grounds of an objective statistical approach to vegetational problems, the extensive use of statistical methods for large-scale work is frequently regarded as so impracticable as to be scarcely worth

96

J . M . LAMBERT AND M . B . DALE

serious consideration; moreover, there is quite rightly considerable suspicion of many of the more empirical methods which have been suggested at one time or another for the solution of particular problems; and again, as Greig-Smith points out (1964, p. 211), there is always the danger of such excessive preoccupation with techniques that the problems themselves may be pushed into the background. Such considerations have probably been at least partially responsible for the rather sceptical attitude of some ecologists as to the general usefulness of a statistical approach. It is also perhaps inevitable that the theoretical advantages may frequently seem to be outweighed by sheer production difficulties: many of the potentially valuable techniques are still under development, progress is often held up through lack of necessary fundamental mathematical research, and few computerprogrammes specifically designed for ecological work have yet passed the stage of the small-scale prototype model. Most working ecologists are naturally more concerned with the question of practicability than anything else : however elegant or powerful a method, it is of little interest to them unless it can actually be used. Unfortunately, this has sometimes led to the use of relatively simple but mathematically unsound techniques ; and, although the use of such methods can never be profitable, a sound but too-elaborate method may defeat its own purpose by never being adopted. The essential need in ecology, therefore, is for a series of well-founded but relatively crude techniques in which the degree of approximation is clearly understood and recognized. Even given the existence of such techniques, however, it is the scale of most ecological data which is frequently regarded as the major obstacle to their extensive use. But here there seems to be a gross underestimation by most ecologists as to the scale of problem it will be possible to tackle with modern computer facilities once the programmes have been made. Advances in computer design have recently been so great that statistical analysis of vegetation on a regional scale can now be contemplated as a practicable possibility within the next few years. For instance, the possibility of a large-scale computer programme for association-analysis, capable of dealing with the whole of the B.S.B.I. Distribution Maps Scheme data of some 2 000 species in some 3 500 10-km squares (Walters, 1954), is now under active consideration; while a programme already exists for the Elliott 803 computer which has performed both normal and inverse analyses for 317 species in 152 vicecomital sites in about 40 h for the two analyses together (M. C. F. Proctor, in Zitt., 1964). Although the time taken to run a.particular computer-analysis is partly a function of the scale of the data, it is also very dependent on

THE USE O F STATISTICS I N PHYTOSOCIOLOGY

97

the efficiencyof the programme itself, and here the ecologist is entirelydependent on advances in machine design and in programming technique. What is frequently forgotten in estimating’the time and cost of employing computer-analysis is that these must be balanced against the time and cost in man-hours of the actual collection of the data: to contemplate an analysis extending over several days-or even weeks-would not be exorbitant if the data themselves had taken several years to collect. The question of the extent to which computer analyses can eventually replace a trained ecologist in the survey of vegetation is naturally controversial. The well-establishedphytosociological schools apparently place great emphasis on the need for experience not only in the collection of the data but also in the subsequent analytical work. However, here we sympathize fully with Webb (1954, p. 365) when he says : “It seems curious that plant-sociologists who collect such large masses of data in quantitative form should then rely merely on their faculty of recognition for sorting out the communities, instead of using a quantitative yard-stick”; and later: “For a scientific taxonomy it is not enough to assure one’s critics that with sufficient experience one can learn to recognize the units; there must be some means of defining them. Without this plant-sociology can only be a craft to which one is apprenticed, and not a science which one can learn. . . . Not only is such a procedure uncommunicable;it is also unreliable.” Although we ourselves are fully convinced of the value and power of appropriate statistical techniques in vegetational analysis, this does not of course imply that we consider that the professional ecologist will himself ultimately become outmoded. Statistics is only a servant, and must not become the master. Its function is to enable a worker to examine a new situation free from preconceived ideas, to aid him in the mechanical task of sorting the data he acquires, and to guide him to the points of greatest significance in a complex situation; but the particular problem to be tackled, the nature of the data to be collected, and the method of assessing the results, must always be determined by the investigator himself.

ACKNOWLEDGMENTS We are extremely grateful to Professor W. T. Williams for muchuseful discussion, particularly on the more statistical points, during the preparation of this paper; and we should also like to thank Mr P. GreigSmith for making the typescript of the 2nd edition of his book Quantitative Methods in Plant Ecology available to us before it was published.

98

J. M . LAMBERT AND M. B . DALE

REFERENCES Anderson, D. J. (1963). J . Ecol. 51, 403-414. The structure of some upland plant communities in Caernarvonshire. 111. The continuum analysis. Bellamy, D. (1962). Przegl. geogr. 34, 691-716. Some observations on the peat bogs of the wilderness of Pisz. Bray, J. R. and Curtis, J. T. (1957). Ecol. Monogr. 27, 325-349. An ordination of the upland forest communities of southern Wisconsin. Curtis, J. T. (1959). “The Vegetation of Wisconsin: an Ordination of Plant Communities.” Madison, Wisconsin. Curtis, J. T. and McIntosh, R. P. (1951). Ecology 32, 476-496. An upland forest continuum in the prairie-forest border region of Wisconsin. Dagnelie, P. (1960). Bull. Sew. Carte phytogdogr. Skr. B. 5 , 7-71 and 93-195. Contribution 8. 1’6tude des communautbs vbg6tales par l’analyse factorielle. Dale, M. B. (1964). The application of multivariate methods to heterogeneous data. Ph.D. thesis, University of Southampton. Goodall, D. W. (1952). Biol. Rev. 27, 194-245. Quantitative aspects of plant distribution. Goodall, D. W. (1953). Aust. J . Bot. 1, 39-63. Objective methods for the classification of vegetation. I. The use of positive interspecific correlation. Goodall, D. W. (19544. Angew. PJISoziol.1, 168-182. Vegetational classification and vegetational continua. Goodall, D. W. (1954b). Aust. J . Bot. 2, 304-324. Objective methods for the classification of vegetation. 111. An essay in the use of factor analysis. Goodall, D. W. (1962). Excerpta bot. Sect. B. 4,253-322. Bibliography of statistical plant sociology. Goodall, D. W. (1963). Vegetatio 11, 297-316. The continuum and the individualistic association. Gounot, M. (1961). Bull. Serv. Carte phytoggogr. Sir. B. 6, 7-73. Les mbthodes d’inventaire de la v6gbtation. Greig-Smith, P. (1964). “Quantitative Plant Ecology”, 2nd Ed. London : Butterworth. Harberd, D. J. (1962). J . Ecol. 50, 1-17. Application of a muItivariate technique to ecological survey. Hopkins, B. (1957).J . Ecol. 45, 451-463. Pattern in the plant community. Hughes, R. E. and Lindley, D. V. (1955). Nature, Lond. 175, 806-807. Application of biometric methods to problems of classificationin ecology. Kendall, M. G. (1957). “A Course in Multivariate Analysis”. London: Griffin. Kershaw, K. A. (1961). J . Ecol. 49, 643-654. Association and co-variance analysis of plant communities. Lambert, J. M. and Williams, W. T. (1962). J . Ecol. 50, 775-802. Multivariate methods in plant ecology. IV.Nodal analysis. Poore, M. E. D. (1955a). J . Ecol. 43, 226-244. The use of phytosociological methods in ecological investigations. I. The Braun-Blanquetsystem. Poore, M. E. D. (1955b). J . Ecol. 43, 245-269. The use of phytosociological methods in ecological investigations. 11. Practical issues involved in an attempt to apply the Braun-Blanquet system. Poore, M. E. D. ( 1 9 5 5 ~ )J. . Ecol. 43, 606-651. The use of phytosociological methods in ecological investigations. 111. Practical application. Poore, M. E. D. (1956). J . Ecol. 44, 28-50. The we of phytosociological methods

THE USE O F STATISTICS IN PHYTOSOCIOLOOY

99

in ecological investigations. IV. General discussion of phytosociological problems. Poore, M. E. D. (1962).Adv. ecol. Res. 1,3548. The method of successive approximation in descriptive ecology. Serensen, T. (1948).K . danske vidensk. Selsk. 5 (4),1-34. A method of establishing groups of equal amplitude in plant sociology based on similarity of species content. Walters, S. M. (1954). Proc. bot. SOC.Brit. Is. 1, 121-130. The Distribution Maps Scheme. Webb, D. A. (1954). Bot. Tidsskr. 51, 362-370. Is the classifhtion of plant communities either possible or desirable? Whittaker, R. H. (1956).Ecol. Monogr. 26, 1-80. Vegetation of the Great Smoky Mountains. Whittaker, R. H. (1962). Bot. Rev. 28, 1-239. Classification of natural communities. Williams, W. T. and Dale, M. B. (1962). Nature, Lond. 196, 602. Partition correlation matrices for heterogeneous quantitative data. Williams, W. T. and Dale, M. B. (in press). Adv. bot. Res. 2. Fundamental problems in numerical taxonomy. Williams, W. T. and Lambert, J. M. (1959). J . Ecol. 47, 83-101. Multivariate methods in plant ecology. I. Association-analysis in plant communities. Williams, W. T. and Lambert, J. M. (1960). J . Ecol. 48, 689-710. Multivariate methods in plant ecology. 11. The use of an electronic digital computer for association-analysis. Williams, W. T. and Lambert, J. M. (1961a). J . Ecol. 49, 717-729. Multivariate methods in plant ecology. 111. Inverse association-analysis. Williams, W. T. and Lambert, J. M. (1961b). Nature, Lo&. 191, 202. Nodal analysis of associated populations.

Litter Production in Forests of the Worldt

.

J ROGER B R A Y

Grasslands Division. D.S.I.R., Palmerston North. New Zealand and E V I L L E GORHAM

Botany Department. University of Minnesota. Minneapolis. Minn., U .S.A . I. Introduction ........................................................ 101 I1. SourcesofData...................................................... 104 A. UnpublishedStudybyJ.R.Bray ................................... 104 B . WorldReview .................................................... 106 I11. Selection and Presentation of Data ..................................... 107 A Criteriafor Acceptance ............................................ 107 B . Arrangement ..................................................... 108 IV. Litter Components ................................................... 118 A Detailed Litter Separation ......................................... 118 B . Percentage of Non-Leaf Litter ...................................... 118 C Understory Litter ................................................ 119 D MineralMaterial .................................................. 121 E . OrganicMaterial.................................................. 125 V Factors Affecting Litter-Fall .......................................... 125 A. Evergreen Gymnosperm and Deciduous Angiosperms ................. 125 B Environment ..................................................... 127 C. Treatment ....................................................... 131 D . TheTimeFactor ................................................. 133 VI. StandingCropofLeaves.............................................. 142 A SeasonalChanges-Intrinsic ...................................... 142 B . Seasonal Changes - Extrinsic ....................................... 144 C . MagnitudeofLeafCrops ........................................... 144 VII . Leaf Litter as an Index to Net Production ............................... 147 References................................................................ 152

.

. . .

.

. .

I . INTRODUCTION The organic debris shed by forest vegetation upon the surface of the soil has long engaged attention . I n the past. branches and twigs were

t This study formed part of the authors’ research programmes while they were on the staff of the University of Toronto. Canada.

102

J . ROGER BRAY A N D EVILLE GORHAM

used as fuel, and leaves as bedding for farm animals or as a soil treatment. I n Germany such utilization prompted concern over site degradation, and provided a stimulus for Ebermayer’s (1876) classic work on the production and chemical composition of forest litter. This study demonstrated conclusively the importance of litter-fall in the nutrient cycle of the forest, at the same time that its significance in soil development was being shown by Miiller’s (1887) pioneer investigation of the types of forest humus layer. More recent studies of the importance of litter-fall in the forest ecosystem have been reviewed by Lutz and Chandler (1946). I n the future, forest litter may assume additional significance. The current rapid increase in human population, with its consequent pressure on food supplies and accelerated depletion of non-renewable resources such as coal and oil, will eventually necessitate much fuller use of the world’s organic production (cf. Gaffron, 1946). To be thoroughly efficient such utilization must be mainly a t the level of green plants, the primary producers in the food web. It will also depend upon cheaper sources of energy and on great expansion of biochemical engineering, with current plant residues of all kinds serving as raw materials in addition to the present mainstays - agricultural crops, tree boles, and fossil plant deposits of coal and oil. Because much of the world’s land is best suited to the growth of trees, and wood will in any case remain a valuable raw material in its own right, forests will probably be a major source of materials for the new biochemical technology, particularly since mature stands can in certain circumstances be managed economically on a sustaining basis by selective felling. Moreover, forests utilize both light and growing season to a much greater degree than most agricultGa1 crops, especially if the trees arel evergreen. Total yields of forest dry matter compare favorably with those of farm crops, even without the constant cultivation and fertilization the latter receive (Weck, 1955 ;Ovington and Pearsall, 1956; Ovington, 1956). If forest production is to be used with maximum efficiency, the leaves and other debris should be utilized along with boles and slash, since they make up an important part of the total yield (see Table XX, p. 148). One promising use of leaf litter is as a source of protein which could be extracted from the leaves and incorporated into palatable foodstuffs. Such protein is “not as good as milk protein, but is as good as, or even better than, fish meal” (Pirie, 1962). Pirie (1953, 1958, 1961) has provided cogent arguments for attempting the extraction of protein from forest and agricultural waste on a commercial scale, and the cultivation of useful micro-organisms on the residue. The cultivation of edible fungi on beds of forest litter or on the forest floor itself (“fungal farming”) is another use of litter which might add to the world food supply.

103 The harvesting of present edible fungal growth is still inefficient and haphazard. Smith (1958) notes that conifer forests and plantations of Michigan, U.S.A., each year “burst with great quantities of relatively few species of Boletus”. He suggests that lumber companies could harvest the fungus crop as a means of paying taxes and other costs while the trees are growing to commercial size. Glesinger (1949), in a popular account entitled “The Coming Age of Wood”, points out that the cellulose in wood wastes is capable of being used not only as natural fibre, but as reconstituted cellulose in rayon and plastics, and as raw material for hydrolysis to sugar. The sugar can then be used to produce alcohol, high-protein yeast fodder, and a variety of other useful products. Presumably the cellulose in litter materials could also be so employed, though not as economically. It seems likely that the lignin in wood waste and litter, like the cellulose, can eventually become the raw material for a wide range of chemical conversions, whose industrial importance will increase greatly as reserves of coal and oil dwindle and demands for industrial raw materials grow. Brauns and Brauns (1960, pp. 742-9), in their book on lignin chemistry, point out that this substance is already used (though not on a large scale, in proportion to its availability) in the production of vanillin, plastics, ion-exchange resins, soil stabilizers, fertilizers, rubber reinforcing agents, tanning agents, stabilizers for asphalt emulsions, dispersants in oil-well drilling and other processes, and in ceramic processing. New and large-scale industrial uses will undoubtedly appear as the chemistry of lignin is further investigated. Other litter components beside cellulose and lignin may have industrial potential, the oils and resins in Gymnosperm litter being perhaps the most probable example. If litter is to be utilized commercially, harvest methods will need to be developed. Mechanized raking might serve in well-spaced plantations with closed canopy and little ground flora. I n other types of forest some form of vacuum collection might be devised, since the litter material is loose and unattached. Should litter utilization become economic, it will inevitably involve replacement of the nutrient elements present in the organic debris harvested. Nitrogen is probably the most important of these, but phosphorus, potassium and calcium will also be significant (Tamm, 1958). I n time, nitrogenous fertilizers may be synthesized from atmospheric nitrogen a t low cost, through the use of small nuclear reactors which could provide local sources of power in forested areas. The increasing use of human excretory wastes as fertilizers through sewage processing may also enable a low cost return of nitrogen, phosphorus, potassium and other nutrients to forested areas, especially since the costly sterilization needed for the agricultural use of LITTER PRODUCTION I N FORESTS O F THE WORLD

104

J . ROGER BRAY A N D EVILLE GORHAM

sewage will not be necessary in forest areas. Aerial application of various fertilizers may become a widespread technique for economically renewing or improving the fertility of forest soils, as the demand for forest products rises. Even if forest litter does not become an economic raw material in the near future, the study of quantitative aspects of litter-fall remains an important part of forest ecology, dealing with a major pathway for both energy and nutrient transfer in this type of ecosystem. And since litter production is easy to measure in comparison with the difficult and expensive techniques for estimating total net production of forest stands, the possibility that litter-fall might serve as a simple and convenient index to net production provided an additional stimulus for this review. The chief aim of the study, however, is to collate available data on the quantity of litter produced by forests in different parts of the world, -and to assess the influence of environment upon litter-fall under different forest communities.

11. SOURCES OF DATA A. UNPUBLISXED STUDY B Y J . R. BRAY Litter production was measured by Bray from 1957 to 1961 in an Angiosperm forest with a slight admixture of Pinus strobus. This forest occurs on the upper slope of the east bank of the Don River valley at Glendon Hall, Toronto, Canada (43" 40' N, 79' 22' W). It is dominated by Acer saccharurn with a density of 247 treestha. The composition is shown in detail in Table I. I n 1957 and 1958, leaves and stem fragments were collected from the forest floor in late autumn at the close of the period of leaf-fall. These collections were made from 1 f t 2 quadrats (30.5x 30-5 cm) placed at equal intervals along a transect. Newly fallen leaves and any stem material included were lifted intact from the decomposed duff layer in each quadrat. Samples were taken to the laboratory, where each leaf was inspected for signs of decay. If a mesic, calcium-rich leaf (e.g. Acer saccharurn, Fraxinus pennsylvanica) showed only slight decay or tiny holes it was retained as representative of the current crop. If a leaf with a tough, leathery surface (e.g. Quercus boreaZiis, Q. alba) was even moderately decayed, it was rejected as belonging to the previous year's crop. On this basis it was possible to separate the leaves of the current season from those of the previous season in all but a few cases. Samples were then oven-dried at 105°C. I n early winter of 1959, Hty galvanized iron litter pails (0.093m2 in area) were placed in the forest in a regular block pattern. No pail was located beneath a shrub or low sapling which would intercept litter

TABLEI Forest Composition, Glendon Hall, Toronto, Canada Basal area Frequency Density Basal area Importance at breast height index* (treeslha) (m2/tree) (m2/ha) (percentage of total) Density

Acer saccharurn Fraxinus pennsylvanica Pinus strobus Prunus serotina Quercus alba Quercus borealis

247 29 15 44 87 160

-

0.044 0.017 0.128 0.115 0.181 0.116

11.0 0.5 1.9 5.1 15.7 18.6

42.4 6.1 3.0 6.1 18.2 24.2

42.5 5.0 2-5 7.5 15.0 27.5

* Sum of frequency, density and basal area percentages.

20.8 0.9 3.5 9.5 29.9 35.2

106 12 9 23 63

87

106

J . ROGER BRAY A N D EVILLE G O R H A M

from the canopy. Each pail was wired to two adjacent steel posts t o hold it level. The bottom of the pail was slightly above ground surface, and was perforated to allow for drainage of rain and snow melt. A copper screen was placed on the bottom of each pail to prevent tiny litter particles such as bud scales from washing through the drainage holes. Samples were taken at three or more irregular intervals during the year, the major collection being made at the close of leaf-fall. A few leaves blew into the pails in winter and early spring after the canopy opened and leaf-fall was complete. These leaves were discarded from the sample. All stem material in the pails was collected along with that portion of any fallen stem lying directly above the inside perimeter of a pail. All material of animal origin, including fecal matter, was rejected from the sample. Samples from 1960 and 1961 were oven-dried at 70". I n 1958 and 1961 leaf samples were ashed in a muffle furnace at around 550°C for 24 h, to measure mineral content. Litter-fall and ash content are shown in Table 11. The similarity of leaf litter values indicates a rather uniform yearly production. Stem data are much more variable.

TABLEI1 Litter Production in Glendon Hall Forest, Toronto, Canada Litter fall (metric tons/ha/yr) 1957 1958 1960 1961 Mean Leaf, incl. bud scales, fruit Stem, incl. bark Total Ash content of leaves (% d r y w t )

2.8 0.6 3.4

3.2 3.2 6.3

3.2 0.5 3.7

0.8

3-1 3.9

3-1 1.3 4.3

-

7.9

-

9.0

8.4

B. WORLD REVIEW Literature containing data on litter-fall was reviewed in Biological Abstracts, Forestry Abstracts and miscellaneous journals. Coverage is undoubtedly incomplete owing to the wide range of journals and annual reports in which data of this kind are published. Biologists in areas for which data could not be found were consulted for literature references and for unpublished material. We are most grateful to many biologists throughout the world who answered our letters of inquiry. The following have very kindly supplied unpublished data or additional information about published data: Dr D. H. Ashton, Dr J. Brynaert, Dr H. R. De Selm, Mr A. Deville, Mr E. J. Dimock, Mr G. S. Meagher, Mr B. A. Mitchell, Dr J. S. Olson, Dr A. M. Schultz, Dr L. J. Webb, Drs F. D. Hole and G. A. Nielsen.

LITTER PRODUCTION I N FORESTS O F THE WORLD

107

We are especially indebted to the librarians in the reference room of the University of Toronto Library for their expert and unfailing bibliographic assistance.

111. SELECTION AND PRESENTATION O F DATA

A.

CRITERIA FOR ACCEPTANCE

Despite wide variation in methods of litter collection (e.g., raking of cleared surface, cloth or wire screen at soil surface, box or bucket with screen bottom above soil surface) and adequacy of sampling, most of the data examined have been included in this review, in order to obtain maximum coverage. Owing t o the difficulty of equating number, area and type of litter traps, length of exposure, etc., studies of very unequal value have had to be given equal weight. I n the case of Japanese forests (Ohmasa and Mori, 1937) the few data based on less than five plots were omitted. A study by Tarrant et ul. (1951) has been excluded because the data refer mainly to one year's growth of leaves or needles sampled from the lower portion of the crown (G. S. Meagher, private communication). All values based on complete or representative sampling of forest tree canopy have been segregated for separate examination of the yearly standing crop of tree foliage (see Table XIX). No attempt has been made to convert air-dry to oven-dry weights, partly because it has not always been possible to ascertain the method of drying, but also for other reasons. There appears to be considerable variation in weight loss upon further drying. Table I11 shows between 7 and 18% loss in weight by air-dry litter after drying either at an elevated temperature or in vucuo. Ebermayer (1876) reported losses

TABLEI11 Loss of Weight by Air-dried Litter upon Further Drying Material

% weight loss

Fagus sihatica litter Picea abies litter Pinus silvestris litter Fagus silvatica leaves Picea abies needles Picea abies needles Picea abies needles Picea ubiecr needles Populus leaves Bet& litter

18 15 14 9.4 10.0 8.1 6.8 6.9 7.8 10.5

Betub litter

10.8

Angiosperm tree litter Heath and moss litter

8.7 9.5

Drying method

100"c

Authority

Ebermayer, 1876 Ebermayer, 1876 100" c Ebermayer, 1876 Burger, 1925 Burger, 1925 I n vacuo, P,O, Lindberg and Norming, 1943 I n vacuo, P,O, Lindberg and Norming, 1943 100-105" C, 5 h. Lindberg and Norming, 1943 I n vucuo, P,O, AnderssonandEnander,1948 I n vacuo, 20" C Knudsen and Mauritz-Hansson,1939 10@-103°C, 2.6 h. Knudsen and Mauritz-Hansson, 1939 105' C Bray, unpublished In vacuo, P,O, AndrB, 1947

looo c

108

J . ROQER BRAY A N D EVILLE QORHAM

between 14 and 18%, all other losses ranged between 7 and 11%. The average for all data is lo%, whether or not each author's data are combined before averaging. Although air-dry litter may retain appreciable amounts of water, the air-drying process may also result in considerable loss of organic matter. Tamm (1955) reported dry weight losses of 1 to 10% by living pine and spruce needles stored upon moist filter paper during 48 h at room temperature, only about 2% appeared to be lost by respiration. White (1954) observed that needles of Pinus resinosa air-dried for six weeks, and then oven-dried, yielded 9% less dry weight than needles oven-dried immediately at 7OoC. If the needles were left on whole branches during air-drying, the dry weight decline amounted to 14%. It thus appears that in some cases, oven-drying and air-drying of living leaves should give fairly comparable results, since the excess water content of air-dry needles may be balanced by their dry weight losses. Whether the same is true of litter remains to be ascertained. I n any case, differences owing to drying techniques are small compared with variations in litter weight from other causes, and are unlikely to affect seriously any conclusions drawn from the material.

B. ARRANGEMENT The collected records of litter-fall are presented in Table IV, as metric tons of leaves, other, and total litter per hectare per annum (1 metric ton/ha = 892 lb/acre). Owing to rounding off original figures, total litter does not always exactly equal the sum of leaf and other litter. An initial grouping of data is given under four major headings based on broad climatic zones : Equatorial, Warm Temperate, Cool Temperate and Arctic-Alpine. The Equatorial forests are all within a 10" band north and south of the Equator, in Colombia, the Congo, Ghana and Malaya. The Warm Temperate group ranges between about 30" and 40" both south and north of the Equator, including Australia, New Zealand, and southern parts of the U.S.A. (Florida, the Carolinas, Tennessee and California). The Cool Temperate forests in North America range from Missouri and the mountains of California to Minnesota and Quebec, or about latitude 37-47' N; and in Europe from Hungary to Finland, or about 47" to 62" N. Japanese forests are included in this group, for although the mean annual temperature is not greatly different from that of New Zealand, the climate is more extreme, with distinctly cool winters. The scanty Arctic-Alpine data come from stands at 3 000 m altitude in the Sierra Nevada of California, at 800m in southern Norway, and from the Kola Peninsula in the U.S.S.R., the last region being the most northerly at approximately latitude 67" N. Within the broad climatic zones, data are arranged alphabetically by country. For each country the presentation is alphabetically by

TABLEI V Annual Production of Leaf,Other a.nd Total Litter by the Forests of the World Authority

Date

Location

Lat. Long. (approx.)

Jenny el al. Bartholomew et at. Brynaert

1949 Colombia 1953 Congo (Yangambi) p.c. (Ituri)

4s 1N 2N

74w 24E 27E

Laudelot and Meyer

1954

(Yangambi)

1N

24E

Nye

1961 Ghana (Kade)

6N

1w

Mitchell

p.0.

3N

102E

Malaya

Alt. (m)

1700 1800 1650

150

c 600 c 230 c 300 c 450

Ashton

p.c.

Australia (Victoria)

Hatch

1956

(Dwellingup)

Stoate

1958

(Western)

37s

145E

33s

ll6E

c 33s

c ll6E

270

Plant community EQUATORIAL FORESTS Rain forest Forest Ewalyptus saligna Cupressus lwrilanica Mixed forest Musanga ceeropioides, young secondary forest Mawolobium forest Mixed forest Brachystegia forest Dioapyros spp., mature secondary forest Dipterocarpue forest, lowland Dipteromrpus forest lowland Dipterocarpua forest: upland undisturbed Secondary forest, apparently never cultivated, moderately disturbed Secondary forest apparently never cultivated, moherately disturbed Secondary forest apparently never cultivated, mohefately disturbed Dipterocarpue baud%% plantation Druobahnops armnatica plantation Fagraea fragrane plantation Shoreu lepr08uh plantation (close planting) Shorea Zep7osllla plantation (wide planting) FORESTS WARMTEMPERATE (including subtropical) Eucalyptus regnans, mature forest with undergrowth, 47 trees/ha Eucalyptus regnaw, spar forest, 217 trees/ha Eucalyptus regnaw, pole forest, 1013 trees/ha Eucalyptus marginata virgin forest Eucalyptus marginata: pole forest Eucalyptus mnrginata sapling forest Eucalyptue diversieoldr, virgin forest, 0.65 canopy

I=indigenous, E=exotic.

* O=oven

dry, A=air dry.

Origin1 Age

(yr)

I I

E E I I I I I I

Drying methoda

Litter-fall (metric tons/ba/yr Leaves Other T o L 10.2 123

0 22 25

40

8.3 2.9 85

0

7.0

14.9

3.5

15.3 12.4 12.3 10.5 7.2 5.5 6.3

I I I

0

I

0

8.3

I

0

10.5

0

14.4

0

0

I I I

0

I

I

28 28 25 30

0 0 0

9.3 10.9 7.7 14.8

I

30

0

10.2

I

200

4.2

3.9

8.1

I

55

4.1

3.9

8.1

I

25

I I I I

36 c 25

0 0 0

34

3.3

6.9

1.2 2.0 1.6 2.8

1.1 1.1 1.0 2.9

2.4 3.1 2.6

6.7

TABLEIV - continued Authority

Webb

Claudot Miller and Hurst Will

Date

P.C.

Lat.

Location

(North N.S.W.)

Long. (approx.)

Alt.

Warm Temperate Forests-continued Eucalyptus diversieolw, regrowth, 0.87 canopy. Subtropical rain forest, Brgyrodendron Fieus Low suhtropkal rain forest emergent Eucalyptus acmenioides e‘tc. Warm temperate rain forest, Ceratopetalum, Schizotperia Warm temperate rain forest, Ceratopetalum, Schizomeriq Tall warm temperate rain forest with l’riatania eonferta Wet sclerophyll forest, Eucalyptus

c 305 c 15OE

1956 Morocco (Rharb) 1957 New Zealand (Wellington) 1959 (Rotorua)

34N 415 385

zrilularis

7W 175W 176W

Biawell and Schults Blow

p.c. U.S.A. (California) 1955 (Tennessee)

39N 36N

123W 84W

915

De Selm et al.

p.c.

(Tennessee)

36N

848

245

Heyward and Barnette

1936

(N.Florida)

30N

83W

Kittredge Mete

1940 1952

(California) (5. Carolina)

38N 35N

122w 82W

Plant community

(4

Eucalyptus camaldulensia Nothofag? twncata Pinus radzata Pinus radiata P i n w nigra Pseudotsuga ~ z i e s i i Pseudotsuga menziesii L a k t decidua Pintu ponderosa, pure stand Mixed Quercue spp cut over a t 62 yr. Larger trees Q .ldccinea and 9.velutina, Panus virgintuna secondary forest Secondary growth Quercue alba, Q . velulina Q . prinus Pinue palustha second growth, 889 trees/ha Pinus palustns second growth, 1161 trees/ha Pinus palustns second growth, 1947 treeslha Pinus palustns second growth, 2 402 treeslha Pinus caribaea second growth, 1277 trees/ha. Pinus cananenas P i n w taeda, P . eehinata Pinue echinata Pinus taeda Pinus eehinata and mixedangiosnerm8 ~~

l

I=indigenous, E =exotic.

a

0 -oven dry, A=air dry.

Origin‘

Age Drying (yr) method*

Litter-fall (metric tons/halyr) Leaves Other Total

2.7

I

0

3.3

I

0

7.3

I

0

4.5

I

0

5.9

I

0

6.8

I

0

6.0

0 A

4.2 3.5 4.2 5.1 1.5 2.2 2.2 2.1

1.5 3.9 2.0 2.8 1.4 0.7 1.5

0 0

3.8 3.5

0.7 1.0

I

0

2.7

I

0

3.4

I

0

2.7

I

0

3.3

I

0

3.9

I I

0 0 0 0 0

5.0 3.3 4.2 3.8

1 I E

E E P E E I I

I I

I

I I

40 28 45

40 33

45

78 15

20-25 30-40 10 1-60

A A

A A A 0 0

6.0

7.2

6.0 5.7

7.4 6.3 7.9 2.9 2.9 3.7 24

1.3 1.3 0.3 1.6

44 4.6

6.7

6.3

46 4.6

64

TABLEIV - continued Authority

Olson

Date

p.c.

Sims

1932

Biihmerle

1906 Austria

Bray Coldwell and De Long

Location

(Tennessee)

(N.Carolina)

Table11 Canada (Toronto) 1950

(Montreal)

Perina and Vintrova 1958 Czechoslovakia Bornehusch 1937 Denmark (Nsdebo)

Lat. Long. (approx.)

36N

Alt. (m)

Warm Temperate Forest-onlinued Mixed angiosperms and Pinue eehinac Mixed angiosperms and Pinusspp. Mixedsnmosoermq Mixed a n ~ o s ~ e - 6 Mixed angiospem Pinus eehinatu north-facing slope Pinu8cehinatu:south-facina slope Pinua eehina&level upland Liriodendrontulipifera Pqpulua, Frminzlsin sinkhold Querctls Carua Liriodendron lulip$era,ndrth-facing slope Q w m s ,Carya, Liriodendron lulipifera,south-f3cin~slope Quercua Carya, Liriaddrmi .!uZip$era, level upland uercu.9,Carya, Liriodendron valley inwr - uercud forest,unburned Pinus - 8uercus forest,burned

84W

36N

83W

48N

16E

44N

80W

47N

74W

B

COOLT E ~ E R A T E FORESTS Pinua niura plantation Pinus nigra plantation 120 Acer saccharurn Quermsboralis alba slight’admixtnre Pi& %obua’ Beer saccharuna

c 49N c l 8 E 56N 12E

Boysen-Jensen

1930

(Sor0)

56N

12E

Moller Aaltonen (data of Svinhnfvud)

1945 1948 Finland (South)

56N c 62N

12E c22E

Plant community

*I=indigenous, E=exotic.

Origin1 Age Drying (yr) method’

Litter-fall Leaves Other Total (metric tons/ha/n)

I

1-60

0

4.3

1.7

6.0

I I I I I I I I

1-40 1-150 1-50 1-45

0

3.6 46 4.1 4.1

1.1 07

4.6 53 4.7 61

0

0 0 0 0 0 0

09

6.6

6.2 3.8 4.7

I

4.0

I

5.0

I

6.4

I

6.3 3.5 2.0

I I

I I I

37 57 60-200

0 0 0

3.1

I

0

3.4

I I I

0

0 0

A

2.2 1.7 1.7 2.2 1.6

Fagus grandifolia Betula populifolia Pwulus arandidenlata. P. trmuloidee Pi+&spl Piceaabie6,stem diam.6 cm., 7 000/ha Picea abies, stem diam. 10 cm., 3 700/ha P i c a abiee, stem diam.21 cm., 1200/ha Fraxanue excelsior.unthinned Fraxinus excelsior,thinned Fagus silvatica PinUS 8dVestl.is

I

A

1.2

I

A

1.6

Picea abiea

I

O=oven dry, A=air dry.

0.6

I

I I I I

94

1.3

3.5 3.8 4.3

1.4

34

12 12 5-200 A

1.0

TABLEI V - continued Authoilty

viro

Danckelmann

Danckelmann

Date

1056

Location

Lat. Long. (approx.)

Alt.

(Eva)

61N

25E

105

(Vilppula) (Eva)

62N 61N

25E 25E

105 105

(Hyytiala) (Vesijako)

62N 61N

24E 25E

150 110

48N

11E

1887a Germany (5.Bavaria)

18871,

(throughout)

c 50N

(m)

c 10E

'I=indigenous, E=exotic.

Plant community Cool Temperate Forests-continued Belula Pinussilvestris Pinuasilve.at+ Pinussilvestrzd Piceaabies Picea abies Pieeaabia Betula Betula Pinus silvestris,plantations on good soils Pinwr silvestri.8, plantations on good soils Pinus silvertris, plantatibns on good soils Pinus silvestris,plantations on good SOilS Pinus silvestri.8, plantations on good soils Pinus silvestris,plantations on moderately good to poor soils Pinus silvestris, plantationson moderately good to poor soils Pinus silvest7i.8,plantationson moderately good to poor SOUS Pinus stluestrzd,plantations on moderately good to poorsoils Pinus siZvestri8,plantations on moderately good to poor solls

Fagus silvatiea, good to moderately good Soil8 Fugwr silvatiea, good to moderately good soils Fagua silvutieu, good to moderately good soils Fagus silvdica, good to moderately good soils Fagus silvdim, good to moderately good soils Fagus silvatica, fair to poor soils Fagwr silvatica fair to poor oils Fugw silvuth: fair to poor soils Picea at+ Pieea a h 8 O=oven dry, A=air dry.

Origin1

I I

I I I I I I

Age Drying (yr) methoda

50 88 58 68

A 0

78

0 0 0 0 0

I I

86 21-40

0 A

91

0

Litter-fall (metric tons/ha/yr) Leaves Other Total

1.7 1.2

0.6 0.6

1.9 1.9 1.7 1.3

1.0 0.5 0.4 0.5

1.9 2.3 1.8 2.7 2.8 2.4 2.2 1.8 1.5 3.3

I

4140

A

3.2

I

61-80

A

3.2

I

81-100

A

3.1

I

100

A

3.0

I

21-40

A

2.4

I

41-60

A

2.3

I

61-80

A

2.2

I

81-100

A

2.0

I

100

A

1.9

I

21-40

A

3.6

I

41-60

A

4.2

I

61-80

A

4.6

I

81-100

A

5.0

I

100

A

4.6

I

4140 61-80 81-100 21-40 41-60

A A A A A

3.9 4.2 3.1 3.7

I I

I I

3.6

TABLEI V - continued Authority

Ebermayer

Ebermayer (data of Hartin)

Date

1876

1876

Location

(Bavaria)

Lat. Long. (approx.)

c 49N

c 12E

Alt. (m)

Plant community

Cool Temperate Forest&-continued Picea abies Pieea dies Picea abies Fagus silvatica Fagua &vat+ Fagus silvatwa P i c a abies P i c a abiea Picea abies Picea abies Pinus si1veatri.a Pinus 6ilvestri.a Pinua eilvestris Faqua silvatica

Origin'

Age Drying (yr) methoda 61-80 81-100 100 30-60 60-90 90 30 30-60 60-90 90 25-50 50-75 75-100 80 100

A A A 0 0 0 0 0

0

0 0

0 0

A A

Litter-fall (metric tons/hs/gr) Leaves Other Total 3.8 3.6 3.4 3.4 3.4 3.3

4.5

3.4 2.9 2.8 2.9 3.0 3.6 4.0 3.8

TABLEIV - continued Authority

Ohmasa and Kori

Date

1937 Japan

Location

(Kunadacs) (Godollo) . (Era) (Kunadacs) (Retsag) (Godollo) (Ugod) (Rallol (Kunadacs) (Kallo) (Matra) (Kallo) (Kallo) (Godollo) (Retsag)' (Godollo)

Witkamp and van der Drift

1961 Netherlands (Amhem)

Bonnevie-Svendsen and Qjem

1967 Norway (Eidsberg)

Lat. Long. (wwox.)

Ale. (m)

Plant community Cool Temperate Forests-continued Populus (Hungarian szurke) Populue nigra hybrid UZmw (Hungarian venic) Betula

48N

48N

22E 201

c 36N c 136W

62N

I

W W U 8 8Ee8aflo7U

I I

uercue robur

Chanzaecyparisobtuea Pinue denaiflora Pinus thunbergii Thujopsid dolabrata Lark kampferc (leptolepia) Abies eaclutlinensur PieeajEZO6nSd Pieea glehnii Caetanea crenata Betula latifolia Quercuarobur, Betula vem(coda, mor

6E

.nil

60N

11E

160

61N 62N 6ON 6ON

11E 11E 12E 11E

69N

10E

170 330 250 80 250 150 170 50

I =indigenous, E =exotic.

mull Boil b r k Sibirieo on brown earth

Lark silririca on brown earth La& sibirica on iron podzol Lark decidua on iron podzol La& l@ptol,e& on brown earth

Pieea abies on iron podzol Pieea a h 8 on brown earth Picea abies on brown earth Pice0 abkd on brown earth (tram. to nnaroii r"-"".,

0 =oven dry, A =air dry.

Age Drying (yr) method' 24 35 40 30 35 45 70

w c u e robur

~ e i & cerruco~)a,&tletnur rob~r,etc., (Ringsaker) (Storelvdal) (Qrue)

Origin'

I E E E I

83

51

60

70 75 76 12-16 28 45

A A A A A A A A A A A A A A A

A

3.8 1.8 2.5 3.6 3.9 2.0 1.1 1.6 1.6 1.9 1.4 2.3 2.8 1.6 2.7

1.0

3.7

A

2.6

1.6

41

I I I

I I

I E E E I E I I I I

4.4 4.0 4.9 3.6 3.3 3.8 2.7 3.4 4.0 44 4.7 4.6 38 41 4.5 6.0

A

I I

I I I I I I .I

Litter-fall (metric tons/ha/yr) Leaves Other Tow

45

0

2.8

36 60 90 30 80 60 30-40 45-66

0 0 0 0 0 0 0 0

2.8 1.2 2.1 3.4 2.0 3.z 2.0 46

TABLEI V - continued Authority

Date

Location

Lat. Long. (approx.)

Origin'

Age

Drying

I I

90-130 80

0 0

I I I I I

26 39 63 62

0 0 0 0 0

2.5 1.6 1.3

0.6

0.3 0.0

3.1 3.1 1.0 1.0

0.1

1.7

(yr) method'

Litter-fall (metric tOns/ha/yr Leaves Other T o L

1939

(Stockholm)

69N

18E

I

0

1.6

1943

(Stockholm)

SON

18E

Pieea abiea

I

0

3.1

1038 1964

(S.W. Dalama)

(Lurid)

66N 60N

13E 16E

I I

A

2.8 1.7

47N

9E

Mixed angiosperms Betula pubeseens, open parkland (46% canopy) Fagus dlvdiea

York

Ehwald (dataof Liebundgut)

1967 Switzerland (Zurich)

Kendrlck Owen Wright

1069 U.K.(Cheshire) 1964 (N. Wales) 1967 (Roxburghahire)

330 60

61N

11E

69N

18E

63N 63N 66N

1033 U.S.A.(Minnesota) 1030 (hfinnesota)

47N 47N

96W 92W

Anonymous Chandler

1960 1941

(Missouri) (New York)

c 39N

c 92W

1944

(NewYork)

43N

42N

80 80 180 180

3w 4w 3w

Alway et al. Alway and Zon

Chandler

Plant community Cool Temperate Forests-continued Pinus silvestrb on iron podzol Fagus silvatiea on brown earth (trans. to podzol) F a g w siludiea on brown earth Pieea abiee Pieea ab-ia Bet& Populust remula, herb-rich, some Betula and Cmylus Betula pubeacend and hybrids

(Storelvdal) (Brunlanes)

Anderason and Enauder b u d s e n and Yauritz-Hanaaon Lindberg and Norming .Lindquist Sjors

Alt. (m)

77w

._..

7RW

I -indigenous, E =exotic.

Fagus eilvdica Pinus silvegtris Picea suchensis Picea abiet?, light low thinning, 460 trees/ha Pieea abies, medium thinning, 237 treestha Picea abies, heavy thinning, 67 trees/ha Pieea abiea, light crown thinning, 152 trees/ha deer saecharum and Tilia ame&ann Pinus banksiana and P . resinosa Pinus resmnosa Pinus bankaiana Pinus resinosa and P. 8trobun Pinus banksiana Pinus echinata Acer saccharurn and some mixed angiosperms Tilia americana and some mixed angiosperms aneiosDerms Tilia-americana, Tilia americana, Q y e w rubra, Carya cordifomzts Acer saccharurn, QUErcuS rubra, R little Fagus grandifolio Pinus strobus

* 0 =oven

dry, A =air dry.

I

P

A

60

2.0 2.6

19

2.8

0 0 0

4.1 2.1

0.7

1-7 48

E E

30 ,46

0 0

E

46

0

4.3

E

46

0

3.7

E

46

0

I I I I I I I I

60

0 0 0 0 0 0

2.2 2.0 2.2 2.1 2.0 2.0

30-70

0

3.3

I

30-70

0

3.1

I

30-70

0

2.9

I

30-70

0

2.9

I

c24

0

3.1

5.7

4.2

38

TABLEI V - continued Authority

Dimock

Hole and Nieleen Jenny el al.

Date

1958

p.c. 1949

Location

(Washington State)

(Wisconsin) (California)

Lat. Long. (approx.) 44N 42N 42N 44N 44N 44N 44N 47N

43N 37N

74w 76W 76W 74w 74w 74w

74w

123W

89W ll9W

Lnnt

1951

(Connecticut)

42N

73w

Scott

1955

(Connecticut)

42N

73w

Ehwald (data of Abramova)

1957 U.S.S.R.(Velikije Luki)

57N

31E

Ehwald (data of Bykova)

1957

62N

391

Ehwald (data of Nesterov)

Ehwald (data of Sacharov)

1967

1957

(Voronezh)

(MOSCOW)

(Brjansk)

66N

63N

Alt. (m)

Plant community

Cool Temperate Forests--eontinued Pinus atrobus Pinua reainosa Picea abies Picea rubens Tauga canadensis Thuja oceidenlalis Abiea balsamea 330 Paeudotauga menziesii unthiuned 300 Paeudotauga menzicsi( light thinning 320 Pseudotsuga menzhii, medium thinning 310 Paeudotsuga menziesii, heavy thinning 290 w c u a alba Q.vdutina, 420 treespa 1200wrcua kedggii 1800 1 200- Pinusponderoaa 2 200 1500 Wxed gymnosperms Pinua resinoaa Pinus strobus Beer aaccharum, Qww rubra and mixed angiosperms Pinus strobua P i c a abiea with Ozalis &a& ground flora PGea abiea with Vmcini:ummurtillw, grouud Eora Pinus advestria

8

Age

Drying method*

I I E I I I I I I I

65 c24 c24 150 150 65 25 45 46 45

0 0 0 0 0 0 0

I

45

0

(yr)

0 0

Litter-fall (metric tonslhaly~) Leaves Other Total 2.9 3.8 3.9

1.9 1.5

0

1.1

I I

100-125 60-100

0

I

is0

o

2.1

0

46

I I I I

50 30-50

I I I I I I I I

0

4.6

1.5

0

4.0 4.0 2.1

0 A

1.7 4.6

A 20

40

60

80

100

6.2 1.3

0 0 0

0 0

37 2.5 2.3 1.9 2.0 1.3

Pinua silvurtria with Qww

I

A

3.0

Pinus silvurtria with Beer Pinus silveatris with mixed spp. Pieeaabies withSambunwr understory Pica a b h Pinua ailveetris with Vaccinium &is-idaea ground flora Pinus ailvurtris with Colylus under-

I I I

A A A

3.6 4.2 8.2

38E

34E

Origin1

at,nro

I=iudigenous, E=exotic.

* 0 =oven

dry, A=air dry.

45-68

I I

A? A

2.7

0.6

6.9 3.2

I

A

4.7

2.2

6.0

TABLEI V - continued Authority

Date

Location

(Velikije Luki)

Lat. Long. (approx.)

I I

Populus tremula, with Cowlus and some ground flora Populus tremula with Cmylu8, Tilia and much modnd flora P o p u l u hemka-w ~ ith TiEia Acer platunoides add much grdund flora Pinus siluestris with Quercwr Pinus siluestris with Acer Pinus siluestris with Vaecinium sitis-idma eronnd . .. . -- -~~flora Quercusplantation Quercus and Frazinus plantation Quercus, Fraxinus and Caragann lnicrophylla Quercusand deer Quercuswith Begopodiumground flora Quernur with Aegopodium and Carex ground flora Qu.ercu.8,solonetz soil Populus, density 0.75,1688trees/ha (thinned) Pomlus densitv 1.0. 2 460 treeslha (thinned) Populus density 0.8, 988 trees/ha (thinnkd) Populus, density 1.1,1464trees/ha

I

10

0

3.9

I

25

0

4.1

I I I I

50

0

4.9

(Kiev)

51N

31E

1953

(Voronezh)

52"

39E

(Derkul steppe)

52N 49N

393

~~

40E

(unspecified) Sviridova

1960

(Voronezh)

52N

39E

"

Mork

Jenny et al. Levina

1942 Norway (Hirkjolen)

1949 U.S.A. (California) 1960 U.S.S.R. (Kola Peninsula)

62N

37N

67N

10E

119w 37E

800

3 000

0 hi

La

Litter-fall (metric tons/ha/yr) Leave8 Other Total

P i c a abks with mixed angiosperms Pinus siluestris

1957

(Voronezh)

Drying methoda

I I

Ehwald (data of Zrazevskij and Krot) Remezov and Bykova

1960

Origin1 Age (yr)

31E

1957

Sonn

Plant community Cool Teniyeralc Forcsls-- continued Picea abies with Vacciniuni mrrtiUu8 and Ozalis aceloselln groiind flora Picea ubies with Relula sp.

Ehwald (data of Smirnova)

57N

Alt. (m)

I =indigenous, E =exotic.

* 0=oven

dry, A =air dry.

A

3.3

1.6

4.9

0

1.5

0.5

2.0

0 10,45,105 0

2.1

0.6

2.7 2.4

1.3 2.0 4.5

30 60

15

3.3 3.1 4.3

50 c 210 130

5.2 4.1 4.1

25

I

ALPINEAND ARCTICFORESTS Picea abies, very slight Betula admixture Pinue siluestri-9,appreciable Belula admixture Betula Pinus conlorta Pinue siluestris with Cladonia ground flora Pinwr silveetri.8 with H~loconzium ground flora

70 38-90

I I I I I I

c 170 30

A

30

A

55

A

5.4

55

A

4.6

1.4 6.0

5.2

el35

0

0.9

0.6

c200

0

0.5

0.3

0.8

c 105 200

0 0

0.6

0.2

0.8

1.5

1.2 0.6 1.0

118

J . ROGER BRAY AND EVILLE GORHAM

author, which involves some separation of data from the same areas within the U.S.A. and the U.S.S.R.

IV. LITTERCOMPONENTS A. DETAILED LITTER SEPARATION Table V shows that leaf material contributed 60-76y0 of litter for the species listed, branches 12-15%, bark <1-14y0 and fruit <1-17%. Trees with loose dehiscent bark produced considerably more bark litter TABLEV Detailed separation of Litter Cmponents Percentage of total litter Leaf Fruit Branch Bark Other*

Pinua Pinua Pinua Picea Picea-Betulu Betula Quercua Eucalyptua

60 62 69 73 76 71 75 60

11 17 2 5 6

-

t l 15t

12 1-21-+] 12 13

14 11

-

k-,l8-----+I

12 <1 15 9 )-2b+]

<1 6 10 16 -

Authority Perina and Vintrova, 1958

Mork, 1942

Viro, 1955 Viro, 1955 Mork, 1942 Viro, 1955 Nieleen and Hole, p.c. Hatch. 1955

* Flowers, bud scales, fragments, epiphytes, insects.

t Including buds.

than did tight-barked trees. For example, the floor of a Eucalyptus forest was often covered with fallen bark; and Pinw forests, particularly of P. resinosa, also showed a high bark-fall. Bark litter from Pagw, Carpinw and other tight-barked trees was negligible. The variation of fruit litter reflects the widely varying fruit size and production of tree species and the usually short period over which litter is sampled. Curtis (1959) found that an Acer saccharurn forest produced from 99 000 seeds/ha/yr to 13 million seeds/ha/yr. Miller and Hurst (1957) noted great annual variation in seed production by a pure Nothofagus truncata forest, and observed that hot dry summers favored flowering. The occurrence of “mast years” in true beech (B’agw)forests is well known. B. PERCENTAGE O F NON-LEAF LITTER The difficulties of sampling tree stem litter have been noted frequently. Nye (1961) observed that timber-fall over a small area was very erratic and difficult to measure, since it was influenced greatly by the fall of even a single large tree. Data from Toronto (Table 11)

LITTER PRODUCTION IN FORESTS OF THE WORLD

119

showed stem litter ranging from 14 to 50% of total yearly litter over four years. Non-leaf litter data in Table IV range from 2.8% in an angiosperm stand of Bonnevie and Gjems (1957) to 55% in an Angiosperm forest of Stoate (1958). These variations emphasize that stem litter sampling requires larger areas and longer time spans than leaf litter sampling if accurate comparisons of the two are to be made. * On average, non-leaf litter makes up about 27 to 31% of the total, as shown by Table VI. The mean of all values is 30%. Values for AngioTABLEVI Percentage of Non-leaf Material in Forest Litter

All species

Angiosperms Gymnosperms

BY

Individual values

author

30 30 29

30 31 27

sperms are slightly higher than those for Gymnosperms, owing mainly to high values for several Australian Warm Temperate species. If the data are divided into latitudinal zones (Table VII), Warm Temperate TABLEVII Percentage of Non-leaf Litter in Different Climates* Climate Gymnosperms Tropical Warm Temperate, Australia and New Zealand (39) Warm Temperate. North America 37 Cool Temperate 23 Arctic-Alpine (39)

Angiosperms (33) 42 23 21 21

~~

*Figures in brackets represent a single author’s data.

areas tend to have a higher non-leaf litter percentage than Cool Temperate areas. This relationship is especially evident in the Gymnosperm species. Among Angiosperms this tendency is less clear, for while the Warm Temperate forests of Australia include many loose-barked trees (e.g. Eucalyptus), the forests of the southern U.S.A. show lower nonleaf litter percentages similar to those of their Cool Temperate counterparts farther north.

c.

UNDERSTORY LITTER

The contribution of understory plants to forest litter is closely related to the density of the forest canopy and light penetration to the under-

TABLEV I I I Understory Litter Species

Understory litter (metric (yoof total tonslhalyr) litter)

Eucalyptus regnans mature forest, 47 treeslha Eucalyptus regnans spar forest, 217 treeslha Eucalyptus regnans pole forest, 1013 trees/ha Robinia pseudacacia, 9 yr Sassafras albidum, 12 yr Larix leptolepis Larix sibirica Picea abies Pinus silvestris Ulmus glabra and mixed Angiosperms Pinus strobus Acer sacchrum Populus tremula, 30 yr, 2 460 treeslha Populus tremula, 30 yr, 1688 treeslha Populus tremula, 55 yr, 1464 treeslha Populus trernula, 55 yr. 988 trees/ha Quercus robur Mixed Angiosperms, cut in 1940

* Probably includea lower story trees. t Living material harvested in mid-summer.

2*0* 0.9*

0.8*

1.1.t 0.210.3 0.2 0.1

0.2

0.3 0.3 0.3 0.4 0.5f 0.3

0.35 0-1 0.8

25 11 11 28 7 10 7

3 7 10 16 15 8 10 8 6 4 20

Authority

Ashton (P.c.) Ashton (P.c.) Ashton (P.c.) Auten, 1941 Auten, 1941 Bonnevie-Svendsen and Gjems, 1957 Bonnevie-Svendsen and Gjems, 1957 Bonnevie-Svendsen and Gjems, 1957 Bonnevie-Svendsen and Gjems, 1957 Lindquist, 1938 Scott, 1955 Scott, 1955 Sviridova, 1960 Sviridova, 1960 Sviridova, 1960 Sviridova, 1960 Witkamp and van der Drift, 1961 Witkamp and van der Drift, 1961

f Improvement cut 1952, sampled 195S58. 8 Improvement cut 1940-42, sampled 1955-58.

LITTER PRODUCTION I N FORESTS OF THE WORLD

121

story. Data in Table VIII reveal that the maximum contribution of understory plants to total litter is 28% in a very young stand of Robinia, while an old open stand of Eucalyptus reaches 25%. Another high value of 20% is exhibited by a mixed Angiosperm stand (height 6.5-10.5 m) opened by cutting. Other values range from 3 to IS%, and average 9%.

D.

MINERAL MATERIAL

Forest litter is not wholly organic, but always contains some mineral matter. Table I X provides average values for the ash content of litter from Angiosperm and Gymnosperm species in North America and Fennoscandia. I n both regions, the Angiosperm group, and especially the non-fagaceous Angiosperms, contained more mineral material than the Gymnosperm group ; with the majority of Gymnosperms having from 2 to 6% ash, the majority of Fagaceae from 4 to 8% ash and the majority of non-fagaceous Angiosperms from 8 to 14% ash content. Mean ash content as a percentage of dry matter of the nineteen Gymnosperm species was 3-7%, of the thirteen species of Fagaceae 6.3%, and of the forty-three non-fagaceous Angiosperms 10.4%. Among genera summarized in Table I X with two or more species, mean per cent ash contents are as follows : Gymnosperms, Pinus, 3.0; Picea, 4.5 ; Juniperus, 4.6 ;Larix, 5.2 ;Fagaceae, Castanea, 4.4; Quercus, 6.6; F q u s , 6-9; non-fagaceous Angiosperms, Acacia, 4.8 ; Populus, 5.5 ; Betula, 5.8; Prunus, 7-7; Acer, 8.4; Diospyros, 9.0; Fraxinus, 10.7; Aesculus, 14.7; Ulmus, 16.1; Morus, 16.3; Celtis, 21-4. Broadfoot and Pierre (1939) present ash contents of leaf litter from trees in West Virginia, U.S.A., indicating a range of from 2 to 3% dry weight for three Pinus species (P. strobus, P . rigida, and P. virginiana). Juniperus virginiana had 5% ash. Leaf litter from fifteen Angiosperm tree species ranged from 3 to 12% ash. It is noteworthy that ash content is usually low for taxa in Table IX such as Acacia, Betulu, Castanea, Juniperus, Pinus, Populw, and Quercus, which are usually pioneer in forest development and which often occur on the more infertile sites. There is a higher ash content in taxa such as Acer saccharurn, Aesculus, .Celtis, Cladrastis lutea, Diospyros,

Praxinus, Juglans nigra, Liriodendron tulipifera, Magnolia macrophytla, Morus, Tilia americana and Ulmus, which usually occur in the more developed (or “climax”) forest communities and on the more mesic and fertile soils. Ash content of litter may vary with region, owing mainly to soil differences. For example, among the Fennoscandian data, those from Finland tend to be rather low. Detailed analyses of the major elements comprising the mineral material in litter are numerous, and have been reviewed extensively

TABLEI X Ash. Content of Litter from Various Trees ~

~

~~~~~~~~

North America

Ash yo of dry weight

1

Scott, 1955 (freshly fallen leaves or foliage)

"

Joffe, 1949 (old leaves)

*Various authors

Frequency

Gymnosperms

Fagaceae Non-fagaceous Angiosperms

Pinus rigida Pinus silvestris

2

4

Fennoscandia*

2

0

0

Pinus rigida

Pinus reainosa Pinus strobus

Pinus silveatris

4

0

0

Pinus banksiana Pinus cari baea Pinw, palustris Quercus borealis

Abies balsamea Picea rubens Populus grandidentata

Larix leptolepia

6

1

1

Acacia angustissima Caatanea sativa Juniperus pinchotii Juniperus utahensia Pinus resinosa

Betula popal~olia Acer rubrum Castanea vulgaris

3

2

3

3

3

4

Acacia roemeriana Acer rubrum Pinus strobus Quercus alba Quercus breviloba Quercus palustris

Betula pabeacens and vermosa Sorbua aucuparia Larix sibirica P i c a abies

* Anderssonand Enander, 1948. Bonnevie-Svendsenand Gjems, 1957. Hesselman, 1925. Knudsen and Mauritz-Hansson,1939. Mork, 1942. Viro, 1955.

TABLEIX - continued North America Ash yo of dry weight 6-

Scott, 1955 (freshly fallen leaves or foliage)

Fennoscandia*

Joffe, 1949 (old leaves)

Fagua grandqolia Quercus virginiana sndifolia mnsylvanica 8

9

10

11

Diospyros virginiana Fraxinus excelsior Fraxinus quadrangulata Liquidambar styraci$ua Prunus serotim

*Various authors

1

3

0

Populus tremzclcl Quercus robur

0

3

4

0

0

6

0

0

6

0

1

3

0

0

3

Carya ovata Catalpa speciosa CornusJorida Diospyros texana Magnolia macrophylla Platanus occidentalis Fraxinua americana

Ulmus americana

Robinia pseudmacia

Fagaceae Non-fagaceom Angiosperms

L a r k decidwz Fagua silvatica

Acer smcharum

Liriodendron tulipij'era Quercus douglaaii Tilia americam

Frequency

Gymnosperms

Acer platanoides

* Andersson and Enander, 1948. Bonnevie-Svendsen and Gjems, 1957. Hesselman, 1925. Knudsen and Mauritz-Hansson,1939. Mork, 1942. Viro, 1955.

TABLEIX - continued North America Ash % of dry weight

-13

Scott, 1955 (freshly fallen leaves or foliage)

-.

Joffe, 1949 (old leaves)

*Various authors

Freauencv

Gymnosperms

Fagaceae Non-fagacsoua Angiosperms

Acer saccharurn

0

0

1

Cladrastis lutea Juglans nigra

0

0

2

Aesculus glabra Robinia pseudacack

0

0

2

0

0

3

0

0

2

0

0

2

Aesculus calijornica Celtis reticulata 16

Fennoscandia*

Fraxinus excelsior

Morus microphylla Morus rubra Celtis occidentalis 26.0

Ulmus scabra 21.3

* AnderssonandEnander, 1948.Bonnevie-SvendsenctndGjems, 1957.Hewelman, 1925.KnudeenandMaurita-Hamson,1939.

Mork, 1942. Viro, 1955.

LITTER PRODUCTION I N FORESTS O F THE WORLD

125

by Lutz and Chandler (1946). Many more recent references may be found in the bibliography at the end of this review. Analyses for several of the minor elements have been made by Scott (1955).

E.

ORGANIC MATERIAL

The organic fractions of forest litter are not a t all well known. Crude proximate analyses of fresh leaf litter from four Gymnosperm and four Angiosperm tree species in eastern North America were carried out according to the methods of Waksman (see Handley, 1954) by Melin (1930), and a similar series of analyses is available for four Angiosperm and four Gymnosperm species in Finland (Mikola, 1954). These may be of interest to persons concerned with the possible utilization of leaf litter. Ether soluble components ranged from 4 to 12% dry weight, coldwater soluble organic matter from 3 to 14%, hot-water soluble organic matter from 3 to 9yo,and alcohol soluble organic material from 3 to 13%. “Hemicelluloses”, estimated crudely from sugars produced by dilute acid hydrolysis of alkali extracts, ranged from 10 to 19% dry weight. “Celluloses”, also crudely estimated by treatment of hemicellulose residues with Schweitzer’s reagent, ranged from 10 to 22%. “Ligninhumus”, the residue remaining after extraction of litter with a mixture containing 10 ml of 18% HC1 and 50 ml of 72% H,SO,, ranged from 5 to 33%. “Crude protein”, estimated by subtracting water-soluble from total nitrogen and multiplying the result by 6.25, ranged from 2 to 15% dry weight. Differences between Gymnosperms and Angiosperms were not consistent. Handley (1954), in discussing such analyses, cites some by Wittich (using methods somewhat different from those above) of German Angiosperm tree litter. These give the following ranges, as percentage of ash-free d q matter: ether soluble 6-14, coldwater soluble 6-25, hot-water soluble P 1 0 , alcohol soluble 2-5, hemicelluloses 25-36, celluloses 7-25, lignin-humus 11-30 and lignin (by acetyl bromide separation) <1-8. Nykvist (1963) has recently investigated the water-soluble aminoacids, sugars and aliphatic acids in Angiosperm and Gymnosperm litter from Swedish forest tree species.

V. FACTORS AFFECTINGLITTER-FALL A.

EVERGREEN GYMNOSPERMS AND DECIDUOUS ANGIOSPERMS

Because of their evergreen nature, Gymnosperms might be expected to be more productive than deciduous Angiosperm trees, although this factor may be countered to some extent by the tendency for Angiosperm forests to occupy more fertile sites. As far as litter-fall is concerned, Table X indicates that when a wide range of sites is considered I2

C.E.R.

126

J . ROGER BRAY A N D EVILLE QORHAM

TABLEX A Comparison of Litter Production by Evergreen and Deciduous Trees in the Northern Hemisphere

Evergreen Deciduous Gymnosperms Angiosperms (metrictons/he/yr) 3.7 3.2 2.6 2.4

No. regions averaged

Total litter Leaf litter

by difference

Other litter observed

8 9

4

1.1 0.7

0.8

0.7

Gymnosperms yield about one-sixth more total litter annually than Angiosperms, the difference amounting to 0-5 t/ha. The difference for leaf litter alone is 8% (0.2 t/ha), and for those stands where other litter was actually collected the Gymnosperm yield was about the same as that of Angiosperm trees, 0.7 t/ha. (In computing averages from combined data, only those countries or states were considered for which both evergreen Gymnosperm and deciduous Angiosperm stand data were available. The averages in Table X, and in Table XI, are based on combined data for each country, with each American or Australian state, TABLEX I Annual Litter Production in Four Major Climatic Zones Leaves Other Total no. regions metric no. regions metric no. regions metric averaged tons/ha averaged tons/ha averaged tons/ha Arctic-Alpine

1

0.7

1

0.4

3

1.0

15

2.5

10

0.9

22

3.5

Warm Temperate

8

3.6

5

1.9

7

5.5

Equatorial

2

6.8

1

3.5

4

10.9

Cool Temperate

and the Werent major areas of the U.S.S.R., treated as countries, except for the following which are grouped - the New England states, the Carolinas.) Data on net production (Table XX) support the generalization that Gymnosperm trees are somewhat more productive than Angiosperm trees, the difference amounting to about one-quarter (2.6 t/ha) for total above- and below-ground production. For leaves the difference is slight (see also data on standing crops of leaves in Table XIX). Specific areas do not always follow the above tendency. For example, within Germany the averages of data in Table I V show somewhat

L I T T E R PRODUCTION I N F O R E S T S OF T H E W O R L D

127

greater total litter production by the deciduous F q u s silvatica, 3.8 t/ha, than by the evergreens Picea abies, 3.5 t/ha, and Pinus silvestris, 3.0 t/ha. If only Ebermayer’s individual data are examined a somewhat different situation is evident, with Fagus averaging 3.3, Pinus 3.2 and Picea 3.0 t/ha (Lutz and Chandler, 1946). B.

ENVIRONMENT

7. Climate and Latitude The predominant influence of climate upon litter production is shown in a general way by Table XI, which summarizes the data for major climatic zones. I n Arctic-Alpine forests total litter production averages 1 t/ha annually, while in Equatorial forests the mean is almost 11 t/ha.i Cool and Warm Temperate forests average 3.5 and 5-5 t/ha respectively. I n round figures the ratios are about 1: 3 : 5 : 10 for the major climatic zones. The range of mean annual temperature spanned by these climatic zones is from below freezing for the most northerly Arctic forests to about 25°C for those of the hottest Equatorial regions. From a biological standpoint the length of the period when temperatures are above freezing is also important. It is of course year-long in the Equatorial forests, and almost so in the Warm Temperate stands; but in Arctic-Alpineregions the mean daily temperature may be above freezing for 6 months or less. The growing season may be considerabIy shorter, for example Paterson (1961) indicates that in parts of northern Fennoscandia the growing season for Gymnosperm forests may be as short as 3 months per annum, while in .Cool Temperate forests it is often about 6 to 7 months in duration. Associated with the higher temperature and longer growing season of the Equatorial zone is the greater amount of insolation during the period of photosynthesis. This must be of considerable importance for primary productivity. Some preliminary calculations based on maps of Black (1956) suggest that the total amount of solar radiation received during the growing season is roughly in the proportion of 1: 3 : 5 for extreme Arctic-Alpine, Cool Temperate and Equatorial sites. I n the Arctic-Alpine sites it is probably less effectively utilized owing to the more open nature of the forest and to the relative coldness of the soil. Data for both leaf and other litter, while less extensive than those for total litter, bear out the general climatic relationship. (In this con-

t I n view of the scarcity of tropical data it may be of interest to note a total annual litter-fall of 14 t/ha by Eucalyptus in Brazil, measured over 8 years (Navarro de Andrade, 1941).This value has not been included in the tables because no further information ia available, particularly as to whether this is dry weight (though it seems unlikely to be otherwise).

128

J . ROGER BRAY AND EVILLE QORHAM

nection it should be borne in mind that the three categories of litter summarized in Table XI do not represent matched sets of data, for some authors collected only leaf litter, and some of those collecting total litter did not separate leaf and other litter.) Leaf litter ranges from 0.7 tjha in Alpine Norwegian forest to nearly 7 t/ha in Equatorial Africa, with the Cool and Warm Temperate zones intermediate a t 2.5 and 3.6 tjha respectively. Data for non-leaf litter are least satisfactory, with less coverage and also less reliability because of the irregular and '4t

NORTH

O R SOUTH LATITUDE ( d e g r e e s 1

FIG.1. Annual production of total litter in relation to latitude. Open triangles equatorial, solid triangles -warm temperate, circles - cool temperate North American (open)and European (closed),squares -Arctic-Alpine. Line fitted visually to means for climatic zones, shown by large crosses. One alpine Californian stand is excluded.

occasional deposition of large branches upon collecting sites. However, even in this instance the annual production in the Norwegian mountains is about one-ninth that in Ghana, 0.4 as against 3.5 tjha. Again the Cool and Warm Temperate forests are intermediate, with 0.9 and 1.9 t/ha. The major role of temperature in controlling litter production is well illustrated in Fig. 1, where total annual litter-fall is plotted versus latitude. The relationship is inverse and linear, with a maximum level of over 11 t/ha at the Equator declining steadily to a little less than 1 t/ha a t latitude 65" N in Europe, where forest grades into tundra. Fig. 1 reveals that litter production in central and north European forests is about the same as that of similar Cool Temperate forests in

129 the northern U.S.A. and Canada, although the European sites are on average about 10 degrees farther north. The warming effect of the Gulf Stream upon European climate is probably the major factor involved, but it is also possible that the intensive management of European as compared with North American forests is important in this connection. LITTER PRODUCTION I N FORESTS OF THE WORLD

2. Altitude and Exposure Ebermayer’s (1876) data for Fagus silvatica, Picea abies snd Pinus silvestris have been summarized by 200 m altitude classes from 250 to 1 250 m. This analysis, in Table XII, indicates a peak litter production at the intermediate elevations from 450 to 850 m, although no species covers the entire altitudinal gradient. The data for Picea are most complete, and show peak litter-fall between 650 and 850 m. There is a tendency in mountainous areas for rainfall and temperature conditions to be optimum for forest growth at intermediate elevations. At higher elevations, temperatures are too low and winds too severe for luxuriant tree growth, and at low elevations, there is often a decreased rainfall. Whether these conditions apply to the data in Table XI1 is not known. TABLEXI1 Litter Production and Elevation in German Porests Elevation

(4 250

650

1050

Fagua ailvatica Picea abies Pinw, silvestris (total litter, metric tons/ha/yr) 3.8

-

3.3

4.1

3.4

6.0t

5.9*

3.9

-

3.6

-

3.1t

1250

* One value.

t Two values.

Ebermayer’s material was also divided into four exposure quadrants, NE, SE, SW and NW. Values from cardinal points (N, E, S, W) were divided in half and assigned to each of the adjacent quadrants. The results of this analysis are given in Table XIII, and indicate a lower litter production in westerly than in easterly quadrants. The highest average litter production was on the NE slope, generally considered to be least exposed to the heating and drying effects of insolation ; while

130

J. ROGER BRAY AND EVILLE GOREAM

TABLEXI11 Litter Production and Exposure in German Forests* Total litter production, metric tonslhalyr

Exposure

Total no.

Fagus silvatica

Picea abies

Pinus silvestris

Mean

Mean weighted by no. of stands

4.1 4.4 3.9 4.1

4.7 3.7 2.5 3.6

4.0 3.7 3.8 3.6

4.3 3.9 3.4 3.8

4.3 3.8 3.2 3.7

NE SE SW

Nw

stands

14.5 10.0 15.0 23.5

* From data in Ebermayer f1876). the lowest average litter production occurred on the most exposed (SW) slope. The greater average exposure effect of SW as compared with SE slopes is presumably owing to maximum reception of insolation on SW slopes in the afternoon, when the air is warmest, driest, and usually clearest. Thus insolation and evaporation maxima tend to be greatest on SW slopes.

3. Soil Fertility The influence of site class on litter production, with special reference to soil fertility, is shown in Table XIV. Site class designations are the European “bonitat” grades of decreasing fertility from I to V. Table XIV generally shows decreasing litter-fall with a decline in site quality,

TABLEXIV Litter Production and Soil Fertility in German Forests Species

Pinus szhestris

Piceu abies Fagus silvatica

Authority

Site class

Total litter (metric tonslhalyr)

I 1-111 I11

3.0 3.2 2.0 2.2 1.0 3-8 3.0 3.5 4.4 2.5 3-9

Zimmerle, 1933* Danckelmann, 1887a Wiedemann, 1948* Danckelmann, 1887a Wiedemann, 1948* Zimmerle, 1949* Wiedemann, 1936* Dietrich, 1925* Danckelmann, 1887b Wiedemann, 1931* Danckelmann, 188713 ~~

m-v

V I I11 I 1-111 I11 111-V ~~

* Cited by EhwaId (1957).

.

LITTER PRODUCTION IN FORESTS OF THE WORLD

131

although for Pagus silvatica the data fiom nineteenth- and twentiethcentury authors require separate consideration. The data for Pinus silvestris provide the clearest indication of a site effect, with litter production on the poorest site (V) one-third that on the best site (I).The intermediate site class (111)produces two-thirds as much littgr as the best class. Data of Bonnevie-Svendsen and Gjems (1957) also indicate higher litter-fall on more fertile soils in Norway. Mean litter production for Larix spp. is 1-6 t/ha on iron-podzol and 2.8 t/ha on transitional and brown-earth soils. For Picea abies the corresponding figures are 2.0 t/ha and 3.2 t/ha respectively. 4. Soil moisture Litter data from Ebermayer (1876) for Pinus silvestris and Picea abies are shown in Table XV from mesic and dry soil moisture sites.

TABLEXV Litter Production and Soil Moisture in German Forests* Species

Pin- silvestris P i n w silvestris Picea abies Picea abies

Soil moisture

No. of sites

Total litter (metric tons/ha/yr)

Mesic

6 11 31 2

4.0

Dry

Mesic

Dry

3.5 3.8 2.3

* From data in Ebermayer (1876). The decrease in litter-fall from mesic to dry conditions is especially marked for Picea, a mesic species, and less noticeable for Pinus, which tends to occur more often on exposed dry sites.

c.

TREATMENT

1. Plantations and Native Forests Only one study (Ohmasa and Mori, 1937) gives sufficient data to allow comparisons of litter-fall in natural forest stands and plantations. I n both Chamaecyparis obtusa and Pinus densijiora there is no significant difference between mean litter-fall in plantaJion and forest. Investigation by Mitchell (private communication) of a variety of indigenous Malayan species yields a mean value of 8-7 t/ha for natural communities and 10.6 tiha for plantations, not necessarily of the same species. 2. Influence of Tree Density and Basal Area Within closed-canopy forests litter production appears to be little affected by differences in tree density, as shown in Table XVI. Com-

132

J . ROGER BRAY AND EVILLE GORRAM

TABLEXVI Litter Production and Tree Density Species

Authority

Fagw silvatica

Moller, 1945

Populus trernula

Sviridova, 1960

Pinus paluatris

Heyward and Barnette, 1936

Pinus ponderosa

Biswell and Schultz (P.c.)

Density (treeslha), 179 233 244 248 281 317 901 908 1173 5 842 6 732 988 1464 1688 2 460 889 1161 1 947 2 402 1196 1495 2 931 3 459

Litter (metric tons/ ham) 2.8 2.3 2.3 2.6 2.7 2.5 2.9 2.8 2.6 2.2 3.0 5.0 4.2 5.4 4.9 2.7 3.4 2.7 3.3 2.6 2.0 1.8 2.1

parison of eleven stands of Fugw silvatica (Moller, 1945) by a rank correlation test showed no significant correlation ( p>0.10) between litterfall and tree density. Similarly, studies by Sviridova (1960), Heyward and Barnette (1936) and Biswell and Schultz (Schultz, private communication) failed to show any consistent relationship between these two variables, although in each of these instances only four stands were compared. I n Norway Bonnevie-Svendsen and Gjems (1957) have shown a distinct correlation between annual fall of leaf litter and stand basal area in a series of Gymnosperm and Angiosperm stands. Litter-fall averaged about 70-75 kg per m2basal area, over a basal area range from 8 to 40 m2/ha. In Missouri (U.S.A.),Crosby (1961) has demonstrated a correlation between total litter-fall of Pinus echinata and stand basal area, but in this case an increase from minimum to maximum basal area coverage of three-fold only doubled the annual litter-fall. The stands of low basal area had been thinned a few years prior to litter collection, and presumably the less productive trees were removed.

LITTER PRODUCTION I N FORESTS O F THE WORLD

133

3. Effect of Thinning If a closed-canopy forest is thinned there is a decrease in litter production which is roughly proportional to the degree of thinning. Data in Table IV demonstrate this relationship for Pseudotszqa menziesii (Dimock, 1958), Picea abies (Wright, 1957) and Fraxinus excelsior (Boysen-Jensen, 1930). I n all cases the control stand has the highest litter-fall and the most heavily thinned stand the lowest. Moller (1945) has shown the effect of thinning upon the standing crop of Fagus silvatica leaves. With no thinning, leaf crop was 2.0 tiha, with Bregentved thinning 1.9 t/ha and with Vemmetofte thinning 1.7 t/ha.

4.'Effect of Litter Removal Rights to utilize forest litter still existed in Germany in 1954, particularly in Bavaria, where it sometimes caused extreme reduction of forest growth (Mayer-Krapoll, 1956). By comparing areas with and without rights to litter utilization, it has been estimated that annual forest output may require forty years to recover from long-continued litter removal. The loss of nitrogen in the litter is believed to be of especial significance in lessening forest productivity. Wiedemann ( 1951) presented data from study plots (Hermeskeil143,146, Table 363, p. 260) indicating that twenty-five years of litter utilization reduced basal area increase by about two-thirds. Thirty years were then required for recovery.

D.

THE TIME FACTOR

1. Seasonal Variation If forest litter is ever to be utilized economically, it will be of importance to know the pattern of litter-fall, whether distinctly seasonal, or more or less continuous. Such knowledge is also of the utmost importance to students of population dynamics in organisms responsible for litter breakdown, and may be of interest to persons concerned with the role of organic matter in soil development. The pattern of litter-fall varies greatly, as demonstrated by Figs. 2-5. I n the Equatorial forests of Ghana, Colombia and Malaya litter-fall is continuous throughout the year, but with a tendency for slightly greater deposition during the first half of the year. I n Ghana a, short dry season in January and February was noted as leading to increased leaffall (Nye, 1961). Laudelot and Meyer (1954) state that at Yangambi in the Congo there are two minima in the wet seasons and two maxima at the ends of the dry seasons. According to Deville (private communication) seven-year plantation of Acacia decurrens showed highs in

134

J. ROGER B R A Y AND EVILLE G O R H A M RAIN FOREST, COLOMBIA,

40

REP.

TOTAL

40

20

20 -I -I

2 !x

0

20

k J

O

2

40

.

0

.

RAIN FOREST, GHANA, TOTAL

40

w I-

.

40 20

.

(I)

.

( 2.)

0

UNDISTURBED DIPTEROCARP FOREST,

MALAYA,

TOTAL

40

20

20

(3)

0

n

w n.

3 I

40

5

20

=

o

0

.

o

40

.

.

.

.

.

SECONDARY FOREST, MALAYA,

.

.

.

.

.

TOTAL

40

20

.

.

.

.

.

.

.

.

.

.

.

SHOREA LEPROSULA [INDIGENOUS 1 PLANTATION, MALAYA

1 JAN.

(3)

0 40

20

20 0

0

(31

/

FEB. MAR. APR. MAY JUNE JULY AUG. SEPT. OCT.' NOV. OEC

0

FIG.2. Seasonal litter-fall in equatorial forests. (1) Jenny et al., 1949;1(2) Nye, 1961; (3) Mitchell (P.c.).

December (18% of total litter-fall) and June (llx), with lows in May (3%) and August to September (5% each). It should be noted that many individual species of indigenous tropical rain forest are distinctly deciduous, though not necessarily tied to any sort of annual cycle. The pattern and timing of leaf replacement is extremely variable, and dependent upon both external and internal factors (see Richards, 1957, pp. 193-8). I n the Warm Temperate forests of eastern Australia, Fig. 3 shows that litter-fall goes on throughout the year, but with a distinct maximum in spring and early summer (October to December). Precipitation generally increases along with temperature during this period in eastern Australia (Walter and Lieth, 1960). I n the Eucalyptus regnans forests of Victoria, Australia, Ashton (private communication) reports about nine times as much leaf-fallin summer as in winter. I n western Australia both young and old forests of Eucalyptus marginata deposit leaf litter mainly from January to March, the warm dry part of the year. Other litter falls irregularly throughout the year, but with slightly greater intensity at about the time of maximum leaf-fall (Hatch, 1955). Eucalyptus diversi-

LITTER PRODUCTION I N FORESTS OF THE WORLD

40

:AST

AUSTRALIAN

FORESTS,

136

LEAVES SCLEROPHYLL

20

2 a

40

w

20

5

0

2 a

40

I-

I-

W

20

0

: o a

3

I I-

z

0

E

0

0

-t

-t

40 20

O 40 20 0

I I R G l N EUCALYPTUS MARGINATA, WEST A U S T R A L I A

-

OTHER

PlNUS NIGRA,

NEW Z E A L A N D

LEAVES

P l N U S RADIATA, LEAVES

-.-,#'

.___-______ --0

NEW Z E A L A N D

40

OTHER

*A,

'---

...

,

A

- 20

(4)

I-

FIG.3. Seasonal Litter-fd in forests of the southern hemkphere. (1) Webb (P.c.); (2) Hatch, 1956; (3)Miller andHurst, 1957; (4) Will, 1959.

color in western Australia shows a similar pattern of litter deposition (Stoate, 1958). I n New Zealand the peak leaf-fall of Nothofagus truncata is connected with the development of new leaves in the spring months of October and November (Miller and Hurst, 1957). The main fall of non-leaf litter takes place 2 months earlier. For the exotic northern Gymnosperms Pinus nigra, P. radiata and Larix decidua in New Zealand, maximum needle-fall occurs in autumn (March to May), according to Will (1959). While precipitation is rather uniform, March is the driest month of the year (9.4 mm) and June the wettest (14.5 mm). Pseudotsuga menziesii shows no definite seasonal trend. I n all species the fall of non-needle litter is more affected by storms than is needle-fall, which does however show some storm influence. The main period of non-needle litter deposition is clearly mid-winter (June to July) for Pinus nigra, the peak coming about 2 months after cessation of needle-fall. The pattern is less regular for Pinus radiata, the average over 4 years showing one peak near the time of maximum needle-fall and another about 5 months later. The Warm Temperate forests of Tennessee in North America also

136

J . ROGER BRAY A N D EVILLE QORRAM

40

20

0 40

rENNESSEE FORESTS, TOTAL UPLAND PINUS

20

0

0 40

i

FINNISH FORESTS,

LEAVES

(

20

3)

LL 4

a

0

-

40

u

-I W

20

13

3 0 z a u Q

40

20

*- I I 0

I

0

40 20

'ICEA

ABlES SCOTLAND, TOTAL

--

DENMARK, LEAVES

------__-h

_ . I -

40

20,

SITCHENSIS,

WALES,

---_--_

LEAVES YEAR OF LOW L E A F F A L L

YEAR ' O F 4IGH L E A F F A L L

20

(6)

~.0

0 'ICEA

.40

'

.40

0 40

20 0

IAPANESE GYMNOSPERM FORESTS, PlNllS IENSIFLORA

LEAVES

CRYPTOMERIA

oOTUSA

.. AN.

FEB. MAR. APR. MAY JUNE JULY AUG. SEPT. OCT. NOV. DEC.

0

FIG.4. Seasonal litter-fall in forests of the northern hemisphere. (1) Witkamp and van der Drift, 1961; (2) Olson (P.c.); (3) Viro, 1955; (4) Kendrick, 1959 and Danckelmann, 1887; (5) Mork, 1942; (6) Wright, 1957 and Bornebusch, 1937; (7) Owen, 1954; (8) Ohmasa and Mori, 1937.

show a seasonal pattern, although some leaf-fall is observed throughout the year (Fig. 4). Angiosperms exhibit a distinct autumnal peak as the weather cools, while the seasonal effect is much less for Pinus echinata (Olsen, private communication). For Pinus echinata farther north in

137

LITTER PRODUCTION IN FORESTS O F THE WORLD

Missouri it is reported (Anon., 1960) that 60% of litter-fall occurs in the 3 months from September to November. I n Cool Temperate forests seasonal patterns of litter-fall are often striking (Fig. 4), with autumnal cooling leading to more or less complete leaf-fall in deciduous species, as shown by the graphs for Dutch Quercus and mixed Angiosperms, and for Finnish. Betula. I n New Hampshire (Anon., 1932) it is reported that four-fifths of the mixed Angiosperm litter-fall over the seven months June to December occurs in October. Among Gymnosperms the pattern of deposition ranges from distinctly seasonal, as in the various stands of Pinus silvestris represented in Fig. 4, to irregular throughout the year, as is the case with Picea abies in the stands shown. Chandler (1944) has noted however that Picea abies planted in the north-eastern U.S.A. shows distinctly lower needlefall in spring and summer than at other times of year. In contrast irregular collections by Lindberg and Norming (1943) in Sweden showed high rates of needle-fall during late spring and early summer. Picea sitchensis in Wales showed a rather irregular pattern of needle-fall whether the total for the year was high (1946-47, with a very cold winter) or low (1947-48). There was however a slight tendency for spring and autumn maxima. Such a bimodal spring and autumn pattern of needlefall was more strongly exhibited by Japanese stands of Cryptmeria japonica, Chumaecyparis obtusa and Pinus densijora. TABLEXVII Speci,fic Differences in Leaf Fall within an Upland Oak Porest in Tennessee, U.S.A. Leaf fall as per cent of total for each species Late August Overstory Understory Miscellaneous

Quercw velutina Quercus coccinea Cornus sp. Oxydendrum arboreum Nyssa sylvatica Quercw alba Quercw montana Quercus falcata Quercw stellata Aeer rubrum Carya sp. Pinw echinata

* Balance of 10% overwinters, 5 % ~

~

0 1 3 1 63 0 0 0 0

0 0 0

MidOctober 3 12 72 99 37

16 4

43

71 32 52 20

Early December 97 87 25 0 0 74* 96 57 29

68 48 80

~~~~

falling by February and the remainder by April.

138

J. ROGER BRAY AND EVILLE OORHAM

The difference in leaf-fall periodicity of different species within mixed forest has been shown clearly by Blow (1955), working in upland oak forest of eastern Tennessee in the U.S.A. Table XVII shows the percentage of leaf-fall during the year 1948-49 which was collected by late August, mid-October and early December. From this table it appears that the overstory species lose their leaves later than the understory species, and that even within these groups there are distinct specific differences. The only species in this forest which retains many leaves (10%) through the winter is Quercus alba, although a few remain REF

40

u. LK

W I-

t

40-

MALAYAN FOREST ANISOPTERA

M A L A Y A N FOREST,

SHOREA

SHOREA

LAEVIS

.

I

11)

.20

QUERCUS

N E T H E R L A N D S FOREST

I

.

- 40

20-

-3

40

+ 60 z

*

0

T R EM U L A

40

20

20 0

J A N . FEB. MAR. APR. MAY JUNE JULY

AUG. SEPT OCT. NOV. DEC.

FIG.5. Seasonal litter-fall of individual species in mixed forests. (1) Mitchell (P.c.); (2) Witkamp and van der Drift, 1961.

on Quercus stelbta. Apparently overstory species may lose their leaves earlier than understory species, or even understory individuals of the same species, in other areas, as for example in mixed savannah forest of NW Thailand (Ogawa et al., 1961). Pig. 5 shows the marked variation in litter-fall periodicity among mixed Angiosperms in the Netherlands (Witkamp and van der Drift, 1961). Populus tremula and Quercus robur show distinct peaks of litterfall in October and November respectively. Betula verrucosa also reaches a peak in October, but litter-fall is quite heavy in August and September as well. Alnus glutinosa is unusual in exhibiting a bimodal pattern, with many green leaves being shed in July in addition to the major fall of brown leaves in October. Tadaki and Shidei (1960) found that in a

139

LITTER PRODUCTION IN FORESTS O F THE WORLD

3-year-old Ulmus parvifolia stand, there was a decreasing attached leaf weight of from 438 g/mz on 20 May t o 297 g/m2 on 27 July. This decrease, which equaled 31% of the total leaf canopy by weight and was the result of a loss of all the lower and more shaded leaves, may be more typical of young sapling stands than of forests of adult trees. Individual species in Equatorial forests may also show seasonal variation, t o a lesser degree than in most Temperate forests but t o a, greater degree than the Equatorial mixed forest as a whole. rig. 5 shows a more or less bimodal pattern for three Malayan species, which may be compared with the data for whole Malayan forests in Fig. 2. 2. Annual Variation Total litter-fall may differ greatly in different years, as illustrated by data, in Table XVIII. Owing to variation in sampling procedures these TABLEXVIII Annual Variation of Total* Litter Production in Stands Sampled for Four or More Years (muximumlminimum ratios) Species EQUATORIAL FOREST Mixed secondary forest, Malaya EVERGREEN ANGIOSPERM Nothofqua trunmta, New Zealand DECIDUOUS ANGIOSPERMS Acer sacccharum mixed forest, Canada Popdua plantations. U.S.S.R. Quercua alba, Q. velutina, U.S.A. Betula, Norway Fagua silvatim, 30 to >90 yra, Germany GYMNOSPERMS EVERGREEN PiGea abies, Germany Picea abies, Norway Picea abies. Germany Pinua silvestris, Germany Pin- silvestris, Germany

Authority

No. yeam Max/& sampled litter-fall

Mitchell (P.c.)

4

1.1

Miller and Hurst, 1957

4

2.7

Bray Sviridova, 1960

4 4

1.8 1.1

Hole and Nielsen (P.c.) Mork, 1942

5 6

1.4 1.3

Ebermayer, 1876

7

1-3

Krutzsch, 1869 4 2.4 Mork, 1942 6 I -4 Ebermayer, 1876 7 1*5 Krutzsch, 1869 4 1-3 Krutzsch, 1869; Ebermayer, 1876 7 2.2 Pinua nigra, Austria Bohmerle, 1906 12 5.1 Pinua reainosa, U.S.A. Lunt, 1951 21 3.4 Pseudotsuga menziesii, U.S.A. Dimock, 1958 6 3.2 * In the few caaes where leaf litter was measuredseparatelyresults were similar.

140

J. ROGER BRAY AND EVILLE GIORHAM

values are not always strictly comparable, and it may be observed that for Angiosperms sampled over 4 years the ratio of maximum to minimum annual litter-fall may be as high as 2.7 (in a New Zealand Nothofagus truncata forest). The highest ratio for Northern Hemisphere deciduous Angiosperms is 1.8 for Acer saccharurn mixed forest. Records over 7 years yield no higher ratios. Data for Gymnospermsinclude some much longer periods of comparison, and the greatest annual variation is recorded for a Pinus nigra stand observed over 12 years, the maximum/minimurn ratio being 5-1. I n general the longer periods of observation yield higher ratios, as expected. Over comparable periods in the North Temperate zone, Gymnosperms tend to show high ratios a little more frequently than Angiosperms. It is well known that the longevity of evergreen Gymnosperm leaves depends upon both internal and external conditions, so that fluctuations from year to year in the environment might be expected to affect the leaf-fall of such species more than that of deciduous Angiosperms whose leaves are necessarily shed every year. The single 4-year ratio for Equatorial forest is very low, at 1.1, as might perhaps be expected where climatic fluctuations are not very severe. The fall of non-leaf litter has not often been measured over several years, but annual variation in the four recorded cases is relatively high, maximum/minimum ratios over 4-6 years ranging from 1.9 for Nothofagus truncata (Miller and Hurst, 1957) to 11.9 for Quercus forest (Hole and Nielsen, private communication). Bray recorded a ratio of 5.8 for Acer saccharurn mixed forest and Mork (1942) a ratio of 3.7 for Betub. Such wide variation is to be expected, since the fall of branches is a very local phenomenon of great weight where it occurs. Among environmental factors mentioned as associated with abnormal litter-fall are storms (Picea abies - Bornebusch, 1937; exotic Gymnosperms in New Zealand -Will, 1959; Betula - Knudsen and MauritzHansson, 1939), which may have a very great effect on twig and branch litter-fall. Insect attack may also be important (Picea abies - Mork, 1942). Dryness has been remarked as a factor in high Picea abies needlefall (Bornebusch, 1932), but in contrast Witkamp and van der Drift (1961) observed that in the Netherlands three surface-rooting deciduous species exhibited distinctly low litter-fall during a dry summer, while a fourth deep-rooting species yielded only slightly less litter than usual. I n New Zealand, Miller and Hurst (1957) reported that hot dry summers favor flowering of Nothfagus truncata, and since leaf bud production for the following year’s foliage varies inversely with flowering, a lower subsequent leaf-fall would be expected. Cold temperatures may also increase litter-fall, for Owen’s (1954) data on Picea sitchensis in Wales show nearly twice as much needle deposition in 1946-47, when the

LITTER PRODUCTION I N FORESTS O F THE WORLD

141

winter was extremely cold, as in the following year. Dimock (1958) also reported for Pseudotsuga menziesii in Washington state, U.S.A., that in the 10 months following a severe cold spell litter-fall amounted to nearly three times that of previous years. Viro (1955) in Finland found that the cold September of 1949 led to early leaf-fall in Betula, while the warm September of 1947 led to late leaf-fall. Mork (1942) has suggested that cool temperatures during leaf development may lead to a light litter-fall later.

3. Age of stand I n no case has litter-fall been followed through a generation in an individual stand, so that all studies of age effects have been based on diverse stands of different ages. The results of such studies are also rather diverse. For example, Danckelmann (1887b) found litter-fall in Fagus silvatica forests on good soils increased about 20% between 30 and 90 years, after which a slight decline was manifested, also evident in Hartig’s data for stands aged 80 to 120 years (cited by Ebermayer, 1876). On the other hand Ebermayer (1876) reported a trifling decline in litter-fall from 30 years on, and Moller (1945) found that standing crop of leaves varies rather little between young (31-60 years), medium (61-119 years) and old (120-200 years) forests, the crops averaging 2.8, 2.5 and 2.7 t/ha respectively. I n the case of Pinus silvestris Danckelmann (1887a) reported a very slight decline in litter-fall from age 30 on, while data from Voronezh in the U.S.S.R. (compiled by Sonn, 1960) showed a more striking drop, from 2.5 t/ha at age 20 to 1.3 t/ha at age 100. Ebermayer (1876) however recorded litter-falls of 2.9, 3.0 and 3.6 t/ha at ages 25-50,50-75 and 75-100 years respectively. For Picea abies Danckelmann (1887b) found maximum litter-fall to occur in middle age (60-80 years), while Ebermayer (1876) observed a steady decline from 4.5 t/ha in stands less than 30 years old to 2.8 t/ha in stands over 90 years old. If all the above-mentioned European data are graphed together and medians taken for successive intervals, little variation in litter-fall is evident from 30 to more than 100 years of age. It would appear that there is no inherent tendency toward higher or lower litter-fall with increasing age, once the canopy becomes closed, but that environmental conditions (including biological agents such as insects or fungi) exert a decisive influence. One additional Australian case may be of interest, in which Eucalyptus regnans stands of 25 years (height 27 m, density 1013 t/ha), 55 years (43 m, 217 t/ha) and circa 200 years (76 m, 47 t/ha) were examined by Ashton (private communication). Total litter-fall amounted to 6.1,

142

J. ROGER BRAY AND EVILLE GORHAM

7.2 and 7.2 t/ha for these three ages, and leaf litter to 3.2, 3.7 and 3.7 t/ha. Litter-fall was thus the same under the 55- and 200-year forests, despite a more than four-fold difference in density. However, undergrowth litter contributed about a quarter of the total litter collected beneath the 200-year forest and only about a tenth of that beneath the 55-year stand.

VI. STANDING CROP OF LEAVES A. SEASONAL CHANGES -INTRINSIC Olsen (1948) found no consistent dry weight change in leaves of Fagus silvatica after the rapid increase accompanying foliage maturation in May. The terminal value for yellow leaves prior t o defoliation was a trifle low, but cannot be specified accurately from his graph. Mitchell (1936) reported rather marked increases in leaf dry weight of six eastern American Angiosperm trees until the beginning of July (and in the two Quercus species until the beginning of August). Carya (Hicoria) ovata leaves in full autumn color weighed only a littlk less than green leaves sampled a month before. Tamm (1955) observed a spring phase of rapid dry weight increase in needles of Swedish Pinus silvestris and Picea abies, followed by a dry weight loss of about 20% in June and July. Viro (1955)sampled different age groups of needles of the same species in Finland during August, and found the dry weight of 1000 Pinus silvestris needles to be 28.6 g in the first year, 28.9 g in the second, 29.5 g in the third, and 34.4 g in older needles ranging up to eight years in age. For Picea abies needles the correspondingweights were 3.51,4-76,4*58and 5.02 g. Weight changes upon and immediately prior to abscission have been investigated by Viro (1955) in a comparison of dry weights of 1000 green leaves or needles collected in August with dry weights of 1000 yellow leaves or needles collected partly from the tree and partly from the ground at the end of September. Weight loss associated with yellowing averaged 21% (range 19 to 23%) for the deciduous Angiosperms Betula (pubescens and verrucosa), Populus tremula, Alnus incana and Salix caprea. If yellow Pinus silvestris needles are compared with needles older than three years, the yellow ones weigh 44% less. For Picea abies the difference is 39%. However, it cannot be assumed that all yellow needles are more than three years old. If yellow and three-year needles are compared, the weight loss on yellowing would appear to be only 34% for both Pinus and Picea. Such a weight loss would still be very high, about one-third of the total dry matter. A much lower difference between brownish and green Picea pungens needles was observed at Minneapolis, U.S.A., on 29 May 1963. Sixty-eight brownish needles were

LITTER PRODUCTION I N FORESTS OF THE WORLD

143

carefully plucked and trimmed along with the same number of immediately adjacent green needles, and a dry weight loss of only 13% was associated with browning. However, these brown needles were not quike ready to fall, for they required a distinct tug with forceps to pluck them. It was also apparent in this case that many needles remained long after abscission upon the branches beneath, where they had become trapped when falling. Presumably considerable weight loss might take place before such needles reached a litter collector. No data could be found for dry weight changes associated with abscission of Warm Temperate or Equatorial tree leaves. However, in 1962, leaflets were sampled from a small leguminous tree, Tamarindus indica, growing in the greenhouse of the Botany Department at the University of Toronto. This tree has compound leaves with from twenty to twenty-six opposite leaflets which yellow and drop off separately. Collections were made of leaflet pairs in which one leaflet was yellow, possessed a well-formed abscission layer, and dropped when lightly touched; while the opposite leaflet was green, healthy and fimly attached. Leaflets were washed with detergent to remove dust and other contaminants, and oven-dried at 80°C for 12h. The results of four leaflet collections from 2 March to 27 April showed the weight of yellow leaflets to vary from 77 to 84% of that of green leaflets, with a mean value of 81%. There was a slight tendency for the relative weight of the yellow leaflets to decrease as the season advanced and solar radiation and greenhouse temperature increased. On average, the Tamarindzls leaves appeared to lose nearly one-fifth of their dry matter before falling. I n another test on 23 May 1963, leaves were collected from a Ficus elastica tree growing in the greenhouse of the Botany Department at the University of Minnesota. A very few of the leaves were yellow, and fell when touched. Ten of these were collected along with immediately adjacent green and healthy leaves of closely similar size. Four 21 mm leaf discs were cut with a cork borer at random (but excluding the midrib) from each pair of washed leaves. The green and yellow leaves of each adjacent pair were superimposed so that disc locations would be the same, and two discs were cut with the green leaf on top, and two with the yellow leaf on top. Discs were then oven-dried a t 80' C overnight. The weight of the yellow discs was 81% of the weight of the green discs, indicating (as in the case of Tamarindus) that Ficus leaves lose nearly one-fifth of their dry matter just prior to abscission. It is interesting that results are identical for two leaves of such very different type, the leguminous leaves being small and rather delicate, the fig leaves large, leathqry and with a waxy surface.

144

J . ROGER BRAY AND EVILLE GORHAM

B. SEASONAL CHANGES - EXTRINSIC Estimates of the percentage utilization of attached leaves of forest trees by primary consumers range from 5.9% for an Acer saccharum Betula lutea - Fagus grandifolia swamp forest to 10.6y0 for an upland Quercus forest (Bray, 1964). The mean of leaf co&.umption values in Lindquist (1938), Rothacher et al. (1954) and Bray is 7.5%. The rate of consumption, decomposition and respiration occurring in leaves immediately after leaf-fall has seldom been studied. Lindquist (1938) estimated that 15% of Fraxinus leaf litter and 10% of Ulmus leaf litter were eaten by earthworms within 1 week after completion of leaf-fall. Loss of weight by respiration may also be high, especially before the leaf becomes desiccated. Studies of weight loss with various techniques of litter collection are badly needed, since such weight loss may be higher than is usually realized, especially when collections are made from the ground. There, leaves are subject to animal consumption, and also, if they remain moist, to microbial decay. Bocock et al. (1960) found that Quercus petraea leaves on a highly acid moder humus layer lost only about 10% of their dry weight in two months. The loss was about 13% on a faintly acid mull. Leaves of Fraxinus excelsior, much more susceptible to decay, lost about 26% of their dry weight in one month on the moder and about 43% on the mull. On the mull earthworms removed many whole Fraxinus leaflets from the nylon nets in which they were held. Lunt (1935) observed that in 7 weeks after leaf-fall the dry weight loss of some eastern North American Angiosperm tree leaves ranged from 3 to 12%, being low for Quercus alba and Fagus grandifolia, and high for Acer rubrum and Cornus jiorida. Earthworms were excluded from the samples by the fine cloth mesh beneath the flat boxes in which the leaves were placed.

c. MAGNITUDE

OF LEAF CROPS

Table XIX shows weights of all attached leaves in various forests, collections being made during the period of maturity but preceding the phase of defoliation. Data are on an annual production basis, to allow comparison of Angiosperms and Gymnosperms. The magnitude of these values for yearly leaf production is similar to that of the cold temperate leaf litter data in Table IV, but variability in site conditions, plant populations and growing season climate makes direct comparison dificult. I n view of the extrinsic and intrinsic losses between the period of petiole attachment and of collection in litter traps, it is likely that the leaf crop values in Table XIX are somewhat higher than litter values for the same community. Direct studies of leaf production and leaf litter within one forest are needed.

TABLEXIX Dry Weight of Xtunding Crops of Leaves on an Annual Production Basis* Authority

Location

Lat. Long. ~~

Bartholomew et al., 1953

Congo, Yangambi

Jacobs, 1936 Burger, 1947 Moller, 1945 Burger, 1947 Zemljanickii, 1954 Ovington, 1962 Potts, 1939 Rothacher et al., 1954

Southern Australia Switzerland, Adlisberg Denmark Switzerland, Winterthur U.S.S.R., Zavetnoe and Kamyshin U.S.S.R. U.S.A., Saugus, Mass. U.S.A., Tennessee

Ovington, 1962 Bray, unpublished

U.S.A., Bethel, Minn. U.S.A., Bethel, Minn.

U.S.A., San Dimas Forest, California Tadaki and Shidei, 1960 Japan Switzerland, Aarburg Burger, 1940 Denmark Mhller, 1945 Boysen-Jensent Denmark Denmark Tadaki and Shidei, 1960 Japan Denmark Mhller, 1945 Jokela and Yliinen, 1956 Finland Ovington, 1962 N. Sweden Ovington, 1962 U.S.S.R.

Kittredge, 1944

1N

56N 47N 47N 43N 45N 46N

Plant community

Age (yr)

~

6 24E Musanga cecropioides pioneer forest Musanga cecropioides pioneer forest 8 Musanga cecropioides pioneer forest 17-18 Evergreen Eucalyptus giganteu forest Quercus robur 13 50 12E Quercus robur 9E Quercus robur - Fagus silvatica forest 65 43E Quercus robur - Fraxinus pennsylvanica 17 shelterbelt Quercus forest 22-200 71W Quercus forest Quercus forest, heavily cut 15 yrs before sampling 93W Quercus forest 57 93W Quercus ellipsoidalis - Q. alba closed forest 62 Quercus elliwoidalis open forest 43 Quercus ellipsoidalis -Q. macrocarpa savannah 60 Quercus chrysolepis 55

35N 136E Quercus mongolica v. grosseserrata 47N 8E Fagus silvatica Fagus silvatica Fagus silvatica Fraxinus excelsior 35N 136E Fraxinus rnandshurica Betula verruco8a Betula verruwsa, B e t d a alba Betula verrucosa Betula vemucosa

* All oven-dried except the U.S.S.R. t From Tadaki and Shidei (1960).

(air-dried) and possibly the Finnish semplee.

80 40

25 26 20-40

Leaf crop (metric tonsblur) 5.6 5.4 6.4 4.4 5.3 4.0 3.1 3.0

3.1 2-6 3.8

3.5 5.2 2-1 0.8 4.1 2.7

3.2

2-7 3-1 2-7 2.2 3.3 2.0 2-7 3.6

TABLEXIX - continued Authority

Ovington and Madgwick, 1959 Ovington, 1962 Tadaki et al., 1961 Tadaki and Shidei, 1960 Bray and Dudkiewicz, 1963 Bray and Dudkiewicz, 1963 Satoo et al., 1956 Auten, 1941 Tadaki and Shidei, 1960 Tadaki and Shidei, 1960 Tadaki and Shidei, 1960 Tadaki and Shidei, 1960 Ovington, 1957

Location

Lat. Long.

Plant community

England, Peterborough

53N

Japan Japan Japan U.S.A., Itasca, Minn.

Betula maximowicziana 43N 144E Betula platyphylla 35N 136E Betula ermanii 47N 95W Populw tremuloidee closed forest

Canada, Dorset, Ontario

45N

Japan, Hokkaido U.S.A., Illinois or Indiana

43N 40N

Japan Japan Japan Japan England, Brandon

35N 35N 35N 35N 52N

Kittredge, 1944

U.S.A., Stanislaus National Forest, California Maruyama and Satoo, 1953 Japan, Makibori, Iwate Mork, 1942 Norway, Hirkjolen Burger* Switzerland Shibamoto* Japan

38N 40N 62N

OW Betula verrucosa

79W Populua tremuloidea - Populua grandidentaka open forest 143E Populua davidiana: 88W Robinia pseudacacia on old field Bassafras albidum on old field 136E Alnus hirsuta 136E A l n w hirsukz v. sibirica 136E S a l k gracilistyla 136E Salix vulpina 1E Pinua dveatris, plantation Pin- silvestris, natural regeneration P i n w nigra, plantation P i n w ponderosa 120W 88W Pinua densiJora forest 10E Picea abies Larix decidua Larix leptolepis

*From Tadaki and Shidei (1960).

Age

(Yd

24-55

Leaf crop (metric tonslhalvr) 1.7

41

2.2 1.2 2.4 3.8

39

1.6

25-40 9 12

2.2 2.9 2.8 2.6 2.6 3.5 2-3 2.9 3.9 3.7 2.0

10

3-55 11-14 31 65-69 40

1.7

0.8 2.6 3.3

LITTER PRODUCTION IN FORESTS OF THE -WORLD

147

The mean leaf crop in t/ha/yr for closed canopy Angiosperm forests in Table XIX is 3.7 for eleven Quercus sites, 3.0 for three Fagus sites, 2.9 for two Salix sites, 2.6 for two Alnus sites, 2.5 for two Fraxinus sites, 2-5 for three Populus sites, and 2-4 for eight Betula sites. The high values for Quercus sites, which range from 2.6 to 5.3 t/ha and include four studies with values of 4-0 t/ha or more, are noteworthy. The Quercus leaf is usually thick and leathery, with a high lignin content (see Handley, 1954, Table XII). It is possible that Quercus forests are among the highest pPoducers of leaf litter within the Temperate regions; and the thick soil litter layers frequently found in Quercus forests may be related to high production of leaves as well as to the slow rate of leaf decay of many species in this genus. Some of the Betula and Populus data &retaken from more northerly areas of the Temperate zone, and are not directly comparable with the Quercus values. Leaf crops of Gymnosperms in Table XIX are similar to those of Angiosperms, with means in t/ha/yr of 2.8 for five Pinus sites and of 3.0 for two Larix sites. Of the nine Angiosperm and Gymnosperm genera with two or more sites, seven genera, including both Gymnosperm genera, have leaf crops between 2.5 and 3-0 t/ha/yr. Mean Angiosperm leaf production by genera is 2.8 t/ha/yr which is similar to the Gymnosperm mean of 2.9 t/ha/yr.

VII. LEAFLITTERAS AN INDEXTO NET PRODUCTION Although leaf litter represents an amount somewhat less than leaf production, owing to intrinsic and extrinsic weight losses prior to and probably also following abscission, it may still be useful as a guide to minimum levels of total net production. There is a logarithmic relationship between the foliage weight of some needle-leaved and a few broadleaved trees and tree stem diameter (Kittredge, 1944; Cable, 1958; Satoo, 1962). It is doubtful if this relationship is of much use in predicting net total production since (1) the constants in the regression equation obtained for a given stand will not, because of varying modes of competition, necessarily apply to other stands of the same species (Satoo, 1962) and (2) the relationship does not apply over the entire life-span of a tree since it is common for many Gymnosperms, including Pinus banksiana and P. contorta, to have a reduced, scraggly, broomshaped crown in later years, which is much smaller than the crown in the earlier part of the life span. Cooper (1960) demonstrated a linear relationship between needle air-dry weight and growth of stem basal area from stump sections in two plots of young (30 yr and 49 yr) Pinus ponderosa. There was no statistical difference in the regression equation for the two plots. Other studies of foliage in relation t o net production are desirable to determine whether a general regression equation can be

TABLEX X Net Leaf Production in Relation to Net Non-leaf Production* ~~

~

Authority

Species

Total

Leaf

Stem

(metric tons /ha/yr) Polster, 1950 Ovington, 1957 Ehwald, 1957 (data of Zhmerle) Polster, 1950 Moller, 1945 Ehwald, 1957 (data of Zimmerle) Polster, 1950 Polster, 1950 Polster, 1950 Moller et al., 1954 Ehwald, 1957 (data of Dieterich) Ovington and Madgwick, 1959 Polster, 1950 Ovington et al.. 1963 Polster, 1950 Satoo et al., 1956 Bray and Dudkiewicz, 1963 Moller, 1945 Nye, 1961 Bartholomew et al., 1953 Means

Below ground

Pinua silvestris Pinus silvestris Pinua mXvestria

9.4 12.4 8.8

1.5 3.3 3.0

6.3 7.2 4.6

( 1.6) 1.9 1.2

Picea abies Picea abies Picea abies

14.2 15.6 14.8

2.1 2.7 3.7

9.7 (10.4) 9.0

(2.4) (2.5) 2.1

Pseudotsuga rnenziesii Larix europaea Fagus silvatica Fagus silvatica Fagus silvatica

17.5 14.2 14.2 11.4 12.6

2.7 5.6 3.2 2.5 3.5

11-9 6.2 8.6 7.4 7.3

(2-9) (2.4) (2.4) 1.5 1.8

Betula verrucosa Betula verrucosa Quercus borealis Quercus robur Populua davidiana Populus trernuloidee Fraxinus excelsior Equatorial forest, Ghana Equatorial forest, Congo 7 evergreen Gymnosperms 10 deciduous Angiosperms 18 cold temperate forests 2 equatorial forests

9.0 8.6 9.1 10.4 7.8 10.9 10.9 24.3 31-5 13.2 10.5 11.7 27.9

1.7 2.0 3.5 2.4 2.2 3.8 2.7 7.0 9.5 2.7 2.8 2.9 8.3

5.1 5.2 4.1 6.3 4.3 5.3 6-7 14.7 19.2 8.4 6-0 7.0 16.9

2.2 (1.4) (1.5) (1.7) (1.3) (1.8) 1.5 2.6 2.8 2.1 1.7 1-9 2.7

* Figures in brackets were estimated (see text).

LITTER PRODUCTION IN FORESTS OF THE WORLD

149

derived which is applicable within species or taxonomic groups. It is likely that amount of foliage will be more easily correlated with current growth than with mean growth over the lifetime of a tree or forest. Table X X lists strdies from which leaf production can be compared to non-lea€and total net production. If below-ground production figures were lacking, a conversion factor of 0.2 times above-ground production was employed to estimate them (Bray, 1963). A value of 12.9 t/ha for stem plus below-ground production in Picea abies (Moller, 1945) was separated into 10.4 and 2.5.t/ha respectively by using the average ratio of stem to below-ground production in six Cool Temperate stands for which observed values were available. (The stand of Betula verrucosa on deep peat, sampled by Ovington and Madgwick (1959), was excluded because it fielded a widely aberrant ratio.) It is clear from Table XX that stem production (4.3 to 19-2 t/ha/yr) exceeds leaf production (1.5 to 9.5 t/ha/yr), which in turn exceeds root production (1.2 to 2.9 t/ha/yr). Total annual production ranges from 7.8 t/ha in a Japanese stand of Populm davidiana to 31.5 t/ha in a Congo forest. A summary of relative litter production by climatic zones together with various indexes of net total production is shown in Table XXI, with Arctic-Alpinevalues being taken as unity in the f i s t three columns, and Cold Temperate values in the last three columns. Bole production was estimated from data in Paterson (1956) by determining the range of C W (climate-vegetation-productivity) indexes for the climatic areas of the stands summarized in Table IV from Tables 16, 17, 18, 19, 20 and 21 of Paterson as follows: Arctic-Alpine,, CVP: 25-100; Cold Temperate, CVP: 100-500; Warm Temperate, CVP: 500-ca 3 000; Equatorial, CVP: 3 000-20 000. Mean bole production in m3/ha calculated from Table XXII of Paterson was 2.0 for Arctic-Alpine, 5.5 for Cold Temperate, 10-3 for Warm Temperate and 14.0 for Equatorial forest areas. Relative bole production for the four climatic areas was 1 to 2.7 to 5-1 to 7.0, which was similar to the ratios for leaf litter of 1 to 3.6 to 5.1 to 9.7, although the relative range for bole production from Arctic-Alpine to Equatorial forest was less than the range for leaf litter production. The other relative production values summarized in Table XXI are for Cold Temperate and Equatorial forest only, and show a range of from 2.3 to 3.0 (mean 2.6) for Equatorial forest over Cold Temperate values. Litter production of Equatorial forest varies from 2.7 (leaf litter) to 3.1 (total litter) times Cold Temperate litter production, which indicated that the use of litter data t o predict total production would slightly overestimate the differencein production between Cold Temperate and Equatorial forest. The ratios in Tables XX and XXI indicate that Equatorial forest is around two to three times as productive as Cold Temperate forest ; that Warm Temperate forest productivity F

C.E.R.

TABLEXXI Relative Litter, Bole and Total Organic Matter Production in Four Major Climatic Zones Total litter (from Table XI)

Leaf litter Bole production Total production Total broad-leaved Total needle-leaved tree production tree production (from (from (from (from Becking, 1962) (from Becking, 1962) Table XI) Paterson, 1956) Table XX)

Arctic-alpine*

1.o

1.0

1.0

-

-

-

Cold temperate?

3.5

3.6

2.7

1.0

1*o

1.0

Warm temperate

5.5

5.1

5.1

-

-

-

10.9

9.7

7.0

2.3

3.0

2.4

Equatorial

* Taken as unity in the firat three columns.

7 Taken aB unity in the last three columns.

151

LITTER PRODUCTION IN FORESTS OF T E E WORLD

is nearer t o Cold Temperate than to Equatorial forest productivity, and that Arctic-Alpine forest is less than half as productive as Cold Temperate forest.

TABLEXXII Ratios of Total, Non-leaf and Stem Production to Leaf Production Total Leaf

~~

Non-leaf Leaf

Stem Leaf ~

Cool temperate evergreen Gymnosperms Cool temperate deciduous Angiosperms Equatorial forest

4.9

3.9

3.1

3.7 3.3

2.7

2.1 2.0

2.4

Ratios of total net production, non-leaf production and stem production to leaf-production are given in Table XXII, for evergreen Gymnosperms and deciduous Angiosperms in the Cool Temperate zone and for Equatorial forests. It is evident that the Gymnosperms exhibit the highest ratios, and the Equatorial forests the lowest. The apparent lower efficiency of tropical tree leaves in producing non-leaf material probably reflects a high rate of respiration (relative to rate of photosynthesis) under higher tropical temperatures. The higher efficiency of temperate Gymnosperm as compared with Angiosperm leaves in producing non-leaf material may be owing to their much longer period (more than two years) of contribution to net production. The significance of the ratios in Table XXII is di%cult to assess because of the small number of studies available. Further productivity research is needed, especiallyin tropical and warm temperate forests, and until such work is done our estimates of net forest production cannot be securely founded. If the ratios of non-leaf to leaf production are approximately correct, then a comparison of them shows that leaf litter cannot be employed directly as an index to net forest production, since evergreen Gymnosperm leaves produce annually over 40 % more non-leaf material than do deciduous Angiosperm leaves, and over 60% more than Equatorial forest leaves. The use of leaf litter values as production indices would therefore greatly overestimate Equatorial forest production and underestimate production by evergreen Gymnosperms. The Cool Temperate Gymnosperms are about 25% more productive than the Angiosperms (13.2 t/ha as against 10.5 t/ha in Table XX). The evergreen nature of the Gymnosperm trees (allowing photosynthesis whenever temperatures are suitable), and the great weight of canopy carried (with several years’ needles present), are presumably responsible for their comparatively high levels of production. Ovington (1956) has

152

J. ROGER BRAY AND EVILLE GORHAM

found that in English forest plantations Gymnosperms may be several times as productive as Angiosperms on the same sites. I n a general comparison over wide areas, however, the tendency of Angiosperms to occur on better soils than Gymnosperms may lessen the differences in production brought about by differences in canopy weight and duration. REFEBENCES Aaltonen, V. T. (1948). “Boden und Wald”, 457 pp. Berlin and Hamburg: Paul Parey. Alway, F. J., Methley, W. J. and Younge, 0. R. (1933).Soil Sci. 36, 399-407. Distribution of volatile matter, lime and nitrogen among litter, duff, and leaf mold under different forest types. Alway, F. J. and Zon, R. (1930).J. For. 28, 715-727. Quantity and nutrient contents of pine leaf litter. Andersson, S. 0. and Enander, J. (1948). Svenska SkogsvF&en. Tidakr. 46, 265-270. Om produktionen av lovforna och dennas sammansiittning i ett mellansvenskt aspbesthd. AndrB, P. (1947). Svenska SkogmForen. Tidskr. 45, 122-131. Bamkens och mossornasfornaproduktion i ett barrskogsbest&nd. Anonymous (1932). Bull. N.H. agric. Exp. Sta. no. 262. Formation of forest soils: rate of deposition of litter. Anonymous (1960).Rep. cent. St. F o r . Exp. Sta., p. 12. Auten. J. T. (1941). Tech. Notes cent. St. For. Exp. Sta. 32, 9 pp. Black locust, pines and sassafras as builders of forest soil. Bartholomew, W. V., Meyer, J. and Laudelot, H. (1953).Publ. INEAC Ser. Sci., 57, 27 pp. Mineral nutrient immobilization under forest and grass fallow in the Yangambi (BelgianCongo) region. Becking, J. H. (1962).I n “Die Stoffproduktion der Pflanzendecke” (ed.H. Lieth), pp. 128-131. Stuttgart: Gustav Fischer Verlag. Ein Vergleich der Holzproduktion im gemiissigten und im tropischen Klima. Black, J. N. (1956). Arch.. Met. Wien, Ser. B, 7 , 165-189. The distribution of solar radiation over the earth’s surface. Blow, F. E. (1955).J. For. 53, 190-195. Quantity and hydrologic characteristics of litter under upland oak forests in emtern Tennessee. Bocock, K. L., Gilbert, O., Capstick, C. K., Twinn, D. C., Waid, J. S. and Woodman, M. J. (1960).J. Soil Sci. 11, 1-9. Changes in leaf litter when placed on the surface of soils with contrasting humus types. Bohrnerle, K. (1906). “Die Streuvemuche im grossen Fohrenwalde”. 22 pp. Vienna: W.Frick. Bonnevie-Svendsen, C. and Gjems, 0. (1957).Medd. Norske Skogsfors~ksv.48, 111-174. Amount and chemical composition of the litter from larch, beech, Norway spruce and Scots pine stands and its effect on the soil. Bornebusch, C. H. (1937). pOr8tl. Fors0bv. Danm. 14, 173-176. Iagttagelser over rpldgranens naalefald. Boysen-Jensen, P. (1930).Forstl. Forsoksv. Danm. 10, 365-391. Undersogelser over Stofproduktionen i yngre Bevoksningeraf Ask og Bog. 11. Brauns, F. E. and Brauns, D. A. (1960). “The Chemistry of Lignin”, Suppl. Vol., 804 pp. New York and London: Academic Press.

LITTER PRODUCTION I N FORESTS O F T E E WORLD

163

Bray, J. R. (1964). Ecology 45, 165-167. Primary consumption in three forest canopies. Bray, J. R. (1963). Canad. J . Bot. 41, 66-72. Root production and the estimation of net productivity. Bray, J. R. and Dudkiewicz, L. A. (1963). Bull. Torrey bot. Cl. 90, 298-308. The composition, biomass and productivity of two Populus forests. Broadfoot, W. M. and Pierre, W. H. (1939). Soil Sci. 48, 329-348. Forest soil studies: I. Relation of rate.of decomposition of tree leaves to their acid-base balance and other chemical properties. Burger, H. (1925). 2. Forst- u. Jagdw. 57, 473-482. Die Transpiration unserer Waldbiiume. Burger, H. (1940). Mitt. Schweiz. ZentAnst. forstl. V e r a u c h . 21, 307-348. Holz, Blattmenge und Zuwachs. IV. Ein 80 j h i g e r Buchenbestand. Burger, H. (1947). Mitt. Schweiz. ZentAnet. forstl. V e r s u c h . 25, 211-279. Holz, Blattmenge und Zuwachs. VIII. Die Eiche. Cable, D. R. (1958). Forest Sci. 4, 45-49. Estimating surface area of ponderosa pine foliage in central Arizona. Chandler, R. F., Jr. (1941). J . Amer; SOC.Agron. 33, 859-871. The amount and mineral nutrient content of freshly fallen leaf litter in the hardwood forests of central New York. Chandler, R. F., Jr. (1944). Proc. Soil Sci. SOC.Amer. 8, 409-411. Amount and mineral nutrient content of freshly fallen needle litter of some north-eastern conifers. Claudot, M. (1956). FAO/SCMIEUI7-B. Influence de 1’Eucalyptussur l’evolution des sols au Maroc. (Mimeographed.) Coldwell, B. B. and DeLong, W. A. (1950). Sci. Agric. 30, 456-466. Studies of the composition of deciduous forest tree leaves before and after partial decomposition. Cooper, C. F. (1960). Ecol. Monogr. 30,129-164. Changes in vegetation, structure, and growth of south-westernpine forests since white settlement. Crosby, J. S. 1961. Tech. Pup. cent. St. For. Exp. Sta. 178, 10 pp. Litter-and-duff fuel in shortleaf pine stands in Southeast Missouri. Curtis, J. T. (1959). “The Vegetation of Wisconsin”, 657 pp. Madison, Wisconsin: University of Wisconsin Press. Danckelmann, B. (1887a). 2. Forst- u. Jagdw. 19, 457-466. Streuertragstafel fiir Kiefern bestlinde. Danckelmann, B. (1887b). 2. Forst- u. Jagdw. 19, 577-587. Streuertragstafel fiir Buchen- und Fichtenhochwaldungen. Dimock, E. J. (1958). Northw. Sci. 32, 19-29. Litter fall in a young stand of Douglas fir. Eberinayer, E. (1876). “Die gesamte Lehre der Waldstreu mit Rucksicht auf die chemische Statik des Waldbaues”, 116 pp. Berlin: Julius Springer. Ehwald, E. (1957). Dtsch. Akad. Landwirtschaftwiesenschaften Berlin. 6 , 1-56. ifber den Nlihrstoffkreislaufdes Waldes. Gaffron, H. (1946). I n “Currents in Biochemical Research” (ed.D. E. Green), pp. 25-48. New York: Interscience. Photosynthesis and the production of organic matter on earth. Glesinger, E. (1949). “The Coming Age of Wood”, 279 pp. New York: Simon and Schuster. Handley, W. R. C. (1954). Bull. For. Corm. London, 23, 115 pp. Mull and mor formation in relation to forest soils.

164

J. ROGER B R A Y A N D EVILLE GOREAM

Hatch, A. B. (1955). LeafE. For. Bur., Canberra 70, 18 pp. The influence of plant litter on the jarrah forest soils of the Dwellingup region, Western Australia. Hesselman, H. (1925). Medd. Skogsforsoksanst 22, 169-552. Studier over barrskogens humustacke, dess egenskaper och beroende av skogsvhrden. Heyward, F. and Barnette, R. M. (1936). Bull. Fla. agric. Exp. Sta. 302, 27 pp. Field chmacteristics and partial chemical analyses of the humus layer of longleaf pine forest soils. Jacobs, M. R. (1936). Commonwe&h Por. Bur. Bull. (Canberra, Australia) 18, 78 pp. The primary and secondary leaf-bearing systems of the eucalypts and their silvicultural significance. J&6, Z. (1958). Erd4szettudodnyi Koslednyek, Sopron 1, 151-162. Alommennyis6gek a magyar erdokben. Jenny, H., Ge~sel,S. P. and Bingham, F. T. (1949). Soil Sci. 68,419-432. Comparative study of decompositionrates of organic matter in temperate and tropical regions. Joffe, J. (1949). “Pedology”, 662 pp. New Brunawick, N.J.: Pedology Publications. Jokela, E. and Yliinen, J. (1956). Metsut. Aikakausl. 73, 131-132. Koivikoiden lehtkadon miiliriistli. Kendrick, W. B. (1959). C a d . J . Bot. 37,907-912. The time factor in the decomposition of coniferous leaf litter. Kittredge, J. (1940). J. For. 38, 729-731. A comparison of forest floors from plantations of the same age and environment. Kittredge, J. (1944). J . For. 42, 905-912. Estimation of the amount of foliage of trees and shrubs. Knudsen, F. and Mauritz-Hansson, H. (1939). Svenska SkogsvForen. Tidskr. 37, 339-347. Om produktionen av lovforna och dennaa sammrtnslittning i ett mellansvenskt bj orkbesthd. Krutzsch. H. (1869). Tharandt.forstl. Jb. 19, 19S227. Untersuchungen iiber die Waldstreu. Lespeyres (1898). 2. Forst- u. Jagdw. 30, 521-537. Der Einfluss der Streuniitzung auf den Holzwuchs in den Kiefernbestlinden des nordostdeutschen Flachlrtndes. Laudelot, H. and Meyer, J. (1954). Act@ et Comptes Rendua du V eCongrbs International de la Science du Sol 2, 267-272. Les cycles d’6lBments mineraux et de mrttihre orgaaique en for6t Bquatorialecongolake. Levina, V. I. (1960). Bot. Zh. 45, 41&423 (cited in For. Abstr. 21, 532, 1960. Determination of the amount of annual litter fall in two types of pine forest on the Kola Peninsula). Lindberg. S. and Norming, H. (1943). Suenska SkogsvForen. TicEskr. 41, 353-360. Om produktionen av barrforna och dennas sammrtnslittning i ett granbesthd invid Stockholm. Lindquiat, B. (1938). Actaphyzogeogr.Suec. 10,273 pp. Dalby Soderskog. Lunt, H. A. (1935). J . For. 33, 607-609. Effect of weathering upon dry matter and composition of hardwood leaves. Lunt, H. A. (1951). Proc. Soil Sci. SOC.Amer. 15, 381-390. Liming and twenty years of litter raking and burning under red (and white) pine. Lutz, H. J. and Chandler, R. F., Jr. (1946) “Forest Soils”, 514 pp. New York: John Wiley. Maruyarna, I. and Satoo, T. (1953). Bull. For. Exp. Sh. Meguro 65, 1-10. Estimation of the amount of foliage of trees and stands. Report 1. On the Akamatu of Iwate District. Mayer-Krapoll, H. (1956). “The Use of Commercial Fertilizers -particularly

LITTER PRODUCTION I N FORESTS OF T E E W O R L D

156

Nitrogen - in Forestry”, 11 1 pp. Translated and published by Nitrogen Div., Allied &em. and Dye Corp., New York. Melin, E. (1930). Ecology 11, 72-101. Biological decomposition of some types of litter from North American forests. Metz, L. J. (1952). Proc. Soil Sci. SOC.Amer. 16, 38-41. Weight and nitrogen and calcium content of the annual litter fall of forests in the South Carolina Piedmont. Mikola, P. (1954). Commun. Inst. for. Finl. 43, 1-50. Kokeellisiatutkimuksia metsiikarikkeiden hajaatumisnopeudesta. Miller, R. B. and Hurst, F. B. (1957). N.Z. For. Res. Notes 8, 14 pp. The quantity and nutrient content of hard beech litter. Mitchell, H. L. (1936). Black Rock For. Pap. 1, 30-44. Trends in the nitrogen, phosphorus, potassium and calcium content of the leaves of some forest trees during the growing season. Moller, C. M. (1945). Forstl. Forseksv. Danm. 17, 287 pp. Untersuchungen iiber Laubmenge, Stoffverlust and Stoffproduktion des Waldes. Moller, C. M. (1947). J . For. 45, 393-404. The effect of thinning, age and site on foliage, increment and loss of dry matter. Moller, C. M., Miiller, D. and Nielsen, J. (1954). F o r d . Forsokew. Danm. 21, 273-301. Respiration in stem and branches of beech. Mork, E. (1942). Me&. Norske Skogfors~ksv.29, 297-365. Om strafallet i v b e skoger. Miiller, P. E. (1887). “Studien uber die natiirlichen Humusformen und deren Einwirkung auf Vegetation und Boden”, 324 pp. Berlin: Julius Springer. Navarro de Andrade, E. (1941). J . Hered. 32, 215-220, and 240. T h e eucalyptus in Brazil. Nye, P. H. (1961). Plant & Soil 8, 333-346. Organic matter and nutrient cycles under moist tropical forest. Nykvist, N. (1963). Stud& for. Suec. 3, 31 pp. Leaching and decomposition of water-solubleorganic subst?nces from different types of leaf and needle litter. Ogawa, H., Yoda, K. and Kira, T. (1961). Nature and Life in8.E. A& 1,21-157. A preliminary survey of thle vegetation of Thailand. Ohmasa, M. and Mori, K. (1937). Bull. For. Exp. Sta. Tokyo-Fu 3, 39-101. The amount of fall and decompositionof the leaf litter of the forest trees of Japan. Olsen, C. (1948). C.R. Lab. Carlsberg, Ser. Chim., 26, 197-230. T h e mineral, nitrogen and sugar content of beech leaves and beech leaf sap at various times. Ovington, J. D. (1956). New Phytol. 55, 289-304. The form, weights and productivity of tree species grown in close stands. Ovkgton, J. D. (1957.)Ann. Bot., Lond. N.S. 21, 287-314, Dry matter production by Pinus sylvestris L. Ovington, J. D. (1962). Adv. Ecol. Res. 1, 103-192. Quantitative ecology and the woodland ecosystem concept. Ovington, J. D., Heitkamp, D. and Lawrence, D. B. (1963). Ecology 44, 52-63. Plant biomass and productivity of prairie, savanna, oakwood and maize field ecosystemsin central Minnesota. Ovington, J. D. and Madgwick, H. A. I. (1959). Plant & Soil 10, 271-283. The growth and composition of natural stands of birch. 1. Dry matter production. Ovington, J. D. and Pearsall, W. H. (1956). Oikos 7, 202-205. Production ecology 11.Estimates of average production by trees. Owen, T. H. (1954). Forestry 27, 7-15. Observations on the monthly litter-fall and nutrient content of Sitka spruce litter. Paterson. S. S. (1956). “The Forest Area of the World and its Potential Pro-

156

J. ROGER B R A Y AND EVILLE GORHAM

ductivity”, 216 pp. and map. Sweden: Dept. of Geography, the Royal University of Goteborg. Patterson, S. S. (1961). Medd. SkogsforsknInst. Stockh. 50, 1-145. Introduction to phyochorology of Norden. Perina, V. and Vintrova, E. (1958). Sborn. Esl. Akad. zeme‘d. (Lesn.)4, 673-688. The effect of fallen pine-needles on the humus conditions of $he pinewoods on the Pleistocene sands (Engl. summary). Pirie, N. W. (1953). Chem. Ind. Lond., No. 19, 442-445. The efficient use of sunlight for food production. Pirie, N. W. (1958). Sci. News, Harmondsulorth 49, 17-38. Unconventional production of foodstuffs. Pirie, N. W. (1961). Econ. Bot. 15, 302-310. Progress in biochemical engineering broadens our choice of crop plants. Pirie, N. W. (1962). J . R . stutiet. SOC.Ser. A. 125, 399-417. Future sources of food supply: scientific problems. Polster, H. (1950). “Die physiologischen Grundlagen der Stofferzeugung im Walde”, 96 pp. Munich :Bayerischer Landwirtschaftsverlag. Potts, S. F. (1939). J . For. 37, 922-923. A method for determining the quantity of foliage per acre of woodland. Remezov, N. P. and Bykova, L. N. (1953). Pochvovedenie 8, 28-41. (Cited in For. dbstr. 16, 334-335, 1955. Uptake and cycle of nitrogen and ash elements in aspen stands.) Richards, P. W. (1957). “The Tropical Rain Forest”, 450 pp. Cambridge Univ. Press. Rothacher, J. S., Blow, F. E. and Potts, S. M. (1954). J . For. 52, 169-173. Estimating the quantity of tree foliage in oak stands in the Tennessee Valley. Satoo, T. (1962). J . Jup. For. SOC.44, 267-272. Notes on Kittredge’s method of estimation of amount of leaves of forest stand. Satoo, T., Kunugi, R. and Kumekawa, A. (1956). Bull. Tokyo Univ. For. 52, 33-58. Materials for the studies of growth in stands. 3. Amount of leaves and production of wood in an aspen (Populua davidiana) second growth in Hokkaido. Scott, D. R. M. (1955). Yule Sch. For. Bull. 62, 73 pp. Amount and chemical composition of the organic matter contributed by overstory and understory vegetation to forest soil. Sims, I. H. (1932). J . For. 30, 90-91. Litter deposition and accumulation ,iathe pine-oak type of the Southern Appalachians. sjors, H. (1954). Actaphytogeogr.w e c . 34, 135 pp. SlBtterSngar i Grangiirde Finnmark. Smith, A. H. (1958). “The Mushroom Hunter’s Field Guide”, 197 pp. Ann Arbor: University of Michigan Press. SOM, S. W. (1960). “Der Einfluss des Waldes auf die Boden” (trans.P. Kundler), 166 pp. Jena: Gustav Fischer Verlag. Stoate, T. N. (1958). Rep. For. Dep. W . Austral. 25, 25. Silvicultural and soils research. A. Karri silviculture. Sviridova, L. K. (1960). Pochvovedenie 4, 68-73. (Trans.in Soviet Soil Sci. 1961, 401-405. Role of improvement cuttings in raising forest soil fertility.) Tadaki, Y . and Shidei, T. (1960). J . Jap. For. SOC.42, 427-434. Studies on productive structure of forest. 1. The seasonal variation of leaf amount and the dry matter production of deciduous sapling stand ( U l m w parviflora). Tadaki, Y . , Shidei, T., Sakasegawa, T. and Ogino, K. (1961). J . Jap. For. Sot.

LITTER PRODUCTION I N FORESTS OF THE WORLD

167

43,19-26. Studies on productive structure of forest. 2. Estimation of standing

crop and some analyses on productivity of young birch stand ( B e t h PktYPhYW. T a m , C. 0. (1955). M a . SkogsforsknInst.Stockh. 45, no. 5, 34 pp. Studies on forest nutrition I. Seasonal variation in the nutrient content of conifer needles. T a m , C. 0. (1958). In “Forest Fertilization, a New Era”, pp. 2-11. h e r . Potash. Inst. Handbook. Forest fertilization in Europe, its research and practices. Tarrant, R. F., Isaac, L. A. and Chandler, R. F., Jr. (1951). J . For. 49, 914-915. Observations on litter fall and foliage nutrient content of some Pacific Northwest tree species. Viro, P. J. (1955). Commun. Inst.for. Finl. 45, 65 pp. Investigations on forest litter. Walter, H. and Lieth, H. (1960). Klimadiagramm-Weltatlas. Jena: Gustav Fischer. Weck, J. (1955). “Forstliche Zuwachs-und Ertragskunde”, 160 pp. Radebeul and Berlin : Neumann Verlag. White, D. P. (1954). Proc. Soil Sci. SOC.Amer. 18, 326-330. Variation in the nitrogen, phosphorus and potassium contents of pine needles with season, crown position, and sample treatment. Wiedemann, E. (1951). “Ertragskundliche und waldbauliche Grundlagen der Forstwirtschaft”, 346 pp. Frankfurt-aml?lctinMain: J. D. Sauerlander’sVerlag. Will, G. M. (1959). N.Z. J . ugric. Rea. 2, 719-734. Nutrient return in litter and rainfall under some exotic conifer stands in New Zealand. Witkamp, M. and van der Drift, J. (1961). Plant &Soil 15,295-31 1. Breakdown of forest litter in relation to environmental factors. Wright, T. W. (1957). Forestry 30, 123-133. Some effects of thinning on the soil of a Norway spruce plantation. Zemljanickii, L. T. (1954). Pochwovedenie 12, 30-35. (Cited in For. Abstr. 16, 469, 1965. Quantity and ash composition of litter in artificial forest plantations in the zone of chestnut soils.)

F2

C.E.R.

Forty Years of Genecology J

. HESLOP-HARRISON

Department of Botany. University of Birmingham. England

.................................... ...................................................... ............................................ ..............................................

I. The Scope and Concepts of Genecology 159 A Introduction 159 B SomeBasicPropositions 160 C EcotypesorEcoclines? 162 D . Genecologicd Categories ............................................ 164 E . GenecologicalTechnique ............................................ 168 F Geneoological Differentiation: Some Illustrative Examples 173 187 G. GeneoologicdDifferentiation: Generalizations .......................... H Genetic Basis of Ecotypic Differences .................................. 189 I1. Evolutionary Aspects of Genecological Differentiation ...................... 193 A. Introduction: the Origin, Storage and Release of Variation ................ 193 B ModesofSelection.................................................. 196 C. Versatile Reproductive Systems...................................... 201 D Isolation and Genecological Differentiation ............................ 204 E . Monotopic and Polytopic Origin and the Retention of Racial Identity ......210 F Plasticity, Genetic Assimilation and Conditioning ...................... 213 111. Physiological Aspects of Genecologioal Differentiation ...................... 217 A Introduction ...................................................... 217 B . Edaphic Adaptation ................................................ 219 C Adaptation to Soil Moisture Stress .................................... 223 D . Adaptation to Light Intensity ........................................ 224 E . Adaptationtoclimate .............................................. 227 I V . Conclusions .......................................................... 237 References................................................................ 240

. . . . .

..............

. . . . .

I. THE SCOPEAND CONCEPTSOF GENECOLOGY A . INTRODUCTION The term genecology was applied originally by Turesson (1923)to the study of the infraspecific variation of plants in relation to environment . Genecology was in this way established as a branch of ecology. comparable with and in various respects complementary to autecology and synecology. the former concerned with the environmental relations of individual species and the latter with those of plant communities. Turesson’s early publications (1922a. 1922b. 1923. 1925. 1930)illustrate the synthesis of approaches -genetical. ecological and taxonomical-

160

J. HESLOP-HARRISON

which he accommodated within the discipline of genecology. The essential unity of these early researches lay in their concentration upon the species as “a genetically complex community, the distribution and the composition of which is largely determined by ecological factors and the genotypical constitution of the individuals . . .” (Turesson, 1923). Their aims, in brief, were to lay bare the patterns of infraspecific” ecological adaptation, and elucidate the mechanisms whereby this adaptation was achieved. The separate identity of genecology depends upon the preservation of this combination of aims; like all synthetic disciplines, it breaks up into its components when purposes different from or more specific than those constituting the original frame of reference are introduced. Thus genecology merges into taxonomy when the primary aim is to systematize for classificatory purposes patterns of infraspecific variation ; into genetics, when the mechanisms of variation and selection form the main targets of study; and into plant physiology when it is the physical responses of the organism to the environment which are of interest. There seems still to be a need to preserve the identity of genecology in the original Turessonian sense, since for the understanding of an important phase of micro-evolution the synthesis is indispensable, in the same way that a synthesis - on a broader canvas -was essential in the development of the argument of “The Origin of Species”. It is the purpose of this article to assess the present position of genecology and to attempt to distinguish a few of the general principles that appear to have emerged since the pioneer days. I n doing this, facts and concepts have been drawn not only from researches which have been expressly oriented towards genecology, but also from related fields, where they contribute to the genecological synthesis in the broad sense.

B. S O M E

BASIC PROPOSITIONS

The basic propositions of Turessonian genecology may be summarized as follows: (1) Wide-ranging plant species show spatial variation in morphological and physiological characteristics ; (2) much of this infraspecific variation can be correlated with habitat differences; (3) to the extent that ecologically-correlatedvariation is not simply due to plastic response to environment, it is attributable to the action of natural selection in moulding locally adapted populations from the pool of genetical variation available to the species as a whole. Turesson’s own work, as exemplified particularly in the papers of 1922, 1925 and 1930, was devoted in large part to the substantiation of propositions (1) and (2), and to showing that ecologically-correlatedinter-population variation was

* “Infraspecific variation” in this article means variation below the level of the average Linnaeaa species (See the discussion pages 165 to 167).

FORTY YEARS OF QENECOLOGY

161

commonly genetically based rather than dependent upon the direct modification of individuals. The experimental methods employed -involving mostly sampling from wild populations and comparison under standard conditions of cultivation -have been reviewed and discussed many times, and need not be described further here. The results of Turesson’s studies were satisfactorily gonclusive. To them may now be added a mass of evidence accumulated subsequently by others, so that it may be accepted as an established fact that plant species in general do show genetically-based ecological differentiation. If this be generally conceded, it cannot be said that there has been as satisfactory an agreement on the nature of the patterns of infraspecific ecological differentiation in the higher plants. Turesson’s method of population sampling and comparative cultivation led him to the view that species commonly constituted a mosaic of populations each adapted to characteristic habitats, each habitat-form being more or less distinct -that is to say, separated by a variational discontinuity -from other (e.g. Turesson, 1936). This conclusion led in turn to the concept of the ecotype, now so fimly identified with Turesson’s work that genecology itself has been called “the doctrine of the ecotype” (Faegri, 1937). Perhaps the most specific statement of the doctrine appears in Turesson’s second paper (192210): “The mass of genetically distinct forms which make up the Linnaean species do not distribute themselves indiscriminately over an area comprising different types of localities, but, on the contrary, are found in nature to be grouped into different types, each confined to a definite habitat. Further, these ‘ecotypes’ do not originate through sporadic variation preserved by chance isolation; they are, on the contrary, to be considered as products arising through the sorting and controlling effect of the habitat-factors upon the heterogeneous species-population.” According to Faegri (1937), Langlet (1934) was the first to point out that, since most of the major habitat factors vary spatially in a continuous and not “stepped” manner, graded rather than discontinuous variation is to be expected in a wide-ranging species as a consequence of adaptation to habitat. There can be no doubt that Turesson’s methods, involving as they generally did sampling from rather remote populations occupying markedly different habitats, were not of a type likely to detect continuous spatial variation did it exist. More recent studies using adequate sampling methods have shown that adaptative variation may be either continuous or discontinuous, and it may now be said that the important questidns concern not the existence of different patterns of variation but the forces acting within species to generate them, matters considered at length in Section 11. Nevertheless, the lively discussion of the relative significance of continuous and discontinuous variation has

162

J. HESLOP-HARRISON

continued up to the present (Clausen, 1951; Clausen et al., 1940, 1948; Faegri, 1937; Gregor, 1939, 1944, 1946a, b, 1956; Gregor and Watson, 1954, 1961; Langlet, 1959, 1963; Sinskaia, 19311, and the immediately succeeding paragraphs survey some of the background. C. ECOTYPES OR ECOCLINES? For reference to graded variation within populations or population systems of organisms the useful word cline is available, introduced by Huxley ( 1938) as.an “auxiliary taxonomic concept”. “Cline” unqualified implies simply the existence of a variational gradient; the prefix ecoindicates that the gradient can be correlated with an environmental variable. The well known researches of Gregor and his associates on Plantago maritima (Gregor et at., 1936, 1950; Gregor, 1938, 1939) revealed several types of graded inter-population variation, including some which was evidently ecologically correlated. I n the discussion of his paper of 1938 and in his review of 1944, Gregor considered various genetical, ecological and taxonomical aspects of such variation, and subsequent work has confirmed the general validity of most of his conclusions. The cline concept is extensively used by Gregor in the recording of graded variation. Geographical clines (which may have an ecological basis where the variation is correlated with regional changes of climate) he refers to as topoclines; where the variation can be associated with the stages of an ecological sequence (which may never appear as an actual physical range of habitats to be encountered in a transect between two geographical points), it is referred to as an ecocline. Gregor emphasizes the need always to consider ecological variation in relation to the habitat range of species viewed as a continuum rather than as a complex of discrete nonoverlapping environments ; this leads him t o the view that the ecotype must be imagined subjectively “as a certain range of variation on a genotypic gradient selectively developed in response to a climatic, edaphic, biotic, or for that matter any environmental gradient we may choose to examine”. The Plantago data contain examples of quantitative characteristics - growth habit, scape length, spike density - which follow the environmental gradient waterlogged-mud drained “saltings”. For reference purposes Gregor divided the populations into three ranges: predominantly decumbent, ascending, and erect, which he referred to as ecotypes. This treatment illustrates his attitude towards the use of the term ecotype : that it should be employed in a subjective sense to refer to “certain populations belonging to an ecocline”. The definitions of certain terms concerned with infraspecific variation advocated by Gregor in 1939 were:

FORTY YEARS O F GENECOLOGY

163

“Cline, any gradation in measurable characters. Topocline, a cline following a geographical gradient. Ewcline, a cline apparently correlated with an observable ecological gradient.

Topotype, a population in a geographical region possessing characters differing from those of another region. A topotype may be extraclinal if it does not fall within a geographical gradient in character expression, or intraclintll if it has reference to a particular range on a geographical gradient. Ecotype, a particular range on an ecocline.” I n Gregor’s usage of this period, then, ecoclines, rather than ecotypes are to be regarded as representing the significant “ecological subcategories” of plant species. The genecological work of the Carnegie group (Clausen et al., 1940, 1948, Clausen and Hiesey, 1958a) reveals a notably different outlook upon infraspecific patterns of ecological differentiation. The study of variation in Potentilla glandulosa along a transect across central California recorded in the publication of 1940 suggested a comparatively simple pattern, the populations falling into morphologicallydistinguishable climatic ecotypes correspondingto the taxonomic subspeciestypica, reJEexa,hanseni and nevadensis. Although these are in a sense ecoclinal in that their distribution is related to a climatic gradient, they were interpreted by Clausen et al., as population complexes replacing each other geographically in rather an abrupt manner, not at all comparable with the “ranges on ecoclines” envisaged by Gregor. An emphasis upon the distinctness of these major intraspecific groupings is a feature of the earlier publication; it also appears in that of 1948 devoted to the Achillea millefolium complex, although here a somewhat different situation is described. The populations sampled across the same California transect were regarded as belonging to eleven distinct climatic races (“or ecotypes”), four of the hexaploid A . borealis and the remainder of the tetraploid A . lanulosa. The races of A. lanulosa from the western slope of the Sierra Nevada are said to form “a graded altitudinal series that shows the close interrelationships between the physiological characteristics of the races and their environments”; they thus form an ecocline in the sense of Gregor. However, while for Gregor the ecotype is no more than an arbitrary section of an ecocline, the climatic race within Achillea is considered by Clausen et al. to be a coherent entity, distinct from others. It is described as follows: “Each climatic race consists of many local populations possessing in common those characteristics essential for survival in their particular environmental zone. Similarity in essential characteristics does not preclude individual variability. No

164

J. HESLOP-HARRISON

two populations are identical in composition, and they are composed of individuals scarcely two of which belong to the same biotype. Although the individuals and the populations vary, the races are statistically and reactionally distinct.” (Clausen et al., 1948, p. 122.) The viewpoints of Gregor’s and Clausen’s school represent, in a sense, the poles of opinion in the period prior to 1950. Before reviewing more recent work, we will consider some methodological problems concerned in the analysis and interpretation of infraspecific diversification.

D.

GENECOLOGICAL CATEGORIES

Turesson’s view of the nature of infraspecific ecological differentiation led him eventually to treat the ecotype as a form of classificatory category, and to appIy a nomenclatural system comparable with that of orthodox taxonomy. I n this, the ecotypes were named according to their habitat predilections -oecotypus alpinus, oecotypus arenarius etc. (Turesson, 1925). There is undoubtedly a strong psychological compulsion to attempt to systematize knowledge of infraspeciflc ecological variation by this kind of classificatory approach, and the existence of orthodox taxonomy as a model encourages the direct transfer of methods. Yet it is now quite apparent that systems based upon the construction of discrete classes are inadequate to accommodate the diversity of genecological data, even if the additional flexibility of a. hierarchical arrangement (“ecotypes within ecotypes”) is permitted. The view that “ecoclines represent the ecological subcategories” has been expressed by Gregor (1944). It is obvious that the “subcategory” envisaged here is different in kind from any orthodox taxonomic unit, since the basis of definition is not character correlation, within a real or imagined population, as it is for example with all the categories of nomenclatural taxonomy, but a selected trend of character variation viewed in relation to an extrinsic factor, either position on the earth’s surface or ecological distribution. There is a sense in which an ecocline can be looked upon as a classificatory unit, when it denotes “a series of habitat populations showing genotypic gradation related to a particular environmental gradient” (Gregor, 1944) ; this usage would bring to-. gether a group of populations because of a particular relationship with each other in the same way that a group of populations may be classed together in a regional subspecies. The criteria for the grouping together are, however, obviously different :with the subspecies similarity is what is significant among the populations, with the ecocline, a particular pattern of difference. There are in any event good reasons for not thinking of ecoclines as classificatory units in the “group of populations” sense, the most compelling being that one and the same local population may contribute to different ecoclines. The essential fact about an eco-

FORTY YEARS OF QENECOLOUY

166

cline is that it is a trend of variation within a major population, namely the species ; in recording the infraspecific variation, what is important is to define the trends and the factors with which they are correlated, not to seek for nameable subcategories. If the concept of ecocline in this more abstract sense (“a variational trend”, rather than “a group of populations”) is accepted, then Gregor’s definition of the ecotype as a range on an ecocline loses much of its meaning. Indeed, it may be asked whether the concept of the ecotype has outworn its usefulness and whether it might not now be discarded with advantage. I n a recent paper from Gregor’s group (Gregor and Watson, 1961), it is concluded that a better understanding “could be achieved through the accumulation of records in which the emphasis has been transferred from the discrete ecotype to the trends of ecotypic differentiation”. The reluctance of several authors to use the term ecotype in recent literature reflects a rather general agreement with this view (Heywood, 1959). Nevertheless, the abandonment of the term would be difficult, and indeed undesirable. As an appellation for a group of populations known to have features in common adaptive to a recognizable habitat it continues to have a respectable use. It only becomes disreputable in misuse, or when its use is taken to imply acceptance of over-simplified conceptions of the nature of infraspecific ecological variation. Turesson’s efforts to elucidate patterns of ecological adaptation in plant populations led him to propose genecological categories at and above the level of the average Linnaean species as well as below, namely the ecospecies and the coenospecies. The definitions of these categories and the evolution in their use since the time of their original suggestion have been reviewed and discussed several times (Gregor et al., 1936; Turrill, 1938, 1946; Clausen et al., 1939, 1940, 1948; Valentine, 1949; Baker, 1952; Heslop-Harrison, 1953b; Cab, 1954). It would be supererogatory to cover this ground again here; but it is important to pay some attention to the relationships between genecological and taxonomic interpretations of the species, since much of this article is concerned with infraspecificvariation, in the broad sense. Nomenclatural taxonomy in its present form depends upon the definition of classes in a hierarchial system, the clmses at each level being based upon an assessment of overall resemblance in the range of features available for study. The basic unit, as defined by the “International Code of Botanical Nomenclature” (1956) is the species, which is to be named with a binary combination consisting of the name of the genus followed by the specific epithet. The principal tasks of the taxonomist concerned with formal nomenclatural taxonomy turn out to be identification and definition of variational units which may conveniently be

166

J. HESLOP-HARRISON

named as species, and the grouping of these units in a category system derived in all essentials from that of Linnaeus. However these activities may be embroidered -for example, by the attribution of lofty motives concerned with phylogeny -it is demonstrably true that they form the major part of taxonomic research. Now it is a peculiarity of Linnaean-style classification that its requirements do themselves generate a “species concept” -one which has come to be accepted almost universally and usually unquestioningly by practising taxonomists (Heslop-Harrison, 1962, 1963). The properties required of a variational unit suitable for recognition as a named species are that its constituent individuals should show overall resemblance, in that any one should have more characteristics in common with any other than would either with an individual of another group; that i t shows distinction from other groups of the same kind ;and that it should have some degree of persistence in time. These are essentially the properties of the morphological species of Mayr (1942) and others. Two of the practical consequences of attempting to apply universally a species concept of this general type are, firstly, that units of very varied biological character are defined as species in different groups in consequence of the diversity of genetic systems prevailing among plants, and, secondly, that in some groups where discrete variational units cannot be distinguished the taste of individual taxonomists becomes the sole determinant of what shall be ranked as species, so that chronic disagreement about nomenclatural treatments is practically inevitable. I n sexually reproducing, outbreeding groups, however, Linnaeanstyle taxonomy has met with reasonable success, and it has become apparent that this has been so because of the existence in these groups of variational units which do, in the main, conform satisfactorily in their properties with those required of the taxonomic species. These units constitute the biological species of Mayr (1942). Mayr’s definition of biological species is as follows : “Species are groups of actually or potentially inter-breeding populations which are reproductively isolated from other such groups.” As he has himself pointed out, this kind of definition contains criteria of two kinds : that concerned with reproductive isolation, and that of collectivity - “species are groups, etc.” Partly in consequence of this duality, no basis is available for the objective definition of the biological species any more than existed for the Linnaean species : indeed it may be argued that the reasoning of Mayr and others simply provides the theoretical endorsement for the application of Linnaean style classificatory methods in sexual groups (HeslopHarrison, 1955,1963). There are several uncertainties arising from Turesson’s definitions and

BORTY YEARS O F GENECOLOGY

167

usages of his own genecological units, some of which arose undoubtedly from linguistic difficulties. Turrill (1946) observes that in aiming at “an understanding of the Linnaean species from an ecological point of view”, Turesson (1922b) was justified in his use of the term ecospecies, since species “as they are realized in nature” are ecologically dFlimited. The whole impression gained from reading Turesson’s early writings is that he was looking upon ecospecies synthetically: as being composed of ecologically differentiated populations, the ecotypes, forming a mosaic throughout the distributional area. The second criterion of the biological species of Mayr, reproductive isolation, did not become part of the dehition of the ecospecies until 1929, when Turesson referred to the ecospecies as “an amphimict population, the constituents of which in nature produce vital and fertile descendants with each other giving rise to less vital or more or less sterile descendants in nature, however, when crossed with constituents of any other population.” Baker (1952)has argued that whatever interpretation may be placed upon this definition, the most appropriate application for the term ecospecies is in reference to population systems that are isolated from each other by both ecological and genetical barriers. Ecotypes of the same ecospeciesare, in contrast, assumed to be inter-fertile when brought into reproductive contact by the elimination of the eco-geographical isolation holding them apart in nature. Leaning as it does upon a genetical criterion, this conception of the ecospecies would at Grst sight appear to be more useful from a genecological standpoint than the seemingly more arbitrary species concept of nomenclatural taxonomy based as it is largely upon comparative morphology. The genetical criterion cannot, however, be readily turned into a definitive test, since no general principles can be enunciated for recognizing barriers to crossing : all integrades exist between failure to interbreed due to ecological isolation and total intersterility (HeslopHarrison, 1955). This limits the usefulness of the ecospecies concept in comparative genecology, since it cannot be assumed that the term will always connote the same thing. Situations in different genera can only be assessed one against the other when all the associated circumstances - breeding systems, nature and effectiveness of isolating mechanisms and the like - can be compared. In the ensuing pages, the word species is used in the usual biological manner, permitting it to take up a meaning largely from the context (Heslop-Harrison, 1963).In general the meaning is that of “ecospecies” in a broadly Turessonian sense; in some instances this involves a direct conflict with nomenclatural taxonomic usage -as in the examples of “species pairs” like Silene maritima and S. vulgaris which stand in relation to each other much as ecotypes of one ecospecies.

168

J. HESLOP-HARRISON

E.

GENECOLOGICAL TECHNIQUE

Phenotypic inter-population variation in a plant species may arise from three principal sources : (i) the direct plastic modification of individuals ; (ii) genetical divergence in consequence of selection, and (iii) fortuitous genetical divergence resulting in varioui ways from sampling errors -e.g. drift in small populations, or the establishment of deviating colonies from small numbers of founders. The variation arising from (i) and (ii) may be expected to show correlation with habitat factors; (iii) will be random with respect to habitat. The essential problem of genecology is to devise ways to distinguish and study variation arising from (ii). For the sake of clarity, we will refer to this variation as genecological digeerentiation, following Harberd (1957). The problem of distinguishing genecological differentiation will be seen to have two parts; it requires the separation of adaptive from random inter-population variation, and the separation of the adaptive variation, in turn, into genetic and non-genetic components. The difference between the requirements of genecological and taxonomic surveys is worthy of note. From the taxonomic viewpoint, ell variation may be regarded as grist for the mill, except that the purist would reject the “modification” as a candidate for nomenclatural recognition. It is of no taxonomic significance whether two subspecies have diverged under selective pressures or merely in consequence of some fortuitous change in gene frequency in the course of an ancient migration; it is the difference that matters, not its source. I n the genecological survey, on the other hand, information about regional variation is useless per se ; it must have the support of other types of evidence before any evaluation of the adaptive differentiation of the species can be attempted. This evidence can be sought in two general ways. The direct approach is to attempt to demonstrate adaptive responses experimentally, the aim being to discover whether or not populations in different parts of the species range do vary inter se in their ability to cope with diverse environmental variables, each being best equipped to deal with its own local circumstances. This essentially physiological approach to the investigation of genecological differentiation is discussed in Section 111. The alternative, indirect, approach is to seek correlations between “plant type” and “habitat type”. If these are consistently found, they may be taken to indicate adaptive divergence without any knowledge of the physiological meaning of the differences observed, since the only reasonable explanation for their existence is the differential effect of selection in the various habitats, save for the possibility of conditioning mentioned on p. 215.

FORTY YEARS O F GENECOLOQY

169

I n most genecological studies of the last forty years the intention has been to detect habitat-correlated, genetically-based variation, using population sampling methods of different kinds followed by comparative cultivation. Many have been open to criticism because the techniques adopted have hardly been adequate to permit the critical distinction between random and non-random variation. As Harberd (1957, 1958) and Wilkins (1959, 1960a) have pointed out in valuable critiques of genecological methodology, it may be extremely difficult - and in some cases, indeed, perhaps impossible - to make this distinction by the orthodox type of genecologicaltrial. Since the whole object of such trials is to seek for differences between populations and relate these to habitat, comparisons between populations taken singly and in groups are constantly required. Unless these comparisons can be placed upon a statistically and biologically sound basis valid conclusions cannot be drawn. In statistical terms, the exercise is to partition the genetical variance of the species in such a way that that part attributable to habitat type can be estimated and its significance evaluated. The ideal situation for analysis would be one where a species had encountered a mosaic of habitats while expanding its range, the selective influences in each habitat acting in turn upon the whole available pool of genetical variation. A comparison of betweenhabitat and within-habitat variances could then be expected to expose as significant those differences which were truly adaptive. All practical cases differ from the ideal one in some respect or other, and some so radically that analysis if not impossible can yield only dubious results. The example of regional subspecies illustrates the point. It may be possible to show that the area of one coincides with a climatic zone different from that of another; each can certainly be said to be “adapted” to its environment to the extent that it survives, but it cannot be said that all the morphological and physiological differences between them have necessarily been the outcome of differential selection in the two areas. As Wilkins (1959)has pointed out, the area of one may have been colonized by a small somewhat unrepresentative invasion from the other, so that adaptive differences rising secondarily in consequence of selection are confounded with the original chance differences between the colonists and the population from which they emerged. In this case evidence of genecological differentiation must be sought by the experimental demonstration of adaptation, using reciprocal transplant methods to test survival capacity or controlled environments to measure response to individual habitat factors. This difficulty is part of a general one, arising from what Wilkins (1959) has termed “a general uncertainty about the variance found within a local population”. Another aspect of this has been stressed by

170

J. HESLOP-HARRISON

Harberd (1957, 1958), namely the danger of attributing false significance to differences between populations in genecological trials in consequence of unsatisfactory sampling procedures. Harberd has been concerned with the particular case of a single genotype being incorporated several times in a sample because of the clonal spread of a parent plant to an unknown degree. By decreasing the within- population variance in a trial this can lead to the attribution of “spurious significance” t o differences between populations. Both Harberd (1957) and Wilkins (1959) are driven to the conclusion that the within- population variance is not a particularly useful statistic in genecological studies, and that the value of a trial is most likely to be increased by increasing the number of populations sampled rather than the sample size. For a trial of given scale, the ultimate is a sample of one plant from each population; then, as Wilkins (1959) says, the variance would automatically. be that within a purely statistical assemblage of unrelated individuals, and there would be no reason to expect any two such assemblages from the same geographical area to show non-adaptive differences. A further source of confusion in genecological studies lies in the form of material collected for experimental garden or other investigation. The investigator in general has two choices: he can transfer mature living plants, or he can grow his material from seed. I n the former case, he is sampling from a selected population; in the latter, he is sampling from a population which has descended from selected ancestors but has not itself suffered selection. The implications of this are sometimes overlooked. When populations occupy unlike habitats but are in sufficiently close proximity to permit gene exchange, differentiation will only proceed if selection pressure is high enough to out-balance gene flow (p. 204). I n the extreme case, adaptation will be a generation-to-generation matter, with an essentially random dispersal of genotypes over the entire area each year and a subsequent stringent selection for adapted genotypes in each habitat “sub-population”. I n this case, to rely upon seed samples for the estimation of genotypic differences is to guarantee that they will not be found. Even when remote populations are being compared, a seed sampling method may not provide an adequate picture of what the actual surviving population in a given site is like. I n the absence of disruptive gene-flow from other populations, recombination is unlikely t o turn up radical deviants in any quantity, but where an adaptive response depends upon a nice balance of polygenes in an outbreeding population amodal types are certain to occur in each generation, normally destined to succumb in the wild unless they happen to be included in a seed packet. The aberrant individuals recorded in various studies on photoperiodic and other responses of ecotypes may have had

FORTY YEARS OF GENECOLOGY

171

this origin. All this leads to the conclusion that the comparison of “seedpopulation” and “plant-population” variances would itself be worth while in genecological trials of perennial plants where technically possible-as it would be in periodicity studies on clonally divisible grasses and the like. A word is necessary here concerning the other important distinction which has to be made in a genecological survey - between the genetical and non-genetical components of inter-population variation. The common practice is to.attempt to eliminate the effects of direct environmental modification of individuals by cultivating population samples side by side in a standard garden. This method (which ante-dates genecology by at least a century) may be quite adequate to permit the kinds of distinction required, given an appropriate design of lay-out and analysis. In some circumstances, however, it may be unsatisfactory. The most important deficiency of the method is that in eliminating environmentally imposed variation it may obscure genetically determined differences in the capacity to react adaptively to special environments. This is particularly dangerous where the aim is not merely to observe morphological differentiation but to test physiological responses. For example, consider the case of two genecologically differentiated populations, one with the capacity to react adaptively to intense sunlight, the other without. Tests made after side-by-side cultivation in a r‘neutra17’environment under moderate illumination could fail altogether to reveal any difference in response to intense light. One might go so far as to say that an “ecotype” is never adapted to its special milieu when cultivated in an experimental garden; it merely carries the ability to become so adapted under the appropriate evocative environment. It may in cultivation show differences from plants from other habitats, but the differences need not relate at all closely to its true adaptive capacities. The work of Bjorkman and Holmgren (1963), described in detail in a later section, is exemplary in giving f d attention to the matter of pre-conditioning. Another deficiency of the comparative cultivation methad lies in the possibility that the test environment while suppressing some environmentally induced characteristics may evoke others never expressed in the natural habitats. A sample from a phenotypically uniform natural population may turn out to be genotypically highly diverse when observed in a different environment, and this artificial enhancement of variance could prejudice attempts to detect and evaluate genecological differentiation. Numerous examples of this effect have now been recorded, particularly in connection with the environmental control of developmental periodicity ; Sinskaia (1958) and Clausen and Hiesey (1958b) discuss some of its implications for genecology. It may be noted

172

J. HESLOP-HARRISON

that this exposure of latent genotypic variation has nothing to do with the uniformity of the test environment; it is simply that this environment increases the penetrance of certain genes compared with the native one. There is another even more subtle way in which the within-sample phenotypic variance may be increased, even under supposedly highly uniform controlled envii-onments: this is when genotypes are carried so far out of their nor& that developmental regulation is disrupted (Heslop-Harrison, 1959a). Here it is not a question of the exposure of genetical heterogeneity -the effect may be seen in what is effectively a pure line - but rather the amplification to a phenotypically obtrusive scale of developmental “noise”. Many of the limitations of the simple transplant technique are removed when methods of varied-environment or reciprocal transplanting are adopted. The classical example of this approach is the work of Clausen et al. (1940 et sq.) using transplant stations along a coastal plain to inland mountain transect in California-work which had a prototype, albeit an unsatisfactory one, in the experiments of Gaston Bonnier. A further technical problem meriting consideration is the selection of characteristics for observation in a genecological study. Morphological characters are those most readily investigated ;yet as Turesson pointed out repeatedly in his pioneer papers, it is the whole reaction of plant to habitat which is of adaptive importance, physiological responses being paramount. I n a paper on recognizing adaptive variants, Wilkins (1960a)points out that as the only ultimate test of adaptation is survival in the given habitat all other kinds of evidence, including that derived from comparative cultivation, is bound to be circumstantial. The best form of such circumstantial evidence, he suggests, is that arising from the study of correlations between measurable habitat factors and plant characters. Frequently there will be no possible way of assessing the biological significance of the features measured, and although the existence of correlations may suggest they are adaptive, an element of uncertainty remains since it is always possible - and indeed for many morphological characteristics, probable -that what is observed is itself no more than a by-product of the physiological process basically concerned. Concentration upon morphologica.1 characters may in fact lead to incorrect conclusions about the “adaptedness” of populations if the characters themselves are neutral enough in respect to selection to vary randomly over the area sampled, particularly when the random variation is taken to indicate the luck of adaptation. Langlet (1963)has quoted as an example of this kind of reasoning Clausen’s conclusion (1951) that since no correlation was detectable between latitude of origin and a small group of morphological features, no ecological clines &represent in Layiaplatyglossa.

173

FORTY YEARS OF OENECOLOOY

F.

OENECOLOOICAL DIFFERENTIATION : SOME ILLUSTRATIVE EXAMPLES

Notwithstanding the technical difficulties mentioned in the foregoing paragraphs, a wealth of data now exists illustrating the patterns of habitat-correlated variation encountered in higher plant species. As a prelude to the discussion of evolutionary aspects of genecological differentiation in Section 11, a number of examples will now be reviewed under headings defining the main types of situation known at present.

1. Major Ecological Races with Vicarious Distributions The most fully studied example of a species containing distinct ecological races with practically exclusive areas is that of Potentilla glanduZosa. I n the most recent publication of the Carnegie group (Clausen and Hiesey, 1958a), this species is discussed again. The distinctness of the four “ecotypic” subspecies in the Californian area studied is maintained, but the additional data available reveal that each is itself heterogeneous.

TABLEI Characteristics of the Ecotypic Subspecies of Potentilla glandulosa along the Central Californian Transect. [Datafrom Clausen and Hiesey, 1958a.l typica Distribution

Coast Ranges and lower Sierra Nevada

refiexa

hamenii

nevadensig

Low and Meadows, mid- High altitudes middle altialtitudes of of Sierra Sierra Nevada tudes of Nevada Sierra Nevada. Moist, sunny Soft chaparral Dryish, open Moist Habitat slopes meadows timbered and open slopes woods Climatic toler- Coastal to Middle and Middle and Coastal to high altimiddle altihigh altimiddle altiance as extudes (poor tudes (poor perimentally tudes tudes survival determined survival near coast) near coast) Winter-dorWinter-active Winter-dorSeasonal perio- Winter- and mant sumdicity at mant s u m summeror -dormer-active Stanford mer -active mant ;sumactive (alt. 30 m) mer-active Wide, at least Moderate, at Internal Wide, proWide, proleast two variation two bably bably “ecotypes” “ecotypes” several several ‘Lecotypes” “ecotypes” Undetermined Self-sterile Self-fertile Self-fertile Self-compatibilitv

174

J. HESLOP-HARRISON

Table I, part of a comparative table given by Clausen and Hiesey, gives the essential information about distribution, ecology, periodicity and internal variation for the four subspecies. The accounts given of the internal variation of the subspecies indicate that each is genecologically differentiated at a lower level. The pattern of this differentiation is related to local climate and microclimate as determined by aspect and altitude, and also t o edaphic features. Subsp. typica is quite evidently extremely heterogeneous;according to Clausen and Hiesey, .“Individuals within each local population differ in minor but distinguishable morphological and physiological characters. Neighbouring populations differ statisticaIly in their average biotype composition even though they belong to the same ecotype. Populations from climatically distinct regions are to a greater degree divergent both morphologically and physiologically and represent ecotypes.” Subsp. refexa is also recorded as being heterogeneous, but seemingly to a lesser degree. Two distinct ecotypes are distinguished within it : one, inhabiting low level forest, being winter active and early flowering in cultivation, and the other, from higher level Pinm ponderosa forest,.being winter dormant and late flowering. Subsp. hanseni, although occupying only a comparatively small total area, is one of the most variable of the four occurring in California. It is described by Clausen and Hiesey as constituting a link between the foothill and high altitude races of the subsp. refexa and nevadensis. I n morphological features and earliness, “with increasing altitude . . . hanseni tends more and more to resemble nevadensis.” Moreover, the same trend is present in life form - “The forms of hanseni are generally hemicryptophytes . . .the hanseni populations from lower altitudes, however, develop short woody crowns similar to the chamaephytes of refexa, whereas those from higher altitudes are more rhizomatous, as in nevadensis.” These observations certainly seem to show that a system of clines does exist in hanseni, but later in the same account $he authors appear to deny that this is so. Subsp. nevadensis, from the highest altitudinal belt, appears to reveal the narrowest range of variation, but even so, genecological differentiation is present. Two climatic ecoty-pes are described, one alpine and the other sub-alpine, differing morphologically and in periodicity. Clausen and Hiesey relate the distinctness of the four central Californian subspecies of P. glandulosa to the sharply differentiated climatic zones across the transect from the Pacific to the Sierra Nevada. They remark that north and south of the area investigated in detail the pattern of zonation becomes obscured; then, “the rather well defined subspecies found in central California break down into other combinations of characters.” Within the area where the subspecies are well distinguished, inter-gradation between them is said to be minimal, and

FORTY YEARS OF GENECOLOGY

176

where it occurs Clausen and Hiesey interpret it as evidence of hybridization. Another example of a Californian species with vicarious, ecologically specialized subspecies is that of Gilia capitata. G . capitata is a polytypic species with a natural range extending from central California to British Columbia. Following an intensive study of population samples from various parts of the range, both as acquired in the wild and in comparative cultivation, Grant (1950, 1952, 1954) has concluded that eight recognizable entities worthy of taxonomic recognition as subspecies are present. Except for subsp. abrotanifolia and pedemontana which have overlapping ranges, these subspecies replace each other geographically in a pattern related to both the north-south climatic gradient of the Pacific coast and the east-west gradient from the coast towards the interior. Grant recognizes two basic types within the complex, subsp. capitata, a plant of rocky hillsides, and the two subspecies chamissonis and staminea, plants of sand dunes or sand plains. These may have had a common parentage in subsp. abrotanifolia. With subsp. tomentosa, subsp. capitata forms a pair of ecotypic vicariads, tomentosa maritime and capitata inland; similarly subsp. chamissonis and staminea can be looked upon as a pair of ecotypes, one coastal, the other inland. The remaining races may have arisen through introgressive hybridization between the primary subspecies. Their success, according to Grant, has resulted because “ecological opportunities” were available to hybrid types. Their populations are thus also ecologically differentiated. As with Potentilla glanddosa, the overall pattern outlined by Grant for Cilia capitata is one of comparatively stable ecological races extending over very considerable areas. Within the subspecies, however, he describes several examples of “inter-colonial” variation, in some cases extending to such characters as seed germination and flowering time. The possibility of clinal variation in the character of glandulosity in subsp. capitata is also considered. Grant does not attribute ecological significance to this local variation, but it is possible that more intensive study might show its adaptive character. In the shrubby California genus Ceanothus, Nobs (1963) describes “species equivalent to ecotypes” in the section Cerastes. They are accepted as taxonomic species because of their morphological distinctness, and “the genetic basis of their differentiation and subsequent stabilization . . . revolves primarily on the coherence of the genotypes balanced against the selective pressures of the natural environments”. All are completely interfertile in experimental crossings. I n the area north of San Francisco Bay, where eleven of these “species” occur, the majority “are geographically separated, and coincide in distribution with changes in ecological habitats, especially with soil type”. I n general, where the

176

J. HESLOP-HARRISON

species do meet, their ecological limitations severely inhibit natural hybridization. Not all of the ecologically-correlated variation within the section Cerastes is accounted for by this sub-division. While some of the narrowly ranging species seem reasonably uniform, Ceanothw gloriosus shows further differentiation into maritime and inland races, also given taxonomic recognition, as varieties. C. cuneatw has “developed numerous recognizable forms”, and “the aggregations of characters that typify each form follow essentially a clinal transition which coincides with a gradual ecological gradient . . .,’. The common feature of the three examples from the Californian flora outlined in the preceding paragraphs is that vicarious ecological races are seemingly present, distinct from each other in morphological and physiological characteristics. The ecological races are all evidently themselves heterogeneous to a greater or lesser degree, showing local variation which is probably both adaptive and non-adaptive. The fact that the situations in the three genera have been given different taxonomic treatment is confusing but irrelevant; it merely indicates that in polytypic species complexes the taste of the individual worker tends to be the arbiter so far as nomenclature is concerned.

2. Ecological Races with Interdigitating or Mosaic Distribution-s There are now several well-documented examples of ecologically differentiated races which do not possess vicarious areas but overlap geographically, each occupying its own characteristic type of habitat within the common area. Again, varying taxonomic treatments tend to obscure the essential similarity of many cases. By the criterion of free gene exchange in experimental crosses sympatric ecological races may merit inclusion in one and the same ecospecies, but their nomenclatural treatment normally depends upon their degree of morphological differentiation. If this is conspicuous and consistent they will, quite justifiably, be generally named as species; otherwise they may be classified as subspecies, or may even escape taxonomic recognition altogether. The example of the Ranunculus lappaceus group in south-eastern Australia is one of the most remarkable yet described (Briggs, 1962). Seven named species occur in the 5 000-6 000-ft altitude zone of the Kosciusko plateau. All are closely similar in karyotype, and are recorded as being freely interfertile ; they may thus be regarded as being part of the same ecospecies. The different races show very narrow ranges of ecological tolerance, and, since the characteristic habitats are scattered throughout the area, they have a mosaic distribution. The faithfulness of the races to their particular habitat is evidently extremely strong. The situation is illustrated by R.millanii and R.dis-

FORTY YEARS O F QENECOLOQY

177

sectifolim, the former characteristic of fen communities and shallow depressions in grassland areas subject to intermittent flooding and the latter of wet grassland, was less than Q m in width; the two forms were entirely restricted to their characteristic habitats, and hybrids were encountered only in the ecotone. Reciprocal transplant experiments were carried out with these two forms; millani showed only moderate growth in the dissectifdim habitat, and dissectifolius failed altogether to survive in the millani habitat. Briggs considers that the identity of the different races is maintained by their ecological specialization, which is sufliciently strong to ensure that hybrid derivatives have no chance of success in the wild except in ecotones, which tend in general to be narrow. This seems to be true also of several taxonomic “species pairs” in the European flora such as Silene maritima and S. vulgaris, Melandrium rubrum and M . album, and Geum rivale and G: urbanum. The common features here are that the entities concerned in each pair are distinguished from each other by several consistent differentiae : they are freely interfertile in experimental crossings, they have distinct ecological tolerances, and they do not hybridize freely in the wild except where ecotones occur or where there has been habitat disturbance. I n addition, the total geographical ranges although overlapping are not coincident, and with some there is evidence that the distinctness has persisted since early post-glacial times (Turrill, 1946). Obviously all of the examples quoted so far are interpretable as cases where two or more races have differentiated allopatrically in both morphological and physiological properties, and have then acquired sympatric ranges by migration, their integrity being preserved subsequently in the common area by strong ecological specialization. There are several essentially similar cases where ecological races have been given taxonomic recognition as subspecies because the differentiae are fewer, or less distinctive, than would normally be required of taxonomio species. The subspecies of Dactylorchis incurnata in the British Isles constitute habitat races showing morphological differentiation principally in floral characteristics (Heslop-Harrison, 1953a, 1956). The differentiae are maintained in cultivation, and are associated with phenological and other physiological differences. Populations referable to the various subspecies often occur in close proximity without loss of identity, and occasionally even intermingled where there are rapid alternations in habitat. Thus in one locality plants referable to pulchella and coccinea occurred together in a field traversed by an old herringbone drainage system. The hollows of this carried a wet-soil vegetation forming a continuation of a lake-side fen. On the drier mineral soil of the ridges, 85% of the plants were coccinea; in the hollows, 96% were

178

J. HESLOP-HARRISON

pulchella. I n this situation the habitats, although adjoining, were distinct. I n others studied, particularly in the East Anglian fens where human intervention has blurred ecological boundaries, many nondescript, intermediate habitats exist. I n these the D. incarnata populations are heterogeneous, and the character combinations typifying the named subspecies are often broken up. The probable role of floral characters in contributing to the reproductive isolation of the ecological races of D. incarnata (Heslop-Harrison, 1958)is mentioned in alater section (p, 212). The pattern of ecological differentiation in D. incarnata is one of distinct yet intermingled races; in other species of the genus a bolder regional differentiation is evident, reminiscent of that described by the Carnegie group in Potentilla glandulosa. The aggregate D. maculata consists of two ecospecies, D. fuchsii and D.maculata,the former diploid and the latter tetraploid, which show different ecological tolerances throughout the European range (Heslop-Harrison, 1951). Each of these in turn contains regional subspecies of an ecotypic nature. D. fuchsii, for example, is represented in the British Isles by three forms, distinct enough to have been acknowledged taxonomically as subspecies,fuehsii, hebridensis and okellyi. Subsp. fuchii is wide-ranging, occurring in meadows and open woodland on neutral or moderately basic soils; subsp. hebridensis replaces subsp. fuchsii along the western seaboard in regions of extreme oceanic climate from Cornwall, through western Ireland and the Hebrides to Sutherland, while subsp. okellyi is restricted to areas of karst-like limestone in Ireland and north-western Scotland (Heslop-Harrison, 1953aJ 1958). The subspecies are distinguished by several correlated morphological features, maintained in cultivation, and also to some extent by phenology. They have the aspect of distinct races which have attained their present distributions independently, but again this can be no more than a subjective judgment. I n some recorded examples of genecological differentiation, the evidence is inadequate to attempt any kind of distinction between possible monotopic or polytopic origins, Thus Habeck (1958) has studied the variation in seedlings from seed samples of T h u y a occidentalis from twenty-nine sites within the distributional area in Wisconsin. The sampling sites were classified into “typical lowland swamps” and “well drained uplands”. Marked differences in response t o three conditions of soil moisture were observed in cultivation, and Habeck concluded that two distinct ecotypes, one adapted to the upland conditions and the other to the lowland, co-exist in Wisconsin throughout the species area. Nothing in the recorded data serves to establish whether two races different in origin are involved, or whether the ecotypes have evolved polytopically in response to local selective pressures. The same kind of difficulty arises in interpreting ecological races oon-

FORTY YEARS O F GENECOLOUY

179

fined to certain extreme soil types occurring locally in otherwise normal terrain, as for example over outcrops of serpentine rocks, or on the spoil heaps derived from lead and other mineral workings. The sudden changes of vegetation encountered with the transition from non-serpentine to serpentine soils is well documented for many parts of the world (Whittaker, 1954; Walker, 1954). Frequently this is seen in-a complete disappearance of intolerant species. I n some cases, a taxonomic species is replaced abruptly by another, very closely related one; in others, edaphic races tolerating serpentine conditions occur which are not sufficiently well differentiated morphologically to have merited taxonomic recognition (Kruckeberg, 1954). These edaphic ecotypes are usually quite sharply demarcated physiologically from neighbouring populations on normal soils, reflecting the abrupt change in habitat. Kruckeberg’s study of 1950 is of special interest, since he included observations on two species already extensively studied genecologically, Cilia capitata (Grant, 1950; see p. 175) and Achillea borealis (Clausen et al., 1948). Within the subspecies capitata of Cilia capitata, considered by Grant to constitute a reasonably homogeneous race with a distribution related primarily to regional climate, Kruckeberg detected local populations adapted to serpentine soils and meriting recognition as edaphic ecotypes. Similarly, in the inner Coast Range-Sierran foothill race of Achillea borealis subsp. californica, regarded by Clausen et al. as a climatic ecotype, serpentine tolerant and intolerant edaphic ecotypes were distinguished. Kruckeberg (1954) comments on the Achillea situation as follows : “There are thus edaphic subdivisions within climatic ones in this species, i.e. ecotypes within ecotypes. Moreover, since the geographical area covered by the foothill climatic ecotype is very diverse lithologically and therefore is a mosaic of different soil types, additional edaphic ecotypes may well be expected. The superimposition of ecotype on ecotype at least suggests that there exists a much more complex genotypical response to habitat than is implied in consideration of single environmental factors.” The differentiation of serpentine tolerant races in wide-ranging species has its parallel in miniature in the emergence of local populations tolerant of the toxic soils formed on the spoil heaps from lead mines and other mineral workings (Bradshaw, 1952). Wilkins (1960a, b) has initiated the study of lead tolerant populations of Festuca owina in Britain. Three “types” were distinguished, in respect to tolerance, one (wide-ranging)intolerant, and two others, restricted to lead-containing soils, of medium and high tolerance respectively when tested for rooting in lead nitrate solutions. Tolerance was not associated with any morphological features, so Wilkins has been reluctant to talk of “lead soil ecotypes”. Evidently there is nothing comparable to a homogeneous race

180

J. HESLOP-HARRISON

involved, the pattern of lead tolerance being superimposedupon all other systems of adaptive and non-adaptive variation within Pe&ma ovina.

3. Clinal Variation Some of the most informative examples of infraspecific ecoclines have been described in woody species. Thus the variation of the Scots pine, Pinw sylvestris, throughout its European range has been the subject of much illuminating research and some controversy (Langlet, 1934, 1936, 1959, 1963; Wright and Baldwin, 1957). I n his earlier work, Langlet described clines spanning almost 25" of latitude, involving several morphological and physiological features such as leaf length, hardiness, dormancy period and shoot extension rate. The 1938 international test involving fifty-two provenances provides some particularly striking evidence. I n the trial grown near Stockholm, Langlet recorded the percentage of dry matter in the needles of 2-4 year seedlings in the Iate autumn. This particular measure showed a close relationship with the length of the growing season, assessed as the number of days with an average temperature of 6" C or more, in the native habitats. A still closer relationship was apparent between dry matter content and the length of the first day of the year with an average temperature of greater than 6°C. The scatter diagram is reproduced in Fig. la: the regression is curvilinear, and r exceeds +0*98 (Langlet, 1959). The clinal nature of the variation is quite beyond challenge. The trial reported upon by Wright and Baldwin (1957) involved forty-six of the fifty-two provenances tested by Langlet, but their treatment and conclusions were radically different. The material was grown in New Hampshire, a t a latitude of 43"N, some 14" south of Stockholm. According to Wright and Baldwin, the evidence suggested that the variation in Scots pine is discontinuous, and they grouped the sampled populations into a number of regional ecotypes. Some of these -in the northern part of the range-occupied latitudinal belts, and bore a clinal relationship to each other ; others further south coincided with no particular latitudinal or climatic zones. This treatment has been heavily criticized by Langlet (1959), who pointed out that the method of grouping the different samples geographically and comparing regional means for the various characters observed would necessarily obscure variational continuity. His own re-plotting of the height data for 17-year trees grown in New Hampshire in relation to day length is reproduced in Fig. lb. There is a remarkable correspondence between the distributibn shown in this figure and that shown in Fig. la; aa Langlet remarks, Wright and Baldwin do seem very satisfactorily to have demonstrated continuous variability in Scots pine, in spite of their own conclusion to the contrary.

181

FORTY YEARS OF QENECOLOQY

The clines in pine revealed by the studies of Langlet and others include examples of latitudinally correlated variation in characteristics which are essentially associated with developmental periodicity. Presumably they reflect the varying response to photoperiod through which adaptation to length of growing season at different latitudes has been attained. Other similar examples, including the "photoperiodic ecotypes" of Vaartaja (1959), are discussed in Section 111.

10

m * am I-

14 (h)

16

18

20

22

24

(h)

FIG.1. (a)Relationship between dry matter content and the length of daylight of the first day in the year with an average normal temperature of + 6' C a t the native habitats of fifty-two provenancesof Pinw s y z w e e t ~ hgrown a t Stockholm. (From Langlet, 1959.) (b) Relationship between tree height at 17 years and the length of daylight of the first day in the year with an average normal temperature of + 6"C at the native habitats of forty-six provenances of P . 8yZweetriS grown in New Hampshire, U.S.A. (Data of Wright and Baldwin, 1957, re-plotted by Langlet, 1959.)

In view of the abundant evidence that the periodicity of growth in forest trees is closely related to the length of the growing season in the habitat of origin, a report by Daubenmire (1950) indicating that cambial activity in Pinus ponderosa is not related to latitude or altitude of origin is of some interest. Forty-one trees, originating from populations span.ning 14" of latitude and occurring up to an altitude of 7 200 ft were studied in cultivation together in northern Idaho. Considerable variation in the duration of cambial activity was observed within the samples from the various populations, but there was no meaningful pattern of variation between the populations, and certainly no obvious relationship with latitude. This observation stands in some contrast with that of Weidman (1939), who grouped the same material into four regional races on the basis of needle number, length and anatomy, and growth rate and hardiness. The data given by Weidman also provide some evidence of clinal variation related both to latitude and altitude. G

C.E.R.

182

J. HESLOP-HARRISON

Since the earlier work of Gregor and his collaborators on Plantago maritima, ecoclinal variation has been described in several herbaceous species. Bocher’s investigations of Prunella vulgaris (1945, 1949) have involved comparative cultivation of large numbers of population samples from throughout the species range, both in standird and varied environments. He considers that units worthy of being termed ecotypes occur within the species, the principal ones being associated with montane-boreal habitats, and with dry, medium dry, wet and shaded soils in lowland habitats. However, “in P . vulgaris there is continuous variation, for which reason the ecotypes . . . represent the most frequent character combinations in the different types of habitat. A large number of biotypes or races may not be conformable to the ecotype system, because in respect of one or more characters they do not agree with the ideal combination (the most characteristic features of the ecotypes).” Bbcher refers to the ecotypes within P. vulgaris as theoretical concepts, since they are, essentially, “composed of ranges within a whole series of continuous or almost continuous character gradients or clines.” His position is thus very close to that of Gregor. Bocher and co-workers have also given evidence of ecoclinal variation in Plantago coronopus (Bocher et al., 1953, 1955). The diploid complex represented by P. coronopus L. and P. macrorhiza Poir. forms a very variable series of populations around the Atlantic and Mediterranean seaboards of Europe. Comparison of population samples from throughout the range in cultivation led to the conclusion that the adaptive trends were best expressed “as through a number of clines running from the north to the south in Europe. Strains of southern origin deviate from those from the north by often being of greater size, with more ascending leaves and scapes, wider leaf-rhachis and longer spikes. They further seem to be more resistant to drought”. Dwarf races occurring in the north-west on exposed rocks and cliffs were considered to constitute a race distinct enough to be called an ecotype, which, however, differed from another similar ecotype on sea cliffs in southern Scandinavia. Like P. maritima, P. coronopzls shows great local variation in Europe, and it is likely that major clinal trends suspected by Bocher et al. overlie local patterns of ecotypic and ecoclinal variation comparable to those described by Gregor for the former speciesin the British Isles (Dodds, 1953). Studies on the phenotypic variation of wild populations have often revealed very distinct clinal trends related to climatic gradients (e.g. Alnus glutinosa, McVean, 1953; Melampyrum pratense, Smith, 1963), and although much of the observed variation may well be in consequence of plastic modification, there seems little doubt that some will be genotypically based. Indeed, the view of Stebbins (1950) that “it is likely that most species with a continuous range that included more

FORTY YEARS OF CJENECOLOQY

183

than one latitudinal or altitudinal climatic belt will be found to possess clines for the physiological characteristics adapting them to the conditions prevailing in different parts of their range” has now been fully justified. A special type of clinal variation occurs in polymorphic species when successive populations show progressive change in the proportion of the morphs (Huxley, 1955). Such ratio clines frequently reflect directly the distribution of allelic genes, and several examples have now been described. Thus Harland (1946)detected a cline in Ricinus comrnunisin Peru in the proportional representation of genes for stem bloom (glaucousness, or waxiness). Bloom is determined by a pair of alleles, B-b. The castor bean is not native to Peru, but it is naturalized as a weed, as well as being in cultivation, up to considerable altitudes. Passing inland from Lima to an altitude of 7 764 f t the proportion of plants bearing bloom increases from 0-15y0to lOOyo.I n experimental cultivation in the Lima district, homozygous B type plants were found to be sterile in the winter, whilst bb plants fruited throughout the cold season, although at a diminished rate. The winter sterility of BB plants Harland attributed to the cold and fog of this season in Lima; the increase of such plants with increasing elevation inland is associated with an increase in the amount of sunlight and the diminution of fog. I n what way glaucousness is physiologically disadvantageous in the Lima climate is not apparent. It may be that bloom is simply a superficial expression of some unknown more profound physiological condition, or, as Harland suggests, selection has been for some closely linked gene of physiological significance. Parallels to Harland’s Ricinus case occur in several species of Eucalyptus in Tasmania, where clines in the incidence of glaucousness have been described by Barber (1955) and Barber and Jackson (1957). The glaucousness of leaves and stems is considered to be under oligogenic control, and the clines, detected by field inspection, are taken by Barber to represent changes in the allelic frequencies at one or more loci controlling wax development. I n three species, E . gigantea, E . gunnii and E , coccifera, parallel clines were observed in a transect passing from altitudes of c. 2 500 f t to 4 000 f t in north central Tasmania. I n E. gigantea which has the lowest altitudinal range, glaucousness is dominant in the lowest populations; thereafter to 3 500 ft glaucous types are absent, when they reappear to achieve a frequency of 60% at 3 750 ft. I n E. gunnii and E. coccifera a simpler clinal pattern appears, with nonglaucous plants at lower elevations giving place quite rapidly to glaucous ones with increasing altitude. I n E. gunnii there is an additional cline in glaucousness of juvenile foliage which is said not to be in step with that in the adult-type foliage. Although the clines described by Barber are described as involving changes in proportion of glaucous and nonglaucous genotypes, he also indicates that growth conditions affect

184

J. HESLOP-HARRISON

expressivity, low temperatures promoting the formation of wax when the genotype allows it. The correlation he deems most significant is that with temperature along the altitudinal transect, increasing frost being correlated with increasing glaucousness. Again, the physiological significance of glaucousness in relation to the factor which is presumed to be selective is not apparent. Two ratio clines have been described in the corn field weed, Spergula akensis, by New (1958, 1959). The characters concerned were the papillation of the seed coat, and the incidence of pubescence on stems and leaves. The inheritance of the seed coat character was found to be dependent upon a pair of alleles, there being no dominance. The inheritance of hairiness was less simple, but no dominance was found, and New concluded that since two classes “medium hairy” and “densely hairy” were usually distinguishable in field populations, the phenotypic ratio must represent gene frequencies. There is no genetic association between the two characters. Sampling and scoring were done directly in the field. I n both characters clines were detected passing across the British Isles from NNW-SSE, the proportions of densely hairy and of non-papillate plants increasing towards NNW. Superimposed upon these regional clines, local clines in the same characters were found related to altitude, the.proportion of non-papillate and of dense-hairy plants increasing with elevation. New carried out experiments under controlled environments to determine whether the characters were associated with any obvious selective factor, The direction of the geographical clines in the British Isles is aIso that of a gradient from a warmer, somewhat drier climate towards a cooler, wetter one, and it is therefore of interest that New found non-papillate, densely hairy plants intolerant of high temperatures and low humidity, suggesting that they would be at a selective disadvantage in the SSW of the country. Nonpapillate plants also show a lower fertility at high temperatures and low humidities, and their seeds germinate better a t low temperatures than high. All of this certainly suggests that the clines are selectively maintained by climate -those related to altitude as well as the larger geographical ones. Again, the physiological basis of the observed differences in reactions of the different genotypes is not clear; as New points out, there is even a certain perversity in the hairy types being less resistant to low humidities, when pubescence is often considered to be an adaptation to reduce water loss.

4. Small Scale QenecologicalDifferentiation Local patterns of genecological differentiation have now been studied in several genera. Bradshaw (1959, 1960) has investigated the common grass species, Agrostis tenuis, over a small area of central Wales. Thirty-

FORTY YEARS O F GENECOLOGY

185

three colonies were sampled, and the samples were compared in standard and varied environments. Substantial differences were observed in morphological and physiological features under standard cultivation. The differences are considered by Bradshaw to be adaptive, and he concludes that it is “the environment, even its local variations, which determines the pattern of differentiation. So where there are sharp changes in environment, . . there are sharp correlated changes in the populations. Where there are gradual changes in the environment. . . the population changes are equally graded. Where in such gradients there are sudden local variations. . . there are sudden population changes”. Evidently A. tenuis shows a mixture of continuous and discontinuous variation on a very local scale, and Bradshaw declines to identify any of the populations as ecotypes, or even to apply the concept of clines. The picture sketched by Gregor and Watson (1954) for variation in Lolium perenne in Britain shows some similarity with that described for Agrostis tenuis by Bradshaw. I n their later investigation of local genecological variation in Plantqo lanceolata (Gregor and Watson, 1961), sample size was restricted in favour of number of populations sampled, in sites scattered throughout a mosaic of different habitats mainly within an area of 70 x 30 miles in southern Scotland. The sites were grouped into three broad classes: A, species-rich pasture; B, speciespoor, lightly grazed vegetation, and C, ungrazed vegetation dominated by Molinia caeruleu. I n cultivation under standard conditions, the plantain samples showed average leaf lengths increasing progressively from habitat type A to habitat type C, the differences between the groups being statistically significant. There is thus a “dispersed” ecocline in the characteristic, in so far as the three habitat types, A, B and C, can be regarded“as forming an ecological gradient in this small geographical area. This study also included an assessment of the role of developmental plasticity in the adaptation of the plantains to minor variations between the habitats. A study of small-scale genecological differentiation by Harberd (196lb) is of particular interest since no fewer than thirteen taxa were investigated simultaneously. The sampling was carried out within one broadly defined type of community, Agrostis-Festuca grassland, and within a comparatively small geographical area. A total of eighty sites were selected in a series ranging from “flushed” to “leached”, and five circular turves c. 7 ern in diameter were taken from each; the plants were grown on from ramets present in the turves. One to ten characteristics were scored in the different species. Using the within-site variance as the error term, an analysis of variance showed that more than 70% of the characters scored differed significantly between the popula-

186

J. HESLOP-HARRISON

tions. These differences need not, of course, be related to habitat differences within the eighty sites. To test for genecological differentiation, Harberd grouped the sites into seven variants of the Agrostis-Festuca community, in the series flushed -+leached. Differencessignificant at the 1 yo level between the grouped samples were found in only four species characteristics. In Carex caryophyllea, leaf length differed significantly between the groups, and this character showed a relationship with the position of the site in the series, shorter leaves coming from the more ~ a trend in flowering time was associated flushed sites. I n P Qtrivialis, with the site-group series, flowering being later in the flushed sites. I n Fatuca rubra and F. ovim, differences were observed which did not relate satisfactorily to the habitat series. I n F . rubra, Harberd considers the observed difference could have arisen from reduplication of genotypes between the sampling sites; he offers no explanation for the anomalous result with F . ovina. Flowering time in Cerastium vulgatum was found to be correlated significantly with the position of the site in the ecological series, although the character did not reach significance in the site-group comparison. This investigation provides a convincing demonstration of the value of extensive sampling methods in genecology, since the design of the programme permitted a clear distinction to be made between fortuitous inter-population variation and that which is likely to be of genecological significance. Ehrendorfer (1953), in a study on a geographical scale similar to that of Harberd, detected genecological differentiation within an essentially continuous population of Galium pumilum. The plant community was Arrhenatherum meadowland, within which various facies could be recognized according to aspect, exposure and soil moisture. The ten sampling sites could be placed in four groups, from warm, dry, open to cool, moist, shaded. The incidence of hairiness was found to vary markedly between the groups, the proportion of hairy individuals in the samples falling progressively from 55 in the driest habitat to 19 in the dampest. Since hairiness is determined by a single dominant gene, this is an example of ratio ecocline. Observations made in the field showed that flowering was earlier on the average in the drier sites. Although the differences in this feature between the sample sites were likely to have been determined in part by direct environmental effects, Ehrendorfer found some association between hairiness and precocity, suggesting that some of the variation between sites in flowering time was genetically based. Although many of the published studies of genecological differentiation in the Californian flora have emphasized the large scale, regional patterns, there have been examples of investigations on a more local scale, and some have offered comparative data. Thus Grant’s analysis of

FORTY YEARS OF GENECOLOGY

187

the variation of Cilia achilleaefolia (1 954) provides an interesting contrast with his study of G. capitata, discussed above. G. achilleaefolia shows very great local variation from colony to colony throughout its range. According to Grant, the local variants “do not group themselves into broad geographical assemblages, as in most species including the related G. capitata, but the broader subdivision in the species is rather along ecological lines. The large-flowered races with dense heads occupy sunny hillsides in grassland and oak savannah ; the small flowered races with loose cymes occur in the shade of oak woodland or redwood forest; and there are numerous transitional forms in the semi-shade of open oak woods”. The situation here described is comparable with that encountered by Bradshaw in Agrostis tenuis, the pattern of differentiation following closely upon local ecology so that sometimes continuous and sometimes discontinuous variation is encountered. Cook’s study (1962) of Eschscholzia californica, the Californian poppy, is of interest since it combined a regional survey of thirty-one populations dispersed throughout the Californian range with a local survey over a 35 mile transect. Some of the data concerning population variation appear to be based upon samples from the wild, so that there is no guarantee that all variation recorded is genotypic; but in so far as it concerned floral characteristics the probability is that direct environmental influence accounted,for little. Longevity was assessed in samples under comparative cultivation. Each of the floral characteristics examined was found to vary gradually, sometimes independently and sometimes in a parallel manner ; no discontinuities were encountered suggesting the existence of regional races. Some of the clines observed could be related in a general way to major ecological gradients, but the evidence presented does not suggest that all necessarily do so. From the local study, it was concluded that much of the variation, particularly in such features as developmental periodicity, is closely related to habitat conditions, so that neighbouring populations may differ abruptly if the habitats are strongly differentiated, or intergrade, in a clinal manner, when the habitats themselves intergrade.

G.

OENECOLOQICAL DIFFERENTIATION : SOME GENERALIZATIONS

The foregoing examples illustrate most of the principal kinds of genecological differentiation known in higher plant species. The interpretation of the different patterns must necessarily depend upon an appreciation of the evolutionary situation each represents, and this in turn must be based upon some understanding of the nature of the response of plant populations to selection and the role of genetic systems in determining that response, matters discussed in Section 11. So far as the observational evidence is concerned the following generalizations may be made.

188

J. HESLOP-HARRISON

(a) The geographical distributional pattern of a species is a determinant of the pattern of genecological differentiation. Where an area is continuous, spanning, say, an appreciable range of latitude, clinal variation is likely to be found, particularly in long-lived species of stable communities; if there are discontinuities, regional races with distinct ecological tolerances may be encountered. ( b ) The total ecological range of a species is significant. I n a species restricted to the modulations of a particular kind of habitat, any ecological variation will tend to be clinal. If a species spans a range of discrete habitats which are themselves sufficiently distinct, ecological races of the nature of the ecotypes may be found, even when the habitats are contiguous. (c) Systems of continuous and discontinuous variation may be combined a t the same and at different leveIs. Thus : (i) A species may show differentiation into major ecological r&es associated with different types of habitat within its area, and within one or more of the races there may at the same time be ecoclinal variation adaptive to smoothly varying climatic factors. (ii) Local variation of an “ecotypic” kind may occur within the framework of a grand system of clinal variation extending throughout the species area. (iii) Conversely, within major ecological races local adaptive clines may be found. (iv) Hierarchical patterns of both ecotypes and ecoclines may exist : thus a regional ecological race may itself be composed of a system of more local ecotypes, and within a major ecocline local systems of clinal adaptation may be present. (v) I n widely ranging species, clinal variation adaptive to different factors may occur, giving in the extreme case, patterns of “intersecting clines”. The existence of complexities such as those listed under (c) above means that the results of any particular genecological study will tend to be related to the scale on which it is conceived. Thus the minutae that engage Bradshaw (1959, 1960) in his investigation of Agrostis tenuis in a small area of central Wales would pass largely unnoticed in a study on the geographical scale of that of Clausen et al. of Potentilla glandulosa, (1940). This is not to suggest that identical methods would reveal corresponding variational patterns in these two particular species, but merely to point out that the results so far obtained with them are not capable of direct comparison. We shall see that much can be deduced about the probable patterns of genecological differentiation within a species from a knowledge of its genetic system, distribution and general ecology, and it would seem an essential now t o take these factors into

FORTY YEARS O F GENECOLOGY

189

consideration in the design of genecological investigations so as to determine the scale and nature of the sampling programmes to be adopted to produce information of value in comparative studies.

H.

GENETIC BASIS O F ECOTYPIC DIFFERENCES

It is a commonplace of descriptive genecology that ecotypes tend to differ in characteristics such as growth habit and dimensions of organs which show continuous variation within populations. Moreover, the gradation between ecoclinally related populations in these features and in others such as developmental periodicity is itself usually found to be continuous, or so finely stepped as to be effectively so when the modulating effects of the environment are superimposed. These two observations are best explained by the hypothesis that ecotypic differences are mostly polygenically controlled (Mather, 1943). Discussing in 1953 the genetical structure of natural populations Mather noted that the direct evidence of polygenic systems mediating continuous variation was not abundant, most likely because the experimental demonstration of their presence is difficult. Among plant examples he quoted the species of Layia and Madia mentioned by Clausen et al. (1940), and the same authors’ comments on the ecotypic variation of Achillea borealis (Hiesey et al., 1942). The analysis now available of the genetics of the races of Potentilla glandulosa from the more recent work of the Carnegie group (Clausen and Hiesey, 1958) provides much the fullest evidence of the genetic basis of ecotypic differences. Most of the differential features are controlled multifactorially ; but the number of loci concerned is seemingly not especially high. I n the study, crosses were made between individuals of the subspecies typica (Californian coast), reJlexa (Sierran foothills) and nevadensis (subalpine form and alpine form). The most fully analysed crosses were between typica and the alpine form of nevadensis, and between rejexa and the subalpine form of nevadensis. I n the first of these, large F, and F, populations were studied in a standard garden, and in the second, F,, F, and F, populations were grown in the standard garden and in addition cloned F, individuals were compared at three transplant stations at different altitudes on the Californian transect. Up to nineteen characters were observed simultaneously in the experiments, and the huge body of data was processed by a punched card technique. Estimates of the minimum number of genes associated with the character differences observed between the different races were based upon comparisons of F, frequency distributions with binomial distributions. These estimates for nineteen characters are summarized in Table 11, from the original of Clausen and Hiesey. For the whole complex of differential features G2

C.E.R.

190

J. HESLOP-HARRISON

observed in the ecotypes studied, they estimate that about 100 or more loci may be concerned. “Considered in terms of the complex array of observed recombinations that can be accounted for in the F, and F, combinations”, they comment, “this number seems to be remarkably small. Thus with a rather limited number of basic genetic building blocks, the diploid species Potentilla glandulosa apparently has been able to evolve an extensive array of ecological races fitted to diverse environments. ”

TABLEI1 Estimate of Minimum Number of Genes Governing the Inheritance of 19 Characters in TwoInter-ecotypic Hybrids of Potentilla glandulosa. [From Clausen and Hiesey 1958a.l Estimated no. of gene pairs

Character, and action of genes 1. Orientation of petals :2 erecting, 1reflexing 2. Petal notch: 1 producing notch, 2 inhibiting 3. Petal colour :2 whitening, 2 producing yellow, 1bleaching 4. Petal width: 4 widening, 1complementary, 1 narrowing 5 . Petal length :4 multiples

3 3 5 6 ca. 4

(plus possible inhibitors)

6. Sepal length: 3 or 4 multiples for lengthening, 1 for shortening, 1

complementary

7. Akene weigh&: 5 multiples for increasing, 1 for decreasing 8. Akene colour :4 multiples of equal effect 9. Branching, angle of 10.

11. 12. 13. 14. 15. 16. 17. 18. 19.

m.5 ca. 6 4

ca. 2 (alsogenes for strict to flexuous branching) Inflorescence,density of ca. 1 (plus modifiers) Crown height ca. 3 (also genes for presence or absence of rhizomes and for thickness of rhizomes, to which crown height is related) Anthocyanin: 4 multiples (1 expressed only a t Timberline), 1 complementary 5 Glanduhr pubescence :5 multiples, in series of decreasing strength 5 Leaf length :transgressive segregation; many patterns of expression in contrasting environments; possibly different sets of multiples ca. 10-20 activated Leaflet nurnber in bracts ca. 1 ’ ‘(plusmodifiers) Stem length: transgressive segregation, 5 to 6 multiples plus inhibitory and complementary genes; many patterns of expression in ca. 10-20 contrasting environments Winter dormancy: 3 multiples of equal effect 3 Frost susceptibility :slight transgression toward resistance ca. 4 Earliness of flowering :strongly transgressive ; many patterns of altitudinal expression; possibly different sets of genes activated many

FORTY YEARS O F GENECOLOGY

191

The comparison of the performance of cloned F, individuals from the reflexa x subalpine nevadensis cross at the different transplant stations revealed notable differences in the effect of environment upon gene expression. Clausen and Hiesey refer to “qualitative” characters, mostly expressed in the flowers, which although subject to some modification at the different stations are yet relatively stable, and to “quantitative” characters modifiable by environment to the extent that genic differences may be totally masked. These modifiable characteristics mostly relate to growth properties and developmental periodicity. I n stem growth, a high correlation (r = 0.48) was observed between performance at Stanford (30 m altitude) and at Mather (1400 m), and a lower one (r=0.12) between Stanford and Timberline (3050 m). I n respect to flowering time, the F, progeny showed scarcely any significant correlation between the stations, although at each the order of flowering between the different individuals was consistent from year to year. It may be surmised that both penetrance and expressivity of the genes segregating in the F, of the reflexa x subalpine nevadensis cross vary in the environments of the three transplant stations, but Clausen and Hiesey comment that the genic systems controlling the quantitative characters have proved too complex for any satisfactorily comprehensive interpretation of the effects of the diverse environments on gene expression. Another feature of great interest in the work of the Carnegie group is the analysis of the correlations between the characters distinguishing the ecological races in F, progeny of inter-racial crosses, (Clausen and Hiesey, 1958a, 1960). These are expressed diagrammatically in Figs. 2a and 2b for the two fully analysed crosses, typica x alpine nevadensis and reJlexa x subalpine nevadensis. In the first of these crosses, of the 91 possible combinations of the 14 characters, 67 showed some degree of association. I n the second, the 12 characters permitted 66 combinations, of which 38 were significantly correlated. Correlations of the type observed could conceivably arise at a physiological level, and it may well be that some of those illustrated in Fig. 2 do represent different expressions of single basic growth processes. However, Clausen and Hiesey consider that evidence points to some form of genetical association between the characters, which they refer to as coherence. The chromosome number of P.glandulosa is low ( n = 7 ) , and they attribute the coherence of racial characters to the high probability that, with so small a number, some at least of the genes governing any pair of quantitatively varying, polygenically governed characters will be linked. They have shown very convincingly that the coherent constellations of characters studied are either themselves concerned with survival in the habitats of the different ecological races, or cohere also with unobserved physiological properties which determine survival.

192

J. HESLOP-HARRISON Winter

length

Petal colour

Petal

Leaf

number

length

FIQ.2 . ( a ) Statistical correlations between pairs of characters in 992 F, progeny of Potentilla glandulosa subsp. 72evadensw x P . glandulosa subsp. tgpicu. ( b ) Statistical correlations between pairs of characters in 570 F, progeny of Potentilla glandulosa subsp. nevademia (subalpine form) x P . glandulosa subsp. rejexa (foothillform). I n each figure, a heavy line indicates r = 0.25-0.80, a light line, r = 0.09-0-25 and a broken line, r = 0-0&0.09 (insignificant).(From Clausen and Hiesey, 1958.)

193

FORTY YEARS OF GENECOLOGY

The evidence for one cross, reftexa x subalpine nevadensis, is given in Table 111. The F, progeny carrying the highest proportion of nevadensis

TABLEI11 Percentages of Long-period Survivors at Three Transplant Stations, Stanford (warm temperate ; altitude 30 m), Mather (subalpine, 1400 m) and Timberline (alpine, 3000 m ) among Three Classes of F Progeny of the Cross Potentilla glandulosa subsp. nevadensis (subalpine form) x P. glandulosa subsp. reflexa (foothill form). The clones were classiJied according to a n indw based upon 12 phenotypic characters in which the parental populations differed. Long-period survivors are ramets of clones that survived over 5 years at Stanford and 9 years at Mather and Timberline. [From Clausen and Hiesey 1958.1 Characterization of index classes Most nevadenais-like Intermediate Most reJexa-like

Percentage of long-periodsurvivors Stanford

Mather

Timberline

17.8 77.0 70.8

49.5 71.5 76-5

75.6 34.3 13.5

No. of clones

90 330 89

features survive best in the subalpine habitat, and those with rejexa characters are most viable in the natural foothill environment of this race. The existence of these closely knit systems of character correlation in the ecotypic races of P. glandulosa has reinforced the view of Clausen and Hiesey that they are distinct, non-intergrading, units, Thus, “combined with natural selection, moderate genetic cohesion within the characters of a race tends to ensure its perpetuation as an evolutionary entity. When morphological characters are included in the coherence systems, morphologically distinct ecotypic subspecies may result, as in Potentilla glandulosa” (Clausen and Hiesey, 1958). Evidently they consider that the genetic cohesion together with the sharp differentiation of the habitats on the central Californian transect preserve the identities of the races, a situation that would be most readily reconciled with a monotopic rather than a polytopic origin of each. The point is reverted to in a later section (p. 212).

11. EVOLUTIONARY ASPECTSOF GENECOLOGICAL DIFFERENTIATION

A.

INTRODUCTION: THE ORIGIN, STORAGE AND RELEASE OF VARIATION

Turesson’s definition of the ecotype (1925) as the product “arising through the sorting and controlling effects of the habitat factors upon

194

J. HESLOP-HARRISON

the heterogeneous species population” shows that he looked upon adaptation to habitat as the outcome of selection ; but, as Baker (1953) has pointed out, he and other pioneer workers were handicapped by imperfect conceptions of the nature of the variability of plant populations, and were accordingly unable to formulate at all clearly how selection was likely to operate in bringing about the differentiation of habitat races. At least in his earlier papers Turesson wrote as if he looked upon the typical plant species as consisting of a generalized population plus specialized “radiations” from it, the ecotypes. The generalized population was rich in “biotypes” (defhed by Faegri, 1937, as groups of individuals with identical genetical constitution), while the ecotypes were “biotype depleted”. This conception was opposed by Faegri (1937), who objected to the idea that biotypes pre-adapted for alpine or other extreme conditions could survive in an undifferentiated lowland population awaiting, as it were, their opportunity to invade a mountain habitat, and argued further that if specialization necessitated genetical impoverishment ecotypes would be populations ultimately destined to become extinct rather than to be the starting points of new species. Much of the early difficulty lay in the impression that genetical variation in plant populations must necessarily be largely overt. The species consisted of biotypes (sometimesreferred to as though they constituted something like self-propagating pure lines, even in outbreeding species); and the biotypes were the units which were selected or rejected in the course of adaptation to habitat. VC’ithout some appreciation of the subtle ways the ebb and flow of variation is governed in plant populations no other standpoint was possible. Current understanding of the way genetical variation may be generated, recombined, exposed, conserved, concealed and lost in the course of sexual reproduction may be dated from two principal publications -Darlington’s “Evolution of Genetic Systems” of 1939, and Mather’s paper on polygenic inheritance and natural selection of 1943. More recent discussions of these processes in higher plant populations are given in Stebbins’ “Variation and Evolution in Plants” of 1950, and in review papers by Baker (1953, 19591, Stebbins (1957,1958) and Grant (1958). In general, adaptive changes in a population must depend upon the accumulation of minor genic changes, mutations, under the directiye action of selection. As Mather (1943) has shown, a compromise will always be involved between fitness for the environment as it exists, and the flexibility which will permit further adaptive change. Fitness is best served by the production of progeny optimal for the immediate circumstances; flexibility by the continuous generation of variants some of which may be optimal for environments only to be encountered in the future, or elsewhere on the earth’s surface. The raw

196 material for variation is genic change ; the devices through which variability is regulated in populations constitute collectively the genetic system (Darlington, 1939). The three principal components of the genetic system may be identified as : (a)the chromosomal system, which establishes the rate of gene segregation and recombination; (6) the breeding system, which governs the level of hybridity; and (c) the intrinsic and extrinsic factors determining in interaction the sizes of breeding groups. We will consider the first two components here ;the third is discussed in a later section (p. 204). Darlington (1939)’hasreferred to the “sum of the haploid number of chromosomes and of the average chiasma frequency of all the chromosomes in a meiotic cell” as the recombination index. A high index must determine a rapid rate of gene segregation and recombination, and so a rapid flow of variation from the concealed to the overt form and back again ;a low index must in contrast establish a slow rate of turn-over of variation. I n terms of the fitness-flexibility compromise, a high index gives flexibility, at the expense of the continuous destruction of any gene combinations giving fitness in the prevailing milieu, while a low index preserves any existing state of fitness at the sacrifice of future adaptive potential. The balance between homozygosity and heterozygosity in a population is established by the prevailing system of breeding; in higher plants, that is to say, by the various properties determining the habitual mode of pollination. Cross-pollination may be promoted by sex differences, dichogamy, heterostyly, and incompatibility mechanisms including certation phenomena; and self-pollination may be imposed by cleistogamy or mechanical adaptation of the flower ensuring direct transfer of pollen from anther to stigma. I n many plants mixed breeding is the rule, either, as Mather (1943)has pointed out, because of the inefficiency of a cross-pollinatingmechanism, or because of the existence of a genetically determined system enforcing a regulated ratio of self- to cross-pollination. Such regulated systems must be subject to selection, but retrospectively, like other components of the genetic system. Heterozygosity is a necessary prior condition for recombination and for the release of concealed genetic variation ; cross-pollinationthus acts like a high recombination index to promote variability. Self-pollination by raising the level of homozygosity restricts recombination and checks the flow of variation, in the extreme case leaving it eventually distributed among homozygous pure lines. Cross-pollination and high recombination index stre complementary in promoting the flow of variability, but low recombination index and selftng constitute alternative means to one and the same end, namely the enhancement of genetic stability over a succession of generations. FORTY YEARS OF GENECOLOGY

196

J. HESLOP-HARRISON

One aspect of the function of the genetic system is the regulation of the pheno5ypic expression of genetic variation. Latent variation may be carried in the heterozygotes of a diploid population in the form of recessive genes, which will reach expression only with the segregation of homozygotes. Where phenotypic characters are governed by several genes each with small individual effects other possibilities for the concealment of genetic variation exist. Mather (1943) has shown how with such polygenic systems a constant phenotypic frequency distribution in a population can be maintained by selection with the simultaneous preservation of a high level of genotypic variation. Because a given phenotypic expression can be determined by several polygene combinations, the selectively favoured group in each generation may be essentially uniform in adaptive characters, but will always transmit genetic diversity to their progeny. Moreover, isolated populations may attain phenotypic similarity under similar selective pressures by the assembly of different polygene complexes each in a more or less homozygous state ;between them they will then preserve a reservoir of genetic variation which will be tapped only when they pass back into breeding contact. I n governing the flow of variability in a population, the genetic system necessarily affects the potentialities for response to selection; it is thus one of the major factors to be taken into account in seeking to understand the origin of adaptive variation in plant species. Baker (1953, 1959) has considered the effects of differences in reproductive methods on race formation, but he acknowledges that it would be equally possible to “examine the consequences of the formation of certain kinds of race upon breeding behaviour of the plants”. Indeed it is barely justifiable to separate intrinsic and extrinsic aspects, for they are always in interaction. The genetic system in the short run may determine the pattern of response to the environment, but environmental selection acting in a retrospective mode will determine what kinds of genetic system will prevail in the ultimGte )mrvivors in any particular ecological situation. Nevertheless, however tangled may be the chain of cause and effect, patterns of infraspecific variation will inevitably show some relationship with genetic systems and the extrinsic factors impinging upon them. B. MODES O F SELECTION Mather (1953) has termed the three basic modes of selection stabilizing, directional and disruptive. Under stabilizing selection the mean of the phenotypic distribution is favoured at the expense of the extremes; with directional selection, one extreme is favoured; and with disruptive selection both extremes. It is obvious that in one and the

197 same population selective forces could act in a stabilizing way on some phenotypic features and, simultaneously, in a directional or disruptive way on others. Moreover, multipolar disruptive selection could be imagined in which more t,han two expressions are simultaneously favoured. I n considering how these selective situations are realized in populations of higher plants, some sense of the kinds of compromise involved in long and short term adaptation is necessary. A plant population which survives in a given habitat must do so in the face of various inimical secular factors and against the predation and competition of other organisms. Since survival depends upon a variety of properties (some, like the needs to maintain gas exchanges and yet control water loss, even antagonistic) a degree of compromise on the physiological level must always be involved. The success of that achieved by any given phenotype could be measured by various criteria, but so far as the persistence of the population is concerned reproductive performance is obviously a principal one. I n the relative sense in which the concept of adaptation is commonly applied, we might say that a population is well adapted to its habitat when it succeeds a t least in maintaining its numbers through successive generations. If this is achieved over an interval of time in an outbreeding population without phenotypic change it may be supposed that the selective forces at work are acting in a stabilizing way to eliminate phenotypes deviating from what is evidently an adaptive mode. The extent of restriction of the phenotypic distribution provides a measure of the intensity of selection. Stringent selection must involve the loss of genetic variation, but as we have seen where polygenic systems are concerned some will always be conserved in a cryptic state. Now a further element of compromise is involved in relation to environmental change. Plant habitats are subject to regular cyclical changes of different periods ; to random fluctuations, again of different periods; and to long term trends of change. Plant populations may be accommodated to these changes in various ways. Taking first the nondirectional changes, a relationship of the first importance is between the duration of the life span and the period of the cycle or fluctuation. If the life span is such as to exceed the period of cyclical change, or the average period of some habitually fluctuating environmental variable, it must be supposed that a level of physiological adaptation has been achieved which permits the effects to be absorbed. The patterns of developmental periodicity seen in perennials, discussed more fully in a later section (p. 227 et seq.), represent physiological solutions to the problem of accomodating annual variation in climate. Forest flora must also preserve FORTY YEARS OF GENECOLOGY

198

J. HESLOP-HARRISON

a capacity to tolerate fluctuating climatic change of longer periods, such as the “sun-spot” cycle recorded in the annuals rings of the redwoods, and the irregular variations in winter temperatures and precipitation that have been a feature of temperate latitudes in post-glacial times (Manley, 1954). Either because the periods favourable to growth are too short, or because the amplitude of variability of environmental factors is beyond tolerance, some habitats will not permit the survival of perennials. However, adaptation to ephemeral habitats or to severe cyclical or fluctuating changes can be attained by curtailment of the life cycle. The annual habit represents a satisfactory solution to the problem of exploiting temporary habitats, or of surviving in a climate where there is extreme but regular seasonal variation in vital habitat factors such as temperature and rainfall. By fitting in with the major cycle of change the annual avoids the requirement for adaptations giving tolerance to the adverse season, which can be safely weathered in the seed ; it is left then to face the longer and shorter period fluctuations. The methods available to achieve this are of three general kinds. The “annual” cycle can be distorted to accommodate irregular longer cycles ;this response is seen in species of arid regions with sporadic rainfall, where the amount of precipitation is itself a determinant of germination time. Or the developmental plasticity of the individuals may be such that the phenotypes produced each year are adapted to match the conditions of that year. Or, finally, the average phenotypes may possess an amplitude of tolerance great enough to accommodate year-to-year fluctuations in climate during the growing season. The ephemeral habit is an adequate solution to survival in habitats subject to severe and irregular disturbance. It commonly combines the ability to suppress growth altogether during unfavourable periods, a property which we may term avoidance, with adaptive developmental plasticity and a high amplitude of individual tolerance. These different kinds of physiological adaptation to change necessarily bear some relationship to community structure and stability. The long-lived perennial habit is that characteristic of the stable climax community such as forest or permanent grassland. I n the case of forest, the moderating and controlling effects of the dominants themselves on soil and microclimate constitute a damping factor stabilizing the environment still further for tree seedlings and the herbaceous members of the community. So far as the tree species are concerned, this is, of course, an aspect of the “independence of environment” designated by Huxley (1 942) as one of the marks of evolutionary progress. The annual habit must necessarily be associated with a less uniformly favourable environment and accordingly with communities of a less stable kind,

FORTY YEARS OF GENECOLOGY

199

while the ephemeral habit, representing as it does mere opportunism, must be associated with the most labile of all kinds of community. These relationships between life-span, non-directional environmental change and community type have important implications for the operation of selection. I n the stable community, it is obvious that selection will tend to act rather consistently in a stabilizing mode. I n the less stable communities of fluctuating environments selective pressures are, in contrast, likely to be erratic -with the annual, because of year to year climatic and other variations, and with the ephemeral because of the seasonal changes themselves and other more catastrophic events. What is to be considered the “optimum” will depend on the period of time over which the selective influences are integrated. The time scale of change in relation to longevity is again of paramount significance where longer-term, directional, environmental change is concerned. A long-lived perennial will be exposed during its life-time to various environmental fluctuations. If the habitat is truly stable, the average experience of each generation will be the same. If there is a trend of some kind, the change will emerge as a progressive shift of the average. The effect of selection will be to favour those phenotypes conforming in each generation to a gradually changing optimum. A model situation may be seen in forest migrations of the post-glacial. Climatic changes meant the shift of climatic belts latitudinally. During the quaternary in the northern hemisphere the changes were slow enough to allow the major forest communities of long-lived species to migrate in step. The migrations took place through the replacement of one species by another in consequence of climatic selection, but, simultaneously, intraspecific selection achieved the necessary adjustments of developmental periodicity to fit the forest dominants to their new latitudes. The situation of short-lived species, with their diverse patterns of physiological adaptation to short-term environmental fluctuations, differs in that the generation time is not sufficient to ensure an automatic averaging over several cycles of change. Yet the average condition may be changing progressively; and survival of the population must depend upon the ability not only to accommodate the fluctuations but to shift in step with the average. If this is to occur, the response to selection must be regulated in such a way that the transient influences are without effect in producing short term changes in genetical structure - which would involve a sort of evolutionary “hunting” like that of a cybernetic system adjusted so as to give excessive feed-back -while the long-term trends are met by progressive, smooth adjustment of the average phenotype. All of the foregoing considerations involve aspects of Mather’s fitnessflexibility compromise, and we may now examine how the different

200

J. HESLOP-HARRISON

situations may be met. The conditions of the stable community, with strong centripetal selection towards a progressively but slowly changing optimum, favour the free release of variation; by this means a perennial with high reproductive capacity can purchase long-term evolutionary advantage at the cost of a great sacrifice of progeny. I n terms of the genetic system, this means high recombination index combined with cross-pollination. I n temporary habitats, or where environmental fluctuations are beyond the tolerance range of plants of a perennial habit, the advantage will lie with the production of large, relatively uniform populations through successive short generations, and a genetic stability permitting the preservation of any adaptive peak achieved without excessive segregation and without too rapid a response to selective influences of a transient nature. This implies a genetic system parsimonious in the phenotypic release of variability, which means a low recombbation index, or inbreeding, or, in the extreme case, apomixis. These various relationships are summarized in Table IV. Most have been documented and exemplified by Stebbins (1950, 1957, 1958) and Grant (1958), and some have been discussed by Baker (1953). It may be noted that some of the conclusions of Thoday (1953) in a discussion of the components of fitness seem to be at variance. Evolutionary change arising sui generis has been ignored in the foregoing discussion. As the history, for example, of the tropical rain forests

TABLEI V Relationship between Some Components of the Genetic System and Longevity, Community Type, Physiological Features and Variation Pattern Free gene recmbinatim resulting from a high recombination index and outbreeding. A880Ck&d with : (a)Protracted life span of individuals reaching maturity. (b) High and selective seedling mortality ;dispersal capacitymoderate to poor. (c) Membership of closed climax communities in stable or but slowly changinghabitats. ( d ) High range of physiological tolerance in the adult, frequently dependent upon precisely adjusted developmental rhythms but not necessarily associated with phenotypic plasticity. ( e ) Continuous, clinal variation.

Reatricted gene recombination resulting from a low recombination index or inbreeding. Associated with : (a)Short life span of individuals. (b) Low seedling mortality during colonization of suitable habitats ; efficient seed dispersal. ( c ) Membership of impermanent communities in ephemeral, fluctuating or shifting habitats. (d) Avoidance of adverse conditionsin the seed state, with or without an associated capacity for adaptive plastic response to environmental stress. (e) Discontinuousvariation.

FORTY YEARS OF OENECOLOGY

201

reveals, a long-maintained, favourable environment with no marked trend of secular change does not mean evolutionary stagnation. This is presumably because there must always be ways in which an organism can steal a march on others even, as it were, in Utopia. Stabilizing selection must favour the “optimum” phenotype from the phenotypes available in each generation, but given a continued release of genotypic variation in phenotypic form some recombinants may attain new and better optima in respect to the capacity for exploiting the very same environment. A trend of change may thus be initiated within the community ;and in so far as a change in one species produces a new environment for others it may be supposed that the pattern of selective forces will be in continuous flux. A final matter meriting consideration in the context of this section concerns selection in the course of migration. There can be little doubt that the genecological differentiation observable in north temperate species is the product of evolutionary change during or following immediately upon fairly recent species migrations of a major scale. I n general terms, it is possible to homologize the progressive change in the character of the environment that a species is likely to meet at the periphery of an expanding area with a temporal environmental trend experienced in one locality. Indeed, for the short-lived species of temporary or fluctuating habitats there would seem to be no essential difference between the two situations ; the problems of matching an existing environment with enough fit genotypes, while at the same time generating sufficient diversity to permit colonization of slightly different ones, remains the same. With the perennial pre-adapted for the stable, closed community there are some significant differences. The principal one is that in the van of a migration centripetal selection is eased to the extent that dispersal is into an open community. The case is analogous to that in which a population is in a phase of rapid increase of numbers : variation may be expected to rise (Mather, 1953). Moreover, the “optimum” phenotype of the open marginal community may not be that of the main closed community, so the phenotypic distribution of the colonists may not only reveal a higher variance but a different mean. The exact implications of this situation have not hitherto been worked out, although it may be supposed that the effect will mainly be transient since the progressive closing of the community may be expected to restore the original balance of selective forces. The matter is considered again in a later section (p. 209).

c. VERSATILE

REPRODUCTIVE SYSTEMS

Obligate self-pollination,by putting an end to recombination, must in the long run jeopardize the survival of a species; it might therefore be

202

J. HESLOP-HARRISON

interpreted as an evolutionary cul-de-sac (Darlington, 1939). It seems probable, however, that autogamy is rarely complete, and many species appear to benefit from an ability to interpolate outbreeding episodes between long periods of selfing, the condition Fryxell (1959) has called cyclical autogamy. The outbreeding episodes result in immediate increase in the level of heterozygosity, leading to the segregation of new recombinant genotypes in future generations. I n offering new pabulum for selection, this facilitates further ecological adaption. The restoration of the autogamous habit rapidly reduces heterozygosity again, but the opportunity is available for the fixation of new adapted variants as pure lines. The sporadic release of variation in this manner is well-known both in wild and cultivated autogams. The control must necessarily be an external one, and several environmental factors, alone or in consort, may be concerned. Environmental control of cleistogamy has been known since Darwin’s day, and Uphof (1938)has listed many examples. Cleistogamy is effective in promoting self-pollination, but it is not necessarily an adaption to this end, since it can also serve as a device permitting reproduction under circumstances adverse enough to preclude normal pollination (Stebbins, 1957).I n many described examples, the norm of a species is chasmogamy and outbreeding, cleistogamy resulting from unfavourable conditions of soil or moisture supply (e.g. stipa leucotricha, Brown, 1952). Of more interest from the evolutionary point of view are cases where the norm is self-pollination, cross-pollination being the rarer alternative. I n the facultatively cleistogamous grass Bromus carinatus, Harlan (1945a, b) found that chasmogamous panicles were formed only under the most favourable growth conditions. Most wild stands sampled in California and Arizona consisted of one or a few practically homozygous lines, but highly heterozygous hybrids were occasionally produced in consequence of chasmogamy and outbreeding. These hybrids Harlan considered could act as the progenitors of “race swarms’’ from which new, uniform populations might be derived. A similar variation pattern to that described in Bromus carinatus typifies many of the short-lived tropical and sub-tropical Andropogoneae, and environmentally governed facultative inbreeding mechanisms have been observed in these also. Bothriochloa decipiens, reported by Blake (1944) as commonly cleistogamous in the wild, possesses an adaptation ensuring the discharge of pollen within the flower. This takes the form of a glume pit which prevents the emergence of the single anther and ensures its dehiscence in contact with the stigmas if the glumes do not open at anthesis (Heslop-Harrison, 1961). The extension of the stem determines whether or not the inflorescence is exserted, and so the possibility of chasmogamy; this is subject to control by photo-

FORTY YEARS O F QENECOLOGY

203

period and temperature. I n another grass of the Andropogoneae, Rottboellia exaltata, both the incidence of cleistogamy and the total production of pollen are affected by the photoperiodic regime in which plants are grown (Heslop-Harrison, 1959b). The tendency towards selfpollination in these grasses is certainly to be associated'with the occupancy of habitats of limited permanency in areas of forest. It is of some interest that they can be brought into a state permitting outbreeding by environmental factors, for their ecologicalhabit may demand not only the ability to build up large homogeneous populations when a favourable habitat is available, but also to achieve, rather regularly, enough recombination to permit the colonization of new habitats suddenly made accessible by local catastrophic change in the forest cover. Rottboellia exaltata shows great geographic variation, sometimes on a very local scale (Hubbard, personal communication), and it would be informative to know to what extent this is ecologically correlated. The matter of ecological adaptation in apomictic complexes merits comment. Obligate apomixis, by suspending entirely the capacity for recombination, must, like obligate inbreeding, presage ultimate extinction once the circumstances for which a,lineage is adapted cease to be available. Yet many major apomictic complexes have achieved considerable ecological success, and have done so by what is essentially genecological differentiation (Turesson, 1943, 1956; Nygren, 1951 ; Clausen, 1954). This must mean that some capacity for attaining recombination remains, and it is probable that in all successful apomictic complexes some sexual potential is retained. This could be through the survival of some sexual races acting as progenitors of apomicts, or through the ability of mainly apomictic lineages to produce some progeny by sexual processes (Gustafsson, 1947; Stebbins, 1950). Haskell (1953, 1959), has convincingly demonstrated quantitative variation in ecologically significant characteristics like flowering time in the progeny of the pseudogamous, largely diplosporous apomict Rubus nitidioides, and has shown a response to selection. The parental clone is highly heterozygous, and Haskell suggests that the limited and irregular segregation observed might be explained by recombination during oogenesis (autosegregation).With aposporous apomicts, complete sexual competence may be preserved, so that the progeny are partly sexual and partly clonal. Apospory is very widespread in grasses of the Paniceae, Andropogoneae and related groups (Brown and Emery, 1958), and it may be surmised that it is commonly facultative. I n Dichanthium aristatum of the Andropogoneae, the balance between sexuality and apospory is subject t o environmental control (Knox and Heslop-Harrison, 1963), and there is evidence that it varies from population topopulation in the wild (Knox, personal communication).

204

J. HESLOP-HARRISON

It is apparent that a combination of sexuality and apomixis could be specially advantageous in some ecological circumstances through offering simultaneously the ability to attain a high level of recombination and the capacity to reproduce en mmse any well adapted genotypes that may emerge (Heslop-Harrison, 1959~). I n allowing the immediate fixation of an adapted genotype whatever its level of heterozygosity the system has advantages over cyclical autogamy, where genetic stability is only achieved with an approach to homozygosity. A closely comparable situation exists where a species is capable of efficient vegetative reproduction. A successful genotype, in achieving vigorous growth, is immediately at a competitive advantage irrespective of its reproductive performance through seed. If propagation is through stolonifery or other direct vegetative means the dispersal capacity may be limited, but an aggressive clone may be expected to spread within the physical limits of the site to produce a highly homogeneous stand (Harberd, 1961a). It is significant that these versatile reproductive systems are commonly found in species - notably the perennial grasses - occupying habitats of moderate permanency but of some ecological diversity. I n successional terms, they are neither ephemeral pioneers, nor yet necessarily components of climax vegetation. The fact accords well with the theoretical advantages to be expected from the special properties of their reproductive systems. D.

ISOLATION A N D GENECOLOGICAL DIFFERENTIATION

The relationships set out in Table I V are obviously of great importance for the interpretation of patterns of genecological differentiation in plant species, particularly in connection with the much-debated matter of ecoclinal as contrasted with ecotypic variation. We see that, in general, the group of circumstances set out in the left-hand column of the table will tend to favour continuity and so an ecoclinal type of variation, while those in the right-hand column will favour discontinuity and so a pattern of variation interpretable in terms of discrete ecotypes. Table IV does not, however, account for all the agencies which may act to generate variational discontinuity within a species, because it does not cover those factors which impinge upon the operation of the genetic system through their effect on the extent of breeding groups. I n the obligate self-pollinator, the breeding groups constitute individuals. In cross-pollinators, the size of the breeding group may be affected by the ecological habit of the species, which establishes the density of individuals and the potential geographical range; by the nature of the terrain, which controls the continuity of distribution; by the dispersal

FORTY YEARS OF GlENECOLOQY

205

range of pollen and propagules, and by influences governing reproductive periodicity. At the one extreme are “colonial” species, occurring in panmictic populations perhaps numerically large but isolated spatially from others. This pattern is necessarily imposed on a species when suitable habitats are widely scattered in comparison with the average dispersal range of pollen and propagules. At the other extreme are species which tend to form continuous wide-ranging communities. Within these, the probability of any two individuals mating depends largely upon their spatial separation, and it may not be feasible to define breeding populations lesser in extent than each major distributional area, as is commonly true with widely distributed forest trees. However, it is possible to overestimate the extent of what might be termed “simultaneous” panmixis in these species. The fact that individual pollen grains can be carried great distances may be of limited significance, since the chance of fusion between gametes of remote provenance is determined by the proportion of foreign and local pollen in the atmosphere, and in a closed community the latter is always bound to predominate overwhelmingly. The effective pollination range may thus be quite small. Colwell (1951) found that extremely little of the pollen of Pinus coulteri released at a height of 12 ft was carried more than 150 f t even downwind; and Bateman (1947), following gene flow rather than pollen movement directly, found that with maize a distance of 60 f t was enough to reduce crossing to 1%. Moreover, the phenological gradients in continuous wide-rangingpopulations of potentially inter-fertile individuals become major barriers to long-range crossing, since they limit the distance over which a grain released at any one point is likely to encounter a receptive stigma. Latitudinal adjustment to photoperiod will thus contribute to limiting panmixis. The conclusion must be that in a continuously dispersed, wind-pollinated species the average area within which there is an appreciable chance of two individuals mating will normally be quite small in relation to the total range. What is important is that pathways for slow, continuous gene migration do exist, uninterrupted by barriers, through large parts of the species area. Most distributional patterns lie between these extremes, with local concentrations of individuals separated by thinly inhabited belts corresponding to topographical barriers or zones of unsuitable ecology. In terms of gene flow, whether such a situation resembles one or other extreme depends again upon the average dispersal ranges of pollen and propagules. With insect pollination, ecological limitation of the vectors may be so strong as effectively to exclude gene exchange between neighbowing concentrations of individuals ; “genetic mobility” (Darlington, 1939) may then IargeIy become a function of seed dispersion. On the

206

J. HESLOP-HARRISON

other hand, the effectiveness of wind-pollination and seed dispersal together may be such as to reduce substantially the isolating effect of local discontinuities, even when the distribution pattern is as fragmented, as in a coastal species like Plantago maritima (Gregor, 1946a). Any form of spatial isolation which is effective in restricting gene flow between two populations of a species will act to facilitate the independent response of the populations to local selective influences. Isolation must therefore always favour the establishment of discontinuities in the variational range of adaptive characters where the separate populations occupy dissimilar habitats. This will be true whatever the intrinsic properties of the genetic system, provided only that there is some release of genetic variation for selection to act upon. By establishing that the distributional pattern will be one of numerous more or less isolated populations, the ecological predilections of species with the “colonial” type of distribution mentioned above themselves create the conditions for ready fractionation into discrete ecotypes adapted for the minor modulations of the characteristic habitat. Even in the case of outbreeding species with tendencies t o form continuous stable communities and to show clinal variation, the occurrence of a topographical barrier imposing some degree of reproductive isolation will favour the appearance of variational discontinuity, producing regional ecological races, or “stepped” ecoclines (Gregor, 1944). These propositions concerning the role of spatial isolation as a determinant of discontinuity can, of course, be traced back to Darwin and Wagner, and they form a well established part of genecological lore. There is, however, the important question as to whether or not spatial isolation is a sine qua non for the development of variational discontinuity. The view of many contemporary evolutionists is that it is, and that divergence in geographical isolation must necessarily precede the evolution of the genetically determined bars to crossing that mark the final step of speciation. This opinion has been especially canvassed by animal systematists (e.g. Mayr, 1942, 1947, 1959; Cain, 1954). There is now, however, excellent evidence from the experimental work of Thoday (1959), Thoday and Boam (1959) and Thoday and Gibson (1962) to show that disruptive selection as defined by Mather (1953, 1955) can produce what is effectively a racial divergence in the face of a very high level of intercrossing. Thoday’s experiments were carried out with Drosophila melanogaster, but the results have immediate interest for the plant situation. One experiment (Thoday and Gibson, 1962) began with eighty flies of each sex from a wild strain. From each sex the eight flies with the highest sternopleural chaeta number and the eight with the lowest were selected. The thirty-two flies were permitted to mate at random during a period

FORTY YEARS OF QENECOLOQY

207

of 24 h, after which the eight high and the eight low chaeta number females were separated. From their progeny, the eight low and eight high flies of each sex were selected and mated a t random, and the process was repeated in each generation. By the fourth generation of selection, almost all the high flies selected came from the progeny of high females and almost all the low flies from the progeny of low females. By the twelfth generation, the distribution curves did not even overlap. I n this experiment, mating was not enforced in any particular pattern, and initially it must have been entirely at random. As the experiment progressed, reproductive isolation developed, limiting the formation of hybrids between the high and low modes. I n an earlier experiment (Thoday and Boam, 1959) mating was enforced between low and high lines so as to ensure the maximum amount of gene flow; significant divergence’ still occurred in consequence of the disruptive selection. Thoday and Boam remark that the two sub-populations in this experiment are “in a formal sense, in the same relative situations as would be two parts of a population in a mosaic environment consisting of two distinct habitats arranged, for example, as a chequerboard. I n such a situation a population could in principle develop two different forms, one adapted to each of the component environments even if there were forced (50%) gene-flowbetween the two forms (ecotypes)”. It will be recognized that with higher plants there could be several forms of the basic disruptive-selection situation. For example, the dissimilar habitats could intermingle in the mosaic fashion envisaged by Thoday, or they could interdigitate, or they could occupy distinct provinces within the potential breeding area of the population. Again, the selective pressures in the habitats could differ in intensity, so that a wide range of phenotypes might survive in one and a very narrow range in the other. Taking an extreme case, it might be that the dispersal capacity of a species was such that throughout an area where two -habitats, A and B, occurred, seeding was essentially random. Selection in each generation would then establish that phenotypes appropriate to the conditions of habitat A survived there and were eliminated from B, and that those adapted to B similarly persisted in B and were eliminated from A. The breeding population would then be effectively dimorphic, although still panmictic. The parallel with Thoday’s (1962) experimental situation is close though not exact, since the dimorphic Drosophila breeding population was derived by selecting the extremes from the total population, whilst in the plant case the morphs are selected from the sample of seeds happening to reach each of the two habitats. Nevertheless, it is evident enough that if Thoday’s results can be extrapolated at all they mean that the occupants of each habitat in successive generations should pro-

208

J. HESLOP-HARRISON

gressively give rise to higher and higher proportions of phenotypes adapted to the same habitat. I n other words, ecotypes should evolve, even in the face of the cross breeding. Some additional factors need to be taken into account. Obviously the intensity of the disruptive selection will be a paramount factor in determining the rate of differentiation. This probably means that this kind of response is only likely to occur at any significant rate where extreme habitats are encountered; and it is noteworthy that some of the best examples of clear-cut ecotypic differentiation are found in habitats where selection is patently intense. The most familiar situation is the asymmetrical one, where a generally favourable habitat adjoins or is interpenetrated by another distinctly less favourable: a coastal belt of extreme conditions adjacent to an equable hinterland, or an area of serpentine soil in an otherwise edaphically normal region. The response here should be such as to produce a specialized but narrow group of phenotypes (at least so far as the adaptive characteristics are concerned), while permitting the survival of a much wider range in the adjoining less demanding environment. There is, of course, no reason to expect divergence in non-adaptive characteristics. It is to be noted that once divergence has been initiated other circumstances will subsequently develop to encourage it. Thus there will necessarily be a movement away from randomness so far as seed dispersal is concerned, since each surviving individual is likely to seed more intensively its own immediate neighbourhood and so its own appropriate habitat than more remote parts of the alternative habitat. I n so far as its adaptive characteristics are heritable, this will ensure that a higher proportion of adapted than non-adapted seedlings arise in each habitat. The same will apply with respect to pollen dispersal; mere contiguity in the individual habitats will guarantee some assortative mating. I n consequence, a higher proportion of recombinants giving more extreme phenotypes will be generated, so allowing a progressive directional change towards better and better adaptation. I n a habitat where selection is intense the product would ultimately be an assemblage of biotypes lying outside of the original distributional range altogether. The situation discussed above in which disruptive selective pressures arise within a population because of its occupancy of a heterogeneous environment represents a special case, since it is envisaged that the" population is capable of a t least surviving in all of the diverse habitats ab initio. This cannot always be true. A capacity for phenotypic plastic adaptation may be important in permitting the invasion of some types of unusual habitat (p. 213), but beyond any extended tolerance range which this may provide, immigration into an environment offering novel and perhaps intense selective pressures will demand new geneti-

FORTY YEARS OF QENECOLOGY

209

cally adapted recombinants, perhaps not immediately available. “his must very frequently be the situation during the migration of a plant species into new territory, and it is one of a good deal more than theoretical interest considering that the north temperate floras have undergone massive oscillatory migrations during glacial and post-glacial times. It may be that the circumstance mentioned above (p. 201) is significant under these conditions, namely that during a period of migration a free release of genetic variation might be favoured in the van. This would arise if the intraspecific competition were heavy enough under the conditions of a closed community to restrict very severely the range of phenotypes reaching reproductive maturity in the territory already occupied. I n a thinly colonized marginal zone, some of the intense selective factors active in the closed community would be relaxed, with the consequent survival of new phenotypes. To the extent that these carried new gene combinations, they might be expected to enhance the segregation range in the pioneer belt. This availability of fresh recombinants would necessarily favour migration into novel habitats and the establishment of adapted races there, provided only that any adaptive gene complexes assembled were not immediately broken up by contamination from elsewhere. Mayr’s (1954) argument that peripheral populations would be inhibited from adapting successfully to new environments by the disruptive effect of gene migration from the interior of the species range has been contested by Thoday and Boam (1959), who on the basis of the experiments described above state that there is no reason in principle why a locally adapted population should not be formed even if one-way gene flow were so great that all progeny were hybrids. Evidently this particular facet of genecological differentiation would merit more observational and experimental study. There is another factor that is likely to be significant in promoting the differentiation of plant populations under disruptive selection, namely the effect of the immediate environment on phenology. I n the British Ecological Society’s series of transplant experiments, summarized by Marsden-Jones and Turrill (1938), some effect of soil type on date of flowering was observed in half of the species tested. Thus in 1932 a sample of Plantago major cultivated on “clay” reached maximum flowering by 19 June, but a corresponding sample of plants of the same parentage cultivated on “sand” in an adjacent plot did not reach a flowering peak until 15 July. Aspect may similarly produce phenological differences. Thus Mitra (personal account) working at Saskatchewan found up to a fortnight delay in the flowering of Stipa species on the north as compared with the south faces of an artificial east-west mound of 10’ slope. Ehrendorfer, in the study mentioned of Galiumpumilum(p. 186) { 1953) observed distinct differences over the small area studied in

210

J. HESLOP-HARRISON

flowering time; plants in the drier, open sites regularly reached anthesis before those in the moister, shadier localities. Here, of course, the possibility of some genetical differentiation for flowering time in the population is not excluded, but the circumstances as discussed by Ehrendorfer do suggest that the environmental control of phenology may have imposed some restriction on free gene exchange between the sub-populations even during the iktial differentiation. The implication of observations such as these is that the mere occupancy of a heterogeneous habitat can fragment a plant population into partly isolated breeding groups by breaking down the synchroneity of flowering. Since each distinct habitat will have its own particular effect, the sub-populations in each will have a characteristic flowering period. The plants in each type of habitat will thus remain panmictic, but will be genetically isolated from others in different habitats to the extent that the flowering periods fail to overlap. This is an irrefutable example of ecological isolation, not to be discounted on the basis of any of the arguments of Mayr (1947).

E.

MONOTOPIC AND POLYTOPIC ORIGIN AND THE RETENTION O F RACIAL IDENTITY

A distinction can be drawn at least on theoretical grounds between mosaic patterns of ecological races where each race has had a monotopic origin, and the superficially similar patterns which have arisen through the polytopic origin of like ecotypes in response to corresponding selective factors in different parts of a species range. It is to be expected that races that have had single independent origins will normally be distinguished by several correlated differentiae, both adaptive and nonadaptive. On the other hand, only the adaptive characteristics are likely to be shared in common between the various populations of an ecotype which has originated polytopically, and the populations may differ among themselves in respect to non-adaptive features. A classical case of presumed polytopic origin is that of the sand-dune ecotype of Hieracium umbellatum described by Turesson (1922b). This occurs in sand-dune habitats widely scattered around the periphery of the species area in Sweden. The different populations share in common various adaptive features such as a capacity for rapid shoot regeneration, but they differ in minor leaf characteristics, each showing a resemblance in these characteristics with neighbouring inland populations. The inhomogeneity of ecotypic populations of polytopic origin is likely to be evident also in the genetic basis of their adaptation. Since similar phenotypes can be established by different combinations of polygenes, there is no reason t o suppose that identical selective pressures in remote sites will necessarily fix the same group of genotypes. I n con-

"

FORTY YEARS O F GENECOLOGY

211

trast, if the various sites have been colonized by a single race the populations may be expected to be genetically homogeneous except in so far as secondary differentiation has progressed. Clausen (1951) reported a study of prostrate maritime populations of Layia platyglossa of the Californian coast in which an attempt was made to discover whether the complex had had a monotopic or a polytopic origin. Representatives of two phenotypically similar populations 140 miles apart were intercrossed, and a large F, family was raised. The adaptive characteristics -prostrate habit, succulence and late flowering - appeared uniformly in the F,, and showed no F, segregation. The F, was, however, variable in respect to other characters such as density of internodes and number of disk and ray florets. This result certainly suggests that the genetical basis of the adaptive characteristics is identical in the remote populations, yet the occurrence of some F, segregation does indicate heterogeneity. Clausen concluded that the maritime race does in fact, represent a single and rather ancient evolutionary entity. If so, the dissimilarities between the populations in non-adaptive characteristics must have arisen through secondary differentiation. Where two or more races, originally allopatric, have attained coincident or overlapping ranges by secondary migration it must be supposed that there are factors operative preventing miscegenation. The sequence is, of course, that widely accepted as being usually involved in speciation (Mayr, 1942) : divergence in geographic isolation, followed by the emergence of some form of reproductive isolation, followed in turn by further migration to produce overlapping distributions. It is not within the scope of the present paper to discuss the broader problems of isolating mechanisms and speciation, but some consideration of the factors acting to preserve the identity of sympatric ecological races is merited. Ecological specialization may itself form the most potent such factor. Where a predilection for highly distinctive habitats has been evolved, this must impose spatial isolation to the extent that the habitats themselves are dispersed geographically. Beyond this, the fact that each race has attained an adaptive peak for its particular habitat will mean that recombinants will be at a disadvantage in competition with their parents and uncontaminated progeny : they will become the victims of stabilizing selection in each habitat. While the pattern of selective forces remains unchanged, hybridization will be disfavoured. Many of the described cases of isolation imposed by strong habitat specialization concern pairs or groups of entities sufficiently distinct in morphology and distribution to have merited recognition as taxonomic species (p. 187). Well-documented examples are Quercus ilicifolia and Q. marilandica (Stebbins et al., 1947) and Silene vulgaris and S. maritima (Marsden-Jones and T u r d , 1957). With both of these pairs hybridiza-

212

J. HESLOP-HARRISON

tion does occur where the parents grow in close proximity, but the hybrids, being ill-adapted to the available habitats, fail to survive in the wild. The limitations of hybrids to intermediate habitats or ecotones has been excellently demonstrated by Briggs (1962)for the Ranunculi of the lappaceus group (p. 176). Here there seems no doubt that the restriction of gene flow depends almost entirely upon selective elimination of hybrids, even though the characteristic habitats may alternate over distances of only a few metres. The same principle is evident on a larger scale with the ecotypic subspecies of Potentilla glandulosa. Clausen and Hiesey (1958a) comment on the situation as follows : “Undoubtedly, at the edges of the distribution of the four subspecies of P. glandulosa a certain amount of hybridization is constantly taking place. Natural selection, however, limits the extent of genic infusion among the subspecies. I n regions where-equilibrium has thus been reached, the subspecies are able to maintain their identity.” There are other factors acting t o limit gene exchange between the P. gtandulosu subspecies ; reJtexa and hunseni, for example, differ in flowering time by as much as a month. Moreover, all of the subspecies except nevadensis are self-compatible and thus probably extensively inbred (Baker, 1953).Nevertheless, the importance of ecological isolation in preserving their independence is beyond question. The significance of isolation dependent upon the flower fidelity of pollinators in entomophilous groups has been emphasized especially by Grant ( 1949).Minor differences in floral characteristics among ecological races of genecologically differentiated species may contribute to maintaining isolation through this so-called ethological means. The specialization of pollinators certainly forms an effective isolating mechanism with the two species of Melandrium, M . rubrum and M . album (Baker, 1947). The ecotypic subspecies of Dactylorchis incarnata (p. 178) are barely distinguishable in vegetative features, and are yet sharply differentiated in flower colour and patterning (Heslop-Harrison, 1956). There is no direct evidence, but the possibility is open that this provides some measure of ethological isolation and has facilitated the independent migration of these races through north-western Europe (HeslopHarrison, 1958). Ecological isolation and isolation dependent upon the behaviour of pollen vectors both arise from genetical properties of the populations’ concerned, and all intermediates exist between these comparatively mild barriers and the ultimate disharmony which results in hybrid sterility or inviability. Ecological and to a lesser extent ethological isolation differ, however, from isolating mechanisms dependent ,upon partial or complete intersterility in not being irremoveable once established. This means that evolutionary fluidity is preserved; to extend

FORTY YEARS O F QENECOLOQY

213

Mather’s phrase, fitness for existing environments is not acquired at the cost of a narrow specialization which sacrifices the flexibility offered by a large gene pool. The change of circumstance which renders the original close adaptation to previously existing habitats valueless may simultaneously permit the survival of hybrid derivatives that otherwise would have been eliminated in favour of the strictly adapted types. What Anderson (1949) picturesquely calls “hybridization of the habitat” may entail a new release of genetic variation, and a new starting point for adaptive differentiation (Anderson, 1949,1953).

F.

PLASTICITY, GENETIC ASSIMILATION AND CONDITIONINQ

I n consequence of their growth by apical meristems, higher plants possess high potential phenotypic plasticity. The continuous serial production of fresh organs means that a capacity t o respond developmentally to environmental influences is present throughout life, not, as in the higher animal, only during an early embryonic period. It is apparent that these responses may or may not be adaptive. Thoday (1953) has suggested the term “developmental flexibility” for the truly adaptive property of generating phenotypes functioning satisfactorily in a range of environments. Mere morphological plasticity is not necessarily evidence of developmental flexibility. For example, the passive response to a factor establishing, say, a particular growth form, does not necessarily mean that a plant is better adapted for the environment in which that factor is present than it would be if it produced some other growth form under the same conditions. Nevertheless, there is good reason to suppose that the plasticity of higher plants does often have adaptive values, within certain limits. It has long been supposed that the wind forms of plants are adaptive, and direct supporting evidence for this view is now available from the work of Whitehead and Luti (1962) and Whitehead (1962, 1963a, b). I n maize and sunflower, exposure to wind speeds approximating those found in mountain regions induced anatomical and morphological changes which were demonstrably advantageous with respect to overall water economy. The phenotype became more xeromorphic, and the change was harmoniously related to the incidence of the adverse conditions. Whitehead points out that neither species is notably plastic, yet there is no doubt that they are equipped with a developmental flexibility which must extend their potential ecological tolerance ranges quite substantially. The work of Bjorkman and Holmgren (1963),discussed at length in a later section (p. 224 et seq.), provides some evidence of the adaptive value of variation in leaf morphology and physiology caused by differing levels of insolation. As part of these experiments, leaves of the different H

C.E.R.

214

J. HESLOP-HARRISON

ecotypes of Solidago were conditioned to various levels of .irradiation. Plants of the “sun” ecotype showed adaptive change, the photosynthetic rate at light saturation being higher following growth in high light intensities. Plants from the “shade” ecotype were incapable of adapting to the highest light intensity used, and the leaves were in fact damaged. Bjorkman and Holmgren’s observations on Solidago virgaurea suggest very clearly that the range of developmental flexibility in this species is somewhat limited. I n respect to the capacity for adapting to different light environments, it would seem that the genotypes tested would not be flexible enough to span more than a part of the observed tolerance range of the species; the breadth of this evidently depends upon the differentiation of ecotypes. There are, however, cases of species which seem to be capable of colonizing unusual habitats either by producing modified phenotypes (developmental flexibility) or by evolving ecotypes (genecological differentiation). Turesson (1922b) mentions the example of the forma nana of Succisa pratensis. This is an extreme dwarf form, attaining a stature of no more than 8.5 cm, occurring in the upper part of salt marshes around the coast of Scania. Samples transplanted from various populations differed in their behaviour under cultivation. All increased in height to some extent, but some populations proved to contain individuals fully capable of achieving the stature of plants from inland populations, whilst others were composed exclusively of hereditary dwarfs. Examples like Succisa pratensis f. nana raise the general problem of the relative values of developmental flexibility and genecological differentiation as devices permitting the extension of ecological range. It is obvious that some direct phenotypic adaptability will always be of value in buffering a plant against minor modulations of environment, either spatial or chronological. Why should tolerance not be extended indefinitely in this manner? Since the evidence shows that genecological differentiation is the “preferred” means of adapting to extreme habitats, it follows that there must be inherent disadvantages in direct phenotypic modification. Turesson’s own conclusion (1922b) probably covers the essential point : “The question may be made clearer by the assumption that the same characteristics which in one form of the species (the resulting modification) requires the exposure to an environmental factor of high intensity in order to become developed, may in another form (the hereditary variation) result as a response to a very much lower intensity of this factor. . . . It is conceivable that the habitat responsible for the development of the characteristic in question may at the same time act as a limiting factor upon general development in the case of the modification, while no such limiting action results in the hereditary variation because of the promptness with which it responds to this same habitat

FORTY YEARS OF GENECOLOGY

215

factor.” I n short, the physiological economy of producing directly an adapted form is of selective value. Turesson’s argument may be linked with that advanced by Waddington (1953a), concerning the “genetic assimilation” of acquired characteristics, the so-called Baldwin effect. According to this, if there is variability in a population in the capacity to react to a particular environment by the production of a modified phenotype which is at a selective advantage in that environment, then the reacting genotypes will necessarily be favoured. Successive generations of selection for the capacity to react to the effective stimuli may be expected to lower the intensity of stimulation required to produce a given phenotype, and ultimately to bring about an approach to a,state where the favoured phenotype is produced spontaneously. The feasibility of this kind of genetic assimilation has been demonstrated in a well-known experiment of Waddington (1953b) with Drosophila, the characteristic being a wing venation defect induced in the normal genotype by a high-temperature shock during pupation. I n the line selected for ready expression of the defect, flies appeared in the 14th generation which revealed it without temperature treatment. Given that there is advantage of the kind envisaged by Turesson in the direct rather than the enforced production of adapted phenotypes, and given further that genetic variability exists in a population in the capacity to respond to the effective environmental factors, then it seems inevitable that some form of genetic assimilation will occur. The significance for genecological differentiation is evident. The ability to exploit an unusual habitat in consequence of inherent developmental flexibility will, as it were, provide a species with a beach-head. To the extent that more facile phenotypic adaptation to the new environment is favoured, genes contributing to this will accumulate in the immigrant sub-population, and so the threshold at which the specialized form appears will be lowered. Since the intensity of selection will decrease as this process i l l never be progresses, the adjustment will be asymptotic, and finality w reached. This means that some adaptation will always depend upon developmental flexibility, and it is significant in this connection that, on transplantation to neutral environments, individuals of ecotypic populations adapted to extreme environments invariably regress in some degree towards the norm of the species. Waddington’s principle of genetic assimilation offers a plausible interpretation of one form of quasi-Lamarckian response to habitat. The possibility of other, more direct, mechanisms is, of course, still to be ke@, open. Dmant’s experiments on environmental conditioning in flax (1958,1962)seem to show unequivocally that one major group of habitat factors, nameIy soil nutrients, can induce heritable changes in pheno-

216

J. HESLOP-HARRISON

type. I n one experiment, eight combinations of nitrogen, phosphorus and potassium fertilizer treatments were given to a parental generation, and effects on weight of progeny were observed for four generations under uniform soil conditions. I n the second generation, the descendants of plants receiving N K but not P weighed only one-third of those receiving PK. The “genotrophs” or conditioned form6 differed in morphological features other than gross production, and revealed various physiological peculiarities. Durrant has referred to them as being adapted to different environments, and it is evident from his data that their performance does not always bear the same relationship to the “normal” under diverse conditions of cultivation in the open and in greenhouses. It is not apparent, however, that the effect of conditioning in flax is necessarily to produce a genotroph better adapted to the soil environment inducing the change. The possibility that environmental conditioning in flax might depend upon the behaviour of cytoplasmic determinants was considered by Durrant (1958), but later abandoned (1962). The suggestion was that under the inducing conditionsthe relative multiplication rate of nuclear genes and plasmagenes might be affected, so establishing a new nuclear-cytoplasmic relationship. Should this new relationship constitute a stable equilibrium, the condition might be propagated through successive generations. Essentially the same suggestion had previously been made by Crosby (1956) as a theoretical basis for the inheritance of acquired adaptations. A model system is described by Crosby in which a plant is presumed to be adapted to a mild environment with a low concentration of a plasmagene favouring resistance to low temperatures. Under cooler conditions, growth rate and so cell division rate is reduced. If it be assumed that the rate of reproduction of the plasmagenes concerned with cold resistance is not reduced proportionately, then their concentration will increase in the cell, so progressively enhancing cold tolerance. If the conditioned state is transmitted through the reproductive cells, then progeny will be better equipped to survive in the cooler environment. Crosby emphasizes that acquired adaptation of this kind is most likely to be found in generalized characteristics like growth and vigour; i t is in these features that Durrant’s genotrophs differ most from their parental strains. The response as envisaged by Crosby and Durrant may be looked upon as a sort of physiological adjustment of phenotype to environment akin to that involved in the conditioning of individuals k i n g growth, but having a transmitted element providing for the partial pre-adaptation of progeny. Certainly there is no need to discount this possibility on any physiological grounds. It is well known that induced states such as those arising from vernalization and photo-

FORTY YEARS OF QENECOLOGY

217

periodic treatment may be transmitted through many cell generations without decay, and may, furthermore, be amplified and transmitted to other plants by grafting. Indeed, having regard to this, what is remarkable is the efficiency of the mechanism which acts in the commonality of cases to restore reproductive cells to a basal, undifferentiated and unconditioned state. That occasionally this restoration should be incomplete so that some proportion of a parental experience is transmitted to progeny seems not at all unreasonable. Durrant’s experiments illustrate one kind of transmission of acquired characteristics; evidently it would be profitable now to reconsider the possibility that some habitat-correlated variation may be due to direct conditioning and subsequent transmission of the conditioned state to progeny. It is interesting to note that there is no a priori reason to expect conditioning invariably to produce adapted phenotypes, except in so far as a principle like that of Crosby might be involved. This opens the possibility of there being non-adaptive,habitat correlated variation. However, there is as yet no direct evidence of such a phenomenon, and in the succeeding discussion of the physiological aspects of genecological differentiation it will not be taken into consideration.

111. PHYSIOLOGICAL ASPECTSOF G E N E C O L O G I C A L DIFFERENTIATION

A.

INTRODUCTION

The detection of genecological differentiation within a species generally depends in the first instance upon the observation of habitatcorrelated variation in morphological features. As we have seen (p. 168) it is presumed that if populations in one type of habitat are regularly found to differ from those in another in any characteristics whatever, those differences (or others unobserved genetically or developmentally correlated with them) must have adaptive significance, because the only plausible cause of the divergence, discounting the kind of conditioning mentioned in the foregoing section, is the differential effect of selection in the two environments. This argument depends not in the least upon any physiological interpretation of the differences. Yet it is difficult to arrest genecological investigation at the point where presumed adaptive differences have merely been identified, for the question of their actual survival value in the habitats concerned then obtrudes itself. An important matter to clear up is the level at which answers to questions of this type would be deemed acceptable. Consider the case of two “ecotypes” A and B, from two habitats, a and b. I n habitat a,A performs better than B as assessed by some arbitrary measure such as dry matter accumulation, and in

218

J. HESLOP-HARRISON

habitat b, the position is reversed. An experimental demonstration of this fact would provide one type of explanation for the occurrence of each ecotype in its characteristic habitat. It might be considered unsatisfactory, (a)because it does not follow that the measure selected is that of primary significance for survival when the two are in competition (some other factor, for example, the powers of seed dispersal, may be more important in these conditions), and ( b ) because the experiment does not reveal which factor or complex of factors in the habitats was responsible for the difference in dry matter production, nor which physiological mechanism in the plants themselves was responding differentially in the two ecotypes. To proceed further with the problems under (a) would demand an assessment of more estimates of performance, each posing its own set of questions under ( b ) . To pursue the queries under (a) would require a factor by factor analysis of the habitats, and a function by function study of the plants. The difficulty under ( a ) might be met in part by comparing the relative fitness of ecotypes directly in competition in the different habitats, the performance measure adopted being simply survival. Other things being equal, the expectation is that the habitat “sieve” will once more sort them out appropriately. An ecotype Derby of this kind would not, of course, represent a recapitulation of the evolutionary processes which led to ecotypic differentiation, since at no time in the course of evolution would the end products of differentiation in specialized habitats come together in direct competition in any one habitat. I n the ideal situation, nevertheless, it might be expected to provide direct evidence to support the original circumstantial case for the ecotypic differences being adaptive. It is this kind of evidence that is sought in reciprocal transplant experiments, and in varied environment experiments such as those conducted by Clausen et al. (1940, 1948). Given this reassurance about the adaptive value of the ecotypic characteristics acting in consort, questions of the ( b ) type still arise, namely: I n what way are they adaptive physiologically speaking? If answers on this level are really required, there is no escape from the need to trace direct causal connections between environment and plant responses. For the higher plant, habitat factors are commonly regarded as being of three general kinds, climatic, edaphic and biotic; yet it is an ecological platitude that survival is determined by the simultaneous effect of all, the environment being “holocoenotic” (Cain, 1944). What possibility is there, then, of linking “character” or “response” with “factor”, when the ontogeny and physiology of the plant are expressions of the reaction between the entire genotype and the entire complex of environmental influences?In the generality of cases the answer to this must be, very little; particularly when the habitat differences are small and multidimensional. Then

FORTY YEARS O F GENECOLOGY

219

truly will the terminus have been reached. Fortunately, however, when it is difference betweenhabitats which is under consideration, the situation may be simpler, because very often it is possible to identify the major determinants. “Adaptation” is still likely to be a holistic process, but it is feasible to measure plant responses to the major environmental variables singly and in interaction in various simple patterns, and obviously this can be done in a comparative manner with populations of different provenqnces. It needs no emphasis that this type of study only becomes useful as a basis for extrapolation to the natural situation when the conditions of the experiments themselves can be related in some meaningful manner to those in nature. Merely to match the list of morphological differentiae observed between ecotypically differentiated populations with another of physiological differentiae demonstrated under some experimental condition or the other is obviously not enough. It is only when a functional difference can be shown to be associated with a differentially acting environmental factor that we are beginning to approach the understanding of a causal sequence. Yet even here it may still not be possible to say with assurance that the selective agent and the specific function upon which it impinges have been identified. The possibility remains that the “real” target of selection is an associated or secondary response, or a function for which the one studied acts as a governor or time-keeper.

B.

EDAPHIC ADAPTATION

There is no particular reason for supposing that the adaptation of plants to their rooting media will involve processes or principles different from those governing adaptation to the sub-aerial environment, but it so happens that there are examples of clear-cut patterns of adaptation to soil type which should permit a more precise kind of analysis than can be given to most examples of adaptation to climate or biotic influence. Such a case is the tolerance of grass populations to soils contaminated with heavy metals, briefly described in a foregoing section (Bradshaw, 1952; Wilkins, 1957, 1960a, b). Wilkins compares this example to industrial melanism in moths : in both situations the selective factor can be identified precisely and its impact can be shown to be severe. I n the case of Pestuca ovina, lead is highly toxic to normal plants, yet on mine spoil heaps in the British Isles tolerant populations occur on soils containing up to 4% lead. Using an assay method based upon the measurement of extension growth in the roots of tillers grown in glass tubes in culture solutions under standard environments Wilkins (1957, 1960b) has obtained evidence suggesting t.hat three types are present in the species : intolerant, medium tolerant and highly tolerant. The tolerant

220

J. HESLOP-HARRISON

plants are restricted entirely to lead-containing soils, and the intolerant ones to leadlfree soils. Genetically the property of tolerance was found to be completely dominant, and although a full analysis has not been possible yet, it is conceivable that the tolerant genotypes may differ from the normal in substitutions at single loci. I n a field trial no morphological features could be shown to be correlated with tolerance, and for that reason Wilkins has been reluctant to look upon the tolerant populations as constituting an ecotype. Whatever terminology is used, BOWever, it is evident that this is an example of genecologicalZfferentiation in the broad sense, and rather an important one in that the selective factor has been identified with some assurance and related to a physiologicd property which should, in principle a t least, be open to precise meaiurement . The examples of ecological races adapted to serpentine soils also discussed in an earlier section are in some respects comparable. Kruckeberg (1954) states that a major criterion for serpentine tolerance must be the capacity for growth on soils of low calcium levels. A direct comparison of the tolerances of serpentine and non-serpentine races of Phacelia californica to calcium deficiency revealed a very clear difference ; the normal soil race showed no growth in un-supplemented serpentine soil, the growth being equalized with supplements equivalent to 2 tons of gypsum per acre. NPKctreatments greatly benefited the tolerant race on serpentine soil, but were without effect on the normal soil race on this medium. The suggestion of Walker (1954) that serpentine races might benefit from a capacity to accumulate calcium in preference to other cations was also verified with the Phacelia races. At each of three soilcalcium levels the tolerant race absorbed greater amounts of calcium and lesser amounts of magnesium, as revealed by tissue analysis. A similar distinction between serpentine and non-serpentine races in tolerance of calcium-deficient soils was observed in twelve other species of annual and perennial herbs, although no serpentine ecotypes were observed among the grasses tested. Although there seems little doubt that the major adaptation shown by the serpentine races is their capacity to tolerate calcium deficiency, Kruckeberg points to the fact that serpentine ecotypes may show other adaptive characteristics. These include in particular the capacity to tolerate the exposure and sometimes drought which may also characterise the serpentine habitat. He concludes “that serpentine plants are physiologically adapted to the open characteristics of serpentine communities as well as to the special soil conditions, and that the physiological adaptations of serpentine plants may have some degree of morphological expression” This multidimensional adaptation of some serpentine races offers a pretty example for consideration in relation to

221

FORTY YEARS O F GTENECOLOGY

the general problem of ecotype evolution. The adaptation of these races must be to the whole environment of the serpentine, yet the dominant factor in this, according to Kruckeberg, is the calcium deficiency. Selection pressure has presumably been greatest therefore for the property of accumulating calcium preferentially ; when this was acquired, the invading populations would become exposed to selection for tolerance to the other conditions in which the serpentine habitat differs less radically from the normal. The inter-locking nature of the various factors becomes apparent when it is appreciated that certain peculiarities of the serpentine habitat arise just because some species -including shade-forming trees -are excluded from it by their intolerance of calcium deficiency, A single major factor is here seen to generate a complex of selective pressures secondarily. The demonstration of adaptive properties in races occurring on extreme soil types raises the further question of whether adaptation has been bought at the cost of the ability to survive in normal soils. I n the case of lead tolerant fescue races, Wilkins (1960b) records a higher mortality among tolerant biotypes in cultivation under conditions of low competition. This observation does indeed suggest that selection for the ability to endure high lead concentrations in the soil has led to theplants’ becoming dependent in some way on lead-rich soils, possibly, as Wilkins suggests, even by establishing a requirement for lead itself. Kruckeberg’s observations (1954) on serpentine endemic species seem to show that their exclusion from normal soils is entirely in consequence of competition. I n culture on calcium-replenished soils free from the suppressive effects of weedy annuals, serpentine endemic Streptanthus species thrived; subject to competition, survival was poor. It is perhaps not justifiable to argue directly from this to the case of serpentine ecotypes, and it may be that the adaptation of these has involved some sacrifice. McMillan (1956a)records that a “strain” of Agrostis halliii from a serpentine soil performed very substantially better, as judged by height and dry weight production, when cultured on Serpentine soil than when grown on a control soil; this may indicate that here adaptation has involved specialization. But the example of Achillea borealis, in which some biotypes from normal soils are recorded by Kruckeberg (1950) as showing tolerance to serpentine conditions, proves that the capacity to survive on calcium-deficient soil does not necessarily put a plant to selective disadvantage in an otherwise unadapted population on a normal soil. The two cases of edaphic adaptation discussed above relating to the tolerance of heavy metals and serpentine soils are unusual in that the selective factors are obviously severe and the circumstances in which they are encountered rather rare. It is obviously important to know H2

C.E.R.

222

J. HESLOP-HARRISON

whether genetic adaptation occurs when species encounter less distinctive soil types. The best evidence so far available, that it does is that given by Snaydon and Bradshaw (1961) for Festuca ovina. I n this study, samples from populations growing on soils with free CaCO, (pH 7.8-8-9) and on soils low in calcium (1.5-6.8 m.e. per lOOg Ca; pH 4.3-4-6) were examined for their response to calcium in culture solution at levels of 5, 10, 20, 75 and 150 p.p..m. An analysis of variance showed that the within-type x treatment interaction was insignificant, justifying the placing of the populations in two groups, “calcareous” and “acidic”. The growth of the two types, as assessed by dry weight production, is related to Ca level in Fig. 3. The striking features are the outstandingly

301 I

,

1

I

Acidic

-0

Id

I

Calcareous/

t

40

0.70 1.0 1.30 1.88 2.M

-

Calcium (log conc.) ,

510

I

20

75

I

I50

Calcium (pprni

FIG.3. The growth ofFestuca ovina of L‘calcareou”’ and “acidic”types at various calcium levels. (From Snaydon and Bradshaw, 1961.)

better performance of the acidic type at low calcium levels and its poorer performance a t the higher level. Plant analysis in an experiment in which genotypes from two contrasting populations were compared indicated a three times better calcium uptake at low calcium levels by plants adapted to acidic soils. At equivalent dry weights, plants from acidic and calcareous soils contained equivalent amounts of calcium, showing that the success of the plants from acidic habitats when grown in low-calcium solutions was not attributable to a lower requirement for the element. Snaydon and Bradshaw point out that the response to calcium is not likely to be the only factor differentiating plant populations from widely different soils, but their results certainly suggest that the wide edaphic tolerance of Festuca ovina has been achieved by the differentiation of races physiologically adapted to different levels of soil calcium, so that calcium availability must have been a major selective factor.

FORTY YEARS OF GIENECOLOGIY

c.

223

ADAPTATION TO SOIL MOISTURE STRESS

In spite of the wide interest in the water relations of plants there have been few experimental studies of ecotypic adaptation to different levels of soil moisture. McKell et al. (1960)have compared two races of Dactylis glornerata, recognized taxonomically as the geographic subspecies lusitanica and judaica, for their performance under differing conditions of soil moisture stress. Subsp. lusitanica is known from the coastal area of central Portugal, where it occurs in a mild coastal climate, while judaica

Time (h)

FIG.4. Rate of water use by two ecotypic subspecies of Dactylis glomerata, SubSp. jud&m and Zusita~iea.(FromMcKell, Perrier and Stebbins, 1960.)

(probably the arid ecotype of Boyko and Tadmor, 1954) occurs in Israel and perhaps neighbouring countries in a dry Mediterranean climate (Stebbins and Zohary, 1959). Plants were grown in pairs in a sandy clay soil in cans. After growth under favourable moisture conditions for three months, they were clipped back and transferred to growth cabinets where they experienced a daylength of 14 h with a light intensity of 1300 fc, with a temperature of 18' f 1°C. and 457/, + 5% R.H. A conditioning period of 1 month was allowed, after which the plants were clipped again and the observations begun. The cans were saturated initially and then allowed to dry out, the soil moisture being monitored continually by observing the electrical resistance of buried gypsum plugs. A difference in soil water utilization was observed between l w i tanica and judaica on a per-plant basis, as shown in Fig..4. Subsp.judaica showed a higher rate of leaf elongation and reached a greater cumdative mean leaf-blade length during the experimental period. On the other hand, lusitanica produced a greater final weight of foliage; evidently cell elongation is inhibited by water deficiency before dry-weight accumulation is curtailed.

224

J. HESLOP-HARRISON

McKell et al. note that subsp. judaica characteristically passes into a state of summer dormancy. This presumably forms its principal means of surviving the main period of moisture stress in its native habitat, rather than any direct adaptations to water deficiency. During the actual growth period the main physiological feature suggesting adaptation to soil moisture stress was the lower rate of water use ; otherwise the overall responses were not very different from those of subsp. lusitanim.

D.

ADAPTATION T O LIGHT INTENSITY

A conspicuous form of environmentally correlated morphological variation in higher plants is that apparent in leaves according t o the intensity of illumination. Differences between sun and shade leaves of the same genotype are a direct consequence of the morphogenetic effects of light, and it is widely supposed that the characteristic features are in some way adaptive, although the direct evidence is slight (Wassink et al., 1956).I n consequence of plasticity, the sun and shade grown individuals of some species may reveal substantia,l phenotypic differences in leaf size and texture, but, as shown by Turesson (192213) with Lysimachia vulgaris, such differences disappear on cultivation in standard environments. Heritable differences in leaf form between populations of a species from open and shaded habitats have, however, been frequently demonstrated, and this may be regarded as an example of genecological differentiation. Although it is the variation in the light environment which is commonly most conspicuous, “sun” and “shade” habitats do inevitably differ also in other correlated factors, notably in temperature and humidity. There is therefore the familiar diEculty of identifying the effective selective factors, and again it must be supposed that several responses have been selected for simultaneously in the course of adaptation. Nevertheless, there is good reason for concentrating upon light as a major measureable differential factor in studying the adaptation of sun and shade ecotypes, particularly since the role of light in the economy of the plant is reasonably well understood. An excellent study of photosynthesis efficiency in sun and shade ecotypes of Solidago virgaurea has recently been carried out by Bjorkman and Holmgren (1963). The sampled populations were from two shaded habitats, oak and beech forest, and two open habitats, dry open meadow at 105 m altitude and alpine heath at 600 m. Ten plants were selected from each environment and propagated to provide two pairs of individuals of each genotype. One pair from each clone was cultivated at low light intensity, 3 x 104 erg/sec/cm2and one pair at high, 15 x lo4 erglsecl ems. Other conditions were identical : photoperiod 16 h, temperature during light period 20.0 3 0.3’C and during dark 10.0 f 0.3”C. Air

225

FORTY YEARS OF GENECOLOGY

humidity was maintained at 70 f 5% R.H. After 4 to 8 weeks of preliminary growth, photosynthesis measurements were made and anatomical differences investigated. Mature rosette leaves were detached and irradiated in temperature regulated, humidity controlled cells with light from a xenon arc source. Photosynthesis was measured by monitoring the utilization of CO, with an infrared absorptiometer. Two expressions of photosynthetic activity are of importance in a comparison of this nature : ( a ) the relationship between photosynthetic rate and light intensity at low light levels (a satisfactory measure being the initial slope of the rate-intensity curve), and ( b ) the photosynthetic rate at saturation light intensity. Comparisons between three clones from each of the four habitats are given in Table V, which is re-arranged from the data of Bjorkman and Holmgren. The two population groups differ

TABLEV ( A ) “Photochemical Capacity” (measured as initial slope of the rate-intensity curves)for Plants of Three Clonesfrom each of Four Localities, Two Exposed and Two Shaded, of Solidagovirgaurea, Grown at Low Light Intensity. (B)Photosynthetic Rate at Light Saturation in Plants from the Same Clones Grown at a High Intensity

Locality 1 clone a clone b clone c Locality 2 clone a clone b clone c

Shaded habit& A

Exposed habitats

B

3.10 2.98 2.58

18.9 18.6 15.2

Locality 3 clone a clone b clone c

2.90 3.04 2.98

18.6 17.0 19.2

Locality 4 clone a clone b clone c

A

B

2.40 2.17 2.27

31.8 17.3 27.6

2.26 2-34 2.44

246 23.8 27.8

Mean values for A : shaded, 2.93h0.08;exposed 2.31&0.04 Mean values for B: shaded, 17-9&0.6;exposed, 25.552-0 [Data from Bjbrkman and Holmgren, 19631

markedly in respect to both measures. The shade ecotypes are evidently capable of a more efficient use of weak light than those from open habitats, while those from open habitats can utilise intense light more efficiently. It is important to note that the pre-conditioning of the plants was such as to ensure a full degree of individual adaptation to low light in the comparison of efficiency at low intensity, and to high light in the comparison of rate at saturation intensity. Undoubtedly, therefore, the differences are genetically based.

226

J. HESLOP-HARRISON

Other comparisons of importance made by Bjarkman and Holmgren are summarized in Table VI. The first observations (A) indicate that the exposed-habitat clones have a marked ability to adapt to the efficient use of strong light, when grown in strong light, lacked by those from shade habitats. The second (B) show that the photochemical response of the clones from open habitats is not affected by light intensity during cultivation, while culture under high light intensity does reduce the initial slope of the rate-intensity curve in clones from shade habitats.

TABLEV I ( A ) Ratios of Photosynthetic Rates at Light Xaturation of Plants of Different Clones of Solidago virgaurea Crown in Strong Light and in Weak Light. (B)Ratios of “Photochemical Capacity” in Plants from the Same Clones Crown in Strong Light and in Weak Light ~

Shaded habitats

Exposed habitats

Locality 1 clone a clone b clone c

A

B

0.96 0.92 1.04

Locality 2 clone a clone b clone c

0.88 0.86 1.19

A

B

0.63 0.63 0.67

Locality 3 clone a clone b clone c

1.86 1.65 1.59

1.12 0.98 1.06

0.65 0.66 0.61

Locality 4 clone a clone b clone c

2.14 1.59 2-19

1.11 0.94 1.00

Mean values for A: shaded, 0.9750.05; exposed, 1+34&0.11 Mean values for B: shaded, 0*64&0.01; exposed, 1.0450.03 [Data from Bjorkman and Eolmgren, 19631

According to Bjarkman and Holmgren, chloroplasts of plants from shade clones grown in strong light are partly destroyed. The results of this investigation seem to indicate unequivocally that the sun and shade races of Solidago virgaurea do differ in their photosynthetic properties in an adaptive manner. Whatever other selective forces may act upon populationssentering these two types of habitat, it is evident that light intensity is of great -probably paramount -importance. The results suggest, moreover, that in this species at least the direct adaptability of the photosynthetic system is too low to permit a single genotype to perform adequately in both open and shaded habitats. It would obviously be informative to compare in this respect the behaviour of a species thought to be more plastic, such, for example, as the Lysimachia vulgaris studied by Turesson. As part of a general investigation of the climatic races of Mimulus cardinalis, the Carnegie group (Hiesey et al., 1959, 1960, 1961 ;Milner et

FORTY Y E A R S O F GENECOLOGY

227

al., 1962) have been examining photosynthetic efficiency in various artificial environments. As reported so far, the work is principally of interest as a study in comparative physioJogy, since it is not always clear in what way the differences observed can be associated with adaptation to habitat. Two examples of responses observed in experiment which may be related to ecotypic adaptation have, however, been recorded (Hiesey et al., 1960, p. 317). A clone from a high-altitude Yosemite population revealed a saturating light intensity for photosynthesis 1-5 times higher than one from a sea level habitat at Los Trancos; it is suggested that this may be related to a difference in average light intensity in the two habitats. Furthermore, the Yosemite plants were found to make less effective use of high CO, concentrations, photosynthesising at optimum temperatures and at light saturation, than plants from Los Trancos, and the possibility is mentioned that this may be concerned with the 15% reduction in CO, in the atmosphere a t the altitude of the Yosemite race compared with that at sea level. Evidence concerning the effects of pre-conditioning comparable with that available for Solidago virgaurea seems not yet to have been obtained, but the first of these observations, so far as it goes, does suggest a similar pattern of racial adaptation to variation in available light to that demonstrated in Solidago by Bjorkman and Holmgren.

E.

ADAPTATION TO CLIMATE

It is evident from the pattern of variation in many of the wide-ranging species discussed in Section I that infiaspecific differentiation must frequently be dominated by the selective effects of regional climates. The analysis of these effects is beset with formidable difficulties. Those arising from the mrdtifactorial nature of climatic differenceshave already been mentioned. I n addition there is the complication that a diversity of solutions is available to plant populations for adaptation to climatic differences. I n the example of habitat adaptation just reviewed, one pervasive environmental factor can be identified as probably having had the dominant selective influence ; survival depends absolutely upon the capacity to adapt to it. Where habitats differ in a multiplicity of seasonally varying climatic factors at one and the same time, adaptation in several features may be required to permit survival, while continued existence may equally well be ensured by more than one pattern of response. Undoubtedly the most powerful means available to the higher plant for adaptation to regionally varying climates is the ability to adjust the developmental cycle. This property is of the greatest significance when a species spans a range of climates differing in the pattern of incidence of unfavourable seasons -hot and dry, or cold. To survive adverse condi-

228

J. HESLOP-HARRISON

tions in a suitably protected state of dormancy can obviously be to close to an ideal solution, necessitating the minimum of direct sdaptive adjustments of the normal prgcesses of growth and nutrition by ensuring that they progress during the most favourable period of the year. The adjustment of developmental periodicity t o the particular requirements of a local climate is a matter of time-keeping, and it may be supposed that where survival depends on re-timing, selective pressures will develop to bring this about. Since seasonal variations in temperature and daylength provide the most reliable clocks in the plant environment, thermal and photoperiodic responses will thus necessarily become a target for selection, and variation in these responses is therefore to be expected in widely ranging species. We may speak of “photoperiodic” or “thermoperiodid’ races or ecotypes in these species, but the adaptation is not of course to the photoperiod or thermoperiod in the sense that these factors are themselves selective; it is that adaptation to local climate has been achieved by the modification of photoperiodic or thermoperiodic reactions. This kind of adjustment by itself may not be adequate to ensure success if more or less inimical conditions prevail even during the most favourable period of the year. Survival may then require the acquisition of characteristics more directly adaptive, as well as the adjustment of periodicity. As might be expected, the best analysed examples of climatic control of developmental periodicity have cdncerned bred strains of economic plants, particularly the cereals and biennial root crops. b o n g perennial species, forage grasses have been extensively investigated, and the work of Cooper (1951,1952)on Lolium and Ryle (1963a, b) and Ryle and Langer (1963a, b) on Phleum is particularly relevant to the problem of analysing the environmental control of developmental cycles in ecotypes. The work of Wareing (1950a, b, 1951, 1953, 1954) on the photoperiodic responses of woody species is similarly of great genecological significance. For the perennial plant of the temperate regions, i t is established that temperature and photoperiod - separately or in conjunction -may affect the duration of bud dormancy, the period of leaf formation and stem extension growth, the timing of flowering and fruiting, the cessation of stem extension, and the onset of leaf-fall and resting bud formation. Response to temperature may be immediate, or inductive, as in the case of vernalization; and the reaction may be to the diurnal as well as the annual cycle of temperature change. Photoperiodic effects are mostly inductive. The usual physiological classification of species into “long day”, “short day”, “day intermediate” and “day neutral’’ with regard to flower initiation has little direct ecological significance since the important factor in nature is the response to the yearly cycle of daylength

FORTY YEARS O F GENECOLOGY

229

change, not the behaviour under a fixed photoperiod. Valuable reviews of temperature and daylength effects in which ecological aspects are taken into consideration are those of Wareing (19'56)and Chouard (1960). Ecotypic variation in temperature responses has been studied experimentally in several species. Part of the current programme of the Carnegie group (Hiesey et al., 1959, 1960, 1961; Milner et al., 1962) is concerned with this aspect of racial differentiation, in Mimulw cardinalis. I n an earlier study, Hiesey (1953) has investigated the thermal responses of the aggregate species Achillea millefolium. The reactions of cloned material of three races originally grown from seed under standard garden conditions were followed under different combinations of day and night temperature in growth chambers. The races were a diploid maritime one, corresponding to the taxonomic subspecies arenicola from a cool Californian coastal region, an interior valley diploid race native to the hot San Joaquin Valley of California, and a high montane race referable to Achillea lanulosa subsp. typica from an altitude of 8 700 ft in the Sierra Nevada. It may be noted that, strictly, this last race must be regarded as belonging to a separate ecospecies from the &st two, having a tetraploid chromosome number, but much of the interest of the study arises from its inclusion. The responses of the races may be summarized as follows : Coastal. Good growth and flowering was attained in night temperatures of 6" C in combination with 20" C or 30" C days; growth and flowering were inhibited with night temperatures of 17" C. Interior valley. Good growth was attained with 6" or 17OC night temperature, while performance was slightly better with day temperatures of 20" C than with 30" C. Flowering was delayed somewhat when 6"C nights were combined with 20" C days. Sierra Nevada. Growth was favoured by 6°C nights compared with 17"C, and apparently by 30" C day temperatures compared with 20"C. Flowering was stimulated by low night temperatures. Hiesey relates these responses to the climates in the regions of origin. The coastal race, occurring in a region of low night temperatures, is clearly intolerant of 17" C nights, although capable of standing warmer days than the average in its native area. The San Joaquin valley race gives evidence of adaptation in its capacity for survival and satisfactory growth in all combinations of day and night temperature, consonant, according to Hiesey, with "the great amplitude of seasonal temperature variation in its native habitat and its all-year activity". The Sierran race evinces rapid development towards flowering with cool nights and warm days, again showing adaptation to climate in its place of origin, where during the growing season the average minimum temperatures range between 5"C and 9" C and the average maximum from 19"C to

230

J. HESLOP-HARRISON

24' C. The rapid development of the Sierran race compared with that from the Californian coast can be related to the much shorter growing season. Hiesey encountered marked variation between individuals in this study, but comments that the differences between the races overshadowed the within-race variation. Clinal variation in the chilling requirements for bud break in Acer rubrum has been demonstrated by Perry and Wang Chi Wu (1960). the red maple has an extremely wide latitude range in North America, from southern Florida to Canada. I n the northern part of the range, the frostfree season is less than 100 days, and in the southern no frosts are experienced at all. Chilling responses were studied by Perry and Wang Chi Wu in progenies reared from seed samples collected at eleven stations spanning the greater part of the latitudinal range. Plants raised from parents from southern Florida showed no chilling requirement, and bud break was in fact delayed by chilling treatments. Progeny from parents from the northern part of New York State mostly failed to break dormancy without vernalization, and where growth did occur it was abnormal. Plants from intermediate latitudes showed intermediate reactions. Among the samples grown without chilling at Gainesville, Florida, a strong correlation (r = 0.96) was observed between date .of bud-break and the duration of the frost-free season at the site of origin. This study reveals clear evidence of a clinal variation in the red maple in dormancy periods and in response to chilling. The experiments so far reported are not, however, adequate to elucidate fully the relationship between latitude, dormancy period and cold requirement. This would require an analysis of the responses of the populations from the different latitudes to a range of treatments, along the lines marked out in Olmsted's full study of dormancy in the sugar maple (1951). Ecotypic variation in respect to chilling requirement for the induction of flowering has been demonstrated by Ketellapper (1960) in Phalaris tuberosa. Plants were grown from samples of caryopses collected from twelve localities throughout the Mediterranean range. Flowering time under a greenhouse temperature of 17-19" C without effective previous vernalization was observed, and also flowering time following chilling at 4' C at the three-leaved stage for periods of 0 , 2 , 4 , 6 and 8 weeks, with subsequent growth in the greenhouse. A high inverse correlation (r = - 0.94)was found between the percentage of plants floweringwithout cold treatment in each sample of c. 75 and the weeks of cold treatment required for full induction to flower. The duration of cold treatment required for full induction was also found to be highly correlated (r = - 0.95) with the average temperature of the coldest month in the locality of origin. Ketellapper is, however, dubious as to the exact role of the cold requirement as a regulating factor for flowering time in the

231 wild. Under field conditions at Canberra, Australia, the time of flowering was found to bear no very close relationship to the vernalization requirement of the different races, suggesting that in this locality other factors -perhaps spring temperatures and photoperiod - may be determinative. He suggests that an interaction of vernalization temperature, spring temperature and daylength may determine the developmental period in the native areas of the races tested. Nevertheless, that genetical variation in cold requirement does exist and that this can be related to one important climatic parameter does suggest that the vernalization response has been subjected to differential selection in various parts of the natural range. The role of photoperiodic responses in the adaptation of species of the grasslands of North America to the climates encountered in various parts of their range has been the subject of several studies in the last twenty years. The work of Olmsted (1944a, b) on Bouteloua curtipendula remains one of the most satisfactory analyses. B. curtipendula has a wide latitudinal range, extending from Mexico in the south to southem Saskatchewan in the north. Olmsted's study was based upon seed samples derived from twelve sites spanning 17" of latitude. Comparisons were made of growth and flowering periodicity under daylengths of 9, 13, 16 and 20 hours and under natural daylength at Chicago. A considerable amount of variation appeared within the samples, but regularities in their behaviour could be distinguished and related to provenance. The periods from germination to f i s t flowering under the various photoFORTY YEARS OF .QENECOLOGY

TABLEV I I Number of Days until first Flowering after Germination on 11 April 1942 of 12 Strains of Bouteloua curtipendula under Different Photoperiods. [Data of Olmsted, 1944a.l Sourceofseed North Dakota Nebraska Nebraska 'Nebraska Kansas Kansas Oklahoma Oklahoma New Mexico Arizona Texas Texas

LatitudeON 464 42i 414 40 394 39 37 35* 323 32 294 294

9h

13h

109

123 109 73 81 95 81 81

-

55

Photoperiod 16h 20h 73 95 95 123 109 123 95 -

-

95

69 81 95 109 109 109 123 123 123

-

-

Nat.day (Chicago) 55 73 95 81 109 109 123 109 109

232

J. HESLOP-HARRISON

periods recorded by Olmsted are shown in Table VII, the strains being ordered according to latitude of origin. The plants of the most southerly origin show a short-day type of reaction, and those from the northern sites are evidently accelerated in development towards flowering by long days. The Oklahoma and New Mexico samples show diversity in response, perhaps as Olmsted suggests because they contain both longday and day-intermediate genotypes. I n respect to vegetative development, the different races revealed a consistent pattern of response:'with increasing latitude of origin, growth was increasingly suppressed under the shorter photoperiod. No internode elongation occurred in the primary axis of plants originating north of 39" N latitude in 9-h days, and none in plants originating north of 414" N in 13-h days. Plants from southern latitudes elongated in all photoperiods. The study of geographical variation in photoperiodic response in Andropogon scoparius (Schizachyrium scoparium) by Olmsted's pupil Larsen (1947) revealed a pattern of responses broadly similar to'that observed in Bouteloua. In a comparison of samples from twelve stations spanning a latitudinal range of Z l " , plants from northern origins were found to be accelerated towards flowering by long days; in the case of Andropogon scoparius, however, the southern strains behaved as dayintermediate plants. Again short photoperiods marked suppressed growth in plants from northern localities. While Olmsted's and Larsen's results were obtained under experimental conditions which cannot be compared directly with any natural situations, they do indicate quite clearly that the differentiation of races adapted to different lengths of growing season in the central grasslands of North America has involved the adjustment of photoperiodic response. The transplant experiments of McMillan (1956a, b, 1959) indicate that several other grass species show similar kinds of ecotypic variation, periodicity being adapted to the south to north decrease in the frost-free growing season and to a similar east to west gradient determined by decreasing precipitation during the later part of the summer. Photoperiodic adaptation has also been demonstrated in tree species which span wide latitudinal ranges. Vaartaja (1954, 1959) found variation in the responses of northern and southern races in several tree species tested for growth rate and onset of dormancy under various photoperiods. Samples were grown from seed derived from native stands, and the seedlings were exposed to natural daylight for 11 h extended by artificial light to give photoperiods of 12, 14, 16 and 18 h. An example of Vaartaja's results is given in Table VIII, where the fresh weights of the tops of plants of Betulapapyrijera from two localities, 17" of latitude apart, grown for 15 weeks under various photoperiods are

233

FORTY YEARS OF GENECOLOGY

compared. Evidently a critical daylength exists for the northern race between 14 and 16 h below which growth is practically suppressed. The southern race, on the other hand, although depressed in growth by 12-h days, is still reasonably active in the shorter days. Similar responses (although none so extreme) were found in most of the sixteen other species tested; the general rule was that the northern races showed growth inhibition in longer photoperiods than those from southern sources. Vaartaja's results are also of interest in the light they throw upon the photoperiodic control.of dormancy in the trees tested. I n some species, continued apical growth in the northern races only proceeded under very long days ; anything less than 16 h led to early dormancy. In

TABLEV I I I Fresh Weight of Tops (in mg) of Seedlings of Betula papyrifera from T w o Localities, Northwest Territories, Latitude 69" N , and Pennsylvania, Latitude 42" N , Grown for 15 weeks under Different Photoperiods. [Data of Vaurtaja, 1959.1 Latitude ON 69 42

12 h 2 93

Photoperiod 14 h 16 h

18 h

8 110

372 209

155 133

contrast, the southern races continued growth at much shorter photoperiods. Photoperiodic control of extension growth in some of the species of Pinus tested prevailed for only a matter of 14 weeks after germination ; thereafter resting buds were formed regardless of treatment. Although most of the published studies on the environmental control of developmental periodicity in ecotypes have stressed the role of one variable -temperature or photoperiod - there seems little doubt that in nature the two will always interact. As mentioned in an earlier paragraph, all phases of the developmental cycle, vegetative and reproductive, may be subject to environmental regulation; in addition, innate timing mechanisms may operate to control the length of certain phases or to determine when sensitivity to environmental control is developed. Given that the genetic control of these different regulatory mechanisms is such as to permit continuous quantitative variation (which implies that it should be polygenic) it is apparent that an appropriate response can be built up for almost any conceivable annual climatic pattern in which some period favourable for plant growth occurs. Moreover, the same kind of behaviour may evolve utilizing different combinations of mechanisms; thus a given phenological pattern may depend in one

234

J. HESLOP-HARRISON

species on a vernalization response, and in another upon spring photoperiod. This kind of variation could account for the exceptions to the general rules observed in the studies of Vaartaja (1959) and McMillan (1959). We may take as a final example of a comprehensive investigation of climatic adaptation the study of Oxyria digyna by Mooney and Billings (1961). This work has combined a thorough study of the physiology of natural populations with the imaginative use of controlled environment cabinets, with the aim of elucidating the adaptive significance of interpopulation differences. 0. digyna in North America occurs in the arctic tundra as far north as 83" N, and as far south as the mountains of Arizona. The populations can be ranked into two primary groups largely on the presence or absence of rhizomes, the rhizomatous forms occurring in the northern part of the range, and the non-rhizomatous from Alberta southwards. Mooney and Billings studied populations from fifteen localities throughout the North American range, and representatives of several of these were examined under controlled environment conditions. The principal comparison was between performances of plants of different provenance in two environments simulating '(arctic" and "alpine" conditons respectively. The '(arctic" chamber provided growing season conditions comparable with those of the Alaskan Arctic Coastal Plain at 71" N, with constant light and alternating 12-h thermoperiods of 55" F and 35" P. The "alpine" chamber simulated conditions in the Medicine Bow Mountains in Wyoming, with a 15-h photoperiod and alternating 12-h thermoperiods of 65" F day and 40" F night. Growth and development were followed under the contrasting conditions, and plants were

0-4

5-9

I

,

,

,

10-14 15-19 20-24 25-29 30-34 Temperature

I

35-39

40-44

("C)

FIG.5 . Average photosynthetic and respiration rates of southern alpine population group of 0.ryria digynn [latitudes 41" 20' N, 35" 42' N and 48" 40' N] and of a northern population group [65O 25' N and 69" 50' N] a t different temperatures. Plants grown initially in the "arctic" chamber, and measurements made in a closed system using an infra-red gas analyser. From Mooney and Billings (1961). 0 photosynthetic rates, northern group; 0 photosynthetic rates, southern group; @ respiration rates, x respiration rates, southern group. northern group;

235

FORTY YEARS OF GENECOLOGY

extracted from the chambers as required for the measurement of photosynthesis and respiration. From the many data presented by Mooney and Billings, a selection may be considered to illustrate some of their principal conclusions. (a) Plants of northern populations have a higher photosynthetic rate at lower temperatures and attain a maximum rate at lower temperatures than do plants of the southern alpine populations, and, ( b ) plants of northern populations have higher respiration rates a t all temperatures than do plants of the southern alpine populations. . The evidence is contained in Fig. 5, showing pooled results from three southern alpine and two arctic population samples. The alpine group shows a temperature optimum for photosynthesis at the light intensity used of 30-40" C, and the arctic group, an optimum around 15-19" C, the maximum rates achieved being much the same. The respiration rate of the northern group is not only generally higher but shows a steeper rise with temperature-than that for the southern group, indicating a lower temperature compensating point for the arctic populations. (c) High-elevation, low-latitude plants attain photosynthetic light saturation a t a higher light intensity than do low-elevation, highlatitude plants. This is evident from Pig. 6, in which rate intensity curves for a southern alpine and a northern population are compared at 20" C. It is notable that the southern, high-altitude population did not achieve complete light saturation even at 5 200 f.c, whilst the arctic population is saturated at 2 000 f.c. ( d ) There is a clinal increase in the photoperiodic requirement for flowering from southern to northern populations. Data illustrating this for four populations from different latitudes are summarized in Table I

I

5-

I

t

1

4-

Y

2

I

I

I

I

I

I

I

500 1000 1500 2000 3000 4000 5000

Foot candles

FIG.6. Photosynthetic light curves at 20" C for a southern alpine population of Ozyn'a d i g y w [latitude 39" 40' N] and a northern population [61° 23' N]. From Mooney and Billings, 1961. southern alpine population; - - - - - - northern population.

236

J. H E S L 0P-H A R R I S 0 N

IX. It is evident that in response to photoperiod the latitudinal variation follows the same pattern as that discussed above for various grass and tree species.

TABLEI X Phenology of Oxyria digyna Plants from Four Populations in the "Arctic" a,nd "Alpine" Growth Chambers. [Data of Mooney and Billings, 1961.1 Latitude of population

38" 42' N 41" 20' N 48' 42' N 69" 60' N

4 weeks Arctic Alpine

20 30 10 30

Percentage flowering after: t? weeks 16 weeks Alpine Arctic Alpine Arctic

100 80 20 0

80 100

60

30 0 0

90

100

100 100 100 100

100 100 60

0

( e ) The northern populations are substantially less tolerant of high temperatures than are the southern ones. This is illustrated by the summer death rate curves of the northern and southern populations compared in Fig. 7. In discussing the general performance of plants from the different populations, Mooney and Billings conclude that the evidence indicates a close adjustment to the specific light climate of the habitat through adaptations in growth and flowering response, perennating bud formation, I

I

I

I

I

Date

FIQ.7. Summer death rate of plants of two northern and two southern populations of O X Y T ~digyna U under greenhouse conditions in Durham, North Carolina. 61" 23' N; -.-. -. 68" 56' N; --------lo 20' N; - - - - - 48" 42' N. From Mooney and Billings (1961).

- --

FORTY YEARS O F GENECOLOGY

237

and photosynthetic light saturation. Thus, the southern populations revealed a clear adaptation to growth under a 15-h photoperiod, while the northern populations reached active growth only in very long days and were brought into a state of dormancy by 15-h days, I n respect to temperature they point to the clear adaptive sigdcance of the lower temperature optimum for photosynthesis evident in the high latitude populations, which normally experience summer temperature maxima of less than 60’ in their natural environment. The higher respiration rate shown by the northern populations they consider also to be adaptive in permitting rapid metabolic rates to be achieved under lower average day temperatures. As a physiological investigation of genecological differentiation, the work of Mooney and Billings is exemplary in asking at the outset the essential question “How are the ecotypes related to their respective environments?”, and in adopting field and laboratory methods and measurements designed specifically to answer it. Their study should set a pattern for others in this field.

IV. CONCLUSIONS

It is appropriate in conclusion to refer again to the synthetic nature of genecology as a discipline combining ideas and methods from genetics, taxonomy and plant physiology, and to emphasize once more the value of this kind of concerted approach to the problems of population differentiation and adaptation in plants. An unfortunate trend during the post-war period has been apparent in recurrent attempts to assimilate genecology into taxonomy. Genecological investigations of infraspecific variation naturally tend to reveal facts of potential taxonomic significance, but taxonomic revision is not in itself an essential part of genecology. Should it be substituted for the original aims of analysing patterns of ecological adaptation and elucidating the means by which they are achieved, genecology becomes inseparable from the discipline commonly termed “experimental taxonomy’’ or “biosystematics” “Experimental taxonomy” seems first to have been used by Clausen et al. (1934), in substitution for “evolutionary taxonomy”, previously used by Clements and others of the Carnegie group. The term was adopted by Gregor et al. (1936),and defined as the classification of evolutionary units on the basis of experimentally derived facts relating to distribution, ecology and cytogenetics as well as to morphology. Valentine (1961)gave a slightly more dynamic version, “the study of evolutionary processes in plants and of the bearings of this study on their taxonomy.” “Biosystematics”, as “Biosystematy”, was introduced by Camp and Gilly (1943),with a broadly similar meaning: “(1) to delimit the natural biotic units and (2) to apply to these units a system

238

J. HESLOP-HARRISON

of nomenclature adequate to the task of conveying precise information regarding their defined limits, relationships, variability and dynamic structure’ ’. By definition, as well as etymologically, experimental taxonomy and biosystematics thus relate to a discipline that is part of taxonomy in the broad sense, which genecology is not. The distinction is no trivial one, since it is now possible to see that a conflict of purposes among students of infraspecific variation in plants has been a continuing source of dispute. Writing with the problems of evolutionary processes in mind, Epling and Catlin (1950) deplore the “conceptual error” which they state has led some workers to express “. . . the inter-relation between population and environment by typification and classification rather than by consideration of the adaptive relations of the interbreeding individuals.” What these authors fail here to acknowledge is that to attempt this type of classification is it respectable enough activity when the purpose of a study i s taxonomic. Conversely, the reservations voiced by Turrill (1938, 1946) about the contribution of genecology to taxonomy are those to be expected from a taxonomist who finds genecological aims and concepts incompatible, in part at least, with the process of perfecting a general system of classification and nomenclature. Biosystematics (or experimental taxonomy) should accordingly be preserved as something distinct from genecology in so far as there is a difference of aims. Re-definitions of biosystematics which have the effect of making the word practically synonymous with genecology (such as that of Clausen et al., 1945) are therefore undesirable. This is not, of course, to argue that genecological and biosystematic aims cannot be pursued at one and the same time in any particular study of infraspecific variation, but simply to urge that if they are, they should be recognized as not being identical. Because the purposes of study are different, different kinds of evidence are required. Thus, ecological data and observations on genetic systems are a sine qua n o n of genecology, although by no means an essential part of taxonomy. Conversely, the nomenclatural and bibliographical studies which are an obligatory part of any taxonomic study are not necessarily significant for a genecological investigation of a species. If the taxonomic aspects of genecology have often been over-emphasized, it is equally true that the genetical and physiological phases have been as frequently understressed. Baker’s regret (1953) that so few studies of race differentiation up to that time had paid adequate attention to the role of breeding systems is less warranted now that publication of his own papers and Stebbins’ fine “Variation and Evolution in Plants” (1950) have alerted authors to the importance of this and other genetical factors, but it is still true that genecological or near-geneco-

F O R T Y YEARS O F QENECOLOGY

239

logical work is being published that reveals a lack of appreciation of some highly relevant facets of population genetics theory. This is reflected in deficiences in technique, particularly in respect t o surveying and sampling methods. I n this connection, critiques such as those of Wilkins (1959) and Harberd (1957, 1958), discussed at some length above, are of special note. Morley (1959),discusses the general limitation of survey methods as an approach to understanding genecological variation. Studiee of population differentiation, he states, “should attempt not only to demonstrate the existence of selection pressure, but also to evaluate selection intensities, to specify mechanisms of adaptation, and to determine the effect of natural selection on population, distribution, structure and number.” Genecological surveys “may disclose relationships between features of the habitat and characteristics of the plants, which hint at, or clearly propose, mechanisms of adaptation. Since selection coefficients cannot be measured by scch relationships, characters of major importance cannot be distinguished from those of minor, though real, significance”. Morley urges the value of the study of natural or artificial populations exposed to selective forces under experimental conditions for the measurement of selection intensities, and refers to work on cultivated plants and man-managed plant communities which has yielded evidence directly relevant to adaptation and population differentiation in wild species. That limitations of space have prevented any consideration of this work in the present review is not to be interpreted as implying that it is not significant for the main theme. Particularly in respect to biotic influences, work on agricultural plants has provided data unmatched as yet from wild populations, and there is every reason to suppose that agricultural experimentation will continue to supply a significant proportion of the evidence relating to selective processes in general. The experimental study of selection in plant populations should obviously be complemented, where feasible, with investigations of mechanisms of adaptation, and we may expect to see much more effort in this field in the next decade or so. This is ensured, if by nothing else, by the increasing availability of controlled environment equipment, and by the steady improvement of the instrumentation necessary to study plant environment and plant responses. If the effort is to be rewarded by commensurate results in the improved understanding of adaptation and adaptive processes, there is no avoiding the necessity for a careful appraisal of the aims of physiological study of genecologically differentiated populations, and of the appropriateness of the techniques adopted. This illustrates again the importance of an understanding of genecology as a synthetic discipline. It will be all too easy to deploy sophisti-

240

J. HESLOP-HARRISON

cated equipment on ill-conceived programmes generating quite the wrong kind of data, and this will certainly happen if physiological investigations come to be based upon too narrow an acquaintance with the many and complex factors involved in the evolution of plant populations.

ACKNOWLEDGMENTS To my colleagues, Professor K. Mather and Dr D. A . W a i n s , I am

indebted for much stimulating discussion of the topic of this article; they may have influenced what I have written, but they are in no way responsible for its deficiencies. My warm thanks are due to Miss Gillian Nolan for assistance with the bibliography.

REFERENCES Anderson, E. (1949). “Introgressive Hybridization”. New York : John Wiley and Sons. Anderson, E. (1953). Biol. Rev. 28,280-307. Introgressivehybridization. Baker, H. G. (1947).J . Ecol. 35, 271-292. Melandrium (Roehlingem.) Fries. Baker, H. G. (1952).Evolution 6,61-68. The ecospecies -prelude to discussion. Baker, H. G. (1953). Symp. SOC. exp. Biol. 7, 114-145. Race formation and reproductive method in flowering plants. Baker, H.G. (1959). ColdSpr. Harb.Symp. quant. Bwl. 24,177-191. Reproductive methods as factors in speciation in flowering plants. Barber, I€.N. (1955). Evolution 9, 1-14. Gene substitutions in Eucalypts. Barber, H. N. and Jackson, W. D. (1957). Nature, Lond. 179, 1267. Natural selection in Eucalyptus. Bateman, A. J. (1947). Heredity 1,235-246. Contaminationof seed crops. I1 Wind pollination. Bjorkman, 0. and Holmgren, P. (1963). Physiol. Plant. 16,889-914. Adaptability of the photosynthetic apparatus to light intensity in ecotypes from exposed and shaded habitats. Blake, S. T. (1944). Univ. Qkensland Paper 2, 1-62. Monographic studies in the Australian Andropogoneae, Part I. Revisions of the genera Bothriochloa, Capillipedium, Chrysopogon, Vetiveria and Spathia. Bocher, T. W. (1945). Dansk. Bot. Ark. 12, No. 3. Some experiments to elucidate the influence of winter conditions on shoot development and floral initiation on various races of Pmnella vulgaris and Ranunculua acer. Bocher, T. W. (1949). New Phytol. 48, 285-324. Racial divergencies in Pmnella vulgaris in relation to habitat and climate. Bocher, T. W., Larsen, K. and Rahn, K. (1953). Hereditas 39, 289-304. Experimental and cytological studies on plant species: I Kohlrauschia prolijera and Plantago coronopua. Bocher, T. W., Larsen, K. and Rahn, K. (1955).Hereditas 41,423-453. Experimental and cytologicalstudies on plant species. I11 Plantago coronopus and allied species.

FORTY YEARS OF GENECOLOGY

241

Boyko, H. and T a b o r , N. (1954).Bull. Res. Counc. Israel I V , 241-248. An arid ecotype of Dactylis glomerutu L. (Orchard grass) found in the Negev (Israel). Bradshew, A. D. (1952).Nature, Lond. 169, 1098. Populations of Agrostis tenuis resistant to lead and zinc poisoning. Bradshaw, A. D. (1959).New Phytol. 58, 208-227. Population differentiation in Agrostis tenuis. I Morphological differentiation. Bradshaw, A. D. (1960).New Phytol. 59, 92-103. Population differentiation in Agmstis tenuis 111Populations in varied environments. Briggs, Barbara G . (1962).Evolution 16,372-390. Hybridization in RanuneuEus. Brown, W. V. (1952). Bot. Guz. 113, 438-444. The relation of soil moisture to cleistogamy in Stipu kucotricha. Brown, W. V. and Emery, W. H. P. (1958).A m r . J . Bot. 45,253-263. Apomixis in the Gramineae :Punicoideae. Cain, A. J. (1954).“Animal Species and their Evolution”. London: Hutchinson. Cain, S. A. (1944).“Foundations of Plant Geography”. New York and London: Harper. Camp, W. H. (1951).Brittonia 7,113-127. Biosystematy. Camp, W.H. and Gilly, C. L. (1943).Brittoniu 4,323-385.The structure and origin of species. Chouard, P. (1960).Ann. Rev. Plant. Phys. 2, 191-237. Vernalization and its relations to dormancy. Clausen, J. (1951).“Stages in the Evolution of Plant Species”. Cornell University Press. Clausen, J. (1954).Cuqobg.ia, VoZ. suppl. 1954,469-479. Partial apomixis as an equilibrium system in evolution. Clausen, J. and Hiesey W.M. (1958a). Publ. Carneg. Instn. No. 615. Experimental studies on the nature of species. IV Genetic structure of ecological races. Clausen, J. and Hiesey, W. M. (1958b).Rep. Scottish Plant Breeding Stat. 1958, 41-51. Phenotypic expression of genotypes in contrasting environments. Clausen,J. and Hiesey, W. M.(1960).Proc. nut. Acad.Sci., Wash. 46,494-506. The balance between coherence and variation in evolution. Clausen, J., Keck D. D. and Hiesey, W. M. (1934).Yearb. Curneg. Inst. 33, 173177.Experimental taxonomy. Clausen, J., Keck, D. D. and Hiesey, W. M. (1939).Amer. J . Bot. 26, 103-106. The concept of species based on experiment. Clausen, J., Keck, D. D. and Hiesey, W. M. (1940).Publ. Curneg. Instn. No. 520. Experimental studies on the nature of species. I Effect of vaned environments on western North American plants. Clausen, J., Keck, D. D. and Hiesey, W. M. (1945).Publ. Curneg. Instn. NO.564. Experimental studies on the nature of species. 11Plant evolution through amphiploidy and autoploidy, with examples from the Madiinae. Clausen, J., Keck, D. D. and Hiesey, W. M. (1948).Publ. Curneg. Instn. No. 581. Experimental studies on the nature of species. 111Environmental responses of climatic races of Achilleu. Colwell, R.N. (1951).A m r . J . Bot. 38, 511-523. The use of radioactive isotopes in determiningspore distribution patterns. Cook, S. A. (1962).Evolution 16, 278-299. Genetic system, variation and adaptation in Eschscholzia cal$brnica. Cooper, J. P. (1951).J . Ecol. 39, 228-270. Studies on growth and development in LoZium. I1 Pattern of bud development of the shoot ccpex and its ecological significance.

242

J. HE SLOP-HARRIS O N

Cooper, J. P. (1952). J . Ecol. 40, 352-378. Studies on growth and development in Lolium. I11Influence on season and latitude on ear emergence. Crosby, J. L. (1956).J . Genet. 54, 1-8. A suggestion concerning the possible role of plasmagenes in the inheritance of acquired adaptations. Darlington, C. D. (1939). “The Evolution of Genetic Systems”. Cambridge University Press. Daubenmire, R. F. (1950). Bot. Gaz. 112,182-188. Geographic races of Pinus. Dodds, J. G. (1953).J . Ecol. 41,467-478. Plantago coronopus L. Durrant, A. (1958). Nature, Lond. 181, 928-930. Environmental conditioning of flax. Durrant, A. (1962). Heredity 17, 27-61. The environmental induction of heritable changes in Linum. Ehrendorfer, F. (1953). Osterreich. Bot. Zeits. 100, 616-637. Okologisch. geographische Mikro-Differenzier einer Population von Galium pumilum M m . s.str. Epling, C. and Catlin, W. (1950). Heredity 4, 313-325. The relation of taxonomic methods to an explanation of organic evolution. Faegri, K. (1937). Bot. Rev. 3,400-423. Fundamental problems of taxonomy and phylogenetics. Fryxell, P. A. (1959). “Advances in Botany”, Vol. 1, 887-891. Toronto Univ. Press. The evolutionary position of inbreeding systems. Biosystematics Symp., 9th Int. Bot. Congr., Montreal. Grant, V. (1949).Evolution, 3,82-97. Pollination systems as isolating mechanisms in angiosperms. Grant, V. (1950). El Aliso 2, 239-316. Genetic and taxonomic studies in Gilia. I Gilia capitata. Grant, V. (1952). El Aliso 2, 361-373. Genetic and taxonomic studies in Cilia. I1 @iliacapitata abrotanifolia. Grant, V. (1954). El Aliso 3, 1-18. Genetic and taxonomic studies in Qilia. IV Gilia achilleaefolia. Grant, V. (1958). Cold Spr. Harb. Symp. quant. Biol. 23, 337-363. The regulation of recombination in plants. Gregor, J. W. (1938). New Phytol. 37, 15-49. Experimental taxonomy, I1 Initial population differentiation in Plantago maritima L. of Britain. Gregor, J. W. (1939). New Phytol. 39, 293-322. Experimental taxonomy. I11 Population differentiation in North American and European sea plantains allied to Plantago m r i t i m a L. Gregor, J. W. (1944). Biol. Rev. 19,20-30. The ecotype. Gregor, J. W. (19464. New Phytol. 45, 254-270. Ecotypic differentiation. Gregor, J. W. (194613). Trans. Bot SOC.Edinb. 34, 377-383. Some reflections on infraspecific ecological variation and its classification. Gregor, J. W. (1956). Proc. roy. Soc. B. 145, 333-337. Adaption and ecotypic components. Gregor, J. W. and Lang, 5. M. S. (1950). New Phytol. 49, 135-141. Intra-colonial variation in plant size and habit in sea plantains. Gregor, J. W. and Watson, Patricia J. (1954). New Phytol. 53, 291-300. Some observations and reflections concerning the patterns of intraspecifk differentiation. Gregor J. W. and Watson, Patricia J. (1961). Evolution 15, 166-173. Ecotypic differentiation. Gregor, J. W., Davey, V. McM. and Lang, J. M. S. (1936). New Phytol. 35, 323350. Experimental taxonomy. I Experimental garden technique in relation to the recognition of smalltaxonomic unite.

FORTY YEARS OF G E N E C O L O G ) ~

243

Gustafsson, A. (1947). Acta Univ. L m d . N.F. Avd. 2 42-3, 1-307. Apo&i~ in higher plants. Habeck, J. R. (1958). Ecology 39,457-63. White cedar ecotypes in Wisconsin. Harberd, D. J. (1957). New Phytol. 56, 269-280. The within population variance in genecological trials. Harberd D. J. (1958). Rep. Scot. Plant Breeding Stat. 1958. 52-60. Progem and prospects in genecology. Harberd, D. J. (1961a). New Phytol. 60, 18k-206. Observations on population structure and longevity of Festuca rubra L. Harberd, D. J. (1961b). New Phytozl 60, 325-338. The case for extensive rather than intensive sampling in genecology. Harlan, J. R. (1945a). Arner. J . Bot. 32, 66-72. Cleistogamy and chasmogamy in Bromus carinatus Hook, and Am. Harlan, H. R. (1945b). Amer. J . Bot. 32, 142-148. Natural breeding structure in the Bromus carinatus complex, as determined by population analyses. Harland, S. C. (1946). Heredity 1, 121-125. An alteration in gene frequency in Ricinus communis L due to climatic conditions. Haskell, G. (1953). Heredity 7, 409-418. Quantitative variation in subsexual Rubus. Haskell, G. (1959). Qenetica 30,240-260. Stabilityandvariation in apomictic Rubus. Heslop-Harrison, J. (1951). Svensk. Bot. Tidskr. 45, 608-635. A comparison of some Swedish and British forms of Orchis maculata L. sens. lat. Heslop-Harrison, J. (19534. Veroff.Geobot. Inst. Rubel fiir 1953,53-82. A synopsis of the dactylorchids of the British Isles. Heslop-Harrison, J. (1953b). “New Concepts in Flowering Plant Taxonomy”. London: Heinemann. Heslop-Harrison, J. (1955). I n “Species Studies in the British Flora”, p. 161. Arbroath: Buncle. The conflict of categories. Heslop-Harrison, J. (1956). Proc. Linn. Soc. Lond. Session 166, 51-83. Some observations on Dactylorchis incarmta L. Heslop-Harrison, J. (1958). Uppsala Uniw. h s s k r . 1958 6, 150-158. EcoIogical variation and ethological isolation. Heslop-Harrison, J. (1959a).Evolution 13, 145-147. Variability and environment. Heslop-Harrison, J. (1959b). Ann. Bpt. N.S. 22,345-349. Photoperiod and fertility in Rottboellia exaltata L.f. Heslop-Harrison, J. ( 1 9 5 9 ~ ) “Advances . in Botany.” Vol. 1, 891-895. Toronto University Press. Apomixis, environment and adaptation. Biosystematics Symposium, 9th Int. Bot. Congress, Montreal. Heslop-Harrison, J. (1961). Phytomorph. 11, 378-383. The function of the glume pit and the control of cleistogamy in Bothriochloa decipiens. Heslop-Harrison, J. (1962). Symp. SOC.Gen. Microbiol. 12, 14-36. Purposes and procedures in the taxonomic treatment of higher organisms. Heslop-Harrison, J. (1963). I n “Chemical Plant Taxonomy”. 17-40. London and New York: Academic Press. Species concepts : theoretical and practical aspects. Heywood, V. H. (1959). Systematics Association Publ. No. 3, 87-112. The taxonomic treatment of ecotypic variation. Hiesey, W. M., Clausen, J. and Keck, D. D. (1942). Amer. Nat. 7 6 , 6 2 2 . Relations between climate and intraspecific variation in plants. Hiesey, W. M. (1953). Evolution 7, 297-316. Comparative growth between and within climatic races of Achillea under controlled conditions. Hiesey, W. M., Milner H. W. and Nobs, M. A. (1959). Yearb. Curneg. Imtn. 58, 350-353. Controlled cabinets for plant growth.

244

J. EESLOP-HARRISON

Hiesey, W. M., Milner, H. W. and Nobs, M. A. (1960). Yearb. Carneg. Instn. 59, 318-319. Responses of Mimulua in controlled environments. Hiesey, W. M., Nobs, M. A., and Milner, H. W. (1961). Yeurb. Carneg. Instn. 60, 381-384. Responses of M6mulua races and hybrids a t the transplant stations. Huxley, J. S. (1938). Nature, Lond. 142, 219. Clines, an auxiliary taxonomic principle. Huxley, J. S. (1942). “Evolution: The Modern Synthesis”. London: Allen and Unwin. Huxley, J: S. (1955).Heredity 9, 1-52. Morphism and evolution. Ketellapper, H. J. (1960). Ecology 41, 298-305. Growth and development in Phalaris. Knox, R. B. and Heslop-Harrison, J. (1963). Bot. Not. 116, 127-141. Experimental control of aposporous apomixiSin a grass of the Andropogoneue. Kruckeberg, A. R. (1950). Amer. J. Bot. 38, 408-418. Response of plants to serpentinesoil. Kruckeberg, A. R. (1954). Ecology 35, 267-274. The ecology of serpentine soils. 111Plant apecies in relation to serpentinesoils. Langlet, 0.(1934). Meddel. Skogsfkksoksanst. Stockh. 27, 87-93. Om variationen hos tallen Pinus sylvestris och dess samband med klimatet. Langlet, 0. (1936). Meddel. SkogsfiCrsoksanst.Stockh. 29, 219-470. Studien uber die physiologische Variabilitiit der Kiefer und deren Zusammenhang mit dem Klina. Beitrlige zur Kenntnis der Okotypen von Pinus sylwestris L. Langlet, 0. (1959). Sylvae Genetioa 8, 13-22. A cline or not a cline - a question of scots Pine. Langlet, 0.(1963). Nature, Lond. 200,347-348. Patterns and terms of intraspecific ecological variability. Larsen, E. C. (1947). Bot. Gaz. 109,132-150. Photoperiodicresponses of geographical strains of Andropogon acoparius. McKell, C . M., Perrier, E. R. and Stebbins, G. L. (1960). Ecology 41, 772-778. Responses of two subspecies of orchardgrass (Dactylis glomerata subsp. lusitanica and judaica) to increasing soil moisture stress. McMillan, C. (1956a). Ecology 37, 330-340. Nature of the plant community. I Uniform garden and light period studies of five grass taxa in Nebraska. McMillan, C. (1956b). Amer. J. Bot. 43, 429-436. Nature of the plant community. 11Variation in flowering behaviour within populations of Andropogon SCOparius. McMillan, C. (1959). Ecol. Monogr. 29, 285-307. The role of ecotypic variation in the distribution of the central grassland of North America. McVean, D. N. (1953). Watsonia 3, 26-32. Regional variation of Alnus glutinosa L. Gaertn. in Britain. Manley, G. (1954). “Climate and the British Scene”. London: Collins. Marsden-Jones, E. M. and Turrill, W. B. (1938)*J. Ecol. 26, 380-389. Transplant experimentsof the British Ecological Society at Potterne, Wiltshire. Marsden-Jones, E. M. and Turrill, W. B. (1957). “The Bladder Campions”. London :Ray Society. Mather, K. (1943). Biol. Rev. 18, 32-64. Polygenic inheritance and natural selection. Mather, K. (1953). S.E.B. Symposium No. V I I “Evolution”, 66-95. The genetical structure of populations. Mather, K. (1955). Evolution 9, 52-61. Polymorphismin disruptive selection. Mayr, E. (1942). “Systematics and the Origin of Species”. New York: Columbia University Press.

FORTY YEARS OF OENECOLOOY

245

M a p , E.(1947). Evolution 1,263-288. Ecological factors in speciation. Mayr, E. (1954). I n “Evolution 88 a Process”. London: Allen and Unwin. Change

of genetic environment and evolution. Mayr, E. (1959). Proc. A m r . Phil. SOC.103, 221-230. Isolation aa an evolutionary factor. Milner, H. W., Hiesey, W. M. and Nobs, M. A. (1,962). Yearb. Carnq. Instn. 61, 313-319. Physiology of climatic races. Mooney, H. A. and Billings, W. D. (1961). Ecol. Monogr. 31, 1-29. Comparative physiological ecology of arctic and alpine populations of 0zy:Yriadigyna. Morley, F. H. W. (1959). Cold Spr. Hurb. Symp. p u n t . Bwl. 24, 47-55. Natural selection and variation in plants. New, June K. (1958). Ann. Bot. N.S. 22,457-477. A population study of Spergulu amem’s. New, June K. (1959). Ann. Bot. N.S. 23, 23-33. A population study of Spergula arvensis. Nobs, M. A. (1963). Publ. Carneg. Inatn. No. 623. Experimental studies on species relationships in Ceanothw. Nygren, A. (1951). Heredity 37, 519-532. Form and biotype formation in Cahmagrostis purpurea. Olmsted, C. E. (1944a). Bot. Gaz. 106, 46-74. Growth and development in range grasses IV. Olmsted, C. E. (194413). Bot. Gaz. 106,382-401. Growth and development in range grasses V. Olmsted, C. E. (1951). Bot. Ouz. 112, 365-393. Experiments on photoperiodism, dormancy, and leaf age and abscission in sugar maple. Perry, T. 0. and Wang Chi Wu, (1960). Ecology 41, 790-793. Genetic variation in the winter chilling requirement for date of domancy break for Acer rubrum. Ryle, G. J. A. (19634. Ann. Bot. N.S. 27, 453-465. Studies in the physiology of flowering in Timothy (Phleum prateme L.). I11 Effects of shoot age and nitrogen level on the timing of inflorescence production. Ryle, G. J. A. (1963b). Ann. Bot. N.S. 27, 467-479. Studiea in the physiology of flowering in Timothy (Phleum pratense L.). IV Effects of shoot age and nitrogen level on the size of the inflorescence. Ryle, G. J. A. and Langer, It. H. M. (1963a). Ann. Bot. N.S. 27, 213-230. Studies on the physiology of flowering in Timothy. I Influence of daylength and temperature on initiation and differentiation of the inflorescence. Ryle, G. J. A. and Langer, R. H. M. (1963b). Ann. Bot. N.S. 27,233-252. Studies on the physiology of flowering in Timothy. I1 Influence of daylength and temperature on size of the inflorescence. Sinskaia, E. N. (1931). Bull appl. Bot. PI-Breed. 25, 50-97. The study of species in their dynamics and interrelation with different types of vegetation. Sinskaia, E. N. (1958). Rep. Scot. Plant Breed. Stat. 1958, 31-40. Investigations on the composition of ecotypical and varietal populations. Smith, A. J. E. (1963). Watsonia 5, 336-367. Variation in Mehmpyrumpratense L. Snaydon, R. W. and Bradshaw, A. D. (1961). New Phytol. 62,219-234. Differential response to calcium in Festuca ovina. Stebbins, G. L. (1950). “Variation and Evolution in Plants”. New York: Columbia University Press. Stebbins, G. L. (1957). Amer. Nat. 91, 337-354. Self fertilisation and population variability in the higher plants. Stebbins, G. L. (1958). Cold.Spr. Harb. Symp. quant. Biol. 23, 365-378. Longevity, habitat and release of genetic variability in the higher plants. I

C.E.R.

246

J. HESLOP-HARRISON

Stebbins,G. L., Matzke, E. B. and Epling, C. (1947). Evolution 1, 79-88. Hybridization in a population of Quercw,mrilandica and Quercw,ilicifolia. Stebbins, G. and Zohary, D. (1959). Univ. Caiij. Publ. Bot. 31,l-40. Cytogenetic and evolutionary studies in the genus Dactylis. Thoday, J. M. (1953). S.E.B. Symp. No. V I I c‘Evolution”,96-113. Components of fitness. Thoday, J. M. (1959). Herediby 13, 187-203. Effects of disruptive selection. Thoday, J. M. and Boam, T. B. (1959). Heredity 13,205-218. Effects of disruptive selection. I1 Polymorphism and divergence without isohtion. Thoday, J. M. and Gibson, J. B. (1962). Nature, Lond. 193, 1164-1166. Isolation by disruptive selection. Turesson, G. (1922a). Hereditas 3, 100-113. The species and the variety as ecological units. Turesson, G. (1922b). Hereditas 3, 211-350. The genotypicrtl response of the plant species to the habitat. Turesson, G. (1923).Hereditas 4,171-176. The scope and import of genecology. Turesson, G. (1925). Hereditas 6, 147-236. The plant species in relation to habitat and climate. Turesson, G. (1930).Hereditas 14, 99-152. The selective effect of climate upon the plant species. Turesson, G. (1936). Bot. Not. 1936 420-427. Rassenokologie und P b e n g e o graphie. Turesson, G. (1943). Bot. Not. 1943, 413-427. Variation in the apomictic microspecies of Alchemilla vulgaris L. Turesson, G. (1956). Bot. Not. 109, 400-404. Variation in the apomictic microspecies of AlchmiUa vulgarie L. Twill, W. B. (1938). Biol. Rev. 13, 342-373. The expansion of taxonomy with special reference to Spermatophyta. Turrill, W. B. (1946).New Phytol. 4 5 , 3 6 4 3 . The ecotype concept: a consideration with appreciation and criticism, especiallyof recent trends. Uphof, J. C. T. (1938). Bot. Rev.4,21-49. Cleistogamio flowers. Vaartaja, 0. (1954). Canad. J . Bot. 32, 392-399. Photoperiodic ecotypes of trees. Vaartaja, 0. (1959). Ecol. Monogr. 29,91-111. Evidence of photoperiodic ecotype in trees. Valentine, D. H. (1949). Acta B i o t h . Leiden 9, 75-88. The units of experimental taxonomy. Valentine, D. H. (1961). I n “Recent Advances in Botany”. University of Toronto Press. Biosystematics:Introduction. Waddington, C. H. (1953a). S.E.B. Symrp. No. V I I “Evolution”, 186-199. Epigenetics and evolution. Waddington, C. H. (1963b). Evolution 7, 118-126. Genetic assimilation of an acquired character. Walker, R. B. (1954). Ecology 35, 259-266. The ecology of serpentine soils. 11 Factors ctffectingplant growth on serpentinesoils. Wareing, P. (1950a). Physiol. Plant. 3, 258-276. Growth studies in woody species. I Photoperiodism in first year seedlings of Pinw, aylveSt&s. Wareing, P. (1950b). Physiol. Plant. 3, 300-314. Growth studies in woody species. IIEffect of day length on shoot-growthin Pinus sylve~trieafter the Grst ye=. Wareing, P. (1961). Physiol. Plant. 4, 41-56. Growth studies in woody species. III Further photoperiodic effectsin Pinus aylvestria.

FORTY Y E A R S O F GENECOLOGY

247

Wareing, P. (1953). Physiol. Plant. 6, 692-706. Growth studies in woody species. V Photoperiodism in dormant buds of Faqw sylvatica. Wareing, P. (1954). Physiol. Plant. 7, 261-277. Growth studies in woody species. VI The locus of photoperiodic perception in relation to dormancy. Wareing, P. (1956). Ann. Rev. PI. Physiol. 7 , 191-214. Photoperiodism in woody plants. Wassink, E. C., Richardson, S. D. and Pieters, G. A. (1956). Acta Bot. Neerl. 5, 247-256. Photosynthetic adaptation to light intensity in leaves of Acer pseudoplantanus. Weidman, R. H. (1939).J. aqric. Res. 59,855-888. Evidences of racial influence in a 25 year test of Ponderosa pine. Whitehead, F. H. (1962).New Phytol. 61,59-62. Experimental studies of the effect of wind on plant growth and anatomy. I1Helianthus annuw. Whitehead, F. H. (1963a). New Phytol. 62, 80-86. Experimental studies of the effect of wind on plant growth and anatomy. I11 Soil moisture relations. Whitehead, F. H. (1963b). New Phytol. 62, 87-90. Experimental studies of the effect of wind on plant growth and anatomy. IV Growth substances and adaptative anatomical and morphological changes. Whitehead, F. H. and Luti, R. (1962). New Phytol. 61, 56-58. Experimental studies of the effect of wind on plant growth and anatomy. I Zeu mays. Whittaker, R. H. (1954). Ecology 35, 275-285. I Introduction; I1 The ecology of serpentine soils; IV The vegetational response to serpentine soils. Wilkins, D. A. (1957). Nature, Lond. 180, 37. A technique for the measurement of lead tolerance in plants. Wilkins, D. A. (1959). Rep. Scottish Plant Breed. Slat. 1959, 92-96. Sampling for genecology Wilkins, D. A. (1960a). Proc. Linn. Soc., Lond. Session 171,122-126. Recognizing adaptive variants. . Wilkins, D. A. (1960b). Rep. Scot. Plant. Bred. Stat. 1960, 85-98. The measurement and genetical analysis of lead tolerance in Festuca ovina. Wright, J. W. and Baldwin, H. I. (1957).Silvae Genetica 6, 2-14. The 1938 International Union Scotch Pine provenance test in New Hampshire.

.

I2

C.E.R.

Author Index Numbers in italics refer to the pages on which references are listed in bibliographies at the end of each article.

A Aaltonen, V. T., 111,152 Away, F. J., 115,152 Anderson, D. J., 75,98 Anderson, E., 213,240 Andersson, S. O., 107, 115, 122, 123, 124,152 AndrB, P., 107,152 Andrewartha, H. G., 9,25,26,54 Auten, J. T., 120,146,152 B Bailey, V. A., 36,49, 50,52,54, 56 Baker, H. G., 165, 167, 194, 196, 200, 212,238,240 Bddwin, H. I., 180, 181,247 Banks, C. J., 35,54 Barber, H. N., 183,240 Barnette, R. M., 110, 132,154 Bartholomew, W.V., 109, 145,148,152 Bateman, A. J., 205,240 Becking, J. H., 150,152 Bellamy, D., 80,98 Bess, H. A., 36, 54 Billings, W. D., 234,235,236,245 Bingham,F. T., 109,116,117,134,154 Birch, L. C., 9,26,54 Bjorkman, O., 171, 213, 224, 225,226, 240 Black, J. N., 127,152 Blake, S. T., 202,240 Blow, F. E., 110,138,144, 145,152,156 Boam, T. B., 206,207,209,246 Bocher, T. W., 182,240 Bocock, K. L., 144,152 Bohmerle, K., 111, 139,152 Bonnevie-Svendsen, C., 114, 119, 120, 122, 123, 124,131,132,152 Bornebusch,C.H., 111,136,140,152

Boyko, H., 223,241' Boysen-Jensen, P., 111,133,152 Bradshaw, A. D., 179, 184, 188, 219, 222,241,245 Brauns, D. A., 103,152 Brauns, F. E., 103,152 Bravenboer, L., 51,54 Bray, J. R., 75,.98, 144, 146, 148, 149, 153 Briggs, Barbara, G., 176,212,241 Broadfoot, W. M., 121,153 Brown, W. V.. 202,203,241 Burger, H., 107, 145,153 Burnett, T., 44, 50,51, 54 Bykova, L. N., 117,156

C Cable, D. R., 147,153 Cain, A. J., 165,206,241 Cain, S. A., 218,241 Camp, W. H., 237,241 Capstick, C. K., 144,152 Catlin, W., 238,242 Chandler, R. F., Jr., 102, 107, 115, 125, 127,137,153,154,157 Chant, D. A,, 44,54 Chitty, D., 9, 54 Chouard, P., 229,241 Claudot, M., 110,153 Clausen, J., 162, 163, 164, 165, 171, 172, 173, 179, 188, 189, 191, 203, 211,212,218,227,237,241,243 Coldwell, B. B., 111,153 Colwell, R. N., 205,241 Cooper, C. F., 147,153 Cooper, J. P., 228,241,242 Crosby, J. L., 216,242 Crosby, J. S., 132,153 Curtis, J. T., 68, 75, 76, 98, 118,153

249

250

AUTHOR INDEX

D Dagnelie, P., 75,98 Dale, M. B., 69, 77, 83, 84,98,99 Danckelmann, B., 112, 130, 136, 141, 153 Darlington, C. D., 194, 195, 202, 205, 242 Daubenmire, R. F., 181,242 Davey, V. McM., 162,165,237,242 Davidson, J., 25, 54 DeBach, P., 35,44,51, 52, 55 DeLong, W. A., 111,153 DeRuiter, L., 48, 55 Dietrick, E. J., 35,55 Dimock,E. J., 116,133,139,141,153 Dodds, J. G., 182,242 Dosse, G., 51,54 Dudkiewicz, L. A., 146,148,153 Durrant, A., 215,216,242

E

Ebermayer, E., 102, 107, 113, 129, 130, 131,139,141,153 Edwards, R. L., 44,51,57 Ehrendorfer, F., 186,209,242 Ehwald, E., 113, 115, 116, 117, 130, 148,153 Elton, C. E., 14,55 Emery, W. H. P., 203,241 Enander, J., 107, 115, 122, 123, 124, 152 Epling, C., 211,238,242,246

F Faegri, K., 161,162,194,242 Fleschner, C. A., 35,55 Franz, J., 9,55 Franz,J. M., 4,55 Fryxell, P. A., 202, 242 G Gaffron, H., 102,153 Gause, G. F., 51, 55 Gessel, S. P., 109, 116, 117, 134, 154 Gibson, J. B., 206,246 Gilbert, O., 144,152 Gilly, C. L., 237,241 Gjems, O., 114, 119, 120, 122, 123, 124, 131,132,152 Glasgow, J. P., 11,55 Glesinger, E., 103,153

Goodall, D. W., 59, 72, 75, 76, 80,98 Gounot, M., 86,98 Gradwell, G. R., 11,28, 29, 52, 53,58 Grant, V., 175, 179, 186, 194, 200, 212, 242 Gregor, J. W., 162, 164, 165, 185, 206, 237,242 Greig-Smith, P., 59, 64, 69, 76, 82, 86, 96, 98 Gunn, D. L., 10, 11,55 Gustafsson, A., 203,243

H

Habeck, J. R., 178,243 Hagen, K. S., 4,57 Hairston, N. G., 17,55 Handley, W.R. C., 125,147,153 Harberd, D. J., 80, 98, 168, 169, 170, 185,204,239,243 Harlan, H. R., 202,243 Harlan, J. R., 202,243 Harland, S. C., 183,243 Haskell, G., 203, 243 Hatch,A. B., 109, 118, 134, 135,154 Heikens, H. S., 48, 56 Heitkamp, D., 148,155 HeslopHarrison, J., 165, 166, 167, 172, 177, 178, 202, 203, 204, 212, 243, 244 Hesselman, 122, 123, 124,154 Heyward, F., 110,132,154 Heywood, V. H., 165,243 Hiesey, W. M., 162, 163, 164, 165, 171, 172,173,179,188,189,191,212,218, 226,227,229,237,241,243,244 245 Honing, C. S., 3, 42, 43, 44, 45, 46, 48, 49, 50, 55 Holmgren, P., 171, 213, 224, 225, 226, 240 Hopkins, B., 82,98 Huffaker, C. B., 35,52,55 Hughes, R. E., 86,98 Hurst, F. B., 110, 118, 135, 139, 140, 155 Huxley, J. S., 162, 183, 198,244 1 Isaac,L. A., 107,157

J

Jackson, W. D., 183,240

251

AUTHOR I N D E X

Jacobs, M. R., 145,154 Jhr6, Z., 113,154 Jenny,H., 109, 116, 117,134,154 Joffe, J., 122, 123,124,154 Jokela, E., 145,154

K

Keck, D. D., 162, 163, 164, 165, 172, 179,188,189,218,227,237,241,243 Kendall, M. G., 86, 98 Kendrick, W. B., 115,136,154 Kershaw, K. A., 82,98 Ketellapper, H. J., 230,244 Kim, T., 138,155 Kittredge, J., 110, 145, 146, 147,154 Klomp, H., 10, 11, 12, 16,26, 50, 51,52, 53, 55, 56, 57 Knox, R. B., 203,244 Knudsen, F., 107, 115, 122, 123, 124, 140,154 Kruckeberg, A. R., 179,220,221,244 Krutzsch, H., 113, 139,154 Kuenen, D. J., 26,56 Kumekawa, A., 146,148,156 Kunugi, R., 146, 148,156 Lack, D., 35,56 Lambert, J. M., 80, 82, 83, 98,99 Lang, J . M . S., 162,165,327,242 Langer, R. H. M., 228,245 Langlet, O., 161, 162, 172, 180, 181, 244 Larsen, E. C., 232,244 Larsen, K., 182,240 Laudelot, H., 109, 133, 145, 148, 152, 154 Lawrence, D. B., 148,155 Leopold, A., 48, 56 Lespeyres, 113,154 Levina, V. I., 117,154 Lieth, H., 134,157 Lindberg, S., 107, 115, 137,154 Lindley, D. V., 86,98 Lindquist, B., 115,120,144,154 Lunt,H.A., 116,139,144,154 Luti, R., 213,247 Lutz,H. J., 102,125,127,154 M MacArthur, R., 14,56 McIntosh, R. P., 68,98 McKell, C. M., 223,244

McMillan, C., 221,232, 234,244 McVean, D. N., 182,244 Madgwick,H. A. I., 146,148,149,155 Manley, G., 198,244 Marsden-Jones, E. M., 209,211,244 Maruyama, I., 146,154 Mather, K., 189, 194, 195, 196, 201, 206,244 Matzke, E. B., 211,246 Mauritz-Hansson, H., 107, 115, 122, 123,124,140,154 Mayer-Krapoll, H., 133,154 Mayr, E., 166, 206, 209, 210, 211, 244, 245 Melin, E., 125,155 Methley, W. J., 115,152 Metz, L. J . , 110,155 Meyer, J., 109, 133, 145, 148,152, 154 Mikole, P., 125,155 Miller, C. A., 4, 7,44, 56 Miller, R. B., 110, 118, 135, 139, 140, 155 Milner, H. W., 226,229,243,244,245 Mitchell, H. L., 142,155 Moller, C. M., 111, 132, 133, 141, 145, 148,149,155 Mook, J. H., 48,56 Mook, L. J., 48,56 Mooney, H. A., 234,235,236,245 Mori, K., 107, 114, 131, 136,155 Mork, E., 115, 117, 118, 122, 123, 124, 136,139,140,141,146,155 Morley, F. H. W., 239,245 Morris, R. F., 2, 5 , 14, 19,20,22, 24,27, 28, 30, 36, 37,41, 56 Mott, D. G., 56 Miiller, P. E., 102,155 N Navarro de Andrade, E., 127,155 New, J u n e K., 184,245 Nicholson, A. J . , 4, 6, 9, 13, 17, 18, 22, 26, 36, 39, 49, 50, 51, 52, 54, 56 Nobs, M. A,, 175,226,229,243,244,245 Norming, H., 107,115,137,154 Nye, P. H., 109,118,133,134,148,155 Nygren, A,, 203,245 Nykvist, N., 125,155

0

Ogawa, H., 138,155

252

AUTHOR INDEX

Ogino, K., 146,156 Ohmasa,M., 107,114,131,136,155 Olmsted, C. E., 230,231,245 Olsen, C., 142,155 Ovington, J. D., 102, 145, 146, 148, 149,151,155 Owen,T.H., 115,136,140,155

P

Paterson, S. S., 127, 149, 150.155,156 Pearsall, W. H., 102,155 Perina,V., 111, 118,156 Perrier, E. R., 223, 244 Perry, T. O., 230,245 Pierre, W. H., 121,153 Pieters, G. A., 224,247 Pimentel, D., 9, 14, 56 Pirie, N. W., 102,156 Polster, H., 148,156 Poore, M. E. D., 77, 88, 89, 90, 91, 93, 98,99 Potts, S. F., 145,156 Potts, S. M., 144,145,156 R Rahn, K., 182,240 Remezov, N. P., 117,156 Richards, 0. W., 2, 11, 31, 32, 33, 34, 35,40,41,56 Richards, P. W., 134,156 Richardson, S. D., 224,247 Rothacher, J. S., 144,145,156 Ryle, G. J. A,, 228, 245 S Sakasegawa, T., 146,156 Satoo, T., 146,147,148,154,156 Schwerdtfeger, F., 12,57 Scott, D. R. M., 116, 120, 122, 123,124, 125,156 Shidei, T., 138, 145, 146,156 Sims, I. H., 111,256 Sinskaia, E. N., 162, 171,245 Sjors, H., 115,156 Smith, A. H., 103,156 Smith, A. J. E., 182,245 Smith, F. E., 26,57 Smith, H. S., 44, 51,52,55 Smith, R. F., 4,57 Snaydon, R. W., 222,245 Solomon, M. E., 4, 6, 7, 10, 14, 17, 18, 22, 26, 42, 53, 57

Sonn, S. W., 117, 141,156 S~rensen, T., 75,76,99 Spitzer, C. H., Jr., 35, 55 Stebbins, G., 223,246 Stebbins, G. L., 182,194,200,202,203, 211,223,238,244,245,246 Stern, V. M., 4,57 Stoate, T.N., 109, 119, 135,156 Sviridova, L. K., 117, 120, 132, 139, 156 Symmons, P. M., 10,55 T Tadaki, Y., 138, 145,146,156 Tadmor, N., 223,241 Tamm, C. O., 103,108,142,157 Tarrant, R. F., 107,157 Thoday, J. M., 200, 206, 207, 209, 213, 246 Thompson, W. R., 6,36,57 Tinbergen, L., 48,50,51,52,53,57 Turesson, G., 159, 160, 161, 164, 167, 193,203,210,214,224,246 Turrill, W. B., 165, 167, 177, 209, 211, 238,244,246 Twinn, D. C., 144,152

U Ullyett, G. C., 44,57 Uphof, J. C. T., 202,246 Utida, S., 51,57 V Vaartaja, O., 181,232,233,234,246 Valentine, D. H., 165,237,246 van den Bosch, R., 4,57 van der Drift, J., 114, 120, 136, 138, 140,157 Varley, G. C., 6, 11, 12, 26, 28, 29, 36, 44, 50,51, 52,53,57,58 Vintrova, E., 111,118,156 Viro, P. J., 112, 118, 122, 123, 124, 136, 141,142,157 Vofite, A. D., 14,58 W Waddington, C. H., 215,246 Waid, J. S., 144,152 Walker, R. B., 179,220,246 Wallace, M. M. H., 7,14,15,16,58 Waloff, N., 2, 11, 31, 32, 34, 35, 40, 41, 56

253

A U T H O R INDEX Walter, H., 134,157 Walters, S. M., 96, 99 Wang Chi Wu, 230,245 Wareing, P., 228,229,246,247 Wasshk, E. C., 224,247 Watson, Patricia, J., 162, 165, 185, 242 Watt, K. E. F., 3, 18,58 Webb, D. A., 97,99 Weck, J., 102,157 Weidman, R. H., 181,247 Welch, J. R., 11, 55 White, D. P., 108,157 Whitehead, F. H., 213,247 Whittaker, R. H., 86, 88,99, 179,247 W i e d e m m , E., 130, 133,157 Wilkins, D. A., 169, 170, 172, 179, 219, 221,239,247 Will, G.M., 110,135,140,157

Williams, E. J., 52,54 Williams, W. T., 69, 77, 80, 82, 83, 98, 99 Witkamp, M., 114, 120, 136, 138, 140, 157 Woodman, M. J., 144,152 Wright, J. W., 180, 181, 247 Wright, T. W., 115, 133, 136,157 Wynne-Edwards, V. C., 14,58

Y

Ylgnen, J., 145,154 Yoda, K., 138,155 Younge, 0. R., 115,152 Z Zemljanickii, L. T., 145,157 Zohary, D., 223,246 Zon, R., 115,152

Subject Index A Aarburg, Switzerland, 145 Abies balsamea, 116 Abies sachalinensis, 114 Abundance (Insect),variation in, 8-10, 14,27

Acacia, spp., 121 Acacia angustissima,122 Acacia decurrens, 133 Acacia roemeriana, 122 Acarids, 51 Acerspp., 116, 117, 121 Acerplatanoides, 117, 123 Acer rubrum, 122,137, 144,230 Acer saccharum, 104, 105, 111, 115, 116, 118, 120, 121, 123, 124, 139, 140, 144

Acleris variana, 5 Achillea spp., 163, 179 Achillea borealis, 163, 178, 189,221 Achilleu borealis ssp. californica, 179 Achillea lanulosa, 163 AchiUea lanulosa ssp. typica, 229 Achillea millefolium, 163, 229 Achillea millefolium ssp. arenicola,229 Adaptation (Genecology) to climate, 227-237 edaphic, 219-222 to light intensity, 224-227 to soil moisture stress, 223-224 Adlisberg, Switzerland, 145 Aegopodium spp., 117 Aesculzcs spp., 121 Aesculus californica, 124 Aesculua glabra, 124 Africa, 10 Agglomerative methods (Statistics), 79-80

Agrostis-Festucagrassland, 185, 186 Agrostw hallii, 221 Agrostistenuis, 184, 185, 187, 188 Alaska Arctic Coastal Plain, 234 Alberta, 234

Alnw glutinosa, 113, 138, 182 Alnw hirsuta, 146 Alnw hirsuta v. sibirica, 146 Alnua incana, 142 Alpine Norwegian forests, 128 America, North, 230, 231,232,234 Amodal distribution (Vegetation), 7 1 Anagasta kuerknielli (Zell.),44 Andropogoneae, 203 Andropogon scoparius, 232 Angiosperm forests (Litter), 104, 107, 111, 115, 116, 119, 120, 121, 125, 132,136,139,148,151-152 Anisoptera laevw, 138 Ants, 35 Aphids, 35 Aphw fabae Scop., 35 Apple blossom thrips, 25 Arctic-Alpine forests, 108, 117, 119, 126,127,150 Argyrodendron spp., 109 Arizona, 202,231 Amhem, Netherlands, 114 Arrhenatherumspp., 186 As, Norway, 114,115 Assimilation, Genetic, 213-217 Atlantic seaboard of Europe, 182 Australia, 14, 108, 109, 119, 134, 141, 145,231 South-Waite Institute, Adelaide, 25 Austria, 139

B Baldwin effect (Genecology),215 Bavaria, Germany, 112, 113 Bay-breasted warbler, 48 Beet webwonn, 44 Behaviour patterns (Insect),14 Bethel, Minnesota, U.S.A., 145 Betula spp., 107, 112, 114, 115, 117, 118, 121, 136, 137, 139, 140, 141, 147 Betula alba, 145

254

SUBJECT INDEX

Betula ermanii, 146 Betula latifolia, 114 Betula lutea, 144 Betula maximowicziana, 146 Betulapapyrifera, 232,233 Betulaplatyphylla, 146 Betulapopulifolia., 111, 122, 123 Betulapubescens, 115, 122, 142 Betula verrucosa, 114, 122, 138, 142, 145,146,148,149 Biological species (Genecology),166 Biosystematics (Genecology),237 Birds, 1, 14, 48, 53 Blarina spp., 45,47,49 Blowflies, 13, 14, 18 Boletus spp., 103 Brachystegia spp., 109 Brandon, England, 146 Brazil, 127 British Columbia, 175 British Ecological Society, 209 British Isles, 177, 179, 182, 184, 185, 219 Brjansk, U.S.S.R., 116 Botanical Nomenclature, International Code, 165 Bothriochloa decipiens, 202 Bouteloua spp., 232 Bouteloua curtipendub, 23 1 Bromus carinatus, 202 Broom, 11,31,33 Broom beetle, 2, 31,32, 34, 35,40,41 Brunlanes, Norway, 114,115 Budworm, black-headed, 5,24,27-28 Budworm, spruce, 2,4,7,14,19,27-28, 30,36,44,48-49 Bupalw spp., 12, 13 Bupaluspiniarius (L.), 12, 16

C

California, 35, 108, 110, 116, 117, 128, 145, 146, 172, 173, 174, 175, 202, 211,229 Callosobruchusspp., 51 Canada, 2,111,128,139,146,230 Canberra, Australia, 231 Canonical analysis (Statistics), 86 Caragana microphylkz, 117 Carm spp., 117 Carex caryophyllea, 186 Carolinas, 108, 110, 111, 126,236

255

Carpinus spp., 118 Carpinus (betulus),113 Caryaspp., 111, 137 Carya cordiformis, 115 Carya ovata, 123 Carya (Hicoria)ovatu, 142 Castaneu spp., 121 Castanea crenata, 114 Castanea sativa, 122 Castanea vulgaris, 122 Catalpa speciosa, 123 Categories (Genecology),164-167 Caterpillars, oak-feeding, 11 Caterpillars, pine, 13 Ceanothus spp., 175 Ceanothus cuneatus, 176 Ceanothus gloriosus, 176 Cellulose in wood wastes, use of, 103 Celtis spp., 121 Celtis occidentalis, 124 Celtis reticulata, 124 Cerastes, 176 Cerastium vulgatum, 186 CeratopetaZumspp., 110 C h a m c y p a r i s obtusa, 114, 131, 136, 137 Chelonus texanus Cress., 44 Cheshire, U.K., 115 Cheyletus eruditus (Schr.), 51 Chicago, 231 Cladoniaspp., 117 Cladraatislutea, 121,124 Classification (Vegetation), 72-77 Classification, inadequacies of, 76 Classification, methods of, 77-81, 89 Classification, systems, 72-73 Climate variability, 13 Climatic zones, 108, 149 Cline, definition of, 163 Coenospecies (Genecology),165 Collembola, 14 Colombia, 108, 109, 133, 134 Competition (Insect), 6-9 Component analysis (Statistics), 74-76 Computer analysis (Statistics), 9&97 Computer design, 96 Computer, Elliott 803, 96 Computer facilities, 63 Conditioning (Genecology),2 13-2 17 Congo, 108,109,145,148,149 Connecticut, U.S.A., 116

256

SUBJECT INDEX

Control (Insect) biological, 2 chemical, 2 natural, 1-4,6,8, 1 6 4 2 Cool temperate forests, 108, 111-1 17, 119,126,127,137,148,150-152 C m u a spp., 137 CornusJEorida,123,144 Cornwall, 178 Corylusspp., 115, 116,117 Cosmia trapezina, 11 Cover abundance scales (Vegetation), 68 Cratichneumon spp., 29 Cryptonzeriajaponiw, 114,136,137 Cryptua inornatua Pratt., 44 C u p e s m luaitanica, 109 Cybernetic principle, 5 Cyzenis spp., 29 Czechoslovakia, 111 D Dactylis glomerata, 223,224 ssp.judaica, 223,224 ssp. luaitanica, 223,224 Dactylis coccinea, 117 Dactylisfuchsii, 178 Dactylis hebridenab, 178 Dactylorch@incarnata, 177, 178,212 Dactylorchis maqulata, 178 DactyZorchia okellyi, 178 Dactylorchis p k h e l l a , 177 Dahlbonzinv fuacipennis (Zett.), 44 Dakota, 231 Dalarna, Sweden, 115 Deer-mice, 42,45 Dendrolimus spp., 13 Dendrolimuapini L., 12 Denmark, 111,136,145 Density (Insect) dependence, 4-9, 14, 17, 19-20, 22-26 independence, 9,17,20,23-24,26 relationships, 26-27 Derkul steppe, U.S.S.R., 117 Dkhanthium arktatum, 203 Didinium spp., 51 Differentiation (Genecology) ecotypic, 193-237 generabtions, 187-1 89 flustrative examples, 173-187

Diospyros spp., 109, 121 Diospyros texana, 123 Diospyros virginiam, 123 Dipterocarpus spp., 109 Dipterocarpus baudii, 109 Discriminate analysis (Statistics), 77 Distribution Maps Scheme (Vegetation), 96 Don River valley forest, Glendon Hall, Toronto, Canada, 104,105,106 Dorset, Ontario, Canada, 146 Drosophila spp., 207,215 Drosophila melanogaster, 206 Dryobalanops aromatica, 109 Durham, North Carolina, 236 Dwellingup, Australia, 109

E East Anglian fens, 178 Eberswalde, Germany, 113 Ecoclines, 162-164 definition of, 163 Ecospecies, 165, 167 Ecotypes, 162-164 definition of, 163 Ecotypic differentiation evolutionary aspects, 193-217 genetic basis, 189-193 physiological aspects, 217-237 Edaphic adaptation, 219-222 Efficiency (Statistics), definition of, 63 Eidsberg, Norway, 114 Elliott 803 Computer, 96 Ellopia spp., 11 Emigration (Insect),6 Encarsia spp., 51 Encarsia formosa Gahan., 50 England, 2, 11, 136, 146, 152 Equatorial forests, 108, 109, 126, 127, 134,148,150-152 Erannis defoliaria, 11 Erd, Hungary, 113,114 Erzgebirge, Germany, 113 Eschscholzia di$ornica, 187 Estimation methods (Insect), 41-42 Eucalyptusspp., 118,119, 121, 127,183 Eucalyptus acmenioides, 110 Eucalyptus camaldulelzsis, 110 Eucazyptus coccifera, 183 Eucalyptus diversicolor, 109, 110, 134-6 Eucalyptus gigantea, 145, 183

SUBJECT INDEX

Eucalyptua gunnii, 183 Eucalyptus marginata, 109,134,135 Eucalyptus pilularis, 110 Eucalyptus regnans, 109, 120, 134, 141 Eucalyptus saligna, 109 Eucosma isertana, 11 Eupithecia spp., 11 Europe, 108,128,182 Evo, Finland, 112 Evolution and variation in plants, 194, 238 “Evolution of Genetic Systems”, 194 Experimental taxonomy (Genecology), 237

F Factor analysis (Statistics), 74-76 Fagaceae, 121 Fagraeafragrans, 109 Fugusspp., 118,121,147 Fagua crenuta, 114 Fqusgrand$’olia, 111,115,123,144 Fagusdvatica, 107, 111, 112, 113, 115, 123, 127, 129, 130, 132, 133, 139, 141,142,145,148 Fall webworm, 20 Fennoscandia, 121,122-124,127 Fertilizers (Forests), 103 Fmtuca ovina, 179,180, 186,219,222 Festuca rubra, 186 Ficus spp., 110,143 Ficus elastica, 143 Finland, 108, 111, 121, 125, 137, 141, 142,145 Fish, 14 Fisheries research, 3 Flea, lucerne, 14, 15, 16 Florida, 108, 110,230 Floristic composition, 70 Floristic recording, 69 Flour beetle, 18 Forests climatic zones, 108, 149 conifer, 24, 103 deciduous, 126-127 Don River valley, 104,105,106 evergreen, 126-127 fertilizers, 103 fir, 2 Japanese, 107 leaf consumption by animals, 144

257

litter production, 101-167 magnitude of leaf crops, 144-147 pine, 13 rain, 109,134 Scots pine, 11, 12 spruce, 2 treatment of, 131-133 Fraxinusspp., 111, 117,121,144,147 Fraxinus americana, 123 Fraxinus excelsior, 111, 123, 124, 133, 144,145,148 Fraxinus mandshurica, 145 Fraxinus penwylvanica, 104,105 Fraxinus quadrangulata, 123 Fungi, edible, 102-103 G Gainesville,Florida, 230 Galiumpumilum, 186,209 Garrulus glandarius, 48 Genecology, 159-247 basic proportions, 160-162 plasticity, 2 13-2 17 scope and concepts, 159-193 surveys, 239 technique, 168-172 Genetic changes (Insect),9 Geometrids, 12 Germany, 102,112,126,130,136,139 Letzlinger Heide forests, 12 Geum rivale, 177 Geum urbanum, 177 Ghana, 108,109,133,134,148 Cilia achilleaefolia, 187 Gilia capitata, 175,179, 187 ssp. abrotanqolia, 175 ssp. capitata, 175,179 ssp. chamissonis, 175 ssp. pedemontanu, 175 ssp. staminea, 175 ssp. tomentosa, 175 Glendon Hall, Toronto, Canada, 104, 105,106 Glossina swynnertoni Aust., 11 Godollo, Hungary, 113,114 Gradient analysis (Statistics), 73,86 Green River, New Brunswick, 19 Green River Project, 2 Grue, Norway, 114 Gulf Stream, effect of, 128

258

SUBJECT INDEX

Gymnosperm forests, 103, 116, 119,

Juniperus utahensis, 122 Juniperus virginiana, 121

H Harvesting methods (Litter), 103, 107 Heath litter, 107 Hebrides, 178 Heterogeneity, 71,89 degree of, 73,76 Heterozygosity (Genecology),195,204 Hieracium umbellatum, 210 Hierarchical classification (Statistics),

K Kade, Ghana, 109 Kallo, Hungary, 114 Kamyshin, U.S.S.R., 145 Kansas, U.S.A., 231 Key factor analysis (Statistics), 27-30 artificial reduction of, 35-37 exclusion of, 35-37 Kiev, U.S.S.R., 117 Kola Peninsula of U.S.S.R., 108, 117 Kosciusko plateau, 176 Knapweed gall-fly, 36 Kunadacs, Hungary, 113,114

121, 125,132,139,148,151-152

78,79

Hirkjolen, Norway, 117,146 Hokkaido, Japan, 146 Homogeneity, 7 1 Homozygosity (Genecology),195,204 Housefly, 44,5 1 Hungary, 108,113 Hylocmium spp., 117 Hyloicwr spp., 13 Hyloicua (or Sphinx) pinaatri (L.), 12 Hypbntria cunea (Dru.), 20 Hypothesis-generation (Statistics), 6062,74-75,81

Hyytiala, Finland, 112

I Ichneumonid parasites (Insect),5 Idaho, U.S.A., 181 Illinois, U.S.A., 146 Imperial College Field Station, 31 Importance indices and value (Statistics), 68 Indiana, U.S.A., 146 International Code of Botanical Nomenclature, 165 Inverse analysis (Statistics), 81-82 Ireland, 178 Isolation (Genecology),204-210 Israel, 223 Ituri, Congo, 109 Iwate, Japan, 146 J Japan, 114,145,146,149 Japanese forests, 107, 108, 136 Jays, 48 Juglam nigra, 121 Juniperwr spp., 121 Juniperus pinchotii, 122

L Lamarckian response (Genecology),215 L a r k spp., 121 Larixdecidua, 110,114,123,135,146 Larix europa, 148 Larix kaempferi (leptolepis),114 Larix leptolepis, 114, 120, 122, 146 Larix sibirica, 114, 120, 122 Lasiocampids, 12 Lasiua niger, 35 Layia spp., 189 Layiaphtyglossa, 172,211 Leaves, standing crop, 142-147 Lepidoptera, 11, 12 Letzlinger Heide forests, Germany, 12 Life cycle (Insect), 13-14 Life tables (Insect),13 analysis, 30-35 Lignin, use of, 103 Lima, 183 Limitation (Insect),4 Linnean species, 161 Liquidambar styracijua, 123 Liriodendron tulipiyera, 111, 121, 123 Litter altitude influence, 129-130 angiosperm, 104, 107, 111, 115, 116, 119, 120, 121, 125, 132, 136, 139, 148,151-152 annual variation, 139-141 ash content, 122-124 climatic influence, 127-129,149 components, 118 factors affecting litter fall, 125-142

SUBJECT I N D E X

259

gymnosperm, 103, 116,119,121, 125, Mimulus cardinalis, 226,229 132,139,148,151-152 Minneapolis, U.S.A., 142 Minnesota,U.S.A., 108,115,145 harvesting methods, 103,107 University, Botany Department, 143 heath and moss, 107 Mirids, 33 index to net production, 147-152 Missouri,U.S.A., 108, 115,132 mineral material, 121-125 Mites, 2,35 organic material, 102,125 phytophagous, 44,52 production, 101-157 predatory, 44,52 protein source, 102 removal, effect of, 133 Model-making (Statistics) derivative seasonal variation, 133-139 and primary, 61 soil fertility influence, 130-131 Molinia caerulea, 185 Monothetic methods (Statistics), 79-80 soil moisture influence, 131 Monotypic origin (Genecology), 210understorey, 119-120 213 utilization, 102-103 Montreal, Canada, 111 Locusts, 14 red, 10 Mormoniella spp., 51 Mormoniella vitripennis (Walk.), 44 Lolium spp., 228 Morocco, 110 Lolium perenne, 185 Morphological species (Genecology), Lozostege sticticalis (L.),44 166 Lucerne flea, 14,15,16 Mortality (Insect) Lucilia spp., 19 density-dependent, 8,18,36-39 Lucilia cuprina Wied., 13,18 density-independent, 37-39 Lund, Sweden, 115 effects, 35 Lymhachia vulgaris, 224,226 factors, 3 percentage, 4 , 2 6 2 7 M rates, 36 Macrolobium spp., 109 Madia spp., 189 regularity, 39-41 tables, 36 Magnoliamacrophylla, 121, 123 variations, 29 Makibori, Japan, 146 winter disappearance, 29 Malaya, 108, 109, 133, 138, 139 secondary forest, 134 Morusspp., 121 undisturbed Dipterocarp forest, 134 Morus macrophylla, 124 Mammals, 1, 14 M o m s rubra, 124 Marschallsruhe, Germany, 113 Moscow, U.S.S.R., 116 Moss litter, 107 Massachusetts, U.S.A., 145 Multivariate methods (Statistics), 66 Mather, 191, 193 Mwanga cecropioidea, 109, 145 Matra, Hungary, 114 Mechanical raking (Litter), 103 N Medicine Bow Mountains, Wyoming, Nebraska, U.S.A., 231 234 Mediterranean Neodiprion sertqer (Geoff.),42 climate, 223,230 Netherlands, 2, 11, 13, 114, 136, 137, seaboard of Europe, 182 138,140 Melampyrum pratense, 182 New Brunswick, 2,5,24 Melandrium spp., 212 Green River, 19 Melandrium album, 177,212 New England States, U.S.A., 126 Melandrium rubrum, 177,212 New Hampshire, 137,180,181 Mexico, 231 New Mexico, 231,232 Michigan, U.S.A., 103 New York, U.S.A., 115

260

SUBJECT INDEX

Peromyscusspp., 45,47 Peru, 183 Peterborough, England, 146 Phmelia spp., 220 Phaceliu catifornica, 220 Phalaris tuberosa, 230 Phleum spp., 228 Phytodecta spp., 31,32,33 Phytodecta olivacea, 11, 31, 41 Phytophagous animals, 8 Phytosociological analysis, 66 comparison of concepts, 87-89,95 nature of data, 64-69 0 Phytosociology Oils in Gymnosperm litter, 103 definition of, 61-62 Oklahoma, U.S.A., 231,232 statistical properties, 70-72 Ontario, Canada, 45, 146 Picea spp., 118,-121, 142 Operophtera brumata (L.), 11,28 Piceadies, 107,111,112,113,114, 115, Optimization (Vegetation), 63 116, 117, 120, 122, 127, 129, 130, ‘‘Originof Species”, 160 133, 136, 137, 139, 140, 141, 142, Ordination (Statistics), 72, 75 146,148,149 methods, 76,89 Picea glehnii, 114 schemes, 72,73 Piceajezoensis, 114 Oscillation (Insect), 7-8, 14 Picea pungena, 142 parasite-host, 24-25,28,49-53 Picearubens, 116,122 Oxalis acetosella, 116, 117 Picea sitchensis, 136, 137, 140 Oxydendrum arboreum, 137 Pine hawk, 12 Oxyria digym, 234,235,236 looper, 16 P sawfly, 42,44-47 Pacific, 174 Pinusspp., 111,118,121,142,147,233 Paniceae, 203 Pinua b a n W n a , 115,122,147,205 Panolis spp., 11, 13 Pinus canariensis, 110 Panolis griseovariegata (Goeze), 12 Pinus caribaea, 110, 122 Paramecium spp., 51 Pinus contorta, 117, 147 Parasites (Insect),3-4, 6-7,24,27, 37 Pinusdenai$ora, 114,131,136,137, 146 density relationships, 42-49 Pinusechinata, 110, 111, 115, 132, 136, fecundity, 4 S 5 0 137 functional and numerical responses, Pinus nigra, 110, 111, 135, 139, 140, 42-44 146 host cycle, 24 Pinuspdustris, 110, 122, 132 host-interaction, 24, 49-53 Pinus ponderosa, 110, 116, 132, 146, hymenopterous, 44,50,51 147,174, 181 ichneumonoid, 5 Pinuaradiata, 110, 135 larval, 29 Pinus resinosa, 108, 115, 116, 118, 122, tachinid, 5 139 Partition correlation analysis (Statis- Pinusrigida, 121,122 tics), 69 Pinussilvestris, 107, 111, 112, 113, 115, Parus spp., 48,53 116, 117, 120, 122, 127, 129, 130, Pathogens, 6 136, 137, 139, 141, 142, 146, 148, Pennsylvania, U.S.A., 233 180,181

New York State, 230 flew Zealand, 108, 119, 132, 135, 139, 140 Noctuids, 12 Nodal analysis (Statistics), 83-85, 94 Nodebo, Denmark, 111 N d a c r i s septemfasciata (Serville), 10 Northwest Territories, 233 Norway, 108,114,117,136,139,146 Norwegian Alpine forest, 128 Nothofagus truncata, 110, 118, 135, 139, 140 Nyssa sylvatica, 137

261

SUBJECT INDEX

Pinus strobus, 104, 105, 111, 115, 116, 120,121,122

Pinus taeda, 110 Pinus thunbergii, 114 Pinus virginiana, 110,121 Plantago spp., 162 Plantago coronopa, 182 Plantago lanceolata, 185 Plantago mucrorhiza, 182 Plantago major, 209 Plantago w r i t i m a , 162, 182,206 Plasticity (Genecology),213-217 Platanus occidentalis, 123 Poa trivialis, 186 Polytopic origin (Genecology),210-213 Polythetic methods (Statistics), 79 Population (Insect), see also Density, Parasites, Predators definition of, 8 dynamics, 1-3, 6, 16,24 Populus spp., 107, 111, 114, 117, 121, 139,147

Populus alba, 113 Populus davidiana, 146,148, 149 Popuha grandidentata, 111,122,146 P o p u l w nigra, 114 Populus tremula, 115, 117, 120, 123, 132,138,142

Populua tremubides, 111, 146, 148 Potentilla glandulosa, 163, 173, 174, 175,178,188,190,191,192,193,212

ssp. hanseni, 163,173,174,212 ssp. nevadensis, 163, 173, 174, 188, 191,192,193,212

ssp. rejlexa, 163, 173, 176, 188, 191, 192,212

ssp.typica, 163,173,174,188,191,192 Predators (Insect) density relationships, 42-40 functional and numerical responses, 42,4547

functional response laboratory model, 42-43 mammalian, 4 P 4 7 other vertebrates, 47-49 pupal, 28-30 Presence-or-absence system (Vegetation), 68-69 Primary analysis (Statistics), 65, 89 Principal component analyses (Statistics), 74

P r u n e h vulgaris, 182 Prunus spp., 121 PrUnuS pennsylvanica, 123 Prunus serotim, 105, 123 Pseudotauga rnenzieeii, 110, 116, 133, 135,139,141,148

Pulse beetles, 51, 52 Pyrus coronaria, 123

Q

Qualitative patterns (Vegetation), 69 Quantitative methods (Vegetation), 92, 93

Quantitative patterns (Vegetation), 69 Quebec, Canada, 105 Quercus spp., 110, 111, 114, 116, 117, 118, 121, 136, 137, 140, 142, 144, 145,147 Quercus alba, 104, 105, 110, 111, 116, 122,137,138,139,144,145 Quercus borealis, 104, 105, 111, 122, 148 Quercus breviloba, 122 Quercus cerris, 114 Quercus chrysolepis, 145 Quercus coccinea, 110, 137 Quercus douglaaii, 123 Quercus ellipsoidal&. 145 Quercusfalcata, 137 Quercus ilicifolia, 2 11 Quercus kelbggii, 116 Quercus mariladica, 2 11 Quercua mongolica v. grosseserrata, 145 Quercus montana, 137 Quercus palustris, 122 Quercus petraea, 144 Quercus pinus, 110 Quercua robur, 114, 120, 123, 138, 145, 148 Quercus rubra, 114, 115, 116 Quercus serrata, 114 Quercua sessilijlora, 114 Quercus stellata, 137, 138 Quercusvelutina, 110, 116, 123, 137, 139 Quercus virginiana, 123

R Racial identity, Retention of (Genecology), 210-213 Ranunculus dissectifolius, 176-177 Ranunculus Zappaceus, 176,212 Ranunculus millanii, 176, 177

262

SUBJECT INDEX

Reconnaisance method (Statistics), 92 Reduction (Statistics), 60, 72, 74 Regulation (Insect) artificial, 17 assessment, 16 degrees of, 2,13,25 density-dependent, 6,&9 detection, 16 modification of, 8-9 processes of, 2 Relationships (Vegetation) plantjplant, 64-65 plantlsite, 64, 66, 88 sitelhabitat, 64, 85, 88 plantlhabitat, 65, 85, 88 plantlsitelhabitat, 65, 85, 88 site groups, 85 Reproduction (Insect), 6 capacity, 13 rates of, 3-4, 9, 18 Reproductive systems, versatile (Genecology), 201-204 Resins in gymnosperm litter, 103 Reticulate classification (Statistics), 78 Retsag, Hungary, 113, 114 Ricinus communiS, 183 Ringsaker, Norway, 114 Riverside, California, 35 Robinia spp., 121 Robinia pseudoacacia, 113, 120, 123, 124,146 Rotorua, New Zealand, 110 Rottboellia exalta, 203 Roxburghshire, U.K., 115 Rubua nitidioides, 203

S

Salix cap-rea, 142 Salix gracilktyla, 146 Salix v u l p ~ n a146 , Sambucua spp., 116 San Dimas forest, California, U.S.A., 145 San Francisco Bay, California, 175 San Joaquin valley, California, 229 Sarothamnus scoparius, 11, 31 Saskatchewan, 209,231 Sassafras dbidum, 120 Saugus, Mass., U.S.A., 145 Scale insects, 35 Scandinavia, 182

Scania, 214 Schizachyrium scoparium, 232 Schizomeria spp., 110 Schweitzer’s reagent, 125 Scotland, 136, 178, 185 Secondary forest, Malaya, 134,138 Selection (Genecology) modes of, 196-201 natural, 160 Shinyanga District of Tanganyika, 11 Shorea curtkii, 138 Shwea laevis, 138 Shorea leprosula, 109,134 Shrews, 42,45 Sierra Nevada, U.S.A., 108, 163, 173, 174,189,229,230 Siknemaritima, 167,177,211 Silenevulgaris, 167,177,211 Sminthurw, viridis (L.), 14 solidago spp., 214, a27 Solidago virgaurea, 214, 224, 225, 226, 227 Swbus aucuparia, 122,146 Sorer spp., 45,47 Sor0, Denmark, 111 Spatial distribution (Vegetation), 6768 Spergulu arvensw, 184 Spiess, Germany, 113 Stanford, 191, 193 Stanislaus National Forest, California, U.S.A., 146 Statistics in Phytosociology, 59-99 analysis methods, 70-87 definition of, 60-61 function of, 62-64 future of, 95-97 hypothesis-generation, 60-62,7675 methods, 66 reduction, 60, 72,74 sequential methods, 91 structuring techniques, 72-80 Stipa spp., 209 Stipa leucotricha, 202 Stockholm, Sweden, 115, 180, 181 Storelvdal, Norway, 114, 115 Streptanthus spp., 221 Subdivisive methods (Statistics), 79-80 Successive appreciation method, 91-92 SucciSa vatensis, 214 Succiaa pratensk f. nuna ,214

SUBJECT INDEX Sutherland, 178 Sweden, 115,142,145,210 Switzerland, 115, 145,146 T Tachinid parasites, 5 Tamarindus indim,143 Tanganyika, Shinyanga District, 11 Tasmania, 183 Taxonomy (Vegetation) categories, 89 experimental, 237 evolutionary, 237 numerical, 65, 77 Tennessee, U.S.A., 108, 110, 111, 135, 136,138,145 Tetranychus t e h r i w , 44 Texas, U.S.A., 231 Thailand, N.W., 138 Tharandt, Germany, 113 Thera spp., 11 Thrips, 26 apple blossom, 25 Thrips imzginh Bagndl, 25 Thuja occidentdw, 116,178 Thujopsk doldwata, 114 Tilth spp., 117 Ti& merimma, 115,121,123 Tiliata cordata, 113 Timberline, 191,193 Tits, 48,53 Topoclines, 162 definition of, 163 Topotype, deiinition of, 163 Toronto, Canada, 111,118 Botany Department, 143 University Library, 107 Tortricidae, 5 Transposed method (Statistics), 80 Trialeurodes spp., 51 Trialeurodes vaporarimm (Westw.), 50 Tribolium cmfuaum Duv., 18 TriStania cOnfeTta, 110 Tsetse fly, 11 T m g a canadensis, 116 Turessonian genecology, 160 l’yphlodromus spp., 44

U Ugod, Hungary, 113,114 U l m w spp., 114,121,144

263

Ulmua americana, 123 U l m w glabra, 120 Ulmuapa&$lora, 139 U h u a ambra, 124 Unimodal distribution (Vegetation), 71 University of Toronto Library, 107 Botany Department, 143 U.S.A., 108, 115, 117, 118, 139, 142, 145,146 north, 119,122-124,125,128,129 south, 119 U.S.S.R., 108, 116, 117, 118, 126, 139, 145

V Vacciniummyrtillus, 116,117 Vaecinium vitie-idaea,116,117 Vacuum collection (Litter), 103 Variation (Genecology) adaptive, 161 and evolution in plants, 194,238 ecologically-correlated, 160 genetical, 160 infraspecific, 160 inter-population, 171 origin, storage and release of, 193-196 phenotypic inter-population, 168 spatial, 160 sporadic, 161 Vegetation chemical composition, 66 clrtssification, 72-77 classificationmethods, 77-81 community, 61 distinction between “vegetation” and “vegetation-habitat complex”, 93 general purpose survey, 67 geometrical form, 66 life-formrecords, 66 nature of, 64-66 nature of measurements, 68-69 patterns, 95 phytosociologicalanalysis, 66 primary analysis, 65,89 relationships, 6&66, 85-87 sampling methods, 67 spatial distribution, 67 studies, 62 tree-regeneration, 66

264

S U B J E C T INDEX

Vegetational continuum, 72 Veldre, Norway, 115 VelikijeLuki,U.S.S.R., 116,117 Vertebrates, 14,47-49 Vesijeko, Finland, 112 Victoria, Australia, 109, 134 Vilppula, Finland, 112 Voronezh,U.S.S.R., 116, I17 W Waite Institute, Adelaide, South Australia, 25 Wales, U.K., 115, 136, 140, 184, 188 Warm temperate forests, 108, 109-1 11, 119,126,127,134,150 ' Washingtonstate, U.S.A., 116,141 Weather, influence of (Insects), 9-10, 14-15,19,22,25-26

Weighting, internal and external (shtistics), 79 Wellington, New Zealand, 110 Whitefly, 50 Winter moth, 29,53 Winterthur, Switzerland, 145 Wisconsin, U.S.A., 116,178 Wyoming, U.S.A., 234 Wytham, 29 Wytham Wood near Oxford, 11, 28

Y

Yangambi, Congo, 110,133,145 Z Zavetnoe, U.S.S.R.,145 ~~~b S e T T d , 114 Zurich, Switzerland, 115