Behavior and Its Neural Control in Gastropod Molluscs

Behavior & Its Neural Control in Gastropod Molluscs

This page intentionally left blank

Behavior & Its Neural Control in Gastropod Molluscs

RONALD CHASE

OXFORD UNIVERSITY PRESS

2002

OXFORD UNIVERSITY PRESS Oxford New York Auckland Bangkok Buenos Aires Cape Town Chennai Dar es Salaam Delhi Florence Hong Kong Istanbul Karachi Kolkata Kuala Lumpur Madrid Melbourne Mexico City Mumbai Nairobi Sao Paulo Shanghai Singapore Taipei Tokyo Toronto and an associated company in Berlin

Copyright © 2002 by Oxford University Press, Inc. Published by Oxford University Press, Inc 198 Madison Avenue, New York, New York 10016 www.oup.com Oxford is a registered trademark of Oxford University Press All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission of Oxford University Press. Library of Congress Cataloging-in-Publication Data Chase, Ronald Behavior and its neural control in gastropod molluscs/Ronald Chase. p. cm. Includes bibliographical references. ISBN 0-19-511314-4 1. Gastropoda—Nervous system. 2. Gastropoda—Behavior. I. Title. QL430.4.C455—2002 573.8'143—dc21 2001036788

9 8 7 6 5 4 3 2 1 Printed in the United States of America on acid free paper

I dedicate this book to the memory of Jim Chase, from whom I learned a love of scholarship.

This page intentionally left blank

Preface This book is about the adaptive behaviors and the efficient brains of snails and slugs. Unfortunately, too many people think that the behavior of slugs is only sluggish, and that snails have no brains at all. In fact, gastropod molluscs are as interesting as any other animal, and we have a detailed understanding of how their brains work. During the past quarter century, there has been a tremendous increase in our knowledge about gastropods, their behavior and their neurobiology. This book is intended to review and celebrate that accomplishment. It is the possibility of gaining detailed knowledge of both behavior and the nervous system that makes gastropod molluscs such attractive animals for biological investigation. Each subject, behavior and the nervous system, can be advantageously studied in gastropods. The behavior of gastropods, for example, is intermediate in complexity compared to other animals. It is neither as complex as the behavior of mammals and insects, nor as simple as the behavior of worms and echinoderms. Also, it has the added feature of being slow, which can sometimes be convenient for the scientist. All things considered, the behavior of gastropods is sufficiently complex to be interesting but not so complex as to be horribly difficult to describe. Similarly, the gastropod central nervous system contains more neurons than does the nervous system of worms and rotifers, for example, but it has fewer neurons than the brains of crustaceans and mammals. The precise number of neurons present in any central nervous system is difficult to determine, but generally speaking, estimates for the three major gastropod groups are in the range of 50,000-200,000 neurons for prosobranch species, 4,000-20,000 neurons for opisthobranch species, and 5,000-100,000 for pulmonate species. A complicating factor for investigators is the presence of a peripheral nervous system, which is difficult to access, but this problem is offset by the opportunity to work with central neurons of exceptionally large size. The study of molluscan behavior from a neurobiological perspective has many early precedents but only a true beginning around 1965. In the writings of Artistotle and Pliny, we see evidence that the ancients paid attention to the most obvious behaviors, such as hibernation in land snails and inking in Aplysia. Illustrations of snail shells commonly appeared in books about natural history published in the 17th century. The first illustrated guide to shells, the Historia Conchyliorum, was published by Martin Lister in England

viii

Preface

in 1685. Interest in the animals that lived inside the shells grew significantly after publication of Linnaeus's great work of classification in 1758. However, it was not until the end of the 19th century that anything was known about the nervous system and, surprisingly, at that time the study of molluscan behavior had hardly progressed since the days of Artistotle. The state of knowledge of molluscan behavior late in the 19th century can be illustrated by quoting a passage from Charles Darwin's The Descent of Man and Selection in Relation to Sex. Although Darwin himself was a reliable observer, and not usually taken in by imaginary animal stories, he accepted as true an account of snail behavior communicated to him by an acquaintance, one Mr. Lonsdale. The latter individual had placed two snails, Helix pomatia, into a garden that was evidently ill-provided with food. One of the snails was strong and healthy whereas the other snail was weak. From evidence of a slime trail, Mr. Lonsdale induced that the strong snail at one point had climbed over a wall and into an adjacent garden. Now quoting Darwin: "Mr. Lonsdale concluded that it [the healthy snail] had deserted its sickly mate; but after an absence of twenty-four hours it returned, and apparently communicated the result of its successful exploration, for both then started along the same track and disappeared over the wall" (Darwin, 1871, p. 325)

What is remarkable here is that Darwin accepted not only the observations but also the interpretation, namely that the healthy snail returned to help its mate because of a sense of "permanent attachment." Today, we know that there is no pair bonding in snails, and no ability to communicate in the indicated manner, although snails do commonly follow slime trails. Eleven years after Darwin published this story, George Romanes recounted it in his own book, Animal Intelligence. Although Romanes wished for an experimental replication, he too was ready to accept the facts and interpretations as given by Lonsdale and Darwin. Another interesting account, in this case obtained third-hand, was reported by the eminent American malacologist William Healy Dall in an article entitled "Intelligence in a snail" (1881). The informant's sister had a pet snail that was remarkable not only for the fact that it could hear voices, but also because it could distinguish its owner's voice from the voices of other persons. It seems that the snail would crawl to the owner when spoken to, but it withdrew into its shell when spoken to by anyone else. After relating what the woman had told him, Dall writes that the facts may be "surprising" but he had no doubts of their "substantial accuracy." While some zoologists were concerned with the mental capacities of snails, others were taking advantage of improved microscopic techniques to learn about the nervous system. One of the first and most significant discoveries was the presence of giant nerve cells. This was reported at least as early as 1863, when Buchholz reported them in freshwater snails (see Bullock, 1965). The anatomical papers that began to appear at this time and that continued to be produced in great numbers for another half century are full of detailed descriptions and beautiful illustrations. Although few tools were yet at

Preface

ix

hand with which to study function—mostly scalpels for producing lesions—the knowledge of neuroanatomy gained from these works laid the foundation for modern neurophysiological studies. However, much of it remained unknown or inaccessible to English-speaking scientists until the publication in 1965 of the monumental two-volume treatise on invertebrate nervous systems written by Theodore Bullock and G. Adrian Horridge. Bullock's chapter on gastropod molluscs alone contains citations to more than 750 papers, many of which are written in languages other than English. Still, the frontier of physiological neurobiology was yet to be truly breached. The years just preceding and just following 1960 saw the establishment of several research laboratories that were to play key roles in the subsequent blossoming of molluscan neurobiology. In each case, individual scientists were at the center of the new laboratories. Professor K.S. Koshtoyants at Moscow State University trained three students who established laboratories in Eastern Europe: Dimitri Sakharov in Moscow, and Janos Salanki and Katalin Rozsa in Tihany, Hungary. Meanwhile, other laboratories devoted to the study of gastropod neurobiology were set up by Professor J. Lever in Amsterdam, by Gerald Kerkut in Southampton, England, and by Ladislav Tauc in Paris. Transplantation of these efforts to North America occurred following cross-fertilizations in Tauc's laboratory. When Eric Kandel visited Tauc's laboratory in 1961, he was a psychiatric resident and research associate at the Massachusetts Mental Health Center in Boston. Having already experienced the frustration of neurophysiological research on the mammalian brain, he was open to new opportunities. He learned about the giant nerve cells of gastropod molluscs from a lecture given by Angelique Arvanitaki at the National Institutes of Health. As a biology student, working at a Mediterranean field station in the early 1940s, Arvanitaki had discovered the potential of Aplysia's giant nerve cells for neurophysiological research. She shared her knowledge and enthusiasm for these cells with Tauc, who then passed it on to Kandel when he arrived for collaborative work. After Kandel's return to Boston, he established his own laboratory devoted to the study of Aplysia. Mixing talk therapy upstairs with slug dissections downstairs, Kandel quickly became acquainted with many identifiable neurons in the abdominal ganglion, and he began to study their roles in the control of behavior. Meanwhile, another American, Felix Strumwasser, had also heard Arvanitaki's lecture and he had also visited Tauc's Parisian laboratory. Thus, with Kandel established in Boston, and Strumwasser at the California Institute of Technology, an enthusiasm for sea slug neurobiology began to sweep across America in the early 1970s. Whereas Kandel's research was specifically directed towards discovering the mechanisms of learning and memory, with its ultimate application to psychiatry, his work led many scientists to realize that gastropod molluscs need to be understood in their own biological context and for their own inherent interest. The accelerating attention paid to gastropods as subjects for neurobiology during the decades of the 1960s and 1970s was part of a broader effort to

x

Preface

discover and exploit animal models that were simpler to understand and more convenient to work with than mammals. Although insects, lower vertebrates and other groups were also being investigated for these purposes, the large size of molluscan neurons was especially attractive to physiologists who required stable intracellular recordings for biophysical work. Others were attracted to giant nerve cells because they allowed for replicable experiments on the role of individual neurons in the control of behavior. Furthermore, and importantly, the rapid success of Kandel and his colleagues in accounting for certain simple behaviors in terms of identified cells, such as in defensive withdrawal reflexes and ink ejection, encouraged a similar approach in a variety of gastropod species besides Aplysia that also have large neurons. Kandel summarized the results obtained by his group and others in two remarkable books published in 1976 and 1979. Although the first book is broader in its approach than the second, both books essentially review progress in understanding the neural control of behavior in Aplysia. With the exception of a two-volume collection of reviews edited by Dennis Willows (see Mpitsos and Lukowiak, 1985; Dorsett, 1986), there has been no comprehensive review of gastropod neurobiology since Kandel's works. In the present book, I include material pertaining to all groups of gastropod molluscs, and I cover both behavior and neurobiology. While the scope is therefore broad, I focus on recent discoveries that explain the neural control of behavior. My approach to the subject assumes that behavior is controlled by the cellular connections between specific neurons that operate within knowable circuits. In this respect, I accept the reductionistic premise that has fueled the recent growth of the field. Whether this approach will continue to be successful remains to be seen, but there is no denying the tremendous amount of knowledge that has so far been generated by adopting this approach, as evidenced in this book. In order to focus on the structure and function of the nervous system, I have not gone deeply into biophysics or molecular aspects, nor have I given justice to the extensive literature on chemical neuroanatomy that reports the use of immunohistochemical methods to localize neurotransmitters and neuromodulators within nervous systems. I have only included mention of these subjects in so far as they bear directly on the mechanisms of behavioral control. I hope that curious persons of diverse backgrounds will read this book. Those readers who have a prior knowledge of either molluscs or neurobiology, but not necessarily of both, will benefit most readily. However, anyone with a basic biology education should be able to understand the book. Because I assume that the reader already knows about neural signals, synaptic transmission, and so forth, I make no attempt to teach neurophysiology. As an introduction to the subject, and for subsequent reference, I begin the book with illustrations of the gastropod species that are most commonly studied in neurobiological research, together with drawings of their central nervous systems. The first four chapters should be particularly helpful to persons with backgrounds in vertebrate neuroscience, since they cover the broader aspects of molluscan biology and draw attention to the special

Preface

xi

features of the gastropod nervous system. Each of the remaining six chapters treats a different type of behavior. In general, I begin these later chapters by reviewing how the demands for behavior differ in diverse taxa, and how the adaptations differ. I then review progress in understanding the mechanisms of neural control, emphasizing cases in which control can be attributed to identified neurons and identified neural circuits. Because the scope of the subject matter is so broad, it has not been possible to include references to all pertinent publications. Although the list of publications is long, it is still greatly abbreviated relative to the actual literature that it represents. I regret having to omit mention of so many worthy authors. Nevertheless, I have tried to provide enough references to key papers to allow readers to quickly find additional publications on any subject that is mentioned in the text. With computer-based resources, it is easy to generate a long list of references on any chosen subject, but the search is greatly facilitated if one can start with a key reference. For this reason, and because credit should be given to those to whom it is due, I have usually cited the original and central publication in respect to all major findings. However, in some cases, for the sake of economy, I have had to cite just a single paper in reference to several related findings, or to refer to secondary sources. One way or another, I have tried to make the list of publications useful to those who wish to find out more. I would like to thank several people and institutions that contributed to the completion of the book: Yutaka Nishioka gave me the initial impetus; McGill University provided the opportunity, and my colleagues in the Department of Biology provided the requisite intellectual climate; my students, postdocs, and assistants throughout the years helped to make my own research on "snail brains" successful and fun; thanks to them for their enthusiasm, and thanks to the Natural Sciences and Engineering Research Council of Canada (NSERC) for consistent funding. The McGill University libraries either had what I needed or they got it; special thanks to Eleanor MacLean of the Blacker-Wood Library for her expertise. Thanks also to Kirk Jensen, my editor at Oxford University Press, for assisting at all stages in the preparation of the book. The illustrations shown in Figures I-VII were drawn by the talented and gracious artist Josee Morin. Wayne Sossin lent me one of his animals for the cover photograph. Finally, I am very grateful to my expert reviewers who provided feedback on earlier version of the manuscript. Irving Kupfermann read the entire manuscript and gave me many useful suggestions. Sadly, Irving died shortly before this book was published. His originality, vast knowledge and knack for discovery were elements that contributed in important ways to the science reviewed here. Other persons who read and criticized one or more chapters are Paul Benjamin, Rhanor Gillette, John Koester, Jon Jacklet, Louise Page, David Rogers, Rich Satterlie, Andries ter Maat, and Terry Walters. Many thanks to all of you.

Figure I. Aplysia californica (Opisthobranchia, Anaspidea). (A) There are about 37 species of Aplysia distributed worldwide, mostly in warm oceans. Since the time of Pliny, in the first century A.D., Aplysia has been known as the "sea hare" because of its resemblance to the terrestrial mammal. A. californica is found in the intertidal and sub tidal zones from central California to northern Mexico. It feeds primarily on red algae. A typical adult is 20cm in length and weighs about 500 g. (B) The central nervous system comprises four paired ganglia plus the unpaired parietovisceral ganglion, or abdominal ganglion, which is the most intensely studied of all gastropod ganglia. The parietovisceral ganglion was formed by the fusion of four ancestral ganglia.

xii

Figure II. Clione limacina (Opisthbranchia, Gymnosomata). (A) This shell-less animal, known as the "sea angel," lives its entire life drifting in the open ocean. Although it frequently swims in an upward vertical direction using highly muscularized "wings," it is unable to counteract the sea's currents. The adult body size ranges from about 20mm in Puget Sound, state of Washington, to about 85mm in the subarctic Atlantic Ocean. The species name is derived from the name of its exclusive prey, Limacina, which is a small planktonic snail (Thecosomata). Polar species feed on one species; temperate zone species feed on another species. (B) The central nervous system comprises five pairs of ganglia. The large posteriorly directed pedal nerve innervates the ipsilateral "wing."

xiii

Figure III. Tritonia diomedea (Opisthobranchia, Nudibranchia). (A) This animal lives on soft sediments in the northern Pacific Ocean, typically in the near-shore subtidal beds of its prey organism, a coral sea pen. Commonly known as a "sea slug," Tritonia is usually found with a uniformly pink color. Prominent respiratory tufts (branchia) protrude from the margin of the dorsal surface. Specimens range in size up to 30 cm in length. (B) The central nervous system is extremely compact and symmetrical. Note the fused cerebral and pleural ganglia.

xiv

Figure IV. Hermissenda crassicornis (Opisthobranchia, Nudibranchia). (A) Like most nudibranchs, Hermissenda is strongly colored. Particularly striking is the bright orange stripe running down the anterior midline and the bright blue outlines around the oral tentacles. The respiratory tufts are modified as cerata containing diverticula of the digestive glands. Hermissenda is carnivorous; when it feeds on anemones or hydroids, the prey's nematocysts end up in the cerata, where they are recycled as defensive weapons. Thus, the coloration of Hermissenda is probably aposematic (i.e., it is a warning to would-be predators). The animal is found in rocky intertidal habitats from Alaska to Mexico, and also in Japan. The average adult size is 3-6 cm in length. (B) The central nervous system is highly condensed, as in Tritonia.

xv

Figure V. Pleurobranchaea californica (Opisthobranchia, Notaspidea). (A) This animal lives along the Pacific coast of California at depths of 10-180 m. The prominent mantle covers an internal shell and a gill situated on the animal's right side. Pleurobranchaea is an indiscriminate carnivore that ingests its prey after capturing it with an explosive projection of the proboscis. When Pleurobranchaea itself is attacked, it releases a highly acidic mucus, pH 1-2. Sizes range from 15 to 50cm. (B) Note that the fusion of the cerebral and pleural ganglia, and the close apposition of the left and right halves, creates a single cerebropleural ganglion.

xvi

Figure VI. Lymnaea stagnalis (Pulmonata, Basommatophora). (A) This snail is a common inhabitant of fresh waters, especially where there is rich vegetation. It is found throughout Europe, North America, and northern Asia. The shell length ranges from 2 to 6cm. There is just a single pair of tentacles, with an eye located at the base of each one. Lymnaea can respire through its skin while submerged, but it must occasionally surface to breathe air. For food, it rasps at plant leaves or grazes on thin algal films. (B) Although the dimensions of the central nervous system are small, the cells are highly pigmented and clearly visible. The dorsal bodies contain endocrine cells.

xvii

Figure VII. Achatina fulica (Pulmonata, Stylommatophora). (A) From its original home in East Africa, this terrestrial snail has spread to tropical and subtropical regions around the world. It is an agricultural pest of economic significance in many places, owing in part to its prodigious capacity to reproduce. Adults generally have shells about 6-10 cm long, but specimens as large as 18cm have been found. All four tentacles carry an olfactory organ at their tips; the posterior pair also has an eye at the tip. (B) All ganglia, excepting the cerebrals and the buccals, lie beneath the esophagus where they form a ring around the cephalic aorta. Note that the right parietal ganglion is considerably larger than the left parietal ganglion.

xviii

Contents 1. The Gastropods 1.1. 1.2. 1.3. 1.4. 1.5. 1.6.

3

General Features 3 Torsion 4 Origins and Diversification of the Gastropoda 7 The "Prosobranchs" 10 The Opisthobranchs 12 The Pulmonates 14

2. The Central Nervous System 17 2.1. 2.2. 2.3. 2.4.

Organization of the Ganglia 17 Taxonomic Trends in CNS Organization 21 Structure of the Ganglia 25 Unique Properties of Gastropod Neurons 29

3. Sensory Systems 34 3.1. Chemoreception 34 3.2. Mechanoreception and Nociception 38 3.3. Sensory Cells for Chemoreception, Mechanoreception, and Nociception 39 3.4. Photoreception 43 3.5. Gravireception 47 3.6. Magnetoreception 51

4. Muscles and the Peripheral Nervous System 53 4.1. 4.2. 4.3. 4.4.

Muscles and Muscle Cells 53 Muscle Physiology 56 Peripheral Contributions to the Control of Reflexes 59 Cellular Elements and Plasticity in Peripheral Neural Circuits 62

5. Regulation of the Internal Environment 66 5.1. Respiration 66 5.2. Control of the Lung in Pulmonates 67 5.3. Blood Circulation 71 5.4. Respiratory Pumping 82 5.5. Water Regulation and Excretion 85

6. Locomotion 93 6.1. Crawling by Ciliary Beating 93 6.2. Crawling by Muscular Contractions xix

95

xx

Contents

6.3. Swimming 99 6.4. Taxes and Other Orientations 113

7. Feeding 124 7.1. 7.2. 7.3. 7.4. 7.5. 7.6.

Adaptive Radiation of a Complex Behavior 124 Food Finding 131 Central Pattern Generators 140 Variations of Buccal Motor Programs 147 Initiation and Modulation of Feeding 151 Plasticity of Feeding Behavior 163

8. Reproduction 170 8.1. 8.2. 8.3. 8.4. 8.5. 8.6.

Modes of Sexuality 170 Sexually Selected Behaviors 175 Finding Mates 181 Nervous Control of Courtship and Copulation 184 Egg Laying 195 Mechanisms for the Control of Egg-Laying Behaviors 209

9. Defense 9.1. 9.2. 9.3. 9.4. 9.5.

214

Dangers from Predators and Conspecifics 214 The Lines of Defense 215 Withdrawal Reflexes 218 Plasticity of Defensive Behaviors 234 Chemical Defenses 244

10. The Temporal Organization of Behavior 249 10.1. Seasonal Cycles 249 10.2. Daily Cycles 254 10.3. Endogenous Circadian Clocks 259 10.4. Mechanisms of Oscillation and Entrainment in the Eye 263 10.5. Multiple Influences on the Selection of Behaviors 266 10.6. Singleness of Action 272

References 281 Taxonomic Index 307 Neuron Index 309 Subject Index 311

Behavior & Its Neural Control in Gastropod Molluscs

This page intentionally left blank

1

The Gastropods 1.1.

General Features

The phylum Mollusca is second only to the Anthropoda in number of species. The molluscan lineage has been extremely plastic, and a great variety of structural plans have appeared. Of the seven classes of molluscs (Fig. 1.1), the largest is Gastropoda, which accounts for about 80% of the extant fauna. Estimates of the total number of living gastropod species is at least 40,000 and perhaps more than 100,000, with about 13,000 named genera (Bieler, 1992); the range of uncertainty indicates our substantial ignorance of this major taxonomic group. The word "gastropod" is something of a misnomer. The term was coined by George Baron Cuvier, in 1797, who was the first of the naturalist classifiers to notice that certain of the shell-bearing molluscs have much in common with other, shell-less molluscs. He gave the name "gasteropode," from the Greek words for stomach and foot, to all those molluscs that appeared similar. But what led Cuvier to construct this particular word? Probably, he picked up on the fact that these animals seem to crawl by means of their underparts, or bellies. It is true that the organ of locomotion in gastropods is located ventrally, where the human stomach is found, but it is not anatomically correct to say that gastropods crawl on their stomachs because the digestive tract of gastropods is mostly situated dorsally, within the shell, not near the foot. Nevertheless, apart from their roles in the naming of the class Gastropoda, the digestive system and the foot have played significant roles in the evolution of the so-named animals. For one, the evolutionary growth of the digestive gland forced a coiling of the dorsal part of the body, which is today evident in the spiral of snail shells. The shell itself is secondarily lost in slugs, but these animals still retain a visceral hump, which is an enlargement owing to the digestive gland. As for the foot, it has undergone frequent modifications to allow locomotion by different means, but since the foot is never very efficient as a locomotor force, the gastropods' early commitment to this form of locomotion accounts for their mostly sedentary habits. The morphological traits (apomorphic characters) that define the Gastropoda in relation to sister taxa (Fig. 1.2) are the larval operculum, the shape of the larval shell and, most importantly, torsion, which is discussed below (Ponder and Lindberg, 1997). Also typical of gastropods is a clearly 3

4

Behavior and Neurons in Gastropods

Figure 1.1. A traditional classification of the gastropod molluscs, based on morphological characters. Many points in the classification are controversial and subject to revision; one alternative is shown in Figure 1.2. The genera listed here are those that have been exploited for neurobiological research.

recognizable head bearing tentacles and eyes, a radula for rasping food, and nerve cells segregated in ganglia. However, none of these latter features is unique to gastropods, and they are presumed to be ancestral. The particular forms taken by gastropod species vary greatly, as do their sizes. Some species of terrestrial snails measure less than 1 mm in their longest dimension, whereas the sea slug Aplysia vaccaria can reach 990mm in length and weigh 14kg. Besides being morphologically diverse, gastropods also have diverse lifestyles. They are one of the few animal groups to successfully occupy marine, freshwater, and terrestrial habitats.

1.2. Torsion Torsion is the single most distinguishing characteristic of the gastropods. Torsion refers only to the twisting of the body; it is entirely different

The Gastropods

5

Figure 1.2. A proposed phylogeny of the gastropod molluscs, derived from a cladistic analysis of 117 morphological characters in 40 taxa (Ponder and Lindberg, 1997). Note that the primary division separates the Gastropoda into two major groups, Eogastropoda and Orthogastropoda, rather than the three subclasses shown in Figure 1.1. Sister groups to the Gastropoda are shown at the top. The listed taxa are those included in the work of Colgan et al. (2000), from whom the figure is taken. Using molecular sequence data, Colgan et al. (2000) found support for some aspects of the illustrated cladogram, but not for others.

from the spiraling of the shell. Fossil evidence suggests that early, non-twisted molluscs already had coiled shells. Conversely, some modern gastropods have uncoiled shells, or even no shell at all. The body plan of all modern gastropods undergoes torsion during some stage in the animal's development. Thus, torsion is a good example of ontogeny recapitulating phylogeny. But the significance of torsion goes beyond taxonomic classification, for it has profound effects on many body structures, including the nervous system. To appreciate torsion, we need to imagine the body plan of the molluscs that existed before the advent of the gastropods. Such speculation rests on the study of extant non-gastropod molluscs, as well as the examination of fossil shells. The early molluscs were probably symmetrical, and they had a alimentary tract running straight down the center of the animal from the mouth at the anterior end to the anus at the posterior end (Fig. 1.3A). The anus was present within a space called the mantle cavity that lay beneath the overhanging shell. Also within the mantle cavity was a pair of ctenidia,

6 Behavior and Neurons in Gastropods

Figure 1.3. Schematic illustration of torsion and its effects on the arrangement of various organs. (A) A hypothetical molluscan ancestor of the gastropods, before torsion. (B) A hypothetical early gastropod, after torsion. From observations of developing gastropods, it is known that torsion occurs in a counterclockwise direction, thus accounting for the twisted appearance of the alimentary tract and the nervous system. The so-called visceral loop comprises the nerves and ganglia connecting the pair of pleural ganglia by way of the visceral ganglia. Some gastropods, particularly the opisthobranchs, have undergone a secondary detorsion during their evolution, so they appear more like (A) than (B).

or gills. At some point during the early Cambrian period, animals appeared in which the shell and the mantle cavity were rotated 180° counter-clockwise towards the anterior end (Fig. 1.3B). These animals were the first gastropods. After further changes, the mantle cavity came to lie at the anterior end, as it does today in most gastropods. The exceptions, notably the opisthobranchs, can be accounted for by the later occurrence of detorsion (Section 2.1). One important consequence of torsion was the displacement of many interior organs. For example, before torsion the two heart auricles were located posterior to the ventricle, but after torsion they lay anterior to the ventricle (Fig. 1.3). The digestive tract became U-shaped, and some nervous ganglia, for example, the intestinal ganglia, moved to new positions. Primitively, there was a left and a right intestinal ganglion (Fig. 1.3A), but torsion caused the former left ganglion to move to the right of the former right ganglion (Fig. 1.3B). Since torsion occurs during development, either during the growth of nerves connecting the ganglia or after they have formed, the nervous system acquires a twisted appearance, a condition known as streptoneury. Note, in Figure 1.3B, that the left limb of the visceral nerve loop runs ventral to the alimentary tract, whereas the right limb is dorsal

The Gastropods 7

(see also Fig. 2.1). Contrasting to streptoneury is euthyneury, in which both limbs of the visceral loop run ventral to the alimentary tract and the loop as a whole is uncrossed; euthyneury is characteristic of some gastropod groups as the result of secondary evolutionary events (see Section 2.1). Authorities disagree about why torsion occurred at all. A widely held view is that the torted condition benefits the larvae, which, in the ancestral forms, were pelagic. When the animals were untorted, the defensive withdrawal of exposed body parts was compromised because the withdrawal of such crucial structures as the head and the velum would have had to wait until the foot was withdrawn. After torsion, the head was able to retract first, and then withdrawal of the foot closed off the shell aperture to prevent access to the head. Also, the early withdrawal of the velum causes the larvae to stop swimming so that it falls in the water column and thus escapes the would-be predator. Although this makes a plausible story, others have pointed out that some pelagic molluscs (e.g., the lamillibranchs) are untorted and yet they survive predation. Alternatively, it has been argued that torsion was an adaptation that allowed the gills better access to water flow, either to take advantage of flow patterns created by evolutionary changes in the shape of the coiled shell or simply to move the gills into the water current created by locomotion (Solem, 1974). Still another idea is that torsion brought the gills to the front end, where they can maximize benefits from small changes in the animal's orientation toward water currents (Morton, 1979). If the reader has correctly understood the description of torsion and its consequences for the rearrangement of organs, it will be apparent that the early gastropods must have been defecating on their own heads. Indeed, the excrement would have been forced upon their heads and their gills by the water current. One way in which evolution looked after this problem was by selecting animals with holes in their shells so that the incoming water could be quickly swept out. Also, to reduce fouling of the gills by excrement, one of the two bilateral gills was lost, and the remaining gill was lateralized opposite the anus. In a subsequent development (the Pulmonata), the gill was entirely replaced by an internal respiratory structure, the lung.

1.3.

Origins and Diversification of the Gastropoda

The Mollusca originated in the late Precambrian period or early Cambrian period about 600 million years ago (Solem, 1974; Runnegar and Pojeta, 1985). They came from a stock of flatworm creatures that also gave rise to the annelids. These ancestors were bilaterally symmetrical, unsegmented but serial in body plan, and acoelomate (i.e., lacking an internal body cavity lined with epithelium). They probably lived at the sandy bottom of the near shore, and they were probably carnivores. One major difference between these early molluscs and modern flatworms, or modern molluscs for that matter, is that the initial molluscs were tiny, measuring at most a few millimeters in length.

8 Behavior and Neurons in Gastropods

The gastropods originated in the early Cambrian period about 550 million years ago. Despite the existence of a fossil record, a great deal of uncertainty remains about the early events. Some authors maintain that the first gastropods were shaped like limpets (Haszprunar, 1988; i.e., with uncoiled shells), whereas others believe that they had coiled shells (Runnegar and Pojeta, 1985; Ponder and Lindberg, 1997). Workers in the latter camp disagree about whether the shells were coiled dextrally (to the right) or sinistrally (to the left). Because torsion is the defining character of the class Gastropoda, much effort has been given to finding the first of the torted molluscs. Possibly the first gastropods originated in the superfamily Bellerophontacea within the class Monoplacophora (Fig. 1.1). The bellerophonts became extinct about 250,000 years ago but they left a rich fossil record. Most of the shells have a perfect bilateral symmetry, as illustrated in Figure 1.4. The case for their inclusion in Gastropoda rests with an indentation that is present, in some forms, at the medial margin of the shell aperture (Fig. 1.4B). Some authorities (Knight et al., 1960; Haszprunar, 1988) believe that the slit is at the anterior end of the animal, and that it indicates movement of the anus and the mantle cavity from the rear of the animal to the front of the animal. As this implies a torsion of 180°, the bellerophonts would therefore qualify as gastropods. Other authorities find evidence in the fossils of bilaterally symmetrical muscle pairs, inconsistent with torsion, and they point out that a rotation of exactly 180° is unlikely; they believe the shell indentation is actually located at the posterior

Figure 1.4. Two fossil examples of the extinct Bellerophontacea, believed by some authorities to be the first gastropod molluscs. (A) Tremanotus, (B) Bellerophon. Note the marginal slit at the midline of the shell aperture. If the slit represents the anus, and if it is present anteriorly in these specimens, then the shells indicate torsion, the hallmark of gastropod identity. Scale bar is 2.5cm for (A), 0.5cm for (B). From Knight et al. (1960) with permission.

The Gastropods 9

end of the animal (Solem, 1974; Runnegar and Pojeta, 1985). According to the alternative view, the first gastropods came from a different group of monoplacophoran molluscs. The classification of gastropods is unsettled. Early taxonomic classifications were based on the observed similarities and differences between taxa, whereas contemporary classifications are assumed to correspond to phylogenies, that is, they are based on hypotheses about the evolutionary descent of the taxa. Bieler (1992) attributes the difficulty of phylogenetic studies in this group of animals to "the long evolutionary history, the often rapid radiations, and the adaptation to many habitats by members of the same evolutionary line and to the same habitat by distantly related forms" (p. 315). The introduction of cladistic approaches, emphasizing the distinction between primitive versus derived characteristics, and computer-based parsimony methods have brought more power to the analyses, but the controversies have not abated. Similarly, molecular studies have not yet proven sufficiently compelling to quell uncertainties. The traditional classification of the Gastropoda recognizes three subgroups of equal rank: Prosobranchia, Opisthobranchia, and Pulmonata (Fig. 1.1). This primary division of the class was originated by Milne-Edwards in 1848, based on the organization of the respiratory system. It later gained wide acceptance after publication of Thiele's treatise on Mollusca (1929-1931). However, as biologists examined more and more morphological characters, in greater and greater detail, the correctness of the traditional classification was increasingly questioned (see Bieler, 1992). Following the publication of two influential papers by Haszprunar (1988) and Ponder and Lindberg (1997), a consensus was reached to reject any tripartite division of the Gastropoda. In these papers, convincing evidence was presented for the paraphyletic nature of the Prosobranchia, meaning that the taxa formerly included in this group do not have a common ancestor. This poses a problem for neurobiologists reading the earlier literature, which makes frequent references to prosobranchs. Therefore, in this book I will continue to refer to prosobranchs, as classified in Figure 1.1, but I will use quotation marks (e.g., "Prosobranchia") to signify that most taxonomists now reject the term. There is still utility in the term "prosobranch" as an indicator of gastropod molluscs that are neither opisthbranchs nor pulmonates. It is generally agreed that the Opisthobranchia are paraphyletic or polyphyletic, but the Pulmonata, alone among the original three subclasses, are monophyletic (Haszprunar, 1988; Ponder and Lindberg, 1997; Colgan et al., 2000). These assertions must be taken as provisional, however, because neither the opisthobranchs nor the pulmonates have received much attention in recent analyses; the most intensely studied taxa belong to the "Prosobranchia." For example, Ponder and Lindberg (1997) used 40 taxa in total, but only four taxa from among Opisthobranchia and Pulmonata combined. In any case, I will use the terms Opisthobranchia and Pulmonata without quotation marks in this book.

10 Behavior and Neurons in Gastropods

A cladogram depicting the results of Ponder and Lindberg (1997) is shown in Figure 1.2. This is based on the most comprehensive cladistic analysis so far conducted, using 117 morphological characters, including those found in the shell, muscles, renopericardial system, reproductive system, digestive system, and nervous system. In comparison with the traditional scheme shown in Figure 1.1, the absence of "Prosobranchia" is evident. So, too, is the introduction of many new names to designate identified clades; many of these names are taken from earlier workers. It can be seen that the Gastropoda are monophyletic and comprise two clades, the small Eogastropoda and the much larger Orthogastropoda. The clades of principal interest to neurobiologists, Opisthobranchia and Pulmonata, are grouped as Euthyneura. Because the validity of the Euthyneura clade is widely supported (Haszprunar, 1988; Ponder and Lindberg, 1997; Colgan et al., 2000), it can be inferred that opisthobranchs and pulmonates descended from a common ancestor. It is noteworthy, however, that not all the conclusions of Ponder and Lindberg (1997) are well supported by a separate molecular study, which analyzed sequences from two segments of 28S rDNA and from the histone H3 gene (Colgan et al., 2000). This suggests that further revisions to the classification of the Gastropoda can be expected.

1.4.

The "Prosobranchs"

The name prosobranch refers to the anterior position of the gills, which is the torted condition of early gastropods. Indeed, the primitive gastropod features are seen most clearly in the "prosobranchs." Contrastingly, many primitive features, including torsion, have been lost through secondary evolution in the opisthobranchs and the pulmonates. Most species of "prosobranchs" occupy marine habitats, but some species have adapted to freshwater and even terrestrial environments. Feeding strategies are very diverse. At the morphological level, major adaptive changes have occurred in the mantle cavity, the alimentary tract and the reproductive system. The treatise of Fretter and Graham (1994) provides a complete summary of this group, including many marvelous drawings. Not surprisingly, there has been considerable controversy concerning the lower levels of classification within the "prosobranch" group. Traditionally, there are three orders: Archaeogastropoda, Mesogastropoda, and Neogastropoda (Fig. 1.1). As the names suggest, these groups were considered a progressive series. Haszprunar (1988) argued that there are just two "grades," the Archaeogastropoda and the Caenogastropoda (Caeno = combined), with the latter grade incorporating the Mesogastropoda and the Neogastropoda. However, the Archaeogastropoda are clearly paraphyletic and it is debatable whether the name should continue in use (Fretter and Graham, 1994; Ponder and Lindberg, 1997). Most of the taxa previously considered in this group are now assigned to the clade called Vetigastropoda (Fig. 1.2). Whereas the morphological analysis of Ponder

The Gastropods 11

and Lindberg (1997) strongly supports the Caenogastropoda, the molecular analysis of Colgan et al. (2000) only weakly supports it. The Archaeogastropoda-Vetigastropoda possess many primitive characters including paired organs such as gills, auricles (diotocardian), kidneys, and osphradia. Figure 1.5 illustrates a representative of this group from the genus Pleurotomaria. According to Graham (1985), the earliest gastropods likely resembled this contemporary snail. By contrast, the Caenogastropoda have just one set of gills, one auricle (monotocardian), and one kidney, each of which is situated on the left side of the animal. The Mesogastropoda are probably paraphyletic (Graham, 1985; Ponder and Lindberg, 1997). The group includes many familiar seashore animals among which are the periwinkles (Littorina), the glossy cowries (Cypraeidea), and the large conches (Strombus). The Neogastropoda comprise exclusively marine snails, and they are probably monophyletic. The advancement of the carnivorous lifestyle in

Figure 1.5. An archaeogastropod, Pleurotomaria. Many of the most primitive features of the class Gastropoda are evident in this modern snail. Note, for example, the paired gills, paired auricles, paired kidneys, and the slit at the anterior margin of the shell. The snail lives in the sea and feeds on detritus and thin algal growths, et, epipodial tentacle; f, foot; la, left auricle; Ic, left ctenidium; Ik, left kidney; 1m, left shell muscle; nl, neck lobe; o, eye on eye stalk; op, operculum; r, rectum re, right ctenidium; rk, right kidney; rm, right shell muscle; st, stomach; t, tentacle. From Graham (1985) with permission.

12 Behavior and Neurons in Gastropods

this group has produced some animals of special interest. Tropical snails of the family Volutacea burrow in the sand and smother prey with their large feet. One of these species, Melo, has a shell that is nearly a half-meter long. Many neogastropods (e.g., the muricids Murex, Busy con, Buccinum) are adept at boring holes in the shells of bivalves and gastropods. The toxoglossans (cone snails) have evolved a particularly dramatic style of feeding utilizing a poison that is injected into the prey via long, hollow, radular teeth. The poison is so potent, and its method of delivery so quick, that Conus snails are able to capture live fish. The Heterostropha are something of a taxonomic dilemma. These animals seem to be mosaics of characters from several of the standard taxa. Their overall organization resembles the mesogastropods, but their radular teeth resemble the neogastropods. The shell, the osphradium, and the sperm are of types seen in the opisthobranchs. Currently, the Heterostropha are considered a separate group, but more aligned with the opisthobranchs and the pulmonates than with the "prosobranchs," primarily because they have an uncrossed (euthyneural) nervous system (Fig. 1.2).

1.5.

The Opisthobranchs

The opisthobranchs (rear placed gills) appear in the fossil record at about the same time as the pulmonates in the late Carboniferous period, that is, some 300 million years after the first gastropods. The morphological feature that best distinguishes the opisthobranchs and the pulmonates from the "prosobranchs" is the uncrossed visceral nerve loop. The Opisthobranchia comprise the gastropod groups that remained in the marine habitat, while the Pulmonata comprise the groups that adapted to terrestrial and freshwater environments. It is possible that both the opisthobranchs and the pulmonates evolved from a common ancestor, as shown in Figure 1.2; alternatively, the pulmonates may have emerged from within the Opisthobranchia. Apart from euthyneury, there are few features that absolutely distinguish the opisthobranchs from other gastropods. In their cladistic analysis involving 117 characters, Ponder and Lindberg (1997) identified only four character changes (apomorphies) associated with the emergence of the Opisthobranchia, among which is the presence of one marginal radular tooth rather than none, and the modification of the cephalic tentacles into rhinophores. However, only the families Aplysiidae and Acteonidae were included in the analysis. Other features generally characterizing opisthobranchs are the reduction or loss of the shell, and benthic habitats. Classification within the Opisthobranchia is difficult because of mosaicism and extensive parallel evolution, which precludes a phylogenetic reconstruction from strictly parsimonious approaches to morphological character analysis. Authorities differ greatly in their classifications at the ordinal level, with the total number of orders ranging from 8 to 14. The classification shown in Figure 1.1 assumes eight orders. Some authors recognize three

The Gastropods

13

higher level groups: the tectibranchs, which have a shell covering their gills; the nudibranchs, which have no shells (and often no gills); and the pteropods, which are planktonic. Although these groupings are not monophyletic, they are often used as indicators of type. When used in the older literature, the term tectibranch generally refers to all gastropods now classified as anaspids, cephalaspids, and notaspids, even though most anaspids have only vestigial shells. Each of the opisthobranch orders is briefly described below. Figures I-V (see pp. xii-xvi) illustrate the species most commonly studied by neurobiologists. The most primitive morphological characters are found in the order Cephalaspidea. The name of this order refers to the presence of a protective head shield. The shell is small and internalized. Because the shell is also thin and bubble-like, these animals are known as "bulloids." Whereas all other opisthobranchs have become detorted, most of the cephalaspids are torted and they retain the primitive condition of a crossed visceral nerve loop. The foot is modifed as parapodial flaps, which serve as wedges for carrying the animal into or through the sand. Primitively, the cephalaspids have an open seminal groove, a non-retractile penis, and a common male/female genital pore. The order Anaspidea (without shields) is characterized by a "migration" of the intestinal ganglia posteriorly and the fusion of these ganglia with the visceral ganglia (Section 2.2). The synonymous term, Aplysiomorpha, by which the order is also known, indicates the best known genus, Aplysia (Fig. I), and signals the importance of the group for neurobiologists. These animals are slug-like, with either a small shell or a vestigial shell. They have two tentacles and two modified tentacles, called rhinophores. In contrast to the Cephalaspidea, which are mainly burrowers, the Anaspidea are browsers that feed on surface algae. A gizzard grinds and strains the algae prior its ingestion. They are called "sea hares" on account of their feeding habits, as well as their appearance. The Thecosomata and the Gymnosomata are together known as pteropods (winged feet). Both orders consist of planktonic swimmers, but they are otherwise quite different. The Thecosomata (encased bodies) have relatively primitive characters including a shell, feeding by ciliary capture of diatoms, and an operculum. The family Limacinidae consists of minute animals only a few millimeters in diameter with transparent spiraled shells, whereas the Cymbuliidae are more elongated and measure up to 5 cm. All these animals have very thin, but large, parapodia, which flap under muscular control and which give thecosomes their popular name, "sea butterflies." The Gymnosomata (naked bodies) are specialized as carnivores, and they feed mainly on thecosomes. They lack a shell, a mantle cavity, and gills; respiration is through the skin. Their bodies are streamlined for fast swimming, and they have a variety of specialized structures to aid in prey capture including, in different groups, hooks, suckers, sticky papillae, and sharp radular teeth. The widely distributed Clione (Fig. II) is a representative gymnosome pteropod, and it is the genus favored by neurobiologists.

14 Behavior and Neurons in Gastropods

The Sacoglossa (referring to a small sac used for storing worn radular teeth) are characterized by a manner of feeding very different from that of the gymnosomes. They are herbivores that suck the juices out of algae, cell by cell. In general, each species specializes on a different algal species. The radula is highly modified with rows of single teeth that are themselves distinctively shaped as blades for piercing algae cells. Once pierced, the contents of the cells are pulled into the opisthobranch's stomach by the pumping action of its pharynx. Interestingly, the ingested chloroplasts become incorporated into cells of the digestive gland, where they continue to photosynthesize. There are two subgroups of Sacoglossa. In one group, the animals have a shell, a mantle cavity and gills, while in the other group they have none of these structures. The Nudibranchia (bare gills) have no shell and no mantle cavity. They are completely detorted, and they have the most compact arrangement of nervous ganglia of any opisthobranch group. In most species, respiration in through branchial tufts or other appendages situated on the animal's broad dorsal surface, as in Tritonia (Fig. III). Some species, for example, Hermissenda (Fig. IV), adorn their backs with cerata containing nematocysts, and they color themselves outlandishly to warn would-be predators. In other species, color is used as camouflage. This is the only group of gastropods in which the cell nuclei contain 13 haploid chromosomes, a number that is found in all species of nudibranchs (Schmekel, 1985). How to classify the Notaspidea (back shields) has been hotly debated for more than 100 years. These animals have a shell, rolled rhinophores, and the anus and ctenidium on the right side. However, some authors do not consider the group monophyletic (Schmekel, 1985). Morphological differences within the group, primarily in the organization of the nervous system and the reproductive system, allow a distinction between two clades. The Pleurobranchomorpha, represented by Pleurobranchaea (Fig. V), look much like nudibranchs, whereas the Umbraculomorpha (e.g., Tylodina, Umbraculum) have primitive features, including limpet-like shells.

1.6.

The Pulmonates

When molluscs moved to terrestrial habitats from the sea, they faced a problem in how to conserve water. A major part of the solution was to seal off the mantle (pallial) cavity. This one innovation greatly reduced evaporative water loss. In pulmonates, the mantle cavity is lined with blood vessels for gas exchange, and a single narrow passage, the pneumostome, provides access to the exterior. New behaviors were also important in facilitating the terrestrial invasion. Especially useful is the ability to retreat into the shell and drastically reduce metabolic activity during extended droughts. Changes to the excretory system also contributed to the terrestrial lifestyle. With the enclosure of the mantle cavity, the ureter was rearranged to open externally, and water was conserved by excreting nitrogenous wastes as almost insoluble

The Gastropods

15

uric acid, rather than as ammonia, which must be diluted with large quantities of water. Like other gastropod groups, the pulmonates have proven remarkably adaptable to diverse niches. Terrestrial molluscs occupy all geographical regions with the exception of the polar extremes. Although they are most common in temperate and tropical areas, they are also found in deserts and high mountains. There are more species of pulmonates, approximately 35,000, than either "prosobranchs" or opisthobranchs. An interesting aspect of the pulmonate radiation is that it tends to occur on a small spatial scale. Because the animals move so slowly, and have little need to migrate, local populations can easily become isolated. One consequence is a high degree of sympatric diversity. For example, a single square kilometer in a Cameroon rainforest may contain as many as 97 species belonging to at least 12 families. Another consequence of spatial isolation in snail populations is the high degree of intraspecific genetic polymorphism. According to one estimate, Cepaea nemoralis has the most extreme intraspecific variation of mitochondrial DNA of any animal (Thomaz et al., 1996). Whether the terrestrial gastropods arrived on land via freshwater intermediary forms, or directly from marine "prosobranchs," is uncertain. Once landed, however, there was a rapid radiation of pulmonate types. Some forms readapted to water and again populated freshwater and marine habitats. The loss of a shell in some groups gave rise to shell-less slugs, especially in environments where calcium is scarce but water is plentiful. In such environments, the chances of desiccation are lessened and the energy saved by not forming a shell can be used for other purposes. The loss of a shell also led to the rearrangement of several internal organs. The pallial cavity of slugs is smaller than it is in snails because it is no longer needed to accommodate a retracted head; the esophagus is shorter, and the stomach is placed more anteriorly. Taxonomists typically divide the Pulmonata into at least three subgroups, or "orders." Two of the subgroups, named Basommatophora and Stylommatophora, are generally recognized (Fig. 1.1), but the name and composition of the remaining group(s) has been much debated (Bieler, 1992). Particularly troublesome are three families of slug-like animals that have been continuously shuffled from one taxon to another; in Figure 1.1 they are represented as Systellommatophora. The Basommatophora (basal eye-bearing stalk) have a single pair of tentacles and eyes located at the bases of the tentacles. The form of the shell varies widely from high-spired in some species, to flat and limpet-like in other species. There are no slugs in this order. The majority of the species are found in freshwater, but there are also terrestrial and marine species. Some of the snails breathe air, others take oxygen from the water, and some do both. The freshwater genus Lymnaea (Fig. VI) includes species exhibiting all three respiratory habits. The Stylommatophora (slender eye-bearing stalk) have two pairs of tentacles, with eyes located at the tips of the posterior, or dorsal, pair.

16 Behavior and Neurons in Gastropods

The monophyletic composition of the Stylommatophora has been confirmed by an analysis of 1,460 nucleotides of the ribosomal RNA gene-cluster from 104 species (Wade et al., 2001). All species are terrestrial, and many are locally abundant (e.g., Helix, Limax, Cepaea, Achatina; Fig. VII). Although the "prosobranchs" and the basommatophores both have terrestrial representatives, about 80% of terrestrial gastropod species belong to the Stylommatophora. The order is the largest and most diverse of the pulmonates with 71-92 families, depending on the classification scheme, and as many as 35,000 species (Bieler, 1992; Wade et al., 2001). These animals have the most advanced characters of all pulmonate groups relative to the ancestral gastropod prototype. Most species have fully sized shells, but the group also includes shell-less slugs and semi-slugs with residual shells that are too small to accommodate the body. Although most stylommatophores feed on plants and detritus, a few are carnivores that feed on other snails.

2

The Central Nervous System 2.1.

Organization of the Ganglia

In segmented animals such as annelids and arthropods, the nervous system is organized with one ganglion or a pair of ganglia per segment, plus an enlarged ganglionic mass (brain) at the anterior end. The absence of segmentation in molluscs allows the nervous system to be organized in a variety of ways. Whereas some molluscs have no nervous ganglia at all (nerve cells are simply scattered along the nerves), one striking feature of the gastropod anatomy is nervous ganglia. The fact that the ganglia occur in a bewildering variety of numbers and positions makes it nearly impossible to discuss subtleties in a brief description. After a thorough review of all published literature on the subject, Bullock (1965) wrote that "the gross anatomy of the gastropod nervous system embraces some of the most difficult problems in comparative anatomy" (p. 1288). The ganglia that comprise the central nervous system (CNS) generally include pairs of buccal ganglia, cerebral ganglia, pleural ganglia, and pedal ganglia, together with a visceral ganglion that is either paired or single. The drawings in Figures I-VII (see pp. xii-xviii) illustrate these ganglia in several commonly studied species. In addition to the ganglia already mentioned, "prosobranchs" also have two intestinal ganglia, one of which is supraesophageal and the other subesophageal (Figs 1.3 and 2.1). All of these ganglia, but especially the intestinal ganglia, appear differently in different taxa and they often move, fuse, or disappear altogether. Furthermore, additional ganglia appear in certain species, namely the accessory ganglion, the labial ganglia, the branchial ganglia, the osphradial ganglia, and the parietal ganglia. As if this were not enough, some of the ganglia are known by more than one name (Table 2.1).

Table 2.1. Synonymous nomenclature for ganglia Abdominal = Parietovisceral (Aplysia) Parietal = Pallial Infraintestinal = Subesophageal Supraintestinal = Supraesophageal

17

18 Behavior and Neurons in Gastropods

Figure 2.1. Representations of distinctive CNS arrangements, shown here for "prosobranchs." (A) Hypoathroid condition, in which the pleural ganglia are close to the pedal ganglion (or the pedal cord). (B) Epiathroid condition, in which the pleural ganglia are close to the cerebral ganglion. Note in both conditions that the visceral nerve loop is twisted or streptoneural. a, anus; bg, buccal ganglion; ccm, cerebral commissure; cpdc, cerebropedal connective; cpp, cerebropleuralpedal connective; Ib, labial ganglion; Ibn, labial nerve; leg, left cerebral ganglion; Ipl, left pleural ganglion; m, mouth; me, mantle edge; osg, osphradial ganglion; pcm, pedal commissure; pd, pedal ganglion; ppd, post pedal commissure; pdc, pedal cord; pdn, pedal nerve; pig, pleural ganglion; vg/lvg/rvg, visceral ganglion; sbg, subesophageal ganglion; sog, supraesophageal ganglion; sbv/suv, sub- and supra-esophageal part of the visceral loop. From Fretter and Graham (1994) with permission.

Member ganglia of each pair are linked by commissural nerves, except for the pleurals and the intestinals. In addition, long connective nerves link certain ganglia. Most importantly, as described earlier (Section 1.2), the pleural ganglia are connected to the intestinal ganglia, which in turn are connected to the visceral ganglion, thus forming the visceral loop. In Figures 1.3B and 2.1, the visceral loops are shown twisted around the gut (i.e., they are streptoneural). The use of the visceral loop as a diagnostic character in taxonomy is discussed below.

The Central Nervous System

19

Figure 2.2. Differences in the innervation of the mantle skirt between "prosobranchs" and opisthobranchs. (A) A typical "prosobranch." (B) An opisthobranch (Akera bullatd). The approximate extent of innervation by ganglia is indicated by arrowed lines. Light stipple indicates hypobranchial gland; dark stipple indicates the part of the mantle fused to the nuchal (neck) portion of the body wall. Anterior is at the top. Note the rotation of the entire mantle skirt. Also note the migration of the supraesophageal ganglion (sup) from the left to the right; this results in an uncoiling of the visceral nerve loop, abv, afferent branchial vessel; afa, afferent gill axis; cs, collecting sinus; ct, ctenidium; ebv, efferent branchial vessel; efa, efferent gill axis; gl, gill; Ip, left pleural ganglion; Ipl, left pallial (parietal) ganglion; os, osphradium; rec, rectum; rp, right pleural ganglion; sub, subesophageal ganglion; sup, supraesophageal ganglion; vis, visceral ganglion. From Brace (1977) with permission of the Zoological Society of London.

The different types of organization in the nervous system reflect adaptations for more efficient communication between neurons and between neurons and peripheral targets. As new peripheral structures appeared, or existing ones moved, the ganglia evolved to better serve them. An instructive example is shown in Figure 2.2, which illustrates schematically some changes in the innervation of the mantle edge that are presumed to have taken place during the evolution from "prosobranch" ancestors to modern opisthobranchs. These changes were driven by an anterior elongation of the body mass, accompanied by a detorsion, or right posterior migration, of the mantle complex. Perhaps the most significant transformation of the nervous system evident in Figure 2.2 is the movement of the supraesophageal ("sup") ganglion from the left side to the right side. Since the visceral loop was twisted when the ganglion was at the left, the move to the right involves an uncoiling

20 Behavior and Neurons in Gastropods

of the visceral loop, that is, a reversion to the euthyneural (uncrossed) condition. Also, after this evolutionary step the two pleural ganglia ("Ip" and "rp") no longer innervate the mantle edge. In the opisthobranch, a new ganglion, the left pallial ("Ipl"), innervates a portion of the mantle previously innervated by the left pleural ganglion, whereas the portion originally innervated by the right pleural ganglion is transferred to the subesophageal ("sub") ganglion. A prominent example of the changes that can occur when ganglia move and fuse is found in the so-called abdominal ganglion of Aplysia. This ganglion actually consists of two hemi-ganglia joined by a commissure. The right hemi-ganglion derives from a fusion of the ancestral right pallial ganglion and the ancestral supraintestinal ganglion, while the left hemi-ganglion derives from a fusion of the ancestral visceral ganglion and the ancestral subintestinal ganglion (see Kandel, 1979). Therefore, if we allow that the terms pallial and parietal are synonymous, it is appropriate and perhaps preferable to refer to the famous abdominal ganglion of Aplysia as the "parietovisceral" ganglion (Table 2.1). Rearrangements of the ganglia during evolution cause individual cells to move from one location to another, with interesting consequences. Two examples concern pairs of cells, which have morphological, physiological and functional symmetry but which do not reside in bilaterally symmetrical ganglia. In Aplysia, the giant neuron R2 is in the right side of the abdominal ganglion, whereas the equally large LP1 is in the left pleural ganglion. According to the interpretation of Hughes and Tauc (1963), these two cells were originally located in the symmetrical pallial ganglia. R2 moved from the right pallial ganglion to the right abdominal ganglion as a result of the fusions discussed above, whereas LP1 moved from the left pallial ganglion to the left pleural ganglion during the change of innervation patterns summarized in Figure 2.2. Another case of two large cells that appear to be homologous but are not symmetrical in their anatomical placements is be found in the pulmonate snail, Achatina fulica. Here, one cell is currently located in the right parietal ganglion, and the other is located in the visceral ganglion, but they probably originated as an homologous pair with one member in each of the two intestinal ganglia (Munoz et al., 1983; Zhuravlev et al., 2001). While streptoneury is a direct, and telling, consequence of torsion, euthyneury resulted from detorsion, but perhaps only in part. Detorsion involved a posterior migration of the anus and the mantle cavity (i.e., from the arrangement shown in Figure 1.3B back to the arrangement shown in Figure 1.3A). To better understand the origins of euthyneury, Brace (1977) examined a graded series of tectibranch (shelled) opisthobranchs. He recognized three major changes in the evolution of the nervous system in early opisthobranchs. First, as noted above, innervation of part of the mantle was transferred from the right pleural ganglion to the subesophageal ganglion. Second, the new pallial ganglia separated out from their origins in the pleural ganglia. Third, the nerves connecting the intestinal ganglia and the visceral ganglia became shorter, eventually leading to the fusions referred to previously. In the earliest

The Central Nervous System

21

stages, the migration of ganglia also resulted in new ganglionic origins for some of the peripheral nerves. For example, the pallial nerve moved from the subesophageal ganglion to the visceral ganglion, although the fields of innervation were unaffected. Thus, in Brace's view, euthyneury was produced not only by detorsion but also by condensation of the ganglia. Although many authorities share this opinion (Bullock, 1965; Haszprunar, 1988; Fretter and Graham, 1994), other authors find evidence to exclude condensation as a significant factor. Several specific ideas accounting for the transition from streptoneury to euthyneury have been extensively discussed and criticized (Bullock, 1965; Page, 1992). These same ideas have generated hypotheses about gastropod phylogeny, but these should be treated with caution (see Dorsett, 1986; Ponder and Lindberg, 1997). It is one thing to attach labels to ganglia in extant gastropods, but far more difficult to prove a history of their transformations. Furthermore, the work of Page (1992) has shown that the reliance on gross examination of adult forms, or even light microscopy, can be misleading. She studied larval development of the CNS by electron microscopy of semi-serial sections. In the particular case of the nudibranch Melibe leonina, her observations led her to conclude that the pleural ganglia are fused with the pedal ganglia, not with the cerebral ganglia as had been previously assumed.

2.2.

Taxonomic Trends in CNS Organization

In Europe, during the late 19th and early 20th centuries, a large effort was devoted to describing the nervous systems of gastropod molluscs. From this work, numerous publications appeared containing precise observations on a wide variety of species, often with beautiful illustrations (see Bullock, 1965). It is in this context that taxonomists began to construct phylogenies for the gastropod class based on nervous characters. Although most authorities agree that phylogenies should be based on the use of multiple organ systems, the discussions of CNS organization by taxonomists help us to appreciate the diversity of its forms, and they provide us with glimpses of its evolution. Spengel (1881) observed that snail species could be divided into those having an uncrossed visceral loop (Fig. 1.3A) and those having a crossed visceral loop (Fig. 1.3B). He used this character as the basis for an influential classification in which all crossed forms are united as Streptoneura, and all uncrossed forms are united as Euthyneura. In terms of the classification shown in Figure 1.1, the Streptoneura generally correspond to the "prosobranchs," whereas the Euthyneura include the opisthobranchs and pulmonates. In Figure 1.2, Euthyneura is given status as a taxon, but there is no taxon Streptoneura. Even in the traditional scheme, the StreptoneuraEuthyneura division is not workable for at least a couple of reasons. First, there are taxa which cannot reasonably be classified using the Streptoneura-Euthyneura division. For example, certain opisthobranch groups, particularly from the Cephalaspidea, have a crossed nervous

22 Behavior and Neurons in Gastropods

system but are traditionally classified as Euthyneury. Certain other groups (e.g., Cingulopsidae, Glacidorbidae, Pyramidelloided) have an uncrossed nervous system but are traditionally classified as "prosobranchs." Second, the condition of euthyneury, although clearly secondary to streptoneury, was arrived at convergently and probably involved different processes (detorsion and condensation) in different groups. Partly in response to the shortcomings of the streptoneury-euthyneury distinction, Haszprunar (1988) proposed that gastropod species be distinguished according to whether they possess parietal (= pallial) ganglia in the visceral loop. Since the number of ganglia in the visceral loop comes to five with the evolutionary appearance of the new parietal ganglia, Haszprunar uses the term Pentaganglionata to refer to the "highest grade" of Gastropoda. However, the Pentaganglionata are essentially the same as the Euthyneura, and Haszprunar has been criticized for unnecessarily introducing new taxonomic names (see Bieler, 1992). Other features of the nervous system have attracted the attention of taxonomists. In Fretter and Graham's (1994) review, several stages in the evolution of the "prosobranch" nervous system are identified, with particular attention given to the position of the pleural ganglia. Two prominent types are emphasized. The "hypoathroid" condition is characteristic of the Archaeogastropoda-Vetigastropoda clade. In this arrangement, the pleural ganglia and the pedal ganglia lie next to each other on the ventral side of the gut, and the pleural ganglia communicate with the cerebral ganglia via long connective nerves (cpp in Fig. 2.1 A). The "epiathroid" condition is found in the Mesogastropoda and the Neogastropoda (=Caenogastropoda). In epiathroidy, the pleural ganglia lie close to the cerebral ganglia. Now the connections between the pleural ganglia and the pedal ganglia are much shorter, and these two ganglia, together with the cerebral ganglia, form a ring around the anterior part of the esophagus (Fig. 2.1.B). The epiathroid condition, with its characteristic concentration of ganglia, is typical of the more derived taxa in all gastropod groups. Haszprunar (1988) has proposed that the hypoathroid-epiathroid difference can separate the "prosobranchs" (Archaeogastropoda in his classification) from the remaining, "advanced" gastropods, with the former characterized by the hypoathroid condition and the latter the epiathroid condition. Moreover, Haszprunar and Huber (1990) used this same criterion, together with the type of cells present in the procerebrum, to distinguish three groups of pulmonates, as shown in Figure 2.3 (see below). However, the value of the hypoathroid-epiathroid distinction for taxonomy is questioned by Page (1992). After studying the development of the nervous system in the opisthobranch Melibe, she concluded that the adult appearance of an epiathroid arrangement can be deceptive. She maintains that, in Melibe at least, the pleural ganglia are in fact fused with the pedal ganglia, not with the cerebral ganglia as previously believed; therefore, this animal has an hypoathroid arrangement of the CNS, not an epiathroid condition. Moreover, there are certain problem groups. For example, an hypoathroid

The Central Nervous System

23

Figure 2.3. A classification of the pulmonates according to CNS organization. Each illustrated genus represents its family and superfamily. Concentration of the ganglia increases from left to right. The three groups are differentiated by the type of nerve cells present in the procerebrum and whether the arrangement of the ganglia is hypoathroid or epiathroid (see Fig. 2.1). (I) Basommatophora; (II) Systellommatophora; (III) Eupulmonata (roughly equivalent to Stylommatophora). Ace, accessory ganglion; C, cerebral ganglion; db, dorsal body; eg, cerebral gland; Os, osphradial ganglion; P, pedal ganglion; Pa, parietal ganglion; PC, procerebrum (small stipple, globineurons; large stipple, large neurons); PI, pleural ganglion; ppc, parapedal commissure; Sb, subesophageal ganglion; Sp, supraesophageal ganglion; sec, subcerebral commissure; V, visceral ganglion. Adapted from Haszprunar and Huber (1990). Reprinted with the permission of Cambridge University Press. nervous system is present in Aplysia and in many pulmonates, especially the Systellommatophora, even though these taxa are considered "advanced" (see Haszprunar and Huber, 1990). The change from streptoneury to euthyneury was arrived at by detorsion in some groups, principally the opisthobranchs, and by condensation of the ganglia, principally in the pulmonates. The latter process resulted in the concentration of ganglia around the anterior end of the esophagus

24 Behavior and Neurons in Gastropods

Figure 2.4. Different routes to the concentration of the ganglia in opisthobranchs. Concentration and cephalization increases from left to right. Note that the Notaspidea is here represented by two groups, the Pleurobranchomorpha and the Umbracumorpha. The Cephalaspidea are shown as ancestral to all other orders, ce, cerebral ganglion; ce-pl-c, cerebropleural complex; ce-pl-g, cerebropleural ganglion; in, intestinal ganglion; pa, parietal ganglion; pa-sbi-vg, parietal-subesophageal-visceral ganglion; pe, pedal ganglion; pi, pleural ganglion; pl-pa-in-c, pleural-parietal-intestinal complex; sbi, subintestinal ganglion; spi, supraintestinal ganglion; vg, visceral ganglion; vl, visceral loop. Adapted from Schmekel (1985).

(cephalization). Four specific routes to cephalization have been identified by Schmekel (1985), as illustrated for the Opisthobranchia in Figure 2.4. (The figure composition reflects Schmekel's view that the Pleurobranchomorpha and the Umbraculomorpha as distinct orders; see Section 1.5.) The pleural ganglia fuse with the cerebral ganglia in some orders (Pleurobranchomorpha, Nudibranchia, Sacoglossa), whereas they migrate toward the pedal ganglia in other orders (Umbraculomorpha, Anaspidea). The intestinal ganglia migrate toward, or even fuse with, the pleural ganglia in most orders, but in the Anaspidea the intestinal ganglia fuse with the visceral ganglion. These and other changes result in different numbers of ganglia in different animals.

The Central Nervous System

25

There may be as many as nine separate ganglia in the order Cephalaspidea (e.g., Accra), whereas there are only four ganglia in some nudibranch species (e.g., Coryphella). The pulmonates also show degrees of concentration of the ganglia along the visceral loop, and here again concentration has been used as the basis for phylogenetic analysis (see Chase, 2001). However, concentration seems to have occurred independently in different groups and by different means in different lineages. The value of this character for phylogenetics is therefore problematic (Haszprunar and Huber, 1990). A more useful character for phlyogenetic analysis in the pulmonates was discovered by van Mol (1974) who observed that the procerebral lobe of the cerebral ganglion is present in all pulmonate gastropods but in no opisthobranch gastropod. He also observed that the size of the procerebrum, and the size of the neurons intrinsic to the procerebrum, varies in accordance with the animal's habitat and lifestyle. Terrestrial pulmonates, which van Mol believes have a greater reliance on olfaction than do aquatic pulmonates, have the largest procerebra. Van Mol arranged the pulmonate families into four orders, based primarily on the type of procerebrum, but also incorporating data on associated neurosecretory structures. A further refinement was made by Haszprunar and Huber (1990), who included the position of the pleural ganglion, (i.e., whether the CNS is of the epiathroid or hypoathroid type). The primitive conditions (shown at the left of Figure 2.3) are characteristic of the three pulmonate groups. The Basommatophora have an epiathroid CNS and a procerebrum with only large cells; the Systellommatophora also have an epiathroid CNS, but a procerebrum with only small cells; the Eupulmonata have an hypoathroid CNS and the procerebrum contains either small neurons alone, or both large and small neurons.

2.3.

Structure of the Ganglia

Each ganglion, as well as every nerve, is surrounded by a sheath that physically confines the neurons. It also protects the neurons, and it forms an interface for the exchange of materials between the nervous system and the blood. The sheath itself is largely made up of collagenous connective tissue fibers and glycoprotein laminae (Coggeshall, 1967). The exact structure of the sheath varies with distance from the ganglion. A membrane-like inner capsule encloses the neurons. Outside the inner capsule, the sheath is most dense close to the ganglion and it then becomes progressively looser. Also, the thickness and density of the sheath increase with age, and its appearance is different for different ganglia. The sheath does not significantly impede the exchange of ions between the interior of the ganglion and the extraganglionic fluids, but it does prevent the influx of larger molecules. It is not clear at what size substances in the hemolymph are excluded from entering the ganglion. Peptides, vital dyes, and perhaps even ferritin can penetrate the sheath, whereas India ink cannot. Within the sheath may be found globular cells, pigment cells,

26 Behavior and Neurons in Gastropods

Figure 2.5. Vascular supply to the ganglia. (A) Ventral view of the abdominal ganglion of the sea hare, Aplysia californica. Branching vessels from the aorta enter the sheath that surrounds the ganglion and end blindly within it. Adapted from Coggeshall (1967). (B) Dorsal view of the subesophageal complex of the European vineyard snail, Helix pomatia. The drawing illustrates the overlapping distributions of blood vessels (dark lines) and serotonergic nerve cell bodies (dark circles). Whereas the neurons lie within the sheath, the vessels lie just outside it. LAA and RAA, left and right anterior arteries; LFA and RFA, left and right anterior foot arteries; LP1 and RP1, left and right pleural ganglia; LPa and RPa, left and right parietal ganglia; VBA, ventral buccal artery; V, visceral ganglion. Copyright 1992. Adapted from Hernadi (1992) with permission from Excerpta Medica Inc.

gland cells, and muscle cells. In terrestrial snails, the glycogen content of the globular cells decreases during the course of the winter's hibernation, while the number of pigmented lipofuscin granules (from oxidated lipids) increases. The ganglia themselves are avascular, but the sheath is supplied with small branches of the arterial vascular system (Coggeshall, 1967; Hernadi, 1992). Nutrients enter these branches and they diffuse into the ganglion or nerve. At the same time, waste products diffuse out of the ganglia, into the sheath, and hence into the hemocoel where they collect in venous sinuses. The pattern of vascularization is different for each ganglion, but quite consistent from specimen to specimen. Two examples, from Aplysia californica and Helix pomatia, are shown in Figure 2.5. The absence of internal blood vessels, and the consequent reliance on diffusion for the two-way movements of nutrients and waste products, may limit the size of the ganglia. On the other hand, it is possible that the transport of nutrients and waste products may be aided by glial cells, some of which are located just beneath the basal lamina at the periphery of the ganglia. From its properties described as above, the sheath can be considered part of the circulatory system. In this regard it is noteworthy that some nerve cell

The Central Nervous System 27

axons end blindly within the sheath. Many of these fibers have their axoplasm filled with granules, and they are presumed to be secretory. In these cases, the sheath functions as a neurohemal organ, receiving and distributing hormones and peptide messengers. The most thoroughly studied of the neurosecretory axons are the neurites of the bag cells in Aplysia, which leave the abdominal ganglion and terminate in the sheath of the connective nerve (Section 8.5.1). As well, there are numerous, small, unidentified nerves that innervate the sheath (Coggeshall, 1967). Other nerve fibers innervate muscle fibers in the sheath, for example, the neuron L7 in Aplysia (Alevizos et al., 1989a). In Helix, the cell bodies of serotonergic neurons aggregate near sites where the blood vessels approach the surface of the ganglion (Fig. 2.5). Presumably, this arrangement facilitates the entry of serotonin into the blood circulation once it is released from the nerve cells. Within the ganglion are nerve cells and glia cells. The neurons are arranged as in most invertebrate ganglia, that is, with the cell somata forming an outer cortex that may be several neurons thick. The central core of the ganglion is largely comprised of neuropil and fiber tracts, although neurons may also invade this region. Figure 2.6 illustrates the foregoing features in the cerebral ganglion of the giant African land snail, Achatina fulica. Glia cells are numerous, both intermixed with the neuronal somata in the cortex of the ganglion and scattered in the neuropil (Fig. 2.6). They are identified in the light microscope by their small, dark, nuclei that are especially apparent after treatment with chromatic stains. With electron microscopy, their cytoplasm is often darker than that of adjacent neurons. In the neuropil, glial processes intermingle with neuronal processes. In the nerves, the glia cell bodies are mostly located at the periphery, and their processes extend inward at right angles to the long axis of the nerve, thus giving nerves the appearance of a sliced orange when cut in cross-sections. The glia cells extend long cytoplasmic processes that include filaments and glycogen particles. These processes loosely surround the neurons where they create large extracellular lacunae. Often, the glial processes invaginate the neural somata and axons forming deep infoldings called trophospongia. Mirolli and Talbott (1972) showed, using morphological measurements and geometrical modeling, that the infoldings allow the axons to stretch 2-3 times their resting length without affecting the conduction velocity for action potentials. They found that the extensive glial infoldings cause the geometrical factor for the length constant (the square root of the ratio between the surface area and the cross-section perimeter) to be independent of the degree of stretch. Consequently, the conduction velocity is unaffected by the animal's state of retraction or extension. There is little direct evidence for the function of glia cells in gastropods. Early notions that the glial infoldings might restrict the diffusion of ions within the perineuronal extracellular space are unsubstantiated. More credible is the idea that glial cells store glycogen as a reserve energy source for neurons, and that glia have a nutritive or trophic function. Some experiments have shown that neuronal activity stimulates the synthesis of glycogen in

28 Behavior and Neurons in Gastropods

Figure 2.6. Histological section of the left cerebral ganglion from the giant African land snail, Achatina fulica stained with toluidine blue. Most of the dark spots within the neuropil are glia cells. Note the small size of the procerebral neurons relative to those in the mesocerebrum and the metacerebrum. Some authors refer to the metacerebrum as the postcerebrum (see Chase, 2000). CC, cerebral commissure; CPC, cerebropedal connective nerve; Olf N, olfactory nerve. Adapted from Chase and Tolloczko (1989). Copyright 1989. Reprinted by permission of WileyLiss, Inc.

glial cells. This effect could be signaled by the elevated levels of extracellular K+ that follow high rates of firing in nerve cells. Another possible mechanism whereby neuronal activity might influence glial function, either nutritive or otherwise, is through synaptic transmission, for there is convincing morphological evidence of conventional chemical synapses from neurons to glia, at least in Aplysia (see Tremblay et al., 1979). The presence of gap junctions between glia cells has led to the suggestion that glia cells may constitute a spatial buffer system for extraneuronal K + , as has been proposed for vertebrates. According to the hypothesis, the glia cells adjacent to active neurons take up K + liberated by neurons during the generation of action potentials, and the gap junctions between glial cells help

The Central Nervous System 29

to redistribute the K + . Patch clamp recordings from glial cells reveal that K+ channels are indeed opened when adjacent neurons are made to fire, but the long delay in K + channel openings makes it doubtful that the glia cells could be involved in siphoning excess K+ released from the neurons (Gommerat and Gola, 1996). Because the K + channel openings are gated by intracellular cAMP, Gommerat and Gola (1996) speculate that the channels might be involved in glial cell proliferation. Gastropod glia have also been implicated in Ca++ and Mg++ transport by the finding of Ca++/Mg++-ATPase enzyme activity in the glial membranes facing the lacunar spaces and in certain glial intracytoplasmic organelles (Fenoglio et al., 1997). A recent study by Smit et al. (2001) has proposed a novel function for glia based on experiments with cultured cells from the pond snail Lymnaea stagnalis. These authors discovered a soluble acetylcholine-binding protein (AChBP) that is synthesized in glia cells and released into the synaptic cleft between presynaptic and postsynaptic neurons. Since the protein binds transmitter molecules before they reach postsynaptic receptors, its presence in the cleft suppresses synaptic transmission. It is suggested that the normal function of glial AChBP in vivo is to buffer acetylcholine concentrations in the cleft, or to modulate the efficacy of transmission.

2.4.

Unique Properties of Gastropod Neurons

The degree to which gastropods are useful "models" for understanding other nervous systems is a function of how similar their neurons are to those of other animals. Because the mechanisms that allow for neuronal signaling have been evolutionarily conserved, gastropod neurons are, in fact, very much like those of other animals. In particular, the membrane composition, ion channels, and mechanisms of synaptic transmission of molluscan neurons do not differ in any fundamental way from that of other invertebrates or even vertebrates (see Kandel, 1976). One widely appreciated property of some gastropod neurons is their large size. Actually, the vast majority of gastropod neurons are small. "Prosobranchs" have no giant neurons at all. Even in opisthobranchs and pulmonates, most of the neurons fall within the range of sizes found in vertebrate brains. For example, in many stylommatophoran pulmonates, the procerebrum contains the majority of all CNS neurons, and these cells have somata that are only about 6 urn in diameter (Fig. 2.6). However, most species of opisthobranchs and pulmonates can claim 10-20 neurons in the category "giant." The largest cells have somatic diameters up to 200 um in pulmonates and up to lOOOum in opisthobranchs. The largest neuronal cell body so far discovered in any animal is R2 of Aplysia (about lOOOum). The significance of giantism, or the answer to why giant neurons mostly occur only in this particular group of animals, is not entirely clear. On one level, it is apparent that neurons with large cell bodies have neurites that

30 Behavior and Neurons in Gastropods

extend for long distances and that branch profusely. The size of the soma can therefore be attributed to the summed metabolic demands of the axonal and dendritic tree, most of which are probably associated with synaptic function. To satisfy these metabolic requirements, the giant neurons have extra copies of genes. The cell R2, for example, has 200,000 times the haploid amount of DNA, which represents 16 doublings of the diploid content (see Gillette, 1991). Presumably, the extra DNA provides an additional production of proteins necessary to support the large neuronal structures. However, it should be noted that even small gastropod neurons (7-10um) have DNA in excess of the diploid amount. Also, total replication of the genome (polyploidy) does not always occur. Many cells appear to replicate only selected genes (Chase and Tolloczko, 1987). In any case, the amount of DNA in giant neurons can only tell us how their size is achieved, not why it evolved in this particular group of animals. Gillette (1991) has presented an intriguing speculation that begins with the observation that the earliest molluscs were minute creatures, whereas presentday gastropods are orders of magnitude larger. Since larger bodies require more servicing by the nervous system, there were two possible accommodations during evolution. One was to increase the number of neurons and the other was to increase the spatial domain of existing neurons. Clearly, "prosobranchs" took the first option, whereas opisthobranchs and pulmonates took the second. Comparing animals of equal size in the groups "Prosobranchia" versus Opisthobranchia-Pulmonata, the former have many times, perhaps thousands of times, as many neurons as the latter. Gillette's argument begins with the suggestion that the behavior of modern opisthobranchs-pulmonates is more simple than that of "prosobranchs." Therefore, if there were no great demands for behavioral subtleties in the early evolution of opisthobranchs and pulmonates, they might have economized on developmental complexity by using a small number of very large neurons to accomplish crudely executed sensory-motor tasks. This scenario suggests that neuronal giantism is a sign of behavioral simplicity, or at least of sluggishness. Certainly it is often true, for nervous systems in general, that behavioral complexity and sensory ability are associated with large populations of small neurons. Nevertheless, it will be difficult to test Gillette's underlying assumptions that opisthobranch-pulmonate behavior is less demanding of the nervous system than is "prosobranch" behavior, and that the reduced energy costs associated with few rather than many neurons has a significant selective advantage. Unlike vertebrate neurons, gastropod neurons seldom have a bipolar form. Rather, they usually have a single large process emerging from the cell body. The process is often referred to as a "neurite" because it combines axonal and dendritic functions. The neurite may bifurcate close to the soma, or branch deeper in the neuropil. In some neurons, additional fine processes can be seen emanating from the soma at positions removed from the origin of the neurite, but whether these have some special function, or indeed any function, is unknown. The predominance of monopolar cells is explained

The Central Nervous System 31

by the fact that the cell bodies are located at the periphery of the ganglion, so their synaptic contacts lie only to the interior. Therefore, in contrast to bipolar cells, where the dendrites occupy one side of the soma and the axon is on the other side, it is more difficult to assign dendritic and axonal functions to portions of the neurite in monopolar neurons. Input and output synapses are often intermixed on the same neuronal process. In some neurons, regions of the neurite with predominately dendritic functions can be identified by extensive branching and the predominance of input synapses. Other regions are identified as axonal because the processes extend into nerves and have few or no branches. Processes with varicosities are usually output regions, since the varicosities typically contain presynaptic release sites. Apart from these generalities, the actual sites of reception and transmission in a given neuron can only be determined by laborious electron microscopy or clever electrophysiology, neither of which is commonly practiced. Another morphological anomaly is the tendency of certain cells to have axons that branch profusely before they enter a peripheral nerve or connective, thus resulting in as many as 25 axons from the same cell lying parallel in the same nerve. If these axons were destined to communicate with targets beyond the end of the nerve, it would obviously be more efficient for them to branch close to their targets. In considering this situation, Pin and Gola (1984) suggested that the fibers are secretory, that the nerve is a neurohemal organ, and that the presence of multiple fibers increases the number of release sites. This is an attractive idea, but it remains to be shown that such multiple fibers do, in fact, secrete messenger molecules en passant. Chemical transmission at synapses is more common than electrical transmission, although many predominately chemical synapses have an electrical component. Apart from neuropeptides, the common gastropod neurotransmitters are all familiar as "classical" transmitters in vertebrates: acetylcholine, serotonin, dopamine, histamine, X-aminobutyric acid, glycine, and glutamate. The neuropeptides, however, seem to be quite different in gastropods and vertebrates. Only a few neuropeptides, for example, insulin and enkephalin, or their variants, are found in both animal groups. It is clear that gastropod neurons synthesize and release a large number of peptides. Descriptions of these peptides, and their functions, appear elsewhere in this book. In addition to the conventional transmitters and the peptide transmitters, the gas nitric oxide (NO) also seems to be a neurotransmitter in gastropods. Histochemical markers have localized the synthetic enzymes for NO production to specific neurons or neuronal groups in several species. While these studies indicate that NO synthesis is most prominent in the feeding system of gastropods, investigations of the signaling properties of NO have just begun, and there is not yet any evidence for a specific functional role (Moroz, 2000). Early electron microscopists (e.g., Coggeshall, 1967) remarked on the rarity of synapses evident in the gastropod neuropil. With further sampling and improvements in technique, structures having much the appearance of vertebrate central synapses are now understood to be commonplace, if not

32 Behavior and Neurons in Gastropods

Figure 2.7. Uitrastructure of synapses in Achatina fulica. The active zones (between arrowheads) are characterized by an accumulation of vesicles, a densification of the perimembranous presynaptic cytoplasm and a widening of the intercellular cleft. Several types of vesicles are evident, suggesting that more than one substance is released, but probably only the small clear vesicles are released at the synaptic site. (A) Asymmetrical, or polarized, synapse. (B) Symmetrical synapse, suggestive of bidirectional transmission. More than 25% of the synapses in the tentacle ganglion of Achatina are of the symmetrical type. Scale bars, 150nm. From McCarragher and Chase (1985). Copyright 1985. Reprinted by Permission of WileyLiss, Inc.

abundant. Most of the synaptic specializations in gastropods differ from their vertebrate counterparts only in the absence of a pronounced postsynaptic density (Fig. 2.7A). However, some contact sites are remarkable for their ultrastructural symmetry, with vesicle aggregations and perimembranous densifications present on both sides of the synaptic cleft (Fig. 2.7B). At these synapses, transmission presumably occurs in both directions across the junction. From studies in the snail Achatina, it appears that the prevalence of symmetrical synapses varies considerably in different neuropils. In the tentacle ganglion, they account for more than one quarter of all detectable synapses (McCarragher and Chase, 1985; Chase and Tolloczko, 1993). Ultrastructural observations suggest that extrasynaptic transmission is important in the gastropod nervous system. Individual neuronal profiles obtained from the neuropil, or from zones of peripheral nerve termination, generally contain a variety of vesicles, as illustrated in Figure 2.7. Three broad types can be distinguished: clear, dense, and dense-core. However, intermediates are also seen, and there is often a large range of sizes within each type (Tremblay et al., 1979; Chase and Tolloczko, 1992). It is generally assumed that the small clear vesicles contain one or another of the "classical"

The Central Nervous System 33

transmitters, whereas the large dense vesicles contain peptides. Physiological evidence, some of which is reviewed in Section 7.5.6, leaves no doubt that multiple transmitters are released from at least some single neurons. The small clear vesicles aggregate at membrane sites that possess the specializations associated with synapses, while the large dense vesicles, or dense-core vesicles, aggregate at morphologically undifferentiated sites (Chase and Tolloczko, 1992). To study the targeting of neuropeptide-containing large dense-core vesicles (LDCVs), Karhunen et al. (2001) stimulated an identified motoneuron to fire at physiological rates. They found that repeated electrical stimulation evoked a large release of peptide (in this case, the small cardioactive peptide (SCP)) and a concomitant redistribution of the LDCVs. More LDCVs were docked close to the membrane after stimulation than before stimulation. Significantly, this redistribution was seen only at membrane regions that were not directly opposed to the target muscle (i.e., they only occurred at non-synaptic sites). Probably the majority of central gastropod neurons release one "classical" transmitter at conventionally differentiated synaptic sites and one or more peptides at non-synaptic sites.

3

Sensory Systems The behavior of every animal depends on its perception of the external world. In the case of gastropods, their world has no sounds and, in most cases, no sights. Gastropods do have eyes, but in only a few species are they used for object recognition. Thus, the distance perception of gastropods usually depends on olfaction, and their perception of near objects is dependent on a combination of chemoreception and mechanoreception. The present chapter contains descriptions of the major sense organs, the receptor cells types found within them, and associated sensory structures. More detailed accounts of these systems, especially regarding electrophysiology, are given in later chapters where the behaviors that they control are discussed.

3.1.

Chemoreception

For the gastropod species that have traditionally interested neurobiologists, olfaction is the paramount sense, being the only means of distance perception as well as an important modality for identifying objects that are physically contacted. Chemoreception controls or influences numerous specific behaviors, including feeding, homing, aggregation, mating, escape, and avoidance (Croll, 1983). The chemoreceptive nature of an anatomical structure is not usually obvious from its appearance. To identify chemoreceptors, investigators have to use a number of approaches, including histological searches for concentrations of putative receptors, testing for chemical sensitivity by directed applications of stimuli, and testing for behavioral deficits after lesions. Electrophysiological studies have not contributed a great amount of information because the receptors are small and inconveniently located in peripheral tissues. Two types of chemoreceptor cells can be distinguished. Those that have a role in the identification of contacted objects (gustation) have relatively high thresholds and are usually concentrated around the mouth. Chemoreceptors that are specialized for the perception of distant chemical sources have lower thresholds and are often located on conspicuous outcroppings of the head. However, chemoreceptive cells are probably present throughout the integument even if at low densities in some regions. Figure 3.1 shows

34

Sensory Systems

35

Figure 3.1. The location of the main chemosensory organs in a variety of gastropods. In every case, a few areas of the body are especially sensitive to chemical stimulation, as indicated. However the eye (marked e) is not chemosensitive. a, anus; at, anterior tentacle; c, ctenidium; ep, epipopdium; eso, epipodial sense organs; et, epipodial tentacles; f, foot; ma, mantle; mo, mouth; o, osphradium; oo, olfactory organ; op, operculum; ot, optic tentacle; p, pneumostome; pg, pallial gills; pt, pallial tentacles; r, rhinophore; s, siphon; t, tentacle. Adapted from Emery (1992). Copyright 1992. Reprinted by permission of Wiley-Liss, Inc.

where the most significant chemosensory structures are located in a variety of gastropods. The osphradium is an unusual chemoreceptive organ because it is neither conspicuous nor located on the head; moreover, it seems to have functions different from other chemoreceptor organs. It is present in all species of "prosobranchs" and aquatic pulmonates, some opisthobranchs, but no

36 Behavior and Neurons in Gastropods

Figure 3.2. A striking example of evolutionary convergence illustrated by the olfactory organs of a marine snail and a fish. (A) The osphradium (Os) of the oyster drill Thais haemastoma canalicula is shown lying adjacent to the siphon (S), the ctenidium (Ct) and the mantle epithelium (Me); R indicates the raphe. (B) Ventral view of the osphradium in the oyster drill. The lamellae are spontaneously active in the living animal. (C) Dorsal surface of the olfactory rosette in the nasal cavity of the channel catfish Ictalurus punctatus. All the images were obtained with a scanning electron microscope. From Garton et al. (1984) with permission.

terrestrial gastropods. It is situated within the mantle cavity, typically at the base of the siphon and often along side the ctenidia (Fig. 3.2A). Because it is present in the pathway through which water enters the mantle cavity from the environment, early authors assumed that the function of the osphradium is to test the quality of the water before it passes over the respiratory organs. Consistent with this view, Stinnakre and Tauc (1969) found that changes in the osmolarity of the seawater are detected by the osphradium and relayed to the CNS in Aplysia. A 5% dilution of seawater is sufficient to cause inhibition of R15, a neuron which contains a peptide that affects water regulation (Kupfermann and Weiss, 1976; Weiss et al., 1989). Other experiments have suggested that the Aplysia osphradium monitors the pH of the seawater. However, in Lymnaea, changes in water pH have no effect on spiking activity in the osphradial nerve, while increases in the carbon dioxide (CO2) content of the water cause an increase in spiking activity; decreases in CO2 have variable effects (Wedemeyer and Schild, 1995). Changes in water quality, usually accompanied by increases in its oxygen content, are effective

Sensory Systems

37

in triggering egg laying in Lymnaea, and these changes are presumably detected by the osphradium (see Hermann et al., 1994). Besides this role in the monitoring of water quality, other results suggest different roles for the osphradium. Deficits in food capture have been reported in several species after experimental lesions of the organ, and a primitive role in the detection of conspecifcs has been suggested (see Haszprunar, 1985). The morphology of the osphradium varies in different species from patches of specialized epithelium to bipectinate lamellar structures (Fig. 3.2). It is larger and more elaborate in carnivorous and predatory species than in herbivorous and detritivorous species (Haszprunar, 1985). A striking example of evolutionary convergence is given by the close morphological resemblance of the olfactory organ in the nasal cavity of fishes and the osphradium in the mantle cavity of "prosobranchs," as shown in Figures 3.2B and 3.2C. Within the central core of the osphradium is an elongated nervous ganglion. Recordings in Lymnaea have shown that the ganglion cells are responsive to both amino acids and organic compounds (Wedemeyer and Schild, 1995), but it is not clear whether the ganglion cells are themselves sensory receptors. Chemoreceptors are thought to be present on the thin lamellae, which extend bilaterally from the central core of the osphradium, of which there may be as many as 200 pairs in some species. The enlarged surface area created by this leaflet arrangement allows for the deployment of more receptors than would be possible with a solid structure. The structure of the lamellae, combined with their spontaneous movements, may also create water currents that facilitate sensory detection. The tentacles are usually, but not always, chemosensitive. Most "prosobranchs" have one pair of chemoreceptive tentacles, but limpets have tentacles distributed around the entire circumference of the mantle (Fig. 3.1). The opisthobranchs have two pairs of tentacles. The posterior pair of tentacles are called rhinophores (literally, bearers of a nose), while the anterior tentacles are called oral tentacles because they constitute the lateral margins of the oral veil. The epithelium of all these tentacles is histologically specialized for chemoreception (see below). The rhinophores of some species are also sensitive to mechanical stimuli and even photic stimuli (Chase, 1979; Hamilton, 1991). All aquatic pulmonates have a single pair of tentacles and in most species they are chemoreceptive. However, in a few genera (e.g., Helisoma, Biomphalarid), the tentacles themselves contain no chemoreceptors and they function only to direct water currents down towards the base of the tentacles where the olfactory organ is located (see Emery, 1992). In terrestrial pulmonates, which have two pairs of tentacles, the anterior (inferior) pair is particularly important for following mucus trails and other chemical cues that lie on the substrate, while the posterior (superior) pair has eyes at the tips and is used for sensing airborne chemical signals (Chase and Croll, 1981). This difference in function between anterior and posterior tentacles probably results from their different positions relative to odor sources, rather than from any essential difference in chemosensitivity. The anterior tentacles lie close to the ground and their

38 Behavior and Neurons in Gastropods

chemosensitive tips often touch it, whereas the tips of the posterior tentacles extend up into the airstream. The organization of the tentacle in terrestrial snails is illustrated in Figure 3.3. The olfactory epithelium comprises only the ventral surface of the terminal knob. The surface of the olfactory epithelium is covered with a brush border of complex ultrastructure that protects the sensory structures from desiccation and may also trap olfactory molecules. Several types of histochemically defined secretory gland cells are found specifically associated with the olfactory organ; these are probably important for conditioning the immediate environment to optimize the capture of odorant molecules (see Chase, 1986). As many as one million sensory neurons are clustered beneath the epithelium. These neurons continuously die and are replaced, as they are in the vertebrate nose. The bipolar receptors have dendrites that extend upwards to the free surface and axons that descend to reach three different targets (Chase and Tolloczko, 1993). Only about 10% of the receptor axons travel directly to the cerebral ganglion. The remainder of the receptor axons terminate in the neuropil of a ganglion that is situated in the tip of the tentacle. The tentacle ganglion has finger-like extensions that contain synaptic glomeruli (Fig. 3.3). Receptor cell axons end either in the glomeruli or in the main part of the ganglion; a stout nerve connects the tentacle ganglion with the cerebral ganglion. Neurons of several morphological types are present in the tentacle ganglion, and a comparison of the neuronal organization in the anterior versus the posterior tentacle found no significant difference, thus supporting the view that the two tentacles serve essentially the same functions but from different vantage points (Ito et al., 2000). Peripheral ganglia are associated with the tentacles in terrestrial pulmonates and opisthobranchs, but not usually in "prosobranchs" or aquatic pulmonates.

3.2.

Mechanoreception and Nociception

The entire skin is sensitive to mechanical stimulation to varying degrees at different places. Mechanoreception serves several important roles in gastropods, including the mediation of defensive behaviors, orienting the animal to wind and air currents, sensing the position of a mating partner, and locating food objects prior to ingestion. In keeping with these functions, the areas of the skin that are most sensitive to touch are the tentacles, the siphon, the perigenital region, and the perioral region. The term "nociception" is reserved for the detection of stimuli that elicit defensive behaviors. Whereas strong mechanical stimuli are noxious, so too are certain types of chemicals. Thus, nociception, and the related concept of noxious stimulation, is defined more by the nature of the response than by the physical nature of the stimulus. As discussed below, some types of receptor cells are especially responsive to noxious stimuli.

Sensory Systems

39

Figure 3.3. The olfactory system of terestrial pulmonates. Composite drawing based on observations in Achatina, Helix, and Limax. All illustrated structures are present bilaterally, although some are shown unilaterally for clarity. The inputs to the procerebrum from the inferior (anterior) tentacles are not shown. Note that three pathways link the procerebrum with other CNS regions: (1) nerve fibers of intrinsic procerebral neurons; (2) nerve fibers from an identified buccal ganglion neuron; and (3) nerve fibers of pedal ganglion neurons. PA, Parietal ganglion; FED, Pedal ganglion; PL, Pleural ganglion; V, Visceral ganglion. From Chase (2001) with permission.

3.3.

Sensory Cells for Chemoreception, Mechanoreception, and Nociception

The sensory cells of gastropods are primary receptors whose cell bodies lie either in the integument or in the CNS. The cell bodies of the centrally located

40 Behavior and Neurons in Gastropods

sensory neurons are large enough to penetrate with a microelectrode, and it is consistently found that the peripheral processes of these cells are sensitive to mechanical stimulation but not chemical stimulation. Presumably the cell bodies of the chemoreceptors are peripherally located. It is very likely that there are also peripherally located mechanosensors, although none has yet been identified physiologically. Most areas of the integument contain several histological cell types with dendritic endings in the skin surface indicative of a sensory function (Fig. 3.4A). If patches of skin could be shown to have sensory properties limited to only a single modality, one could identify the cells present in this patch of skin as receptors specialized for that particular sensory modality. However, no unequivocal examples of this type of limited sensory capacity have been reported for any tissue, including the osphradium and the tentacle tips. Thus, it is necessary to combine intracellular recordings with morphological labeling to identify a specific morphological type with a particular sensory function. Unfortunately, the small size of these cells and their inconvenient locations discourage such experiments. A final complication is that there are more morphological cell types than there are sensory modalities, if we count as modalities only mechanoreception and chemoreception. Some of the receptors may therefore have photic sensitivity or temperature sensitivity. Alternatively, there may be subspecializations within the mechanical and chemical modalities. In any case, the sensory roles of the various types of receptors have not yet been determined with certainty. Using Golgi impregnations and scanning electron microscopy of skin from the head of the pond snail Lymnaea stagnalis, Zaitseva and Bocharova (1981) described at least six types of subepithelial receptor cells, and possibly as many as nine types, depending on whether one judges certain variations of terminal dendritic structure as continuous or discontinuous. Three types have sensory endings extending to the epithelial surface, while three other types have endings within the epithelium, as shown in Figure 3.4A. One of the latter type of receptor endings was not found associated with any subepithelial cell body, so these endings presumably belong to centrally located cell bodies. Probably all areas of the skin, including regions specialized for sensory reception and those not so specialized, contain collections of these receptor types in different combinations and in different concentrations (Dorsett, 1986; Emery, 1992). Although authors disagree about exactly how to type the receptors, they generally agree about how many types there are. Thus, there are 6-9 types in the aquatic pulmonate Lymnaea, 3-5 types in the "prosobranch" osphradium (Haszprunar, 1985) and 4-6 types in the tentacular organ of terrestrial pulmonates (see Chase, 1986). Zaitseva and Bocharova (1981) estimate that in the tentacles and lips of Lymnaea the total density of receptors reaches about 5 x 105 per 1 mm2. The cell bodies of the receptor cells are small (<20um), bipolar, and generally subepithelial in position. A slender axonal process (0.2-0.5 um) arises from the central pole of the cell body. This process is often directed

Sensory Systems 41

Figure 3.4. Morphology of epithelial receptors. (A) Representation of the epithelium in the tentacle of the pond snail Lymnaea stagnalis, based on scanning electron microscopy and light microscopy after silver impregnation. Six types of receptor endings (RE) are indicated, three of which rise to the epithelial surface and three of which terminate within the epithelium. Cell bodies associated with all but one type of ending are present beneath the epithelium; cell bodies for the remaining type are probably located in the CNS. In addition to sensory cells, there are also ciliated epithelial cells (CEC) and microvillous epithelial cells (MEC). From Zaitseva and Bocharova (1981). (B) Classification of dendritic endings according to Dorsett (1986, adapted), c, cilia; mv, micro villi.

towards the CNS. In other cases the receptor cells appear to communicate only with other neural elements lying within peripheral tissues, that is, with processes of interneurons and motoneurons. In this way, local circuits are formed that are capable of mediating reflex responses. Collateral processes on some axons allow for both peripheral and central connections. Many axons of receptor cells have prominent varicosities after Golgi silver impregnation (Fig. 3.4A). Most of the receptor endings bear cilia, of which there may be as few as one or as many as 40. The cilia vary in thickness and length, with

42 Behavior and Neurons in Gastropods

some as long as 30 um. There are also large variations in the rootlets, with some extending as long as 20 um into the dendrite (Fig. 3.4B, type Ib). The arrangement of microtubules is usually standard, namely nine peripheral pairs plus two central singlets. Microvilli often appear together with cilia, or even without cilia, and these also vary considerably in length, thickness, and branching. Dorsett (1986) classified the endings of epithelial receptors according to ultrastructure, based on the types that are broadly encountered in various organs representing many different taxa (Fig. 3.4B). His class I endings bear one or more cilia emerging from a slightly domed membrane at the extremity of the dendrite, plus microvilli in some cases. Class II endings are distinguished by a depression of the apical membrane from which variable numbers of cilia and microvilli emerge; in contrast to class I, the cilia have no rootlets. Class III and class IV endings have only microvilli; they differ from each other in the number and length of the microvilli. In the absence of electrophysiological data, workers have speculated on function based on structure alone. For the most part, attention has focused on the receptors with ciliated endings (Emery, 1992). Some authors assume that short cilia must be mechanosensory because their small surface area could capture few odor molecules, while their short stature should confer mechanical stiffness. Others argue that mechanoreception is indicated by cilia with long roots or single cilia surrounded by specialized microvilli (stereocilia). In any case, if ciliated endings are indeed mechanosensory, then endings that have only microvilli, or which are dominated by long microvilli, must be chemosensory. Such arguments typically rest on comparisons with known mechanoreceptors in non-molluscs, or on the prevalence of a receptor type in skin not noted for its chemosensitivity, but none of the identifications has been clearly established in gastropods. One likely exception to the above scheme is found in certain receptors of "prosobranchs," where the cilia have dilated or paddle-like tips ("discocilia") suggestive of a surface area expansion for chemoreception (Haszprunar, 1985). On the other hand, the relatively common endings known as "free-nerve endings" that terminate within the epithelium are probably mechanosensory because their processes should be inaccessible to odor molecules and because they probably belong to centrally located cell bodies, which are known to be mechanosensory. No morphological feature has been identified to distinguish ordinary mechanoreceptors and nociceptive receptors (nociceptors). Physiologically, mammalian nociceptors are characterized by high thresholds for activation and graded responses for suprathreshold responses over a range of stimulus strengths; maximal responses are recorded for stimuli that are injurious. By these criteria, certain clusters of mechanically sensitive cells in Aplysia are identified as nociceptors (Section 9.3.3; Illich and Walters, 1997), but it is not clear whether these cells are excited by noxious chemicals, such as salt, acid or predator-specific substances.

Sensory Systems 43 3.4.

Photoreception

Most gastropods have a pair of eyes, although a few genera have none at all. The eyes may be situated at the bases of tentacles (basommatophores and opisthobranchs), on short stalks ("prosobranchs"), or at the tips of retractable tentacles (stylommatophores). In some nudibranchs (e.g., Hermissenda), the eye is situated on the surface of the brain and the photoreceptors are only about 100|im away from central neurons. Although gastropod eyes have been examined morphologically, and more recently in studies of circadian pacemaker generation, their functions in the control of behavior have received much less attention, a point that has been forcefully made by Hamilton (1991). One function of the eyes is to mediate phototaxis during locomotion, a topic which is discussed in Section 6.4.4. Another function is to regulate the timing of behavioral activities on both a daily and a seasonal basis, as discussed in Chapter 10. However, the eyes are not the only organs capable of detecting photoperiod. Extraocular photoreceptors are present in the skin epidermis in many regions of the body, but especially in the tentacles, rhinophores and siphon (Stoll, 1976; Chase, 1979). In Aplysia, the extraocular photoreceptors are capable of entraining the circadian locomotor rhythm even after the eyes have been removed (Lickey et al., 1977). It can also be noted that some neurons in the central ganglia show direct membrane potential responses when exposed to light. The most common response is a K+-dependent hyperpolarization that is probably mediated by light-sensitive carotenoid pigments. For species with unpigmented or thin skins, enough sunlight might penetrate though the skin to affect nervous activity in some circumstances. The extent to which gastropod eyes may mediate true vision (i.e., the detection of forms) is still open to investigation. One remarkable eye that needs to be studied more in this respect is that of the heteropod "prosobranchs." The heteropods are pelagic animals with a predatory lifestyle. Their eyes are large relative to the body length (4mm in an animal of 15cm), and the lens is well separated from the retina, which allows for good optical resolution. When Land (1981) reconstructed the retina in one animal, he found an array of receptors comprising six rows and 286 columns laid out on a flatbed surface. Each column contains a stack of five receptors piled on top of each other to a total depth of lOOum. Consequently, light reaching the retina passes through, and presumably excites, all five receptors. There are other peculiarities, including stacked membrane disks in the receptive segments of the photoreceptor cells, as in vertebrate photoreceptors, and prominent grooves in the retina and the adjacent epithelium. One interpretation of these unusual structures is that they provide information on the distance of prey objects (Fretter and Graham, 1994), but Land (1981), who made a detailed study of the optics, found no ready explanation except to suggest that they are probably used for sighting prey. Regrettably, despite the indications of an extraordinary eye, the visual capabilities of heteropod molluscs have never been studied under controlled conditions.

44 Behavior and Neurons in Gastropods

Recent experiments with the freshwater snail Lymnaea stagnalis have demonstrated an ability to distinguish between a black and white check pattern and a gray background that is matched to the check pattern in overall luminance (Andrew and Savage, 2000). The ability of the snails to discriminate visual stimuli was used by the experimenters to study associative learning. Animals learned to feed when they saw the check pattern but to withhold feeding when they saw the gray background (see Section 7.6). Interestingly, while the results clearly depended on the snails' ability to resolve the checkered pattern and to discriminate it from the gray background, only the checked pattern was effective as the conditioned stimulus; the animals did not learn to respond to the gray background while withholding responses to the checkered pattern. Perhaps the best evidence for form vision comes from the common marine snail Littorina, also known as the periwinkle snail. The optics of the eye provide excellent resolution. In L. irrorata the average separation between photoreceptors is 3.44um and the angular separation of neighboring photoreceptors is 1.7°(Hamilton, 1991). By comparison, the blur circle produced on the retina by the convergence of rays parallel to the optic axis is 1.1 urn. Periwinkles normally inhabit the upper intertidal zone amid vertical plant stems. If the animals are displaced from this region, they move towards the area that contains the plants, as opposed to moving horizontally along the beach or towards the shore. The snails can also locate the closest vertical plant stem if they are released on a substrate surrounded by plants. In both cases, appropriate control experiments indicate that the behavior is guided by visual cues. For standardized tests of visual detection, Hamilton and Winter (1984) used black shapes drawn on white test cards (see also Hamilton, 1991). Since most snails prefer to move towards darkness and away from light (negative phototaxis; Section 6.4.4), visual acuity could be measured by their attraction to the test cards. The experiments showed that Littorina irrorata begins to respond to vertical bars when they subtend 0.9°. The archaeogastropod Turbo castaneas performed even better. This snail, which inhabits shallow sea grass beds, has a threshold of only 0.3°, and it can discriminate horizontal and vertical stripes. Altogether, Hamilton and Winter (1984) studied four species, and found a good correlation between detection thresholds and anatomical resolution in the eyes. Apart from the heteropods, discussed above, most gastropod eyes conform to a common structural form, as illustrated in Figure 3.5 with the eye of the bubble snail Bulla. Details of structural variations in different species can be found in Dorsett (1986) and Hamilton (1991). Generally, the entire eye is enclosed by a collagenous capsule. The cornea is overlaid by a thin layer of skin, and the lens, which may be soft or hard in different species, is usually spherical. The cup-shaped retina may lie close to or distant from the lens. When the retina is distant from the lens, a vitreous body fills the intervening space. An optic nerve joins the eye with the cerebral ganglion. The cellular composition of the retina shows a great deal of variation between species but the functional significance of these differences has

Sensory Systems

45

Figure 3.5. The eye of the bubble snail Bulla gouldiana as photographed in a thick histological section. Approximately 100 large photoreceptors (PC), known as type R cells, surround the lens. The basal retinal neurons (BRN) surround the neuropil (NP) near the origin of the optic nerve. It is not certain whether the BRNs are themselves photosensitive or whether their responses to light depend on receiving inputs from primary photoreceptors. In any case, the BRNs are very interesting because each cell of this type is a circadian pacemaker, as described in Section 10.4. The axons of photoreceptors and BRNs are present in the optic nerve. From Block et al. (1993) with permission.

hardly been addressed (see Hamilton, 1991). The greatest number of photoreceptors, nearly 100,000, has been reported in the large conch snail Strombus, whereas certain nudibranchs (e.g., Hermissenda, Tritonid) have as few as five photoreceptors. In most eyes, there are several types of photoreceptors, and much effort has been devoted to describing the types in different species. Classification focuses on the ultrastructural specializations that are present at the luminal surface because phototransduction is assumed to occur here. Most gastropod photoreceptors are microvillous, and the individual microvilli are loosely bundled to form rhabdomeres; some photoreceptors bear cilia in addition to microvilli. The eyes of two opisthobranchs, Aplysia and Bulla, have been studied in detail because each one operates as a circadian clock (Section 10.4). In most

46 Behavior and Neurons in Gastropods

respects, the eyes are very similar in the two species. There are four or five photoreceptor types, and each type is found in a characteristic location, either dorsally near the lens or ventrally near the origin of the optic nerve (Fig. 3.5). One type of photoreceptor (R cells) is large and lies close to the lens. The R cells have elaborate microvillous distal segments, in contrast to the other photoreceptor types, which are primarily ciliary. Whereas the R cells are depolarized by light, other smaller cells in the same region, namely the H cells, are hyperpolarized by light. Both the R cells and the H cells send axons into the optic nerve. Another cell type, known in Bulla as the basal retinal neurons (BRNs), also sends axons into the optic nerve. Although the BRNs respond to light, this may be a secondary response driven by input from the R cells. In any case, the BRNs produce action potentials spontaneously, in both light and dark. There are about 100 BRNs in Bulla and about 1,000 equivalent cells in Aplysia (Jacklet, 1989; Block et al., 1993). Because the BRNs are strongly coupled to one another by electrotonic junctions, their synchronous firing creates compound action potentials in the optic nerve. The BRNs are especially interesting because within each one of them lies a circadian oscillator. The number of action potentials that they generate varies with the time of day (Section 10.4). In the terrestrial pulmonates Helix and Limax, two types of photoreceptors have been identified based on the ultrastructure of the microvilli and the presence or absence of microvesicles. A physiological correlate of these two receptor types was found by recording extracellular electroretinograms (Suzuki et al., 1979). Two spectral sensitivity curves were recorded, one associated with dark adaptation (peak at 480 nm) and the other with light adaptation (peak at 460 nm). This duplex system may allow the animals to perceive relative light intensities over a larger range than would otherwise be possible. The most thorough analysis of synaptic interactions in a gastropod retina has been carried out in the nudibranch Hermissenda. The eye of this animal measures only 75 urn in diameter and contains only five photoreceptors (Fig. 3.6A; Alkon, 1980). Two of the receptors are type A cells; they have weak interactions with each other and with type B cells. The three type B cells mutually inhibit each other and they also inhibit the type A cells (see Fig. 6.11). Functionally, the type A cells respond best to highintensity light, while the type B cells respond best to low-intensity light. Because each photoreceptor has its own receptive field, the inhibitory synaptic interactions between type B cells amplify the differences between the electrical responses of light-stimulated and non-stimulated receptors, thereby aiding in the detection of light-dark edges. This has behavioral significance because the animal avoids dark borders. Further processing of visual signals occurs in a small optic ganglion, which measures only 40|am across and contains just 14 neurons (Fig. 3.6A). Activity in type B photoreceptors causes inhibition in 13 of the ipsilateral optic ganglion neurons, while the type A neurons apparently synapse not in the optic ganglion but only in the brain. The majority of the optic ganglion neurons also project

Sensory Systems

47

Figure 3.6. Photographs of the statocyst as it appears in a fresh (unstained) preparation from the nudibranch Hermissenda crassicornis. (A) The dorsai surface of the brain showing the statocyst (S) lying next to the eye (E) and the optic ganglion (OG). Other labels indicate the cerebral ganglion (CG), the pleural ganglion (PG) and the pedal ganglion (PdG). (B) The statocyst shown at a higher magnification. The statoconial mass (SC) comprises many individual calcareous stones. It is contained within the statocyst wall (SW) and surrounded by fluid. The sensory hair cells are not visible in this photograph. Scale bar is 20mm in (A), 20 um in (B). Adapted from Detwiler and Alkon (1973) by copyright permission of The Rockfeller University Press.

to the brain, but one cell in each ganglion, the C cell, projects to the contralateral optic ganglion where its dominant postsynaptic action is inhibitory. Additional second-order visual neurons are found within a small region of the cerebropleural ganglion. Interestingly, each type of photoreceptor, whether A or B, medial or lateral, seems to project to a different set of cerebropleural neurons (Crow and Tian, 2000). This arrangement suggests "labeled-lines" that could allow each photoreceptor type to have a different influence on behavioral functions. 3.5.Gravireception

Gastropods sense gravity by the use of statocysts, which are simple spherical chambers formed by an epithelial wall that contains ciliated receptor cells

48

Behavior and Neurons in Gastropods

(hair cells) and supporting cells (Fig. 3.6). The interior of the cyst is filled with fluid and one or more calcareous stones, called statoconia, that are slightly more dense than the fluid. In species that have many statoconia (Aplysia has about 1,000), the statoconia move around in the fluid as a mass. In other species, as well as in immature animals which later develop statoconia, there is a single large stone known as a statolith. Regardless, as the stones are pulled by gravity against the cilia, a nervous signal is generated, as shown in Figure 3.7. Since signals are produced by rotations around the horizontal axis but not around the vertical axis, and since no response is given by either acceleration or deceleration, the statocyst is a pure gravity sensor. Each animal has a pair of statocysts, and these occupy an unusual place for a sense organ—within the connective tissue sheath that surrounds the pedal ganglion. In opisthobranchs and pulmonates they are located on the dorsolateral surface close to the pleural ganglion (Figs 3.6A and 3.7), while in "prosobranchs" they are on the ventral side of the pedal ganglion. The receptors are primary sensory cells that send their axons into a nerve, called the "static nerve" or the "acoustic nerve," that connects to the cerebral ganglion. The nerve is very short in opisthobranchs (about 30 (am in Hermissenda), but much longer (about 5mm) in pulmonates, owing to the fact that in the latter animals the pedal ganglion lies below the esophagus and the cerebral ganglion lies above it. The largest statocysts are found in

Figure 3.7. The statocysts of the planktonic species Clione limacina and their responses to tilting. (A) The location of the statocysts and the morphology of a receptor cell. Abbreviations: St, statocyst; SRC, statocyst receptor cell; Cer G, cerebral ganglion; PI G, pleural ganglion; Pe G, pedal ganglion. (B) Schematic drawing of the device used to tilt the experimental preparation while recording from receptor cells with a microelectrode (ME). For the traces shown in this figure, rotations were made around the horizontal axis (Axl). (C-E) Examples of responses to tilting. The initial positions and the final positions of the recorded cells relative to the position of the statolith are indicated. Note that the cells are active only when they are underneath the statolith. Adapted from Panchin et al. (1995b).

Sensory Systems

49

"prosobranchs," where they have diameters 250-500 um compared to 80-200 jam in opisthobranchs and pulmonates. "Prosobranch" statocysts have hundreds or even thousands of small hair cells, while those of pulmonates and opisthobranchs have only 9-14 hair cells (Wolff, 1975). Correspondingly, the hair cells of "prosobranchs" are small (6-9 um), while those of pulmonates and opisthobranchs are large (up to 150 jam) and disk-like in shape. Except for the occasional interposed supporting cell, the luminal surface of the statocyst is entirely covered with receptor cells and their projecting cilia. As many as 700 cilia extend from a single hair cell. The mechanism of sensory transduction is simple but interesting because it differs from that of the vestibular apparatus in vertebrates. Studies of hair cell physiology are easy to perform using the large opisthobranch hair cells. The cilia within the gastropod statocyst are of the kinocilia type and their micro tubules are characteristically arranged in a 9 + 2 pattern. By contrast, the cilia of the vertebrate otolith organ are typed as stereocilia; they are stiff structures packed with actin fibers but containing no micro tubules. Whereas transduction occurs at the tips of stereocilia, in kinocilia it occurs neither at the tip nor along the axoneme (the bundled microtubules), but at the base. In other words, the role of the cilia in kinocilia is to transmit the mechanical signal, not to transduce it. This conclusion was reached by Stommel et al. (1980) from experiments in which both the rigidity of the cilia and their movements were manipulated pharmacologically. When the cilia were rendered flaccid, gravitational stimuli failed to produce any generator potentials in the hair cell. On the other hand, if rigid cilia were inhibited from moving, transduction still occurred. The generator potential is produced by an influx of sodium across the hair cell plasma membrane. The mechanically sensitive ion channels are thought to lie near the junction of the cilia and the plasma membrane. Here, membrane particles and microtubule connections form a complex structure known as the ciliary necklace. The membrane in this region is probably deformed by stresses transmitted to it by rootlets radiating from sites where the cilia are anchored. The cilia beat constantly at a frequency of about 10 Hz. Stommel et al. (1980) proposed that the beating amplifies the mechanical signal by adding a reaction vector to the gravity vector. When a beating cilium encounters the statoconial mass, the force that the cilium exerts on the mass should be transferred by leverage to a point at the base of the cilium. Even as the gravitational stimulus persists, and the cilia remain stationary under the load of the statoconial mass, each beat pulse should exert a force and produce a transduction event. As predicted by this hypothesis, the electrical responses of the hair cells show no adaptation under conditions of constant loading, although a phasic response typically precedes the tonic activity (Fig. 3.7C). Individual hair cells can respond to rotations in any plane and in any direction around the horizontal axis, provided that they are contacted by the statoconial mass. Whether or not a given cell responds to a given stimulus depends on the position of the cell relative to the position of the statoconial

50 Behavior and Neurons in Gastropods

mass at the onset of rotation (Fig. 3.7C-E). In Aplysia, a ring of receptor cells located on the equator of the statocyst are crucial for detecting the direction of tilts (Janse, 1983). Janse estimates that five receptor cells constitute the equatorial ring, that each is responsive to rotation over approximately 70°, and that there is a functional overlap of about 30%. When the animal is in a horizontal position, the equatorial cells are just barely touched by the statoconial mass. Depending, then, on the direction of tilt, different cells will be stimulated. The resulting pattern of activity provides a code by which the CNS can initiate appropriate compensatory motor responses. An intriguing aspect of gastropod hair cells is their innervation. This was discovered by Detwiler and Alkon (1973) when they evoked generator potentials and spiking in hair cells of Hermissenda. They observed that many spiking events were followed by membrane hyperpolarizations. The source of hyperpolarization was traced to other hair cells in either the same statocyst or the contralateral statocyst. Because the probability of reciprocal inhibition was greatest for pairs of cells that were located at 180° to each other around the circumference of the cyst, it was suggested that the function of inhibition is to sharpen a difference signal for gravitational direction. In related work with the terrestrial slug Arion, Wolff (1975) recorded from the peripheral stump of the statocyst nerve after it was detached from its sensory organ. Activity in the nerve increased when the preparation was tilted provided the contralateral statocyst was left intact, implying that the nerve response was a centrifugal signal initiated in the contralateral statocyst. The result suggests the possibility of cross-talk between the two statocysts under normal conditions. Individual hair cells are evidently contacted not only by other hair cells, but also by peripheral or CNS elements associated with other sense organs ("efferent" signals). In Hermissenda, the hair cells respond to light stimuli as well as to chemical stimuli applied to the tentacles, and in both cases the responses seem to be mediated by synaptic inputs (Alkon, 1980). Conversely, some statocyst hair cells synapse on to photoreceptors (see Section 6.4.4 and Fig. 6.11). In Lymnaea, the hair cells are sensitive to mechanical stimulation of the entire skin but especially near the lips, the tentacles, and the mantle edge. Again, skin stimulation affects the hair cells through synaptic contacts, causing excitatory, inhibitory, or biphasic responses (Janse et al, 1988). Also, some hair cells in Lymnaea become hyperpolarized when the oxygen content of the water contacting the mantle edge is reduced; this could be important for regulating the animal's movements towards and away from the air-water interface (Section 6.4.3). Finally, from experiments in Clione, in which intracellular stimulation of certain cerebral neurons either excited or inhibited statocyst hair cells (Panchin et al., 1995b), and from similar experiments in Aplysia and Hermissenda, it is evident that peripheral sensory neurons are not the only sources of direct synaptic inputs to hair cells. These reports of apparent synaptic inputs on to hair cells leave at least two questions for further investigation. First, how do the efferent signals propagate to the statocyst? This question arises from observations that the number of axons in the statocyst nerve, at least in pulmonate and

Sensory Systems

51

opisthobranch species, is no greater than the number of hair cells (Wolff, 1975). Second, what is the function of the efferent innervation? It has been suggested that the gravitational response of hair cells may be modulated in certain behavioral or environmental situations. However, tilt-induced afferent discharges in the statocyst nerve appear to be unaffected by efferent activity (Wolff, 1975). Furthermore, in one case where gravitational responses of the whole animal are clearly affected by temperature, the locus of neural modulation appears to be central, not peripheral (see Section 6.4.3).

3.6.

Magnetoreception

The capacity of animals to detect the earth's magnetic field, and their use of this information for purposes of navigation, has been persuasively demonstrated in birds and fish. Somewhat less convincing evidence has been reported for a variety of invertebrates, including molluscs. It is hypothesized that cells containing permanently magnetized iron-rich crystals, magnetite, may constitute the magnetic sense organ. Magnetoreception was extensively studied in the mud snail Nassarius obsoleta by F. A. Brown and his colleagues (Brown et al., 1960, 1964). They observed the snail's angles of turning as they emerged from a corral into an open space. On some trials, a horizontal bar magnet was placed beneath the test apparatus. The main finding was that the presence of the magnet increased the amount of turning and caused the snails to turn more to the left, regardless of whether the magnetic field was oriented east-west or north-south. In addition, there were circadian and lunar rhythmicities in the magnetic responses. Still other experiments were said to demonstrate compass directional biases in the snails' behavior, for example, a preference to turn toward the east, and it was reported that these biases could be experimentally manipulated with magnets. While Brown's experiments with Nassarius are provocative, they produced variable results and small effects. Until the experiments are replicated by other workers, his conclusions will remain in doubt. The influence of magnetic fields on spatial orientation was studied in the sea slug Tritonia diomedea by dropping animals into a circular tank in complete darkness (Lohmann and Willows, 1987). The animals were left undisturbed for 17 hours. On alternate days, they either remained for an additional 90 minutes in the ambient geomagnetic field or they were exposed to a magnetic force that cancelled out the horizontal component of the ambient geomagnetic field. At the conclusion of the 90-minute experimental period, the body axis orientation was assessed. It was found that under ambient conditions the animals showed a strong preference for eastern headings (as in Brown's experiments), but this preference disappeared when the horizontal component of the geomagnetic field was cancelled. In a second series of experiments, in which Lohmann and Willows recorded turning responses in a Y-maze apparatus, the ratio of left turns to right turns was significantly influenced by a reversal of the geomagnetic field. Again, as in

52 Behavior and Neurons in Gastropods

Brown's experiments, the response to magnetism varied in a pattern that correlated with the lunar cycle. The possible use by Tritonia of geomagnetic cues for directed locomotion is discussed further in Section 6.4.5. A magnetic effect has also been reported in the terrestrial snail Cepaea nemoralis (Kavaliers and Ossenkopp, 1991). In this case, while the phenomenon is measured behaviorally, it does not necessarily imply a sensory process. The experimental paradigm involves placing snails on a surface that has been warmed to 40°C. Because the heat is aversive, the snails lift the anterior portion of their foot, and the latency of the response is measured. The influence of magnetic fields is evaluated in terms of the ability of morphine to attenuate the avoidance response. Thus, injections of morphine ordinarily induce analgesia, but the analgesic effect is reduced if the injection is accompanied by a slowly rotating magnetic field or a slowly oscillating magnetic field. However, when a pulsed magnetic field is used, an analgesic effect of the field itself is reported. These results are interpreted as implying an influence of magnetism on opioid signaling in the nervous system, perhaps mediated by calcium channels, G proteins or protein kinase C (Kavaliers and Ossenkopp, 1991). Still another twist is added by examining how light modulates the magnetic field effect (Prato et al., 1998). Since the effect is greater in darkness than in light, while the effect of naloxone (an opioid antagonist) is equal in both situations, the authors conclude that light exerts its modulatory action on the magnetic detection mechanism (i.e., prior to its coupling with the opioid system). In summary, a number of novel mechanisms are implied by these experiments with Cepaea. It remains to be seen whether the phenomena described are specific to opioid pathways, and if so, why.

4

Muscles and the Peripheral Nervous System 4.1.

Muscles and Muscle Cells

The renowned slow behavior of gastropods can be attributed to the lack of a hard skeleton, which limits the amount of amplification and antagonism that can be achieved from muscular actions. One of the few examples of direct muscular antagonism in gastropods occurs in the buccal mass where retractors and protractors operate to move the structure back and forth during feeding movements. However, to some extent hydrostatic pressure created by a partially open blood circulatory system compensates for the absence of a bony skeleton by creating what can be called a hydrostatic skeleton. Furthermore, muscular contractions can increase the flow of hemolymph to selected regions and thereby increase the pressure exerted against the body wall in those regions. This can cause movements of body parts to produce elongation and bending, protraction and eversion. For example, when the superior tentacle of Helix pomatia is everted, the blood pressure increases about 16cm in the tentacle, while surrounding areas experience little or no increase in pressure. Protraction of the buccal mass during feeding is also thought to be accomplished in this way, as is the elevation of the parapodial wings in opisthobranchs during swimming. Further, hydrostatic pressure can antagonize muscular actions, for example, in controlling the length and stiffness of a tentacle. The muscles of gastropods come in a variety of sizes and shapes. The most conspicuous are the retractor muscles, which form either thick ribbons or cylindrical bundles. Individual retractor muscles are responsible for withdrawing individual organs, for example, the tentacles, the penis and the gills. These movements can be surprisingly fast and vigorous. In snails, the retractors form a common muscle, the columellar muscle, which originates on the apex of the shell (the columella) and attaches to the foot and the body wall. The columellar muscle system of the vineyard snail Helix pomatia (also called the Roman snail and the escargot de Bourgogne), is shown in Figure 4.1. When the entire columellar muscle system is activated, the body is withdrawn into the shell. In opisthobranchs, most of which have no external shell, retractor muscles originate in the body wall. Even in pulmonates, 53

54 Behavior and Neurons in Gastropods

Figure 4.1. The columellar muscle system of the European vineyard snail, Helix pomatia, as drawn by Trappman (1916). The system comprises a set of radiating retractor muscles (retr.) anchored to the apex of the external shell (columella, col.). When individual muscles are contracted, for example, retr.t.mai., only certain structures are withdrawn, (in this case, the superior tentacles). When the entire system is activated, the whole animal withdraws into its shell. Several pharyngeal muscles (p.vn.s., lev.lab.ext., lev.phar.) and the cerebral ganglion (g.c.) are also shown.

certain retractor muscles, such as those of the buccal mass, originate in the body wall. Other muscles confined to the body wall itself are responsible for causing bending movements. The musculature of the body wall is thicker in terrestrial gastropods than in aquatic taxa because greater hydrostatic pressures are required to support the body in terrestrial animals. The muscles of the foot, used for locomotion in most gastropod species, are especially bulky. Finally, many of the internal organs, including the CNS and its nerves, are covered with a thin sheet of muscle mixed with connective tissue. It would be erroneous to conclude from the foregoing that all body movements must be caused either by the bulk flow of fluids or by the contraction of muscles, with some structures subject to the antagonistic actions of both forces. Another mechanism, the muscular hydrostat, has been proposed to

Muscles and the Peripheral Nervous System 55

operate based on an antagonism inherent in individual muscles (see Thompson et al., 1998). The hypothesis assumes a muscle that consists of fibers oriented in two or more orientations, that is essentially incompressible, and that has a constant volume. The idea is that, if the fibers of one orientation contract, thereby shortening the structure in one dimension, the structure must then lengthen in another (usually opposite) dimension. In principle, therefore, a muscle like the columellar could function not only as a retractor, but also as a protractor. Some authorities believe that the emergence of a snail from its shell and the eversion of organs such as the penis and the tentacles owes more to muscular hydrostats than to fluid inflation. Indeed, the morphology of the columellar muscle seems to meet the requirements of a hydrostat in that it is a densely packed three-dimensional array of muscle fibers and connective tissue fibers. Moreover, muscle and connective tissue fibers in the columellar muscle of "prosobranch" snails are arranged differently in different zones, suggesting to Thompson et al. (1998) that each zone might be specialized for a different function. The region near the origin of the muscle on the columella appears to function in retraction of the body, while the region just posterior to the head and extending ventrally appears to function in protraction of the snail and in elevation of the shell during locomotion. Individual muscle fibers are small, and they usually have a fusiform shape. The maximum diameter, up to about 20 um, is present in the central nuclear region, while the total cell length can reach up to to several millimeters. Although the fibers are typically unbranched, for example, in the accessory radula closer muscle (ARC) of Aplysia (Cohen et al., 1978), they are extensively branched in the vasoconstrictor muscles of the abdominal aorta in the same animal (Alevizos et al., 1989a). In the latter case, because the fibers branch in a direction perpendicular to the main axis of the muscle, it is hypothesized that the branching serves a signaling function, possibly to provide electrical coupling across muscle fibers in regions where they are not in close physical contact. Histologically, gastropod muscle cells have been described as "smooth" because they lack the precise alignment of Z discs and dense bodies that characterizes vertebrate skeletal muscle. However, molluscan muscles have abundant myofilaments that are often highly organized. The myofilaments are staggered to make oblique angles with respect to the longitudinal axis, often giving a spiral appearance in tissue sections. Thus, Hoyle (1983) prefers to call the molluscan muscles "obliquely striated," rather than "smooth." The myofilaments are predominately composed of paramyosin protein, with very little ordinary myosin. Gastropod muscles are slow to generate tension and even slower to relax, but they are able to reach high maximal tensions. Whether these physiological features are due to the oblique arrangement of the thick filaments or the predominance of paramyosin is unclear, as both are associated with gastropod muscles. Only a few muscles have been examined in any detail from a physiological or functional perspective (see Hoyle, 1983). Most of our information comes

56 Behavior and Neurons in Gastropods

from muscles of the buccal mass. However, some generalities apply to at least the larger of the internal muscles. The small size of gastropod muscle cells makes it difficult to perform physiological studies of individual cells in situ, but recent studies have utilized dissociated, functionally intact, muscle fibers (Brezina et al., 1994a). What follows is an overview of muscle physiology as it pertains to the neural control of behavior, with further details given in later chapters.

4.2.

Muscle Physiology

The majority of muscles are polyneuronally innervated and even individual muscle fibers are often innervated by several neurons (Cohen et al., 1978; McPherson and Blankenship, 199la). It is important to note that innervation is typically provided both from motoneurons with cell bodies in the CNS and from motoneurons with cell bodies in the periphery, as further discussed in Section 4.4. Synaptic inputs to individual muscle fibers are usually integrated across the entire muscle by electrical coupling between muscle cells (Cohen et al., 1978; but see Alevizos et al., 1989a, as discussed above). Most inputs are excitatory, but inhibitory junction potentials (IJPs) are also been seen in some muscles (Sawada et al., 1981). Kozak et al. (1996) have advanced the view that the functional role of IJPs is to limit the amount of depolarization and keep the membrane potential within a range that is optimal for controlling graded outputs. They discovered that the so-called excitatory transmitter, acetylcholine (ACh), activates simultaneously both inward currents and outward currents, the latter of which, they argue, is functionally equivalent to an IIP because it limits the amount of ACh-induced depolarization (see below). In addition to ACh, other known or putative excitatory neuromuscular transmitters include L-glutamate, dopamine, and the peptide, FMRFamide. Despite the fact that gastropod muscles show no morphological striation, the sliding filament mechanism of contraction still applies at the molecular level. An unusual aspect of muscle biochemistry in molluscs, contrasting with most other animal groups, is that the regulation of myosin-actin interactions is governed by the direct binding of calcium to myosin without any intermediate phosphorylating pathway. Several investigations have attempted to identify the source of calcium that triggers contraction. Although in vertebrates the usual source is internal cellular stores, the situation in molluscs is less clear. Possibly, L-type voltage-gated channels in the plasma membrane provide all the calcium necessary for triggering muscle contractions (see Scott et al., 1997). In general, the membrane currents of molluscan muscles resemble those of molluscan neurons (Brezina et al., 1994a; Laurienti and Blankenship, 1996; Scott et al., 1997). One significant exception is that most gastropod muscle cells lack voltage-sensitive sodium currents, although a few gastropod muscles generate calcium-dependent action potentials. Scott et al. (1997) made a detailed comparison of ion currents in two antagonistic muscles of Aplysia,

Muscles and the Peripheral Nervous System

57

one which spikes and one which does not. While the two muscles support similar currents, shown in Figure 4.2, there are subtle differences. Both the radula opener muscle, which spikes, and the radula closer muscle, which does not spike, have at least four outward currents that oppose depolarization to the threshold voltage for generating spikes. One of these, the 7k(A) current,

Figure 4.2. A comparison of ion currents in two antagonistic muscles of Aplysia, the accessory radula closer muscle (ARC) and the radula opener muscle. Two critical differences account for the fact that the opener muscle generates action potentials ("spikes") but the closer muscle does not. One difference is the reversal potential for the current induced by acetylcholine (^ACH); the other is the activation voltage for /K(A> which causes the combined curve for all potassium currents (7K) to be shifted to the right in the opener muscle. As a result of these differences, the ARC muscle has a smaller operating range than the opener muscle. The effects of modulatory substances are also shown, e.g., MM, myomodulin, SCP, small cardioactive peptide. From Scott et al. (1997) with permission.

58 Behavior and Neurons in Gastropods

activates at membrane potentials about 9mV more positive in the opener muscle than in the closer muscle, and the acetylcholine-activated chloride current is weaker in the opener muscle than in the closer muscle. These two differences are apparently sufficient to alter the balance between inward and outward currents, so that the critical voltage for spike generation is reached in the opener muscle but not in the closer muscle. Spikes cause the opener muscle to contract more strongly and in a more twitch-like manner than the closer muscle. In the absence of action potentials, contraction strength is graded in proportion to the degree of membrane depolarization (Cohen et al., 1978; McPherson and Blankenship, 199la). In muscles that are capable of producing action potentials, the spikes are variable in amplitude and they do not usually overshoot the resting membrane potential. These phenomena suggest an active suppression of the spike-generating mechanism. It has already been mentioned that ACh induces outward currents as well as inward currents. Direct evidence for the suppression of spiking has come from the study of ion currents in dissociated muscle cells obtained from two muscles that do not normally exhibit spiking, the buccal mass retractor muscle of Philine aperta (Dorsett and Evans, 1991) and the accessory radula closer muscle of Aplysia californica (Brezina et al., 1994a). When clamped with depolarizing voltages, fibers from both muscles generated three separate potassium currents. Significantly, when the K + currents were pharmacologically blocked, the latent spiking capability of the muscle fibers was unmasked. The authors of these papers believe that spiking is undesirable because it would produce pulsatile contractions where a finer graded control is required. In gastropods, all the fibers of a muscle are innervated by the same few motoneurons and the muscle response is integrated by electrical coupling. By contrast, vertebrate muscles comprise many independent motor units. Thus, while vertebrate muscles can achieve graded responses through a variable number of motor units being activated in an all-or-none manner, molluscan muscles usually produce graded responses by synaptic actions that are generated in parallel and integrated across fibers. The synaptic inputs to muscle cells are integrated both spatially and temporally. With repeated stimulation of the inputs, frequency-dependent effects such as facilitation, depression and post-tetanic potentiation are commonly observed (Cohen et al., 1978; McPherson and Blankenship, 1991a). The usedependent plasticity of neuromuscular transmission allows for the integration of neural signals beyond simple summation, and it further reflects the sensitivity of the contractile response to the membrane voltage. Muscle activity is also commonly modulated through presynaptic or postsynaptic mechanisms. Numerous transmitter substances have been implicated in muscle modulation including serotonin, dopamine, and several peptides. Some effects of modulation on muscle cell ion currents are shown in Figure 4.2. Because most of the detailed investigations of muscle modulation have been carried out on the feeding muscles comprising the buccal mass, this subject is treated more fully in Section 7.5.6.

Muscles and the Peripheral Nervous System 59 4.3.

Peripheral Contributions to the Control of Reflexes

Whereas the CNS is defined as comprising the ganglia and connective nerves in the interior of the animal (Figs I VII, 2.1, 2.3 and 2.4), the peripheral nervous system (PNS) is defined as comprising all other neurons and neural processes, whether situated in smaller peripheral ganglia, in nerves connecting the CNS with tissues or organs, or in tissues that are not predominately nervous. Most behaviors are initiated within the CNS and they are primarily controlled by the CNS. The PNS has sensory and motor functions but it has little capacity for coordinating or integrating behaviors. A good indication of the role of the PNS is given by the fact that some mechanically elicited reflexes can be executed in the absence of the CNS even though, as noted in Chapter 3, many mechanoreceptor neurons have centrally located cell bodies. This is possible because there are nerve cells and neural circuits in the periphery that are fully capable of mediating reflexes, at least in some areas of the body. Although there is an abundant early literature on the morphology of the peripheral nervous system (see Bullock, 1965), many details of its organization remain unknown. Various terms have been employed to describe the PNS. Bullock (1965) often refers to the peripherally located neural "plexus," by which he means a layer of nerve fibers comprising a nerve net, a tangle of sensory endings, or a collection of ganglion cells. The term "nerve net," according to Bullock, denotes a specific type of plexus, namely, a system of neurons dispersed in a plane and connected by fusion or synapses. A good example of a molluscan nerve net is found in the sole (foot) plexus of terrestrial snails, as illustrated in Figure 4.3. Here, the plexus provides the neural substrate for the propagation of peristaltic contractions during forward locomotion (see Section 6.2). Other authors have been less fastidious than Bullock, with the result that the words "plexus" and "net," as well as other similar terms, are often used vaguely where precise knowledge is lacking. When considering the relative control of the CNS and the PNS on reflex movements, it is useful to bear in mind the distinction between local reflexes and remote reflexes (Perlman, 1979). A local reflex is one exhibited by the organ that is stimulated; a remote reflex is one elicited by stimulation of a site different from that which exhibits the movement. For local reflexes, a large degree of peripheral control permits the expression of reflex actions even in the absence of the CNS. For remote reflexes, the situation is less predictable. If the response site is very distant from the stimulus site, mediation by the CNS is no doubt required, but if the stimulus and response sites are close together, and especially if they are related functionally, they may be linked by peripheral neural connections. The relative contributions of the CNS and the PNS can be assessed by comparing reflex amplitudes before and after lesions of the CNS. The lesions may be permanent, as when a peripheral nerve is cut, or they may be reversible, as when synaptic transmission is blocked by bathing the CNS in either a sucrose solution or a high Mg ++ /low Ca++ saline solution. Several local reflexes have been studied with these techniques. In the case of the siphon

60 Behavior and Neurons in Gastropods

Figure 4.3. Innervation of the foot in Helix pomatia. (A) The radiation of peripheral nerves from the pedal ganglia, shown here in a ventral view of the foot. (B) The nerve net, or "plexus," in the sole. The peripheral nerves shown in (A) are continuous with the nerve net; they enter the net at the places marked as "nerve entrances." Nodes are occupied by small ganglia. Adapted from Schmalz (1914).

withdrawal reflex of Aplysia, the results show that the CNS contributes about 55% to the total reflex amplitude (Perlman, 1979; Antonov et al., 1999). Roughly the same contribution of the CNS is found in the tentacle withdrawal reflex of Helix, except that here the CNS contribution increases with increasing stimulus strength, giving a range of contributions from 25 to 55% (Prescott and Chase, 1996; Prescott et al., 1997). Since these examples come from just one opisthobranch (Aplysia) and one pulmonate (Helix), it is interesting to note that "prosobranchs" evidently lack any peripheral control over comparable reflexes in the absence of the CNS, according to Bullock (1965), citing earlier sources. A failure to recognize the distinction between local and remote reflexes was in part responsible for a controversy about the mechanisms that control the gill withdrawal reflex of Aplysia. In 1969, Kupfermann and Kandel published an initial description of neuronal elements in the parietovisceral (abdominal) ganglion that were said to mediate the gill withdrawal reflex. This finding

Muscles and the Peripheral Nervous System 61

initiated an intensive study of the central nervous mechanisms responsible for the reflex and for some simple forms of behavioral plasticity, namely habituation and dishabituation, that are associated with the reflex (see Section 9.4). Soon after Kupfermann and Kandel's paper appeared, however, Peretz (1970) reported that the gill withdrawal reflex, as well as habituation and dishabituation, could be elicited even in the absence of the CNS. Kupfermann's group responded by pointing out that the reflex studied by Peretz involved only a single pinnule, not the entire gill (see Kandel, 1976). The pinnules, of which there are 16 in A. californica, are elementary components of the gill. To elicit the pinnule withdrawal reflex, a weak tactile stimulus is applied to a pinnule, which then contracts. Therefore, the pinnule withdrawal response is a local reflex. By contrast, the gill withdrawal reflex, as commonly referred to, is a remote reflex because it involves tactile stimulation of the siphon (or the mantle shelf). By testing the pinnule reflex in preparations that either included the CNS or excluded it, the conclusion was reached that the peripheral circuitry is both necessary and sufficient for the pinnule reflex (see Kandel, 1976), but controversy continued over the gill withdrawal reflex. The contribution of the CNS to the remote gill withdrawal reflex in Aplysia has been investigated many times (for reviews, see Kandel, 1979; Mpitsos and Lukowiak, 1985). Although early studies by Kandel and colleagues tended to ignore any possible contribution by the PNS, other workers reported that the amplitude of the response after removal of the CNS was, on average, no different than when the CNS was attached (Peretz et al., 1976). Subsequently, several factors were identified that influence the amount of central participation in the reflex, including the extent to which the experimental preparation is "reduced" from the intact state, the device used to deliver tactile stimulation (electromechanical tapper, servo-controlled probe, or water jet), the device used to measure gill withdrawal (photocell vs. strain gauge), the stimulus intensity, and the response criteria (Carew et al., 1979). Kandel's group chose to work under conditions that maximize the CNS contribution, principally the use of a minimal response amplitude criterion. Meanwhile, an opposing group, led by Peretz and Lukowiak, drew attention to conditions that maximize the PNS contributions, namely, stimulation of the siphon at its base rather than at its rim and the acceptance of all responses regardless of their amplitude. After performing experiments specifically directed towards identifying the sources of the different results obtained by the two groups, Carew et al. (1979) reported that the CNS contributes at least 85% of the total reflex in reduced preparations using experimental procedures favored by Kandel's laboratory. Also, with intact animals, Carew et al. (1979) found that the CNS mediates a similar 90% of the reflex if a stimulus of "moderate" intensity is used and very small responses are ignored. Nevertheless, 10 years later, Leonard et al. (1989) would still write that "[the] monosynaptic reflex arc through the parietovisceral ganglion is not necessary, and has not been shown to be sufficient, to produce any of the various types of gill withdrawal reflex" (p. 602). Eventually, a new preparation was developed for laboratory investigations in which the siphon is surgically isolated from the gill, except

62 Behavior and Neurons in Gastropods

for nerve connections via the CNS, thus ensuring that the CNS mediates the reflex (Cohen et al., 1997). For reflexes triggered in the intact animal over the full range of stimulus intensities, different interpretations of CNS versus PNS control are still possible. It is worth noting that stimulus intensity affects the CNS contribution differently in the tentacle withdrawal reflex of Helix and the siphon-elicited gill withdrawal reflex of Aplysia. In the former case, increasing stimulus intensity recruits the CNS (Prescott and Chase, 1996), whereas in the latter case increasing stimulus intensity causes a reduction in the CNS contribution (Carew et al., 1979). However, it must be remembered that the tentacle withdrawal reflex is defined as local, whereas the siphon-elicited gill withdrawal reflex is denned as remote. Perhaps, as the strength of siphon stimulation increases, larger siphon withdrawal responses are elicited; these, in turn, might excite additional peripheral pathways connected to the gill, thus reducing the relative contribution of the CNS. Unfortunately, there have not yet been any reports in Aplysia describing how stimulus intensity affects the CNS contribution to local withdrawal reflexes, for example, the siphon withdrawal reflex.

4.4.

Cellular Elements and Plasticity in Peripheral Neural Circuits

Because the PNS is difficult to study, we lack a detailed knowledge of its cellular components. It is clear, however, that nerve cell bodies can be found throughout the periphery. To study this neuronal population, Xin et al. (1995) applied radioactive amino acids to the skin of Aplysia in several body regions. After allowing time for the amino acids to be incorporated into macromolecules, and for their transport from cell bodies into nerve fibers, the distribution of labeled fibers within the central ganglia was studied by autoradiography. The results indicated that all regions of the skin contain substantial numbers of cell bodies and that they have fibers projecting centrally. Presumably, these peripheral neurons are sensory in function, and they include the subepithelial receptor cells described in Chapter 3. Muscles can also harbor nerve cells, as, for example, in the gill of Aplysia (see Peretz et al., 1976) and the penis of Helix (Eberhardt and Wabnitz, 1979). Still other neurons reside in peripheral ganglia. These ganglia may be large and well developed, as in the tentacles of pulmonates where the ganglia contain about 5,000 neurons (Fig. 3.3), or they may be small, as in the branchial ganglion of Aplysia, which contains only a couple dozen neurons. In addition to the tentacles and the gills, ganglia are also commonly associated with the reproductive organs, the rhinophores, the eye, and the osphradium. Some ganglia have purely sensory functions, others purely motor functions, and some have mixed sensory and motor functions. Another place where neurons are frequently found is along the peripheral nerves. Sometimes the neuronal somata are present within the nerve together with the axons; in other cases the somata lie outside the nerve per se but

Muscles and the Peripheral Nervous System 63

within a common sheath (Bailey et al., 1979). Nerve branch points are frequently occupied by nerve cell bodies. In the sole (foot) plexus of Helix (Fig. 4.3B), small ganglia are distributed throughout the neural net at junctions and along small nerve fibers. It is not difficult to imagine an evolutionary history in which scattered peripheral cells coalesced into junctional aggregates and eventually into ganglia. The most thorough study of peripheral neurons is the investigation by Bailey et al. (1979) of the siphon skin in Aplysia. Here, neurons lie along the siphon nerve, as shown in Figure 4.4A. One cell type (type I) is represented by a single large cell that has a diameter of about 150 um. This cell is presumed to be neurosecretory on the basis of its white color under epiillumination, its spontaneous and highly regular pattern of firing, and its morphological detail. The remainder of the neurons (type II) number 30-50 in different preparations. They have cell bodies about 50-100 um in diameter, and overall morphologies that include regions of rich arborization, probably dendritic in function, as well as unbranched neurites. Intracellular stimulation of these cells elicited contractions of the siphon, suggesting that the type II cells are motoneurons. They receive considerable synaptic input, and most of them are electrically coupled to one another. The presence of sensory neurons, motoneurons and interneurons in both the CNS and the PNS allows for several different kinds of reflex circuits. To simplify the possibilities, highly schematic drawings, without interneurons, are shown in Figure 4.5. On the left side of Figure 4.5A we see a simple reflex circuit completely contained in the periphery; on the right is a single motoneuron within the CNS. This arrangement can account for the peripheral control over the pinnule withdrawal response in the gill of Aplysia (Peretz et al., 1976), while also permitting involvement of the CNS under some conditions. Figure 4.5B depicts circuitry that can explain control of the local siphon withdrawal reflex in Aplysia. Here, peripheral motoneurons are excited by central mechanoreceptor cells through monosynaptic connections (Bailey et al., 1979). Since these same mechanoreceptors also excite central motoneurons (see Section 9.3.3), they evidently control the siphon musculature through parallel downstream pathways. A somewhat different example of parallel motor control is found in the gill of Aplysia, where at least one central neuron, L7, synapses directly with gill muscle cells as well as with a peripheral motoneuron located in the branchial ganglion (Kurokawa et al., 1998). The tentacle withdrawal reflex of Helix represents still another type of arrangement (Fig. 4.5C; see also Fig. 9.2). In this case, there is no evidence for peripheral motoneurons but, nevertheless, the local reflex still can be elicited after a CNS lesion (Prescott and Chase, 1996). To explain this result, it has been proposed (Bullock, 1965; Prescott et al., 1997) that the peripheral reflex pathway depends on sensory signals being transmitted from sensory elements in the tentacle tip to peripheral parts of axons belonging to central motoneurons. The central motoneurons seem to be used only for remote reflexes. In the early 1970s, much attention was given to whether peripheral reflex circuits, especially in Aplysia, are capable of simple forms of learning.

64 Behavior and Neurons in Gastropods

Figure 4.4. Peripheral neurons in the siphon of Aplysia californica. (A) The approximate locations of nerve cells shown in relation to mantle cavity organs and the siphon nerve. The type I cell is a unique individual and it is probably neurosecretory. The type II cells, of which there are about 50, are smaller than the type I cell. (B) Evidence for parallel connections of central mechanoreceptors with central and peripheral motoneurons. (1). Mechanical stimulation of the skin triggers synaptic excitation first in the peripheral motoneuron, then in the central motoneuron. (2). Intrasomatic stimulation of a sensory neuron results in a reversed order of excitation (i.e., first in the central motoneuron, then in the peripheral neuron). These data, and others, imply the neural circuit for siphon withdrawal shown in Fig. 4.5B. From Bailey et al. (1979) with permission.

While Kandel and his colleagues were making major advances in the investigation of habituation learning by studying central neural mechanisms, the experiments of Peretz, Lukowiak and Jacklet showed that the isolated PNS was equally plastic (for review, see Mpitsos and Lukowiak, 1985). First, Peretz (1970) reported that habituation and dishabituation of the pinnule withdrawal response occurred in the absence of the CNS. Next, the locally

Muscles and the Peripheral Nervous System 65

Figure 4.5. Three kinds of neural circuits that can underlie peripheral motor competence in the absence of the CNS. The schemes shown here are based on interpretations of experimental studies that are cited in the text. (A) Dual control of the motor response by a local PNS circuit and a central motor pathway. Based on the gill of Aplysia and the penis of Helix. (B) Parallel outputs from central mechanosensors provide conjoint central and peripheral motor control. Based on the siphon of Aplysia. (C) Hybrid circuit in which peripheral sensory neurons contact axons of central motoneurons. Based on the tentacle of Helix (see also Fig. 9.2). M, motoneuron; S, sensory neuron.

elicited siphon withdrawal reflex was shown to habituate to a photic stimulus (incandescent white light) and to dishabituate to a tactile stimulus, and vice versa (Lukowiak and Jacklet, 1972). These findings initially impacted on the concurrent investigations of central nervous mechanisms, but their influence attenuated when experiments demonstrated that remote reflexes, at least, use primarily central pathways (Carew et al., 1979; Cohen et al., 1997). Meanwhile, workers continued to study the peripheral mechanisms of plasticity. With intracellular recordings, Bailey et al. (1979) were able to show that homosynaptic depression (the physiological correlate of behavioral habituation) has similar kinetics whether between sensory neurons and central motoneurons, or between sensory neurons and peripheral motoneurons. In the gill, Peretz and colleagues used progressive lesions to demonstrate that an isolated pinnule will habituate to repeated tactile stimulation, but dishabituation of the response requires an intact peripheral ganglion. Since the stimulus for dishabituation in their experiments was electrical stimulation of the ctenidial nerve, it would appear that dishabituation of depressed pinnule synapses is initiated by central commands but relayed through, or otherwise transformed in, the peripheral ganglion. Further light on the plasticity of peripheral synapses was provided by experiments on the locally elicited tentacle withdrawal reflex of Helix (Frescott and Chase, 1996). When a tactile stimulus was repeated at 4-minute intervals, the response initially increased in amplitude, then it gradually decreased. This result suggests the interplay of two plastic processes, one facilitatory (sensitization) and one depressive (habituation). When the experiments were repeated using tentacles that were isolated from the CNS, only habituation was seen. Evidently, synapses in the tentacle of Helix are capable of depression with repeated use, but they only express sensitization when it is induced centrally.

5

Regulation of the Internal Environment Some of the functions carried out by the nervous system are not strictly speaking behavioral because they do not involve observable movements. Visceral functions and physiological activities serving to maintain a constant internal environment, or homeostasis, fall into this category. In gastropods, the nervous system plays a major role in the regulation of respiration, blood circulation and water balance. It not only directly regulates these functions, it also coordinates them with ongoing behavioral activities. The central ganglia that are most involved in regulating the internal environment are the visceral ganglion and the paired parietal ganglia or, in Aplysia, the fused parietovisceral ganglion. Peripheral ganglia and isolated peripheral somata may also participate, depending on the organ being regulated.

5.1.

Respiration

The exchange of gases (principally O2 and CO2) between the animal and its external environment occurs at three different sites in gastropods. All gastropods respire at least partially through the skin. In addition, the marine forms have specialized structures, ctenidia in "prosobranchs" and gills (or branchia) in opisthobranchs, that are essentially outfoldings of the skin that create an enlarged surface area for gas exchange. The pulmonates, of course, are characterized by their lungs, which again constitute expansions of the respiratory interface. Regardless of the structure involved, whether the skin, a gill or a lung, each is highly vascularized to allow for the transfer of gases to and from the site of exchange. Oxygen is typically carried in the blood by one or more respiratory pigments, although some species have none at all. The most common oxygen-carrying pigment in gastropods is hemocyanin. Hemoglobin is used by some species, including several aplysiades and planorbids; myoglobin is also used in some species.

66

Regulation of the Internal Environment 67 5.2.

Control of the Lung in Pulmonates

The lung is a large sac, or cavity, enclosed within the mantle. Opinions differ as to whether it is a modification of the mantle cavity present in "prosobranchs" or a novelty appearing at a later evolutionary stage. The sac has a volume of about 6 ml in Helix pomatia. In all species it is richly supplied with blood vessels, which protrude into the volume of the sac and increase the respiratory surface. The interior location of the lung is an advantage not only because it gives protection against damage but, more importantly, because it minimizes the loss of water. The link between the lung and the ambient air is made through a small opening, the pneumostome, which is situated on the right anterior underside of the mantle. The pneumostome is ordinarily closed except when commanded open by neural signals. Most studies on the neural control of respiration have been undertaken on freshwater species, especially the pond snail Lymnaea stagnalis. The presence of a lung and the absence of ctenidia in freshwater pulmonates is indicative of the fact that these animals evolved from terrestrial forms, not from marine forms. Many freshwater species are primarily air breathing, while others are entirely aquatic. In the latter cases, the lung is continuously filled with water. Species in the widely distributed families Physidae, Planorbidae, and Lymnaeidae are intermediate in their utilization of the lung (i.e., their breathing is bimodal). They rely mostly on cutaneous gas exchange while submerged, but they also come to the water's surface to breathe air. These snails maintain a fairly constant rate of oxygen consumption over a wide range of dissolved pO2. They do so by respiring mostly through their skin when water pO2 is high, and mostly through their pneumostome, at the water's surface, when pO2 is low. Therefore, as the water pO2 declines they come to the surface with increasing frequency, a change in behavior that is possibly mediated by a functional switch in the output connections of the statocyst organs (see Section 6.4.3). Recent experiments have shown that the frequency of air breathing in Lymnaea can also be influenced by learning experiences in which aversive stimuli are associated with aerial breathing. Lukowiak and colleagues (1996) subjected Lymnaea to an operant conditioning paradigm in which a weak tactile stimulus was presented to the mantle tissue in the area around the pneumostome every time the snail attempted to breathe. As a consequence of these pairings, the animals learned not to breathe. They opened the pneumostome less frequently and for shorter durations than did control animals that received the tactile stimulus unpaired with pneumostome openings. When the snails were given five training sessions over 2.5 days, the memory persisted for 24 hours. With eight training sessions spaced out over 2 months, the memory persisted for as long as 4 weeks (Lukowiak et al., 1998). Preliminary studies of the neural changes associated with learning in this paradigm indicate that the memory is distributed among multiple sites within the neuronal network that is responsible for the behavior, as described below (see Benjamin et al., 2000).

68 Behavior and Neurons in Gastropods

The actual breathing behavior of Lymnaea, and its motor control, has been described by Syed et al. (1991). When the snail reaches the water's surface, it positions itself so that the pneumostome is exposed to the air while most of the rest of the body remains submerged. The pneumostome opens slowly and then, after an interval of approximately 20-30 seconds, it closes somewhat more quickly. The respiratory cycle may be repeated. It is believed that muscle contractions are responsible for the expulsion of air from the lungs when the pneumostome first opens, whereas the inspiration of fresh air occurs passively, by diffusion. At least three muscle groups are involved in the respiratory movements. The opener muscle and the closer muscle are present in the mantle tissue itself, but an additional bundle, called the mantle cavity muscle, is inserted at the shell column. The mantle cavity muscle runs through the lung roof, and terminates in the dorsal body skin. Motoneurons for the opener and closer muscles are present found in the visceral ganglion, intermixed with cells of other functional descriptions; motoneurons of the mantle cavity muscle are in the right parietal ganglion. In all animals, rhythmical behaviors, including respiration, are driven by networks of interneurons (sometimes including motoneurons) that convert tonic inputs into oscillatory outputs, typically with alternation between outputs to antagonistic muscles systems. Such networks are known as central pattern generators, or CPGs. According to Syed and co-workers (Syed et al., 1990; Syed and Winlow, 199la), the central pattern generator for respiration in Lymnaea consists of just three interneurons as shown in Figure 5.1 A. More accurately, the CPG contains three interneuronal types, since one element of the ensemble, Ip3, comprises several electrically coupled cells. To monitor the activity of Ip3, sometimes one of the coupled neurons, referred to as Ip3I, is recorded from, assuming that its activity is representative of the entire Ip3 ensemble. In other experiments, however, activity in Ip3I cannot be recorded because it is on the opposite surface of the ganglion from other neurons of interest; in these circumstances, activity in VJ cells, follower neurons of Ip3I, are recorded as a proxy. Experiments suggest that one of the unique neurons in the CPG network, RPeDl, may be important for initiating the respiratory rhythm. RPeDl is a giant dopaminergic cell found in the pedal ganglion. To study its role in initiating breathing, Inoue et al. (2001) used a semi-intact preparation in which the gas contents and pH could be controlled separately in the CNS and the periphery (pneuomstome, mantle, lung, kidney and heart). They found that the periphery exerts a background suppressive effect on respiratory discharges. However, when the periphery, but not the CNS, was made hypoxic (90%N2; 10%O2), the respiratory discharge frequency increased significantly. Since RPeDl is the only CPG neuron with peripheral axon projections, RPeD 1 probably provides the CPG with the signal that initiates breathing. When chemical synaptic transmission in the periphery was blocked, the CPG no longer responded to hypoxia, indicating that the axon of RPeDl is excited by chemosensory cells in the periphery. The precise location of these chemosensory cells, however, is unknown. Janse et al. (1985)

Regulation of the Internal Environment

69

Figure 5.1. Generation of a respiratory rhythm by a central pattern generator. (A) Synaptic connections among cells constituting the central pattern generator in Lymnaea. Open triangle, excitation; closed triangles, inhibition. (B) Initiation of the respiratory rhythm in an isolated brain preparation (1) and in culture (2). Since simultaneous recordings from VD4 and Ip3I are not possible with the brain preparation, the VJ cell is used as a proxy for Ip3I in part 1. Direct firing of RPeDl (horizontal bar) initiates cycles of bursts in VD4 and VJ (Ip3I). Note that there is no electrical coupling between RPeDl and the other two cells (hyperpolarization, asterisk). Reprinted with permission from Syed et al. (1990). Copyright 1990 American Association for the Advancement of science. noted that the RPeDl axon ramifies among the blood vessels that supply the anterior part of the lung roof, and they also found that RPeDl responded to changes in pO2 in the vicinity of the mantle, thus suggesting that the mantle in the source of the hypoxic signal. On the other hand, Inoue et al. (2001) suggest that the chemosensory cells may be located in the osphradial ganglion, but they present no evidence.

70 Behavior and Neurons in Gastropods

The oscillatory nature of the respiratory movements in Lymnaea is believed to result from reciprocal inhibitory connections between Ip3I and VD4, as shown in Figure 5.1 A. Reciprocal inhibition is a common anatomical basis for oscillation in central pattern generators. Ip3I and VD4 both connect to the same pneumostome motoneurons, but with opposite synaptic effects in every case. To determine which phase of the respiratory cycle is governed by which of the two interneurons, Syed and Winlow (199la) first analyzed the pattern of synaptic connections between the CPG interneurons and the pneumostome motoneurons, both openers and closers. The results suggested that activity in Ip3I should cause the pneumostome to open, while activity in VD4 should cause it to close. Experiments were then conducted using a semi-intact preparation that permitted simultaneous recordings of neuronal activity and muscle tensions. These experiments demonstrated, as predicted, that Ip3I activity is correlated with pneumostome openings and that VD4 activity is involved in pneumostome closures. However, earlier experiments by Janse et al. (1985) found that activity in VD4 was correlated with pneumostome openings. While the predominance of evidence seems to favor the view of Syed and Winlow (199la), these seemingly contradictory results suggest that the outputs of the respiratory CPG are not organized in a simple manner. Because of the numerous synaptic connections of the pattern generator neurons, and the impracticality of recording from Ip3I while simultaneously recording from RPeDl and VD4, owing to their positions on opposite surfaces of the CNS, the sufficiency of these three neurons for generating the respiratory rhythm has not yet been demonstrated in isolated brain preparations. Therefore, to determine whether these cells can sustain an oscillation, Syed et al. (1990) extracted the neurons from their respective ganglia and cultured them in vitro. When cultured individually, none of the cells had intrinsic bursting properties. Nor could any cyclical bursting be produced when only two of the three cells were co-cultured. When all three neurons were cultured together, their neurites formed synaptic connections that were functionally identical to those seen in semi-intact or isolated brain preparations. Now, when RPeDl was stimulated to increase its firing, the entire ensemble showed patterns of cyclical bursting that was strikingly similar to those observed in the isolated brain preparation. The intracellular records shown in Figure 5.IB reveal that the basic physiological mechanism responsible for the alternation of bursts in Ip3I and VD4 is postinhibitory rebound excitation, which is mediated anatomically by reciprocal inhibitory connections (Fig. 5.1 A). The neural control of respiration has been much less thoroughly studied in terrestrial species. The pneumostome is normally closed, but it opens repeatedly when a snail or slug is exposed to low ambient oxygen or elevated ambient CO2. Openings are correlated with muscular movements that inflate the surface area of the lung cavity, but there is little or no ventilation, so gas exchange probably occurs by diffusion aided by differences between internal and external partial pressures. For most terrestrial animals, respiration is

Regulation of the Internal Environment 71

controlled by the level of CO2 in the blood, not by O2 as in aquatic animals. With this fact in mind, Erlichman and Leiter (1997) used a microperfusion technique to search for CO2-sensitive neurons in the CNS of Helix pomatia. The technique permitted focal stimulation with a hypercapnic saline solution containing 6% CO2, in contrast to the 2-3% solution that bathed the rest of the preparation. Using a semi-intact preparation comprising the CNS and the mantle, they found an area on the dorsal surface of the visceral ganglion, at the border with the right parietal ganglion, that is capable of evoking pneumostome openings when focally stimulated with CO2. When individual cells within the area were probed with a micropipet, some cells, but not all, responded to the hypercapnic saline with increased rates of firing. A subset of these neurons (about 65%) caused the pneumostome to open when fired by intracellular electrical stimulation. While these results only begin to describe the neural network responsible for controlling the pneumostome in Helix, they are important because they suggest that respiration in terrestrial snails is triggered by the direct neuronal detection of elevated blood CO2 levels, a mechanism that has also been proposed for mammals but remains controversial (see Erlichman and Leiter, 1997). In contrast to these results in Helix, comparable experiment in Lymnaea demonstrate the absence of centrally located CO2-sensitive receptors, and even peripherally located CO2-sensitive receptors are doubtful in Lymnaea (Inoue et al., 2001).

5.3.

Blood Circulation

The blood (hemolymph) of gastropods serves the usual functions of supplying oxygen and nutrients to the tissues, and removing waste products. Also, as we saw in the last chapter, it plays an important role in the putative hydrostatic skeleton. The circulatory system of Aplysia, shown in Figure 5.2, is fairly representative of all gastropods. Only molluscs and vertebrates have chambered hearts. The single auricle collects blood and the ventricle pumps it out. Blood is distributed to organs throughout the body via large arteries or aorta, of which there are three in Aplysia but often fewer in other taxa. Routing of the blood supply is achieved by contracting or relaxing the circular muscles that enclose major arteries. In addition, at least in Aplysia, the abdominal aorta possesses its own sphincter. There are no true capillaries. Instead, blood is released from fine arterial branches into interstitial spaces, or lacunae, some of which are nearly as small as mammalian capillaries (10-20(im). Blood diffuses from the lacunae to the body cavity where it is known as hemocoel. It then collects in various muscular sinuses from which it is distributed to the kidney and the gills (or lung) in various proportions; from these organs it returns to the heart. Thus, the circulatory system is enclosed within vessels on the arterial side, but it is relatively open on the venous side. The heart is myogenic, so its normal spontaneous beating does not require input from the nervous system. However, the frequency and the strength of

72 Behavior and Neurons in Gastropods

Figure 5.2. The circulatory system of Aplysia. The heart has chambers, as in vertebrates. Note that blood leaves the heart "closed" within arteries, but it returns through an "open" hemocoel within the body cavity. White arrow, blood flow; back arrows, ultrafiltrate. Morphological evidence suggests that an additional site for formation of the ultrafiltrate is through the crista aortae surrounding the anterior aorta. From Koester and Koch (1987) with permission.

the heartbeat is regulated by neural and hormonal signals. Arterial flow rates are regulated independently of the cardiac functions. When Dieringer et al. (1978) looked for stimuli and behavioral states that affect heart rate in Aplysia, they found that the rate increased during eating and following a meal, during spontaneous increases in activity, with increases in temperature, and in response to foods or noxious stimuli. Similarly, in Clione, the heart rate varies during different behavioral activities (Arshavsky et al., 1990). In this animal, the heart rate is inhibited during defensive withdrawals and it ceases altogether during brief pauses in swimming (5-10 seconds), whereas it increases rapidly during escape locomotory reactions. When Clione encounters its natural prey, the heartbeat also accelerates rapidly to increase hydraulic pressure in the buccal region, so that the animal's formidible hunting apparatus can be everted (see Section 7.1.2). When Aplysia is aroused by food stimuli, the systemic blood pressure increases two-fold (see Koester and Koch, 1987). As the buccal mass moves forwards and back to grasp and ingest seaweed, phasic changes in vascular resistance cause the blood to flow preferentially towards the head when the buccal mass is being protracted and towards the digestive system when the buccal mass is being retracted. A summary of these observations is shown schematically in Figure 5.3. It appears that retraction of the buccal mass causes a mechanical constriction of the anterior artery that results in blood being diverted towards the digestive system. The subsequent diversion of blood flow to the head during protraction not only facilitates the movement of the buccal mass but it also compensates for the loss of anterior flow

Regulation of the Internal Environment

73

Figure 5.3. Phasic regulation of blood flow during feeding in Aplysia. Blood is directed toward the head when the buccal mass is protracted and it is directed toward the viscera when the buccal mass is retracted. The resistive changes in the anterior aorta are probably caused by mechanical factors. Resistance in the abdominal artery is controlled by three identified motoneurons that contact muscles near the base of the artery. Neural regulation of the abdominal artery facilitates protraction of the buccal mass and it compensates for the loss of blood flow to the head during retraction. From Skelton et al. (1992) with permission.

during retraction. Whereas the changes in resistance of the anterior artery can be attributed to mechanical causes, blood flow in the abdominal artery is regulated neurally. During protraction, three motoneurons in the abdominal ganglion, the LBVc cells, fire in phase to cause a contraction of the sphincter that surrounds the abdominal aorta (see Fig. 5.5B). The LBVc cells are themselves partly under the control of higher order neurons in the cerebral ganglion. Thus, constriction of the abdominal aorta during protraction of the buccal mass may be caused in part by activity in CC7, a cerebral interganglionic coordinating neuron that has monosynaptic excitatory connections with the LBVC cells (Xin et al., 2001). The blood circulation can be significantly affected by major postural adjustments, such as moving from an elongated form to a contracted one. Because the internal organs of gastropods lack the protection of a rigid skeleton, movements such as these could kink or twist blood vessels, but

74 Behavior and Neurons in Gastropods

there seems to be a mechanism, at least in Aplysia, to prevent this from happening. Some of the major arteries are supplied with longitudinal muscles that contract during certain body movements that might otherwise jeopardize blood flow. For example, the longitudinal muscle of the rostral anterior artery is innervated by a bilateral pair of cells, the pedal arterial shortener (PAS) motoneurons (Skelton et al., 1992). When the neck of the animal shortens, these cells fire in unison to shorten the artery. The cells also innervate the longitudinal muscles of the left and right pedal-parapodial arteries, with each member of the pair innervating either the left or the right artery. When the animal turns its head to one side, firing occurs only in the neuron that is ipsilateral to the direction of the turn. Two interganglionic control cells in the cerebral ganglion, the bilateral CC5 cells, provide a major source of excitatory input to the PAS neurons (Xin et al., 1996). The CC5 cells are involved in producing head movements in a variety of situations including locomotion, head turning, defensive head withdrawal, local tentacle withdrawal, feeding and head lifting. In all these behaviors, firing of the CC5 cells appears to be necessary and sufficient for contracting the pedal-parapodial arteries (Xin et al., 1996). Egg laying may be another activity that affects blood circulation in Aplysia. When hormones are released to cause egg laying, strong contractions begin in the anterior aorta and the gastroesophageal artery (Ligman and Brownell, 1985). Although these contractions are predominately longitudinal, and blood flow has not been measured under these conditions, the contractions may be sufficient to restrict blood flow through the affected arteries. If this were true, the blood would be rerouted to the reproductive organs via the genital and abdominal arteries. The contractions may be triggered in part by hormones acting directly on the vascular muscles and in part by hormonal modulation of central neurons, for example, the LByc cells and R15 (see below). Circulatory patterns do not return to normal until 30 minutes to several hours after release of the hormones. 5.3.7.

Chemical Mediators

The neural and hormonal mechanisms of cardiovascular regulation are still poorly understood. Although numerous substances are known to be cardioactive, in few cases is it known from which cells the substances derive, whether they function as hormones or neurotransmitters, and under what circumstances they are released. The first molluscan cardioactive peptide to be identified was the tetrapeptide FMRFamide, which was isolated from clam ganglia by Price and Greenberg in 1977. Many additional FMRFamide-related peptides (RFamides) were discovered later, for example, Helix alone expresses seven members of the RFamide family (Price et al., 1990). This group of peptides is so widely distributed in molluscs that it is sometimes referred to as "the molluscan peptide." FMRFamide itself has been localized to central neurons, to peripheral neurons, and to the heart. Two identified neurons in

Regulation of the Internal Environment

75

the visceral ganglion of Lymnaea, called Ehe cells, are known to both express FMRFamide and to regulate the heart (Buckett et al., 1990a). Bursts of spikes in these cells cause increases in the heart rate and the beat amplitude. They appear to be motoneurons because their spikes produce excitatory junction potentials in auricle muscle fibers at constant latencies. Perfusion of the heart with exogenous FMRFamide mimics the effects of intracellular stimulation, thus implicating FMRFamide as the excitatory transmitter. However, the conclusion is not straightforward. The exon of the FMRFamide gene that is expressed in the Ehe cells codes for two additional RFamide peptides that are also found in the Ehe cells, namely, FLRFamide and a 22-amino acid peptide abbreviated as "SEEPLY." Voltage clamp studies on the heart cells indicate that both FMRFamide and FLRFamide activate calcium channels, thus increasing contractions, while "SEEPLY" has no effect on the muscles when applied alone but it can modulate the effects of the tetrapeptides (Brezden et al., 1999). Evidently, the Ehe motoneurons release all three peptides, and the combined effect results in channel open times that yield optimal contractions. Other members of the RFamide family, and perhaps even FMRFamide itself, circulate in the blood to act as hormones. The tetrapeptide pQDPFLRFamide is particularly interesting. Its N-terminal extension provides some protection against enzymatic degradation, perhaps accounting for its detection at pharmacologically active levels in the blood of Helix. While it is about 100 times more potent than FMRFamide in exciting the Helix heart (Price et al., 1990), its actions on somatic muscle are opposite to those of FMRFamide. Whereas FMRFamide contracts somatic muscles, pQDPFLRFamide relaxes the same muscles. Lehman and Greenberg (1987) suggest that pQDPFLRFamide may have a special behavioral function. Since the emergence of a snail from its shell requires both an increase in cardiac output to increase hydrostatic pressure, and a relaxation of somatic muscle to eliminate opposing contractile forces, pQDPFLRFamide is suited to mediate extensions of the head and body. Another group of cardioactive peptides was discovered in Helix aspersa by Lloyd (see Lloyd et al., 1985). Because he found multiple peptides of different molecular weights, Lloyd initially classified them as small, medium, and large cardioactive peptides, or SCPs, MCPs, and LCPs. Presently, three of the SCPs have been sequenced, and they all contain nine amino acids (Price et al., 1990). Although all the peptides originally described by Lloyd are strongly cardioactive, only the LCPs are strongly implicated as neurohormones. Lloyd showed, in Helix, that the LCPs are released from the subesophageal ganglia and they are present in the blood circulation in concentrations sufficient to affect contractions of the heart ventricle. However, in Helix, none of the cardioactive peptides has been localized to single neurons. In Aplysia, where the SCPs are potent cardioexcitors, SCP-containing neurons are located mostly in the buccal ganglion. Some buccal motoneurons release SCPs together with ACh and the peptide modulates the postsynaptic actions of ACh on feeding muscles (see Section 7.5.6). It has not

76 Behavior and Neurons in Gastropods

been determined whether the SCPs released from the buccal ganglion also enter the circulation to affect cardiac function. In addition to the RFamide peptides, the SCPs and their relatives, other peptides also have cardioexcitatory properties. Noteworthy are myomodulin and R15a2, both of which have been localized to nerve terminals in the heart of Aplysia (Skelton et al., 1992) and Archidoris (Wiens and Brownell, 1995) using immunohistochemistry. Peptides that are structurally similar to myomodulin and R15a2 have been similarly localized to cardiac nerve terminals in a variety of gastropods including "prosobranch" and pulmonate species (Kerkhoven et al., 1993). Many non-peptide neurotransmitters are potently cardioactive. Acetylcholine, in concentrations as low as 10~ 10 M, has an inhibitory effect on the gastropod heart, just as it does in vertebrates. The cardiovascular actions of several identified neurons are mediated by ACh, for example, the two heart inhibitor motoneurons LDHn and LDHI2, and the vascular muscle inhibitors LByci-3 in Aplysia, and the heart inhibitor Khi in Lymnaea (Buckett et al., 1990b). Serotonin is equally potent as a cardiac exciter. In Aplysia, the identified neuron RBHE uses serotonin, as do the neurons Hhe and She in Lymnaea (Fig. 5.4). Dopamine excites the heart, but its functional role is uncertain because no cardioactive neuron has been shown to release dopamine. The neurons named R3-R14 in the abdominal ganglion of Aplysia innervate multiple vascular tissues including, but not limited to, the heart and the pericardium. They are thought to use glycine as a transmitter. They may also release peptides because they express a gene that appears to be a precursor for multiple peptides. An homologous peptide precursor is found in the "light yellow" cell system of Lymnaea, as well as in other pulmonate snails, both freshwater and terrestrial (Boer and MontagneWajer, 1994). Since the pulmonate cells innervate peripheral targets similar to those innervated by R3-R14, it is possible that the same system is present in all pulmonates and all opisthobranchs. The cell R14 is particularly interesting because it has a vast axonal projection, innervating vascular tissues associated with the digestive gland, the crop, the stomach, the abdominal ganglion, the pericardium, and the body wall (Rittenhouse and Price, 1986). In the nudibranch Archidoris, glycine can potentiate the contractions of muscles at the base of the anterior aorta but it has little potency for influencing the heart (Wiens and Brownell, 1995). The most likely function for the R3-14 neurons is to exert a broad influence on the contraction (or dilation) of vascular tissues. Clearly, there are many cardioactive messengers, but why so many? In Helix, for example, there are at least seven RFamide peptides, two SCP peptides, one peptide from yet another family, plus acetylcholine and serotonin. Are the various messengers producing the same effects or different ones? Are the peptides acting as hormones, neurotransmitters, or modulators? Future work should be directed toward continuing the identification of cellular sources of cardioactive messengers, clarifying their functional roles, and determining how they interact.

Regulation of the Internal Environment

5.3.2.

77

Control by Neurons and Neural Circuits

The cardiovascular motoneurons are often found in the visceral and parietal ganglia (abdominal ganglion in Aplysia), where they intermix with respiratory interneurons and motoneurons (Koester et al., 1974; Buckett et al., 1990b; Zhuravlev et al., 1999). However, the pair of arterial shortening neurons in Aplysia, mentioned above, and one cardioexcitatory neuron in Clione (Malyshev et al., 1999), are found in the pedal ganglia. In the nudibranch Archidoris, some cardioregulatory neurons are present in the pleural ganglion (see Wiens and Brownell, 1995). The actions of a representative heart exciter motoneuron is illustrated in Figure 5.4. Here, cell She of Lymnaea is seen to exert a strong effect on heart rate and heartbeat amplitude. Four other excitatory motoneurons have

Figure 5.4. An excitatory heart motoneuron, She, in the visceral ganglion of Lymnaea. (A) The cell has a single axon projection in the nerve that innervates the heart. (B) Four superimposed sweeps illustrate that excitatory junction potentials (ejps) are evoked in a heart muscle at a constant latency after spikes in She(C) Single action potentials in She induce single heartbeats. (D) A brief burst of action potentials in She induces increases in beat frequency and beat amplitude, which persist for almost 1 minute after termination of the burst. Adapted from Buckett et al. (1990b).

78 Behavior and Neurons in Gastropods

been identified in Lymnaea, two of which use the peptide FMRFamide as a transmitter (Buekett et al., 1990a,b). Only one inhibitory motoneuron, Khi, has been identified. Seven cardiovascular motoneurons of Aplysia were identified in an important early study of the neural control of circulation (see Koester et al., 1974). The initial set of motoneurons included two heart exciters, two heart inhibitors and three vasoconstricters, as shown in Figure 5.5B. Later, additional motoneurons were identified, including one exciter of the longitudinal muscle of the anterior aorta and two inhibitors of the same muscle (Sawada et al., 1981). Also, the multifunctional motoneuron L7 was shown to innervate auricle muscle and vasoconstrictor muscle surrounding the abdominal aorta; in both cases, L7 has an excitatory function (Alevizos et al., 1989a). Since no synaptic interactions have been described between the motoneurons, their activities appear to be wholly controlled by higher level interneurons and possibly by circulating factors. Several cerebral interneurons with influence on cardiovascular motoneurons have been described (Xin et al., 1996, 2001). As already mentioned, CCS commands contraction of the pedal-parapodial artery by monosynaptic excitation of the PAS motoneurons. CC7 exerts a strong excitatory influence on vasoconstrictor motoneurons and on the serotonergic heart motoneuron RBHE, in both cases by monosynaptic connections. Another cell, CC3 also excites RBHE, but only indirectly. When Kandel and his colleagues first began to record from identified neurons in the isolated abdominal ganglion, they noticed recurrent patterns of synchronous postsynaptic potentials that occurred synchronously in numerous cells, called "follower cells." Most of the follower cells were later identified as motoneurons of the cardiovascular, respiratory, and renal systems. The synaptic inputs were assumed to be from interneurons, which were given Roman numbers (I, II, III, etc.) corresponding to the different patterns of activity in the follower cells (see Kandel, 1976). Koester et al. (1974) used such physiological criteria to identify five interneurons coordinating the activity of cardiovascular motoneurons. Only later, and gradually, were these interneurons identified morphologically, at which time they were given Arabic numbers designating their positions in the ganglion (e.g., Interneuron I = L10; Interneuron II = R25/L25; Interneuron XI = L24). Interneuron I (L10) and Interneuron II have particularly interesting connections with the cardiovascular motoneurons. As shown in Figure 5.5B, L10 excites the heart excitatory motoneuron RBHE> inhibits the two heart inhibitory motoneurons and inhibits the three vasoconstrictor motoneurons. As a consequence, when L10 fires the heart rate goes up, vascular resistance goes down and cardiac output increases. No behavioral context for L10 firing has been ascertained, but it could be imagined that it fires when the animal needs to extend its body or in circumstances of metabolic stress. However, it is also possible that the main function of L10 is to regulate the kidney, in which case its ability to increase cardiac output could be seen as serving to increase the rate of blood filtration (see Section 5.4).

Figure 5.5. Stages in the elaboration of neural circuit models to explain respiratory pumping in Aplysia. (A) The first neural circuit for any molluscan behavior based on identified neurons. The cell labeled INT is Interneuron II, later shown to be the R25/L25 network. Neurons excited by Interneuron II are shown as thin circles; cells inhibited by Interneuron II are shown as thick circles. Reprinted with permission Copyright 1969 American Association for the Advancement of Science. From Kupfermann and Kandel (1969). (B) Numerous cardiovascular motoneurons receive synaptic input from either Interneuron II or L10 (Interneuron I), which themselves are connected by reciprocal inhibitory synapses. Open triangles represent excitatory synapses. Reprinted with permission from Koester et al. (1974). Copyright 1994 American Association for the Advancement of Science. (C) Bursts of activity in the R25/L25 network trigger respiratory pumping and orchestrate its expression. Neurons R15 and R20 are important modulators of R25/L25. Open circles, excitatory chemical synapses; closed triangles, inhibitory chemical synapses; closed circles, electrical coupling. Additional details of the motoneurons and premotor interneurons are shown in Figure 9.5. Reprinted with permission from Koester (1989). Copyright 1989 American Association for the Advancement of Science.

80

Behavior and Neurons in Gastropods

Interneuron II affects the heart motoneurons oppositely to L10, although it affects the vasoconstricters similarly (Fig. 5.5B). In addition to the connections shown in Figure 5.5B, Interneuron II also excites a pair of motoneurons that inhibit muscles at the base of the anterior aorta (Sawada et al., 1981). Thus, when interneuron II fires, the heart rate decreases and the blood vessels are dilated. As described in the next section, these cardiovascular actions form part of a broader behavior, respiratory pumping, that involves stereotyped movements in several structures involved with respiration and circulation. The cardiovascular component evidently allows for a flushing of blood through the heart and arteries. Reciprocal inhibitory connections between L10 and Interneuron II (fig. 5.5B) ensure that the incompatible cardiovascular actions commanded by the two neurons will not be initiated simultaneously. Three different types of cardioactive neurons have been discovered in the planktonic pteropod Clione limacina (Arshavsky et al., 1990; Malyshev et al., 1999). One type, present in the intestinal ganglion, is inhibitory. A second type is excitatory and seems to be represented by just a single neuron in the pedal ganglion. The third type, also excitatory, is represented by two pairs of bilaterally symmetrical cells in the intestinal ganglia. While all these cells apparently innervate the heart, significant differences were discovered by combining intracellular stimulation experiments and immunohistochemical investigations of the CNS and heart. Most interesting are the differences between the excitatory neurons (Malyshev et al., 1999). The pedal neuron affects only ventricular contractions, and it appears to use serotonin as a transmitter, whereas the intestinal neurons primarily affect auricle contractions and they appear to use pedal peptide as a transmitter. These findings are consistent with observations of episodes of spontaneous contractions in either the auricle or the ventricle, independent of contractions in the other chamber. Thus, it appears that the auricle and the ventricle can be independently regulated in Clione, but it is not known why. The neural control of cardiac function in the giant African snail Achatina fulica presents yet another type of organization (Furukawa and Kobayashi, 1987; Zhuravlev et al., 2001). In this animal there are eight putative cardioactive motoneurons; five are excitatory, two are inhibitory, and two produce biphasic inhibitory-excitatory junction potentials. The biphasic potentials result from activity in the two giant neurons, d-VLN and d-RPLN, which are known to constitute a bilaterally homologous pair even though the two cells are situated in non-symmetrical ganglia (Munoz et al., 1983). Several of the motoneurons are synaptically connected through either electrical junctions or chemical junctions, but some appear to act independently. Certain pairs of motoneurons are reciprocally inhibitory. An unusual feature of the network is the inclusion of a bilateral pair of cerebral neurons (Furukawa and Kobayashi, 1987). The cerebral cells appear to have both peripheral cardioexcitatory effects and central excitatory effects on the other heart regulators. The functional significance of these complex interactions remains to be discovered.

Regulation of the Internal Environment

81

An aspect of cardiac regulation that is no doubt functionally important, but difficult to study, is the role of peripheral neurons. From studies in several gastropod species, peripheral sensory neurons appear to control central regulatory responses, and there may also be peripheral motoneurons (Lymnaea, Buckett et al., 1990b; Achatina, Furukawa and Kobayashi, 1987). Interestingly, the peripheral sensory neurons appear to exert an inhibitory influence in most cases. Zhuravlev and colleagues recorded inhibitory postsynaptic potentials (IPSPs) in the myocardium of Helix using extracellular suction electrodes (see Zhuravlev et al., 1999). Barrages of IPSPs were frequently observed in the auricle and the lower part of the ventricle. The IPSPs occurred spontaneously and in response to mechanical stimulation of the heart. When the nerves from the central ganglia were severed, the number of IPSPs was reduced but not eliminated, thus indicating both central and peripheral sources of inhibition. Another phenomenon observed by Zhuravlev and colleagues is a ventricular-auricle reflex that produces IPSPs in the auricle at every systole of the ventricle; this reflex is actually enhanced when central connections are cut. One striking feature to emerge from studies of cardiovascular interneurons is the apparent multifunctionality of some of the cells. In order to coordinate visceral functions, and to link these to behavioral states, the interneurons typically synapse on effector neurons associated with more than one functional system. This is apparent in Lymnaea, for example, where the proposed respiratory pattern generator (Fig. 5.1 A) comprises three interneurons that influence heart rate, namely Ip3I, VD4 and RPeDl. These interneurons excite the heart exciter motoneurons, Hhe, while inhibiting the heart inhibitors, Khi (Buckett et al., 1990b). Further, they affect motoneurons for whole body withdrawal and locomotion (Syed and Winlow, 199la). Another pair of neurons in Lymnaea, named L/RPeD 11 for their locations in the left and right pedal ganglia, is linked in complex ways to several functional systems. These cells excite heart motoneurons, whole body withdrawal neurons and locomotor neurons (Syed and Winlow, 1991b; Inoue et al., 1996). They are also linked to the respiratory network because they excite putative motoneurons in the VJ and VK cell clusters, and they are inhibited by VD4, the pneumostome closer interneuron (Syed and Winlow, 1991b). RPeDl 1, but not LPeDl 1, inhibits the respiratory pattern generator neurons VD4 and RPeDl (Inoue et al., 1996). Lastly, L/RPeDll excites VD1 and RPD2, a pair of large electrically coupled neurons whose activity is strongly correlated with respiration and implicated in the control of circulatory, but whose precise function is unknown (Janse et al., 1985; Smelik, 1995). To summarize, in Lymnaea there is a neural network linking, on the one hand, respiration and circulation with each other, and on the other hand linking both of these systems to overt behaviors. The simplest interpretation of the synaptic connections made by L/RPeDll is that they prevent respiration during whole body withdrawal. Because respiration requires the body to be extended beyond the shell, respiration and whole body withdrawal are incompatible behaviors (Inoue et al., 1996; see Section 10.6).

82

Behavior and Neurons in Gastropods

Similar complex networks linking multiple visceral functions to overt behaviors are present in other gastropods. Several well-characterized neurons of Aplysia exhibit a striking degree of multifunctionality. The motoneuron L7, for example, has a field of innervation that includes the gill, the siphon, the sheath-contracting muscles of the pleuroabdominal connectives, the heart auricle and the vasoconstrictor muscle fibers of the abdominal aorta (see Alevizos et al, 1989a, 199la). The multifunctional roles of the interneuron L10 in cardiovascular control have already been mentioned; additional roles for this neuron in respiration and water regulation are discussed below. In terrestrial snails, some large neurons in the visceral and parietal ganglia that appear to have a role in commanding defensive responses, such as Pa3 in Helix and d-VLN in Achatina, have pronounced influences on cardiac function. Intracellular stimulation of these cells causes biphasic junctional potentials in heart muscle and also produces changes in the beat rate. The latter influence is itself modulated by other neurons within the subesophageal ganglia (Zhuravlev et al., 1999). The paired giant neurons, d-VLN and d-RPLN, are evidently multifunctional (Zhuravlev et al., 2001). While they innervate the myocardium (see above), they also innervate the ventricle, mantle, body wall, and lung vessels. In all cases, the giant cells evoke biphasic junction potentials in the target tissues. On the afferent side, the cells are excited by mechanical stimulation of the skin covering almost the entire body wall. The concept of multifunctionality, as illustrated by the examples above, is ambiguous. It might refer, on the one hand, to a cell which actually has multiple, independent functions, but this could occur only if there were multiple spike initiation sites and independent signalling in different branches of the axonal tree, as has been proposed for the metacerebral giant cell (Antic et al., 2000; see Section 7.5.3). Although the metacerebral giant cell is involved in regulating feeding behavior, and probably has no visceral functions, it is noteworthy that many of the neurons that are thought to have multifunctional roles in the visceral systems are similarly very large. If independent signalling to different target organs is a property associated with large neurons (see Antic et al., 2000), then there is at least the potential for true multifunctionality in the visceral interneurons. On the other hand, in some cases at least, the term "multifunctional" might only imply that a neuron has a single, high level function that encompasses actions on diverse effector systems.

5.4.

Respiratory Pumping

Respiratory pumping is a multifaceted stereotyped behavior that has been extensively studied in Aplysia. It is related to jet propulsion and the behaviors referred to as ventilatory pumping in other aquatic molluscs (see Kanz and Quast, 1990). In Aplysia, respiratory pumping involves withdrawal of the siphon, contraction of the gill and the mantle shelf, and simultaneous closure

Regulation of the Internal Environment

83

of the parapodia. As a consequence of these actions, sea water is expelled from the mantle cavity and fresh sea water is drawn in; simultaneously, as noted above, the circulatory system is flushed. The total duration of these associated activities is 3-6 seconds. Respiratory pumping occurs spontaneously at variable rates. The basal rate is about three pumps per hour, but much higher rates can be triggered by certain stimuli, as discussed below. Sometimes animals produce a patterned burst of pumping actions in which a high initial rate is followed by progressively slower rates over a period of 30-90 minutes; this has been called a "respiratory pumping seizure" (Kanz and Quast, 1990). Graded withdrawals of the mantle organs, similar or identical to those seen during respiratory pumping, can also be elicited by local tactile stimulation (see Leonard et al., 1989). The same motoneurons mediate withdrawal whether it is initiated centrally (i.e., during spontaneous respiratory pumping), or by peripheral stimulation (i.e., during a defensive reflex; see Section 9.3.3). It is noteworthy, however, that the LFS (left side, cluster F, siphon) motoneurons, some of which contribute significantly to siphon contractions in response to tactile stimulation, are not activated during respiratory pumping, regardless whether it is triggered or spontaneous (Frost and Kandel, 1995). In looking for variables that influence the rate of respiratory pumping, studies have focused on water quality. Several specific conditions that increase the probability of respiratory pumping include hypoxia, hypercapnia, low pH, hypertonicity, and hypotonicity (Levy et al., 1997). These conditions could arise if the animal encounters sea water of poor quality, suggesting that the function of respiratory pumping is to increase respiratory exchange in such situations. However, other results suggest that a broader notion of respiratory stress may be appropriate because feeding, escape locomotion, high temperature, light onset, reimmersion in sea water, and egg laying are additional triggers for respiratory pumping (Koester, 1989; Kanz and Quast, 1990). Because respiratory pumping flushes the mantle cavity, it can be used to remove debris or noxious materials lodged within the cavity. A defensive function is also indicated by the fact that pumping can be evoked by a local tactile stimulus. Finally, respiratory pumping facilitates the expulsion of defensive secretions, for example, ink and opaline (see Section 9.5.2), and it contributes to jet-propulsive swimming in some species (see Fig. 6.2A). In summary, respiratory pumping appears to be a multifunctional behavior. Some of the neuronal elements responsible for respiratory pumping were described as early as 1967. One particular cell, known as Interneuron II, was already considered important because it was found to mediate synaptic actions, mostly inhibitory, on a large number of follower cells. Many of the follower cells later turned out to be motoneurons innervating the organs involved in respiratory pumping (i.e., the heart, blood vessels, gill and siphon; see Kandel, 1976). At about the same time, the first motoneurons were identified in Aplysia. Kupfermann and Kandel (1969) described five gill motoneurons and provided a diagram showing the neural circuit for gill withdrawal (Fig. 5.5A). Their drawing, which is the first cellular description of any

84 Behavior and Neurons in Gastropods

centrally initiated molluscan behavior, shows Interneuron II distributing a mixture of excitation and inhibition to the gill motoneurons. In the same year, Peretz (1969) reported similar findings, again describing spontaneous gill movements in a semi-intact preparation, identifying motoneurons, and calling attention to the crucial involvement of Interneuron II. The decade of the 1970s produced detailed descriptions of cardiovascular motoneurons and interneurons (Koester et al., 1974; Kandel, 1976), as described in the preceding section and illustrated in Figure 5.5B. The fact that Interneuron II fires in spontaneous bursts in the isolated abdominal ganglion allowed researchers to monitor respiratory pumping in reduced preparations by recording characteristic patterns of synaptic activity in appropriate motoneurons. Eventually, Interneuron II was unmasked as two clusters of neurons, designated R25/L25, that lie at approximately symmetrical positions on the ventral surface of the abdominal ganglion (Fig. 5.5C). Each cluster contains about 14 neurons, and together R25/L25 constitutes a small network whose member cells fire in synchronous bursts during respiratory pumping (Koester, 1989). Direct electrical stimulation of a single L25 cell or a single R25 cell can trigger bursting in the R25/L25 network and an all-or-none sequence of respiratory pumping. Tactile stimulation of the mantle organs can also drive the R25/L25 cells to bursting. Koester (1989) showed that the bursts comprise two phases. The early phase, which is characterized by relatively lowfrequency firing, is driven either by an endogenous pacemaker or by afferent synaptic input. Possibly the rate of the pacemaker depolarization, and thus the rate of respiratory pumping, is modulated by the intensity of sensory inputs, which might reflect the quality of sea water in the mantle cavity. The late phase is characterized by high-frequency firing; it is driven by excitatory feedback within the network (Fig. 5.5C). Activity in the key R25/L25 network is modulated by interneurons (Fig. 5.5C). One pair of electrically coupled neurons, the R20 cells, can slowly excite the R25/L25 cells (Alevizos et al., 1989b). The R20 cells contain the small cardioactive peptide and presumably they use it as a transmitter. Activity in the R20 cells can initiate respiratory pumping or increase its spontaneous rate. Because no source of significant synaptic excitation has been found for the R20 cells, it is thought that they are driven by circulating factors (e.g., hormones, blood gases, or metabolic products). Phares and Lloyd (1996) suggested that one possible source of excitation for R20 cells might be the cerebral peptide 2 (CP2). This peptide is transported to the abdominal ganglion and, since it is slow to degrade, it may also circulate in the blood. Application of CP2 to the saline solution bathing an isolated abdominal ganglion causes direct (i.e., non-synaptic) depolarization in the R20 cells, as well as in the R25 cells. Consistent with these electrophysiological effects, an injection of synthetic CP2 into a freely behaving animal causes a five-fold increase in the rate of respiratory pumping (Phares and Lloyd, 1996). Another modulator of R25/L25 activity is the well-known neuron R15 (Alevizos et al., 199la). The cell exhibits a striking pattern of endogenous

Regulation of the Internal Environment

85

bursting when recorded in an isolated abdominal ganglion. Although extensively studied since first described by Arvanitaki and Chalazonitis (1958), the function of R15 has eluded investigators. Evidence for its function in osmoregulation and egg laying is given in Sections 5.5 and 8.6, respectively. In presenting an argument for its role in respiratory pumping, Alevizos et al. (199la) begin by noting that the synaptic actions of R15 are mediated by a peptide, R15cd, that rapidly desensitizes its receptors. Following dissection, R15 fires bursts of action potentials and presumably there is a substantial release of R15cd. This release leads to a desensitization of receptors and the consequent blocking of R15's effects. Therefore, to minimize the release of R15's transmitters, the authors performed their dissections using a saline solution containing high Mg++/low Ca++, and once the dissection was completed, they hyperpolarized R15 to limit its firing. If they then released R15 to fire in bursts, the frequency of R25/L25 bursts (hence, respiratory pumping) increased significantly during the first 5-10 minutes of R15's activity. Furthermore, bath applications of R15al also increased the rate of bursting in R25/L25. It is interesting, in light of the variety of conditions under which respiratory pumping occurs, that R20 and R15 modulate R25/L25 in different ways. R15 activates both strong and weak episodes of respiratory pumping, whereas R20 activates only strong episodes (Alevizos et al., 199la). It is also curious that the R20 cells exert strong inhibition on gill motoneurons located in the branchial ganglion, the same cells that are excited by R25/L25 (Alevizos et al., 1989b). Evidently these different modulatory effects reflect the multiple functions of respiratory pumping. A significant finding of Alevizos et al. (199la) is that R15 is not spontaneously active in the intact animal. This discovery was made by recording the cell's activity using an extracellular electrode chronically implanted around one of the peripheral nerves carrying an axon of R15. Several manipulations of the animal its environment failed to trigger activity in R15, thus leaving unanswered the question of what stimuli cause R15 to fire in vivo. Frost and Kandel (1995) described two neurons, L33a and L33b, that are strongly excited by the R25/L25 cells at the peak of their bursting activity. These neurons are not shown in any of the circuit diagrams of Figure 5.5, but they are illustrated in the circuit for siphon withdrawal, Figure 9.5. Spikes in L33a produce a monosynaptic IPSP in the LBs motoneurons, and spikes in L33b produce a monosynaptic IPSP in the gill motoneuron L7. Thus, the L33 cells seem to play a role in selecting the subset of motoneurons that will drive peripheral contractions during respiratory pumping, similar to the actions of R20 as noted above.

5.5.

Water Regulation and Excretion

Gastropods differ markedly in their requirements for internal water regulation depending on their ecological niches. The skin of marine gastropods is generally permeable to water and these animals are usually in osmotic

86 Behavior and Neurons in Gastropods

equilibrium with the sea water. Terrestrial gastropods, on the other hand, lose water by evaporation and they are subject to large swings of hydration states. Freshwater gastropods (Basommatophora) have very dilute blood but they still bale out copious amounts of water as urine. The pond snail Lymnaea ingests a significant amount of water through its mouth. In one study that used a tracer molecule placed in the ambient medium, as much as 30% of Lymnaea's urine production was attributed to oral ingestion (De With, 1996). Terrestrial snails and slugs exhibit a number of behaviors that appear adapted to the conservation of water. Perhaps most importantly, they largely limit their behavioral activity to the dark phase of the photoperiod so they avoid the heat of the day. This strategy allows some species of snails to live even in the hot and dry Negev Desert. In general, however, terrestrial pulmonates are found only in moist habitats. Slugs are especially vulnerable to desiccation because they are not provided with a hard shell. They lose water from their integument, from their lungs, and from the copious secretions of mucus that are required for locomotion. Losses as great as 16% of body weight per hour have been recorded. To minimize water loss from the lung, pneumostome openings are regulated by the animal's state of hydration. Dehydrated slugs open the pneumostome less frequently, for shorter durations, and to lesser extents than do hydrated slugs (Prior et al., 1983). When tested in a Y-maze that has one dry arm and one moistened arm, fully hydrated slugs show no preference for either arm, but slugs that have been dehydrated to 70% or less of their initial body weight orient toward the moistened arm (Banta et al., 1990). Dehydration also causes inactive animals to aggregate. They come to rest in closely apposed clusters, or huddles, which reduce evaporation by reducing surface area. Measurements of weight changes in huddled versus unhuddled slugs show that huddling reduces evaporative losses by about 40% (Prior et al., 1983). Observations of young snails indicate that the animals do not attain their aggregations either by orienting to individual mucus trails or to stationary physical cues, but rather they follow an olfactory cue (see Chase, 1986). This cue could be a spatially integrated signal from the aggregate of all mucus trails, or it could be a volatile substance released from the snails but not laid down in the trails. When the weather becomes hot and dry, snails climb up vertical or inclined surfaces to avoid the heat of the ground. They attach themselves to a tree trunk, a rock or a wall, and they secrete mucus from the skin of the mantle. After the mucus has sealed the shell to the substrate, more mucus, or a thin calcareous coating (epiphragm), is then secreted over the remainder of the mantle. By these means the snail encloses itself within a nearly watertight structure in which it can remain until reactivated by rainfall. Because the snail is elevated above the ground, it is relatively safe from potential flood zones. Snails have been reported to remain in this condition for up to 6 years without water. When these behaviors occur in the summer months, they are known as estivation; a similar response to cold weather, hibernation, is observed during the winter months (see Section 10.1).

Regulation of the Internal Environment

87

Some authors have suggested that terrestrial gastropods drink water to combat dehydration. There is no doubt that they ingest fluids through their mouths, and these may hydrate the animal, but Prior and colleagues (1989) showed that there is no oral uptake when feeding stimulants (e.g., glucose or fructose) are removed from the fluid by dialysis. Hence, according to Prior et al., oral ingestion of water is driven solely by nutritional needs, and rehydration per se is normally achieved by absorption of water through the skin or the integument of the foot. Notwithstanding the many behavioral strategies employed by terrestrial gastropods to remain hydrated, they cannot avoid significant changes in water content. Since there is little regulation of either blood osmolarity or blood ionic composition, as blood volume changes, so, too, do these variables. Blood osmolarity, for example, can vary more than two-fold. Despite this, terrestrial gastropods can tolerate large variations in water content, with some animals able to survive a loss of 80% of their body weight. It is reasonable to expect that changes in blood volume will affect nervous activity, for all cells will undergo changes in cell volume when blood osmolarity changes, and electrochemical gradients are affected by changes in blood ionic composition (especially Na+ and Cl~). While individual neurons have different responses, and long-term effects are sometimes quite different from short-term effects, in general, neural activity seems to be greater in hypotonic solutions than in normal or hypertonic solutions (Hasegawa and Takeda, 1986). Thus, the direct influence of dilute blood on nerve cell excitability can possibly explain the greater activity levels of snails and slugs in wet environments compared to dry environments. The organs of water regulation and excretion do not greatly differ among the gastropods. The renal system of Aplysia is illustrated in Figure 5.2. Hydrostatic pressure created by cardiac contractions causes blood to pass through certain regions of the heart's epithelial wall and the crista aortae, and causes it to be filtered in the process. The so-called ultrafiltrate then passes through a pore to move from the pericardium to the lumen of the kidney. The kidney secretes nitrogenous waste, in the form of ammonia, which is added to the ultrafiltrate. At the same time, mineral salts are reabsorbed into the blood. Urine leaves the kidney through a second muscular pore. The renal system of "prosobranchs" is similar to Aplysia's except that pericardial filtration is even more pronounced, owing to special epithelial structures. The site of filtration in pulmonates is controversial. Some authors maintain that the heart is the main site of filtration, whereas other authors deny any role for the heart and cite evidence indicating that the kidney has assumed this function. In any case, one adaptation of terrestrial species that undoubtedly conserves water is the conversion of ammonia to uric acid, which is excreted as almost insoluble crystals. The urine of pulmonates is also greatly reduced in volume, or made highly hypotonic, by secondary reabsorption of water and salts in the ureter. Basommatophores have largely retained the pulmonate adaptations, despite their return to an aquatic environment.

88 Behavior and Neurons in Gastropods

Innervation of the renal system by fibers immunoreactive to serotonin, FMRFamide, and other peptides is well established (Rittenhouse and Price, 1986; Koester and Alevizos, 1989; Giardino et al., 1996). FMRFamiderelated peptides are thought to be especially important for osmoregulation in pulmonate snails. One study using the intertidal snail Melampus bidentatus found that the osmolarity of the hemolymph adjusts to a broad range of environmental salinities (Khan et al., 1999). Body weights were initially affected by changes in the surrounding water, but they returned to their original values after about 24 hours. An immunohistochemical survey of various tissues in Melampus revealed that FMRFamide-related material is present at high level in the CNS and the kidney. The visceral nerve, which innervates the kidney, is also immunoreactive. Interestingly, levels of FMRFamide immunoreactivity in the CNS, kidney, and hemolymph were found to be highest in animals that had previously been exposed to hypertonic sea water. A clear example of how the nervous system exerts control over water regulatory functions comes from the work of Koester and Alevizos (1989) in Aplysia. These authors show that the renal pore opening (Fig. 5.2) is controlled by a pair of antagonistic muscles. One identified neuron, L10, is an important regulator of the renal pore opening. It directly excites the pore opener muscle and directly inhibits the pore closer muscle. Therefore, renal pore openings are highly correlated with spiking activity in L10, as shown in Figure 5.6. Curiously, LlO's control over pore opening appears to be aided by a single identified peripheral motoneuron, which is excited by L10 and which in turn excites the opener muscle. As noted above, L10 also has numerous synaptic connections with cardiovascular motoneurons. By means of these contacts, L10 can increase cardiac output, which may, in turn, increase the rate of blood filtration through the heart and the crista aortae. Because L10 simultaneously opens the renal pore, and perhaps has additional influences indicated by axon projections to regions of the kidney distant from the pore, L10 effects a coordinated enhancement of renal function. The control of the renal pore by LI0 is antagonized by a small number of neurons located in the left upper quadrant of the abdominal ganglion, the LUQ cells. There are five identified LUQ cells, named L2-L6. A subset of at least three LUQ cells excites the renal pore closer muscle, perhaps directly, and opposes excitation of the muscle by the peripheral opener motoneuron, perhaps by heterosynaptic inhibition of L10. Interestingly, this antagonism between L10 and the LUQ neurons in the periphery is supported by reciprocal inhibitory connections centrally. The axons of the LUQ neurons ramify extensively on the kidney, as do those of L10, so it is likely that all these cells have renal functions additional to micturition. Since the individual LUQ cells synthesize two different peptides, and each cell has a distinctive field of innervation on the kidney (Giardino et al., 1996), it would be useful to know how the renal functions of individual LUQ cells are differentiated. Finally, it is noteworthy that the effects of L10 on the renal system and the cardiovascular system are independently antagonized by the LUQ neurons

Regulation of the Internal Environment

89

Figure 5.6. One role of the multifunctional neuron L10 in Aplysia is to regulate kidney function, in part by controlling the release of urine from the kidney. The neuron makes direct excitatory connections with the renal pore opener muscle and direct inhibitory connections with the renal pore closer muscle. These records illustrate the correlation of spiking activity in L10 and renal pore opening. (A) Periodic spontaneous openings of the pore are invariably associated with high-frequency bursts in L10. (B) Single twitches of the opener muscle follow brief bursts of action potentials, but not single action potentials, in L10. From Koester and Alevizos (1989). Copyright 1989 by the Society for Neuroscience.

and by Interneuron II, respectively. Koester and Alevizos (1989) point out that this arrangement allows individual components of LlO's function to be overridden. A similar type of organization might apply to certain integrative neurons in other animals (e.g. RPD2 and IP3 in Lymnaed) and it might explain some of the inconsistent behavioral effects observed during the activation of these cells. A higher level of control of the renal pore is exerted by at least two neurons in the cerebral ganglion of Aplysia. CC2 has monosynaptic excitatory connections to the LUQ cells, and CC3 has monosynaptic connections with L10 (Xin et al., 2001). Thus, CC2 should promote closing of the pore, while CC3 should promote opening of the pore, although neither effect has been reported. Since both CC2 and CC3 respond to tactile stimulation of the lips, and they receive non-phasic inputs during the expression of buccal motor programs, they may coordinate visceral responses with feeding responses.

90 Behavior and Neurons in Gastropods

The secretion of hormones by neurons has long been suspected to play a role in osmoregulation. Studies in Lymnaea focus on three groups of neurons that possibly secrete peptides when the animal is exposed to hypotonic conditions (Soffe et al., 1978). These cells are found scattered within the visceral, parietal and pleural ganglia and are referred to as Yellow, Yellow-Green, and Dark Green, based on their reactions to the Alcian Blue-Alcian Yellow staining procedure. When the snails are removed from their normal pond water and exposed to deionized water for periods of one day or more before using the histological procedure, the number of neurosecretory cells that stain is less than when the snails are taken directly from pond water, thus indicating either enhanced release of material, reduced synthesis, or both. Although each type of neurosecretory cell is presumed to release a different neurosecretory product, it is unclear whether the responses of the groups to deionized water are in any way different. Soffe et al. (1978) suggest that one or more types is responsible for regulating blood volume after exposure to hypotonic water, because the histological signs of neurosecretory release follow changes in blood ion concentrations but precede changes in blood volume. In Aplysia, the neuron R15 has been implicated in several behaviors, including respiratory pumping, discussed above, and water balance. The most compelling evidence linking R15 to water balance comes from experiments in which either homogenates of the cell or doses of its peptide are injected into whole conspecific recipients. In the first report of such experiments, Kupfermann and Weiss (1976) injected homogenates of single cells and observed weight gains of 3-10% within 90 minutes. The minimum effective dose was found to be 10-20% of the contents of a single cell. Controls using other identified cells were ineffective. Later, 820 cell bodies of R15 neurons were removed from ganglia and the peptide composition of the cells was analyzed (Weiss et al., 1989). It was found that a single RNA message generates four peptides, and one of these, R15cd (38 amino acids), has a physiological potency similar to that of whole cell homogenates. Thus, if an Aplysia is injected with either a cell fraction containing the native peptide (obtained by high-pressure liquid chromatography; HPLC), or a synthetic peptide with the same sequence as the native peptide, the animal gains weight. The gains are equivalent with both types of injected material, amounting to 4% after 45 minutes and 7% after 90 minutes. These weight changes after injections could be caused either by an increase in water uptake or by an anti-diuretic action; water uptake itself could be secondary to an increase in blood tonicity. Regardless, the results are consistent with an earlier report that, when the osphradium is exposed to diluted seawater, the spontaneous spiking activity of R15 in semi-intact preparations is inhibited (Stinnakre and Tauc, 1969). With R15 inhibited, less of its anti-diuretic peptide messenger will be released, the animal will not take up water, and its internal osmolarity will be maintained. Despite all the evidence just reviewed, it remains unclear how R15 functions in the freely behaving animal. Bablanian and Treistman (1985) used a

Regulation of the Internal Environment

91

minimally dissected preparation that allowed them to measure the chemical composition of the hemolymph, while at the same time manipulating the spiking activity of R15. They first replicated the inhibitory effects of hypotonic seawater on the activity of the cell. Then, when R15 was hyperpolarized through an intracellular electrode to prevent its spiking, they observed that the hemolymph concentration of potassium rose 13%, ammonia rose 70%, and taurine rose 367%, in addition to other lesser effects. These findings led the authors to propose that, in nature, the inhibitory effect of hypotonic sea water on R15's activity might serve to protect cells from swelling when the dilute seawater enters the body. The inhibition of R15's activity could have this effect if it caused osmolytes (ions and compounds) to leave cells and enter the hemolymph. The problem with the hypothesis is that R15 is probably not spontaneously active in the fully intact animal (Alevizos et al., 199la), and even if it were spontaneously active, the hypothesis does not address the function of that activity, as opposed to the function of its inhibition. The large size and prominent bursting activity seen in R15 have prompted searches for homologous cells in related gastropod species. In land snails (Achatina, Helix, Otala), there is a large cell whose pattern of spontaneous firing in isolated CNS preparations almost exactly matches the bursting behavior of R15. The cell has been given several names, including Cell A, Big D, Br, Fl, RPal, 11, and PON. Its position in the rostral portion of the right parietal ganglion is equivalent to the position of R15 in the right half of the abdominal ganglion of Aplysia. On the basis of these similarities the pulmonate bursting cell has been considered homologous to R15 (see Furukawa and Kobayashi, 1987). I attempted to obtain functional evidence to confirm the homology by injecting conspecific snails with homogenates of the cell and then placing the snails in a shallow pool of water. However, there were no reliable differences between the weight gains recorded in experimental animals versus control animals that received saline injections (R. Chase, unpublished observations). Another cell that has been proposed as a homologue to R15 is RPD2 in Lymnaea stagnalis (Kerkhoven et al., 1993). As with the pulmonate cell discussed above, RPD2 has electrophysiological properties similar to those of R15, and it is located in an homologous ganglion (right parietal). It has a role, yet undefined, in the control of respiration and circulation (Smelik, 1995; Section 5.3.2). Both RPD2 and its electrically coupled mate, VD1, have prominent axon projections to the auricle of the heart, suggesting either a neurosecretory function or a cardiomodulatory function. Furthermore, RPD2 expresses a gene that shares considerable sequence homology with the R15 gene and, like the R15 precursor, the message in RPD2 is alternatively spliced to give multiple pep tides. The gene, however, may be widely expressed; an antibody generated against the alpha domain of the preprohormone labels numerous cells in the CNS of Lymnaea. The same antibody labels cells in ganglia from stylommatophores and opisthobranchs, but none of these cells corresponds to either R15 or its putative homologue in

92 Behavior and Neurons in Gastropods

Lymnaea (Kerkhoven et al., 1993). Nor is there any direct evidence indicating an osmoregulatory function for RPD2. In summary, there is currently insufficient evidence to conclude that RPD2 is homologous to R15, even if it is similarly involved in homeostasis through neurosecretion. As a final point, it is interesting to note that experiments using chronically implanted fine wire electrodes reveal that VD1 is only about 10% as active in vivo as it is in vitro (Smelik, 1995). Because RPD2 is electrically coupled to VD1, it, too, is probably much less active in vivo than in vitro. In this respect, at least, RPD2 behaves like R15.

6

Locomotion Not all gastropods are slow moving. The pelagic species, such as the pteropod Clione, are strong swimmers that can swim at speeds up to 50 cm/second. However, the vast majority of gastropods locomote not by swimming, but by crawling, and they are indeed slow. The fastest of the crawlers, Lymnaea peregra, has been clocked at speeds up to 17.5 cm/second. Crawling is accomplished by two different mechanisms, depending on the animal. Cilia are present on the soles of all gastropods, and their use for crawling is probably primitive, but ciliary locomotion is generally limited to small animals that carry light loads or larger marine animals that are buoyant in water. Other gastropods crawl by producing waves of muscular contractions on the sole of the foot. Ciliary locomotion is used by nearly all "prosobranchs" and basommatophores, many opisthobranchs (including the facultative swimmers), but only a few of the smaller species of stylommatophores. Swimming, by various mechanisms, occurs in many species of opisthobranchs and in a very few species of benthic "prosobranchs." Numerous studies have confirmed the key role of the pedal ganglia in the control of locomotion. An obvious indication of this is the fact that nerves connecting the CNS with the foot arise from the pedal ganglia. In addition, the pedal ganglia contain interneurons that generate rhythmical motor programs and motoneurons that innervate the muscles involved in locomotion. The cerebral ganglia provide the signals that initiate locomotion, while modulatory neurons are found in the cerebral ganglia and perhaps the pleural ganglia.

6.1.

Crawling by Ciliary Beating

If the sole of the foot of a freshwater snail or a nudibranch is viewed through a glass plate during locomotion, it will be seen that the sole remains adhered to the plate at all times and there are no visible contractions. Movement is achieved by the coordinated beating of cilia that cover the entire epidermal surface of the foot. The rate of ciliary beating is controlled by neurons in the pedal ganglion. Deliagina and Orlovsky (1990) found that the foot of the snail Planorbis corneus is divided into six zones, each of which is innervated by a different pedal nerve. Several neurons were identified that have axons 93

94 Behavior and Neurons in Gastropods

projecting into one or another of these nerves. The investigators studied the motor influence of the cells in a reduced preparation comprising only the pedal ganglion and the foot. By attaching a small disk to the surface of the foot they were able to measure the effect of neuronal activity on the rate of rotation of the disk, and thus measure directly a neuron's influence on locomotion. When the firing rate of one neuron, Al, was manipulated within its normal physiological range, Deliagina and Orlovsky were able to obtain nearly the entire range of spontaneous locomotor speeds, 0.2-1.3 mm/second. Because of the restricted anatomical projections of the cells, each one controls ciliary beating in just a single zone of the foot epithelium. This differential control is presumably useful during turning movements and when the animal is feeding on the surface of the water, at which time the anterior part of the foot is active but the posterior part of the foot is inhibited. The nudibranch Tritonia diomedea is an ocean bottom dweller that ordinarily moves by ciliary locomotion, although it swims to escape danger. It crawls slowly but smoothly at a rate of about 2 mm/second. Three bilateral pairs of neurons in the pedal ganglia regulate the ciliary beat frequency (Willows et al., 1997). When a single cell is stimulated intracellularly, it will either initiate ciliary beating or increase its frequency. Because the effect is seen even within small patches of the foot that have been isolated from neighboring regions by surgical cuts, it is thought that the pedal neurons may act by releasing messengers that diffuse across the epidermis, rather than by synapsing on every individual ciliary cell. Several investigations have attempted to identify the chemical or chemicals that mediate the nervous control of ciliary beating. Serotonin has been found to be one such mediator in at least Planorbis, Tritonia, and Lymnaea (Deliagina and Orlovsky, 1990; Willows et al., 1997; McKenzie et al., 1998). However, of the three pairs of identified cells that are known to control ciliary locomotion in Tritonia, only Pd21 contains serotonin. The other neurons, Pd5 and Pd6, do not contain serotonin but they synthesize three peptides, each of which is chemically related to the pedal peptide (Pep) of Aplysia (Willows et al., 1997). From experiments in which exogenous substances were applied directly to the foot, or to isolated ciliary cells, it is clear that both serotonin and the Pep-like peptides can modulate beat frequency. Dopamine is also implicated as a possible modulator of beat frequency (McKenzie et al., 1998). Whether the actions of these several messengers differ in detail, and how they interact, is unknown. It is interesting that pedal peptide may also have a role in regulating locomotion in Aplysia even though this animal does not use a ciliary mechanism. There is evidence suggesting that pedal peptide influences locomotion in Aplysia by modulating the contractions of muscles in the foot (Hall and Lloyd, 1990). Ciliary beating is important for the movement and rotation of many gastropod embryos. Experiments with the pond snail Helisoma have shown that serotonin is again an important mediator (Diefenbach et al., 1998). As in the adult Tritonia, there appear to be only a few neurons controlling the cilia. In fact, the pedal ganglia of early embryos contain only a single pair

Locomotion

95

of serotonergic neurons. These cells first express serotonin immunoreactivity at about the developmental stage when rotational behavior begins, and they lose their serotonin immunoreactivity at a later stage, approximately when this behavior ceases.

6.2.

Crawling by Muscular Contractions

Waves of pedal contractions propel many species of gastropods, but there are interesting differences in the mechanisms. In terrestrial gastropods, the waves move in the same direction as the animal, so they are known as anterograde waves, or direct waves. In many opisthobranchs and "prosobranchs," on the other hand, the waves are said to be retrograde because they move in the direction opposite to that in which the animal moves. Both anterograde waves and retrograde waves may occur either as contractions that extend continuously across the width of the foot (monotaxis) or as separate but parallel contractions on the two lateral halves (ditaxis). Even more complex are the waves in the limpet Cypraea, where each half of the foot can have waves progressing in different directions at the same time; this allows the animal to rotate in place without locomoting. Pedal contractions can be easily observed by first moistening a snail or a slug to activate it, and then placing it on a glass plate. When the crawling animal is viewed from underneath the plate, dark bands in the sole of the foot will be seen moving continuously forward. The number of bands that is simultaneously visible varies from 2-3 in small snails, to nearly 20 in a large animal such as an adult slug, Limax maximus. Although even early workers recognized that these bands were due to muscular contractions, an adequate understanding of the mechanics of locomotion required detailed investigations of the relevant shearing forces, pressures, muscular anatomy, and adhesive properties of mucus (reviewed in Trueman, 1983). A diagrammatic explanation of how the direct waves work is given in Figure 6.1. Generally speaking, forward motion in terrestrial gastropods is achieved at the sites of muscular contraction on the foot. At any one time, bands of tissue about 50-100 um wide are contracted, and these translate anteriorly as waves. The active regions are contracted forward, while the uncontracted regions remain in place. Thus, while a single point on the surface of the foot alternates between forward movement and status quo as it experiences successive waves of contraction, the movement of the whole animal is smooth because there are multiple concurrent waves. Some authors (e.g., Trueman, 1983) believe that the waves involve the contraction of oblique muscles as well as longitudinal muscles, with the former slightly elevating the foot to create a small concavity underneath the wave. Mucus is secreted under the foot from a narrow gland that lies along the midline in the anterior half of the body. Thus, by lifting the foot the friction of the mucus might be overcome. However, Denny (1980) showed that the mechanical properties of the mucus make it unnecessary for the foot to be lifted. Denny found that the strain

96

Behavior and Neurons in Gastropods

Figure 6.1 . Diagram illustrating the passage of two contractile waves in the foot of a snail or slug during locomotion. The muscle types are labeled. Contracting muscles are shown as heavy lines; resting or passively affected muscles are shown as thin lines. Mucus is represented as stipple beneath the foot. In regions marked with light stipple, the mucus is relatively fluid, allowing these parts of the foot to move forward. By contrast, in the regions marked with heavy stipple, the mucus is viscous, causing these parts to act as anchors. Since, in this example, both the animal and the waves move in the same direction (arrow), the waves are said to be "direct" or "anterograde." Adapted from Trueman (1983).

exerted by each contractile wave is sufficient to break down the adhesiveness of the mucus. When the mucus yields, it is converted to a liquid that permits forward movement. Once the wave passes, and the stress is eliminated, the mucus once again becomes an adhesive elastic gel. The switch from liquid to adhesive requires only about 100 milliseconds. This cycle of yielding followed by rapid healing allows the mucus to serve as a material ratchet, facilitating forward movements while resisting backward movements. Crawling in Aplysia involves retrograde waves. Generally, a single monotaxic wave of muscular contraction sweeps from head to tail before a second wave is initiated. Locomotion begins with a lifting of the anterior foot, which is then extended forward. Once this leading portion of the foot has again been anchored to the substrate, successively more posterior regions of the foot are likewise raised and extended forward. Thus, locomotion proceeds in the forward direction as the wave of contraction is seen to move backwards. A variant form of locomotion occurs only in response to noxious stimulation and may be regarded as an escape response. Here, the initial movement of the anterior end lifts the head higher than in normal crawling and extends it farther forward. After the anterior foot is reanchored, the entire remaining portion of the body is lifted from the substrate and extended forward. Thus, what is normally achieved by a propagated wave is, in the escape mode, achieved by a single step. Each step is accompanied by a transverse contraction along the length of the body and a flaring of the parapodia. This manner of accelerated locomotion in Aplysia is called "galloping" (Hening et al., 1979; Jahan-Parwar and Fredman, 1979), and similar variants

Locomotion

97

of crawling are exhibited by terrestrial snails. While galloping, Aplysia can reach speeds of 30 cm/minute. From the drawing in Figure 4.3, it is apparent that both the central nervous system and the peripheral nervous system are involved in the control of pedal waves, at least in terrestrial gastropods. Interest in how the central and peripheral components interact began early in the 20th century, but enduring insights were slow to accumulate. Considering that the pedal waves are a good example of stereotyped yet complex motor outputs, it is surprising that the neural control mechanisms have still not been thoroughly elucidated. Early studies reported that lesions of the pedal ganglion produced complete locomotor paralysis in Helix and Aplysia, but not in Limax and Arion (reviewed by Bullock, 1965). Later investigations performed in Aplysia (Hening et al., 1979; Jahan-Parwar and Fredman, 1979) showed that the pedal ganglion houses a pattern generator that is responsible for contractile waves in the foot. The waves disappear when the peripheral nerves of the pedal ganglion are cut, but patterned neural activity can still be recorded from neurons in the ganglion. When Hening et al. (1979) recorded intracellularly from pedal ganglion neurons, they found that in many cases the activity of the cell was correlated with locomotor contractions. Each neuron was related to contraction in only a limited region of the foot along the anterior-posterior axis, as shown in Figure 6.2B. Although the cells recorded by Hening et al. were probably motoneurons, this was not conclusively demonstrated. Jahan-Parwar and Fredman (1979) showed that, in addition to controlling locomotion per se, the central ganglia also mediate certain proprioceptive and tactile reflexes of the foot and parapodia, including a suction reflex necessary for locomotion. This was demonstrated by using segments of the foot that were isolated from each other except for nerve connections via the CNS. When a stimulus was delivered to one segment, a response could be measured in the other segment. Because pedal reflexes are thus routed through the CNS, and given that the basic locomotor rhythm has a central origin, there would seem to be no role left for the peripheral nerve net, at least in Aplysia, although some authors speculate that the nerve net "coordinates" the contractile waves. It is interesting in this regard that in terrestrial slugs (e.g., Arion and Limax} the pedal waves can still be generated even when the foot is completely isolated from the CNS (see Bullock, 1965). Thus, it seems that the peripheral nervous system plays significantly different roles in the control of crawling in different taxa. Experiments on locomotion in Aplysia were among the first to establish the decisive role of the cerebral ganglia in initiating behaviors. Because the cerebral ganglia collect information coming to the CNS from sensors concentrated at the anterior end of the animal, they are often crucially involved in evaluating which behaviors need to be performed at any given moment. In Aplysia, the cerebral ganglia trigger locomotion by sending appropriate neural signals to the pleural ganglia and the pedal ganglia. When the connections between these ganglia and the cerebral ganglia are

98 Behavior and Neurons in Gastropods

Figure 6.2. Locomotion in Aplysia. While some species, including A. californica, locomote only by crawling, others are capable of swimming. (A) Drawings from a film sequence showing swimming in A. fasciata. Adapted from Behavioral Biology of Aplysia by Eric R. Kandel. Copyright 1979 by W.H. Freeman and Company. Used with permission. (B) The paired pedal ganglia contain central pattern generators for locomotion, as well as motoneurons. Each ganglion is subdivided by connective tissue septa, as indicated at the left. The approximate locations of motoneurons are shown at the right where they are classified according to their target regions in the foot or parapodia (see legend). Four of the cells (P1-P4) are individually identifiable. Copyright 1974. Adapted from Hening et al. (1979) with permission from Excerpta Medica Inc.

cut bilaterally, all types of locomotion are disrupted (Jahan-Parwar and Fredman, 1979). Selective lesions of the cerebropedal connectives prevent both crawling and galloping, even though weak waves persist in the foot. By contrast to lesions of the cerebropedal connectives, lesions of the cerebropleural connectives prevent only galloping, leaving crawling

Locomotion

99

unaffected. Galloping is also eliminated after lesions of the pleural-pedal connections, thus confirming a crucial role for the pleural ganglia in mediating this form of locomotion. A modulatory role for the pleural ganglia in the control of crawling was shown by the fact that movements are slowed, but not otherwise affected, by cuts of the pleural-pedal connectives (Jahan-Parwar and Fredman, 1979).

6.3.

Swimming

A variety of adaptations evolved within the opisthobranchs to permit swimming. In most taxa this involved remodeling parts of the foot to serve as either paddles or wings. However, a form of swimming evolved in some species of nudibranchs and notaspids, despite the absence of either paddles or wings. In these species, swimming is achieved by undulations of the whole body. The majority of these animals are benthic and they ordinarily locomote by ciliary beating, but they swim as an escape response. For example, Hexabranchus, a coral reef dorid, swims by producing a retrograde wave along the lateral margin of the dorsal body surface. The waves are combined with a slower flexion of the body in the sagittal plane to simultaneously drive the animal forwards and upwards. Another group of nudibranchs, comprising small slug-like animals in the family Phylliroidae, has a planktonic lifestyle and swims by lateral retrograde undulations of the whole body. Escape swimming in the nudibranch Tritonia diomedea has been extensively studied from a neurobiological perspective (Fig. 6.3). For ordinary goal-directed locomotion, Tritonia uses ciliary mechanisms. However, when disturbed by secretions from the tube feet of a predatory sea star or, in the laboratory, with crystals of salt, the animal executes a swimming escape response (Willows, 1967, 1971). Figure 6.3A illustrates how this is done. The behavior begins with a local reflex withdrawal, followed by a longitudinal extension and then a strong ventral flexion of the entire body that propels the animal off the substrate. About 4 seconds later, there is a strong dorsal flexion. Altogether, the animal executes 1-8 cycles of ventral-dorsal flexion (average, five cycles), at the end of which it settles again on the substrate but at some distance from the source of the original stimulus. The notaspid Pleurobranchaea has a similar escape response, in this case often initiated by contact with a predatory conspecific. The anaspids swim using their parapodia, which are muscular lobes derived from the foot. In most species, the parapodia are reflected dorsally so that they lie on top of the shell and mantle. In the genus Akera, the parapodia are greatly enlarged to form a broad skirt surrounding the visceral mass. When the parapodia are opened and closed by muscular actions, the undulating movements gracefully propel the animal through the water. In Aplysia saltator, water is squirted through a funnel-like opening of the parapodia to produce jet propulsion. In A. fasciata and A. brasiliana, the parapodia are large, symmetrical structures that rest folded on top of

100 Behavior and Neurons in Gastropods

Figure 6.3. Escape swimming in the nudibranch Tritonia diomedea. (A) Swimming involves several cycles of alternating dorsal and ventral flexions. From Willows (1971). (B) Schematic model of the central pattern generator (CPG) and related circuitry responsible for escape locomotion. See text for abbreviations and explanation. Adapted from Frost and Katz (1996) Copyright 1996 National Academy of Sciences. (C) Simultaneous intracellular recordings from four of the cell types indicated in (B). The recordings were obtained from an isolated brain preparation. Electrical stimulation of a cerebral nerve (arrowheads) initiates a sequence of fictive swimming representing four cycles of dorsal ventral flexions. The dotted line in the DSI trace emphasizes the tonic ("ramp") depolarization that is crucial for generating the swim cycle. From Getting et al. (1980) with permission.

the posterior body. As shown in Figure 6.2A, swimming in these species occurs when the parapodia are engorged with blood and flapped while the lateral margins of the foot are drawn together across the longitudinal midline. The outwardly opening movements (downstrokes) begin bilaterally at the anterior part of the parapodia and then proceed posteriorly as a metachronal wave. The closing movements (up strokes) similarly begin at the anterior edge and progress posteriorly. The animals swim near the surface of the water, often continuously for long periods, but whether they swim to locate food, to find mates, or obtain shelter is not known. Although A. californica has parapodia similar to those present in A.fasciata and A. brasiliana, it cannot swim. Some gastropods have adopted holoplanktonic lifestyles, meaning that they spend their entire lives in the water column. Among "prosobranchs," however, there is only one group of pelagic species, the heteropods. These animals are lively visual predators that have modified a portion of the foot to create a single fin on the ventral side. They swim by going upside down and undulating the now upright fin. Speeds up to 50 cm/second are reached by the

Locomotion

101

larger heteropods when escaping predators. The Opisthobranchia include two orders, Thecosomata and Gymnosomata, that contain brilliant swimmers (Lalli and Gilmer, 1989). Here, portions of the foot have been modified to create wings (hence, pteropods). Whereas the thecosomes are shelled, the gymnosomes have discarded the protection of the shell in favor of more efficient swimming. Together, these animals combine beauty and efficiency in a variety of forms. Here is Morton's (1979) description of the thecosome family Cavoliniidae: "Through the needle-like Cresis, the broader and compressed Diacria, the tripod Euclio to the plump inflated Cavolinia, every detail of the design facilitates rowing, equilibration and maintenance of vertical position with a minimum of muscular work" (p. 47). The thecosomes are small animals that have a unique mode of feeding involving the deployment of freely floating mucous webs to capture small planktonic organisms. Limacina beats a pair of muscular wings in a sculling motion. It swims almost constantly to remain high in the water column. However, its jerky upward movements are occasionally interrupted when it stops swimming. Then, as it slowly sinks, it releases a mucus web, which causes the animal to become neutrally buoyant. After feeding at the web for a few minutes while remaining buoyant, it once again swims upwards. Other species make twice-daily vertical migrations in the water column, sometimes traveling several hundred meters in one direction. Swimming rates in the thecosomes range from 2 to 45 cm/second, with the higher rates achieved during escapes. The gymnosomes have adapted to the planktonic lifestyle by eliminating the shell, streamlining the body, and reducing its size. Because they live mostly in the open oceans and many of the 40-50 species are rare, the group as a whole is not well known. A notable exception is Clione limacina, which is the only pteropod species to have been studied to any great extent by neurobiologists. This animal is widely distributed in temperate and polar ocean waters. It ranges in size from about 20mm in length at the Friday Harbor Laboratories in Washington to about 85mm in the subarctic Atlantic. Clione limacina preys preferentially, perhaps exclusively, on two species of the small thecosome Limacina, hence the species name. Clione, in turn, is eaten by baleen whales. Swimming in Clione generally alternates with periods of inactivity, during which time the animal hangs upside down and motionless. Clione swims by moving its wings, as illustrated in Figure 6.4. The action has been described as symmetrical sculling (Morton, 1979). Rather than simply flapping up and down, the wings twist and turn to decrease the angle of attack (Fig. 6.4A, C). Muscular contractions on the downstroke drive the wing tip at velocities approaching O.lm/second. As the wing tip crosses the lateral midline of the body on the downstroke (Fig. 6.4A, frames 32 and 34), the muscular contractions stiffen and straighten the wing (Fig. 6.4B). Individual swimming episodes tend to be in either a slow mode or a fast mode, with top speeds of six body lengths per second obtained in the latter instances. In addition, there is a supercharged escape mode that overrides all other forms of swimming (Norekian, 1997). Escape is usually

102 Behavior and Neurons in Gastropods

Figure 6.4. Tracings of cine frames to illustrate swimming in the pteropod Clione limacina. The direction of swimming is ordinarily upward, toward the water surface. (A) One complete stroke as seen from the animal's right. Frame numbers are indicated; the time between successive frames is 10 milliseconds. Note the twisting of the wing during both the upstroke (frames 12, 14) and the downstroke (frames 32, 34). (B) Tracings of the leading edge of the left wing as seen from a head-on view of the animal. The horizontal line is the animal's midline. The dot at the right represents the attachment site of the right wing. (C) Tracings of the wing tip positions in a side view. Notched regions in the figure of eight represent times when the wing tips bend behind the rest of the wing. The vertical bar indicates where the wing attaches. Adapted from Satterlie et al. (1985) with permission of Company of Biologists Ltd.

triggered by tactile stimulation of the tail. It begins with a startle phase in which one or two wing strokes propel the animal forward at an extrapolated speed of 18 body strokes per second. Thereafter, swimming continues in the normal fast mode. Otherwise, swim speeds typically change from slow to fast after encounters with food prey, and following tactile stimulation on body

Locomotion

103

surfaces other than the tail. Clione usually swims upwards toward the water surface to prevent downward gravitational drift, but in warm waters it swims downwards. Considerable progress has been made in understanding the neurobiological bases for the switches from slow swimming to fast swimming, and from upward swimming to downward swimming, as will be discussed later in this chapter. 6.3.7.

Central Pattern Generators

When locomotion is achieved by rhythmical contractions in antagonistic muscle groups, as for example in swimming, central pattern generators create the motor pattern. We saw in Section 5.2 that the CPGs for respiration are comprised of networks of interneurons (sometimes including motoneurons) that convert tonic inputs into oscillatory outputs. The pattern generators for swimming have been studied in detail in Tritonia, Pleurobranchaea, and Clione. Investigations into the neural basis for swimming rely on two experimental approaches. The first uses nearly intact animal preparations such as those pioneered by Dennis Willows using Tritonia (Willows 1967, 1971). The animals in these experiments are intact except for a small slit cut into the dorsal skin of the animal to expose the brain. Hooks are then inserted around the margin of the cut to support the animal in a tank of sea water. Since the animals are buoyant, or nearly so, they can be kept high in the water with only a small amount of tension applied to the supporting lines. Animals prepared in this manner perform normal swimming movements even though they remain stationary. For microelectrode work, the brain is secured to a small platform placed beneath it. The second experimental approach uses a reduced preparation in which there is no bodily movement but only the expressed output of the central nervous system as recorded in efferent nerves. In these preparations, the cyclical bursts of spikes substitute for swimming per se; they are said to represent "fictive" swimming. The cellular network that generates the motor pattern for swimming in Tritonia is one of the best understood of all molluscan CPGs. It consists of just 12 interneurons (Fig. 6.3B), six in each cerebral ganglion. On each side of the brain there are three dorsal swim interneurons (DSIs), two ventral swim interneurons (VSIs) and one unique cell, C2. Every cell is electrically coupled to its contralateral counterpart. Peter Getting and his colleagues characterized the monosynaptic connections within the CPG and figured out how it works (Getting and Dekin, 1985). Individual synaptic actions are mostly inhibitory, with effects lasting from 0.65 to 20 seconds or more. None of the interneurons possesses intrinsic rhythmicity, but bursts of spikes are generated alternately in the DSIs and the VSIs, partly through the effects of reciprocal synaptic inhibition and postinhibitory rebound excitation, and partly by a delayed excitation of the VSIs by the DSIs acting through C2. Bursts in the interneurons produce bursts in the efferent neurons, with alternation between dorsal flexion neurons and ventral flexion neurons (Fig. 6.3C).

104 Behavior and Neurons in Gastropods

Strong support for the cellular organization of the swim CPG in Tritonia comes from a computer simulation of the network performed by Getting (1989). In this work, Getting modeled a simplified network comprising only one C2 neuron and one representative each of the groups DSI and VSI. He first defined the input-output relationship for each neuron by specifying passive membrane properties and repetitive firing characteristics. Then he modeled the time course and strength of each synaptic action. Finally, the network was provided with an input corresponding to the sensory signal that normally activates swimming. When the simulated network was thus stimulated, the three neurons fired bursts of action potentials in a pattern that did not differ in any fundamental way from that shown in Figure 6.3C. Importantly, all parameters in the model were established from experimental data prior to operation of the assembled network, that is, the success of the simulation did not depend on any post hoc tinkering of parameters to improve the model's performance. However, by selectively removing spike frequency adaptation and the transient potassium current, 7A, Getting discovered that, while neither of these cellular properties is critical for production of the correct order of bursting within a swim cycle, spike adaptation is important for determining the number of cycles produced and 7A is important for delaying the excitation of VSI relative to C2. Tritonia, a nudibranch, and Pleurobranchaea, a notaspid, have common ancestors in early pleurobranchomorph opisthobranchs (see Section 1.5). Not surprisingly, therefore, modern representatives of these genera have similar styles of escape swimming based on similar pattern generators. Jing and Gillette (1999) identified the main elements of the CPG in Pleurobranchaea californica; their results are summarized in Figure 6.5. As in Tritonia, most of the synaptic actions are inhibitory (Fig. 6.5A). Specific similarities between the CPGs of Pleurobranchaea and Tritonia are apparent, as is partly evident by a comparison of Figures 6.5B and 6.3B. First, both CPGs are composed of premotor neurons (interneurons) only (i.e., motoneurons play no role in rhythm generation). Second, the basic mechanism for alternation between dorsal flexion and ventral flexion is reciprocal inhibition between two groups of cells (Tritonia, DSI and VSI; Pleurobranchaea, Asl-4 and IVs)Third, certain neurons in the two CPGs are possibly homologous given similarities in their synaptic connections, morphology, immunostaining, and function. Particularly striking is the parallel between the As 1-3 neurons in Pleurobranchaea and the DSI neurons in Tritonia, both of which mediate dorsal flexion and all of which are serotonergic. Also, Al in Pleurobranchaea and C2 in Tritonia occupy unique positions in their respective CPG circuits, and both aid the transition from dorsal flexion to ventral flexion. On the other hand, the command components (Tritonia, DRI; Pleurobranchaea, A10) are quite dissimilar, and many circuitry details are obviously different. The neural control of swimming in Clione limacina was first described by two groups of investigators, one led by Richard Satterlie, working at the Friday Harbor Laboratories in Washington, and the other led by Yu. I. Arshavsky, working at the White Sea Marine Biological Station "Kartesh."

Locomotion

105

Figure 6.5. Synaptic connections among interneurons comprising the central pattern generator (CPG) for escape swimming in the notaspid slug Pleurobranchaea. (A) includes all single cells, whereas part (B) is simplified by grouping cells with similar functions. The cell labeled 7Vs is known only by its synaptic effects; its cell body has not been located. Note that the general organization of the CPG, shown in (B), is similar to that of the CPG in Tritonia (Fig. 6.3B), principally because of reciprocal inhibition between As 1-4 and IVs> and the recurrent inhibition of Al by A3 and IVsInteractions between the swim CPG and the feeding CPG are shown in Figure 10.8. From Jing and Gillette (1999) with permission.

The results obtained from the two groups (Arshavsky et al., 1991; Satterlie and Norekian, 1996) are generally consistent and complementary, except for one minor disagreement that will be discussed below. In contrast to Tritonia and Pleurobranchaea, where the CPG is found in the cerebral and cerebropleural ganglia, respectively, the CPG of Clione is located in the pedal ganglia. In all three cases, bilateral coordination is achieved by electrical coupling of interneurons across the appropriate commissure. The organization of the CPG in Clione is simpler than in the two previous examples. It consists of just two types of interneurons, one of which (type 7) is active during dorsal flexion of the parapodia (D-phase) and one of which (type 8) is active during ventral flexion (V-phase), as shown in Figure 6.6.

106 Behavior and Neurons in Gastropods

Figure 6.6. Schematic drawing of the central pattern generator for swimming in Clione (cell groups 7 and 8), shown together with several classes of motoneurons that control wing flexions. D-phase, upward stroke; V-phase, downward stroke. Adapted from Arshavsky et al. (1985).

There are about ten interneurons of each type in each ganglion, and neurons of the same type are electrically coupled. Each interneuron fires only a single action potential at the beginning of the appropriate phase. These action potentials are unusually long, with durations of 50-150 milliseconds, compared to 1-5 milliseconds in swim motoneurons. Although the circuit diagram of the Clione CPG includes reciprocal synaptic inhibition between the two types of rhythm-generating interneurons (Fig. 6.6), as in the other swim CPGs previously discussed, Arshavsky's group and Satterlie's group disagree about the necessity of reciprocal inhibition for creating the alternation of firing between D-phase interneurons and V-phase interneurons. Satterlie (1985) maintains that reciprocal inhibition is necessary because, when he bathed his preparations in high Mg++ to block synaptic transmission, there was no cyclical electrical activity, even though the cells could still be induced to fire by postinhibitory rebound after direct hyperpolarization. On the other hand, Arshavsky et al. (1986) found that rhythmical activity could still be recorded from the interneurons even after they had been removed from the ganglion and were completely isolated one from the other. In further experiments, Sadreyev and Panchin (2000) used a pharmacological agent to suppress the inhibition of V-phase interneurons by D-phase interneurons. They found that, despite synaptic suppression, an alternating pattern of rhythmic activity could still be induced in the two groups of neurons, but only if both were constantly depolarized. Since synaptic transmission was not completely blocked in these experiments, and since depolarization increased excitation in the interneurons, the extent to which these last experiments demonstrate that rhythm generation is independent of synaptic interactions is questionable. Nevertheless, it might

Locomotion

107

be safest to conclude, as do Sadreyev and Panchin (2000), that redundant mechanisms underlie pattern generation in the Clione swim CPG and that reciprocal inhibition, postinhibitory rebound, and endogenous oscillations all have roles to play. 6.3.2.

Muscles and Motoneurons

Swimming requires the coordinated movements of many muscles and even more motoneurons. To a large extent, these components of the motor system for swimming have been identified in the major model species. The findings indicate that perhaps a majority of pedal ganglion neurons is active during every episode of swimming. For example, Hume et al. (1982) studied swimrelated neurons in the pedal ganglion of Tritonia, where efferent cells were identified by their patterns of activity during the swim cycle. If a cell fired in synchrony with one or both of the major motor phases, it was considered a putative motoneuron, even though its direct connection with muscle cells was not established. From the samples taken by Hume et al., 56% of the neurons on the dorsal surface of the pedal ganglion are dorsal flexion neurons (50 cells total), 16% are ventral flexion neurons (12 cells), and 10% fired bursts either during both motor phases or during the interval between the dorsal flexion phase and the ventral flexion phase. Only 18% of the neurons sampled in the pedal ganglia (12 cells) did not fire any bursts during the swim cycle. The paired wings of Clione are short and broad. The movements of each wing are governed by a set of seven muscle groups. The two outermost dorsal and ventral groups produce bending movements; two of the three inner groups are involved in wing retraction; the remaining inner group is thought to control wing thickness and it may be involved in wing protraction (Satterlie et al., 1985). There are about 40 motoneurons in each pedal ganglion. About 20 motoneurons fire in phase with dorsal bending (D-phase) and an equal number fires in phase with ventral bending (V-phase). Arshavsky et al. (1985) described six types of motoneurons (Fig. 6.6) based on the timing and nature of their activities in respect to wing movements. Most synergistic motoneurons are electrically coupled. Satterlie (1993) recognized that, within each broad class of motoneuron (D-phase or V-phase), there is a single cell whose cell body is considerably larger than the others and whose axon, exceptionally, innervates the entire wing. All motoneuron axons exit the pedal ganglion via a single nerve running to the ipsilateral wing. The parapodial musculature in Aplysia brasiliana is even more complex than in Clione, with eight functional groups present in an equal number of dorsal-ventral layers (McPherson and Blankenship, 199la). The orientation of the fibers is different for each group. Using paired intracellular recordings, McPherson and Blankenship showed that the muscles are directly innervated by motoneurons that have somata in the pedal ganglia. From a sample of 410 parapodial motoneurons they identified 16 functional types based on axon projections, motor fields, directions of muscle contraction when the cells were

108 Behavior and Neurons in Gastropods

stimulated, and phasic firing patterns. Another 14 types were incompletely characterized. Interestingly, the cells of any given type tend to be found together in the same region of the pedal ganglion, but the amount of focus varies with type. This arrangement of parapodial motoneurons is similar to the arrangement of motoneurons innervating other body wall regions (Fig. 6.2B), wherein neurons innervating the anterior, middle, or posterior body wall tend to be found in different sectors of the ganglion. However, there seems to be no actual somatotopy (i.e., no orderly mapping of the body space on to the ganglion space). 6.3.3.

Initiation and Modulation

The role of the cerebral ganglia in initiating locomotion is demonstrated by two findings. First, as already mentioned in Section 6.2, spontaneous locomotion completely disappears after lesions of the cerebropedal connective nerves. Second, in experiments in vitro, nerve activity patterns indicative of fictive swimming can be elicited from the pedal ganglion by electrical stimulation of the severed cerebropedal connective nerves. Studies in several species have identified cerebral neurons that either initiate or modulate locomotor behavior. To study the roles of individual cerebral neurons in initiating different kinds of behaviors in Aplysia, Fredman and Jahan-Parwar (1975) began by identifying several groups of neurons on the dorsal surface of the ganglion. Their nomenclature designating these groups, shown in Figure 6.7A, was subsequently adopted by other workers studying the cerebral ganglion, as will become evident in later sections of this book. Fredman and Jahan-Parwar discovered that as few as two neurons within each of the C clusters is perhaps responsible for triggering crawling in Aplysia californica (Fredman and Jahan-Parwar, 1983). These cells are excited by the same stimuli that trigger crawling, and when stimulated by intracellular current injection, the cells are individually capable of initiating pedal waves, as monitored by discharges in pedal nerves. Furthermore, hyperpolarization of any individual cell blocks generation of the pedal wave motor program. Neurons with similar properties, located in the C and F clusters, seem to be responsible for triggering swimming in Aplysia brasiliana (Gamkrelidze et al., 1995). Three pairs of cells have been identified in this species and, again, each cell is capable of turning on the pedal pattern generator. A third pair of cells causes only weak activation of the swim program but it excites the other command cells. The trigger for swimming in Tritonia has not been so easy to find. It has long been understood that swimming begins only after the DSIs undergo a strong depolarization, or "ramp," that develops rapidly from transient afferent activity and then decays slowly over many seconds, all the while sustaining the swim cycle by increasing the excitability of the interneurons. The ramp depolarization can be seen by noting the dotted line under the DSI cell in Figure 6.3C. Initially, Willows and co-workers thought that the ramp depolarizations were caused by input from a group of "trigger" cells in the

Locomotion

109

Figure 6.7. Nerves and cellular organization of the cerebral ganglia in Aplysia californica. (A) Dorsal side; (B) ventral side. The identification of the cell groups shown here (labeled with letters) was an important step in elucidating functional roles for the cerebral ganglion. All cell groups with the exception of group H have bilateral representations. Nerve abbreviations: AT, anterior tentacle; C-B, cerebrobuccal connective; C-P, cerebro-pedal connective; C-PL, cerebro-pleural connective. LLAB, lower labial; O, optic; PT, posterior tentacle; ULAB, upper labial. Copyright 1975, and 1980. Adapted from Fredman and Jahan-Parwar (1975; part A), and Ono and McCaman (1980; part B) with permission of Excerpta Medica Inc.

pleural ganglion (see Willows, 1971). Later, when the cerebral neuron C2 (Fig. 6.3B) was identified, workers believed that it was the source of ramp depolarizations and the initiator of the swim cycle because intracellular stimulation of C2 at least sometimes started the swim cycle. However, in 1996, almost 30 years after the search for trigger cells first began, Frost and

110 Behavior and Neurons in Gastropods

Katz (1996) described a new single neuron, DRI, which has properties even better suited to the role of swim initiator. DRI stands for "dorsal ramp interneuron." Its cell body is located in the dorsal pleural ganglion, and it clearly does drive the ramp depolarization in the DSIs. Further indication of its role as swim initiator is the fact that DRI receives convergent input from afferent neurons. Crucially, experiments show that spiking in DRI is both necessary and sufficient to elicit the swim motor program. Once it has initiated activity in the CPG, DRI may recruit itself into the patterngenerating circuit because it receives phasic input from C2 (Fig. 6.3B). An important contribution of Peter Getting was to call attention to the fact that the same neuronal ensemble that constitutes the swim circuit in Tritonia also executes reflex withdrawals (Getting and Dekin, 1985). It must be the case, then, that the stimuli that elicit swimming switch the circuit from a resting condition in which it produces single-action reflexes to an oscillatory mode appropriate for swimming. In other words, the circuit can be "reconfigured" to produce a different kind of behavior. Getting proposed that a critical step in reconfiguring the network is the excitation of C2 because the synaptic actions of C2 on DSI, VSI, and other neurons seem to be necessary for causing oscillation in the circuit. The excitation of C2 depends on ramp depolarization and spiking in the DSIs (Fig. 6.3B). Katz and Frost (1995) showed that DSI influences C2 not only by conventional chemical synaptic transmission, but also through two separate modulatory actions mediated by serotonin. First, serotonin acts presynaptically on C2 to enhance its release of transmitter, and second, serotonin increases the excitability of C2 to allow it to fire at faster rates (Katz and Frost, 1997). In summary, the circuit can be switched from a reflex configuration to a pattern-generating configuration through the excitation and modulation of C2, and both influences are mediated by serotonin when it is released from the DSIs. Other studies have also indicated an important role for serotonin in controlling swimming in gastropods. Injections of serotonin (5-HT) at dosages of 10~ 5 M can cause swimming in Aplysia brasiliana and in Tritonia diomedea (McClellan et al., 1994). Similarly, swimming is initiated in Aplysia fasciata when the metabolic precursor of serotonin, 5-hydroxytryptophan (5-HTP), is injected, but only after a latency of 20-90 minutes (Sakharov et al., 1989). When 5-HT or 5-HTP is injected into species that do not swim, namely Aplysia californica and Aplysia depilans, the animals begin to crawl (Sakharov et al., 1989). In Clione, serotonin not only has a role in initiating swimming but it also modulates the speed of swimming (Satterlie, 1991). Panchin et al. (1995a) found ten groups of neurons in the cerebral ganglion of Clione that influence the pedal neuronal machinery for swimming. Most of the neurons either activate or accelerate the motor program, and some also intensify the firing rates of motoneurons. A few cells inhibit the motor program and/or the motoneurons. Immunohistochemical staining showed that two of the cell groups are serotonergic, and these latter groups were

Locomotion

111

studied more thoroughly because they were found to produce physiological changes like those observed during spontaneous accelerations of swimming (see Satterlie and Norekian, 1996). As previously mentioned, swimming in Clione is either slow, with a wing-beat frequency of 1-4 Hz, or fast, with a beat frequency of 3-8 Hz. The switch from slow swimming to fast swimming requires two key changes: an increase in the CPG oscillation frequency, and an increase in muscle contractile force (Satterlie, 1991). Satterlie and Norekian discovered that the serotonergic neurons recruit a special class of motoneurons, the so-called general exciter neurons (Satterlie, 1993), which increase muscle contractions by activating fast-twitch parapodial muscle fibers that are used exclusively for fast swimming. Several different mechanisms are involved in increasing the rate of CPG oscillation for fast swimming. One is the activation of a special group of interneurons, the type 12 cells (Arshavsky et al., 1989). These cells, possibly consisting of just a single member in each pleural ganglion, produce plateau potentials during the V-phase of the swim cycle, but only during fast swimming. The plateau potentials inhibit discharges in V-phase interneurons while exciting D-phase interneurons. Thus, when the plateau potentials are present, they change the dynamics of the CPG to increase its cycle frequency. Since intracellular stimulation of the serotonergic cerebral neurons can cause the type 12 interneurons to produce V-phase plateau potentials, this cellular pathway is presumably critical for bringing about changes in swimming speed. Another mechanism is spike narrowing in CPG interneurons (Satterlie et al., 2000). Stimulation of a cerebral neuron in the group that causes swim acceleration, or application of serotonin, leads to a shortening of spike durations by almost 20 milliseconds. Because each spike produces an IPSP in the antagonistic interneuron, the narrower spikes result in briefer IPSPs. Now recall that during each swim cycle the interneurons receive one IPSP and fire just one spike upon rebound depolarization. Thus, if spike narrowing did not occur, the combined durations of the IPSP and the spike would limit the maximum cycle frequency to about 4 Hz. With spike narrowing, however, cycle frequencies of at least 5 Hz can be achieved. Other findings by Satterlie et al. (2000) suggest that spike narrowing in Clione depends on a serotonininduced increase in potassium conductance. It was mentioned earlier that in Tritonia, serotonin increases spike frequency rates in a key CPG neuron by increasing its excitability, presumably by a decrease in potassium conductance (Katz and Frost, 1997). Thus, it appears that the mechanisms by which serotonin modulates swimming may be quite different in Tritonia and Clione. In addition to acting centrally, serotonin also acts at peripheral sites. The peripheral actions are attributed to a type of large serotonergic neuron that is found in the pedal ganglia of both Aplysia and Clione. Although the axons of these cells branch and terminate within the parapodial muscles, the cells are not capable of causing any motor effect when acting alone. However, they can modulate muscle contractions by releasing serotonin, which acts presynaptically to increase the release of acetylcholine from motoneurons

112 Behavior and Neurons in Gastropods

innervating the parapodial muscles. In Clione, activity in the pedal neurons enhances the size of excitatory junctional potentials (EJPs) and it facilitates spike-like events, but only in fibers of the slow-twitch type (see Satterlie and Norekian, 1996). In Aplysia brasiliana, where similar cells are known as POP neurons because they fire during the parapodial opener phase of swimming, activity in a single cell can double the amplitude of muscle EJPs and facilitate the amplitude of muscle contractions by 300%, as illustrated in Figure 6.8. In addition, the rate of muscle relaxation is increased by 40% (McPherson and Blankenship, 1991b). Using a voltage clamp, Laurienti and Blankenship (1997) found that serotonin increases the muscle's voltage-gated calcium current, which probably accounts for the change in contraction amplitude. In this respect, and others, the details of muscle modulation by serotonin turn out to be strikingly similar in the swimming motor system and the feeding motor system (see Section 7.5.6). The ultimate functional effect of the POP cells on the expression of swimming behavior has not been determined. Since physiological experiments indicate that the POP cells are active during the opener phase of swimming, they could contribute to changes in swim speed. Alternatively, or perhaps additionally, they could cause changes in swim direction when activated unilaterally. It is curious that large serotonin-containing neurons like the POP neurons of Aplysia brasiliana are present in the pedal ganglia of A. californica, Tritonia diomedea, and Pleurobranchaea californica (see Jing and Gillette, 2000), even though these latter species do not swim. It is thought that the cells may

Figure 6.8. Modulation of swimming speed in Aplysia brasiliana by serotonergic pedal ganglion cells that increase contractions in muscles responsible for the parapodial opening phase (POP). In these experiments, the motoneuron is fired by intracellular current injections. (A) The POP cell is silent. (B) With the addition of POP cell activity, muscle depolarizations (EJPs) and muscle tensions are enhanced. Spikes in the POP cell are distorted by current injection artifacts. Adapted from McPherson and Blankenship (1991b).

Locomotion

113

modulate crawling behavior in these species. However, in A. californica it is unclear whether crawling is mediated by 5-HT, by the pedal peptide (Pep), or by both substances (Hall and Lloyd, 1990; McPherson and Blankenship, 1992).

6.4.

Taxes and Other Orientations

The direction an animal takes when locomoting depends in part on whether or not it is attempting to reach a goal object such as food or a mate. If it is trying to reach a goal, it must continually perceive that object, judge its own position relative to the goal and make appropriate adjustments in the direction of its locomotion. Even in the absence of any goal object, many animals none the less move in one direction or another under the influence of a constant environmental stimulus such as light, gravity, wind, or odor. These movements, which may be directed either towards or away from the stimulus, are known as taxes. In still other cases, animals may use environmental stimuli as aids to navigation regardless of the compass direction in which they need to move. Locomotion in gastropods is strongly influenced by sensory stimuli in several modalities, and regrettably, because a variety of mechanisms is used in orientations and taxes, the terminology can be confusing. However, before discussing the sensory influences on locomotion, I will first review the motor mechanisms that determine the direction of locomotion. In crawling animals, turning is often achieved by contracting the body wall on the side towards which the turn is made. The body wall muscle is innervated by nerves and motoneurons of the pedal ganglion (Hening et al., 1979). If one or more of these nerves is cut, animals are unable to turn toward the side of the lesion (see Jahan-Parwar and Fredman, 1979; Murray and Willows, 1996). In the swimming species Clione limacina, turning is accomplished by flexions of the tail that produce moments of force, as with a rudder. Turns during locomotion can also be made by asymmetrical modulation of the left and right central pattern generators. This is an obvious strategy for crawling animals that use ditactic waves. However, even in Aplysia californica, which uses monotactic waves, unilateral lesions of the cerebropedal and cerebropleural connectives demonstrate the importance of the balance between left and right central pattern generators because such lesions cause the animals to circle strongly towards the intact side (Jahan-Parwar and Fredman, 1979). Another possibility for turning is modulation of the locomotor system downstream of the pattern generators, either at the level of the motoneurons or at the level of the muscles. This is easily imagined for animals that swim by beating their appendages, especially since certain cerebral neurons affect pedal ganglion neurons only unilaterally, that is, with either ipsilateral or contralateral influences (Gamkrelidze et al., 1995; Panchin et al., 1995a). One of the neurons contacted asymmetrically by the serotonergic cerebral neurons in Aplysia brasiliana is the POP neuron, which,

114 Behavior and Neurons in Gastropods

as described above, modulates contractions of the parapodial muscle. Thus, sensory signals converging on to the cerebral ganglion could cause turns by increasing the thrust produced by one parapodium relative to the other. 6.4.1.

Wind and Water Currents

The tendency for gastropods to orient upwind or upstream is common, and possibly universal. The former orientation is known as anemotaxis, the latter, rheotaxis. Locomotion upwind or upstream is a strategy adapted to bring an animal into contact with a goal object whose presence is indicated by chemical cues borne by air or water. Thus, anemotaxis and rheotaxis are typically triggered only after a chemical cue is detected in the air or water (Croll, 1983). In the land snail Achatina fulica, the posterior tentacles are sensitive to both wind and odors, and lesion studies demonstrate that these tentacles mediate anemotaxis (Chase and Croll, 1981). The posterior tentacles are extremely sensitive to stimulation by winds. When puffs of air with velocities as low as 25 mm/second are directed to the tip of the tentacle, transient discharges are recorded in the peripheral ganglion at both stimulus onset and stimulus offset (see Chase, 1986). These results demonstrate that the snail's tentacle is approximately as sensitive to wind as the cereal organ of cockroaches. In aquatic gastropods, water flow can be important for respiration, for flushing wastes, and for stabilizing position. These considerations apply especially to animals subject to tidal flows. Experiments performed by Murray and Willows (1996) tested how Tritonia diomedea orients to water flow in the absence of any contributing olfactory or thermal cues. Their results provide unequivocal evidence for orientation to, and locomotion towards, directional water flows. By examining the animals' performances after lesions of various nerves, it was determined that rheoreceptors (sensory cells sensitive to water flow) are probably concentrated in the lateral and/or ventral region of the oral veil. In Aplysia, by contrast, the structure most sensitive to water currents seems to be the rhinophore. The offshore orientation of A. brasiliana is mediated by rhinophores detecting the direction of the waves (see Hamilton and Russell, 1982). Lesion experiments have shown that neither rheotaxis (Tritonia, Murray and Willows, 1996) nor anemotaxis (Achatina, Chase and Croll, 1981) is affected by unilateral lesions of the relevant sensory structures. These types of taxes require only that a single sensor detects the directional component of a mechanical stimulus. 6.4.2.

Chemical Stimuli

To locomote toward an odor source, an animal must first detect the concentration gradient that is established by the source. Then there are two possible ways of moving up the gradient. With the first method, klinotaxis, the animal makes successive assessments of the chemical concentration and it

Locomotion 115

compares the concentration determined at one moment with that determined at a subsequent time. If it senses that the concentration is increasing, then it will continue moving in the same direction; otherwise, if the concentration is decreasing, it will change its direction. With the method of tropotaxis, the animal uses spatial integration rather than temporal integration. It senses the direction of the gradient by simultaneously sampling the environment at two or more places using multiple olfactory organs. The rhinophores and the tentacles are well suited for this purpose because they are bilaterally paired organs and their tips, which are chemosensory, are widely separated. I demonstrated that the snail Achatina uses tropotaxis by fitting each posterior tentacle with a glass tube through which I circulated odors (Chase, 1982). When I delivered an odor at unequal concentrations to the two tentacles, the snails consistently turned towards the side of higher concentration, indicating that they were using bilateral comparisons. However, in some situations, Achatina orients to odors using klinotaxis instead of tropotaxis. We reached this conclusion after watching snails with unilateral lesions of the posterior tentacles orient upwind to an odor source as well as intact snails (Chase and Croll, 1981). The result implies the use of klinotaxis because the animals had no contralateral tentacle with which to make spatial comparisons of odor concentration. In this situation, orientation is accomplished by klinotaxis in combination with anemotaxis. By contrast, in the absence of any ambient air currents, the animals could not orient unless both posterior tentacles were intact, implying the use of tropotaxis. Regardless of what method is used, movements of the chemosensors aid chemotaxis by permitting more efficient sampling of the environment. The head and tentacles of gastropods clearly do move during directed locomotion, but details of the sampling strategies employed by different species have not been carefully studied. In the common garden snail Helix aspersa, the tentacles show two kinds of subtle movements during olfactory orientation, neither of which is dependent on head movements (Lemaire and Chase, 1998). Twitches are brief retractions of the tentacle tip lasting only about 4 seconds. They may refresh the sensory process by removing odor molecules trapped in the liquid covering of the olfactory epithelium. Quivers are rapid lateral movements (350 milliseconds) unaccompanied by retraction; these may facilitate the access of odor molecules to the olfactory receptors by decreasing the boundary layer at the surface of the tentacle. 6.4.3.

Gravity

Gastropods rely on afferent signals from their statocysts to direct them either downwards (positive geotaxis) or upwards (negative geotaxis). Terrestrial snails are strongly driven to move upwards when wet. If a snail, Helix aspersa, is showered with water and placed at the base of an inclined plane, it will invariably move up the plane at the steepest possible angle (Wolff, 1975). In the field, snails are commonly seen stuck to walls, poles, or the undersides of elevated plants. There are probably several factors underlying

116 Behavior and Neurons in Gastropods

the negative geotaxis of terrestrial snails. Snails that remain on the ground in a heavy rainfall can drown in small puddles, so climbing is preventative. On the other hand, snails risk desiccation in hot dry weather. By climbing, the snails avoid the high temperatures associated with the ground surface, and consequently they reduce the loss of water by evaporation. In a study of Helix pomatia in Denmark, Lind (1988) observed that the snails climbed to an average height of 60.8cm during the hot summer months compared to only 29.3 cm in the earlier cooler months. In summary, negative geotaxis is favored by terrestrial gastropods that are about to enter a protracted period of inactivity, since an elevated position will provide protection against flooding, desiccation, and predation. When conditions change, and the animals need to return to the ground for foraging and mating, the geotactic response becomes positive. Aquatic pulmonates, such as Lymnaea, move upwards to breathe air at the water surface, but they move downwards in oxygen-rich waters. These observations led Janse et al. (1988) to suggest that an oxygen-dependent sensory mechanism might change the direction of geotaxis by efferent control of the statocyst hair cells (see Section 3.5). Considerable progress has been made in understanding how geotaxis switches from positive to negative in the pteropod Clione. As already noted, this animal is usually found maintaining itself near the water's surface in a head-up position. It does so by beating its parapodial wings, counteracting downward drifts resulting from either water currents or its own heavy head. However, every few minutes the animal ceases to swim and it is passively pulled down in the water column by gravity. After 5-10 seconds, swimming again resumes and the animal returns to its vertical head-up position. Experiments performed by Panchin and co-workers (1995b) have shown that the tail is used as a rudder to steer the animal in the upward direction. When reorienting after a passive fall, the tail must bend in a particular direction, either laterally or dorsoventrally, to achieve an upward movement. Instructions for the direction of bending come from the statocysts. When a statocyst hair cell is excited by gravity, it exerts differential excitation and inhibition on the cerebral interneurons that were described above (Panchin et al., 1995a). One group, named CPB3 (Deliagina et al., 2000), synapses on tail motoneurons, in the pedal ganglion. Some CPB3 cells excite left tail motoneurons, while others excite right tail motoneurons. A second group of interneurons, named CPB2, mediates excitation of motoneurons innervating the wing muscles. Finally, two additional groups of cerebral interneurons mediate excitation of the central pattern generator for swimming. Thus, the statocysts simultaneously drive the swim motor program and steer negative geotaxis by orchestrating appropriate adjustments of motoneuronal outputs. The foregoing scheme was tested using an ingenious in vitro preparation (Deliagina et al., 1998). In these experiments, stable nerve recordings could be obtained during imposed tilts because the electrodes consisted of small pieces of filter paper soaked in sea water, rather than conventional wire or suction

Locomotion

117

electrodes. An electrical motor was used so that the direction and amplitude of the tilts could be precisely controlled. In a closed-loop condition, shown in Figure 6.9, the motor was driven by the computed difference between spike rates recorded in left and right pedal nerves carrying efferent signals to antagonistic tail muscles. When the preparation was tilted slightly to one side, activity increased in one of the nerves and then, after a postural adjustment, activity increased in the other nerve. With the motoneurons thus firing in alternating bursts, the head-upright position was stabilized with periodic oscillations. This result confirms that stabilization is based on the antagonism between two groups of neurons, most probably the orientation-sensitive cerebral interneurons mentioned above (Panchin et al., 1995a). When the preparation was tilted under open-loop conditions, one of the efferent nerves fired continuously until the loop was again closed by intervention of the experimenters (Fig. 6.9). Geotaxis in Clione becomes even more interesting with knowledge that the direction of orientation is reversed in some circumstances and eliminated in others. So far, we have only considered negative geotaxis, with the head upwards, which occurs in cold waters. However, when Clione swims in water at warmer temperatures, its geotaxis becomes negative, with the head oriented downward. In laboratory experiments, 70% of the animals swim with their heads up at 5°C, whereas only 10% swim with heads up at 20 °C (Panchin et al., 1995b). At the higher temperature, most of the animals aggregate at the bottom of the aquarium as if they were trying to swim downwards to cooler waters. The direction of geotaxis also switches

Figure 6.9. The control of geotaxis in Clione. Recordings were obtained using a robotic in vitro preparation with closed-loop feedback. Spiking in the left and right tail motoneurons (LTN, RTN) drives a motor, which tilts the entire preparation to the left or right. The orientation of the animal relative to vertical (0°) is shown in the trace labelled /3. Note that the orientation is stabilized by alternating bursts in the two neuronal groups. When a large rotation is made under open-loop conditions (horizontal line), neurons in the left tail nerve fire continuously until the loop is again closed, whereupon stability is quickly re-established. From Deliagina et al. (1998). With copyright permission of Nature magazine.

118 Behavior and Neurons in Gastropods

during one stage of hunting. In addition, during defensive reactions, when the animal ceases to swim and falls passively downward, there is no geotaxis at all. Several mechanisms have been proposed to explain the temperaturesensitive reversal of gravitational reflexes in Clione. One idea is that the responses of statocyst hair cells are modified through their efferent innervation (see Section 3.5; Panchin et al., 1995b). However, recent experiments indicate that the responses of hair cells do not change, but instead, the connections of hair cells with central interneurons are modified. Using their stable in vitro preparation, Deliagina et al. (2000) were able to record neural activity in critical nerves during tilts produced at different temperatures. Figure 6.10A shows that axons in the right tail nerve are excited by tilts to the right at 10°C, whereas at 20 °C they are excited by tilts to the left; opposite results are seen in the left tail nerve. These findings are consistent with the effect of temperature on the behavioral reflex, but they do not provide insight into the mechanism of directional reversal. More revealing is the change of response in the cerebral interneurons, CPB3, which was seen while monitoring the activity of these neurons in the subpedal commissure. As shown in Figure 6.10A, the CPB3 cells fire in phase with right tilts at 10 °C, but they fire in phase with left tilts at 20 °C. From these changes in CPB3 responses, and from other results, Deliagina et al. conclude that the reversal of geotaxis depends on a reconfiguration of the connections between hair cells and interneurons, as shown schematically in Figure 6.1 OB. Essentially, the scheme relies on the selective suppression of certain synapses coupling statocyst hair cells and cerebral CPB3 interneurons. The authors contend that a similar but more pervasive suppression of hair cell to interneuron connections accounts for the loss of geotaxis during defensive responses. However, the mechanism of synaptic suppression has not been identified; nor has its dual control by temperature and behavioral state been explained. 6.4.4.

Light

Some gastropods are positively phototactic, some are negatively phototactic, and still others change their orientation to light depending on conditions. Terrestrial taxa, for example, Helix and Otala, which are mostly active at night and which need to avoid the heat of the sun, are negatively phototactic. Similarly, most aquatic "prosobranchs" orient preferentially toward dark stimuli (Hamilton and Winter, 1984). On the other hand, Lymnaea and Aplysia californica are positively phototactic. In low light environments, Lymnaea orients using its eyes, while in bright light it uses extraocular photoreceptors (Stoll, 1976). An interesting example of the use of visual cues has been reported for Aplysia brasiliana, which orients to the shore while swimming at the water's surface. Hamilton and Russell (1982) reported that the animals become disoriented when they are denied an unobstructed view of the sky, but whether the pertinent visual cue for A. brasiliana is the sun's

Locomotion

119

Figure 6.10. Condition-dependent expression of gravitational reflexes in Clione. The animals swim upwards in cold water, downwards in warm water, and not at all during defensive reactions. (A) Nerve recordings show that motoneurons and cerebral interneurons can fire during either left or right tilts depending on temperature. CPB3s, cerebropedal interneurons; LTN, left tail nerve; RTN, right tail nerve. (B) Proposed scheme to account for behavioral and electrophysiological observations. The statocyst receptor cell shown in black activates right tail motoneurons (black cells) at low temperatures, but left motoneurons at high temperatures. During defensive reactions, none of the motoneurons is excited. Gray arrows indicate suppressed synapses between receptor cells and interneurons. Adapted from Deliagina et al. (2000). With copyright permission of Nature magazine.

position in the sky, the polarity of the light or something else, has not been determined. Some clever experiments performed by Fraenkel (1927) demonstrated that phototaxis can be either positive and negative in the same animal at different times, just as geotaxis is variable as described above. Fraenkel's experiments were performed using the intertidal snail, Littorina. Under usual conditions, this snail is negatively phototactic, and light causes the animal to move onshore to the high intertidal zones where it can find its preferred plant

120 Behavior and Neurons in Gastropods

species. Also, when above water, negative phototaxis causes the snail to go into rock crevices where it can avoid the desiccating effects of direct sunlight. What Fraenkel showed was that, when Littorina is under water and upside down, it becomes positively phototactic. The interpretation offered by Fraenkel is that the switch in phototaxis makes it possible for a snail to first enter a dry crevice and later, after the tide has risen, to exit from the crevice. Imagine a U-shaped crevice that is horizontally orientated and lit by sunlight at its entrance. Using negative phototaxis, a snail should enter the crevice and crawl to the cul-de-sac; then, using negative geotaxis, it should proceed upward. However, at this point, if it remained negatively phototactic, it would remain stuck in the cul-de-sac. Instead, by reversing its response to light, it can leave the crevice and return to the open water. Phototaxis in the nudibranch Hermissenda crassicornis is another modifiable response. This example is particularly interesting because the mechanisms of modification by associative conditioning have been extensively investigated. Associative conditioning of phototaxis in Hermissenda is demonstrated by pairing light with rotation (Alkon, 1980). The animals are individually placed in clear plastic tubes that are arranged radiating from a central core. Once in the tubes, they are subjected to a training regime, consisting of exposure to one or more sensory stimuli, and then they are allowed to move towards the illuminated core. If there is no training, the animals consistently move to the light. If the entire array of tubes is simultaneously illuminated and rotated, the animals' movement towards the lighted core is suppressed relative to controls. Since the control groups include animals that experienced rotation alone, light alone, and random presentations of light and rotation, the modification of behavior is shown to depend on a specific temporal pairing of rotation and light. This example of associative learning may be related to the animals' behavior in its natural intertidal habitat in the following way (Alkon, 1980). Positive phototaxis could be adaptive under normal conditions because it would bring the animal to shallow waters where it can encounter its preferred food, small hydroids. In turbulent seas, however, the animal will be buffeted, perhaps violently, by the wave action when it is in shallow water. Because turbulence is inherently aversive, the animal may learn to avoid light, and hence waves, provided it has experienced the association of light and turbulence. In the laboratory, rotation of the illuminated radial tube apparatus may simulate this natural phenomenon. Since the learning persists for several days after training, and it increases with the amount of training, it could be acquired during one stormy period and retained until the next. What are the cellular mechanisms responsible for the conditioned modification of phototactic behavior in Hermissenda1. As described in Chapter 3, the eye contains only five photoreceptors, two of which are type A and three type B. Figure 6.11 shows some of the numerous synaptic interactions that exist between statocyst hair cells and photoreceptors. According to the explanation initially offered by Alkon (1980), associative conditioning causes a disinhibition of the type B photoreceptors, an effect mediated by the

Locomotion

121

Figure 6.11. Interactions between neural elements in the eye and statocyst of the nudibranch Hermissenda. There are two type A photoreceptors and three type B photoreceptors. The optic ganglion contains a total of 13 neurons, two of which are identifiable as individuals. The statocyst contains 12 hair cells, although only three are shown here. Clear axons are excitatory, while stippled or dark axons are inhibitory. Behavioral training using paired presentations of light and rotation results in higher rates of firing in type B photoreceptors. From Alkon (1980) with permission.

intersensory connections. Alkon and his colleagues observed several electrophysiological changes in the B cells that could account for their disinhibition and that were correlated with conditioned behavioral changes. These B cell responses include a cumulative membrane depolarization, an increase in input resistance, and a decrease in voltage-dependent K+ currents (Alkon, 1980; Alkon et al., 1985). Since the excitability of the B cells increases as a net result of the cellular changes, they fire more spikes in response to a light stimulus. Owing to the numerous inhibitory outputs of the B cells, notably those on to the type A photoreceptors (Fig. 6.11), the increased response of the B cells causes downstream inhibition, which ultimately results in a lessened activation of motor circuits in the pedal ganglion and a suppression of phototaxis. While the foregoing account of cellular changes provides a plausible explanation for behavioral learning, other findings complicate the story. Crow (1985), for example, called attention to the fact that, whereas all earlier biophysical measurements of the B-cell responses to light were obtained from dark-adapted eyes immediately after light onset, the behavioral measurements of phototaxis were obtained after the animals had adapted to the light for several minutes. He therefore looked at the B cells' steady-state responses to light, rather than their phasic responses immediately after light onset. He discovered that the amplitude of the generator potential and the rate of spiking were both substantially reduced 5 minutes after light onset in

122 Behavior and Neurons in Gastropods

conditioned animals relative to controls. Thus, Crow found that the responses of the B cells did not increase after conditioning, they actually declined. Since the strength of phototactic behavior is closely related to light intensity, his results suggest that the suppression of phototactic response after conditioning can be explained in part by a change in the adaptation properties of the photoreceptors, without invoking any network interactions. Other studies have shown that the light response changes not only in the type B photoreceptors but also in the type A photoreceptors (Farley and Han, 1997). Also, the effects of training are not limited to changes in generator potentials and excitability. One particular synaptic interaction, the inhibitory connection between type B photoreceptors and type A photoreceptors, is enhanced after conditioning (Frysztak and Crow, 1994). It is interesting that the mechanism of facilitation at this synapse appears to be an increase in the duration of presynaptic action potentials caused by a reduction in one or more presynaptic K + currents (Gandhi and Matzel, 2000). This indicates that at least one component of the cellular mechanism for learning in the eye of Hermissenda is similar to the main mechanism responsible for learning in the withdrawal reflexes of Aplysia (Byrne and Kandel, 1996; Section 9.4.2). In summary, the phototactic behavior of Hermissenda exhibits a fascinating plasticity involving interactions between two sensory systems. Several cellular correlates of associative conditioning have been described, but which of these phenomena is most important for causing the change in behavior is at present uncertain. It can also be noted that nearly all studies relating to this issue have focused on the photoreceptors and the optic ganglion, with little attention having been given to the motor circuitry. Recent work indicates that each of the five photoreceptors makes different synaptic connections with downstream interneurons (Crow and Tian, 2000). This allows for the possibility that the distinctive effects of conditioning on each of the photoreceptors might be preserved at the point of interaction between the visual system and the motor system, perhaps allowing for learning to affect photo taxis while leaving other responses to light unchanged. 6.4.5.

Magnetic Fields

As discussed in Section 3.6, there is evidence that gastropods can detect geomagnetic cues. A possible role for a magnetosense in locomotor orientation has been studied in Tritonia using both behavioral and electrophysiological approaches. The influence of magnetic field on body axis orientation and Y-maze choice behavior was reviewed in Section 3.6. One neuron in the pedal ganglion of Tritonia, Pd5, is thought to play a role in magnetic orientation. This cell contains peptides that accelerate ciliary beating when applied directly to the foot (Section 6.1); its firing rate is correlated with the rate of locomotion as measured on a treadmill, and its axon innervates the foot (Popescu and Willows, 1999). Thus, Pd5 may be able to modulate crawling locomotion. Significantly, experiments performed in vitro show that the spiking activity of Pd5 increases significantly after brief rotations of

Locomotion 123

an earth-strength magnetic field, while no such responses were observed in 50 other neurons tested under the same conditions (Popescu and Willows, 1999). Because the magnetic response in Pd5 has a very long latency, 6-16 minutes, and because the response is abolished when all peripheral nerves are cut, it is probably triggered by receptors outside the CNS. From these and other experiments, Willows and colleagues conclude that Pd5 uses information received from a peripheral magnetosensor to modulate ciliary beating during crawling locomotion. To investigate how magnetotaxis might be adaptive for Tritonia, Willows (1999) conducted simple experiments in the field using scuba. In one, he picked up individual animals from the ocean bottom, transported them to new locations and located them 24 hours later. The results of this experiment provided little evidence for "homing" because only a few of the animals returned to the sites at which they were initially collected. Instead, Willows found that displaced animals usually moved in a shoreward direction. A possible reason for shoreward movement is that the animal's favorite food, the sea pen Ptilosarcus, is most plentiful in patches running parallel to the shore at depths of 5-30 m. In a critical experiment, Willows moved animals to a spot where the shore was now situated in a direction opposite to where it had been relative to the original collection site. When next observed, the animals had moved offshore from the new site, suggesting that shorespecific cues had not been used to orient. Instead, Willows proposes that the animals in this experiment oriented using geomagnetic cues. Although not all the relevant variables could be controlled in these field experiments, the results are nonetheless consistent with the idea that Tritonia uses learned geomagnetic cues to stay close to the shore or to reorient to it after displacements by strong water currents.

7

Feeding 7.1.

Adaptive Radiation of a Complex Behavior

As early gastropod forms dispersed geographically, natural selection favored adaptations that permitted exploitation of diverse habitats. The adaptive radiation of the Gastropoda is nowhere more evident than in the variety of feeding methods that have evolved in different species. In each of the diverse habitats in which gastropods reside, specializations of the feeding apparatus permit the animals to harvest whatever foods may be available, be it floating microorganisms, algal films growing on rock surfaces, large plant leaves, sedentary animals, or even live fish. Each food material demands its own means of acquisition, thus giving rise to several basic feeding types, named as grazers, raspers, suckers, collectors, cutters, and hunters. Obviously, different buccal structures and different neural programs are required for rasping versus sucking, collecting versus hunting. Less obvious, but no less interesting, are the different requirements for rasping on soft algae versus rasping on hard coral, for ingesting a worm versus ingesting a shelled mollusc. Because feeding behavior is fairly stereotyped and well defined, it offers an excellent opportunity for comparing neural control mechanisms in related species. Comparisons of this kind contribute to our understanding of how the nervous system evolves. The principal structures of the buccal mass, and their movements during feeding, are shown in Figure 7.1. The figure is based on observations in the pulmonate snail Lymnaea stagnalis, but its main elements apply at least approximately to other gastropods. An oral tube connects the mouth with an enlarged space called the buccal cavity. An important structure within the buccal cavity is the radula, which is one of the most characteristic of all structures in the gastropod body. It is a flexible chitinous ribbon studded with projecting teeth. The size, shape, and number of teeth, as well as the manner in which they are employed, determines the type of food that an animal can eat. The radula is attached to the odontophore (tooth bearer), whose function is to manipulate the radula into a position where it can rasp (or bite) the food, capture it, and pass it to the esophagus. The odontophore is a complex mass of cartilage and muscle that projects from the posterior wall of the buccal cavity. Some of its muscles are extrinsic, meaning that they originate in the body wall; when these muscles contract, the odontophore is 124

Feeding

125

Figure 7.1. Movements of buccal mass structures during feeding in Lymnaea stagnalis. At rest and between cycles, the odontophore holds the radula away from the mouth. The feeding rhythm comprises three phases. In the first phase, the odontophore is protracted and rotated, thus bringing the radula into contact with the food. The two subsequent phases both involve retraction. In the second phase, retraction of the odontophore causes the radula to rasp against the food and draw it into the buccal cavity. The third phase consists of a stronger retraction that pushes the food into the esophagus. Adapted from Benjamin (1983).

pulled or rotated within the buccal cavity. Other muscles are intrinsic, and some of these are responsible for moving the radula relative to the rest of the odontophore. In gastropods that bite their foods (e.g., Aplysia, Tritonia, Pleurobranchaed), the intrinsic odontophore muscles are used to open and close the two halves of the radula, thus allowing the radula to grasp food. Collectively, the oral tube, buccal cavity, odontophore, anterior portion of the esophagus, and all the associated muscles comprise a structural and functional unit known as the buccal mass. Feeding in freshwater gastropods such as Lymnaea and Helisoma begins when the animal opens its mouth and protracts its odontophore (Fig. 7.1). Simultaneously, the radular ribbon is also protracted and pulled over the tip (bending plane) of the odontophore. The radular teeth, which are arranged in a series of transverse rows along the radula, are erected during the process of protraction by their contact with the odontophore cartilage. As the odontophore moves forward over the food, the radula is now retracted and its teeth rasp the food. Next, the odontophore is retracted and the food particles are pulled into the anterior portion of the esophagus, where they

126 Behavior and Neurons in Gastropods

are transported to the digestive organs by peristalsis and ciliary action. Thus, the radula, assisted by the odontophore, acts as both a rasp and a conveyor belt. There are many variations on the foregoing scheme. For example, some gastropods have jaws, either as pairs or as single mid-dorsal structures; these are used to assist in securing large prey. Another structure, the proboscis, is common in carnivorous neogastropods and opisthobranchs; it permits an extension of the mouth to better reach the prey. Graham (1973) has given a complete account of the mechanics of feeding in "prosobranchs." A major evolutionary trend within this group is the progressive reduction in the number of buccal mass muscles. Monodonta, for example, which has a preponderance of primitive features, has 33 buccal muscles, whereas Urosalpinx, which has more advanced characters, has only 15 buccal muscles. This reduction is mainly attributed to an increasing reliance on the power of the odontophore's movements, as opposed to the control of its position and tension. Since many small tensor muscles are needed for fine control, these gradually disappear. Variations in the radular teeth signal important phylogenetic trends. In all animals, the teeth are arranged in a series of rows. In helicid land snails, for example, there are about 170 rows with about 160 teeth per row, or a total of 27,000 teeth. The total number of teeth in different species ranges from two to more than 150,000. Typically, each row has a single central tooth flanked by several lateral teeth and many more marginal teeth. The size, shape, and number of teeth is so characteristic of each species that the radula is often used as a key for species identification and classification. Several distinctive types are recognized in the "prosobranchs," as described below. The next three sections provide a survey of gastropod feeding strategies. For a more complete account, see Kohn (1983). 7.7.7.

"Prosobranchs"

The most primitive radular type (rhipidoglossan) is found in the Archaeogastropoda-Vetigastropoda clade. It consists of a large central tooth, flanked laterally on each side by five smaller ones and then a fan-like array of many more slender teeth. These animals (Haliotis, Monodonta, Nerita) mostly feed on microscopic or filamentous algae growing on rock surfaces. The radula protrudes from the mouth and licks at the rock. Limpets (Patella, Acmaea) use fewer, but heavier, teeth (docoglossan) to rasp at hardened growths such as coralline algae. Mesogastropods usually have a single central tooth and three lateral ones on each side (taenioglossan). Some species are raspers (Littorina), while others (Hydrobid) swallow detritus from the ocean floor and digest the microorganisms found within it. Another group (Crepiduld) captures phytoplankton suspended in the water. The plankton is drawn into the mantle cavity where it sticks to a mucus web formed on the ctenidium; it is then carried to the mouth by ciliary action. The carnivorous species in this group

Feeding

127

feed mostly on sessile invertebrates such as sponges, cnidarians, ascidians, and molluscs, but the heteropods (Carinaria, Pterotrachea, Cardiapodd) are exceptional because they are pelagic swimmers with well-developed eyes (Lalli and Gilmer, 1989). The heteropods actively hunt in the water column looking for fish, polychaetes, chaetognaths, and salps. The carnivorous lifestyle is even more widespread among the Neogastropoda. There are two radular types. The rachiglossate neogastropods have only three teeth per row, but each tooth is narrow and sharply cusped. These snails (Buccinum, Murex, Busycon) typically have a highly extendable proboscis. They feed on a wide variety of invertebrates, but they are especially adept at exploiting bivalve molluscs, which they overcome using a mixed strategy of prying, drilling, and chemical attack. The toxoglossate neogastropods have an extremely specialized radula and a very interesting mode of predation. The genus Conus is unusually large, containing 500 species of snails with cone-shaped shells. There are just two teeth in each radular row. The teeth are large, hollow, and endowed with as many as five barbs near the pointed ends. The teeth have both the appearance and the function of harpoons. While all Conus species are carnivorous, some species are generalists, whereas others specialize on molluscs, worms, or fish. In all cases, a single tooth is hydrostatically forced from the extended proboscis, and it pierces the skin of the prey. The victim is then injected, via the tubular teeth, with a potent mixture of toxic venoms. Once the prey becomes paralyzed, it is pulled into the Conus buccal cavity by retraction of the proboscis and swallowed whole. Cone snail venoms typically contain a vicious cocktail of 50-200 pharmacologically active components, most of which are small, disulfide-rich peptides (Mclntosh et al., 1999). The majority of the peptides seem to be highly specific ligands for ion channels. They are classified according to their actions on ion channels. For example, or-conotoxins inhibit ACh receptors; /Lt-conotoxins block sodium channels, 5-conotoxins delay sodium channel inactivation; /c-conotoxins block potassium channels, and &>-conotoxins block voltage-sensitive calcium channels. Acting together, these peptides rapidly immobilize the prey by blocking transmission at central synapses and at neuromuscular junctions. Both presynaptic and postsynaptic mechanisms are affected. Not only does each species use a large number of peptides, but the package of toxins differs greatly from species to species (Oliyera et al., 1999). Post-translational modifications are partly responsible for the great molecular diversity, helping to generate an estimated tens of thousands of Conus peptides. These findings raise many interesting questions. For example, why are so many chemicals required to immobilize the prey, and why do different species use different combinations? 7.7.2.

Opisthobmnchs

Although this subclass includes a range of feeding styles nearly as great as among the "prosobranchs," the majority of the animals are carnivorous

128 Behavior and Neurons in Gastropods

grazers. Favored prey include sponges, worms, and other sessile invertebrates. Although the radula vary greatly in appearance in different opisthobranch species, they are not strictly classifiable as to type. In many species, the radula is absent altogether, and the prey is swallowed whole. On the other hand, the coral-eating Umbraculum has 150,000 radular teeth. The following survey emphasizes the species of greatest interest to neurobiologists, and those with the most spectacular feeding habits, with some species satisfying both criteria. The sacoglossans are small animals specialized for fluid herbivory. They possess a single central tooth in each radular row. The size of the tooth varies slightly in different species but in all cases it precisely matches the size of cells in one particular species of algae. After the cell wall is pierced by the tooth, suction is applied by muscular actions in the buccal cavity until the cell's contents are drawn into the esophagus. Adding further interest to this mode of feeding is the fact that chloroplasts ingested by the gastropod are incorporated into cells lining the digestive gland where they continue to photosynthesize, thus satisfying a large fraction of the host's energetic requirements. The thecosomes are shelled, swimming pteropods. In the Limacinidae and related families, feeding animals float motionless in the water while they produce massive external mucus webs, which may grow to ten times the size of the animal itself (Lalli and Gilmer, 1989). The web spreads out above the pteropod to capture tiny phytoplankton, zooplankton, and protozoa. The web is eventually withdrawn and brought to the mouth by ciliary beating, and the food is ingested. Whereas Limacina uses mucus webs to capture food, this thecosome gastropod is itself hunted and eaten by the gymnosome Clione, as described below. The anaspids are known as "sea hares" because they look like hares and they eat large quantities of plant leaves (macroherbivorous). The feeding habits of Aplysia have been studied in the field by Kupfermann and Carew (1974) and Carefoot (1987). Observations in the Pacific waters of Southern California reveal that Aplysia californica eats nine species of seaweed including red, brown, and green types, but the red algae, Laurencia, is preferred. A. Juliana, on the other hand, prefers the green algae, Ulva, and A. dactylomela prefers the green algae, Cladophora. In all three species, the food that is eaten most by animals in nature is the one, amongst those available, that provides the best nutrition as measured by growth rates in the laboratory. Aplysia begins to feed by grasping a frond of algae with its lips (Kupfermann, 1974). The mouth is then opened, and the jaws and buccal mass are protracted. Just prior to the point of maximum protraction, muscles pull laterally on the radular membrane to cause a vertically oriented medial division of the radula. With further protraction, the open radula receives the algae, and when the radula then closes and retracts, a piece of food is ingested. A subsequent, slightly different sequence of protraction-retraction is required to bring the food into the esophagus; this swallowing action is performed without opening the mouth. Aplysia can also feed in a rasping

Feeding

129

mode, as when it scrapes encrustations from the surface of leaves (Kupfermann and Carew, 1974). The opisthobranchs Pleurobranchaea and Tritonia also bite with an open radula, but these animals are carnivores. They bite their prey by clasping it between the two halves of a protracted radula. Pleurobranchaea is an aggressive indiscriminate predator equipped with a proboscis. Tritonia prefers to feed on cnidarians (sea whips, sea pens, or soft coral); it does not have a proboscis but it strikes using a rapid forward thrust of its buccal mass through an open mouth. The order Cephalaspidea includes both herbivores and carnivores. Philine is a carnivore that feeds on polychaete worms and bivalve molluscs in the sublittoral sand. It has only one pair of teeth, which are shaped as hooks. When the radula is protracted, the teeth diverge. Prey become caught in the teeth when the radula is retracted, and they are ingested whole. More actively predacious is Navanax, a large bottom dweller that preys on other opisthobranchs. Like Philine, it also swallows its prey whole, and its buccal cavity has become highly adapted to this mode of feeding by the elimination of teeth, radula, buccal mass, and jaw. The buccal bulb (pharynx) is especially large and muscular. After locating the prey by following its mucus trail, Navanax approaches, then partially everts its buccal bulb. Immediately thereafter, it expands the buccal bulb to create a strong vacuum. As a result, the prey is sucked in and swallowed. Perhaps the most remarkable feeding habits are found in the Gymnosomata. Lalli and collaborators (Lalli, 1972; Lalli and Gilmer, 1989) have described an impressive array of structural adaptations in this order, some of which are illustrated in Figure 7.2. The gymnosomes are shell-less pelagic animals that number only about 50 species in six families, but their diversity of feeding organs is impressive. Most species have a radula and jaws, and some have a proboscis. In addition, there appears among the gymnosomes one feeding organ, the hook sac, that is unique to molluscs, and another feeding organ, prehensile arms, that is otherwise found only in cephalopods. The hook sac (Fig. 7.2) is part of the buccal mass. It contains an armory of curved, chitinous hooks, numbering 5-100 in different species, that are used to grab the prey and pull it into the buccal cavity. The arms, when present, are long flexible organs much like the tentacles of octopuses. In one family, the Pneumodermatidae, the arms are even equipped with suckers. The six arms of Clione, known as buccal cones (Fig. 7.2), have no suckers but they secrete a sticky substance from epithelial glands to help seize the prey. Clione limacina feeds exclusively on two species of thecosome snails, Limacina helicina and Limacina retroversa. As soon as a snail is contacted, Clione everts its cones by hydrostatic pressure and the prey is enveloped. The cones then manipulate the snail's shell so that its opening is aligned with the mouth of Clione. At this point, the hooks are ejected, again by hydrostatic pressure, and they penetrate into the flesh of the snail. Now the hooks and the radula are retracted, pulling the prey animal from its shell. The food is swallowed whole, and the intact empty shell is discarded.

130 Behavior and Neurons in Gastropods

Figure 7.2. Drawing of the pteropod Clione limacina feeding on its namesake, Limacina helicina. The cones and hooks are special adaptations for feeding on the live shelled prey. The cones secure the snail while the hooks remove it from its shell. Copyright 1989. Adapted from Lalli (1972) with permission from Elsevier Science.

7.7.3.

Pulmonates

Most pulmonates are herbivores, but they eat flesh when it is available. A few families stand out as predacious carnivores. The neural control of feeding has been especially well studied in the freshwater snails Lymnaea and Helisoma, as described below. These snails feed either by repeatedly rasping on stationary leaves or by grazing on thin algal growths. The radula of Lymnaea and Helisoma has about 100 rows of teeth, each with about 130 small teeth. During feeding, the buccal mass moves back and forth as shown in Figure 7.1 (see Section 7.3 for further description). When grazing, the head moves from side to side as the whole animal moves forward. Consequently, the snail leaves a zigzag pattern of small scrape marks on the algal film. Terrestrial species, (e.g., Helix, Limax and Achatina) usually feed by stationary rasping on plant leaves. Although pulmonates are often described as feeding generalists, studies in the field indicate that certain plants are preferred over others. Helix, for example, likes the coarse leaves of nettles (Urtica) and burdock (Arctium), and these plants are eaten to a far greater extent than predicted from their relative local abundances (Iglesias and Castillejo, 1999). Chemically, these plants contain high levels of protein, ash and calcium, all necessary for the growth of snails. Even when feeding on a single species, Thymus vulgaris

Feeding

131

(thyme herb) in the south of France, Helix selects one of six available phenotypes (Linhart and Thompson, 1995), each of which synthesizes a different monoterpene oil. When given a choice in controlled trials, snails prefer the linalol-containing phenotype over all others. Experiments showed that the phenotype bias is due to the avoidance of plants containing monoterpenes other than linalol, rather than an attraction to linalol. Other studies support the conclusion that food selection in pulmonates is usually based on the avoidance of unpalatable foods. This is because many plants produce chemicals to deter predation by herbivores. Wholly carnivorous pulmonates are uncommon. However, some snails, most famously Gonaxis and Euglandina, are voracious predators of other terrestrial molluscs. They have fewer teeth than the herbivores, but larger and more slender ones for slicing into their prey. Euglandina rosea has been intentionally introduced into some regions as a biological agent to control pulmonate pest species such as Achatina fulica. In most areas where Euglandina was introduced the results have been disastrous, with little effect on Achatina but decimation of indigenous species, especially tree snails.

7.2.

Food Finding

Hungry animals are faced with difficult perceptual problems relating to the detection and location of their potential food. Nearly all gastropods use their chemical senses to solve these problems regardless of whether they live in water or on land. The only group of gastropods known to rely on vision while hunting is the heteropods, which use large eyes to locate prey in ocean waters (see Section 7.1.1). The heteropods detect their small prey against a bright daylight background, and their acuity is good enough to allow them to attack from a distance of 60cm (Lalli and Gilmer, 1989); otherwise, little is known about how they find food. The first task of the olfactory system is to detect molecules (odorants) that are associated with acceptable foods. The animal must then direct its locomotion toward the source of the odor either by detecting a chemogradient or by moving upwind/upstream. Once close to the food, the animal combines olfaction, contact chemoreception and mechanoreception to maneuver its mouth directly in front of the food. A final taste test is performed just before ingesting the food to confirm its acceptability. 7.2.7.

food Odors

Foods, whether plant or animal, typically release hundreds of compounds into the air or water. In the aquatic environment, amino acids seem to be especially effective in attracting gastropods, as indeed they attract fish and other animals. In one study, tests were performed in an olfactometer using the freshwater snail Biomphalaria glabrata, an important host of the human parasite Schistosoma. Twenty-nine amino acids and several related compounds

132 Behavior and Neurons in Gastropods

were evaluated. Adult snails were strongly attracted to aspartic acid, glutamic acid, proline, and hydroxyproline (Thomas et al., 1980). Surprisingly, juvenile snails were not attracted to any of these amino acids, but instead they were attracted to citrulline, asparagine, and homoserine. On the other hand, similar tests with adult Aplysia californica gave results like those seen using adult Biomphalaria, that is, they demonstrated attraction primarily to glutamic acid and aspartic acid (see Fredman and Jahan-Parwar, 1980). Since glutamic acid is abundant in animal tissues, one might expect that carnivorous species would be especially sensitive to it. This seems to be true of the nudibranch Phestilla sibogae, which feeds exclusively on corals. Murphy and Hadfield (1997) recorded from afferent nerves and observed that when agonists of glutamic acid and glutamate (DNQX, NMDA, and kainic acid) were applied to the animal's tentacles, spike discharges were evoked in the nerves. Proteins are also attractive to aquatic gastropods, in some cases at concentrations even lower than for amino acids (see Kohn, 1983). On land, animals are attracted by volatile chemicals of low molecular weights. However, little attention has been given to the specific compounds that attract gastropods. Slugs readily feed on potato tubers, and one compound that smells very much like tubers, 2-ethyl-3-methyloxypyrazine (EMOP), attracts Limax maximus in behavioral assays and elicits neurophysiological responses (Gervais et al., 1996). Besides potatoes, beer has long been recognized as a strong attractant for slugs, and since many gardeners make practical use of this fact, efforts have been made to identify the active compound(s). Evidently alcohol is not itself attractive, nor are any of the aldehydes or ketones that are present in beer. The attractive chemicals, still unidentified, are mostly present in the unrefined fermentation products, namely yeast and wort. Eight volatile compounds considered likely to attract Achatina fulica were tested by Chase (1982) using a hand-held olfactometer that made it possible to deliver an odorant unilaterally to either one of the two superior tentacles. Only hexanoic acid, butyric acid, 2-octanol and amyl acetate caused consistent turning towards the side of stimulation. Surprisingly, a common leaf aldehyde, fraw.s-2-hexanal, was ineffective. The compound to which the snails were most sensitive was hexanoic acid. Achatina can detect hexanoic acid at a threshold concentration of 10~ 8 M, but humans detect the same compound at 10~ 10 M, and dogs respond at concentrations as low as 10~ 17 M. 7.2.2.

Plasticity of Olfactory Responses

Young gastropods, like all other young animals, must have an innate ability to sense food odors or they would quickly die. On the other hand, it is clear from experiments that gastropods can learn about food odors. What part of a animal's foraging behavior can be attributed to innate sensory responses and what part to learned associations is difficult to determine. Avila (1998) approached the problem by culturing specimens of the nudibranch Hermissenda crassicornis in the laboratory. Hermissenda is a generalist

Feeding

133

carnivore that feeds on cnidarians, timicates, small crustaceans, and nudibranchs, including conspecifics. However, the animals in the experiment were fed just a single food from the time of their larval metamorphosis, the cnidarian hydroid Tubularia crocea. When tested in a Y-maze for orientation based on olfactory cues (chemotaxis), the experimental animals were able to detect T. crocea and a second hydroid, Pennaria, but they were unable to detect a tunicate (Ciona), a mussel (Mytilus), and the anemones Metridium and Haliplanella. These results indicate the animals' innate abilities. Subsequently, after 2 weeks of conditioning in which the animals ate exclusive diets consisting of one of the previously undetected foods, they could detect most of the conditioned foods in the Y-maze trials. If conditioned to Ciona or Haliplanella, they now oriented exclusively to those prey odors, even ignoring the previously preferred Tubularia. These results indicate that learning is required before Hermissenda can detect certain prey items. However, certain other species of would-be prey, namely Mytilus and Metridium, never elicited orientation behavior in the Y-maze, even after conditioning. A Y-maze was also used by Croll and Chase (1980) to study plasticity in the olfactory orientation of land snails, Achatinafulica. The snails were fed an exclusive diet consisting of a single vegetable (carrot, lettuce, or cucumber) for various lengths of time before they were given a choice between the odor of the food that they had been eating and the odor of an alternative food. For juvenile snails, 5-50 days of age, exposure to a food for as little as 12 hours was sufficient to significantly bias their orientation. The duration of the memory for the learned food-odor depended on the length of exposure and, to a lesser extent, on the age of the snail. In one experiment with adult snails, animals were fed an exclusive diet of either cucumbers or carrots for 86 days. When they were then presented with a choice in the Y-maze between the odor of cucumber and the odor of carrot, the snails preferentially oriented toward the vegetable to which they had been conditioned. To study the duration of the memory, tests were given in subsequent weeks while the snails ate only lettuce. A significant preference for the conditioned odor continued to be expressed for 120 days after the end of the conditioning period. In addition to affecting the direction of locomotion, food-odor conditioning produces subtle changes in the bearing of the tentacles. Snails raise their tentacles when they perceive an odor that is unknown to them, but they lower their tentacles after perceiving a conditioned odor (Peschel et al., 1996). The learning in these situations is associative and comparable to pavlovian conditioning, because, if the snail is not allowed to ingest the food at the same time as it smells it, learning is prevented (Croll and Chase, 1980; Ungless, 1998). In addition to appetitive learning, discussed above, herbivorous gastropods also learn to avoid foods that are experienced as unpalatable. Avoidance is based partly on gustatory cues, discussed below in Section 7.6, and partly on olfactory cues. Sahley et al. (1981) demonstrated that the usual orientation of slugs, Limax maximus, to the odor of potato could be markedly reduced by a brief conditioning procedure. Slugs were allowed to

134 Behavior and Neurons in Gastropods

feed on a potato concentrate for 2 minutes before they were exposed to the bitter taste of quinidine sulfate. When tested 24 hours later in an open field olfactory arena, the slugs avoided the potato odor, although they continued to orient to carrot roots and other foods. Just a single pairing of potato and quinidine was sufficient to produce avoidance. 7.2.3.

Orientation to Foods

Feeding, like other motivated behaviors, is often viewed as comprising two phases, appetitive and consummatory (Kupfermann, 1974). The appetitive phase is preparatory and primarily involves orienting the animal to the food. Ingestion of the food occurs in the consummatory phase. All gastropods except heteropods rely on olfaction to detect distant food sources. Finding the food then involves either orienting to the odor gradient, called chemotaxis, or simply moving upwind or upstream, called either odor-gated anemotaxis or odor-gated rheotaxis depending, respectively, on whether the medium is air or water. Although the entire head region is abundantly supplied with chemoreceptors and mechanosensors, it is the tentacles and rhinophores that are principally involved in the orientation to distant food sources (see Croll, 1983). This is easily demonstrated by lesion experiments. For example, removal of the posterior tentacles prevents snails from determining which side of a Y-maze contains the food (Chase and Croll, 1981). The special chemosensitivity of the tentacles and rhinophores is also established by functional tests using local applications of chemical stimulants, and by histological counts of receptors (reviewed in Croll, 1983; Emery, 1992; Chase and Tolloczko, 1993). Since the tentacles are present as bilateral pairs, locating an odor source could be accomplished by an animal making simultaneous comparisons of the odor concentration at the left and right tentacles, with consequent movement in the direction of the highest concentration, called tropotaxis. Alternatively, the tentacular sensors could function independently to monitor the change in concentration over time, adjusting the direction of locomotion to maintain a perception of ever increasing concentrations, called klinotaxis. To determine which of these methods is used by Achatina, Chase and Croll (1981) observed olfactory orientation after unilateral lesions of the posterior tentacles. The results depended on environmental conditions. When the snails were required to orient in the absence of any winds, they needed the pair of tentacles, implying the use of tropotaxis. However, when orienting to an upwind odor source, a single tentacle was sufficient, implying the use of klinotaxis. Those gastropods that hunt molluscan prey can sometimes find a prey animal by following mucus trials. This method is used by some opisthobranch carnivores (e.g., Navanax and Naticd) and probably by all pulmonate carnivores. The process has been well described in the terrestrial snail Euglandina rosea. This animal is a voracious predator of snails and slugs, and it relies heavily on the mucus trails of its victims to find them. Cook (1985a) observed that Euglandina hardly ever passes up an opportunity to follow a trail, and

Feeding

135

once it begins it may continue to follow the same trail for 20 cm or more. The lips of Euglandina are unusually extended and mobile, and sensitive to both touch and chemicals. Thus, Euglandina uses its lips repeatedly to probe the ground to maintain contact with the trail. Interestingly, experiments indicate that Euglandina, as well as Haplotrema concavum, another terrestrial snail predator, are able to detect which species of gastropod has laid the trail (Pearce and Gaertner, 1996). The evidence suggests that predators forage preferentially for those species that offer the best nutrition at the least cost. Obviously, it would be advantageous if a predator could detect the direction in which a trail was laid, but neither the terrestrial species nor Navanax appear to have this ability. Once an animal has approached its food object, it must then place its mouth directly in line with the food. Although chemical stimuli are paramount in the early stages of appetitive orientation, their importance for proximal orientation is less certain. In Aplysia, for example, the mouth moves more accurately towards a purely tactile stimulus (a glass rod) than towards local concentrations of seaweed extract, provided the animal has first been aroused with food chemicals (Teyke et al., 1990a). This result is consistent with the fact that a tactile stimulus will usually provide better information about locus than a chemical stimulus. However, high threshold chemoreceptors, also called contact chemoreceptors, could provide information about locus, as well as about identity, if they responded to chemicals at threshold concentrations that were equivalent to those present at the physical surface of the food. There is no histological or electrophysiological evidence for high threshold chemoreceptors in gastropods, but their presence is suggested by observations that animals sometimes respond to a food when it is touched to their skin, whereas the same animals fail to respond to a purely tactile stimulus, for example, a piece of paper, even in the presence of food chemostimulants. To summarize, the early stages of appetitive arousal and locomotor orientation are mediated exclusively by olfactory cues in nearly all cases. Chemoreception is probably less important in the final stages of food localization, but once again it becomes crucial for initiating the consummatory phase of feeding, as discussed in Section 7.5.2. The foregoing scenario is supported by a study of the innervation of the lip in Lymnaea stagnalis (Nakamura et al., 1999). Two cerebral nerves innervate the lip region. One of these, the superior lip nerve, innervates the immediate circumference of the mouth, and fibers in the nerve fire action potentials when sucrose is applied to the lip. The other nerve, the median lip nerve, innervates a slightly more peripheral region around the mouth and does not respond to sucrose. Thus, it is proposed that the median lip nerve, but not the superior lip nerve, mediates the localization of food based on mechanosensory signals, while the superior lip nerve mediates the consummatory phase based on chemical cues. The final stages of orientation to a piece of food has been examined most thoroughly in Aplysia californica. The head of Aplysia is highly flexible and can be bent in all directions. The main muscles used in turning the head are

136 Behavior and Neurons in Gastropods

longitudinal bands that lie dorsally and laterally in the neck. Afferent signals from food stimuli travel through the cerebral ganglia, both directly and indirectly, to excite motoneurons located in the pedal ganglia and the pleural ganglia; other afferent pathways may converge on motoneurons located in the cerebral ganglia. When the animal encounters food it first raises the anterior two-thirds of its body so that the head can rotate free of the substrate (Kupfermann, 1974). If a small piece of seaweed is now briefly touched to the head, the animal turns towards the stimulus but it usually overshoots the target (Teyke et al., 1990a). The amplitude of the overshoot depends on the location of the target, with positions most distant from the mouth associated with the largest overshoots and positions near the mouth associated with the smallest. Although overshooting is common when the stimuli are only briefly presented, animals can accurately bring their mouths to the food if the food stimulus is continuous, or when the animal is provided with feedback about its changing position relative to the animal. These observations led Teyke and co-workers to suggest a model in which initially strong motor commands are progressively substituted for weaker ones as the stimulus is moved across the head towards the mouth. Carnivorous predators eating shelled gastropods have the special problem of accessing the prey's soft parts. As mentioned previously, Clione limacina must manipulate the shell of its prey so that the aperture is pressed against the mouth of Clione; only then can it ingest the snail's body while rejecting the shell (Lalli and Gilmer, 1989). Euglandina rosea employs a strategy akin to that of a lion's attack when it hunts the large snail Succinea campestris. It strikes a biting blow on the neck of the prey to sever the columellar muscle and thus disable any defensive withdrawal into the shell. In 35% of initial attacks, this strategy is successful in rendering the victim incapable of escape (Cook, 1985b). 7.2.4.

Central Neural Mechanisms

The neural pathways that mediate orientation to distant food sources are those that begin in the tentacles and terminate in or pass through the cerebral ganglion (see Section 3.1; Fig. 3.3). Some of the signals generated in the tentacles are transmitted directly to the CNS, whereas others are routed through peripheral ganglia. Synaptic integration occurs in the peripheral ganglia, as shown by the reduction of centripedal electrical responses when the ganglia are bathed in zero-calcium saline solutions (Fredman and Jahan-Parwar, 1980; also see Chase and Tolloczko, 1993), but the significance of peripheral integration, and its relation to central integration, is unknown. The procerebrum is a remarkable structure found only in the brains of stylommatophoran pulmonates, that is, terrestrial snails and slugs (Van Mol, 1974). Its olfactory function is indicated by the fact that it receives axons from the tentacle nerve, which innervates the posterior tentacle, and from the medial lip nerve, which innervates the inferior tentacle

Feeding

137

(Chase and Tolloczko, 1993). The procerebrum is one of three recognizable lobes in the cerebral ganglion (see Figures 2.6, 3.3, and 7.3). The number of neurons present in the combined left and right cerebral lobes is estimated to be 40,000 in both Helix aspersa and Achatina fulica, and 100,000 in Limax maximus (Chase, 2000). These numbers suggest that the procerebrum contains at least half of all neurons in the central nervous system. The cells are small, 6-8 um in diameter, and densely packed in a mass lying adjacent to an associated neuropil, as shown in Figure 7.3. Although the cell bodies of procerebral neurons appear uniform in histological sections, when individual cells are injected with biocytin large variations in neurite morphology are revealed (Ratte and Chase, 2000). Some cells possess long neurites that extend into the main neuropil of the cerebral ganglion. Since these neurites have mostly output synapses on their distal processes, but mostly input synapses near the cell body, the centrally projecting cells probably carry processed outputs from the procerebrum to motor control neurons, including those responsible for food finding.

Figure 7.3. The procerebral lobe of terrestrial molluscs is unlike any other central nervous structure in gastropods. In the slug Limax maxima, shown here, each lobe contains approximately 50,000 neurons. The procerebrum is involved in processing olfactory information, but its precise role is unknown. Limax is advantageous for neurophysiological and functional studies of the procerebrum because the lobe in this species is large and substantially separated from the rest of the cerebral ganglion; compare, for example, Achatina, shown in Figure 2.6. (A) and (B) Photomicrographs of the desheathed lobe. (C) The cell layer seen at higher magnification in a section 6 um thick. The cell membranes are stained with a voltage-sensitive dye; scale bars are 100 um. From Kleinfeld et al. (1994) with permission.

138 Behavior and Neurons in Gastropods

Much of the interest in the procerebrum stems from the discovery of its electrical oscillations (see Gelperin, 1999). The oscillations can be routinely recorded as field potentials using any type of low impedance electrode placed in or around the procerebrum, as shown in Figure 7.4. They occur continuously, even in the absence of chemical stimulation, and they have been recorded in vivo as well as in vitro. They occur with different waveforms and with slightly different periodicities in different species, but the period is generally about 0.7 Hz. Optical recording methods have been used to study the spatial distribution of the oscillations. In Limax maximus, they propagate as a continuous wave from the apical end of the lobe to the basal end, at a rate of about 1 mm/second (Gervais et al., 1996). Kleinfeld et al. (1994) used a combination of optical measurements and perforated patch whole-cell recordings to elucidate the cellular mechanisms responsible for the waves.

Figure 7.4. Spontaneous electrical oscillations in the procerebral lobe of Limax, and their perturbation by olfactory stimuli. Local field potentials (LFP) were recorded in the procerebrum (PC) while an electro-olfactogram (EOG) was recorded from the surface of the olfactory epithelium in the tip of the superior tentacle. Saline-filled glass electrodes were used at both sites. (A) A control trial in which moist room air was gently puffed over the olfactory epithelium. (B) A puff of air containing amyl acetate, an aversive odor for Limax. (C) A puff of EMOP, a synthetic substance that has the smell of potatoes and to which Limax is attracted. From Gervais et al. (1996) with permission.

Feeding

139

They found that about 1 % of the procerebral neurons are of a type that fires periodic bursts of action potentials. These bursting cells are slightly larger than neurons of a second type that are synaptically inhibited by the bursters (Watanabe et al., 1998). The field potential oscillations arise from the summed synaptic currents in the inhibited cells, while the phase gradient along the length of the procerebral lobe is caused by differences in the excitability of bursting cells in apical and basal regions. The precise function of the procerebrum, and the role of its oscillations, is controversial. In some experiments, stimulation of a tentacle with odors produces changes in spontaneous oscillations, including amplitude and frequency effects, as illustrated in Figure 7.4. In addition, when Gervais et al. (1996) studied the system using optical recordings, they found that odor stimulation causes a collapse of the phase gradient across the procerebrum, that is, the oscillations momentarily cease to propagate as a wave. However, none of these effects is consistently reported by all workers (Chase, 2000). Schiitt et al. (1999) carried out a spectral analysis of odor-induced effects. They report that five pure chemicals and two mixed odors each cause specific changes in the oscillations, with each odor producing oscillations at a characteristic peak frequency. Despite all these reports of odor effects on spontaneous procerebral oscillations, it needs to be noted that suppression of the oscillations does not completely eliminate odor-evoked responses in efferent cerebral nerves (see Gelperin, 1999). This last result, plus others, led Gelperin (1999) to propose that the procerebrum is primarily a learning machine, that it is more involved in odor discriminations than in odor recognition, and that the function of its oscillations is to produce spatial segregation of learned odor representations. By using a fine wire recording electrode implanted in the procerebrum of freely behaving slugs, Cooke and Gelperin (2001) were able to establish that the oscillations occur in vivo. This is important because it shows that the oscillations are not an artifact of the in vitro preparations in which they are usually studied. Odor presentations caused increases in the frequency and the amplitude of the oscillations, but no differential effects were detected with different odors. Moreover, there were frequent changes of frequency and amplitude that could not be explained by any environmental factor nor any observable movement of the animal. The authors suggest that these "spontaneous" variations reflect an on-going modulation of the local field potential, but for what purpose is unknown. There is no procerebrum in opisthobranchs, and our current knowledge about the central processing of olfactory signals in these gastropods is only sketchy. Studies by Fredman and Jahan-Parwar (1980) suggest that the A and B clusters of the cerebral ganglion (see Fig. 6.7) may be important sites for mediating orientation to foods. When the anterior tentacle was stimulated with extracts of seaweed, cells in these clusters became depolarized by synaptic excitation and they fired action potentials. When the cells were driven by intracellular stimulation, they evoked contractions of the tentacles, the foot, and the parapodia. Neurons in the A and B clusters also interact with circuits

140 Behavior and Neurons in Gastropods

controlling locomotion. Thus, these cells could mediate the postural and locomotor adjustments necessary for food localization. Significantly, neither the A cells nor the B cells affect the feeding motor program. Thus, their roles differ from that of other identified cerebral neurons that also respond to food chemicals, but that initiate or modulate the consummatory phase of feeding (see Section 7.5). Neurons possibly homologous to those described above have been found in Helix, in a similar posterior location within the cerebral ganglion (Kemenes, 1994). Some of these cells are excited when chemical or tactile stimuli are applied to the lips, whereas others are inhibited. Stimulus specificity and receptive field location differs from cell to cell. The axons of these cells project from the cerebral ganglion into the lip nerves. Although the cells are clearly not primary sensory neurons, the fact that their axons are present in nerves that carry afferent signals from the periphery suggests that they might be part of a feedback loop that is responsible for positioning the mouth relative to contacted food objects.

7.3.

Central Pattern Generators

The consummatory phase of feeding is driven by motor commands generated within the buccal ganglion. The central pattern generators that are responsible for feeding are similar to those underlying respiration and locomotion in that they consist of a small number of neurons, mostly interneurons, that interact with one another to produce phasic outputs. Studies of feeding CPGs have been most informative in the Basommatophora, in part because the investigations demonstrate a high degree of similarity between the CPGs of different species. One general feature of the feeding motor program in basommatophoran snails that appears to be consistent across species is the triphasic character of the basic rhythm (Murphy, 2001). This is unusual because the motor programs for respiration and locomotion are biphasic. It remains to be seen whether the feeding CPGs of other gastropod taxa are also triphasic (see below). The CPG of Lymnaea has been described in detail by Paul Benjamin and colleagues. It will be discussed here in detail because the methods and results from Lymnaea are generally applicable to related gastropod species, albeit with qualifications to be highlighted below. The analysis of feeding in Lymnaea began with electromyogram recordings and cinephotography (Rose and Benjamin, 1979). From these records it was apparent that ingestion comprises a repeated sequence of three distinct buccal mass movements, or phases, with a quiescent period of variable length interposed between the end of one cycle and the beginning of the next cycle (Fig. 7.1). Protraction of the odontophore occurs first, followed by two different types of retraction, specifically, an initial rasping phase in which the food is brought into the buccal cavity, and a subsequent swallowing or hyper-retraction phase in which the food is taken to the esophagus. Altogether, the buccal mass uses 46 muscles

Feeding

141

to effect these movements. However, most of the muscles are present as bilateral pairs and many act in concert, so it is not necessary to investigate the actions of each individual muscle. Also, most muscles are active during only one phase of the feeding motor program. Therefore, in semi-intact preparations protraction is indicated by electrical activity recorded from the posterior jugalis muscle, rasping is indicated by activity in the lower anterior jugalis, and swallowing is indicated by activity in the upper anterior jugalis. Control of the muscles rests with about 400 neurons situated in each of the buccal ganglia. Rose and Benjamin described ten putative motoneuron cell types, suggesting that there are at least that many functional muscle groups. The three phases of feeding behavior can be elicited in semi-intact preparations by presenting sucrose to the animal's lips. In fully reduced preparations consisting of the cerebral and buccal ganglia, rhythmic activation of motoneurons and interneurons (fictive feeding) is usually produced by intracellular driving of the identified interneuron SO (see Section 7.5.2), although a number of other neurons are also effective. The core of the CPG in Lymnaea consists of three groups of interneurons, named Nl, N2, N3, as illustrated in Figures 7.5 and 7.6. Each group fires during one of the motor phases described above (i.e., protraction, rasping, or swallowing). Cells were identified as members of one or another group if they fired in the appropriate phase, had the correct synaptic effect on motoneurons, and were capable of resetting the rhythm when injected with electrical currents (Elliott and Benjamin, 1985a). In early studies, the motoneurons were considered to be entirely downstream of the CPG, but Staras et al. (1998) later made the case for including at least four motoneurons, representing all three phases of the feeding program, in the CPG. The argument rests on the finding that manipulation of the spiking activity of these motoneurons during fictive feeding resets the feeding rhythm. The motoneurons are electrically coupled to the interneurons in the CPG (DC coupling coefficients, 12-36%), and feedback to the interneurons from the motoneurons is functionally important. For example, firing of the protraction phase motoneuron B7a excites the protraction phase interneurons N1M and thereby contributes to a key build-up of activity during the protraction phase. In another case, the phase 3 motoneuron B4 indirectly inhibits the N1M cells via its coupling to a phase 3 interneuron, thus delaying the onset of a new feeding cycle. It is suggested that this latter interaction is responsible for limiting the feeding motor program to a range of frequencies sustainable by the buccal muscles. Models of the central pattern generator in Lymnaea have become progressively more complex as new cell types are discovered. In a recent version (Brierley et al., 1997b), all of the original interneuron groups, Nl, N2 and N3, have been subdivided to yield a total of six cell types. The firing patterns and synaptic connections of these cells are illustrated in Figure 7.6. Most of the synaptic interactions are presumed to be monosynaptic based on paired intracellular recordings. Also shown in Figure 7.6 is SO, which is an interneuron that can initiate rhythm generation in the network when it is depolarized.

142 Behavior and Neurons in Gastropods

Figure 7.5. Cells and circuits that control feeding in Lymnaea stagnalis. (A) The pair of buccal ganglia is shown twisted, with the ventral surface illustrated at the left and the dorsal surface at the right. In physiological experiments, such twisted preparations allow simultaneous recordings from both sides of the ganglia. Cells designated "B" are motoneurons, "N" cells are central pattern generator interneurons. All cells are present bilaterally except for SO, which is present as a single cell in either the left ganglion or the right ganglion. From Brierley et al. (1997b). (B) Schematic illustration of functional compartments in the neuronal network. Locations of the identified buccal ganglion cells are shown in (A); MGC and CV1 are cerebral ganglion neurons. Solid circles, inhibitory chemical synapses; bars, excitatory chemical synapses; crooked lines, electrical connections. Abbreviations ("motor" compartment): P, protraction; R, rasp; S, swallow. From Benjamin et al. (2000) with permission.

Figure 7.5A shows the locations of the CPG interneurons, SO, and some of the motoneurons in the buccal ganglion. A plausible model of the operation of the feeding CPG can now be described (refer to Figs 7.5 and 7.6). Each cycle of the feeding motor program begins with the protraction phase, which is driven by a burst of

Feeding

143

Figure 7.6. Summary of activity in the central pattern generator (CPG) for feeding in Lymnaea. Six types of interneurons are shown, plus the SO cell, which is not itself part of the CPG, but which can initiate the feeding motor program when tonically depolarized, as in this example. Four feeding cycles are shown separated by vertical bars. Phases within each cycle (dotted lines) are defined in terms of the active interneurons (Nl, N2, N3, at the top) and the types of odontophore movements (P, protraction; R, rasp; S, swallow, at the bottom). Synaptic interactions are indicated by arrows: e, excitatory; I, inhibitory; resistor, electrical connection. The black boxes illustrate multiple connections of the SO with CPG interneurons. Different connections are shown in each cycle, as indicated by the boxes and the associated connections with SO. From Brierley et al. (1997b) with permission.

firing in the Nl cells. Excitation of the Nl cells comes primarily from synaptic inputs, although weak endogenous bursting has also been reported. The bilateral pair of NIL cells is excited by the SO cells through electrical synapses; the NIL cells are themselves electrically coupled to each another. The N1M cells, one on each side, are excited by SO via strongly facilitating chemical synapses, and by NIL cells, also via chemical synapses.

144 Behavior and Neurons in Gastropods

The N2 cells, responsible for the rasping phase (retraction), fire next. They comprise a single pair of dorsal cells (N2d) and a single pair of ventral cells (N2v). The N2 cells are inhibited early in the protraction phase by SO and NIL, but later in the same phase SO and NIL excite the N2 cells. A plateau potential is triggered in the N2v cells and it quickly spreads to the N2d cells through electrical coupling. N2v fires throughout the rasping phase, whereas N2d fires only briefly at the beginning. Differential synaptic connections of the N2v and N2d cells with motoneurons are responsible for coordinating the sequence of contractions in the transition from rasping to swallowing (Brierley et al., 1997a). The motoneurons for rasping are excited during the N2 plateau potentials, while the motoneurons for swallowing are strongly inhibited during the N2 plateau potentials and they fire only afterwards by postinhibitory rebound. In addition to controlling rasp phase motoneurons, the N2 interneurons inhibit both the Nl interneurons and the N3 interneurons, thereby terminating bursts in the former cells and delaying bursts in the latter cells. The interneurons that drive the swallow phase (hyper-retraction) comprise the phasic cells N3p, the tonic cells N3t, and a newly discovered group of octopamine-containing cells (Vehovszky and Elliott, 2001). The first two types are represented by at least one pair of neurons, the octopamine group by three electrically coupled cells. The N3 cells fire, by postinhibitory rebound, at the termination of the plateau potentials in the N2 cells. The N3p and N3t cells differentially excite specific groups of motoneurons responsible for swallowing. Their activity also provides feedback inhibition to SO and the Nl interneurons. As activity in the N3 cells wanes, depolarization predominates in SO and Nl, leading to the next cycle of activity in the network. The octopamine (OC) interneurons are not considered part of the CPG, but they are rhythmically active during the swallowing phase. They modulate the feeding motor program by interacting synaptically with many of the identified neurons in the buccal ganglion, including both interneurons and motoneurons (Vehovszky and Elliott, 2001). The nature of their synaptic influences are diverse: electrical and chemical; excitatory, inhibitory and biphasic; short duration and long duration (>10 seconds in some cases). Although they fire in phase with the N3p and N3t cells, with which they are electrically coupled, their output connections are different. Activity in the OC cells significantly increases the rate of the feeding motor program by decreasing the Nl and N2 phases. In addition, the OC cells selectively affect motoneuron involvement, for example the OC cells strongly inhibit B3, while weakly exciting other motoneurons. Feeding in another freshwater snail, Helisoma trivolvis, is also characterized by three phases of buccal mass movements. As is the case for Lymnaea, the CPG of Helisoma consists of three subunits, each of which contains a semi-independent neuronal oscillator that drives sets of motoneurons during appropriate phases of the feeding cycle (Quinlan and Murphy, 1996; Murphy, 2001).

Feeding

145

Generalizations about the neural organization of the feeding motor program in taxa other than the Basommatophora are problematic due to variations in feeding methods and incomplete neuronal characterizations. For example, while many elements of the neural circuit have been described in Aplysia, the operation of the CPG is not entirely understood. Some of the identified neurons involved in the generation and expression of feeding motor programs in Aplysia are illustrated in Figure 7.7, together with their synaptic connections. Several of the illustrated cells are motoneurons, including B31 and B32, which innervate the major protractor muscle, 12. B31/B32 may be key players in initiating feeding because they receive extensive convergent inputs (Fig. 7.7), and their activity seems to be required for expression of feeding motor programs. The identified cell B8 is a radula closer motoneuron. A higher level of control is exerted by interneurons such as B63 and B34, which initiate protraction by exciting the B31/B32 cells and inhibiting retraction neurons. Neuron B64 has a reciprocal function. It excites retractor motoneurons, while inhibiting protractor interneurons and motoneurons, including B31/B32 (Hurwitz and Susswein, 1996). This cell, B64, is particularly interesting because it shares several properties with the N2v cells of Lymnaea (Brierley et al., 1997a). Most importantly, both cells appear to control the phase shift from protraction to retraction. Also, both B64 and N2v produce plateau potentials and both fire continuously while depolarized. Finally, the two cells have similar morphologies. Thus, B64 in Aplysia and N2v in Lymnaea appear to be functional analogues, and they are possibly homologues. While the buccal ganglion controls movements of the buccal mass, other ganglia control movements of the lips, mouth, and body wall during feeding. The need for coordination amongst these organs is apparent from the fact that efferent copies of the buccal motor program are transmitted to other ganglia, mainly the cerebral and pedal ganglia. In Lymnaea, this probably occurs via axonal projections of the Nl interneurons. Because the buccal input to the pedal ganglion is mostly inhibitory (Kyriakides and McCrohan, 1988), it is thought that this input may prevent movements that could interfere with feeding. An alternative interpretation of the buccal input to the pedal ganglion is that it could drive contractions of the foot and body wall, which, if appropriately timed, might aid the ingestion of food particles. In Aplysia, each cerebral ganglion (left and right) has a CPG that interacts with the buccal CPG to provide coordinated movements of the lips and buccal mass (Perrins and Weiss, 1996). The fact that many CPGs are clearly triphasic (e.g., Lymnaea, Helisoma, Tritonia) led Murphy (2001) to propose that the rasping mode of feeding in gastropods is based, in all cases, on a triphasic CPG. However, some CPGs of rasping snails have been described as biphasic. For example, Arshavsky et al. (1988) found seven groups of rhythmic neurons in the buccal ganglion of Planorbis corneus, but only two of these were categorized as "influential" in the sense that current injection into the cells could reset the feeding cycle. These workers therefore concluded that the CPG in Planorbis is biphasic, not triphasic. Murphy believes that, in this case as well as others (e.g., Aplysia,

146 Behavior and Neurons in Gastropods

Figure 7.7. Components of the neural system responsible for feeding in Aplysia. Buccal neurons active during the protraction phase are shown as dark circles; neurons active during the retraction phase are shown as white circles. Note that synaptic connections are generally excitatory between cells that are active during the same phase, but inhibitory between cells active during different phases. Neurons that can bias activity toward either ingestion movements (B51) or rejection movements (B34 are shown with bold circles. The cell labeled MCC is the same cell referred to as MGC in the text. From Lechner et al. (2000), who caution that the drawing is "not a comprehensive description of available data." Copyright 2000 by the Society for Neuroscience.

Clione, Helix, Pleurobranchaea), the descriptions of biphasic CPGs simply represent incomplete characterizations. To support his proposal, Murphy (2001) documents several instances of apparent interspecific homology involving CPG interneurons and motoneurons. Candidate homologies are identified for each of the three phases.

Feeding

147

For example, in phase 1 (protraction), the neurons NIL in Lymnaea, Nla in Helisoma, and B65 in Aplysia appear to be homologues. Also, the pair of buccal-cerebral interneurons called N1M in Lymnaea have morphological and physiological properties very similar to the three pairs of BCN1 neurons in Helisoma, as do B34 and B63 in Aplysia. In phase 2 (rasp), the interneurons N3v in Lymnaea and B2 in Helisoma are likely homologues, and all interneurons in both species use glutamate as their primary transmitter (Brierley et al., 1997b). Murphy rejects any homology of these last mentioned neurons with neuron B64 of Aplysia and "group 2" neurons of Planorbis, even though both are active during phase 2, because neither is capable of resetting the feeding rhythm; therefore, neither is a true member of the CPG. Comparisons among phase 3 (swallow) interneurons are even more difficult. Neither the N3t cells nor the N3p cells of Lymnaea match the N3a cells of Helisoma, and no interneuron in Aplysia is similar to any of the preceding ones. Vehovszky and Elliott (2001) remark on the fact that the OC cells of Lymnaea "reconfigure" the buccal system, as does N3a, but the details of their actions do not seem to match those of N3a. Thus, further comparative studies, especially of phase 3 neurons, are necessary to further substantiate Murphy's idea of a universal feeding CPG.

7.4.

Variations of Buccal Motor Programs

Even after an animal has bitten into what it thinks is a piece of food and even after it has begun to ingest the apparent food, signals may still be generated signifying that the object is, in fact, unpalatable. Kupfermann (1974) noted that in such a circumstance Aplysia will actively reject the would-be food item. During rejection, the radula closes as it protracts, whereas during ingestion, the radula closes as it retracts. Also, rejection involves a predominance of protraction over retraction, whereas the opposite is true for ingestion (Morton and Chiel, 1993). Studies of feeding motor program generation have shown that a variety of rhythmical patterns can be recorded in a single species, and even within a single preparation. At least some of these patterns control the rejection of foods, but certain authors prefer to speak only of "normal" feeding patterns and "non-standard" patterns (Quinlan and Murphy, 1996). Arshavsky et al. (1988), working with Planorbis, described three patterns: mode A is assumed to be the normal feeding pattern; mode B is characterized by the absence of a protraction phase; and mode C involves repeated cycles of protraction and retraction with no intervening rests. The buccal ganglion of Aplysia is described as generating "at least six distinct but related programs" (Hurwitz et al., 1999). To understand the behavioral significance of multiple buccal programs, it is necessary to relate the activity of nerve cells and muscles recorded in vitro to buccal movements and food passage recorded in vivo. Unless this is done, the results from in vitro work can be misleading, especially when the CNS

148 Behavior and Neurons in Gastropods

is completely isolated, because proprioception, which is absent in such preparations, plays an important role in regulating buccal movements (see Section 7.5). The first correlational studies of this type were done by Croll and colleagues using Pleurobranchaea (Croll et al., 1985a). When Pleurobranchaea is presented with a piece of squid, the animal ingests the food; however, if a small amount of detergent or alcohol is put into the mouth, the substance is ejected (rejected). In addition to ingestion and ejection, Pleurobranchaea exhibits an ambiguous type of buccal mass movement, called "neutral" activity, that occurs spontaneously but has no known function. Croll and colleagues studied the control of these movements initially with electromyograms recorded in intact animals, then neurophysiologically in "reduced" preparations consisting of the buccal mass with attached buccal and cerebral ganglia, and ultimately in isolated buccal ganglia. Their results, summarized in Figure 7.8, show that ingestion is correlated with a motor pattern dominated by retraction, whereas egestion is correlated with a motor pattern

Figure 7.8. Schematic of nerve and muscle activity during three types of feeding motor programs in Pleurobranchaea californica. Position and length of bars represent position and duration of activity; thickness represents relative intensity or amplitude. Muscles are identified as "m"; the salivary duct is "s.d." All other abbreviations refer to nerves. The structures indicated by arrows are those whose activity is most useful for distinguishing motor programs. The approximate period length of the ingestion program is 5.5 seconds; the period lengths of the other programs are relatively longer. Adapted from Croll and Davis (1982).

Feeding

149

dominated by protraction. Quantitatively, the motor program type is predictable from the protraction phase duty cycle because duty cycles of less than 50% are characteristic of ingestion, while duty cycles greater than 50% are characteristic of egestion. The activity of certain individual muscles also distinguish between ingestion and egestion. For example, the buccal constrictor M5 is active during ingestion but suppressed during egestion. Lastly, activity recorded in the salivary duct (and its nerve) is also highly predictive of the motor pattern because the duct is active during ingestion but not active during egestion. Similar criteria have been developed for distinguishing patterns of ingestion and egestion in Aplysia (Morton and Chiel, 1993), and these criteria have been employed in recent studies of operant conditioning (see Section 7.6). The feeding motor programs provide an additional example of the multifunctionality of central pattern generators, an idea that was first applied to swimming CPGs (Getting and Dekin, 1985; Section 6.3.3). In Lymnaea, rasping movements of the buccal mass are performed not only during feeding but also during egg laying; in the latter case, they prepare the substrate for attachment of the eggs (see Section 8.5). When Jansen et al. (1999) recorded from buccal nerves while snails were rasping in these two different behavioral contexts, they found significant differences in the patterns of neural activity. The bursts of activity contained fewer spikes and had more variable durations during egg laying than during feeding. Also, the interburst intervals were generally longer during egg laying than during feeding. Because the organization of the CPG is incompletely understood in most species, it has been difficult to learn how, exactly, the feeding CPG is reconfigured to produce different patterns. Attention has focused on cellular activity that might be necessary and/or sufficient to switch the buccal output from one pattern to another. In Planorbis corneus, the level of activity in group 1 neurons seems to be important since sustained depolarization of even a single group 1 neuron can cause the feeding rhythm to switch from mode A to mode C, whereas sustained hyperpolarization can trigger a switch from mode A to mode B (Arshavsky et al., 1988). In Aplysia, some identified buccal ganglion neurons are thought to be important for initiating rejection patterns, namely B34 and B4/5, while others, namely B65 and B51, are considered important for initiating ingestion rhythms (see Figs 7.7 and 7.10; Hurwitz et al., 1997; Kabotyanski et al., 2000). Another approach to understanding the multifunctionality of feeding CPGs has been to study how certain neurotransmitters and peptides can elicit specific types of motor programs. Numerous substances influence the CPG output, including acetylcholine, dopamine, egg-laying hormone (ELH), FMRFamide, octopamine, small cardioactive peptide B (SCPB), and serotonin (Sossin et al., 1987; Quinlan and Murphy, 1996; Kabotyanski et al., 2000). All these substances are known to be present in either the cerebral ganglion, the buccal ganglion, or both. When delivered exogenously, they generally increase activity in what is described as the

150 Behavior and Neurons in Gastropods

feeding motor program, but the results are often difficult to interpret because one does not know when the substance is normally released in the intact animal, at precisely which locations, and at what concentrations. Most interesting is the finding that some substances appear to be selectively associated with certain types of CPG activity. For example, dopamine is associated with ingestion rhythms (Kabotyanski et al., 2000), whereas serotonin is associated with swallowing-phase activity (Quinlan and Murphy, 1996; Kabotyanski et al., 2000). Figure 7.9 illustrates the fact that dopamine, 5-HT, and several peptides exert their influences at different levels and in different ways in the feeding control system of Aplysia. The complexity of these mechanisms underscores the difficulty of learning about CPG reconfigurations. Still another approach has sought to discover higher level identified neurons that are able to initiate particular buccal motor programs. For example, Croll et al. (1985c) described three cell types in the buccal ganglion of Pleurobranchaea, each of which is capable of eliciting the egestion motor program, namely the ventral white cell, the anterior ventral cells, and the B3 cells. Other cells in the cerebral ganglion, including the metacerebral giant cell and the paracerebral cells, were shown to command the ingestion motor program. Similarly, in Aplysia, Rosen et al. (1991) described several cerebral neurons that have axons projecting to the buccal ganglion. Certain of them (e.g., CBI-1) trigger egestion, whereas others (e.g., CBI-2) trigger ingestion. The functions of command and modulatory neurons are discussed at greater length below.

Figure 7.9. Multiple sites of modulation in the feeding system of Aplysia. The box at the left (CPG) represents the central pattern generator. The box at the right (ARC) represents the accessory radular closer muscle. Three buccal ganglion cells (circles) are also shown. Synaptic connections are indicated as excitatory (triangles), inhibitory (circles) or polysynaptic (arrowheads). Modulatory actions are shown as either excitatory (plus sign) or inhibitory (minus sign). Abbreviations (modulators): Bucc, buccalin peptides; CP-2, cerebral peptide 2; DA, dopamine; ELH, egg-laying hormone; FMRFa, tetrapeptide; 5-HT, serotonin; MM, myomodulin peptides; SCP, small cardioactive peptides. Updated from Sossin et al. (1987). Copyright 1987 by the Society for Neuroscience.

Feeding 7.5. 7.5.7.

151

Initiation and Modulation of Feeding Behavioral States

Sometimes, when an animal is presented with food, it will eat ravenously. At other times, it will not eat at all. The difference in the two cases can be attributed to differences in motivation. Motivation is an example of an "intervening variable," a postulated internal influence that mediates responses to stimuli. Modern neurobiological studies in gastropods, in particular by Kandel, Kupfermann, Weiss, and colleagues have greatly contributed to the concept of motivation by providing examples of detailed neuronal and hormonal mechanisms, in many cases involving single identifiable neurons. The behavioral state in which animals are motivated to eat is known as "hunger." Numerous internal and external factors influence the degree of hunger, but probably the most significant is the amount of food already present in the animal, or the level of satiation. Thus, most animals stop eating at some point even though attractive foods may remain nearby. Kupfermann (1974) studied satiation in Aplysia, and he reported that it occurs in both the laboratory and the field. Satiated animals become inactive and show no appetitive responses to food stimuli. One signal known to regulate satiation in Aplysia californica is mechanical distension of the anterior portion of the gut (Susswein and Kupfermann, 1975). Blood glucose does not appear to influence satiation in Aplysia californica, even though the hemolymph concentration is regulated and it increases following a meal (Horn et al., 1998). When hungry animals are exposed to food stimuli, a state of arousal is induced. As characterized by Kupfermann and colleagues (1991), foodinduced arousal in Aplysia comprises both appetitive and consummatory phases. The initial contact with a food stimulus usually produces a withdrawal response. Appetitive arousal is manifested when a hungry animal responds to a subsequent presentation of food by raising the anterior part of its body and moving its head to and fro. At the same time, the heart rate and the blood pressure increase. Further contact with the food, especially when localized to the mouth, triggers consummatory behaviors, in particular biting. Consummatory arousal is manifested by a progressive increase in the rate and magnitude of biting (Kupfermann, 1974). An experimental measure of feeding arousal that incorporates both the appetitive and the consummatory aspects is the latency to bite a piece of food once it has been brought to the animal's lips. Kupfermann (1974) found that a single presentation of a small piece of food reduces the latency to bite when a second piece is subsequently presented, compared to controls that receive no prior food. This effect persists for up to 5 minutes after the first presentation. Similar experiments in Lymnaea found an effect persisting for about 15 minutes (Tuersley, 1989). These results imply that the intervening variable, arousal, has a memory component. Stimuli other than food can also influence feeding behavior. In Aplysia, pheromones are released

152 Behavior and Neurons in Gastropods

from the atrial gland during mating. Studies by Susswein and colleagues have demonstrated that the atrial gland pheromones facilitate sexual activities and they also facilitate feeding activities (Blumberg and Susswein, 1998). 7.5.2.

Modulatory Neurons and Command-Like Neurons

The physiological mechanisms of motivation lie in the modulation of nervous function through chemical synaptic transmission and, possibly, circulating hormones. Modulation may involve changes in membrane potential, synaptic transmission, excitability, or muscle excitation-contraction coupling. By definition, neurons having a modulatory function can affect the quality or magnitude of a behavioral output, but they do not, by themselves, initiate behavior. Modulatory neurons are often contrasted with "command" neurons, where the latter are capable of triggering a behavior or causing an abrupt change in the type of behavior expressed. Although a distinction between modulation and command is generally maintained in the sections that follow, most neurons are not easily categorized as either purely modulatory or purely command-like, as the examples will illustrate. Even the definition of a command neuron is somewhat controversial. Some authors maintain that neurons should be so designated only if their activity is both necessary and sufficient to cause some particular behavior. However, in this section the term "command neuron" is used more loosely to refer to any cell whose spiking activity can initiate behavior, or which can elicit "fictive behavior," that is, patterns of efferent neural activity characteristic of that recorded during actual behavior. Thus, in the feeding system, command neurons are usually identified by their ability to trigger the feeding motor program. An additional requisite property of a feeding command neuron is that it fires when food chemicals are applied to the animal's lips, because chemostimulation is the ultimate natural trigger for the feeding motor program. In a study specifically addressing the roles of tactile versus chemical stimulation in Lymnaea, Staras et al. (1999a) found that tactile stimulation alone does not initiate feeding, but sucrose stimulation does. Once feeding begins, however, continued touches to the lips cause the feeding movements to become faster and more regular. This was an interesting finding because repetitive touches similar to those delivered by the researchers occur naturally during the rasping phase of feeding. We saw earlier (see Section 7.2.3) that chemical stimulation and tactile stimulation both contribute to the appetitive, or food finding, phase of feeding. Now, from the results of Staras et al., it is clear that the two modes of stimulation again contribute to the consummatory phase of feeding, albeit in different ways. Accordingly, most of the interneurons mentioned in the discussion below respond to both chemical stimulation and tactile stimulation, and they integrate the two types of signals (Delaney and Gelperin, 1990; Staras et al., 1999a). Several neurons with modulatory functions have been identified in the buccal ganglion. The SO cell of Lymnaea is typical (Elliott and Benjamin,

Feeding

153

1985b; Kemenes et al., 2001). Although SO fires in a bursting pattern during the Nl phase of the feeding rhythm (Fig. 7.6), perturbation of its burst cycle by current injection does not reset the feeding rhythm, so it is not part of the CPG. However, injections of depolarizing current into SO elicit oscillations of the feeding CPG, and this property is frequently exploited in experiments. Although activity in SO is sufficient to elicit buccal rhythms, the patterned bursts that occur after sucrose stimulation of the lips does not require activity in SO. Instead, the cell whose activity appears to be critical for initiating the feeding rhythm is the CPG interneuron, N1M (Kemenes et al., 2001). The normal role of SO seems to be to modulate the frequency of the rhythm as do the OC cells (Vehovszky and Elliott, 2001). Thus, when the contribution of SO is eliminated, by hyperpolarizing the cell during an ongoing rhythm, the cycle period of the rhythm increases signficantly (Kemenes et al., 2001). A curious feature of SO is that it is present only as a single cell. While the two buccal ganglia are symmetrical in nearly all other respects, the SO cell appears randomly in either the left or right member of the pair (Fig. 7.5A). The buccal ganglion of Aplysia contains several neurons with properties similar to SO, namely, they participate in feeding-related neural rhythms and they initiate rhythmic patterns when depolarized, but neither the sufficiency nor the necessity of their activity for the initiation of feeding in the whole animal has been established. Included among these cells are B31/B32, B63, and B65 (Hurwitz and Susswein, 1996; Hurwitz et al., 1997; Kabotyanski et al., 2000). A pair of white cells on the ventral surfaces of the buccal ganglia in Pleurobranchaea are again similar to those described above, but remarkable in some respects. When they fire in bursts they trigger a buccal motor program. The bursts are driven by recurrent spontaneous depolarizations that resemble plateau potentials but have unusually long durations in the range 30-240 seconds. The action potentials broaden during the course of a burst from about 15 milliseconds at the beginning to about 250 milliseconds at the end, and experiments have shown that spike broadening is more important than spike frequency for initiating the motor program. Evidently the long duration spikes facilitate the release of a peptide transmitter, which acts on unidentified neural elements critical for generating the rhythm. The exact nature of the motor program triggered by the ventral white cells is disputed. Early workers said that it triggers a feeding, or ingestion, program, but Croll et al. (1985b) says it triggers an egestion program (see Section 7.4). Unfortunately, no recent studies have addressed the issue. Since the chemical and tactile stimuli that elicit natural feeding are channeled into the CNS through nerves of the cerebral ganglion, but the feeding motor program is generated in the buccal ganglion, considerable interest attaches to cerebral neurons that respond to food stimuli and that have axons projecting to the buccal ganglion (Figs 7.5 and 7.7). Most interesting of all are the neurons that can trigger feeding motor programs. Cerebral "command" cells fitting this description have been reported in several

154 Behavior and Neurons in Gastropods

gastropod genera including Aplysia, Limax, Lymnaea, and Pleurobranchaea (see Croll et al., 1985b,c; Delaney and Gelperin, 1990; McCrohan, 1984; Rosen et al., 1991). The data suggest that some of the cells that are separately described for different species may be homologous. It is at least apparent that all the studied species have cerebral cells with strikingly similar properties. In each species, there are approximately 20 cerebral neurons on each side that have axons in the cerebrobuccal connective nerve, including the metacerebral giant cell (see below) and a number of mechanosensory neurons. All the cells are true interneurons, that is, they have no axon in any cerebral or buccal nerve except the cerebrobuccal connective nerve. Hence, they are known as cerebrobuccal interneurons, or CBIs. Some of these cells have been considered command-like because they are able to trigger buccal rhythms. The neurons characterized as command-like are generally located at the edge of the ganglion near where the chemosensory nerves insert. In many cases, their monosynaptic synaptic connections with buccal neurons have been documented. However, while the cells may be able to activate the feeding rhythm when stimulated intracellularly, and they may respond to food stimuli, their firing does not appear to be necessary for the initiation of food-evoked feeding rhythms (Kemenes et al., 2001). An interesting feature of the CBI cells is that they typically receive synaptic feedback from the buccal pattern generator. This causes their firing to be confined to a particular phase of the feeding cycle. It also reinforces their excitatory function by increasing the rate of firing, thus ensuring that a rhythm will be initiated and sustained. Each population of cerebrobuccal interneurons contains a variety of individual types. Specifically, they differ in their activity patterns (tonic or phasic) and their connections with neurons in the buccal ganglion, some of which are shown in Figure 7.7. Also, while all the CBI cells can initiate the feeding motor program, for some of them the program stops as soon as intracellular stimulation stops, whereas for others the program continues even after the end of stimulation. Each command-like CBI is sufficient to trigger a buccal rhythm, but none has been shown conclusively to be necessary for triggering the rhythm. Individual cells seem capable of modulating feeding episodes in specific ways. Some CBI cells alter phase relations in the feeding program or intensify activity in the CPG, while others differentially recruit specific motor components. Perhaps most interesting are the CBI cells that bias the CPG to express one or another alternative form of the motor programs. Two particular cells, CBI-2 and CBI-3 have been studied in detail to determine how they influence the expression of feeding motor programs. Stimulation of CBI-2 can elicit both ingestive and egestive motor programs. By contrast, CBI-3 does not elicit any motor program when active alone, but when active together with CBI-2 it converts the CBI-2elicited egestive programs to ingestive ones. Jing and Weiss (2001) found that CBI-3 exerts its effects through two CPG elements, B20 and B4/5. When these CPG cells are active, as they are during egestive programs,

Feeding

155

they ensure that the radula closes during protraction, not during retraction. By suppressing both B20 and B4/5, CBI-3 is able to switch the program from egestion to ingestion. Since CBI-3 is only active during the protraction phase, it inhibits B20, also active during protraction, by means of fast IPSPs, whereas it inhibits B4/5, active during retraction, by slowly reducing the excitability of B4/5. The slow inhibition, at least, appears to be mediated by the peptide APGWamide. 7.5.3.

The Modulatory Neuron, MCC

Special among all the cells identified in gastropod molluscs is the metacerebral giant cell (MGC), in part because homologues of the cell have been reported in all species of opisthobranchs and pulmonates so far examined (Pentreath et al., 1982). Numerous names have been given to the cell, some highlighting the fact that it is the largest cerebral neuron (in most species), others emphasizing its placement in the metacerebrum (Fig. 2.6), and still others emphasizing its use of the transmitter serotonin. The MGC is discussed here because its function is to modulate feeding behavior. The MGC receives inputs from all the major sensory nerves of the cerebral ganglion. The inputs are distributed throughout an extensive and richly arborized dendritic tree located within the neuropil of the cerebral ganglion (Chase and Tolloczko, 1992; Chase, 2000). The branching of the MGC axon varies somewhat between species but generally includes one branch that travels to the contralateral cerebral ganglion, another that enters the cerebrobuccal connective nerve and others that enter one or more of the lip nerves. Antic et al. (2000) used optical recording methods to study the initiation and propagation of action potentials in the axonal branches. They found that spikes could be initiated at two separate sites, either in the commissural axon or in the branch leading to the cerebrobuccal connective nerve. Interestingly, spike conduction consistently blocked at certain locations in the axonal network where the diameter of the fibers increases. Thus, because there are at least two spike initiation zones, and conduction into some branches is limited, the possibility exists for the MGC to send signals independently to one target but not to another (e.g., to the ipsilateral buccal ganglion but not to the contralateral ganglion, or vice versa). The MGC axon in the buccal ganglion branches profusely. Some fibers travel into the buccal nerves and go to muscles that control the buccal mass, while most terminate within the buccal ganglion. There is one MGC in each cerebral ganglion and both cells project to both buccal ganglia. Curiously though, the pattern of projection from cerebral ganglion to buccal ganglion is variable. In some species the MGC projects contralaterally within the cerebral ganglion, while in other species the MGC withholds its contralateral projection until its axon reaches the buccal ganglia, and in still other species there are contralateral routes through both the cerebral ganglia and the buccal ganglia (Pentreath et al., 1982). No functional correlate of these different anatomical arrangements has yet been discovered.

156 Behavior and Neurons in Gastropods

A consistent picture of MGC function has emerged from studies in several species using the combined methods of electrical stimulation, in vivo recordings and lesions. The main effects of its activity seem to be to increase the rate and intensity of buccal movements during rasping or biting, but it also contributes to the initiation of feeding. The cell is excited by food stimuli during the appetitive phase of feeding and then its activity wanes as the consummatory phase begins (Morgan et al., 2000). Given that rate and intensity both continue to increase during the course of a feeding episode, despite the fact that activity in MGC declines, Morgan et al. (2000) propose that the modulatory influence of the MGC is slowly replaced by a similar influence imposed by one of the cerebral-buccal interneurons, perhaps CBI-2 in Aplysia. Intracellular stimulation of the MGC is reported to trigger feeding motor programs, at least sometimes and at least in some genera (Helisoma, Lymnaea, Pleurobranchaed). However, as Yeoman et al. (1994b) point out, the rate of firing elicited in these in vitro experiments was probably many times greater than that which the cell actually experiences under natural conditions in the intact animal. Using fine wire electrodes chronically implanted in Lymnaea, Yeoman et al. (1994b) discovered that the rate of firing in dissected preparations (30-120 spikes/minute) is 2-8 times greater than that which is recorded in intact animals. Most importantly, the MGC is silent in quiescent animals, but fires tonically when food is presented to the lips and the animal begins to rasp. The cell also fires when rasping occurs during locomotion, even in the absence of food. When the effects of MGC stimulation were tested in isolated ganglia using rates of firing equivalent to those encountered in vivo, fictive feeding was never elicited by MGC stimulation alone. However, MGC activity was critical for the ability of the SO neuron to drive the feeding rhythm. If the MGC was kept silent, no rhythm could be elicited by stimulating SO. When the MGC was stimulated to fire within the range encountered in the intact animal (7-20 spikes/minute), the frequency of the feeding rhythm was correlated with the frequency of MGC firing. Lesion experiments likewise indicate a role for the MGC in modulating the rate of feeding. When the two MGCs (called MCCs in these Aplysia experiments) were killed, by injecting them with proteases, the lesioned animals were subsequently found to bite at a rate that was 40% slower than control animals (Rosen et al., 1989). Similar effects were seen in Lymnaea when the MGCs were functionally inactivated by either photoablation, serotonin depletion, or serotonin antagonism (Yeoman et al., 1994a,b). The physiological actions of the MGC are mediated by serotonin, which is present in the soma at a concentration of approximately 4 x l O ~ 4 M (Pentreath et al., 1982). In an important early investigation, Weiss et al. (1978) found that serotonin from the MGC modulates the accessory radula closer muscle of Aplysia in several different ways. First, it synaptically depolarizes motoneurons. Second, it potentiates the amplitude of excitatory junction potentials. Third, it increases the size of muscle contractions by

Feeding

157

mechanisms intrinsic to the muscle itself (i.e., independently of EJP amplitude). Later work showed that the mechanism of muscle contraction facilitation depends on cyclic AMP, and it likely involves an increase in voltage-dependent calcium currents (Brezina et al., 1994b; Section 7.5.6). Very similar effects of MGC activity have been seen in another muscle of Aplysia, the anterior intrinsic buccal muscle, as well as in the radular retractor muscle of the terrestrial snail, Achatina fulica. Together, these effects may account for the increase in bite size seen during the course of a meal, or at least during the beginning of the meal when the MGC is most active. The MGC-dependent increase in bite frequency can be accounted for by the ability of the cell to increase the frequency of the feeding motor program and shorten the duration of its protraction phase (Weiss et al., 1978; Morgan et al., 2000). Specific synaptic influences of the MGC on buccal CPG neurons were examined in Lymnaea by Yeoman et al. (1996). They found that the MGC excites many, but not all, members of the CPG. Interestingly, the evoked excitatory postsynaptic potentials (EPSPs) are brief in some cells (~200 milliseconds), but very long (MO seconds) in others. MGC activity also affects the intrinsic properties of some CPG cells, for example, it shortens the plateau phase of N2v bursts. Together, these modulatory effects of MGC activity increase the rate of CPG oscillation by shortening the Nl and N2 phases, while leaving the N2 (swallow) phase largely unaffected. 7.5.4.

C-PR, A Command Neuron for feeding Arousal?

A persistent idea in gastropod neurobiology is that single neurons may command significant behaviors through anatomically distributed synaptic connections. Kupfermann, Teyke, Weiss, and colleagues claim that the C-PR neuron of Aplysia is one such cell. Its functions are said to include not only the initiation of feeding behaviors but also the establishment of a motivated state of feeding arousal. C-PR is present as a bilateral pair in the M cluster of the cerebral ganglion (Fig. 6.7B). The acronym stands for "cerebro-pedal regulator," referring to the observation that its synaptic outputs appear confined to the pedal ganglion (Teyke et al., 1990b). When C-PR fires, an estimated 2000 pedal neurons are affected through direct or indirect synaptic pathways. Some of the neurons affected by C-PR excite neurons with an identified role in elevating the head. In this respect, C-PR might be seen as a postural command cell. Adding further interest to the cell, however, is the fact that it also evokes synaptic responses in cells associated with several other functional systems, namely, the cardiovascular system (cells L10, RBHE, LBVcX the defensive system (cells Bn, R2, Pll, L7), and the feeding system (MGC and CBIs). The synaptic actions are such that C-PR should, in the intact animal, raise the head, increase the heart rate and blood pressure, inhibit defensive responses, and facilitate buccal feeding motor programs. In sum, C-PR seems to be able to induce a state of arousal for feeding, as described in Section 7.5.1. It is also interesting that the effects

158 Behavior and Neurons in Gastropods

of C-PR on the cerebrobuccal command neurons, the CBIs, are specific to each cell (Hurwitz et al., 1999). C-PR excites CBI-1, CBI-12, and CBI-8/9, but it inhibits CBI-3 and CBI-5/6. As already noted, individual CBI cells have different effects on the buccal pattern generator. Therefore, from the connections made by C-PR, Hurwitz et al. (1999) suggest that the behavioral state elicited by C-PR may specifically facilitate head-up feeding, while being irrelevant to, or even suppressing other forms of feeding, for example, grazing with the head down. 7.5.5.

Sensory Influences on the Feeding Motor Program

Once feeding has begun, the behavior is subject to an ongoing modulation, which comes in part from consummatory arousal but which may also involve selective adjustments in the intensity or timing of particular motor phases. Both exteroceptors and proprioceptors contribute to modulation, and several participating afferent neurons have been identified. However, investigators have not found it easy to identify the precise roles for these sensory influences. Food in and around the mouth contributes to arousal and helps to stabilize the feeding motor program (Staras et al., 1999a). Among the cells excited by perioral stimulation are the paired C2 cells in the cerebral ganglia of Aplysia. They are unusual because they use the transmitter histamine as well as nitric oxide. Although excited by food, they continue to be active once the food is removed from the lips, possibly because they can sense movements of the buccal mass (proprioception). Their numerous postsynaptic targets include the MGC and motoneurons that innervate extrinsic muscles of the buccal mass, suggesting that C2 may contribute to arousal as well as fine tune the retraction phase of feeding (Chiel et al., 1986). Certain identified neurons in the buccal ganglia of Aplysia have unique properties that closely correspond to those of neurons found in the terrestrial slug Incilariafruhstorferi (Borovikov et al., 2000). They are mechanoreceptors that innervate a thin layer of muscle underneath the radula. They are presumably activated when food contacts the radula, thus qualifying as exteroceptors, but they may also be activated by movements of the radula, in which case they would also be proprioceptors. The cells are especially sensitive to distortion of the posterior part of the radula, which is the region in which food is grasped. From studies in reduced preparations, it appears that the cells are mainly excited during the retraction phase, and their probable function is to aid in grasping the food as it is swallowed. In Aplysia, numerous identified inputs and outputs have been identified for two particular mechanoafferents, named B21 and B22 (Borovikov et al., 2000; Rosen et al., 2000). The output targets, some of which are shown in Figure 7.10, include motoneurons, interneurons, and other sensory neurons. These connections are consistent with the idea that the cells can enhance the radula closing phase of feeding. In addition to receiving afferent excitation, they receive significant inputs from central neurons including, indirectly, the cerebrobuccal command cell, CBI-2. Most of their central inputs occur in

Feeding

159

Figure 7.10. Synaptic connections of the radular mechanoafferent neuron, B21, in Aplysia. When the radula is touched, as by food, action potentials propagate centrally. Output connections of B21 include five motoneurons innervating three different muscles. The muscles 14 and ARC attach to the radula, whereas 12 does not. B21 is electrically connected to interneurons B64 and B19, and it is synaptically inhibited by the multifunctional neuron B4. The neuron may act to modulate the force of radula closure during biting movements. Some of the cells shown in this figure also appear in Figure 7.7. From Rosen et al. (2000) with permission.

phase with the feeding cycle and they seem to gate the cells' outputs so that they do not occur during inappropriate times in the cycle. A further interesting feature has been reported for one of the cells, B21, whose action potentials have a tendency to block when propagating in a region of the neurite near the soma. Because of the location of this region of low safety relative to regions where B21 makes synaptic contacts with its follower cells, inhibitory synaptic inputs to the cell from central sources can selectively gate the synaptic outputs of B21 to different follower cells during different phases of the feeding cycle. Again, this may be part of a mechanism to ensure efficient swallowing. Another type of mechanoreceptor, in this case, one whose neurites terminate in the esophageal wall, has been described in Lymnaea (Elliott and Benjamin, 1989). These cells are excited by distension of the esophagus,

160

Behavior and Neurons in Gastropods

and their activity inhibits all the identified pattern generating interneurons in the groups Nl, N2, and N3. If a single neuron is fired by intracellular current injection, it can cause an ongoing feeding rhythm to stop. It is therefore suggested that these cells might be responsible for terminating a meal. The observations are consistent with the fact that satiation in Aplysia is triggered when ingested bulk causes expansion of the alimentary tract (Susswein and Kupfermann, 1975). However, at odds with this interpretation is the fact that these same cells also excite motoneurons that are active during the rasp phase. 7.5.6. Peripheral Modulation Modulation of the motoneurons' control over muscle contraction is a final, powerful means of regulating the feeding system. Current discussions distinguish two types of modulation, which are conveniently invoked here. Intrinsic modulation comes from neurons that are themselves part of the circuit that controls the behavior (i.e., either the CPG or motoneurons). By contrast, extrinsic modulation comes from neurons that lie outside the controlling circuit. Serotonin is a transmitter that has potent modulatory effects but it is not present in any of the intrinsic neurons. It must therefore be released from extrinsic neurons, one of which is known to be the MGC. The MGC releases serotonin into the buccal muscles where it causes an increase in the size of contractions evoked by motoneuron activity; it also increases the rate of relaxation in the same muscles (Weiss et al., 1978). The intrinsic modulators are the primarily motoneurons of the feeding circuit, which typically release multiple co-transmitters. One transmitter, usually acetylcholine or glutamate, is fast acting and responsible for depolarizing the muscle fiber. In addition, the motoneurons release one or more peptide transmitters that do not, by themselves, cause any muscle contractions, but modulate the contractions caused by the fast transmitter. Intrinsic peptide modulators have been studied most extensively in Aplysia, although they have also been reported in the buccal ganglia of other gastropods, including Achatina and Lymnaea. The prevalence of peptide co-transmitters in the feeding circuit is indicated by one survey of 17 motoneurons in Aplysia (Church and Lloyd, 1991). The peptides included in the survey included myomodulin, small cardioactive peptide (SCP), buccalin and FMRFamide. Every cell was found to express at least one peptide; some cells contained two peptides and others had as many as three peptides. The particular combination of peptides is different in different cells. Biochemical extractions of peptides from buccal muscles yield similar results, namely that every muscle contains several peptides. As with the motoneurons, not all muscles contain the same complement of peptides. For example, buccalin is abundant in the accessory radula closer muscle, but it is absent from the accessory radula opener muscle. Adding to the complexity is the fact that the genes coding for modulatory peptides generate large families of related molecules. The myomodulin gene, for example, codes

Feeding

161

for nine peptides; the buccalin gene, 16 peptides. Many of these peptides, if not all, have modulatory functions, and their physiological actions are not necessarily identical. Important insights into peripheral modulation have come from studies of the accessory radula closer muscle of Aplysia, particularly with respect to how two of its motoneurons, B15 and B16, control the muscle (Fig. 7.9; Cohen et al., 1978). Whereas both motoneurons use acetylcholine to generate muscle contractions, B15 uses the SCPs and the buccalins as co-transmitters, while B16 uses the myomodulins and buccalin as co-transmitters. For both cells, the two peptide co-transmitters are co-localized within the same individual vesicles and they are co-released in fixed ratios during physiological rates of motoneuron firing (Vilim et al., 2000). The effects of peptides on muscle contractions can be studied by applying the peptides directly to a muscle while firing a motoneuron using intracellular stimulation. When examined individually, the peptides either increase or decrease the size of evoked contractions, and some increase the rate of relaxation. To investigate the cellular mechanisms of peripheral modulation, Brezina and colleagues (1994a,b) studied the modulation of contractions in single dissociated muscle fibers. Figure 7.11 shows that acetylcholine contracts the fibers when it is delivered alone, but if 5-HT or SCP is delivered together with acetylcholine, the contractions are potentiated. Myomodulin usually potentiates at low concentrations but depresses at high concentrations. It was found that potentiation is caused by the enhancement of an L-type Ca++ current, whereas depression is caused by activating a K + current. The peptidergic regulation of the rate of relaxation is apparently mediated by a cAMP-dependent phosphorylation of contractile proteins. These results demonstrate that, for the most part, the peptides act directly on the muscle (see also Fig. 4.2). Only buccalin acts on the motoneuron terminals (Fig. 7.9), where it depresses the release of ACh, thus reducing the size of contractions. An intriguing differential action of FMRFamide was discovered in the medial intrinsic muscle of Aplysia, named I3m (Keating and Lloyd, 1999). Here, where muscle contractions are evoked by two motoneurons, B3 and B9, FMRFamide increases EJPs and contractions evoked by B3, but it decreases EJPs and contractions evoked by B9. Interestingly, only B3 expresses FMRFamide; in addition, while B3 uses glutamate as a fast transmitter, B9 uses acetylcholine. Whether these differences in transmitter types explain the differential modulatory actions of FMRFamide, or indeed whether FMRFamide acts presynaptically or postsynaptically at these synapses, has not been determined. The obvious complexity of peripheral modulation raises questions about how the multiple effects of the various peptides interact and what are their underlying functions. From theoretical and experimental investigations (Brezina et al., 1996, 2000), it emerges that modulation of the rate of muscle relaxation is extremely important. Unless a muscle (e.g., the radula closer) is sufficiently relaxed, contraction of its antagonist (e.g., the radula opener) will

162 Behavior and Neurons in Gastropods

Figure 7.11. Postsynaptic modulation of contraction demonstrated in a single muscle fiber from the accessory radula closer muscle of Aplysia. The fiber is held by suction

while perfused with a fast-flowing stream of sea water. The muscle contracts when a puff of acetylcholine (ACh) is added to the sea water. When serotonin (5-HT) is added, the contraction is potentiated. For these photographs, fast green dye was added to the sea water to enhance the appearance of the stream after addition

of acetylcholine. From Brezina et al. (1994b). Copyright 1994 by the Society for Neuroscience.

create little movement of the structure, in this case the radula. So, the rate of relaxation limits the rate of rhythmical movement and hence limits the functional capacity of the feeding mechanism. Intrinsic modulation works well to increase the rate of relaxation because, the faster the muscles are

asked to move by higher rates of neuronal firing, the greater will be the release of modulating co-transmitters (Vilim et al., 2000). It is interesting that the contraction-potentiating effects of SCP and myomodulin are often countered

by the contraction-depressing effects of buccalin, with the consequence that there may be little net change in the size of contraction in a particular muscle, yet the rate of relaxation may be significantly affected. Even when buccalin is effective in depressing the size of contractions, and therefore causes smaller

absolute movements of the mouth, it still facilitates the functional output of the radular muscle system because it allows the mouth to open and close more quickly (Brezina et al., 2000). In other words, the net effect is still a faster rate of feeding.

It is also clear that particular combinations of peptides can provide graded modulation of muscle contractions in respect to both size and rate. Since peptide release is sensitive to the rate of neural firing, when different peptides

Feeding

163

are released from different motoneurons innervating the same muscle, as in the case of B15 and B16, different ratios of concentration can be achieved by varying the rates of firing in the motoneurons. In this example, both neurons release peptides that modulate muscle contraction size and muscle relaxation rate, but the net effects produced by each cell are different. The advantage of mixing the peptides over a range of concentration ratios is that combinations of contraction size and relaxation rate can be achieved that would not otherwise be possible (Brezina et al., 1996). In this way, presumably, the nervous system can adjust details of the muscle's movements as necessary, perhaps depending on the physical properties of the food being consumed.

7.6.

Plasticity of Feeding Behavior

The majority of gastropods are generalist feeders, meaning that they are not tied to any particular food, but can accommodate to seasonal variations and patchy distributions. Because an individual animal may encounter many potential foods, and because no two foods are likely to have equal nutritional value, or to be equally digestible, it is obviously advantageous for an animal to have an informed basis for choosing from among the available foods. Probably many such mechanisms are used, some operating during food finding, as discussed above in Section 7.2.2, and others operating during the consummatory phase, to be discussed now. The latter are more interesting to neurobiologists because our detailed knowledge of buccal motor program generation provides a good opportunity for investigations at the cellular level. By contrast, the olfactory pathways and the locomotor circuits are much less accessible. An optimum diet will see an animal avoiding foods that provide poor nutrition while being attracted to, and eating, foods that provide good nutrition. Although numerous studies have confirmed that learning by experience is one mechanism by which gastropods are able to select their foods, it is important to note that other mechanisms may also be at play. A comparison of two studies on terrestrial slugs illustrates how difficult it can be to interpret simple behavioral observations. Both studies examined how slugs responded to artificial diets that had been made nutritionally deficient. The animals' standard, nutritionally complete diet was altered by substituting some of its constituents while making only minimal changes to its physical and chemical properties. Delaney and Gelperin (1986), working with Limax maximus, selectively removed single amino acids, whereas Cook et al. (2000), working with Deroceras reticulatum, used foods that contained either too little protein or too little carbohydrate. In both studies, the animals that ate the nutritionally deficient foods quickly responded by avoiding the deficient food while choosing instead another food that contained the previously missing ingredient. In both studies, the change in the slug's choice of food could be prevented by injecting the missing nutrients directly into the animal's hemolymph.

164 Behavior and Neurons in Gastropods

Interestingly, while the results from the two studies summarized above are very similar, the authors give different interpretations. In the view of Delaney and Gelperin (1986), the slugs learn an aversive association between the taste of the nutritionally deficient food and the postingestive consequences of consuming it. On the other hand, Cook et al. (2000) question whether learning is important in this context, mostly because they conducted experiments in which certain "flavorants" were added to the food to see if the slugs would associate the flavor with the nutritional content. Because no strong associations were detected, the authors concluded that the slugs chose their foods on the basis of the nutritional content alone irrespective of the flavor. Thus, Cook et al. (2000) believe that nutritional imbalance influences food choice directly, perhaps by altering sensory responses to specific components of the food. For example, the sensitivity to carbohydrates might be heightened in animals that need to ingest more carbohydrates. Such a mechanism could account for the phenomenon of "neophilia," the tendency to eat novel foods, which is commonly observed in snails and slugs. Since animals that consistently feed on a single food may develop nutritional imbalances, any mechanism that increases the sensitivity to missing ingredients will make it more likely that the animal will ingest foods that contain the missing ingredients. Whether heightened sensitivity for a component of food involves learning remains to be seen, however, because the experiments of Cook et al. (2000) do not rule out the possibility that the slugs may have formed associations to chemical stimuli other than those intentionally inserted. Alternatively, it is also possible that learned associations might have been demonstrated had different artificial flavors been used. Pavlovian or classical conditioning methods have been used by several investigators to establish conditioned aversions to foods. The conditioned stimulus (CS) is either a food normally eaten by the animal (potato for Limax', carrot root for Helix', squid for Pleurobranchaed) or a phagostimulant such as sucrose (Lymnaed). Training consists of pairing the CS with an unconditioned stimulus (UCS), typically either electrical shock or a noxious chemical. After training, the animals respond to the CS by either eating less or, in some cases, actually withdrawing from the food. Several studies of this kind were conducted in Pleurobranchaea, where care was taken to include appropriate controls for non-associative processes, and where one-trial learning and differential conditioning were demonstrated (reviewed in Mpitsos and Lukowiak, 1985). Subsequently, a neural correlate of food-avoidance learning in Pleurobranchaea was discovered. Whereas food stimuli ordinarily excite an identified group of paracerebral neurons that have a role in commanding the feeding motor program, in conditioned animals the same stimuli inhibit the paracerebral cells. It was suggested that aversive conditioning enhances the excitability of another group of interneurons, known as 12, which then suppress feeding by indirectly inhibiting the paracerebral command neurons and by excessively driving the retraction phase of the CPG (see Fig. 10.8, where the command neurons are labeled as PCP and PSE; Jing and Gillette, 2000).

Feeding

165

A neural correlate of aversive conditioning has also been described in Lymnaea. Kojima et al. (1997) found that pairing sucrose (CS) with KC1 (UCS) caused a reduction in the number of bites elicited by the sucrose stimulus. When the conditioned animals were evaluated electrophysiologically and compared to control animals, the most significant difference was that the MGC inhibited the NIM cells more in conditioned animals than in control animals. Since early firing of the NIM cells is important for starting up the feeding motor program (Fig. 7.6; Section 7.3), enhancement of this inhibitory pathway might explain, at least in part, the conditioned suppression of feeding. However, Kojima et al. found no evidence to support the idea that the MGC itself is the point of convergence of the CS and UCS, since neither the intrinsic electrical properties of the MGC nor its responses to the CS were changed by conditioning. From a direct comparison of appetitive food conditioning and aversive food conditioning in Lymnaea, Whelan and McCrohan (1996) concluded that appetitive protocols are much more effective than aversive protocols. Recent studies of appetitive conditioning in gastropods have reinforced this view while providing novel insights into the cellular mechanisms. An appetitive training method using a chemical CS was first developed by Teresa Audesirk and colleagues for use with Lymnaea (Audesirk et al., 1982). The CS, amyl acetate, is first added to the snail's ambient water. Fifteen seconds later the UCS, sucrose, is added. Learning is evaluated by comparing the number of rasps elicited by the CS before and after training, where training consists of pairing the CS and the UCS as described. Using this paradigm, Whelan and McCrohan (1996) found a correlation between the extent to which an animal demonstrated behavioral learning and the ability of amyl acetate to evoke fictive feeding in the same animal after reduction to an isolated CNS-lip preparation. Also, certain cerebrobuccal interneurons, named CV1 (Fig. 7.5B), were excited by amyl acetate in 13 of 17 preparations from conditioned snails, but in none of the preparations from control snails. In a further analysis of conditioned feeding in Lymnaea, a tactile stimulus was substituted for the previous CS, amyl acetate, in order to better control the stimulus and to more easily track its neural representation (Staras et al., 1999b). Again, neurons were studied in semi-intact preparations derived from previously conditioned snails; these records were compared with those obtained from non-conditioned control animals. Neurophysiological investigation revealed that cellular changes occurred at several specific sites in conditioned animals. First, by recording early in the response pathway, in the cerebrobuccal connective nerve, Staras et al. (1999b) found that the touch-evoked, short-latency discharges were larger in conditioned animals than in control animals. This result presumably reflects an enhancement of the sensory response in first- or second-order neurons. Second, they found that the CPG-driven fictive feeding responses increased in conditioned animals, and lastly, they saw facilitation of a CS-evoked EPSP that was recorded in the motoneuron B3. To test whether this latter effect was due

166 Behavior and Neurons in Gastropods

specifically to conditioning, or whether it was secondary to facilitation of the central pattern generator, Staras et al. (1999b) fed conditioned animals to satiation before dissecting them. In these sated animals there was little fictive feeding (because the animals were not hungry) but facilitation of the EPSP in B3 was still present, indicating that the EPSP effect was independent of the CPG effect. In summary, appetitive conditioning in the feeding network of Lymnaea appears to be mediated by cellular changes at multiple sites (see Benjamin et al., 2000). In previous paragraphs, two procedures for appetitive conditioning in Lymnaea have been described, one using a chemical CS and the other a touch CS. Recent experiments demonstrate that visual stimuli can also be associated with food (Andrew and Savage, 2000). In one experiment, it was shown that snails can be conditioned in a single trial to feed (rasp) in the presence of a black panel. In a second experiment, snails distinguished a checkered black and white pattern from a 50% gray panel of equal luminance, and they learned to associate food exclusively with the former stimulus. Given that the same response, namely feeding, can be conditioned to chemical, tactile, and visual stimuli, and that the neural circuitry is relatively well understood in Lymnaea, it will be interesting to see whether learning involves the same mechanism(s) in all three modalities and, if so, whether there is a common point of anatomical convergence. An appetitive conditioning paradigm similar to that used in Lymnaea is also effective in Aplysia. The CS is a touch to the lips and the UCS is a piece of food (seaweed). Lechner et al. (2000) studied the neurophysiological correlates of conditioned learning in vitro after the animals had been trained in vivo. They found that more cycles of the feeding motor program were elicited by electrical stimulation of a lip nerve in animals that had received paired training of the CS and UCS than in animals that had received the same stimuli in an unpaired protocol. Also, stimulation of the lip nerve evoked larger polysynaptic EPSPs in B31/B32 cells when examined in preparations from animals receiving paired training compared to preparations from control animals. As previously mentioned (see Section 7.3), the B31/B32 cells are thought to play a key role in initiating feeding motor programs (Fig. 7.7). The increase in synaptic excitation was much less in the neuron B4/5, which, like B31/B32, is also a member of the CPG but which fires during the retraction phase. It is especially interesting that conditioning evidently produced no changes in the intrinsic properties of any of these cells, thereby implying that conditioning results in a pairing-specific strengthening of one or more synapses in the polysynaptic pathway from lip sensory neurons to buccal CPG neurons. Thus, studies of appetitive learning in Lymnaea and Aplysia both suggest that modulation in the afferent pathway encoding the CS is necessary for behavioral modification, but whether changes at these sites are sufficient remains to be determined. It will also be interesting to see exactly how similar are the mechanisms of appetitive classical conditioning in Aplysia and Lymnaea.

Feeding

167

In addition to classical conditioning, the feeding system of Aplysia has also provided a rare opportunity to investigate the cellular mechanisms of operant conditioning. The initial behavioral observations were reported by Susswein et al. (1986), working with A. fasciata and A. oculifera in Israel. They found that an animal's feeding behavior depended on its previous success or lack of success in obtaining food in a given situation. Some animals were negatively reinforced by giving them access to food only through a plastic meshing that allowed them to taste the food (green algae or lily), and even bring it into their buccal cavities, but not swallow it. Faced with the netted food, the animals made progressively fewer and less effective attempts to ingest it. After a few minutes of trying to deal with the netted food, these animals stopped responding to it altogether. By contrast, animals that were allowed to swallow their food (i.e., were given positive reinforcement) exhibited progressively more frequent and more vigorous feeding responses. These experience-dependent modifications were specific to the foods encountered during reinforcement, and the memories persisted for at least 48 hours. Neurophysiological investigations of operant conditioning were made possible by the discovery that negative reinforcement is mediated by the esophageal nerve, which innervates the gut (Schwarz and Susswein, 1986). Bilateral sections of the nerve prevented the response decrement attributed to learning that a food is inedible, but the nerve sections did not block another type of decrement due to sustained lip stimulation. Nargeot et al. (1997) took advantage of this finding to develop an in vitro analog of operant conditioning. Their preparation consists of the isolated buccal ganglia and associated nerves, as illustrated in Figure 7.12A. Tonic electrical stimulation of the peripheral nerve n.2,3 elicits rhythmic motor activity that can be recorded in other nerves, and two particular patterns, named I and II, were studied in detail (Fig. 7.12B). To study the effects of reinforcement, the esophageal nerve E n.2 was briefly stimulated immediately after the appearance of either pattern I or II. Whereas pattern I and II were expressed in a randomly mixed temporal sequence, the frequencies of their occurrence were altered by reinforcement, as described below. In control experiments, the non-specific effects of stimulation were identified either by omitting stimulation altogether or by stimulating with inappropriate timing ("yoked" control). Tests showed that the selective reinforcement of either one of the two feeding motor patterns significantly increased the frequency of appearance of that pattern but did not change the appearances of the non-reinforced pattern. Curiously, stimulation of E n.2 was effective as a positive reinforcer in these experiments, even though the same nerve mediated negative reinforcement in the behavioral experiments of Susswein et al. (1986) and Schwarz and Susswein (1986). A cellular analysis suggests that the CPG neuron B51 plays a key role in operant conditioning. The cell has a tendency to fire in bursts and in phase with the retraction phase of the feeding motor program (Fig. 7.7). Nargeot et al. (1999) observed that B51 fires much more commonly in association with pattern I than with pattern II, but it fires probabilistically only during some

Figure 7.12. Design of experiments to study operant conditioning of feeding in vitro. (A) Schematic drawing of the isolated buccal ganglia ofAplysia. Sites of recordings are shown by open triangles, sites of stimulation by filled triangles. Tonic stimulation (4 Hz) of n.2,3 induces rhythmic activity in the recorded nerves. Contingent reinforcement was delivered by stimulation of E n.2. Adapted from Nargeot et al. (1997). (B) Two patterns of activity in nerves and cells. The timing of the retraction phase, identified by activity in a nerve not shown here, is indicated by the dashed vertical lines. In pattern II, motor activity in radula closure neurons occurs during the protraction phase, whereas in pattern I most of the closure motor activity occurs during the prolonged retraction phase. Bursting activity in B51 is primarily associated with pattern I, and contingent reinforcement of pattern I modifies several intrinsic properties of B51 that may be causally related to the increased expression of pattern I. From Nargeot et al. (1999). Copyright 1999 by the Society for Neuroscience.

Feeding

169

cycles. Contingent reinforcement of pattern I, by electrical stimulation of the esophageal nerve, caused several changes in B51, namely an increase in its input resistance, an increase in its excitability, and an increase in the frequency of its bursts. To see if these changes in the intrinsic properties of B51 actually contribute to the selection of pattern I after contingent reinforcement, Nargeot et al. (1999) manipulated the excitation of B51 by intracellular current injection. Depolarization caused B51 to fire more often in phase with pattern I, and it also increased the occurrences of pattern I. Conversely, hyperpolarization silenced B51 and caused a reduction in the frequency of pattern I. Thus, the ability of operant conditioning to bias the system towards expression of pattern I may depend on changing the firing behavior of B51. In summary, the results obtained to date in the feeding system of Aplysia suggest a fundamental difference in the mechanisms underlying appetitive classical conditioning and operant conditioning. While classical conditioning seems to rely on changes in transmission of the afferent signal (the CS), operant conditioning seems to involve changes in the intrinsic properties of neurons in the CPG. On the other hand, comparable studies of appetitive conditioning conducted in Lymnaea suggest that critical changes occur at multiple sites including sensory pathways, cerebrobuccal interneurons, the CPG, and motoneurons. The latter picture is more in line with evidence from studies of non-associative forms of learning in gastropods, which indicate that learning occurs at multiple sites that are distributed in parallel and in series within the stimulus-response pathway (see Frost et al., 1988; Prescott and Chase, 1999). The unique opportunities provided by the gastropod feeding system for studying different forms of learning at both behavioral and cellular levels should yield many more insights in the future.

8

Reproduction The variation in gastropod reproductive behaviors is even greater than the variation in feeding behaviors. The internal reproductive structures are often amazingly complex, and in many cases they are sufficiently different among closely related species to constitute the main criteria for taxonomical distinctions. Some of the mating behaviors of gastropods are said to be "unusual" or even "bizarre," but such descriptors, of course, reflect an anthropocentric point of view. When reading the paragraphs below, it is important to bear in mind that sexual traits are subject to a special evolutionary influence not encountered by most other traits. Charles Darwin himself distinguished natural selection, in which environmental factors drive evolution, from sexual selection, in which the competition for mates is responsible for evolution. Darwin (1871, pp. 324-327) dismissed the possibility of sexual selection in gastropod molluscs, in part because of their hermaphroditism, but in this he was probably wrong, as we will see.

8.1.

Modes of Sexuality

There are two fundamental types of sexuality found in organisms: gonochorism, and hermaphroditism. The gonochorists produce either eggs or sperm, and have just a single functional reproductive system, either female or male. The hermaphrodites possess, at least during some part of their lives, both a functional male system and a functional female system. Hermaphroditism is extremely common among flowering plants and it is present in 20 of 28 animal phyla, including seven phyla that are exclusively hermaphroditic. Within the Gastropoda, hermaphroditism occurs in 3% of "prosobranchs," 99% of opisthobranchs, and 100% of pulmonates (Baur, 1998). 8.1.1.

Sexuality in "Prosobranchs''

The great variety of reproduction in "prosobranchs" is consistent with the large variations in their body forms, habitats, and lifestyles (Fretter, 1984). Cross-fertilization occurs by two methods. In the archaeogastropodsvetigastropods fertilization is external, by spawning, whereas in the Caenogastropods fertilization is internal, by copulation. Sexual dimorphism 170

Reproduction

171

is sometimes evident in structures other than those found in the reproductive system, for example, in the shell, the radula, and certain soft tissues. Males, for example, are smaller than females in some species; females may have more radular teeth, and larger teeth. In archaeogastropods vetigastropods the tubules of the male gonad join a duct of the right kidney and the sperm are discharged into the sea water through an opening in the duct. The female system is similar, with eggs discharged singly, each in a gelatinous covering. The synchronization of spawning and the fusion of the gametes is controlled by pheromones that are produced by the gametes themselves. In caenogastropods, where fertilization is internal, both sexes have genital ducts that open to the exterior at the right anterior of the animal, usually near the base of the tentacle and often within the mantle cavity. In some species the males possess no penis (aphallic). In these cases, the sperm are delivered into the water current as it enters the mantle cavity of the female, and they are then directed by cilia to the oviduct. In all other species, copulation occurs when the male approaches the female from behind and mounts her shell. After mounting, the male positions himself on the female's right side, where he intromits into a vaginal sac or pore. The sperm are either packaged in a spermatophore or delivered in a seminal fluid, depending on the species. Copulation durations vary greatly, in part depending on whether a spermatophore is used, because time is required to fabricate and transfer the spermatophore. Even different species in the genus Littorina, none of which uses a spermatophore, copulate for widely different durations, for example, less than 10 minutes in Littorina pintado but up to 4.5 hours in Littorina obtusata. Moreover, males of the latter species sometimes remain on the female's shell without genital contact for up 2.5 hours after copulation, perhaps to guard against her remating. Multiple matings between the same pair of animals is common, sometimes extending over several days. Hermaphroditism in "prosobranchs" occurs as two types, sequential (sex reversal) and simultaneous. Sequential hermaphroditism is typically protandric, meaning that the animal is first male, then female. Sex reversals occur in many "prosobranch" families, usually once at the end of the first breeding season, but annual sex reversals have also been reported. Simultaneous hermaphroditism, which is less common, presents the possibility for self-fertilization, but whether this occurs in any "prosobranch" species is uncertain. A few "prosobranchs" reproduce by parthenogenesis; most of these occupy freshwater habitats. Much remains to be learned about the reproductive behavior of "prosobranchs," and hardly anything is known about its nervous control, so there is a need for further research. 8.7.2.

Simultaneous Hermaphroditism

The fact that hermaphroditism occurs in so many diverse taxa, both animal and plant, has given rise to speculation about what conditions predispose

172 Behavior and Neurons in Gastropods

a species to this mode of reproduction. Hermaphroditism is favored whenever the overall reproductive success (fitness) of the hermaphrodite exceeds that of either a pure female or a pure male. The most likely situation in which this might occur is where the reproductive possibilities for one of the pure sexes is limited. Males could easily be limited because their success typically depends on how many females they can inseminate. When population densities are low or the organisms are immobile, the chances of encountering a mate are reduced and there is a ceiling on male fitness. Charnov (1979) modeled the evolution of hermaphroditism by showing that when nothing is to be gained by producing more sperm, fitness can be increased by shifting some resources into the female function. The transition to hermaphroditism will be least costly when the same structure can serve both male and female functions as, for example, in the gastropod's ovotestis. It is difficult to imagine hermaphroditism as an adaptation to any particular environment, for example, one which forces population densities to be very low, since there seems to be no environment common to all gastropod hermaphrodites but unknown to gastropod gonochorists, given that opisthobranchs (hermaphrodites) and "prosobranchs" (gonochorists) largely share the same ocean habitats. On the other hand, given that hermaphroditism occurs only in the Euthyneury (Opisthobranchia and Pulmonata), and that this clade appears to be monophyletic (see Section 1.3), it is conceivable that hermaphroditism appeared in some ancestral species as an adaptation, but then persisted in all descendent species as an evolutionary constraint, that is, few euthyneuran species have been able to eliminate the trait (see Leonard, 1990). While nearly all opisthobranchs are simultaneous hermaphrodites, they nevertheless exhibit diverse forms of mating (Hadfield and Switzer-Dunlap, 1984). Reciprocal copulation is especially common in animals where the penis and the female genital aperture are located close together, as in nudibranchs, pleurobranchomorphs, and sacoglossans. In this arrangement, the animals position themselves face-to-face with their right sides opposed. Each animal inserts its penis into the other's vaginal duct, and in return it receives the other's penis. Unilateral copulation is more common in animals where the male and female genital pores are distant from one another, as in Aplysia. Simultaneous reciprocal matings are occasionally observed in Aplysia, but mostly in laboratory aquaria; in the field, reciprocal matings are rare (Pennings, 1991). In unilateral matings, one animal acts as a male, the other acts as a female. After one mating, reciprocation may occur, with the animals reversing roles. In any case, an individual may act as a female on one occasion, then as a male on another. Chain matings also occur, in which the first animal in the chain acts as a male, the last acts as a female, and all animals in between act simultaneously as males and females. Sometimes the chains are closed, with the linking animal acting simultaneously as a male to one partner and a female to another. In a study of 190 mating pairs of Aplysia vaccaria that were observed along the beach in San Diego, 22 matings comprised chains of three animals, and three matings comprised

Reproduction

173

chains of four animals (Angeloni and Bradbury, 1999). Chains containing as many as ten animals have been observed in Aplysia californica. From detailed observations of groups of Aplysia, it is known that mating combinations can be highly dynamic, with roles and partners changing frequently (Kupfermann and Carew, 1974; Susswein et al., 1984; Pennings, 1991). In a study of A. fasciata (Susswein et al., 1984), four individuals were continuously observed in the laboratory for 3 hours. During this time, a total of 18 different mating configurations were seen involving open, closed, and even branched chains containing two, three, or four animals. Pulmonate gastropods are simultaneous hermaphrodites for the greater portion of their adult lives, although they usually develop with some degree of protandry. Mating may be either reciprocal or unilateral, depending on the shape of the shells. Snails that have flat shells, and slugs with no shells, mate face to face in the case of snails (e.g., Helix, Cepaed), but often hanging, in the case of slugs (see below). Snails that have shells with tall spires (e.g., Achatina, Lymnaea, Physd) mate by shell mounting. In these latter cases, mating is usually unilateral, with the snail that mounts acting as the male. Chain matings may occur in some families of basommatophores (Planorbidae, Lymnaeidae, Physidae) but not in terrestrial forms. An interesting relationship has been discovered between the mating behaviors of snails and the direction of the coil in their shells (Asami et al., 1998). Most snails have dextrally coiled shells. Sinistral coils are generally rare, but they are much more common in some species than in others. It was found that chiral reversal, that is, sinistral shells in species that are mostly dextral, is more common among tall spired snails than among flat spired snails. The explanation, according to Asami et al. (1998), is that flat-shelled snails mate face to face and reciprocally so, in this configuration, mating is physically impossible between snails of dissimilar chirality. By contrast, tall-shelled snails mate unilaterally, so they can still mate even if their partners have differently coiled shells. Consequently, selection against the chiral minority is relaxed in tall-shelled species, and populations of tall-shelled species contain more left-handed individuals than do populations of flat-shelled species. Although cross-fertilization is the rule for both pulmonates and opisthobranchs, quite a few pulmonate genera benefit from facultative selffertilization (e.g., Lymnaea, Rumina). When mates are available they copulate and cross-fertilize, but when they cannot find mates, they self-fertilize. Selffertilization is costly because it leads to an accumulation of deleterious homozygous genotypes (inbreeding depression). However, if it is mixed with outbreeding within a population, self-fertilization can sustain high rates of reproduction without serious genetic disadvantage. 8.7.3.

Mating Strategies in Simultaneous Hermaphrodites

The reproductive interests of males and females are typically different, which gives rise to sexual conflict. Theoretical considerations give plenty of scope

174 Behavior and Neurons in Gastropods

for sexual conflicts even in hermaphrodites (Michiels, 1998). Conventional theory predicts that, because sperm are cheap to produce relative to eggs, the male role should usually be preferred. Obviously, if both partners insist upon acting as males, and only as males, nothing will happen. However, the need to donate or to receive sperm will change from time to time depending on the animal's recent history. If an animal is already storing a large amount of received sperm (allosperm), it is not likely to need more. On the other hand, if an animal has not recently mated, it may not have enough allosperm with which to fertilize its eggs. Thus, if two animals both need to receive sperm, but neither is ready to donate, there will again be a conflict. One way to resolve these conflicts is to transfer sperm reciprocally. In fact, simultaneous reciprocity is obligatory in most species of land snails. In these animals, mutual insemination involves both members of the mating pair functioning as male and female. For most other gastropods, which mate unilaterally, conflicts between individuals are bound to occur, and it is not always clear how they are resolved. Some snails (e.g., Lymnaea, Par tula} practice serial reciprocity with role alternation whereby an animal will first mate as a male (sperm donor), and shortly thereafter it will mate again with the same partner but this time as a female (sperm recipient). In other species, the patterns of role playing are not so clearly defined. Individual Aplysia, for example, are said by some authors to consistently behave in just one sexual role, but other authors report frequent role switching (see Baur, 1998). Laboratory studies of Aplysia tend to report individuals that specialize in a single sexual role, whereas field studies tend to report frequent switches. In one field study (Yusa, 1996) involving Aplysia kurodai, 26 tagged animals were observed over a period of 3 days. On average, each animal mated 1.2 times per day in each sexual role, and the standard deviation (0.60) was identical for both sexual roles. Some authors maintain that there is a preferred sexual role and that, when partners alternate their sexual roles, the animal which first performs the less preferred role does so only when it expects the partner to reciprocate. Leonard argues that the female role is preferred in simultaneous hermaphrodites (Leonard, 1990; Leonard and Lukowiak, 1991). She points out that females have better control over fertilization than do males, they control the production of offspring, and they alone can be assured of reproductive success. Males, on the other hand, face competition from other males, so they risk having their sperm digested or lost. Thus, simultaneous hermaphrodites face a "dilemma," according to Leonard (1990), because both mating partners want to perform the female role. When she observed matings in the opisthobranch Navanax inermis, she noted that pairs mated successively in alternate sexual roles and there were often repeated bouts of matings between two animals (see Leonard and Lukowiak, 1991). Bouts of three or more copulations occurred in 19 of 59 observation periods, usually with alternating roles. In one case, two animals exchanged roles six times. Furthermore, in some situations, the female role seemed to be preferred (Leonard and Lukowiak, 1991).

Reproduction

175

Leonard's contention that hermaphrodites face a dilemma founded on their universal preference for the female role is flawed on several grounds. First, the empirical evidence for a preferred female role is weak in Navanax and absent from any other hermaphroditic gastropod species. Second, it cannot be true that all Navanax individuals prefer the female role or even that a substantial majority does because, if that were the case, then any individual willing to perform the male role would encounter many opportunities to reproduce, and hence the number of these animals in the population would increase, that is, the fitness of each sexual role is inversely proportional to its frequency in the population. Third, while it may be true that the risk to individual male actors is greater than the risk to individual female actors, the number of offspring produced by males and females is identical in every generation because each individual has both a mother and a father. Since the variance in reproductive success associated with each sexual role is therefore the same over time, the fitness of male actors is exactly equal to that of female actors (Greeffand Michiels, 1999).

8.2.

Sexually Selected Behaviors

A trait is said to be sexually selected when there is competition over mates and the expression of the trait is associated with differences in reproductive success. Whereas Darwin thought of sexual selection only in terms of precopulatory mechanisms, and it is fair to say that there is little evidence for precopulatory mate choice in gastropods apart from selection for size (see below), it is likely that hermaphroditism actually promotes postcopulatory sexual selection (Charnov, 1979; Michiels, 1998; Rogers and Chase, 2001). In gastropods, the complex and variable internal reproductive structures probably reflect the influence of sexual selection. A good example is the organs that receive, store, and digest allosperm in terrestrial pulmonates. These structures are so varied between taxonomic families, and even between species of the same family (Tompa, 1984), that no adequate description can be given here. More important than the details of this variation is the understanding that the principal evolutionary process that led to structural variation is the competition among sperm donors (males) to fertilize eggs. In taxa other than Gastropoda, males compete to gain the favor of females who might allow them to mate. In gastropods, mate selection is at best weak and not likely to be a major factor. Probably much more important is competition closer to the fertilization event, that is, competition among sperm transferred from different donors for access to eggs, commonly known as sperm competition. Two conditions, especially, favor the evolution of sperm competition: the capacity of females to store sperm and the occurrence of multiple matings prior to fertilization (promiscuity). As explained below, both conditions apply to gastropods. Lind (1988) studied mating patterns and egg laying in a marked population of Helix pomatia living in a Danish wood. From observations averaged over

176 Behavior and Neurons in Gastropods

6 years, he concluded that individual snails mate about 4.5 times per season, but they lay only one clutch of eggs. Sperm received in one season may either be used in the same season or stored until the next season. Similarly, the banded land snail Cepaea nemoralis is able to store allosperm for at least 4 years. Because of the promiscuity of these snails, and their ability to store sperm, a single clutch of eggs usually yields offspring sired by two or more fathers (Baur, 1998; Landolfa et al., 2001). The promiscuity of Aplysia has already been mentioned, and additional evidence can be added here. Thus, Aplysia fasciata is estimated to spend 25-30% of its time in sexual activity during the mating season (Susswein et al., 1984). Aplysia californica is so preoccupied with reproduction in early summer mornings around Santa Catalina Island that 50% of the animals can be found engaged in either courtship, copulation or egg laying (Pennings, 1991). In summary, since promiscuity and sperm storage are common in gastropods, it is likely that sexual selection through sperm competition is also common in gastropods. However, firm evidence that particular traits, whether behavioral or otherwise, have arisen through sperm competition is just beginning to appear (Baur, 1998). In the few cases where mate choice has been shown to occur in gastropods, it is based on size, in this case the body size of the prospective partner. In Aplysia, for example, larger animals are more frequently seen performing the female role than the male role (Yusa, 1996; Angeloni and Bradbury, 1999). The larger the size difference between two members of a mating pair, the greater the likelihood that the larger animal will be the female actor. Since large animals of all species typically produce more eggs (greater fecundity), choosing large female-acting mates could be a way for male-acting animals to increase their fitness. The advantage achieved by donating sperm to large mates is thought to explain certain mate-rejection behaviors seen in the freshwater snail Physa, which mates unilaterally and without reciprocation. The mate-rejection behaviors are only observed when a snail is mounted by another snail that is larger than itself. In such cases, the mounted individual commences to swing its shell vigorously to the left and right while keeping its foot stationary. The mounted animal also bites the genitals of its would-be suitor. Following these and other similar maneuvers performed by the larger snail, the smaller animal usually dismounts. It is easy to infer from these materejection behaviors that the intended sperm recipient would prefer to donate sperm to its larger partner. Consistent with this interpretation is the finding that a significant majority of uncontested copulations in Physa involve relatively small sperm donors copulating with larger sperm recipients, as in Aplysia (De Witt, 1996). Notwithstanding the foregoing, most sexual selection in gastropods is on traits that act after sperm transfer, not before. One such trait is the production of large numbers of parasperm, or sterile germ cells (Tompa, 1984; Baur, 1998). Present in many families of "prosobranchs," the parasperm are thought to help their fertile siblings, the eusperm, to outcompete

Reproduction

177

rival sperm. Several mechanisms have been proposed, including assisting in the transport of eusperm to the female's storage site, creating a false sense of fulfillment and thus delaying further matings by the female, secreting products that can harm sperm transferred subsequently by other males, and preventing later inseminations by plugging the female tract with secretions. Evidence in support of the postulated parasperm function has come from a study in the freshwater snail Viviparus ater in which the relative numbers of fertile eusperm and non-fertile parasperm were measured in males exposed to different degrees of competition from local rivals (Oppliger et al., 1998). It was found that, when competition increased, either through laboratory manipulation or from natural variations in the field, the production of parasperm increased significantly while the production of eusperm remained constant. Samples taken from six sites around Lake Zurich showed that the parasperm: eusperm ratio increased as a nearly linear function of the male: female ratio in the local population. The results indicate that the production and transfer of parasperm is an adaptive response to sperm competition. An important factor driving sexual selection in hermaphroditic gastropods is sperm digestion. Most species possess a special gameolytic organ that enzymatically digests many of the received sperm before they can reach the storage site. The immediate benefit for the female function is the nutrient value of the digested cells, but there are other important consequences. Since digestion limits the opportunities for fertilization and since, in promiscuous species, the sperm from different donors compete to fertilize eggs, sperm digestion selects for individuals that can transfer large quantities of sperm in each ejaculate. From these assumptions, Greeff and Michiels (1999) predict that, for simultaneous hermaphrodites that mate reciprocally, the result should be the investment of ever increasing resources into the male function at the expense of the female function until, ultimately, the allocation is biased towards the male function. Also, as the demand for sperm increases, the ability to transfer large numbers of sperm becomes more important than the ability to obtain frequent matings. In contrast to gonochorists, therefore, the selection on traits that increase the number of matings may be weak in reciprocally mating hermaphrodites. Lastly, the presence of sperm digestion should select traits, other than increased sperm production, that can negate or circumvent digestion. Some of the traits thought to have been selected in this manner are discussed below. One way to get around the problem of sperm digestion is to inject sperm hypodermically, a method of insemination that is seen in sacoglossan opisthobranchs (see Hadfield and Switzer-Dunlap, 1984). In these animals the penis is equipped with a hard cuticular style, an extension of the vas deferens, which is used to facilitate penetration of the penis through the skin of the partner. Sperm is then inserted either directly into a sperm storage chamber (bursa) that is attached to the inside of the body wall or more randomly into the hemocoel. In the latter case it is not clear how the sperm find their way to the storage site. However, both routes of insemination avoid the vaginal tract

178 Behavior and Neurons in Gastropods

altogether, and hence they avoid the dangers of obstruction, interference, and digestion that may be lurking there. In some species, sperm is exchanged externally. Emberton (1994) has suggested that, when sperm are transferred by external exchange, they might be able to bypass the gameolytic organ, and thus swim unharmed to the storage organ. Whether sperm exchanged in this manner do, in fact, avoid digestion remains to be proven, but the behavior can be observed in four families of terrestrial pulmonates representing both snails (Polygyridae, Succineidae, Discidae) and slugs (Limacidae). The penes of the two mating animals are first everted, then they intertwine, and eventually sperm is reciprocally transferred between the tips of the penes. Possibly the nudibranch Glaucus atlanticus also mates like this (see Fig. 16 in Lalli and Gilmer, 1989). The slug penis bears a chalice-shaped structure for molding and concentrating the mass of autosperm prior to transfer, and ridges for retaining the received allosperm. After sperm transfer, the penes are retracted, by inversion, and the mass of allosperm is presumably released into the female tract. The mating behavior of slugs is particularly interesting. They do it at night while hanging vertically from walls or tree branches on a mucus thread, as shown in Figure 8.1. During courtship the partners circle around each other

Figure 8.1. Two slugs, Limax cinereoniger, mating while they hang from a branch of a small bush. Their bodies and their penes are intertwined; they exchange sperm externally at the tips of their penes. Adapted from Solem (1974), Copyright 1974. Reprinted by permission of John Wiley & Sons, Inc.

Reproduction

179

secreting copious amounts of mucus and mouthing each other's tail. Langlois (1965) describes what happens next in the species Limax maximus: As each slug crawls onto the body of the other slug they crawl off the wall and do a slow-motion head-dive together, suspended by the tips of their tails on a doubled slime string. They pause when they have dropped from 6 to 12 inches below the point of attachment, and they dangle, head down and rotating on the end of their string.

Next, they evert their penes and transfer sperm at the tips. Following this, the slugs climb back up their mucus ropes or drop to the ground. Some species that exchange sperm externally have evolved penes of extraordinary length (Fig. 8.1). Limax corsicus, for example, has a body length of 12-15 cm and a penis length of about 60 cm; Limax redii has a body length of 13-15 cm and a penis length of 85cm (Baur, 1998). Emberton (1994) suggests that the exaggerated length of the penes in these species results from selection for devices to place the sperm at ever increasing distances from the digestive processes that have themselves evolved to counter external sperm exchange (an evolutionary arms race). But the penis in these animals is an organ that receives sperm as well as delivers it, so Emberton's hypothesis requires the impossible assumption that a trait (long penis) that benefits the sperm donor could be selected in the sperm recipient. Alternatively, the long penes might have evolved through the process of "runaway" selection on an arbitrary trait. A bizarre behavior known as apophallation occurs after some matings involving slugs of the genus Ariolimax. The penis in these animals is long and very thin at its end. Possibly because of its flexible structure, the penis sometimes becomes tied in a knot while inside the partner. Consequently, the penis becomes lodged in the female tract and the only way the mated slugs can disengage is for the penis to be severed. This service is rendered by the partner when it gnaws through the other's penis (see Baur, 1998). The disattached penis does not regenerate. Although both slugs survive the ordeal, the one which has lost its penis cannot mate again as a male and, since the matings are reciprocal, probably not as a female either. Some terrestrial snails, representing about 10 of roughly 65 pulmonate families, pierce their partners with "love" darts during courtship (Tompa, 1984). The darts are made of calcium carbonate crystals and they are typically very sharp, although their size, shape, and even their number varies considerably from species to species. Structural details of the dart from Helix aspersa are shown in Figure 8.2. The dart is contained within a muscular sac. Towards the end of courtship, when appropriately stimulated by the partner, first one snail and then the other snail forcefully expels ("shoots") its dart in the direction of its partner (see Fig. 8.6). Further details of dart shooting behavior in Helix will be given in Section 8.4.2. Here I note that numerous hypotheses have been proposed for the function of this bizarre behavior. The most popular hypotheses are species recognition, stimulation of the partner, signaling that the shooter is ready to mate or that it is ready to mate as a male, and presentation of a gift of calcium. A telling fact, however, is that copulation takes place with little or no regard to whether a dart has been shot

180 Behavior and Neurons in Gastropods

Figure 8.2. The "love" dart of the garden snail Helix aspersa is made of calcium carbonate (argonite). During mating, each snail shoots a single dart. The vanes of the dart hold a mass of mucus that is deposited on the dart from a special gland. Experiments by the author suggest that the dart functions as a hypodermic device to inject mucus into the mating partner, with the result that the partner stores more of the shooter's sperm than if it had not been hit by a dart. The act of dart shooting is illustrated in Figure 8.6. From Tompa (1984) with permission.

or whether it has found its target. Virgin snails mate and reproduce even though they do not possess darts, and adult snails may either fail to shoot or shoot so ineffectively that the expelled dart misses its intended target, but still they copulate. It is likely, therefore, that the dart is another example of a trait that has been selected by a postcopulatory mechanism. My colleagues and I have suggested that dart shooting evolved in the context of sperm competition as a way to increase the survival of the shooter's sperm in the recipient's female tract, and thus increase the shooter's success in fertilizing eggs. Specifically, we hypothesize that the dart acts as a hypodermic device to inject into the recipient a substance that contracts the female tract in such a manner as to promote the uptake of the spermatophore and prevent the digestion of transferred sperm (Koene and Chase, 1998). Just prior to its expulsion, the dart is covered with mucus from a special gland. Our experiments in vitro showed that application of the mucus, in amounts that are transferred on the dart, causes stereotypical contractions in the female reproductive tract. In a more direct test of the hypothesis, we mated virgin snails with partners that either shot their darts successfully (i.e., the dart remained embedded in the recipient throughout copulation) or that shot unsuccessfully (missed entirely). Seven days later we counted the number of donated sperm that were stored in recipient snails. As reported by Rogers and Chase (2001) the once-mated former virgins stored 116% more sperm from the successful dart shooters than from the unsuccessful shooters. Since the ultimate test of sexual selection is differential reproductive success, we also investigated whether dart shooting is associated with male reproductive success. In these experiments, a single mother (egg layer) was allowed to mate sequentially with two potential fathers (sperm donors). By chance, in some triads, one of the would-be fathers shot its dart successfully and the other did not. Once the mother had laid eggs, a genetic assay was used to determine the proportions of offspring in each clutch that were fathered by the two sperm donors. We found that the successful shooter sired significantly more offspring than the unsuccessful shooter, regardless of which snail mated first

Reproduction

181

(Landolfa et al., 2001). The result, therefore, supports the idea that the dart evolved in the context of sperm competition. 8.3.

Finding Mates

Mate finding, like food finding, depends on chemical cues. Whenever adults of the same species are seen aggregated together and engaged in sexual activity, a role for pheromones is suggested. However, aggregations may also result from an attraction to food sources or to other features in the environment; indeed, mating often occurs on feeding grounds. For these reasons, pheromones are implicated most strongly when aggregations and matings occur during times of non-feeding, or when food is unavailable. Pheromones allow for conspecifics to be recognized at a distance in cases where there is little or no capacity for visual or acoustic recognition, but chemical cues are not always entirely specific. Many authors have observed interspecific matings or interspecific courtships, usually between two closely related species. Only one pheromone, used by Aplysia, has so far been identified (see below). Aggregations of enormous size occur in the pelagic pteropod, Clione antarctica. This species is the southern hemisphere counterpart to Clione limacina, which lives in northern seas. Bryan and Slattery (1996) reported on swarms of C. limacina that were observed in McMurdo Sound, Antarctica. The average swarm was estimated to contain 1.3 million individuals, with mean densities of about 7.9/m3. The swarms were located in the first 6m of depth, and they moved with the current. When the scientists dove beneath the thin icy surface, they saw frequent copulations, especially near the surface. At 1 m depth, approximately 8% of the animals were copulating at any given time. It was noted that when animals happened to be separated by 2.5cm or less they often swam towards one another, whereas animals spaced farther apart rarely did so. Once a pair made contact, copulation followed shortly thereafter, with simultaneous reciprocal insemination. Of 40 pairs that came as close as 1.25cm, 38 pairs coupled. The authors were unable to determine for certain what sensory cues were used for mate recognition but they ruled out random contacts because seemingly accidental bumps were never seen. They conclude that chemical signals are most likely used for mate recognition, and they suggest, based on preliminary biochemical work, that water-soluble conjugates of progestin might be the specific cue. The aggregations of Aplysia, while not nearly as massive as Clione's, are even more obviously related to reproduction. Pennings (1991) observed aggregations of A. californica at Santa Catalina Island, California. They usually comprised 4-6 mature individuals and sometimes more than 12 individuals. Some aggregations persisted for a week or more, but the specific composition of a given aggregation changed over time, with some animals leaving and others joining. Only about one-third of the animals present on one day were again present on the next day. While animals belonging to an

182 Behavior and Neurons in Gastropods

aggregation may graze on algae from time to time, their main occupation is reproduction. Kupfermann and Carew (1974) observed Aplysia at one study site in southern California that they designated the "brothel," owing to the intense sexual activity that occurred there. Noting that the site was "dark and murky," the authors made the important observation that it contained numerous fresh egg cordons. Later, in laboratory experiments, Audesirk (1977) found that animals were most likely to mate if they were in a tank with another mating pair, especially if the female copulant was at the same time laying eggs. Blankenship et al. (1983) reported that 67 of 101 egg-laying animals were simultaneously copulating as females. These and similar observations led Audesirk (1977) to suggest that mating is triggered by a pheromone that is present in, and released from, the egg cordon. Painter and colleagues set out to identify the Aplysia pheromone (Painter et al., 1991). They first used T-maze experiments to confirm the attractive properties of recently deposited egg cordons and to demonstrate that the animals themselves were no more attractive after laying eggs than before laying eggs. When eluates of egg cordons were separated by high-pressure liquid chromatography (HPLC), four fractions were identified that were attractive to animals when presented in the T-maze (Painter et al., 1998). These fractions were then sequenced, and primers were derived that were used to generate a gene product from a cDNA library. This procedure yielded a 58-residue peptide, named "attractin," which was shown to be expressed in the albumen gland. When tested in the T-maze, attractin molecules were effective in attracting Aplysia. The attractive property of egg cordons is explained by these findings because the pheromone should be inserted into the egg cordon as it is constructed by the albumen gland. The presence of the four smaller peptides in the eluate, which are also attractive, is accounted for by extracellular degradation of the full-length attractin peptide. In sum, these experiments implicate the peptide attractin in the formation of naturally occurring aggregations of Aplysia, but whether it is also responsible for inducing mating behavior remains to be determined. So far, the evidence supporting the latter proposition is suggestive but inconclusive (Painter et al., 1998). The evidence for pheromonal attraction in "prosobranchs" is anecdotal but persuasive in the case of the spindle-shaped miter snail Mitra idac (Gate, 1968). Specimens were collected off the coast of Southern California and observed in an aquarium. The key observation is that males only approached females after the latter had exuded "a cloud of clear mucus." The mucus cloud, which hung about the shell of the female, often contained trapped air bubbles and/or sand particles. As soon as the mucus was released, several males (identified by their smaller size) would rapidly approach the female. To reach her, they sometimes took short cuts up suspended air hoses following paths not otherwise seen in 8 months of observations. Once a male had located the female, he aligned himself at her right side and intromitted. Copulations lasted about 1.5 hours.

Reproduction

183

We saw earlier that gastropods sometimes follow mucus trails to locate prey (see Section 7.2.3). Evidence for trail following to find mates comes mainly from terrestrial gastropod species. For example, Chase et al. (1978) found that Achatina fulica follows conspecific mucus trails but not those of Otala vermiculata. Moreover, mature snails that were housed in individual containers for 30 days had a greater tendency to follow conspecific trails than did snails that were housed collectively with opportunities for mating. These results suggest that trail following in Achatina is used to find mates. Further evidence comes from a study of the terrestrial slug Deroceras reticulatum in which trails were laid and followed in an outdoor arena (Wareing, 1986). Significantly, trails were followed in the same direction in which they were laid on 16 of 17 occasions, and the follower slug usually courted the trailblazing slug after catching up with it. Gastropod trails are made of mucus that comes from a gland embedded in the foot. The mucus is released through a pore at the anterior end as the animal locomotes. In some terrestrial slugs (Arionidae, Veronicellidae) there is an additional mucus gland situated on the dorsal surface of the foot just above the tail (=caudal gland). When present, secretions from this gland seem to be targeted during trail following. As courtship ensues the secretions become ever more copious, and they are eagerly ingested by the partner as soon as they are released. Some snail species may use airborne pheromones to find mates, either as a complement to trail following or as an alternative. A study of locomotor orientation in a T-maze olfactometer provides evidence for airborne chemical cues (Chase et al., 1978). Air was blown into one arm of the maze from a flask loaded with snails of the species Achatina fulica, while air was blown into the other arm from an identical flask that contained snails of the species Helix aperta. When specimens of the two species were tested alternately in the same apparatus, the snails oriented selectively to the odor of their own species. The so-called "head wart" is a candidate source of pheromones in the pulmonate families Bradybaenidae and Camaenidae (Tompa, 1984). The head wart is a tubercular outgrowth located on the dorsal aspect of the head directly between the posterior tentacles. An association with reproduction is indicated by the fact that the head wart is found only in sexually mature individuals. Histologically, it contains neither glands nor any other special cell type, but the epithelial cells are highly elongated. The function of the organ has been studied in the large Asian snail Euhadra peliomphala (Takeda and Tsuruoka, 1979). When the snails engage in courtship behaviors the wart repeatedly swells and contracts, and the partners' tentacles brush up against it. Experiments were done using alcoholic extracts of the wart to see how they affected behavior. The results indicate that the snails' overall activity increased after exposure to the wart product(s), but the extent to which such activity was sexual is unclear. Unusual formations of various kinds are occasionally found on the skin of other species in exactly the same place as the head wart (i.e., between the posterior tentacles). One of these, the "frontal organ," is present in snails of the genus Gymnarion from central and western Africa. It is an

184

Behavior and Neurons in Gastropods

intriguing structure that may serve an unusual function in reproductive behavior. According to Binder (1976), the frontal organ contains numerous small hooks, as many as 40 in one species. As the blood pressure at the anterior end of the snail increases during sexual arousal, the hooks are thrust outward by erection of the organ. Because the courting snails spend long periods of time face to face with abutted frontal organs, it is suggested that the hooks act like Velcro to secure the mate. Once contact is made with another gastropod, additional sensory mechanisms may come into play to confirm the species identity of the potential mate, similar to the way in which contact chemoreceptors are used to confirm the identity of a potential food. Many species have elaborate courtship behaviors involving repeated mutual contacts that could provide either tactile or chemical signals adequate for species recognition. However, courtships between related species are commonly observed, for example, between various species of Aplysia, suggesting that in many cases there may be no proximal mechanism for species recognition. The mechanical requirements for copulation are often sufficient in themselves to preclude interspecific matings, and postcopulatory mechanisms are often available to prevent cross-species fertilization.

8.4.

Nervous Control of Courtship and Copulation

Our knowledge of the neural control of mating behavior is derived almost entirely from work in three species, namely the pond snail Lymnaea stagnalis, the garden snail Helix aspersa, and the sea hare Aplysia californica. In addition to reviewing neurobiological studies, the following treatment also adds further detailed descriptions of courtship and copulation in these model species. It will be seen that, despite the considerable differences in behavior, there is a common basis for neural control. 8.4.1.

Lymnaea stagnalis

Mating in this snail involves little or no courtship. The female role appears to be entirely passive. Prior to intromission, the snail that will adopt the female role goes about her usual business of feeding, locomoting and air breathing. In contrast, the animal that will act as the male behaves in a stereotyped manner, as shown in Figure 8.3. He first climbs on to the shell of the prospective female (Fig. 8.3A), and he then travels around the shell while always following a counterclockwise direction. He may make one or more turns before he reaches the edge of the shell on the right hand side of the female, at which position he stops (Fig. 8.3B, C). Here he assumes a characteristic posture (Fig. 8.3D) in which the head-foot shortens and the tentacles droop. At this time the preputium is partially everted through the male pore at the animal's anterior right side. The preputium is a large muscular cylinder that carries the penis, as shown in Figure 8.4. As the preputium is further

Figure 8.3. The mating behavior of Lymnaea stagnalis. (A) The prospective male (dark line) mounts the shell of the prospective female (light line). (B) Circling. (C) Positioning. (D) Partial eversion of the penial complex. (E) Total eversion of the preputium; probing for the female's genital pore (dotted lines). (F) Intromission. Adapted from De Boer et al. (1996).

Figure 8.4. Drawings of the preputium and the penis in Lymnaea to illustrate how the organs are everted. The preputium is everted by being turned inside out, whereas the penis is everted by the combination of forward thrust and contraction of the inner preputium wall (the outer penis sheath). Eversion of the penis requires peptides to contract longitudinal muscles and to divert blood into the space between the penis and its inner sheath. The sensory neurons at the tip of the preputium, which express the peptide conopressin, are thought to provide feedback to the CNS while the preputium probes for the female gonopore. Adapted from De Lange et al. (1998a). Copyright 1998. Reprinted by permission of Wiley-Liss, Inc.

186 Behavior and Neurons in Gastropods

everted (Fig. 8.3E), it begins to probe for the female gonopore, which is located posteriorly under the shell. Once the tip of the preputium is inserted into the gonopore, the penis is everted and intromission occurs (Fig. 8.3F). The intromission phase of mating is quite constant in length, 36 ± 4 minutes, but each of the preliminary (appetitive) stages is variable in length, giving an overall duration for mating of several hours in some pairings. At the end of intromission, the penis is retracted and the snails separate. The motor control of the penial complex (preputium and penis) poses two special problems. First, the preputium and the penis must move independently of one another and without the benefit of any skeletal support. Second, the preputium needs to execute finely controlled probings. Furthermore, while the preputium probes for the female gonopore, it must provide sensory feedback to trigger penial eversion at the right moment. The performance of these tasks depends on the complex structure and physiology of the penial organ, which is shown schematically in Figure 8.4. The preputium is shaped like a flattened cone whose closed end remains within the body while the open end is extended for probing. Its wall contains primarily longitudinal muscles and circular muscles, but also some dorsoventral muscles. The penis is a muscular tube consisting of longitudinal and circular muscles. The wall of the preputium continues as a folded, or two-layered, penis sheath. Between the inner layer of the penis sheath and the penis wall there is blood space. The penis has its own retractor muscle and its own artery. Figure 8.4, taken from De Lange et al. (1998a), provides an explanation of how the preputium and the penis are everted. According to the detailed scenario presented by De Lange et al., the preputium is everted as the result of several coordinated actions which may be summarized as follows. First, the circular muscles around the gonopore relax; second, the hydrostatic pressure and/or blood pressure increases; third, contractions of the longitudinal muscles inside the body wall alternate with relaxation of the circular preputium muscles at the level of the body wall. The penis is everted by combining contractions of the longitudinal muscles in the outer penis sheath with increased pressure in the penial blood space caused by contractions of arterial muscles. By examining the penial complex using immunohistological methods, De Lange et al. (1998a) found that it is richly innervated by fibers containing several transmitters. Furthermore, various parts of the penial complex receive their own, specific, innervation. The transmitters thought to be involved in producing eversions of the preputium and the penis include the FMRFamiderelated peptides SEEPLY and DEILSR, myomodulin, Lymnaea inhibitory peptide, APGWamide, and 5-HT. In addition, there is a population of neurons in the distal tip of the preputium that stains for the peptide conopressin. These latter cells, which have neurites extending upwards through the epithelium, may provide the sensory feedback required during the probing actions of the preputium. When two snails meet, the time elapsed since the last mating as a male determines which animal will play the male role in the current encounter.

Reproduction

187

All snails are receptive in the female role at all times. Thus, the frequency of mating within a population is a function of the collective male motivation, or the male sexual drive. Experiments performed by De Boer et al. (1997a) succeeded in identifying the neural mechanism responsible for sexual drive in Lymnaea. First, they discovered that the volume of the prostate gland increases after a period of social isolation and decreases after copulation. Next, they found that if a single nerve that innervates the vas deferens is cut, male sexual drive, as expressed in male mating behaviors, is significantly reduced. Finally, experimental inflation of the prostate gland produced a major increase in spike activity as recorded from the vas deferens nerve (NP1). Together, the results point to the conclusion that male sexual drive is limited by the availability of seminal fluid. A similar manner of control is likely in at least some other gastropods, but not in species that use spermatophores because in these cases the prostate is greatly reduced in size and importance. To identify neurons that control the male copulatory apparatus (penis, preputium, vas deferens, and prostate gland), the penis nerve was backfilled with nickel-lysine (De Boer et al., 1996; De Lange et al., 1998b). Not surprisingly, given that the penis nerve is on the right side, all five cell groups that were identified by this procedure are on the right side of the CNS. Three groups form distinct morphological entities, namely the anterior lobe of the right cerebral ganglion (rAL), the ventral lobe of the right cerebral ganglion (rVL), and the Ib-cluster of the right pedal ganglion (Pelb). By contrast, backfilled neurons in the right pleural and parietal ganglia are more dispersed. Numerically, rVL contains the largest number of backfilled cells, 173 on average, but rAL has the largest percentage of neurons projecting into the penis nerve, about 75% (De Lange et al., 1998b). By combining nerve backfills with immunocytochemistry and in situ hybridization, De Lange et al. (1998b) discovered that neurons innervating the male sexual organs contain a total of nine different transmitter substances. The correspondences between backfilled neurons, the five anatomical units, and transmitter content are shown schematically in Figure 8.5. These intriguing results suggest that different functional aspects of male mating behavior may be controlled by different groups of neurons, each of which is chemically and spatially defined. Note that while the majority of cells in rAL and rVL were backfilled from the penis nerve, other cells in these areas were not backfilled despite having the same transmitter profiles as the backfilled cells. It is possible that the backfill method failed to label some of the projecting cells adequately. Alternatively, some cells in these lobes may innervate structures such as the body wall or foot that play ancillary roles in the expression of mating behavior. Note also in Figure 8.5 that multiple transmitters are co-localized in most rAL neurons, and that about 60% of the cells contain APGWamide and conopressin. These two transmitters have antagonistic effects on vas deferens muscles, with conopressin causing contractions and APGWamide inhibiting contractions. From these results, obtained in vitro, it was suggested that conopressin and APGWamide might act together to

188 Behavior and Neurons in Gastropods

Figure 8.5. Venn diagrams illustrating the correspondence between transmitter phenotypes and axonal projections in the CNS of Lymnaea stagnalis. Shading represents neurons backfilled from the penial nerve. The arrows indicate how neurons of the right anterior lobe innervate other cell clusters and other central ganglia. From De Lange et al. (1998b). Copyright 1998. Reprinted by permission of WileyLiss, Inc.

produce peristaltic contractions of the vas deferens and ejaculation during intromission. However, injections of APGWamide and conopressin into intact animals, either singly or in combination, failed to produce ejaculations (De Boer et al., 1997b). A final significant finding illustrated in Figure 8.5 (see also Croll and Van Minnen, 1992; De Boer et al., 1997b) is that large numbers of APGWamide-containing nerve fibers project from rAL to other cell groups innervating the male sexual organs. This result, suggestive of a hierarchical arrangement, prompted a search for the functional role of the right anterior lobe. Experiments performed by De Boer et al. (1997b) demonstrate that one role for the right anterior lobe is eversion of the preputium. The authors reached this conclusion after successfully implanting fine wire electrodes into the rAL, from which they recorded nervous activity during natural mating behavior. The records obtained with this method show that activity in the rAL is negligible in the absence of sexual behavior but it increases significantly when the preputium is everted (Fig. 8.3D). The spiking activity continues throughout intromission (Fig. 8.3F) until the preputium is retracted. On average, in four successful recordings, neuronal firing rates

Reproduction

189

increased from 21 spikes/minute during the 10 minutes preceding eversion to 147 spikes/minute during the first 10 minutes after eversion. The implanted fine wire was also used to stimulate rAL neurons in quiescent animals. Three of 24 animals responded by fully everting their preputium, while another 20 animals responded with partial eversions. One unique aspect of the rAL is that it contains the majority of CNS neurons with positive APGWamide immunoreactivity (Croll and Van Minnen, 1992; De Lange et al., 1998b). Even the homologous left anterior lobe contains only 8-15 immunoreactive cells, compared to 120-150 immunoreactive cells in the right anterior lobe (Croll and Van Minnen, 1992). APGWamide-containing fibers and axon terminals, presumably belonging to rAL somata, are present in muscles of the male copulatory apparatus, the prostate gland, the vas deferens, and the body wall surrounding the male gonopore (Croll and Van Minnen, 1992; De Lange et al., 1998a). In light of these findings, De Boer et al. (1997b) decided to look for behaviors that could be elicited by injections of APGWamide into intact animals. As expected, injections of APGWamide at concentrations of 10~~ 5 M to 10~ 6 M reliably produced eversions of the preputium. To test whether the injected APGWamide was acting directly at peripheral sites or through CNS pathways, some injections were made after lesions of the penis nerve. Since the lesions did not affect the ability of APGWamide to elicit eversions of the preputium, the injected peptide must have acted directly on its peripheral targets. Presumably additional actions of APGWamide are mediated centrally through fibers that connect rAL to other CNS sites; these other functions might include controlling the probing actions of the preputium or triggering penial eversions. Subsequent to its discovery in Lymnaea, APGWamide has been further reported in a number of gastropod species. Signficantly, the peptide is consistently found within groups of neurons located in the anteromedial region of the cerebral ganglion. In the "prosobranch" Littorina and the basommatophoran Bulinus, immunocytochemical labeling identifies APGWamide within a bilateral pair of cell clusters located in the left and right cerebral ganglia (De Lange and van Minnen, 1998). In Lymnaea, Aplysia, Avion, Limax, and Helix, APGWamide cells are primarily located in a single cluster on the right side of the cerebral ganglion (see Fig. 8.7; Li and Chase, 1995; Fan et al., 1997; De Lange and van Minnen, 1998). When taken together with data from several species implicating the right anteromedial region of the cerebral ganglion in the control of mating behaviors (see below), APGWamide emerges as a peptide that is especially associated with the control of such behaviors. 8.4.2.

Helix aspersa

Because Helix mates as a simultaneous reciprocal hermaphrodite, male and female roles are not partitioned between the partners; instead, each role is performed simultaneously by both individuals. The following account of mating in the brown garden snail Helix aspersa is based on Adamo

190 Behavior and Neurons in Gastropods

and Chase (1988) and supplemented by other observations in my laboratory. As in Lymnaea, matings in Helix increase after a period of social isolation, but even after several weeks of isolation no more than 10-20% of the snails will mate on the first day after their isolation. By contrast, comparable isolation of Lymnaea causes 100% of the animals to copulate when given the opportunity (De Boer et al., 1997a). When two mature Helix snails meet, they may begin to touch each other with the tips of their maximally extended posterior tentacles. A pair of snails will typically exchange touches head to head for tens of seconds, then break away and move in circles until they find each other again, at which point further touching ensues; or they may separate entirely and quit the courtship. Gradually, as courtship continues, the genitals progressively swell and protrude from the common opening at the anterior right side of the animal. First the penial lobe is everted, and then the vaginal lobe. The degree of eversion, in which six stages are identified, is indicative of the extent of arousal and predictive of copulation. As courtship progresses, the animals may bite each other on the genitals and lips. An animal withdraws sharply when bitten but it recovers within a few seconds to resume courting. Separations followed by reunion, or in some cases by abandonment, may still occur. The aforementioned behaviors typically continue for about 35 minutes, but courtships of several hours are not uncommon. The imminent approach of dart shooting is evident when the snails align themselves with their heads facing in opposite directions and their genital openings apposed on the right sides. An animal prepares for dart shooting by stopping its locomotion and contracting the anterior portion of its body, presumably to build up hydrostatic pressure. Dart shooting appears to be triggered when the animal senses its partner pressing against its own everted genitals, or against the skin immediately surrounding the genital pore. Muscles then pull on the dart's sac to create a sudden and forceful expulsion of the dart. Figure 8.6 is a photograph of a snail in the act of shooting its dart. About one-third of all shot darts miss the partner entirely and are retracted. Another one-third of the darts strike the partner with only a glancing blow before falling to the ground. The remaining one-third of shot darts penetrate deeply enough to remained lodged in the partner for the duration of copulation, although few of these are fully internalized. Most of the successful shots hit the partner on the right flank, often near the genitals, but about one-quarter of the darts penetrate the partner's foot. Other results are more bizarre, such as darts that go into the partner's genitals or through the partner's head. Once, when dissecting a snail for a physiological experiment, I found a dart stuck in the cerebral ganglion. We have also watched snails shoot themselves in the foot or even in the head. For the reasons explained in Section 8.2, I think that the function of the dart is to increase sperm storage in the dart recipient. Within about 30 minutes of the first snail shooting its dart, the second animal shoots. The action, and the consequences, are the same in the second shooting as in the first. Thereafter, the animals begin to intromit. Since each penis must be inserted into the other animal's female tract,

Reproduction

191

Figure 8.6. Dart shooting during courtship in Helix aspersa. The animal at the bottom of the photograph has contracted the anterior part of its body to push out its dart (dark arrow); note that its tentacles are pinched inwards. The targeted animal has everted genitals (open arrow) indicative of sexual arousal. Although darts are often received in the vicinity of the genitals, in this case the recipient was too far away from the shooter so the dart missed; it was retracted by the shooter and digested. Photograph by Michael Landolfa. Used with permission.

several intromission attempts (up to 50) are usually required before both have simultaneously intermitted. The spermatophores are now formed, and the copulating pair remains motionless until both spermatophores have been completely transferred. Although the foregoing account of mating in Helix aspersa applies fairly well to other terrestrial pulmonates that likewise mate face to face, details differ even for closely related species. For example, it has already been noted that most land snails do not use darts. Helix pomatia, which does shoot a dart, mates much like H. aspersa but it does so in an upright posture that allows the sticky soles of the partners' feet to rub against each other as in an erotic slow dance. The drama and splendor of these actions is captured in an unforgettable scene from the commercial film Microcosmos (1996). An additional difference between the mating behaviors of H. aspersa and H. pomatia is that copulation lasts for about 7 hours in H. aspersa, but only about 5 minutes in H. pomatia. These and other differences in the mating behavior of H. aspersa and H. pomatia have led some taxonomists to reassign the two species to separate genera. The male sexual organs of Helix aspersa are innervated by two peripheral nerves, the penis nerve and the nervus cutaneus pedalis primus dexter, or NCPPD, which innervates the dart sac. Backfills of either nerve label cells mainly in the right cerebral ganglion and the right pedal ganglion (Fig. 8.7;

192 Behavior and Neurons in Gastropods

Figure 8.7. Conserved locations for CNS neurons involved in the control of male mating behaviors. Results from morphological, electrophysiological, and immunohistochemical studies implicate the right anteromedial portion of the cerebral ganglion and the right pedal ganglion (shaded) as key areas. H cluster; Ib cluster; Meso, mesocerebrum; rAL, right anterior lobe. In each illustrated species, the penial complex is innervated by one cerebral nerve and one pedal nerve. Adapted from Koene et al. (2000). With permission of Company of Biologists Ltd.

Eberhardt and Wabnitz, 1979; Li and Chase, 1995). The backfilled cells in the pedal ganglion have not been extensively studied, but some of them are penis retractor motoneurons (Eberhardt and Wabnitz, 1979) and others are probably motoneurons involved in dart shooting. The backfilled cells in the cerebral ganglion are primarily found in the mesocerebrum (see Fig. 2.6). The right mesocerebrum contains many more cells than the left mesocerebrum, and the cells are much larger on the left side than on the right side (Chase, 1986). The axon projections of right mesocerebral cells into the penis nerve and the NCPPD was confirmed by intracellular injections of tracers (Li and Chase, 1995). Approximately 25% of 145 filled mesocerebrum cells had an axon in the penis nerve and another 25% had an axon in the NCPPD. Most mesocerebrum cells have projections into two or more nerves, and about 12% of them have axons in both the penis nerve and the NCPPD. In virtually all cases (139 of 145 filled cells), however, the right mesocerebrum cells project an axon into the right pedal ganglion through the cerebropedal connective. Some of these axons continue directly into the NCPPD to gain access to the dart sac, while others presumably terminate in the pedal ganglion where they synapse with pedal motoneurons or premotor neurons. Immunocytochemical studies show that the right mesocerebrum reacts to antisera raised against a number of peptides. Li and Chase (1995) found that APGWamide immunoreactivity is most closely associated with cells projecting into the penis nerve, while FMRFamide is primarily associated with cells projecting into the NCPPD. Consequently, we proposed that APGWamide and FMRFamide have different behavioral functions, with APGWamide involved in penial eversion and FMRFamide involved in

Reproduction

193

dart shooting. Injections of APGWamide into intact animals support the first part of the hypothesis because they result in penial eversions, but injections of FMRFamide do not produce dart shooting (Koene et al., 2000). Another peptide found in the mesocerebrum is GFAD (Poteryaev et al., 1998). When GFAD is injected into an intact snail, the animal stops locomoting, its heartbeat increases, its head swells, and the genital pore opens. These effects are similar to those seen prior to both penial eversion and dart shooting, but neither of these behaviors was observed after GFAD injections, perhaps because sensory inputs were lacking. Electrical recordings show that right mesocerebral neurons are quiescent in the absence of inputs from cerebral peripheral nerves (Chase, 1986). Tactile stimulation of the skin between the posterior tentacles, or in the region surrounding the genital pore, evokes discharges in the right mesocerebrum that can be recorded in vivo using an implanted fine wire (Koene et al., 2000). When the mesocerebrum was stimulated through the same implanted electrode, penial eversions were elicited. When direct electrical stimulation of the right mesocerebrum was combined with tactile stimulation of the genital pore area, the animal both everted its penis and shot its dart. These results are consistent with behavioral observations, suggesting that physical contact with the partner triggers dart shooting (see above). They also corroborate results from reduced preparations showing that individual right mesocerebral cells are capable of evoking movements of either the penis or the dart sac when stimulated with an intracellular electrode (Chase, 1986). One animal with an implanted electrode engaged in a completely normal episode of courtship and copulation, during which time extracellular recordings were continuously obtained from the right mesocerebrum (Koene et al., 2000). A computer program was used to sort and recognize action potential waveforms. By correlating spiking activity with recorded video images, it was shown that 10 of the 11 identified units increased their discharges during physical contacts with the partner. Several of these same units increased their firing during either penial eversions or during dart shooting; three units increased their activity during both penial eversions and during dart shooting. The studies summarized above suggest that the right anterior lobe of Lymnaea and the right mesocerebrum of Helix are homologous structures. In addition to being similar morphologically, as shown in Figure 8.7, both structures contain neurons that have the necessary anatomical and physiological properties appropriate for roles in the control of mating behavior. Another common feature is the prominent chemical signature from the peptide APGWamide. It is likely, as we see below, that homologous neurons are also present in opisthobranchs. 8.4.3.

Aplysia

The mating behavior of Aplysia is rather simple and it more resembles mating in Lymnaea than in Helix (Kupfermann and Carew, 1974; Susswein et al., 1984). As previously discussed (see Section 8.1.3), it is unclear what

194 Behavior and Neurons in Gastropods

factors determine whether a particular animal will take the male role or the female role in a given encounter (see also Pennings, 1991). In any case, when a pair of animals first meet, they use their heads and tentacles to explore the upper body surfaces of their prospective mates. After a variable length of courtship lasting from several minutes to an hour or more, the prospective female eventually becomes immobile and fixed to the substrate as the prospective male climbs on to her. Either animal may occasionally wave its head from side to side, much as it would do when searching for food. The animal in the male role probes with his head inside and under the anterior portion of the female's parapodia searching for her genital pore. His partner may assist him by spreading open her parapodia; he may then wrap the anterior part of his foot around her exposed mantle to secure a hold. Intromission follows quickly once he has found her genital pore, and sperm transfer (without a spermatophore) is usually completed within 1 hour. A common feature of mating in Aplysia is the formation of chains with multiple animals (see Section 8.1.2). The conglomerate has been described as a "raveled knot" (Susswein et al., 1984), in which the exact arrangement of animals and the sexual roles of each individual are difficult to make out. However, Susswein et al. (1984) observed that group matings in A. fasciata commonly occur when an animal that is already mating as a female begins to court an adjacent animal that is mating as a male. When the outsider joins an existing group it usually does so as a male (10 of 11 occasions). Thus, the male joins on the back end of the chain rather than in the front or in the middle. Pennings (1991) made similar observations in A. californica. In an early study of function in the nervous system of Aplysia, Bottazzi and Enriques (1900) reported that electrical stimulation of the right cerebral ganglion causes peristaltic contractions in the penial sheath. Curiously, there have been no more recent investigations in Aplysia using either electrical stimulation to evoke components of sexual behavior or implanted electrodes to monitor neural activity in mating animals. However, Fan et al. (1997) discovered APGWamide immunoreactivity in a cluster of 30^0 neurons at the anteromedial margin of the right cerebral ganglion. This finding gives credence to the work of Bottazzi and Enriques (1900), and it fits well with results obtained from Lymnaea and Helix. The position of the labeled cell cluster corresponds to the position of the previously identified H-cluster (see Fig. 6.7A), which is the only cell cluster in the cerebral ganglion of Aplysia without bilateral representation. Neurons in the H-cluster send axons into the lower labial nerve, from which one branch emerges to innervate the penial complex. Given that the APGWamide signal is stronger in the H-cluster than in any other region of the Aplysia CNS, that a similar association is true of the right anterior lobe of Lymnaea and the right mesocerebrum of Helix, and that anatomical and electrophysiological evidence implicates all three structures in the control of mating behavior, Koene et al. (2000) proposed that the structures are homologous (Fig. 8.7).

Reproduction

195

A small number of APGWamide immunoreactive cells is present in the right pedal ganglion of Aplysia intermixed with unlabelled cells (Fan et al., 1997), but there are apparently none in the left pedal ganglion. Again, this pattern of labeling implicates a CNS region that is known to have connections with the male organs in Lymnaea and Helix, and it is consistent with knowledge that the right pedal ganglion of Aplysia contains motoneurons of the penis retractor muscle. Additional APGWamide immunoreactivity is present in the penial complex and in the male ducts that carry autosperm to the penis, suggesting that APGWamide may be involved in mediating ejaculation. In summary, evidence supports the conclusion that two evolutionarily conserved regions in the gastropod CNS contain neurons that control male mating behaviors: the right anteromedial cerebral ganglion, and the right pedal ganglion (Fig. 8.7). This is especially interesting in light of the diversity of mating behaviors seen in the representative species. However, more evidence is needed to substantiate the putative homology, particularly in respect to Aplysia where, for example, the consequences of APGWamide injections into intact animals have not been tested, although APGWamide was shown to cause penial eversions in reduced preparations (see Fan et al., 1997). Also, further studies are needed in other taxa to define the roles of APGWamide, and to examine the patterns of cellular specialization that account for behavioral diversity.

8.5.

Egg Laying

Apart from the Archaeogastropoda-Vetigastropoda, whose gametes mix in the sea, all gastropod molluscs reproduce by internal fertilization. The vast majority of species are oviparous, meaning that fertilized eggs are encapsulated but otherwise externalized with little or no embryonic development. In some species the eggs may be retained either to wait out unfavorable environmental conditions or for nourishment (brooding). Rarely, however, are eggs retained so long that the young hatch in the reproductive tract (ovoviviparity). True viviparity, in which the eggs are not encapsulated and the embryos undergo full development in the female tract, is extremely rare and present only in a few species of land snails (Tompa, 1984). Before fertilization, oocytes and allosperm are stored separately within the female tract. Therefore, the fusion of gametes occurs only after the gametes have been released from storage and allowed to mix. Typically, fertilization occurs in a special region of the female tract called, appropriately, the fertilization pouch. In oviparous species, the gametes are released and mixed just minutes before the fertilized eggs are externalized. "Prosobranchs" typically package their eggs in capsules that are either deposited on the sandy ocean floor or attached to a hard surface (Fretter, 1984). The number of eggs per capsule ranges from one to several thousand, and multiple capsules can be deposited. Thus, different species of Conus snails

196 Behavior and Neurons in Gastropods

in the Indian Ocean deposit between 30,000 and 1,500,000 eggs per occasion. Some caenogastropod species release their eggs in gelatinous strings. The large Strombus snail, for example, releases gelatinous strings that attain lengths of at least 20 m and that contain as many as 430,000 eggs. Terrestrial pulmonates lay eggs that are individually covered with either jelly-like shells or calcified shells. As the embryos develop, the calcium in the egg shell is slowly incorporated into the shells of the babies. Egg sizes span a remarkable two orders of magnitude, from a diameter of 0.5mm in Vallonia to a diameter of 50 mm in Strophocheilus (Tompa, 1984). In most cases the eggs are deposited in moist soil to prevent desiccation and predation. Snails use their feet to excavate a shallow subsurface chamber in soft soil (see Fig. 10.1). Only after the chamber is fully dug out will the snail initiate the processes of ovulation and fertilization, during which time it remains in position with its head hanging downwards into the chamber. Thus, in Helix there is a delay of 6-12 hours before any eggs appear in the nest, and another 24 hours or more before all the eggs are deposited. Egg numbers are quite variable, ranging from one to several hundred depending on the species. Helix lays 40-100 eggs per clutch. Once the eggs are laid, the mother abandons the nest but leaves it covered by soil. Freshwater snails (Basommatophora) lay eggs in gelatinous masses that are firmly attached to hard surfaces. The egg laying behavior of Lymnaea stagnalis is described in detail here because it has been studied neurobiologically (see below). Three phases of egg laying behavior can be distinguished (Hermann et al., 1994). The first phase is a resting phase in which the animals stop locomoting and slightly contract their foot; the tentacles droop and the shell is drawn down on the body. Two behaviors, turning and rasping, dominate during the next phase, which lasts 1-1.5 hours. Turning behavior prior to egg laying involves only the shell and mantle of Lymnaea; the foot remains in place. A single turn moves the shell 60-90°, always in a counterclockwise direction relative to the resting position. The new position is maintained for several minutes before the shell is returned to the resting position. An animal typically performs 2-A such turns before oviposition. Their function is probably to aid the transport of the egg mass through the reproductive tract to the genital pore. The function of the rasping behavior that occurs during this same phase is to clean a surface for better attachment of the eggs. Even when the surface happens to be an edible leaf, only a thin superficial layer of potential food is actually removed. From an analysis of buccal movements, it has been shown that the action of rasping is slightly different when performed during egg laying and when performed during feeding (see Jansen et al., 1999). For example, the duration of the protraction phase is significantly longer during egg laying than during feeding (means 1.72 and 0.81 seconds, respectively). The observed differences in rasping behavior in the two contexts are correlated with differences in central pattern generation and metacerebral giant cell activity, as shown by electrical recordings in vivo (Jansen et al., 1999). The final phase of egg laying behavior in Lymnaea consists of oviposition. For this, the animal moves slowly forward while extruding the egg mass

Reproduction

197

and pushing it against the substrate. About 75 eggs are included in each egg mass, and oviposition is usually completed within about 15 minutes. Opisthobranchs generally live for just 1 year, and they breed seasonally. Their eggs are laid within gelatinous masses that take different forms in different species. Depending on the size of the mother, the number of eggs can be very large. A frequently cited figure for Aplysia californica is 140,000,000 eggs laid by one mother in a single mass, as reported by MacGinitie (1934). Although Aplysia releases its eggs in a string that is manipulated to create a fairly compact mass (see below), it is calculated that the total length of all strings laid by a single animal in 1 day can reach 606 meters (Kandel, 1979). An individual animal lays eggs repeatedly during the spawning season, typically at intervals of 1 or 2 days. One specimen of Aplysia that was observed in captivity by MacGinitie (1934) laid 4.78 x 108 eggs in 27 separate episodes over 4 months; however, only eggs in the first 15 spawns were fertile. Detailed quantitative descriptions of egg laying in Aplysia californica take advantage of the fact that egg laying can be elicited by injecting an animal with hormone or by causing discharges of the bag-cells (see below). The component behaviors that are associated with egg laying appear in a stereotyped sequence. An animal that will lay eggs first slows its locomotion and begins to move its head and neck in certain characteristic ways. These movements become increasingly frequent so that, by the time the animal is ready to oviposit, it is immobile but far from quiescent. Ferguson et al. (1989a) described four types of head and neck movements that occur in Aplysia before and during ovipositon. Head waving consists of large side to side movements like those seen during feeding; these might be used to search for a suitable substrate. In nature, eggs are typically laid at spots where others have previously been laid, thus creating large mounds of multicolored egg masses (Kupfermann and Carew, 1974). Once a site has been selected, the substrate is prepared by small up and down undulations of the head. Weaving movements appear next, but usually only after the eggs have begun to emerge; these are side to side movements that serve to distribute the egg string. The eggs come out from the genital pore, which is located near the base of the right tentacle. As soon as the egg cordon leaves the genital pore, it enters an external groove that lies within the skin on the right side of the neck; this directs the cordon towards the mouth. As the head moves from side to side, the egg string is brought into contact with the mouth, where it receives a sticky secretion that facilitates its attachment to the substrate. Since the posterior portion of the animal's foot remains attached to the substrate during head weaving, the eggs tend to pile up or form a knot. Tamping movements of the head also occur during oviposition, intermixed with weaving movements. Tampings are dorsal-ventral movements that are used to press the eggs on to the substrate. For purposes of description, and with possible implications for neural control (see Section 8.6), Ferguson et al. (1989a) designated waving and undulation as appetitive behaviors, and weaving and tamping as consummatory behaviors.

198 Behavior and Neurons in Gastropods

The total duration of the foregoing events is approximately 3 hours, with the consummatory phase beginning at about 1 hour. Feeding behaviors are suppressed during egg laying, especially during the consummatory phase, probably as an adaptation to prevent the eggs from being consumed by the egg layer. 8.5.7.

Neuroendocrine Cells Control Egg Laying

Kupfermann (1970) made a key discovery when he found that an extract of the bag-cells from Aplysia will cause egg laying after injection into an intact animal. Subsequent work, described below, demonstrated that, under natural conditions, an egg-laying hormone is released by the bag-cells. Injections of ELH cause not just the expulsion of eggs but also the expression of ancillary behaviors as summarized above (Bernheim and Mayeri, 1995). Besides ELH, the bag-cells release additonal peptide hormones that help to orchestrate egg laying activities. A similar command structure for egg laying exists in Lymnaea, where a comparable group of neurosecretory cells, the caudodorsal cells, plays the central role. The overall scheme of control in these animals, involving neural pathways, hormones and peptides, is shown in Figure 8.8. The remainder of this chapter treats the various elements of Figure 8.8, with emphasis on the bag-cells of Aplysia and the caudodorsal cells of Lymnaea. Many aspects of function in the bag-cells have been reviewed by Conn and Kaczmarek (1989). The bag-cells, shown in Figure 8.9, part Al, are so named because they are found in two bag-like clusters on the rostral edge of the parietovisceral (abdominal) ganglion at the points of origin of the pleurovisceral connective nerves. In mature animals there are about 400 multipolar cells in each cluster (Coggeshall, 1967). Some of the cells' neurites enter the connective nerve bundle while others extend over the surface of the parietovisceral ganglion, but the majority of the neurites ramify in the nerve's connective tissue sheath where they form a cuff around the central core of axons. The neurites, which contain electron-dense neurosecretory granules, terminate near capillary spaces in the shealth, and it is here that the peptidergic contents of the granules are released. For this reason, the nerve sheath is said to be a neurohemal organ. Electrophysiologically, the bag-cells are ordinarily silent. However, when they are adequately stimulated, the cells produce sustained discharges of action potentials that generally last about 30 minutes (Fig. 8.9, part A2). Electrotonic coupling between cells synchronizes the discharges of individual bag-cells within a cluster and also synchronizes the discharges of left and right clusters. The action potentials have unusually long durations, 30-150 milliseconds. To investigate the function of bag-cell discharges, Pinsker and Dudek (1977) videotaped the spontaneous behaviors of Aplysia while simultaneously recording bag-cell electrical activity using an implanted "cuff" electrode wrapped around the connective nerve (Fig. 8.9, part Al). They found that spontaneous egg laying never occurred in the absence of a bag-cell discharge,

Reproduction

199

Figure 8.8. A schema of mechanisms and pathways responsible for egg laying, based on data from Aplysia and Lymnaea. Only the presence of the atrial gland in Aplysia distinguishes the two species, and its role in spontaneous egg laying, if any, is uncertain. Overt behaviors differ in the two genera, but include head movements, changes in locomotion, inhibition of feeding, and rasping radular movements.

and that every bag-cell discharge was followed by egg laying about 30 minutes later (see also Dudek et al., 1979). Thus, these experiments convincingly demonstrated that bag-cell discharges cause egg laying. No correlation was found between the duration of the bag-cell discharge and the number of eggs released. Since discharges even as brief as 3 minutes evoked normal egg laying, it is evident that there is a large margin of safety in the amount of ELH released by the bag-cells. Cells similar to the bag-cells are found in at least five genera (seven species) of basommatophoran snails (Roubos and Van de Ven, 1987). These cells, the caudodorsal cells (CDCs), are located at bilaterally symmetrical medioposterior positions in the cerebral ganglia, as shown in Figure 8.9, part Bl. In Lymnaea stagnalis, there are about 25 cells in the left cerebral ganglion and about 75 cells in the right cerebral ganglion. CDCs on the dorsal surfaces of the clusters are electrically coupled through junctions located within a distinct

200 Behavior and Neurons in Gastropods

Figure 8.9. Electrical discharges in specific clusters of neuroendocrine cells trigger egg laying. (Al) Morphology of bag cells in Aplysia, shown with implanted cuff electrode. From Dudek et al. (1979). (A2) A spontaneous discharge of the bag-cell population recorded in vivo using a cuff electrode of the type illustrated in (Al). Reprinted with permission from Pinsker and Dudek (1977) Copyright 1977 American Association for the Advancement of Science. (Bl) Cerebral ganglion of Lymnaea showing the dorsal and ventral caudodorsal cells (CDCs). Electrotonic coupling between CDCs occurs in the loop area. (B2) An electrical discharge evoked in one CDC by intracellular electrical stimulation of another CDC. (B1,B2) Copyright 1988. Adapted from Brussaard et al. (1988) with permission from Excerpta Medica Inc.

region of axonal looping. On the ventral surfaces there are about seven CDCs that are slightly larger than those on the dorsal surfaces and that have an additional axon, which crosses the cerebral commissure to synchronize electrical activities between left and right clusters. The ELH is released into the hemolymph from sites at the periphery of the cerebral commissure, much as in the pleurovisceral connective nerve of Aplysia. Additional sites of release, at least for the ventral cells, are located within the core axon bundle of the cerebral commissure. These latter sites lack synaptic specializations but presumably allow for communication with elements of the central nervous

Reproduction

201

system. Like their counterparts in Aplysia, the CDCs are normally silent, but they can be induced to fire synchronous, long-lasting discharges (Fig. 8.9, part B2). Recordings obtained from the CDCs in vivo are consistent with recordings from the bag-cells of Aplysia in that CDC discharges invariably precede egg laying. Since other electrical properties of the two cell types are likewise similar, and the secretion products of the CDCs and the bag-cells have a high degree of molecular homology (see below), it can be concluded that the CDCs and the bag-cells are homologous. Efforts to find homologues of the bag-cells in other gastropods have met with mixed success. Cross-species injections of bag-cell extracts from Aplysia have been reported to induce egg laying in a few other opisthobranchs, and intraspecific injections of material from the pleural and visceral ganglia produce egg laying in other species of Aplysia, consistent with the location of the bag-cells in A. californica, but no cell clusters morphologically equivalent to the bag-cells have yet been found in any opisthobranch other than Aplysia (see Hadfield and Switzer-Dunlap, 1984). In contrast to Aplysia, the cells that control egg laying in Pleurobranchaea seem to be located in the medial lobe of the pedal ganglion. To search for homologues of the bag-cells in pulmonate and "prosobranch" species, immunocytochemical investigations have used antibodies raised against either bag-cell peptides or CDC peptides. In Helix aspersa and Limax maximus, antibodies raised against a CDC peptide label a few large cells in the visceral ganglion and the right parietal ganglion; in Limax, a few smaller immunoreactive cells are also present in the cerebral ganglia (Van Minnen et al., 1992). At odds with these results are those of Croll (R.P. Croll, unpublished observations), who surveyed the CNS of Helix aspersa for immunoreactivity to an antibody raised against ELH; he found that most of the reactive cells were in the pleural ganglion. Three genera of "prosobranchs," namely Busycon, Concholepas, and Tegula, were examined with antibodies raised against both bag-cell ELH and CDC ELH (Ram et al., 1998). In all cases, the antibodies labeled cell clusters of about 300 neurons along the medial margins of the left and right cerebral ganglia. Collectively, these studies suggest that egg laying in gastropods is always controlled by neuroendocrine cells, but the identity of the relevant cells remains to be established in most species. 8.5.2.

Secretory Products of Neuroendocrine Cells

Studies on the neuroendocrine cells responsible for egg laying in Aplysia and Lymnaea show that there exists a small family of evolutionarily conserved genes. In Lymnaea, the CDCs themselves express two, or possibly three, such genes (Vreugdenhil et al., 1988; Hermann et al., 1997). All the genes code for a similar precursor (preprohormone) molecule. After proteolytic cleavage of the precursor, one hormone and several peptide sequences are produced, as illustrated in Figure 8.10. The secreted products of the bag-cells of Aplysia include an ELH of 36 residues and at least four peptides (Scheller et al., 1983). That the ELH causes ovulation was shown in vitro

202 Behavior and Neurons in Gastropods

Figure 8.10. Comparison between hormone precursor molecules of Lymnaea and Aplysia based on amino acid sequences deduced from cDNA nucleotide sequences. Black bars mark sites for proteolytic cleavage. Percentages indicate the degree of homology between corresponding regions. Although the overall homology between the sequences is low (29%), there is a significant homology (51%) between the caudodorsal cell hormone (CDCH) and the egg-laying hormone (ELH). Also, several of the predicted peptide sequences are highly conserved. Adapted from Vreugdenhil et al. (1988). Copyright 1988 by the Society for Neuroscience.

by observing that eggs are released from the ovotestis after application of ELH. It is not clear, however, whether the hormone acts by contracting the muscles that surround the egg-containing follicles, or by disconnecting the mature eggs from the epithelial lining of the follicles. Muscle contractions have not been directly observed. The amino acid sequence of ELH is highly conserved in all species of Aplysia in which it has been examined, showing homologies of 70% or more between any two species (Li et al., 1999). Furthermore, one of the two ovulation hormones secreted by the CDCs of Lymnaea stagnalis, CDCH-I, is very similar to ELH of Aplysia. Although the overall homology between the cDNA nucleotide sequences of the Aplysia ELH precursor and the CDCHI precursor is only 29%, the region of the Lymnaea precursor that codes for CDCH-I is 51 % homologous with the region of the Aplysia precursor that codes for ELH (Fig. 8.10). The ammo acid sequence identity between CDCH-I and ELH from Aplysia californica is 39%; it is 44% between CDCH-I and ELH from Aplysia vaccaria (Vreugdenhil et al., 1988; Li et al., 1999). Another egg-laying hormone has been isolated from the DNA of the abalone Haliotis rubra, a "prosobranch." Remarkably, the predicted amino acid sequence of the abalone hormone is 97% homologous with CDCH-I and 47% homologous with ELH of Aplysia californica (Li et al., 1999). Thus, the gene that codes for the ELH is highly conserved in all three subclasses of Gastropoda. Regrettably, there have been few reported tests of the functional equivalence of the ELHs across species (see above). It is especially curious that there have been no reports of egg laying induced in Lymnaea by injections of Aplysia ELH, or egg laying induced in Aplysia by injections of CDCH. I myself have tried without success to induce egg laying in two species of pulmonate snails, Helix aspersa and Cepaea nemoralis, by injecting

Reproduction

203

bag-cell extracts from Aplysia. The general absence of positive results from cross-species injection experiments suggests that, even though the ELHs may be conserved, the receptors for these hormones may not be. The genes that encode the ELHs are expressed not only in bag-cells and CDCs, but also at other central and peripheral sites. Small clusters of cells expressing the genes are present within the cerebral and pleural ganglia. As discussed below, these cells may play a role in initiating discharges in the bag-cells and CDCs. Moreover, two of the genes that code for ovulation hormones in Lymnaea, CDCH-I and CDCH-II, are widely expressed in peripheral tissues (Van Minnen et al., 1989). Neurons expressing CDCH-I are found in all parts of the female reproductive tract except the albumen gland. These cells are probably involved in transporting and processing eggs. Also, both CDCH-I and CDCH-II are expressed in epithelial exocrine cells in the male system. Since the secreted products of these genes enter the seminal fluid before it is transferred to the mating partner during copulation, they could play a role in postcopulatory sperm competition. In Aplysia, the atrial gland is a major site of expression for two genes whose sequences are closely related to that of the ELH gene. The atrial gland is an exocrine organ that secretes into the ovoduct. Although both atrial gland genes yield a precursor molecule that is very similar to the ELH precursor, ELH itself is not a product of either gene. Instead, the peptides A and B are produced. Interestingly, the A and B peptides are potent releasers of egg laying behavior when injected into intact animals, despite the fact the atrial gland secretes into the ovoduct so its products should not ordinarily have access to the CNS. The peptides are effective in causing egg laying because they evoke discharges in the bag-cells, but they do not directly excite the bag-cells. Instead, they seem to excite neurons in the cerebral and/or pleural ganglia, which in turn excite the bag-cells (Fig. 8.8; see below). Some species (e.g., Aplysia parvula) do not have an atrial gland, nor do they have genes that encode the A and B peptides. Several additional cleavage products of the egg-laying genes may be involved in the control of egg laying behaviors, but the number of signal peptides that is actually released from the neuroendocrine cells, and the functional roles of each, are not precisely known. There are as many as ten peptide messengers encoded on the ELH gene of Aplysia, and as many as nine encoded on the CDCH-1 gene of Lymnaea; some of these are indicated in Figure 8.10. In Aplysia, ELH and at least four additional cleavage products are released from the bag-cells. One of these is an acidic peptide (AP), which has no known function. Three closely related peptides are also released, of which two are pentapeptides (/3-BCP and y-BCP) and one is a peptide comprised of nine amino acids (a-bag-cell peptide, or a-BCP). The coding regions for a-BCP and /2-BCP are highly conserved across species, suggesting that these peptides may be especially important for reproduction.

204 Behavior and Neurons in Gastropods

8.5.3.

Triggers for Sustained Electrical Discharges

Since the all-or-none discharges of the neuroendocrine cells are both necessary and sufficient for egg laying, control over the electrical activity of these cells is obviously crucial, as indicated in Figure 8.8. The following discussion focuses on four aspects of regulation in these cells: what triggers the electrical discharges; how they are maintained; how they are terminated; and what processes account for the refractory period. Perhaps the most difficult problem is to identify the stimuli that trigger electrical discharges in neuroendocrine cells or, more generally, what stimuli trigger egg laying. Presumably, egg laying requires certain conditions, both internally within the animal, and externally in the animal's environment. It is obvious, for example, that the animal must be sexually mature and, in many species, the season must be summer. One idea is that the proximal trigger for egg laying is the presence of mature oocytes in the ovotestis. However, Ferguson et al. (1986) recorded spontaneous bag-cell discharges in Aplysia, which had previously had their ovotestes removed. As for the dependency on season, this fact alone does not identify the critical environmental stimuli because photoperiod, temperature, the availability of food, and even maturity itself all vary with season. In fact, all these variables have been shown to separately influence egg laying in one or another species. A particularly effective stimulus for eliciting egg laying in Lymnaea is to change the water in which the snails are living from "dirty" to "clean," or simply to increase the oxygen content. Although this procedure causes 95% or more of the individuals in a population to oviposit, the exact nature of the stimulus has not been determined. Terrestrial snails are particularly finicky about when, or where, they lay their eggs. While food, lighting, temperature, moisture, calcium, and suitable substrate are all thought to be relevant, no combination of parameters yields consistently high levels of oviposition in the laboratory. The conclusions that can be drawn from the foregoing discussion are that multiple stimuli influence the readiness to lay eggs, that the stimuli differ between species, and that more work is needed to identify them. It was mentioned in Section 8.3 that some signal from egg-laying Aplysia appears to induce mating in nearby conspecifics. Some authors have suggested that the converse may also be true (i.e., that mating might trigger egg laying). This could happen if, for example, the penis stimulated the female's atrial gland to release an egg laying pheromone into the sea water. However, Blankenship et al. (1983) showed that this probably does not occur. They observed a laboratory population of Aplysia for 41 days and recorded all egg laying events and all occasions when an animal mated as a female. They found that the number of egg laying events (84) was only about 13% as great as the number of times that any animal was seen mating as a female (640). Thus, mating does not appear to trigger egg laying in nearby animals. However, egg laying may trigger egg laying in other animals. This idea is

Reproduction

205

supported by the observation, originally by Kupfermann and Carew (1974), that large egg masses comprising numerous egg cordons are often laid together at individual sites in the field. Since each cordon has eggs of slightly different colors, they are probably contributed by different individuals. The occurrence of these composite egg masses suggests the possibility that egg cordons contain a spawning pheromone, as proposed by Audesirk (1977), but no molecule with such properties has yet been identified. Temperature is one variable that strongly affects the frequency of egg laying in Aplysia. Pinsker and Parsons (1985) kept A. californica and A. brasiliana at 15°C for 16 days, then raised the temperature to 20 °C for another 16 days. For A. californica, which is normally subject to relatively large seasonal changes in water temperature, the number of egg laying episodes increased from 8 to 57 as the temperature increased from 15°C to 20 °C. For A. brasiliana, which is less seasonal in its reproductive habits, the number of episodes increased somewhat less, from 71 at 15 °C to 125 at 20 °C. Several attempts have been made to identify the site at which temperature acts to affect egg laying. While higher temperatures promote the maturation of oocytes, changes in the ovotestis are not likely to be relevant because, as noted above, the bag-cells will discharge even after the ovotestis has been surgically removed (Ferguson et al., 1986). Also, while ELH synthesis in the bag-cells is temperature dependent, neither the duration of bag-cell discharges nor the amount of ELH secreted during a discharge is affected in a consistent manner by temperature (Wayne et al., 1996). One possibility is that temperature affects egg laying by influencing critical trigger neurons. Ferguson and colleagues found evidence for the presence of egg laying command neurons in the cerebral ganglion (Ferguson et al., 1989b). In one experiment, they used focal electrical stimulation in various regions of the CNS to search for sites from which bag-cell discharges could be evoked. They found that stimulation was especially effective when delivered in the region of the F cluster of the cerebral ganglion (see Fig. 6.7A). Also, when they investigated the effects of nerve lesions on spontaneous egg laying, they found that lesions of the cerebropleural connective nerves were the most potent, causing a complete elimination of egg laying. These results suggest a chain of command from the cerebral ganglion to the pleural ganglion to the bag-cells. The hypothesis supposes that the relevant sensory cues for egg laying first converge on to F cluster neurons in the cerebral ganglion, further integration then occurs in the pleural ganglion, and finally a command is sent to the bag-cells to cause them to discharge. Further support for the idea of a cerebral-pleural-abdominal line of command comes from the discovery that a small number of cells in the cerebral and pleural ganglia are immunoreactive to the bag-cell peptide a-BCP. Most of these cells are clustered in the pleural ganglion near the origin of the cerebropleural connective. An intriguing aspect of their distribution is that the right pleural ganglion is consistently found to have more immunoreactive cells than the left pleural ganglion (Brown et al., 1989). Electrical recordings from the a-BCP immunopositive pleural cells shows

206 Behavior and Neurons in Gastropods

them to have properties similar to the bag-cells, that is, they are electrically coupled and they exhibit prolonged, synchronous discharges. Especially compelling is the finding that intracellular electrical stimulation of just one a-BCP immunopositive pleural cell can elicit a burst discharge in the bag-cells (Brown et al., 1989). One obvious explanation of this action would be if axons of the pleural cells project down the pleurovisceral connective and synapse on to bag-cell somata or bag-cell neurites close to the abdominal ganglion. Another possibility is presented by the demonstration that some of the bag-cell neurites extend sufficiently far up the pleurovisceral connective to reach the pleural ganglion and even the cerebral ganglion (Shope et al., 1991). This latter finding suggests that the pleural cells might excite the bag-cells through contacts in the pleural ganglion. Wayne et al. (1996) were interested in finding the cells that mediate the temperature dependency effect. They did so by testing the efficacy of atrial gland extracts in evoking egg laying at different temperatures. They found that the extracts evoked significantly more bag-cell discharges at 20 °C than at 15°C, and that the latencies to discharge were significantly shorter at the high temperature than at the low temperature. Since previous experiments had shown that the atrial gland peptides act at sites in the cerebral and pleural ganglia, these results again implicate the cerebral and pleural ganglia in the initiation of egg laying. However, the idea that sensory information must first reach the cerebral ganglion before a command for egg laying can be issued is difficult to reconcile with the fact that spontaneous egg laying can still be observed after all peripheral nerves of the cerebral ganglion have been severed (Ferguson et al., 1989b). Obviously temperature is one environmental signal that does not require nervous conduction to reach central sites, but it would be surprising if this were the only natural cue for egg laying. More work is needed, not only to identify the sensory stimuli responsible for triggering egg laying in Aplysia; but also to clearly establish the critical neural pathway. 8.5.4.

Changes in Excitability during Prolonged Discharges

The prolonged discharges of the neuroendocrine cells are associated with some striking changes in ion channel function (see Conn and Kaczmarek, 1989). Very early in a discharge, the action potentials increase in amplitude and duration due to frequency and voltage-dependent effects acting on the sodium current. Later in the discharge, potassium currents and calcium currents are affected by other mechanisms. The delayed rectifier K+ current is inhibited in bag-cells by cAMP, probably acting through protein kinase A. Concurrently, protein kinase C mediates an increase in Ca++ currents; this latter effect involves the recruitment of previously covert calcium channels (Strong et al., 1987). The combined result of all these effects is a three-fold increase in the amplitude and duration of action potentials within 2 minutes of the start of a discharge, relative to values measured before the discharge.

Reproduction

207

Presumably, the larger action potentials maximize the secretion of the messenger molecules that are required for egg laying. Quite separate from the changes in the form of individual action potentials are the changes in excitability that underlie and sustain the discharge. Two mechanisms, both attributed to an increase in cAMP concentration, have been identified in the bag-cells. One mechanism involves inhibition of the IA current, which is a potassium current activated at very low membrane potentials. Ordinarily, cells that possess the IA current have low excitability because even small depolarizations rapidly activate the IA current and hold in check any further depolarization. Therefore, the inhibition of the IA current in bag-cells allows them to fire at a faster frequency than would otherwise be possible. A second mechanism that seems to be important for sustaining discharges is the activation of a slow, non-specific, cation channel that depolarizes the cells (Wilson et al., 1996). The whole-cell current generated by this channel has so far been studied only in isolated bag-cells, and it has not yet been examined in cells that are triggered to discharge by electrical stimulation. A toxin from the Conus snail is used to activate the whole-cell current. Single-channel recordings suggest that the channel can be switched from a relatively inactive mode dominated by closed states to a "high-activity" mode dominated by open states, and that regulation is mediated by a tyrosine phosphatase (Wilson and Kaczmarek, 1993). Because the slow inward current is activated even in membrane patches that are placed inside toxin-stimulated cells, so the membranes are not directly exposed to the toxin, normal activation of the channel must depend on intracellular messengers rather than the direct binding of a ligand. A further interesting feature of this channel is that it is sensitive to intracellular concentrations of calcium, which may help to explain how the bag-cell discharges are terminated (see below). Once the neuroendocrine cells have begun to fire, two types of feedback help to intensify, spread, and sustain the discharge. One route for feedback is through the electrotonic junctions that are known to be prevalent within the populations of bag-cells and CDC cells, as already noted. A second route is through chemical autoexcitation. Many studies, to be reviewed below, have demonstrated that the peptide secretions of the neuroendocrine cells can produce electrical effects in ordinary central neurons. The same peptides influence the neuroendocrine cells themselves. The influence is most easily demonstrated in experiments where protease inhibitors are added to the bath solution, because otherwise the peptides are quickly broken down. In Aplysia, a chemically mediated feedback excitation was demonstrated using an experimental set-up in which the secreted products from a source ganglion are transferred to an assay ganglion by serial perfusion (Brown and Mayeri, 1989). If the bag-cells in the source ganglion are electrically stimulated to cause a discharge, the bag-cells in the assay ganglion will also begin to discharge after a short delay due to the perfusion. The same system was used for tests of the a-, /?-, and y-BCPs, each of which was delivered separately via the perfusion. Three of these peptides were found to depolarize the bag-cells.

208 Behavior and Neurons in Gastropods

Most impressively, arterial perfusion of a-BCP at concentrations equal to or below those that would be expected to occur during a bag-cell discharge produced bag-cell depolarizations of 2-16 mV in 28 of 32 preparations (Brown and Mayeri, 1989). Contradicting the results from the experiments just described, which indicate that the a-BCP depolarizes bag-cells, are reports from other workers that a-BCP hyperpolarizes and inhibits bag-cells (see Redman and Berry, 1991). To explain the contradiction, Redman and Berry (1991) conducted tests at different temperatures and discovered that the effects of a-BCP on bag-cells are temperature dependent. At 15°C, a-BCP reduces the levels of cAMP in bag-cells and causes hyperpolarization, whereas at 20 °C, a-BCP increases cAMP and causes depolarization. Since egg laying normally occurs at relatively high temperatures, a-BCP is likely to have a positive feedback effect in vivo. To summarize, several mechanisms have been identified that can induce and sustain an electrical discharge, namely inhibition of the IA potassium current, activation of a slow inward current, electrotonic coupling and chemical autoexcitation. Because these mechanisms appear to be redundant, they seem to ensure that any suprathreshold stimulus will produce a discharge that is fully capable of causing egg laying. 8.5.5.

Discharge Termination and Refractoriness

After discharging steadily for about 30 minutes, the neuroendocrine cells stop firing and enter a refractory state, during which time they cannot again be stimulated to produce a long lasting discharge. Refractoriness lasts for about 6 hours in Lymnaea and about 18-20 hours in Aplysia. Presumably, this time is sufficient to allow the regeneration of mature oocytes. Redundant mechanisms may again be at play, although somewhat different factors may account for termination of the discharge and for refractoriness. Termination could be caused either by depletion of autoexcitatory peptides or by desensitization of the cells' responses to them (Brown and Mayeri, 1989). Biochemical studies suggest that the duration of a discharge may be limited either by the cells' ability to maintain high levels of cAMP, or by their ability to maintain a physiological connection between cAMP levels and electrophysiological responses (see Conn and Kaczmarek, 1989). Increasing concentrations of intracellular calcium may also be important. Calcium enters discharging cells through the non-specific cation channel and through voltage-dependent Ca++ channels. Moreover, calcium is released from intracellular stores. Magoski et al. (2000) investigated the role played by calcium in inducing refractoriness in the non-specific channel. They found that the channel becomes refractory for a period of about 24 hours after a single dose of Conus toxin. After testing separately the three different sources of calcium, they concluded that it is the calcium entering through the non-specific channel that is critical in causing refractoriness, possibly because

Reproduction

209

entering calcium uncouples the channel's receptor from its downstream effectors. Whether these findings are relevant to the control of egg laying behavior depends upon whether the non-specific cation channel is actually involved in generating normal bag-cell discharges, something which has not yet been demonstrated (see above).

8.6.

Mechanisms for the Control of Egg-Laying Behaviors

Egg laying involves more than simply the extrusion of an egg mass because, as described in Section 8.5, the concomitant overt behaviors include changes in locomotion, buccal rasping, and head movements. There are also covert "behaviors," such as ovulation, the inhibition of feeding, and cardiovascular adjustments (Ligman and Brownell, 1985). Early commentators (e.g., Scheller et al., 1983) introduced the idea that each aspect of egg laying behavior is separately controlled by a different molecule from among the many that are secreted by the bag-cells and the CDCs. However, Bernheim and Mayeri (1995) reported that all, or at least most, of the egg laying behaviors of Aplysia can be induced by injections of ELH alone, that is, by acting downstream from the bag-cells and thus presumably without the assistance of any bag-cell peptides. Furthermore, other experiments, to be described below, have shown that some control of egg laying behaviors is mediated by purely neural signals operating through reflex pathways. The evidence to be reviewed in this section suggests that hormones, peptides, and neural reflexes work together to orchestrate the full complement of behaviors associated with egg laying (see Fig. 8.8). Each mechanism of control is appropriate to specific needs. The ovulation hormones, are released into the general blood circulation and they are slowly inactivated, so they can affect multiple targets and distant targets. The peptides are more quickly broken down, and their actions are more limited. Given the absence of morphological and physiological evidence for any conventional synapses between neuroendocrine cells and follower neurons, it seems likely that the peptides are released at parasynaptic (non-synaptic) sites from which they enter the interstitial space and diffuse for tens or hundreds of micrometers before becoming inactivated by proteases. If there are neuronal membranes within the interstitial space with appropriate receptors, then transmission is possible between neuroendocrine cells and ordinary neurons. Transmission in this manner probably occurs both in the cerebral commissure of Lymnaea (see Section 8.5.1), and in the pleurovisceral connective of Aplysia. Moreover, if the bag-cell neurites extend into the cerebral and pleural ganglia (see Section 8.5.3), then peptides are probably released parasynaptically at these locations as well, and they have the potential to influence circuits that control behavior. Finally, pathways that rely on conventional synaptic transmission, as in the reflex circuits to be described below, provide the most specific form of control over behavioral expression.

210

Behavior and Neurons in Gastropods

Table 8.1. Neurons in Aplysia modulated by egg-laying hormone (ELH) or bag cell peptides (BCPs) Neuron

Effect

Molecule

Putative Function

B16 L10

Depolarization; excitation Transient excitation; prolonged inhibition Transient excitation Long-lasting inhibition Inhibition Inhibition Prolonged excitation Inhibition Inhibition Inhibition Burst augmentation Depolarization; excitation

ELH a-BCP

Radula closer motoneuron Increases cardiac output

/3-BCP a-BCP + ? a-BCP a-BCP ELH a-BCP a-BCP a-BCP ELH BCPs

Mechanoreceptors Kidney functions Ink gland motoneurons Visceral motoneurons

LI, Rl L3, L6 L14A, B, C LB, LC clusters R2 R3-14 R15

Bag cells

Mucus secretion Regulation of vascular muscle Aspects of egg laying Egg laying (autoexcitation)

All listed neurons, except the buccal neuron B16, are in the abdominal ganglion. Data from Sigvardt et al. (1986), Sossin et al. (1987), Brown and Mayeri (1989).

It remains to be seen exactly how these various forms of cellular communication combine in the egg laying systems. One possibility is that hormonal and/or peptidergic modulation readies appropriate neural circuits to respond in a reflex manner. Alternatively, some behaviors might be controlled entirely by hormones and/or peptides, while others are controlled reflexively. Ter Maat and Ferguson (1996) have suggested that appetitive behaviors are largely mediated by the former mechanism and consummatory behaviors are largely mediated by the latter mechanism. Mayeri and colleagues have identified several neurons whose electrical behavior is modulated either by ELH or by one or more of the bag-cell peptides. The usual assay involves delivery of the test substance(s) by arterial perfusion using concentration of 0.1-1.0 JIM, and with protease inhibitors added to potentiate the effects. Table 8.1 summarizes the data. Note that all but one of the listed cells are in the abdominal ganglion. In some cases, the effects have been shown to persist after surgical isolation of individual cells, or after blockage of fast synaptic transmission, thus indicating that they are direct, but this has not been rigorously tested in all cells. The specificity of the nerve cell's responses to peptides has been convincingly demonstrated by Sigvardt et al. (1986). They showed that many cells will respond to a-BCP when it is delivered alone, but the same cells will not respond to a mixture containing ELH and four BCPs other than a-BCP. The responsiveness of neuron R15 to both ELH and a-BCP is particularly interesting. Alevizos et al. (1991b) present a case for this cell having an important role in integrating aspects of egg laying behavior. As mentioned in Sections 5.4 and 5.5, R15 seems to be involved in regulating water balance and activating respiratory pumping. Alevizos and colleagues have

Reproduction

211

shown that R15 also activates anterograde peristaltic movements in the large hermaphroditic duct. This last finding is pertinent to egg laying because fertilized eggs must pass through the large hermaphroditic duct on their way to the genital pore. All these actions of R15 are apparently mediated by its peptide transmitter, R15cd. Together, the data suggested to Alevizos et al. (1991b) that R15 might be responsible for coordinating the many visceral and behavioral requirements of egg laying. They discovered that R15 is especially sensitive to ELH when it is electrically silent, as opposed to when it is bursting, and they found from recordings obtained in vivo that R15 is indeed normally silent (Alevizos et al., 199la). However, the critical test of their hypothesis is whether R15 becomes active during episodes of spontaneous egg laying, or when egg laying is experimentally induced, and so far no results from such tests have been reported. The evidence summarized in Table 8.1 supports the conclusion that at least some of the secretory products of the bag-cells (e.g., ELH and a-BCP) can influence neuronal activity, but they do not allow the assignment of specific molecules to specific egg laying behaviors. With one exception, all the cells known to be affected by bag-cell products are located in the abdominal ganglion, whereas lesion experiments show that the commands for egg laying behaviors (subsequent to a bag cell discharge) come from the cerebral ganglion, with motor outputs coming mostly or entirely from the pedal and pleural ganglia (Ferguson et al., 1989b). A somewhat better association between specific molecules and specific egg laying behaviors has been worked out in Lymnaea. Hermann et al. (1994,1997) studied the control of shell turning, which is a behavior that is uniquely expressed during the second phase of egg laying (see Section 8.5). After identifying a group of motoneurons in the right pedal ganglion (RPeN cluster) that are necessary for shell turning, experiments were performed to examine the effects of messenger molecules encoded on the CDCH-I gene. It was found that the peptide /33-CDCP excites the RPeN motoneurons, whereas the ovulation hormone CDCH-I inhibits the same cells. Injections of /33-CDCP into intact animals led to shell turning within 1-2 minutes, and this behavior continued at levels significantly above controls for 10-15 minutes. The frequency of other behaviors associated with the turning phase of egg laying also changed significantly following /63-CDCP injections; for example, rasping increased and locomotion decreased. Similar behavioral results were obtained with injections of a-CDCP. It is important to note that none of the animals injected with /J3-CDCP or a-CDCP laid eggs, at least in the 120 minutes following an injection. By contrast, injections of CDCH-I caused the full complement of egg laying behaviors, including oviposition. From the foregoing results, Hermann et al. (1997) proposed a scenario for the control of spontaneous egg laying behaviors in Lymnaea. Phase 1 behaviors, including ovulation, the packaging of ripe eggs and the inhibition of shell turning are coordinated by CDCH-I. Phase 2 behaviors, including shell turning and rasping, are initiated by a-CDCP and /33-CDCP. Crucial to the interpretation of Hermann et al. (1997) is the idea that the actions of

212 Behavior and Neurons in Gastropods

a-CDCP and /33-CDCP during phase 2 are due to their release from CNS neurons expressing the CDCH-I gene (Van Minnen et al., 1989), but not from the CDCs themselves. The idea is that these neurons release the CDC peptides only when triggered by an afferent signal caused by the passage of eggs through the female tract. A further account of this hypothesis, which is based on reflex circuitry, is presented below. In respect to the activities of the CDC peptides, however, it is curious that the authors discount the possibility that phase 2 behaviors could be caused by CDC peptides released from the CDC cells, since presumably they are released at the same time as the ovulation hormone, CDCH-I. If their scenario is correct, it can only be assumed that the peptides that are released directly from the CDCs are either insufficient in quantity, poorly timed, or too remote from their targets to initiate phase 2 behaviors. Some evidence suggests that internal sensory feedback plays a role in mediating egg laying behaviors. The general idea is that, while fertilized eggs are being transported through the female tract to their eventual release into the environment, they excite neurons along the way that send signals to the CNS, which then trigger egg laying behaviors. Observations consistent with this idea come from the work of Ter Maat and colleagues, who found that the total duration of turning and oviposition behaviors in Lymnaea is highly correlated with the number of eggs in the egg mass (see Hermann et al., 1997). They also noticed, in Aplysia, that weaving and tamping movements

Figure 8.11. The elimination of internal sensory feedback prevents the expression of certain egg-laying behaviors in Aplysia. (A) Interventions involved lesions of nerves as well as ligations of the genital groove to block the passage of eggs; these were performed either separately in different animals or together in the same animal. (B) Two types of head movements, tamps and weavings, were severely affected in animals with combined ligations (LIG) and lesions (LES). The data shown here indicate that both internal and external portions of the genital tract provide feedback necessary for triggering tamping. Two other types of head movements, performed prior to the external appearance of the egg cordon, were not affected by the same interventions (data not shown here). (C) Feeding latencies were affected only by ligations, indicating that feeding is inhibited as the eggs become externalized. Adapted from Ter Maat and Ferguson (1996).

Reproduction

213

of the head only begin once the extruded egg string has reached the substrate, and thereafter the movements continue at a nearly constant rate but only until the egg string is fully deposited (Ter Maat and Ferguson, 1996). Experimental evidence supporting the idea of reflex control of egg laying behaviors was first presented by Cobbs and Pinsker (1982) who ligated the small hermaphroditic duct near the ovotestis to prevent the packaging and movement of the egg string. They reported that, when the ligated animals were injected with bag cell extracts, egg laying behaviors were markedly reduced relative to controls. Especially noticeable was the absence of feeding inhibition. In follow-up work by Ferguson et al. (1986), the surgical intervention removed the ovotestis instead of just ligating the tract, and the behavioral analysis focused on head and neck movements instead of feeding. To see whether the head movements normally associated with egg laying behavior could be elicited in these animals, the bag cells were electrically stimulated through an implanted wire. Since the lesioned animals showed a significant decrease in head and neck movements relative to intact animals that were able to produce egg cordons, the result supports the hypothesis that internal sensory feedback triggers egg laying behaviors. Ter Maat and Ferguson (1996) attempted to identify the site or sites where feedback signals are generated. They cut nerves to prevent feedback through a specific pathway already implicated in triggering egg laying behaviors, and they ligated the external genital groove to prevent movement of the egg string to peripheral sites. These procedures, shown in Figure 8.11 A, were performed either separately in different animals or combined in the same animal. The animals were then induced to lay eggs by electrical stimulation of the bag cells, and the individual egg laying behaviors were observed. Since the interventions that included disruption of sensory signals from both the genital pore and the external genital groove prevented head weaving and head tamping, these areas are implicated in triggering weaving and tamping (Fig. 8.1 IB). By contrast, two other head movements, waving and undulation, which are performed before the egg cordon is externalized, were not affected by the same combined interventions, suggesting that they are triggered by bag cell products acting directly on central neurons. In the same study, the authors tested whether the disruption of peripheral feedback loops would prevent the inhibition of feeding that is associated with egg laying. The experimental measure was the latency to bite a piece of algae. Since they found that ligation of the genital groove was sufficient to block the inhibition of feeding (Fig. 8.11C), they concluded inhibition is triggered when the eggs contact the lips or mouth. In summary, the evidence suggests that hormonal, peptidergic, and neural signals all combine to generate egg laying behaviors. However, the details of their integration are far from clear, which leaves plenty of challenge for future research.

9

Defense 9.1.

Dangers from Predators and Conspecifics

Apart from their shells, which are even absent or insignificant in some species, gastropods provide easily digestible meat for predators. Moreover, humans partake in the feast. Gastropods are defenseless against humans, and human consumption is significant, especially in regard to snails. In France, for instance, snails both collected and farmed constitute an industry with sales of 50 million Euros ($46 million US) per year. In central Africa, the very large snails of the family Achatinidae contribute a considerable fraction of protein to the local diet. Some gastropod populations have been eliminated, or nearly so, by overzealous collection. Examples of species endangered by human consumption include, in Europe, the vineyard snail Helix pomatia and, in the Caribbean, the giant marine conch Strombus gigas. In Asia, humans eat Aplysia, probably other opisthobranchs, and a variety of marine snails. Mead's (1979) well-documented list of predators of terrestrial snails and slugs includes ants, beetles, birds, centipedes, millipedes, crabs, larval insects, toads, lizards, rats, mice, squirrels, shrews, turbellarian worms, scorpions, and other gastropod molluscs. The source of predation varies with location. In some "marginally tenable environments" of southern California, it has been estimated that three of every four individuals of the snail species Helix aspersa die from predation by small mammals. At other locations, where large numbers of empty snail shells are found with holes through the top of the shell, one can infer heavy predation by birds. This is because the birds peck at, and sever, the snail's columellar retractor muscle so the snail cannot retreat into its shell. Some birds prefer to crush the snail's shell on a stone anvil or drop it from a height on to a hard surface. Predator species have been intentionally introduced to certain locations in an attempt to control invasive snail pests, especially Achatina fulica. Mead (1979) describes many such experiments, including the introductions of thousands of glowworm beetles and carabid beetles to southern Pacific islands, and the release of 70,000 hermit crabs on the South Andaman islands. The most consequential of all these introductions were those involving predatory snails. Of the 11 snail species that were introduced to Hawaii for the purpose of controlling Achatina, the most efficient predators are Euglandina rosea and Gonaxis quadrHateralis. Unfortunately, it turned out 214

Defense

215

that the introduced species caused more harm than good because they ate not only Achatina fulica, but also the indigenous gastropods. Gastropods in aquatic environments also face a variety of predators. From Snyder (1967), the known predators of freshwater snails include leeches, insects in both adult and larval stages, crabs, crayfish, fish, amphibia, turtles, birds, and shrews. Additional threats are found in the sea, especially for opisthobranchs that lack a protective shell. Johnson and Willows (1999) list the predators of sea hares (Anaspidea) as including anemones, crabs, lobsters, sea stars, turtles, and gastropod molluscs. The neogastropods (e.g., Buccinum, Busy con, Conus) and certain opisthobranchs (e.g., Clione, Navanax, Pleurobranchaea, Conus) are especially important molluscan predators, not only of sea hares but of many other gastropods including shelled ones. Attacks from conspecifics represent a different kind of threat. Although agonistic interactions have only been reported in a few species, they can result in serious injuries or even deaths. Some terrestrial slugs (Limax, Arion) commonly pounce on conspecifics with open mouths causing visible skin damage. Because of the seasonal nature of these attacks and their frequent occurrence near shelters, Rollo and Wellington (1979) interpreted the aggressive behavior of slugs as assertions of territoriality. Regardless, the responses of the victims are similar whether attacked by another slug or by a predator, namely they secrete slime and they either crawl away or contract. The victim may also sideswipe the attacker with its tail. Agonistic encounters between gastropods may also involve cannibalism. Individuals of many species will eat conspecifics if the victim is sick, injured, or otherwise unable to escape. Zack (1974) studied agonistic behaviors in the nudibranch Hermissenda crassicornis. He found that 20% of all encounters between animals were agonistic, that is, they involved attacks in which an animal lunged at, and bit, a conspecific. Interestingly, when the animals were starved for 3 days, the proportion of agonistic encounters increased to an average of 37% on each day of deprivation, and then they fell to 15% 1 day after the animals were fed. These results suggest that for Hermissenda, at least, agonistic behavior is allied with hunting behavior.

9.2.

The Lines of Defense

Some important defenses have the function of preventing attacks. For the most part, these are passive defenses that do not involve overt motor behaviors. An exception is behavior that puts the animal in a place where it is either undetectable or unreachable. The homing behavior of gastropods, discussed in Section 10.2, is one such example. Another example of active prevention is burrowing, which is common among bivalve molluscs and some "prosobranchs." Aplysia also burrows into soft sand (Aspey and Blankenship, 1976). An example of a passive defense is the possession of a shell, or being large, since either attribute may be sufficient to dissuade a potential predator. Crypsis is also an important passive defense used by

216 Behavior and Neurons in Gastropods

gastropods, especially cryptic coloration. On the other hand, bright coloration, as in nudibranchs, can be aposomatic, warning predators of the unpalatability of their prospective prey. Luminescence, which is an unusual feature present in only three genera of nudibranchs and three genera of pulmonates, is thought by some authors to be a warning signal. Odor might also be an aposomatic signal, at least in the Anaspidea (Johnson and Willows, 1999). The first line of active defense is to escape before a potential predator can make contact. This requires an early detection system, and since vision is usually poor in gastropods, and audition is never a factor, olfactory cues are best suited for this purpose. Firm evidence for predator avoidance based on olfaction is scanty, but persuasive enough in a few cases. Some marine species have been reported to show chemically mediated escape reactions either when contacted by specific predators or when exposed to extracts of the same predators. Limpets (Acmaed) sometimes move rapidly away from approaching starfish (Pisaster), although Bullock (1953) found that the response was less than reliable. Also, the "prosobranch" snail Strombus gibberulus has been observed to respond with strong "kicking" movements when specimens of the predatory snail Conus marmoreus are placed nearby, or even when sea water conditioned by the presence of Conus snails is introduced. On the other hand, when Snyder (1967) performed experiments involving numerous species of freshwater snails exposed to water conditioned by known predators including fish, turtles, crayfish and water bugs, he found little evidence for flight responses based on olfactory cues. Observations in a few species provide strong support for the conclusion that gastropods can be alerted to imminent dangers by chemical signals released from conspecifics following attacks by predators. Snyder accumulated convincing evidence for such alarm reactions by crushing freshwater snails to obtain "juices" that were then filtered and delivered to test aquaria in serial dilutions. Nineteen of 30 species showed reactions to the juices, and there was a considerable degree of species specificity. The responses varied from burrowing into the substrate, common in "prosobranch" species, to crawling out of the water, common among pulmonates. It is possible that Aplysia brasiliana is similarly responsive to conspecific alarm signals since burrowing can be induced in swimming animals that are placed into an aquarium that already contains burrowed conspecifics (Aspey and Blankenship, 1976). The large opisthobranch predator Navanax inermis exhibits escape responses when a certain yellow secretion is released from conspecifics. It has recently been discovered that the secretions contain chemicals that are closely related to secondary metabolites found in the cephalaspidean prey species of Navanax, and that these metabolites are used as alarm pheromones and defensive allomones in those prey species (Marin et al., 1999). It is suggested, therefore, that Navanax recycles the prey's secreted signals for its own use to avoid predation by cannibalistic conspecifics. Once contacted by a predator, gastropods have three main defenses: they can withdraw, secrete chemicals, and locomote away. The different types of

Defense

217

defenses may be employed singly or in combination. In general, animals without shells tend to use multiple defenses, whereas those with shells rely on retreat alone. Aplysia has a coordinated response involving local withdrawal of body parts, ejection of secretions, and escape locomotion (Walters and Erickson, 1986). Terrestrial snails, on the other hand, always withdraw into their shells when threatened or attacked. I studied how snails, Achatina fulica, responded to an irritant chemical that was delivered under controlled conditions to one posterior tentacle (Chase, 1982). Weak concentrations of the chemical caused the snail to retract just the stimulated tentacle. As I increased the concentration the snail retracted progressively more body parts until eventually it withdrew its entire body in response to high concentrations. It was particularly interesting that, in contrast to what one would expect of Aplysia in comparable circumstances, the snails never responded to the noxious stimulus by turning away from it, even though they consistently turned towards attractive odors. Certain marine gastropods, even some with shells, respond to predators with dramatic escape behaviors. Bullock (1953) describes how limpets react to a starfish by rapidly raising their shells from the substrate "like a mushroom" and then scurrying away at a rate of about 20 mm/second. If the tube feet of the starfish have already gripped the limpet, it rocks violently back and forth in its mushroom stance to shake off the predator, and then it locomotes away. This escape behavior is seen only when the limpets are attacked by starfish; other forms of disturbance elicit a clamping down of the shell. The reaction to the starfish is mediated by chemical recognition, and some predator species (e.g., Pisaster ochraceus) trigger significantly greater reactions than do others (e.g., Patina miniatd). Also, only certain species of limpets appear to have acquired the ability to detect the chemical signatures of starfish predators. Even within the single genus Acmaea, three species tested by Bullock responded "well," whereas two other species responded "almost never." The escape behavior of the marine snail Nassa mutablis bears some resemblance to the limpets' escape response, and it is equally striking (Lemmnitz et al., 1989). Again, it is triggered by chemical signals, typically from starfish, but only when the echinoderms contact the dorsal posterior part of the snail's long foot. When contacted, the snail rotates its body so strongly to one side or the other that the foot becomes detached from the substrate. Lying then on its side, it begins a sequence of violent twisting movements involving the foot and the shell that cause the snail to "leap" away from the starfish. A single touch from a starfish can trigger as many as 12 "leaps," each lasting about 80 seconds. Other marine gastropods showing chemically mediated escape responses include strombid snails, turban snails (Bullock, 1953), and certain nudibranchs (Willows, 1971). The neural control of escape locomotion has already been reviewed in Chapter 6. Therefore, the remainder of the present chapter deals mostly with withdrawal behaviors and chemical secretions.

218 Behavior and Neurons in Gastropods 9.3.

Withdrawal Reflexes

In general, withdrawals are graded according to the intensity of the stimulus. As already noted, weak stimuli elicit only local withdrawals of the affected appendage or body wall. As stimulus strength is increased, progressively larger areas of the body are withdrawn. Very strong stimuli elicit whole-body withdrawals. However, in the case of visual stimuli, even small shadows can trigger complete whole-body reactions, particularly in aquatic snails. A shadowy presence is apparently too threatening and too uncertain to allow anything but a full retreat. As explained in Section 4.3, it is useful to distinguish local responses, in which only the stimulated structure is withdrawn, from remote responses, in which structures at some distance from the stimulated structure are withdrawn (Perlman, 1979). Peripheral neural circuits play a major role in mediating local withdrawal reflexes. For example, in the case of the snail's tentacle withdrawal reflex, the PNS is responsible for as much as 75% of the response strength when weak stimuli are used (Prescott and Chase, 1996). However, stronger stimuli recruit a greater contribution by the CNS, up to 55% for the tentacle withdrawal reflex, which increases both the size and the duration of the responses. Recruitment of the CNS is also essential for generalizing responses to areas not directly affected by the stimulus. Whole-body withdrawals can require the participation of neurons from all the central ganglia (see below). Stimulus intensity also affects plastic properties of the response, including manifestations of learning and memory. The repetition of weak stimuli results in habituation of the response, whereas the repetition of strong stimuli or even the delivery of a single strong stimulus, can sensitize responses (see Section 9.4.1). Also, as we shall see, sometimes different groups of receptors are recruited to mediate withdrawal responses to weak and strong stimuli. 9.3.1.

Whole-Body Withdrawal

Generalized withdrawal responses are fast events that are triggered at very short latencies. The latency for whole-body withdrawal in Lymnaea stagnalis is 0.2 or 0.5 seconds, depending on whether it is elicited by a tactile stimulus or a photic stimulus (Ferguson and Benjamin, 199la). In Helix lucorum, the latency is 0.6 seconds (Zakharov, 1994). Whole-body withdrawal is mediated primarily by the columellar muscle, the large muscle that inserts on the column of the shell. The organization of the columellar muscle system in Helix is illustrated in Figure 4.1 (described in Zakharov, 1994). Several prominent bands branch off from the main muscle and attach to key structures at the anterior end of the animal. Thus, the tentacle retractors, the penis retractor, the buccal mass retractor, and the pedal retractors (attaching to the sole of the foot) are all part of the columellar system. Each of the muscles is independently innervated, so that individual structures, for example, a tentacle, can be withdrawn separately. However, the whole

Defense

219

system can also be activated in a coordinated manner. The basal part of the columellar muscle (i.e., the part that is common to all branches) is innervated by the left and right pleural nerves. When the entire columellar muscle contracts, the distance between the animal's snout and the shell aperture is shortened. Thus, if the animal's foot is well attached to the substrate, and the shell is light, it appears as if the shell is being brought down on the animal (Lymnaea)', conversely, if the foot is not well attached, and or shell is heavy, the animal appears to be drawn into the shell (Helix). The columellar muscle of Lymnaea is similar to that of Helix except that it is less obviously branched (Ferguson and Benjamin, 199la). Two lateral bands extend forward from a common point of insertion in the shell until they join and interdigitate anterior to the buccal mass. Numerous small fibers branch off from these bands and attach diffusely in the foot, while one larger branch attaches exclusively along the posterior midline of the foot. Innervation of the columellar muscle in Lymnaea is by two nerves, called columellar nerves, that originate in the pedal ganglion. Another muscle, the dorsal longitudinal muscle, works in conjunction with the columellar muscle in Lymnaea to shorten the body (Ferguson and Benjamin, 1991a,b). This muscle comprises three bands on either side of the midline, which together cover the entire dorsal surface of the animal. It is innervated by two pairs of cervical nerves, from the pedal ganglion, and one pair of nerves from the parietal ganglion. Even though few nerves innervate the muscles involved in whole-body withdrawal, the motoneurons that contribute fibers to those nerves are widely scattered in the CNS. This situation has been described in detail by Ferguson and Benjamin (199la). They found, in Lymnaea, a total of four motoneurons innervating the columellar muscle and 19 or more motoneurons innervating the dorsal longitudinal muscle. The motoneurons are distributed in seven different ganglia, but there is a concentration in the posterior portion of the A-cluster of the cerebral ganglion. Each motoneuron innervates a limited region of muscle, typically one band, and each region of muscle receives innervation from motoneurons in several different ganglia. During whole-body withdrawal, a unity of action is achieved by electrotonic coupling within the network of motoneurons, including coupling between neurons in different ganglia and coupling between neurons innervating the two major muscle systems. Figure 9.1 illustrates how contractions of the columellar muscle and the dorsal longitudinal muscle are coordinated at the level of motoneurons. Nerve backfills in other species of pulmonate snails yield a similar picture of motoneurons that are widely distributed in all, or most, of the central ganglia (Helix, Zakharov, 1994; Planorbis, Arshavsky et al., 1994). It is surprising that motoneurons innervating the same muscle and serving the same function should be scattered in this manner, and the pattern is not readily explained. By contrast, the "great majority" of neurons innervating the columellar muscle of the "prosobranch" snail Nassa mutabilis reside in the left pleural ganglion (Lemmnitz et al., 1989). Whether this is an anomaly

220 Behavior and Neurons in Gastropods

reflecting the peculiar escape behavior of this animal, or whether it is representative of "prosobranchs" in general, remains to be seen. Despite the scattered distribution of motoneurons in Lymnaea, the wholebody withdrawal response is executed quickly and smoothly. Coordination is achieved in part by electrotonic coupling between motoneurons, as already described, and in part by patterns of divergent sensory inputs. Ferguson and Benjamin (1991b) found that tactile stimulation of the skin in any small area of the body evoked synaptic depolarizations simultaneously in many motoneurons, including those in different ganglia, although ipsilateral EPSPs were larger than contralateral EPSPs (Fig. 9.1). Conversely, sensory signals generated by tactile stimulation on all parts of the body converge on to individual motoneurons. Photic stimulation (light of!) is also effective in exciting motoneurons. Interestingly, the shadow reflex involves neither the

Figure 9.1. Synchronous activation of motoneurons and muscles by tactile stimuli that elicit whole-body withdrawal in Lymnaea stagnalis. Coordination of the behavior is due in part to extensive electrical coupling within the network of motoneurons. Even cells innervating different muscles and located in different ganglia are electrically coupled. CM, columellar muscle; DLM, dorsal longitudinal muscle. From Ferguson and Benjamin (1991b). With permission of Company of Biologists Ltd.

Defense

221

eyes nor the optic nerve. Rather, the response is mediated by photoreceptors distributed in the epidermis, as first suggested by Stoll (1976). Although no interneurons were described by Ferguson and Benjamin (1991a,b), their involvement in whole-body withdrawal is inferred from the fact that the authors could not find any direct connections between sensory neurons and motoneurons. Therefore, the pattern of connection whereby sensory inputs first diverge from the source of stimulation and then converge on to individual motoneurons is likely mediated by interneurons. One candidate higher order interneuron in Lymnaea is the pedal cell PeD 11. This cell excites a large number of foot and body wall motoneurons, including some in the cerebral A-cluster, and its firing is associated with whole-body withdrawal behavior (Syed and Winlow, 1991b). However, stimulation of PeD 11 cannot initiate the behavior. In Helix, Balaban (1979) described nine interneurons with apparent "command" functions for withdrawal behavior. The group includes several giant cells on the dorsal anteromedial surface of the parietal ganglia, which are excited by noxious stimuli in all modalities and which fire prior to withdrawal reactions. Intracellular stimulation of some cells in the group causes muscular responses that are limited to the tentacles or the pneumostome, but stimulation of other cells causes a widespread contraction of the columellar system including contractions of the foot and the mantle. From experiments with the pond snail, Planorbis corneus, Arshavsky et al. (1994) concluded that interneurons do not just coordinate the withdrawal response, they also constitute a central pattern generator. In support of this idea, they cite the fact that patterns of motoneuronal firing and muscle contraction characteristic of the whole-body withdrawal response were sometimes recorded spontaneously in reduced preparations that lacked normal sensory inputs. The spontaneous patterns are typically biphasic, as in the reactions of the intact snail, with body-shell retraction in the first phase followed by detachment from the substrate and expulsion of pulmonary air in the second phase. The CPG is thought to consist of two components, representative of the two phases of the behavioral response. One identified pleural neuron, DRN1, seems to drive the first phase since its activity is highly correlated with phase 1 motor activities and intracellular stimulation of the cell can elicit those activities. Strong stimulation of DRN1 triggers phase 2 behaviors, suggesting a link between the two parts of the CPG. Still a different view of coordination is presented by a study of whole-body withdrawal in the pteropod Clione (Norekian and Satterlie, 1996). It is surprising to find any withdrawal response at all in this animal because it has no shell and its main defensive strategy is to swim rapidly away from danger (see Section 6.3). However, when the threatening stimuli are perceived as very strong, or when they cannot easily be localized, the animal's response is to cease swimming altogether, contract its body (withdraw) and sink slowly in the water. Because there is no shell, and no columellar muscle, the response is dependent on muscles in its body wall. Four neurons in each pleural ganglion, named Pl-W cells, seem to coordinate the response. The Pl-W cells have an unusual morphology in that the axons travel through all the central

222 Behavior and Neurons in Gastropods

ganglia except the buccal ganglia, and they project into all the major peripheral nerves that innervate the body wall. There is no evidence for either synaptic connections or electrical coupling between any Pl-W cells. The most striking feature of the Pl-W cells is that intracellular stimulation of any one of them in a semi-intact preparation is sufficient to cause simultaneous retractions of the body wall, tail, wings, and head. Part of this capability of the Pl-W cells is attributed to their role in activating local withdrawal circuits, for example, one in the cerebral ganglion for withdrawing the buccal cones and one in the pedal ganglion for withdrawing the wings. Because dye injections showed that at least some of the Pl-W axons terminate in peripheral muscles, Norekian and Satterlie (1996) describe the cells as motoneurons. One could therefore say that Clione coordinates its wholebody withdrawal response through the parallel activation of a few widely acting motoneurons, but since the Pl-W cells have premotor connections in addition to the putative muscle connections, they appear to possess features common to both motoneurons and interneurons. 9.3.2.

Tentacle Withdrawal

The tentacles are protruding appendages that are at risk of being contacted first by predators. Since they are also important sensory structures for olfaction, photoreception, and mechanoreception (Chapter 3), it is clear why the tentacles are quickly withdrawn when disturbed. The response is particularly fast and vigorous in stylommatophoran pulmonates because the tentacles in these animals have their own retractor muscles. Also, stylommatophores have motoneurons that exclusively innervate the tentacle retractor muscle and the nearby tegumental muscle. By contrast, the motoneurons for tentacle withdrawal in Aplysia, which comprise a subpopulation in the cerebral B-cluster, have a broad motor field that includes the head, the foot, and the parapodia (Teyke and Kupfermann, 1989). The neural control of tentacle withdrawal has been studied in detail in Helix aspersa (Frescott and Chase, 1996; Prescott et al., 1997). As illustrated in Figure 9.2, both the PNS and the CNS contribute to the withdrawal reflex (see also Sections 4.3 and 4.4). The peripheral and central contributions add linearly when the stimuli are weak, but the relative contribution of the CNS increases with stimulus strength. Since the CNS alone innervates the proximal portion of the retractor muscle, which contracts more tonically than the distal muscle, recruitment of the CNS is accompanied by an increase in the duration of the response. The initial phase of muscle contraction is largely caused by firing in a group of about 22 small motoneurons whose somata lie in the metacerebrum. Thereafter, most of the central contribution to muscle contraction, approximately 85%, is delivered by a single giant neuron, C3 (Fig. 9.2). This cell responds to mechanical and chemical stimulation of the tentacle over a wide dynamic range and it responds maximally to strong stimuli, as is typical of nociceptive neurons. Because it has axon projections into multiple peripheral nerves, C3 is able to coordinate a distributed

Defense

223

Figure 9.2. The neural circuit mediating tentacle withdrawal in the snail Helix aspersa. The reflex is triggered by tactile stimulation at or near the tip of the tentacle. The afferent signal, and part of the efferent signal, is routed through the tentacle ganglion. The motoneurons comprise the identified neuron C3 and about 22 unidentified cells, all located in the cerebral ganglion. Note that these cells innervate multiple muscle sites through four different peripheral nerves. Peripheral nervous circuits (not shown) mediate 45-75% of the reflex, depending on stimulus strength. O1N, olfactory nerve; PtNe, PtNi, external and internal peritentacular nerves; TegM, tegmental muscle; TRM, tentacle retractor muscle; TRN, tentacle retractor nerve. From Prescott et al. (1997) with permission.

muscle response. The muscle contractions evoked by C3 are mediated by acetylcholine, but C3 also contains high levels of FMRFamide; indeed, C3 was the first molluscan neuron shown to contain FMRFamide. Although no role for FMRFamide has so far been discovered, the tentacle withdrawal reflex is affected by habituation and sensitization, and C3 appears to be at least partly responsible (Prescott and Chase, 1999), so it is reasonable to think that FMRFamide may play a modulatory role in mediating behavioral plasticities. Tentacle withdrawal movements can also be elicited in Helix by direct stimulation of certain giant neurons in the pleural ganglion, which have been characterized as "command" cells, but these cells evidently operate independently of C3 because no synaptic connections have been found between them and C3 (see Zakharov, 1994).

9.3.3.

CHI and Siphon Withdrawal

Nudibranchs need to be especially defensive of their gills because, as the name says, their gills are exposed. The gills take different forms in different families. In the dendrotonid nudibranchs (e.g., Tritonia), they appear as two rows of turfs on the dorsolateral edges of the body; in the dorids, the tufts form a ring

224 Behavior and Neurons in Gastropods

around the anus. In early experiments, Willows (1967) found that intracellular stimulation of a few single neurons in the pedal and pleural ganglia of Tritonia produced branchial tuft contractions. Later, Dickinson (1980) used similar methods to look for comparable neurons in several species representing nudibranchs, notaspids, and anaspids. She found that every species possesses between six and eight gill control neurons. In none of these experiments was there sufficient testing to verify that the identified cells were motoneurons. Nevertheless, Dickinson made the interesting observation that in every species of nudibranch and notaspid the gill control neurons are synaptically connected, whereas in Aplysia the comparable neurons are not connected but they receive common synaptic inputs. In Aplysia, the gills, the siphon, and the mantle shelf are all withdrawn when either the siphon or the mantle shelf is touched. This bit of behavior, often referred to simply as "the defensive withdrawal reflex," was one of the first to be subjected to a thorough neurobiological analysis at the cellular level, and it remains one of the most intensively investigated neural circuits in any animal. The diagram of Figure 9.3, thought to represent the neural control of the reflex, was a fixture in nearly every neurobiology textbook published in the last two decades of the 20th century. The main reason for this was that Eric Kandel (1976, 1979) used it as the platform upon which to build his analysis of mechanisms underlying learning and memory (see Sections 9.4.1 and 9.4.2). Despite an ever-accumulating knowledge of the central nervous components involved in the defensive withdrawal reflex of Aplysia, the goal of total explanation remains unattained. From the beginning, Kandel and his colleagues have asked, "Can a behavior be fully explained in terms of the sum of its neural parts?" (Kandel, 1976, p. 374; italics in original). But even after investigators analyzed a new simplified preparation in which 85% of the gill withdrawal reflex is attributed to a single motoneuron, the authors had to admit that "additional experiments.. .will be necessary" to achieve a full quantitative explanation (Cohen et al., 1997). It is useful to begin by considering separately the two main motor components of the combined reflex, gill withdrawal, and siphon withdrawal. As Figure 9.3 shows, these movements are served by different groups of motoneurons. Although the gill and the siphon usually contract together, surgical procedures can separate the gill and the siphon so that each component of the reflex can be studied individually. In early experiments (Kupfermann and Kandel, 1969), the stimulus-evoked gill withdrawal response was described as a simple defensive reflex (see also Kandel, 1979). Later, Leonard et al. (1989) re-examined the gill's response to stimulation of the siphon and concluded, controversially, that the response is "neither simple, defensive, nor a reflex" (p. 600). These authors said the behavior is not simple because it consists often different actions (e.g., general contractions, contractions of individual pinnules, flaring, rolling, lifting) that combine to yield four categories of coordinated gill movements (e.g., large unimodal gill movements, small multimodal gill movements). Further, they

Defense

225

Figure 9.3. An iconic representation of the circuit mediating the gill and siphon withdrawal reflex in Aplysia, first published by Kupermann et al. (1974) and shown here as it appeared in Kandel (1976). The box marked "sensory neuron" represents the LE cluster of about 24 mechanosensory cells. Tactile stimulation of the siphon excites the LE cells, which in turn activate the identified motoneurons through both direct and indirect synaptic connections. PS, peripheral siphon neurons. An updated version of the circuit for siphon-mediated siphon withdrawal is shown in Figure 9.5. For locations of the cells in the abdominal ganglion, see Figure 9.6.

said that because the gills are protected by the parapodia, and therefore not ordinarily exposed to danger, the gill movements are not really defensive. Finally, Leonard et al. argued that the response is not a reflex because it is too variable and because all variations of the stimulus-evoked response occur spontaneously as well. Neither the data nor the interpretations of Leonard et al. (1989) detract from the fact that the gill withdrawal response of Aplysia, as originally described by Kupfermann and Kandel (1969), is a reflex. Their experiments only confirmed earlier observations in regard to spontaneous contractions and variable components of the total response (Kupfermann et al., 1974). While it is thus true that the gill exhibits a variety of movements, the simple components of the reflex response can be studied in isolation under suitable experimental conditions. Furthermore, under some natural conditions, for example, when the oxygen content of the sea water is low, the gill is exposed and hence defensive withdrawal is very important.

226 Behavior and Neurons in Gastropods

Walters and Erickson (1986) interpret gill withdrawal in Aplysia in the context of a broader defensive response. Gill withdrawal is thus linked, for example, to the ejection of defensive secretions because gill contractions increase the capacity of the mantle cavity to fill with these secretions, while eliminating an obstruction that could interfere with their directed ejection. Also, by contracting the gill prior to secreting ink or opaline, the animal protects its gill from becoming fouled by the secretions. In a similar vein, several authors have pointed out that gill withdrawal normally occurs in Aplysia as a component of respiratory pumping (see Section 5.4), and that respiratory pumping, in turn, occurs during the ejection of defensive secretions from the ink and opaline glands (Dickinson, 1980; Walters and Erickson, 1986). It has therefore been suggested that both ink ejection and respiratory pumping are mediated by a single network of neurons, one that includes Interneuron II (R25/L25). Leonard et al. (1989) generally agree, saying that touches elicit respiratory pumping and gill withdrawal as part of a broader defensive response in which respiration and circulation are increased to prepare the animal for escape locomotion and other exertions. The siphon of Aplysia protrudes from the parapodial flaps, so it is more exposed to dangers than is the gill. Withdrawal of the siphon thus clearly serves a defensive function. Indeed, the siphon is very sensitive to touch, and its sensory neurons have smaller receptive fields than do the neurons innervating any other region of the mantle cavity (Dubuc and Castellucci, 1991). When the siphon is touched it shortens and constricts, thus producing a local withdrawal. However, the movements of the siphon that are triggered remotely by stimulation elsewhere on the body are quite different from those triggered locally, as illustrated in Figure 9.4. The responses of the siphon to remote stimulation seem to prepare the animal to direct its defensive secretions toward the site of contact (Walter and Erickson, 1986). For example, the response to stimulation of the tail is a leveling of the siphon involving little contraction of the siphon itself but strong contractions of the mantle shelf and the parapodia; this causes ink and opaline to be ejected posteriorly through the siphon. By contrast, the response to head stimulation is a severe contraction and vertical movement of the siphon, which effectively closes it shut and causes the defensive secretions to be directed anteriorly through the flared anterior portions of the parapodia (Fig. 9.4). Evidently the availability of chemical defense in Aplysia has resulted in the modification of certain remotely triggered withdrawal reflexes to serve a more specific function (see Section 9.5.2). Before we consider the central nervous elements that participate in the control of gill and siphon withdrawal, it is important to remember that the PNS also contributes to these behaviors. Depending on which stimulus parameters are used and what properties of the response one wishes to measure, the PNS contribution to gill withdrawal can be as much as 100% (Peretz et al., 1976) or as little as 5% (Carew et al., 1979). The PNS contribution to local siphon withdrawal has been estimated at 45% (Perlman, 1979; Antonov et al., 1999). Because the CNS is much more accessible than the PNS

Defense

227

Figure 9.4. The ejection of defensive secretions by Aplysia is directed toward the source of the threatening stimulus. The drawing at the top shows the locations of the ink gland and the opaline gland in the roof and the floor of the mantle cavity, respectively. Although ink is known to irritate would-be predators, the function of the opaline secretion is unknown. After the secretions have been released into the mantle cavity, contractions of the siphon, mantle, and parapodia influence whether they are ejected posteriorly (drawing at left) or anteriorly (drawing at right). Identified LFS motoneurons mediate the two types of siphon movements shown here. From Walters and Erickson (1986) with permission.

for experimental work, most studies have focused on the former. For studies of the siphon-elicited gill withdrawal reflex, involvement by the PNS can be minimized by surgical cuts that ensure that all communication between the siphon and the gill is necessarily through the CNS (Cohen et al., 1997). By 1974, when the now-classic circuit diagram of the "defensive withdrawal reflex" appeared (Fig. 9.3), all the central gill motoneurons had been identified, or at least no additional ones have been identified since then. Each central motoneuron mediates a different movement through the innervation of different muscles. Early hyperpolarization experiments ("reversible lesions") demonstrated that L7 and LDGi (left side, cluster D, gill, cell 1) each contribute about 35% of the total response, but the photocell that was used in these experiments measured mostly the contraction of pinnules. Later experiments used an isotonic transducer, which is less sensitive to gill contraction and more sensitive to movements of the efferent vein. Using this method, 85% of the response was shown to be mediated by LDG1 (Cohen et al., 1997). As noted elsewhere (see Section 5.3.2), L7 has connections with multiple effector organs. Because L7 also has an excitatory synaptic connection with at least one peripheral motoneuron, it is an

228 Behavior and Neurons in Gastropods

interneuron as well as a motoneuron (Kurokawa et al., 1998). Its overall function may be to coordinate the cardiovascular and respiratory aspects of defensive responses. A recent circuit diagram of the network controlling the siphon withdrawal reflex ofAplysia is shown in Figure 9.5, and the locations of the gill and siphon motoneurons are shown in Figure 9.6. Note that there is an unknown number of peripheral motoneurons. Note also that the LFS motoneurons included in Figure 9.5 are absent from the earlier version of the circuit, shown in Figure 9.3. The LFS motoneurons are present as a single cluster but they comprise two functionally distinct groups. The four LFSe cells are collectively responsible for almost the entire tail-evoked siphon response, and they contribute strongly to the head-evoked siphon response (Hickie and Walters, 1995). As few as two LFSB motoneurons may be responsible for the entire CNS contribution to the siphon-evoked siphon withdrawal reflex (Antonov et al., 1999). By contrast, the three LFSA cells contribute only weakly to any of the siphon withdrawal responses (Hickie and Walters, 1995; Fang and Clark, 1996). Two other motoneurons, RDS and LD S i, are responsible for more than 80% of the head-evoked siphon response (the contributions are, respectively, about 65% and 15%; Hickie and Walters, 1995). At least five different clusters of mechanosensory cells are involved in gill and siphon withdrawal reflexes (Fig. 9.5; Dubuc and Castellucci, 1991).

Figure 9.5. The circuit for siphon-mediated siphon withdrawal in Aplysia. Several additional gill motoneurons are not shown here but are shown in Figure 9.3. Cells and synapses shown in black are inhibitory; all others are excitatory. Electrical synapses are indicated by resistance symbols, indirect connections by broken lines. "L.T. Unid." represents an unidentified group of low-threshold mechanosensory neurons. PMN, peripheral motoneurons. From Frost and Kandel (1995) with permission.

Defense

229

The one most commonly used in studies of plasticity is the LE cluster (see Fig. 9.6 for location), which was the first group of sensory neurons to be identified in Aplysia. There are approximately 25 neurons in the LE cluster, and all of them have receptive fields that are either entirely within the siphon or that include the siphon. In three other identified sensory clusters, namely

Figure 9.6. The locations of identified cells and identified cell clusters in the abdominal ganglion of Aplysia that are known to be involved in withdrawal reflexes involving the gill and siphon. Adapted from Frost and Kandel (1995).

230 Behavior and Neurons in Gastropods

rLE, RE and RF, there is a variable proportion of cells with receptive fields in the siphon. However, of the approximately 100 neurons in these four clusters, only about 44 have receptive fields that include the siphon; the other cells innervate the gill, the mantle, and the floor of the branchial cavity. The existence of an additional but still unidentified group of sensory neurons is implied by experiments showing that tactile stimulation produces EPSPs in motoneurons even before the LE neurons begin to fire. The EPSPs appear 77-89 milliseconds before the first LE spike, depending on the type of stimulation (Frost et al., 1997). Furthermore, stimuli too weak to excite the LE cells are nonetheless able to elicit withdrawal reflexes. Since the missing sensory neurons have not appeared in exhaustive searches of the CNS, they may be in the periphery (see Frost et al., 1997). Evidence for peripherally located sensory neurons was provided by Xin et al. (1995), who tracked the transport of radiolabeled amino acids from the skin of the siphon to the CNS. Since labeled axons were present in the CNS, the amino acids must have been incorporated into macromolecules before being transported centrally. It is therefore possible that at least some of the peripheral cells implicated in this study might be the missing mechanosensors. Illich and Walters (1997) examined the responses of LE neurons to siphon stimulation at varying strengths using a preparation in which the siphon was free to move, in contrast to the usual preparations in which the siphon is restrained. They found that the cells have high thresholds and they require a strong stimulus (15-35g/mm2) to fire. Furthermore, the number of spikes evoked by the stimulus increases with stimulus strength and maximum firing occurs only when the stimuli are strong enough to cause obvious injury to the skin. These results indicate that the LE cells are not typical mechanoreceptors, but rather they belong to a special class of mechanoreceptors known as nociceptors. The centrally located sensory neurons connect both monosynaptically and polysynaptically with motoneurons. It is estimated that 75-89% of the LE sensory cells converge on to the motoneuron L7, and the convergence on to LFS motoneurons is nearly as great (Frost and Kandel, 1995). Since several motoneurons each receive inputs from a large fraction of the LE population, it is reasonable to conclude that each LE sensory neuron excites many motoneurons (i.e., the sensory pathway is divergent). It would be interesting to know whether other clusters of sensory neurons have similar patterns of divergent outputs, or whether it is only the sensory cells with receptive fields on the siphon that have a widespread influence on motoneurons. From a different the point of view, it is apparent that all motoneurons do not receive the same afferent inputs, even if they innervate the same structure. For example, LE inputs converge on to L9 to a lesser degree than they do on to L7, perhaps because L9 has a more limited motor field (Fig. 9.3; Kandel, 1976). As well, different siphon motoneurons must receive inputs from different groups of sensory neurons because their responses depend on the location of the stimulus. Thus, LFSB cells, which bend the siphon towards the tail, are preferentially activated by stimuli delivered to the tail, whereas RDS and

Defense

231

LDS1, which constrict the siphon and bend it forward, are preferentially activated by stimulation of the head (Hickie and Walters, 1995). The LB and RD cells, which bend the siphon toward the head, are preferentially activated by stimulation of the mantle (Fang and Clark, 1996). The complex postsynaptic potential evoked in motoneurons by a sensory stimulus includes both monosynaptic and polysynaptic components. The proportion of the complex EPSP attributed to monosynaptic connections has been estimated to be as low as 5% and as high as 50%, or more, depending on the method used (Antonov et al., 1999). Two recent investigations used saline solutions containing high concentrations of calcium and magnesium to raise spike thresholds and thus block the polysynaptic pathways. The results from these latter studies are consistent in that both found that about 75% of the complex EPSP was blocked by the high cation solution, that is, only about 25% of the postsynaptic potential (PSP) is due to the monosynaptic pathway (Trudeau and Castellucci, 1992; Cohen et al., 1997). Regardless of the exact proportions by which synaptic drive is attributable to monosynaptic versus polysynaptic pathways, it is clear that interneurons play important roles in defensive withdrawal circuits. Not only do interneurons contribute to the compound EPSP, they also enable elements of complexity and plasticity that would otherwise be impossible. Although many interneurons no doubt remain unidentified, in some cases they have been characterized in detail. One particularly interesting example in Aplysia is the interneuron L29 and the mini circuit in which it participates, as illustrated in Figure 9.7. Several specific roles for interneurons in withdrawal circuits have been commented upon by others (Trudeau and Castellucci, 1992; Cleary et al., 1995; Frost and Kandel, 1995). A summary of these roles will now be given, with examples taken from the siphon withdrawal circuit of Aplysia as shown in Figures 9.5 and 9.7. (1) As already stated, interneurons contribute to the complex EPSP that is evoked in motoneurons. In this sense, the interneurons are said to be "recruitable" neuronal elements that increase the gain of the sensory to motoneuron connection. However, their participation in the reflex is not constant; it can be modified from sources both intrinsic and extrinsic to the reflex circuit. (2) Interneurons function as "gain control" elements. For example, the inhibitory interneurons LI6 and L30 control the recruitment of L29 (Fig. 9.7). Note that two additional interneurons, L21 and L35, also inhibit L29, but their actions can be removed by the mechanism of disinhibition, since L21 and L35 are themselves inhibited by L30. (3) A third role of interneurons is to maintain a "singleness of action." The activation of L30 prevents L21 and L35 from interfering with the action of siphon withdrawal, as could happen if they were concurrently activated from an extrinsic source (Fig. 9.7). (4) Sometimes interneurons mediate the inhibition of a reflex. One example is the transient inhibition of siphon withdrawal that follows noxious stimulation of the tail. This inhibition is mediated, in part, by the interneuron LI6. Another example is the inhibition of certain siphon movements following

232 Behavior and Neurons in Gastropods

Figure 9.7. A detailed summary of synaptic inputs and outputs involving the interneuron L29. Note that L29 both mediates and modulates the siphon withdrawal reflex. It's activity is heavily influenced by inhibitory influences, shown here in black. From Frost and Kandel (1995) with permission.

exposure to the animal's own ink (see Johnson and Willows, 1999); it is reasonable to think that the effect is mediated by one of the intrinsic inhibitory interneurons shown in Figure 9.7. (5) Interneurons prolong the duration of the reflex withdrawal by prolonging the duration of the motoneuronal discharge. While neither the sensory neurons nor the interneurons normally fire much beyond the duration of the stimulus, the EPSPs produced by L29 can last tens of seconds, in contrast to those of the sensory neurons, which last only a few milliseconds. The prolonged postsynaptic depolarizations attributable to L29 can cause motoneurons to fire even after the sensory neurons have stopped responding to the stimulus. (6) Some interneurons cause heterosynaptic facilitation of sensory neuron to motoneuron synapses (Fig. 9.5). In certain cases, for example L29, such

Defense

233

interneurons are activated by a local stimulus (i.e., one delivered to the siphon), whereas in other cases (not shown in the figures) they are activated only by strong remote stimuli. (7) Some recruited interneurons, R25/L25 in the present example, trigger behaviors that are functionally related to the withdrawal reflex, here respiratory pumping. Note that the recruitment of R25/L25 causes excitement of interneuron L33, which in turn inhibits the LBS siphon motoneurons and the L7 gill/mantle motoneuron (Fig. 9.5). (8) Interneurons distribute information about sensory stimulation to other circuits within the CNS. This function is not evident in any of the cells shown in Figures 9.5 and 9.7, but relevant interneurons have been described in regard to whole-body withdrawal (Section 9.3.1). In Aplysia, which does not have a true a whole-body withdrawal response, tactile stimulation nonetheless triggers generalized withdrawals beyond the point of stimulation if sufficiently strong (Teyke and Kupfermann, 1989). Furthermore, very strong stimuli elicit defensive responses that include not only widespread withdrawal but also mucus secretion, ink secretion, opaline secretion, turning, and locomotion (Walters and Erickson, 1986). A role for interneurons in the coordination of these responses is implied. 9.3.4.

Jail Withdrawal

Tactile stimulation of the tail of Aplysia causes withdrawal of the tail and then forward locomotion. The neural circuit for tail withdrawal is not as well understood as that for siphon withdrawal and gill withdrawal (see Cleary et al., 1995). The afferent signal for the reflex is carried by a population of about 40 sensory neurons belonging to two large clusters of about 200 cells each that are located on the ventrocaudal surfaces of the pleural ganglia (Walters et al., 1983). Each cluster, called ventrocaudal or VC, comprises mechanosensory cells that together innervate most of the body surface. Those neurons which respond to tail stimulation are grouped together within the larger cluster, and within the tail group itself there is a rough somatopic organization. Cells with receptive fields on the dorsal portion of the tail tend to be placed at anterior positions, while cells with ventral fields are placed posteriorly. Also, cells with very posterior fields are found mostly on the medial edge of the cluster. Like the LE sensory neurons discussed above, the VC cells require a strong tactile stimulus before they fire, and in general they have the characteristics of nociceptors (see Illich and Walters, 1997). The VC cells obviously serve an important role in the defensive behaviors of Aplysia. Tail withdrawal is mediated in part by three identified motoneurons in the pedal ganglion and two identified interneurons in the pleural ganglion. One interneuron, LP117, is particularly interesting because it seems to be involved in coordinating the multifaceted defensive response that is elicited by strong stimulation of the tail (Cleary and Byrne, 1993). LP117 fires when the tail is stimulated, and it has excitatory connections, either directly or indirectly,

234 Behavior and Neurons in Gastropods

with motoneurons responsible for retractions of the tail, the siphon, the gill, and the tentacles. It also excites neurons responsible for the release of mucus and ink. Since strong stimulation of the tail usually triggers forward locomotion as well, one would expect LP117 to also activate neurons in the pattern-generating system for locomotion, but these connections have not yet been reported.

9.4.

Plasticity of Defensive Behaviors

The current popularity of gastropod molluscs as subjects for behavioral neurobiology can be credited in large part to Eric Kandel's early vision of their potential for studies of learning and memory. Already by 1963, he had observed heterosynaptic facilitation in the abdominal ganglion of Aplysia. In 1970, he published, together with Vincent Castellucci, Irving Kupfermann, and Harold Pinsker, a set of three ground-breaking papers in Science magazine that introduced the gill withdrawal reflex of Aplysia as a model system that could be used to investigate the cellular mechanisms of learning (see Kandel, 1976). His enthusiasm for this system energized an entire generation of scientists, who advanced our knowledge of molluscan neurobiology and behavior even while they studied learning and memory. The choice of defensive behavior as the initial subject for studies of learning and memory in Aplysia was probably opportunistic, but it was clearly fortunate. Because defense is important for survival, and costly of energy, evolution will select neural mechanisms that optimize its function. For this reason, and from other similar arguments, Walters (1991) has suggested that the cellular mechanisms of behavioral plasticity might have first appeared in primitive mechanosensitive cells responding to nociceptive stimuli. The modifications of behavior that are dependent on experience are commonly referred to as "learning." My treatment of this vast subject focuses on the advantages of learning for defense. In addition, I shall briefly review what is currently known about the cellular and network mechanisms at play in the neural systems responsible for defensive behaviors. I have limited the discussion to modifications that last no longer than a few hours (short-term memories) because thereafter molecular mechanisms of memory storage dominate, and these lie beyond the scope of the present work (see Kandel, 2001). 9.4.7.

Habituation

The repetition of stimuli typically leads to a progressive diminution of the response, a process known as habituation. Provided that the stimulus is not too strong, it probably represents an insignificant threat and, if so, continuing to respond to it is maladaptive. Thus, habituation is an example of learning because it is dependent on experience and it is adaptive. It is a non-associative form of learning, and hence simple, because only the repetition of a single type of stimulus is required to modify the behavior.

Defense

235

Since habituation is ubiquitous among animals in general, it is not surprising to find that all withdrawal reflexes in gastropods show habituation. Given these features, it is fitting that Kandel's seminal work, cited above, should deal with habituation of the gill withdrawal reflex in Aplysia. These early papers also investigated dishabituation, the active process by which, following a strong stimulus, an habituated response increases in strength. Dishabituation is closely related to sensitization, the process by which, following a strong stimulus, a non-habituated response increases in strength (see Antonov et al., 1999). It has long been postulated that the cellular mechanism of habituation may be the depression of synaptic transmission between sensory neurons and follower cells, and indeed, repeated activation of mechanosensory neurons does lead to synaptic depression (Kandel, 1976). By reducing the excitatory drive in motoneurons, synaptic depression causes a reduced motoneuronal discharge and thus a smaller muscle contraction. The process is homosynaptic, meaning that it depends only on the repeated use of the synapses. Mechanosensory neurons in Aplysia also exhibit heterosynaptic depression, in which depression results from the activation of interneurons that produce long-lasting presynaptic inhibition at sensory neuron terminals. Only recently has the association between synaptic depression and behavioral habituation been rigorously tested in Aplysia. Hawkins and colleagues did this using new simplified preparations that permit simultaneous measurements of synaptic transmission and withdrawal contractions (Frost et al., 1997; Antonov et al., 1999). The results show that for siphonelicited gill and siphon withdrawals, homosynaptic depression of monosynaptic PSPs from sensory neurons to motoneurons largely accounts for habituation, although the contribution of this mechanism to behavioral habituation was not quantified. By contrast, a similar study of the tail-elicited siphon withdrawal reflex found that, while depression again occurs at sensory neuron to motoneuron synapses during behavioral habituation, heterosynaptic facilitation also occurs at the same synapses at the same time (Stopfer and Carew, 1996). Consequently, in the latter instance, monosynaptic transmission from sensory neurons to motoneurons actually increases during habituation, leaving synaptic depression at interneuronal synapses as the most likely explanation for habituation. The results indicate that plasticity is handled differently in the circuits mediating tail-elicited withdrawals and siphon-elicited withdrawals, but differences in experimental procedures might also have contributed to the different results (Antonov et al., 1999). The fact that both synaptic depression and synaptic facilitation occur simultaneously in the neural circuit for siphon withdrawal explains the non-monotonic appearance of the learning curve during repetitive tactile stimulation (Prescott and Chase, 1996; Stopfer and Carew, 1996). When learning experiences engage the dual processes of response increment and response decrement, the resulting competition between these processes determines the final behavioral outcome (Prescott and Chase, 1999). Typically, synaptic depression occurs at upstream loci while facilitation

236 Behavior and Neurons in Gastropods

occurs at downstream sites. With repeated stimulation using stimuli of intermediate strength, interactions between depression and facilitation cause the response to first increase (sensitize) and then decrease (habituate). Dualprocess learning is widespread among both invertebrates and vertebrates. The cellular explanation for homosynaptic depression is generally thought to be the depletion of transmitter from the pool of readily releaseable vesicles (Kandel, 1976; Byrne and Kandel, 1996). Quantal analyses of transmission support this idea because they consistently find that the locus of depression is presynaptic. Furthermore, when vesicles at the synaptic membranes of sensory neurons are counted by electron microscopy, fewer vesicles are found in samples taken from habituated animals than in samples taken from control animals. However, not all data are consistent with a simple depletion model, and it has been suggested that depression might even occur independently of transmitter release (Royer et al., 2000). It must be recalled that, under most normal circumstances, withdrawal reflexes are mediated conjointly by both the CNS and the PNS (see Sections 4.3 and 4.4). Furthermore, synaptic depression is known to occur in the PNS. Thus, if the PNS is experimentally isolated from the CNS, it alone can mediate habituation of withdrawal reflexes (Lukowiak and Jacklet, 1972; Prescott and Chase, 1996). When the system is intact, synaptic depression no doubt occurs at both central and peripheral loci, and the relative contributions of depression at these two loci depend on the intensity of stimulation. Interactions between the center and the periphery are also important, as emphasized by Lukowiak and colleagues (Mpitsos and Lukowiak, 1985). For example, the central gill motoneuron L9 is able to modulate the level of habituation in the gill withdrawal reflex, or even prevent its occurrence, even though firing L9 does not appear to affect activity in any other central motoneuron. One can infer from these data that L9 probably modulates peripheral circuits, perhaps presynaptically at neuromuscular junctions as suggested for other neurons by Peretz et al. (1976). 9.4.2.

General Sensitization

This section deals with a type of behavioral plasticity in which strong or noxious stimuli delivered to one part of the body cause increases in withdrawal responses elicited by stimuli delivered to a different site. As the term is used here, general sensitization is a form of non-associative learning because it does not depend on a training regime in which the stimuli are paired in time, in contrast to some of the phenomena that are discussed in the next section. While it is obvious that general sensitization is adaptive, because it prepares the animal for an imminent attack anywhere on its body, general sensitization is only part of a much broader response, one which Walters (1991) calls "nociceptive sensitization." The term refers to the state of readiness that an animal enters after noxious stimulation. It involves physiological changes that prime the animal to perform a number of related defensive behaviors that includes withdrawal, but is not limited to withdrawal.

Defense

237

The ways in which different body parts are affected by nociceptive sensitization can differ depending on the site of noxious stimulation. For example, noxious stimulation of the tail of Aplysia actually inhibits the siphon withdrawal reflex for about 10 minutes following the stimulus. A detailed examination of this phenomenon revealed that inhibition affects only those aspects of siphon withdrawal that can interfere with the siphon's role in directing ink towards the tail (i.e., towards the site of attack; see Fig. 9.4; Illich et al., 1994). Furthermore, the enhancement of withdrawal reflexes by general sensitization is not necessarily equal at all body sites. From the work of Walters (1987), it is known that nociceptive sensitization involves not only general sensitization, but also site-specific sensitization, in which the enhancement of withdrawal is greater at the site where the noxious stimulus is received than at other sites. Because site-specific sensitization shares features with associative conditioning, it will be discussed in the next section. The procedure for sensitizing withdrawal reflexes in Aplysia requires only a strong pinch of the animal's head, neck, or tail. However, for reliable stimulation in experiments, a brief train of electrical shocks is usually delivered to the body wall through implanted wires. Alternatively, in reduced preparations, one of the pleuroabdominal connectives or a peripheral nerve is shocked. The ensuing sensitization can last for hours or even weeks, depending on the number and patterning of shocks (Kandel, 1979). Although the following discussion deals only with central mechanisms, it must again be noted that peripheral neural circuits are undoubtedly involved (see Section 4.3; Mpitsos and Lukowiak, 1985). Because the periphery is largely inaccessible to neurophysiological investigation, nearly all of our detailed knowledge comes from studies of the central components. Early studies focused on changes at the monosynaptic connection between LE sensory neurons and identified gill motoneurons, typically L7 or LDG1 (see Kandel, 1979; Kandel and Schwartz, 1982). The electrical shocks that produce sensitization of the reflex in the whole animal also produce facilitation at these synapses in the isolated ganglion. Facilitation is caused by a heterosynaptic mechanism, and some of the interneurons that are capable of mediating facilitation have been identified, for example, L29 (Figs 9.5 and 9.7). Serotonin is implicated in the mechanism primarily on the basis of the observation that application of serotonin to the isolated ganglion can mimic the effects of firing L29, even though L29 itself does not contain detectable amounts of serotonin. Furthermore, stimulation of a serotonergic interneuron in the cerebral ganglion, named CB1, can facilitate synapses between siphon sensory neurons and motoneurons (Mackey et al., 1989). Numerous individual studies, reviewed by Kandel and Schwartz (1982), led to an explanation of reflex sensitization that can be briefly summarized as follows. Shock excites L29 and other interneurons, which then release serotonin and other modulating substances possibly including peptides. These transmitters affect presynaptic sites at the sensorimotor connection, causing the sensory neurons to release more transmitter when excited by subsequent tactile stimuli. Consequently, there are larger EPSPs in the

238 Behavior and Neurons in Gastropods

motoneurons, more spikes, and a greater motor response. Most of the early studies of sensitization were conducted using the gill withdrawal reflex, but later experiments with siphon withdrawal and tail withdrawal produced analogous results, that is, behavioral sensitization was again correlated with presynaptic facilitation of the sensorimotor synapses (see Frost et al., 1988; Cleary et al., 1995). While facilitation at sensorimotor connections is widely attributed to the actions of serotonin, inhibitory effects on defensive reflexes are mediated by other pharmacological agents. The peptide FMRFamide mediates presynaptic inhibition of sensory neurons in Aplysia (see Walters, 1991). In the terrestrial snail Cepaea nemoralis, opioid-like substances have been reported to suppress defensive responses (Kavaliers and Ossenkopp, 1991). Ordinarily, a snail lifts its foot when placed on a warmed surface. However, this nociceptive response does not occur, or it is reduced, when the snail is treated with morphine or enkephalin; by contrast, the responses are increased after treatment with the opioid antagonist naloxone. The exact mechanism of presynaptic facilitation has been debated. In early studies by Klein and colleagues it was found that the stimuli that cause facilitation also cause a reduction of current flow through voltage-sensitive K + channels, which in turn leads to a prolongation of the action potential in sensory neurons. Longer-duration action potentials cause an increase in the release of transmitter because they permit a greater influx of Ca++. However, later experiments, also by Klein (1994), indicated that the amount of facilitation caused by action potential prolongation (5%) is negligible compared to the amount caused by serotonin (300%). This result, together with data from other investigators, implies that facilitation relies at least in part on a mechanism that is independent of spike duration. Still unidentified, this second mechanism could involve vesicle mobilization, vesicle exocytosis, or the modulation of Ca++ channels localized at the release site. Meanwhile, it was discovered that serotonin activates two separate intracellular signaling pathways. One pathway uses cyclic AMP to activate the A-kinase (PKA), and the other uses diacylglycerol to activate the C-kinase (PKC). Suspicions also arose over differences between facilitation at rested synapses versus facilitation at previously depressed synapses (dishabituation). Eventually, Byrne and Kandel (1996) were able to reconcile all the available data by proposing that PKA and PKC both cause spike prolongations and engage the spikeindependent mechanism(s), but PKA mostly does the former and PKC mostly the latter. Furthermore, again according to Byrne and Kandel (1996), different kinases are activated depending on the state of the synapse, with PKA dominating at rested synapses, and PKC dominating at depressed synapses. Another variable that should be borne in mind is possible differences in the properties of pleural versus abdominal sensory cells, differences that could account for some of the discrepant results summarized in this paragraph. The mechanisms already discussed can partly explain sensitization of defensive withdrawal reflexes, but the extent to which behavioral sensitization

Defense

239

can be attributed to presynaptic facilitation at the monosynaptic connection between sensory neurons and motoneurons has not yet been quantified. Indeed, the first direct evidence that plasticity at this synapse is even partially responsible for behavioral plasticities was provided by Antonov et al. (1999) using a new preparation designed for this purpose. The findings clearly demonstrate that heterosynaptic facilitation at sensorimotor synapses contributes to dishabituation and sensitization, but the percentage of that contribution relative to other possible mechanisms is not specified. Recent studies have revealed additional cellular mechanisms, operating at various sites, that also contribute to sensitization (Frost et al., 1988; Trudeau and Castellucci, 1993; Cleary et al., 1995; Cohen et al., 1997). While none of these has received the detailed attention given to presynaptic facilitation at the sensorimotor synapse, they are all potentially important. The following list includes all mechanisms that are thought to play a role in the sensitization of gill and siphon withdrawals (refer to Figs 9.3, 9.5, and 9.7 for cells and circuits). (1) Presynaptic facilitation of the sensory neuron to motoneuron synapse. As described above. (2) Presynaptic facilitation of the sensory neuron to interneuron synapse. It has already been noted that the polysynaptic pathway from sensory neurons to motoneurons accounts for about 75% of the stimulus-evoked complex EPSP. Plasticity at this polysynaptic pathway was estimated by Trudeau and Castellucci (1993) during experiments in which the monosynaptic pathway (1, above) was also monitored. They found that the synapses between sensory neurons and excitatory interneurons facilitated about onehalf as much as did the synapses between sensory neurons and motoneurons. This is interesting because it implies that different branches of the same sensory neuron may be modulated to different extents by extrinsic influences. (3) Increased excitability of sensory neurons. Sensitization is often accompanied by an increase in the number of action potentials fired by sensory neurons when responding to a tactile stimulus. This effect results from a PKA-mediated reduction in K+ currents, that is, the same mechanism that causes an increase in action potential durations (see above). (4) Post-tetanic potentiation at synapses between excitatory interneurons and motoneurons. Although heterosynaptic facilitation has not been seen at interneuron to motoneuron synapses, a form of homosynaptic potentiation, known as post-tetanic potentiaion, is evident in the LF$ motoneurons after high rates of firing in either L29 or L34 (Frost and Kandel, 1995). The significance of this phenomenon is brought into question, however, by the results of a separate study in which no change in the amplitude of EPSPs was observed after connective nerve stimulation equivalent to that which produces sensitization (Trudeau and Castellucci, 1993). Thus, post-tetanic potentiation at these synapses may require a rate of presynaptic activity that does not necessarily occur during behavioral sensitization. Note that a different form of potentiation may occur at the sensory neuron to motoneuron synapse, one that is mediated by TV-methyl-o-aspartate receptors (Murphy and Glanzman, 1997).

240 Behavior and Neurons in Gastropods

(5) Reduction of feedback inhibition in the interneuronal pathway. Some pathways in the gill and siphon withdrawal networks go from sensory neurons to excitatory interneurons to inhibitory interneurons and then back to excitatory interneurons. An example of such a feedback loop, shown in Figure 9.7, is the pathway LE to L29 to L30 to L29. Trudeau and Castellucci (1993) recorded simultaneously from inhibitory interneurons and excitatory inteneurons before and after shocking a connective nerve to simulate sensitization. They found that transmission between the inhibitory interneurons and the excitatory interneurons was significantly reduced after connective nerve stimulation. This implies that sensitization disinhibits the excitatory interneurons and allows them to fire at higher rates in response to tactile stimulation. It was also observed that connective nerve stimulation reduced the excitability of the inhibitory interneurons. (6) Neuromuscular facilitation. Since stimuli that sensitize gill and siphon reflexes also increase the basal firing rates of motoneurons, the conditions necessary for peripheral facilitation are established (Frost et al., 1988; Cohen et al., 1997). Jacklet and Rine (1977) were the first authors to show that it actually occurs, using the gill motoneurons L7 and LD G i- Neuromuscular facilitation is the most likely explanation for the finding that gill contractions can increase even when the complex PSP in a key motoneuron does not increase (Cohen et al., 1997). The explanation in this case for why the complex PSP does not increase, despite presynaptic facilitation of connections from sensory neurons, is that the stimulus transiently activates inhibitory interneurons (Cohen et al., 1997; Antonov et al., 1999). The pre-eminence of Aplysia californica as a model organism for neurophysiological studies of behavioral plasticity raises a question about the generality of the mechanisms described above. Unfortunately, studies of comparable problems in other species are often hampered by the absence of mechanosensory neurons with readily accessible cell bodies. Nevertheless, the results obtained in Helix lucorum by Balaban and colleagues are at least broadly consistent with the mechanisms described for Aplysia (Zakharov et al., 1995). There is a population of cells on the mediorostral edge of the pedal ganglion in Helix that have properties like those of the facilitating interneurons in Aplysia. They contain serotonin, they are excited by noxious stimulation of the skin and, when activated, they heterosynaptically increase the excitatory input to identified "command" neurons. By thus increasing the responses of the command cells to sensory stimulation, the pedal interneurons may cause a general sensitization of withdrawal behaviors. A novel approach to understanding the evolution of general sensitization was taken by Wright et al. (1996), who looked for the appearance of neuromodulation in five anaspidean species and one outgroup species, Bulla gouldiana. Putative homologues of the tail mechanosensory neurons were identified in all six species, and experiments were performed to test for two types of neuromodulation induced by serotonin: spike broadening and increased excitability. Recall that both spike broadening and increased excitability are implicated as mechanisms of general sensitization in Aplysia.

Defense

241

Wright et al. then mapped the neurophysiological data on to a phylogenetic tree that was constructed from character states. The results, summarized in Figure 9.8, suggest that neuromodulation dependent on increased excitability is an ancestral trait, present even in the distantly related genera Akera and Bulla, whereas spike broadening first appeared in Aplysia or its close relatives. Interestingly, both increased excitability and spike broadening are absent from the most recently evolved member of the clade, Dolabrifera. This last result suggests that some mechanism common to both types of modulation was discarded during the evolution of the lineage that led to Dolabrifera.

Figure 9.8. A proposal for the evolution of sensory neuron plasticity in Aplysia and related opisthobranch genera. The branched pattern at the left is a hypothetical phylogenetic tree developed from cladistic analysis of anatomical and morphological traits. The figures in the middle show superimposed action potentials before and after treatment with serotonin (5-HT). The figures at the right show responses to depolarizing current injections before and after treatment with 5-HT. A mapping of the physiological data on to the phylogenetic tree suggests that the ancestors of all modern anaspideans had the ability to modulate excitability, but this trait was lost in the lineage leading to Dolabrifera. Spike broadening evidently appeared later in evolution and it, too, was subsequently lost. From Wright et al. (1996) with permission.

242 Behavior and Neurons in Gastropods

The authors suggest that the nocturnal habits and the protected habitat of Dolabrifera might provide it with sufficient protection against predators. If sensitization is no longer required, the costs associated with maintaining it can be recovered by eliminating sensitization from the animal's repertoire.

9.4.3.

Site-specific Sensitization and Associative Plasticities

In 1983, two groups independently reported an important discovery concerning the plasticity of sensorimotor synapses in Aplysia (see Kandel and Schwartz, 1982). Simply put, presynaptic facilitation was found to be greater when the sensitizing stimulus is delivered immediately after the sensory neuron has fired action potentials than when the sensitizing stimulus is delivered while the sensory neuron is silent. I will follow Walters' (1987) usage in referring to this effect as "activity-dependent extrinsic modulation," or ADEM. The significance of ADEM lies in its ability to mediate learned associations between two stimuli, one of which causes the sensory neuron to fire and the other which causes the facilitating interneuron to fire. The underlying mechanism is thought to involve the convergence of two signals affecting adenylate cyclase. Calcium enters the sensory neuron through voltage-dependent channels when the cell fires an action potential. The calcium primes adenylate cyclase in the sensory neuron's axon terminals so that when the cyclase is subsequently activitated by serotonin, even more cyclic AMP is produced than when serotonin acts alone. Terry Walters discovered ADEM while studying a phenomenon that he calls site-specific sensitization. He realized that, if the skin of Aplysia receives a noxious stimulus, the defensive responses of the animal should be directed toward the site of stimulation, because this area is most likely to receive any follow-up attack and it is also the most vulnerable to serious injury owing to damage caused by the initial attack. Walters further recognized that ADEM is perfectly suited for mediating site-specific sensitization because the noxious stimulus should excite both sensory neurons and modulatory neurons specifically in the region of stimulation. To test this idea, Walters (1987) delivered a noxious stimulus to a specific place on the tail during training. He then compared the ability of a weak stimulus to elicit siphon withdrawal when applied either to the trained site or to a control site on the opposite side of the tail. The results from this experiment, and from similar ones performed with stimuli delivered to other body regions or using other measures of response, demonstrated a large difference in the amount of withdrawal elicited at the trained site relative to the control site. In some experiments the superiority of the trained site continued to be evident for several days after training. Responses elicited from the control site were also sensitized (general sensitization), but not so much as at the traumatized site (site-specific sensitization). Figure 9.9 schematically illustrates the pathways for these effects.

Defense

243

Figure 9.9. A model of the plastic phenomena induced in mechanosensory pathways following noxious stimulation. A noxious stimulus causes firing in modulatory neurons (MOD) that results in increased cell excitability and increased transmitter release, as indicated by arrows. If sensory neurons are co-active with modulatory neurons (shaded), the former are subject to activity-dependent extrinsic modulation (ADEM), which accounts for site-specific Sensitization and classical conditioning. Sensory neurons that are not activated by the noxious stimulus (e.g., number 2) may also be sensitized, but to a lesser extent. US, unconditioned stimulus; CS, conditioned stimulus (classical conditioning). From Walters (1987). Copyright 1987 by the Society for Neuroscience.

A further intriguing aspect of site-specific Sensitization relates to its possible function in compensating for injuries sustained at the site of attack. This could happen if afferent fibers are damaged in the attack. Later, the enhanced signal provided by ADEM in the surviving afferents could compensate for the loss of signal due to the damaged fibers (Walters, 1991; Illich and Walters, 1997). The initial descriptions of ADEM drew attention to the fact that the main requirement for ADEM closely resembles the main requirement for associative classical conditioning, namely, that a weak stimulus (conditioned stimulus, CS) must immediately precede a strong stimulus (unconditioned stimulus, US). According to the usual interpretation of this learning paradigm, the animal learns that the CS predicts the US. For defensive reflexes, the survival value of such a learning mechanism is obvious. Not surprisingly then, it has proven possible to classically condition the gill and siphon withdrawal reflex of Aplysia. For the CS, investigators typically use a tactile stimulus to the siphon as the CS, and for the US they usually use an electrical shock to the tail. From the schematic representation in Figure 9.9, it is evident that ADEM could account for classical conditioning of the siphon withdrawal reflex. Antonov et al. (2001) used a simplified semiintact preparation to examine this tissue. While they were able to establish that activity-dependent activity at synapses between LE sensory neurons and LFS motoneurons almost certainly contributes to behavioral conditioning,

244 Behavior and Neurons in Gastropods

they were unable to say whether this is owing entirely to an ADEM mechanism or whether other mechanisms, for example hebbian long-term potentiation (LTP) mediated by 7V-methyl-D-aspartate (NMDA) receptors (Murphy and Glanzman, 1997), also contribute. Quite possibly, both presynaptic mechanisms (ADEM) and post-synaptic mechanisms (LTP) participate in producing the synaptic plasticity that underlies classical conditioning. Defensive responses can also be elicited or enhanced by classical conditioning procedures that use conditioned stimuli in modalities other than touch. Experiments in Helix demonstrate that if food (e.g., carrot) is paired with electric shock, a general aversive response is established such that subsequent presentations of the carrot alone elicit withdrawals of the head and tentacles, and closure of the pneumostome (Balaban, 1993). In Aplysia, chemicals added to the seawater can serve as conditioned stimuli. In these experiments, the chemical itself does not necessarily elicit a defensive response but its presence seems to provide a context in which the animal learns to expect the noxious stimulus. For example, in the original experiments of Walters et al. (1981), training consisted of adding an extract of shrimp to the seawater 60 seconds before the onset of an electrical shock. After nine such pairings, the shock elicited stronger and more long-lasting withdrawal behaviors, more inking and more locomotion than the same shock stimulus elicited in control animals that had received the shrimp extract and the shocks in an unpaired protocol. Animals trained in this manner can be said to have acquired a conditioned fear, similar to that which is seen in vertebrates. Experiments such as this have therefore led investigators to comment upon the similarity of learning in gastropods and vertebrates, especially in the realm of defensive behaviors. The existence of these common features validates the gastropod model for studies of the cellular mechanisms underlying associative learning (Walters et al., 1981; Balaban, 1993; Hawkins et al., 1998).

9.5.

Chemical Defenses

Although withdrawal reflexes are very attractive for neurobiological study, and they do play a role in defense, they offer no real protection to a small, slow, and shell-less gastropod in an encounter with a large and hungry carnivore. Chemical defenses come to the fore in such situations. In principle, one can distinguish passive defenses, in which tissues of the body contain substances that are unpalatable or toxic to potential predators, from active defenses, in which chemicals that are released by the animal allow it to escape an attack. In practice, however, the distinction is often difficult. Much work is required to determine the identity of defensive chemicals, the tissues in which they are found, and the circumstances, if any, in which they are released. Even for a putative passive defense, it must not only be demonstrated that an animal sequesters a particular metabolite and that the metabolite is unpalatable to potential predators, but also that the metabolite is stored in the skin or some other tissue where it can be an effective deterrent,

Defense

245

rather than in an organ like the digestive gland where the predator is unlikely to encounter it, or at least not until after it has killed the gastropod. Anyway, chemists have documented the extraordinary abilites of opisthobranchs to accumulate selected metabolites from dietary sources, to transform them, and to synthesize unusual compounds de novo (Cimino et al., 1999). Aplysia dactylomela, for example, sequesters about 20 secondary metabolites derived from dietary algae (Carefoot, 1987). The focus of the discussion below is on the secretions from skin and glands, since these have the clearest defensive functions and they are often controlled by nervous circuits. 9.5.7.

Secretions from the Skin

When provoked by tactile or chemical irritants, nearly all gastropods release slime from the dorsal surface of their skin. The slime is mostly water, but it contains sulphated acidic mucopolysaccharides, proteins, and exogenous substances derived from the blood. Several epidermal gland types can be differentiated by morphology, chemical content, and distribution on the skin; as many as 11 types have been described in Lymnaea stagnalis. Many of the glands secrete continuously to seal the body wall and to protect the skin. By contrast, the main function of the evoked secretions is to dilute and wash off irritants. In some cases, however, the slime plays a more active role in defense. For example, terrestrial slugs secrete a slime, either yellowish or milky white, that can be highly repellent to insects and agonistic conspecifics (Rollo and Wellington, 1979). The notaspid opisthobranchs secrete an especially acidic mucus, with pH as low as 1-2 (Gillette et al., 1991). This slime is an effective deterrent to predators, but it also produces temporary blisters in the skin of the secreting animal. Experiments performed in Pleurobranchaea suggest that the release of acid is controlled jointly by the CNS and the PNS, because it can be triggered either by electrical stimulation of peripheral nerve stumps or by mechanical stimulation of the skin in de-ganglionated preparations (Gillette et al., 1991). The release of mucus from the body wall of Aplysia is generally interpreted as defensive, even though its effectiveness in preventing predation or other dangers has not been reported. What makes mucus release in Aplysia interesting is that the two largest neurons in the nervous system are apparently involved in its control. The neurons R2 and LP1, found respectively in the abdominal ganglion and the left pleural ganglion, constitute an homologous pair (Hughes and Tauc, 1963). After injecting them with horseradish peroxidase, Rayport et al. (1983) traced their branching axons to subepidermal glands. Intracellular stimulation of R2 or LP1 resulted in the release of mucus from glands discharging to the body surface. However, the amount of mucus released in this manner was small compared to the amount released by noxious stimulation of the skin, which led the authors to suggest that local body wall contractions ordinarily work together with R2 and LP1 to effect release. The control of mucus release, therefore, seems similar in Aplysia and in Pleurobranchaea in so far as interactions between the CNS

246 Behavior and Neurons in Gastropods

and the PNS are required in both cases. To further investigate the interaction in Aplysia, it would be interesting to see how much mucus is released by mechanical stimulation of the skin after lesions of R2 and LP1. 9.5.2.

Secretions from the Ink and Opaline Glands

The miserable specimens of Aplysia californica, commonly seen lying on the beaches of southern California surrounded by a heavy swirl of purple ink, are typically the victims of curious and mischievous children, not predators. They have released their ink under the duress of poking and prodding. In its natural environment, Aplysia is rarely seen to release ink. According to Carefoot et al. (1999), only one author has reported seeing ink released in the field under natural provocation from another animal. In the laboratory, release can be triggered by noxious stimulation, but some observations suggest that distributed stimulation such as that brought about by picking up an animal is even more effective (Nolen and Johnson, 2001). Although there is ordinarily a high sensory threshold for release (Carew and Kandel, 1977), some circumstances have been noted in which the threshold is low (Illich et al., 1994; Nolen and Johnson, 2001). Field observations suggest a higher threshold for release in A. fasciata than in A. californica (Susswein et al., 1984). A defensive function for inking has long been suspected, but the effectiveness of ink in anti-predator defense was first tested by Nolen et al. (1995). These authors facultatively de-inked a group of juvenile Aplysia californica by confining their diet to foods that contained none of the pigments necessary to make ink (see below). The animals were then exposed to encounters with a natural predator, the anemone Anthopleura xanthogrammica, which found the inkless Aplysia no less palatable than their fully armed conspecifics. Once the anemone had engulfed a sea hare with its tentacles, and positioned its mouth over the potential meal, the experimenters used a syringe to deliver 6ml of freshly collected ink to the anemone's tentacles, or as a control, an equal volume of seawater. The anti-predator value of the ink was demonstrated by the fact that when the anemone was given ink, 71 % of the Aplysia survived, whereas only 7% survived if the anemone was given sea water. Although this result confirms a role for inking in defense, the exact mechanism by which it works is not known. In the experiments just described, ink caused the anemones to recoil their tentacles; sometimes an anemone would evert its digestive cavity and then assume a balled-up defensive posture. In other experiments, with crabs and urchins, ink from Aplysia caused general irritation, induced expansion of bristles, grooming, mucus production, and temporary cessation of the heartbeat (Carefoot et al., 1999). Besides acting as a sensory irritant, other defensive roles for Aplysia's ink have been proposed but none has been proven. These include camouflage, distraction, distastefulness, chemical signaling, and metabolic inhibition. It should be noted, however, that ink has no apparent deterrent effect against two known predators of Aplysia, namely lobsters and Navanax (see Nolen et al., 1995).

Defense

247

The ink gland is located at the edge of the mantle, directly over the gill (Fig. 9.4). Ink is stored in vesicles that are surrounded by a matrix of collagen and muscle fibers. When the muscles contract, ink is released ventrally into the mantle cavity. From there it is delivered into the environment by contractions of the parapodia and the body wall, an activity that closely resembles respiratory pumping. As already mentioned, specific movements of the siphon and nearby body parts ensure that the ink is directed towards the source of the eliciting stimulus (Fig. 9.4). For example, when the eliciting stimulus is received on the tail, the ink is ejected mostly through the siphon. The stimulus that causes ink to be released also modifies or inhibits local siphon reflexes so that these do not interfere with the directed ejection of ink (Illich et al., 1994). While all sea hares (anaspids) have an ink gland, and most of them release purple ink, a few species (e.g., Aplysia Juliana and Aplysia vaccaria) release white ink instead. The chromophore in the purple ink of Aplysia californica is mostly phycoerythrobilin, which is extracted from pigments in the red algae eaten by the animal. A protein of high molecular weight is also present in the ink, constituting about 35% of the dry mass. This protein is synthesized within the ink gland and not derived from algae; its function is unknown (Prince et al., 1998). The central nervous circuit for ink release in Aplysia was thoroughly described in early papers by Carew and Kandel (1977) and Byrne et al. (1979). These authors identified five "motor cells" comprising three L14 cells on the left ventral surface of the abdominal ganglion (Fig. 9.6) and two R19 cells on the right dorsal surface of same ganglion. While intracellular stimulation of these cells causes ink release, their direct connection with muscles or vesicles in the ink gland has not been established. Prince et al. (1998) found that isolated ink gland vesicles could be induced to release ink by application of low concentrations of acetylcholine. Since the identified central motor cells are not cholinergic, the fact that acetylcholine causes release suggests that the central motor cells act by exciting peripheral nervous elements. Sensory neurons in the LE, RE, RF clusters, and presumably in the VC cluster, all excite the central motor cells through direct and indirect pathways. Several interneurons have been identified, including those with both inhibitory and excitatory connections to the motor cells; all but one of these (L31) is also a member of the gill withdrawal circuit. The early characterization of inking as a high threshold, all-or-none behavior (Kandel, 1976; Carew and Kandel, 1977) encouraged researchers to identify neural mechanisms that might be responsible for these features, especially as they differed from the more graded responses of the gill withdrawal reflex. However, it has already been noted that inking does not always have a high threshold, and Leonard et al. (1989) have questioned whether gill withdrawal is really graded. Furthermore, recent experiments have shown that ink is not necessarily released in an all-or-none manner. In two species of Aplysia, Nolen and Johnson (2001) found that animals consistently released approximately 52% of their total ink store in response to the first stimulus, no matter what type of stimulus was used. When further stimuli

248 Behavior and Neurons in Gastropods

were presented, at intervals of 5-10 minutes, the animals continued to release ink for three or four more trials or, in one case, 11 more trials. Interestingly, the animals tended to release 30-50% of their available ink reserves on each occasion, even after the glands had been depleted to half of their initial contents. The difference between these results and those of Carew and Kandel, in which all-or-none release was described, is striking. Nolen and Johnson (2001) suggest that the discrepancy may be explained by the use of different procedures for estimating the glands' ink content, by different diets or by differences in the life histories of the experimental animals. All authors agree that the trigger mechanism for inking in Aplysia operates as an all-or-none device, and usually with a high threshold, even if there is disagreement about whether the release itself is all-or-none. In addition, there is no doubt that several properties inherent in the inking circuit can account for the stated properties of the trigger mechanism (Carew and Kandel., 1977; Byrne et al., 1979). Firstly, the ink motor cells have high resting membrane potentials and high thresholds for excitation, requiring about 30 mV depolarization to initiate spikes, in contrast to the gill motoneuron L7, which requires only 5-1 OmV. Secondly, they have a low input resistance due to electrical coupling; this reduces the amount of depolarization produced by synaptic activity. Thirdly, they exhibit a fast potassium current, 7A, that is activated at membrane potentials just above resting levels. This outward current effectively negates brief or small synaptic depolarizations. Together, these properties contribute to the requirement for strong sensory stimuli to trigger inking behavior. Nevertheless, stimuli that are effective in causing at least some neurons to fire at least some spikes can quickly induce high frequency bursts in all the motoneurons due to feedback excitation. Feedback is mediated by electrical coupling in combination with a late and slowly developing depolarization created by a conductance-decrease EPSP. Lastly, mention must be made of the opaline gland, which is a collection of vesicles found on the floor of the mantle cavity in Aplysia and other anaspids. Like the ink gland situated above it, the opaline gland discharges its product into the mantle cavity (Fig. 9.4). The gland releases a white substance, opaline, that is much more viscous than ink, but neither the composition of opaline nor its function is known. Strong stimuli evidently cause the combined release of both ink and opaline, although weak stimuli can also sometimes trigger the release of opaline (Walters and Erickson, 1986). One idea is that when opaline is combined with ink the high viscosity of opaline slows the dispersal of ink and therefore prolongs its effect. Alternatively, opaline itself be aversive because of its tendency to stick to the appendages of predators. Evidence supporting the notion that opaline has a defensive function comes from anecdotal observations that anemones withdraw their tentacles after having been exposed to it (Johnson and Willows, 1999), and crabs spend long periods of time grooming their mouth parts after they have become entangled with opaline/ink mixtures (T. Walters, unpublished observations).

10

The Temporal Organization of Behavior In the previous chapter we saw that even with the simplest withdrawal reflexes, where the causal relationship between stimulus and response leaves little room for complicated neural circuitry, the expression of behavior is affected by an animal's past experience through the processes of habituation, sensitization, and associative learning. Earlier, we saw that the more complex behaviors of feeding and mating are influenced by motivational states. In this chapter we consider some additional factors that influence the expression of behavior in gastropods, where the common element is the constraint of time. There are times of the day and times of the year when a gastropod will not respond, or responds only poorly, to stimuli that usually elicit behaviors, simply because the time is not right. In general, time constrains behavior because it brings a predictable sequence of environmental events, only some of which are suitable for particular behaviors. Virtually all animals have developed physiological timing mechanisms to predict and anticipate changes in the environment. However, unpredictable day-to-day changes in weather also influence the expression of behavior, especially in terrestrial species. The constraint of time can be seen as a demand that decisions be made continuously as to which behavior or behaviors is/are to performed, or alternatively, whether to do nothing but rest. The need for mechanisms to make such decisions is especially apparent when multiple simultaneous stimuli excite neural systems that trigger mutually incompatible behaviors. One of the fruits of cellular analysis in gastropod molluscs is that some of the neural mechanisms responsible for the temporal organization of behavior can be approached experimentally.

10.1.

Seasonal Cycles

The activity levels of gastropods, like those of all animals, are sensitive to conditions in the local environment. Temperature is obviously important because gastropods have no internal means of regulating their body temperature. Since their skin is semi-porous, gastropods are also greatly affected by external water conditions, in both the aquatic and the terrestrial 249

250 Behavior and Neurons in Gastropods

environments. Light, and the heat that often accompanies it, explains the nocturnal habits of terrestrial gastropods, allowing them to avoid evaporative water losses from daytime heat, while also providing protection against visual predators. Because all of these factors—temperature, water, and light—vary in a predictable manner according to the earth's movements in relation to the sun, the activity of gastropods exhibits seasonal and circadian variations. As we shall see, gastropods not only respond to changing conditions, they are also able to anticipate them using internal timing mechanisms. The dependency of behavior on season is greater on land and in temperate regions, where environmental variations are greater, than in aquatic and tropical regions, where the seasons are less pronounced. For this reason, and because it is easier to maintain terrestrial species in the laboratory for extended periods, and easier to study them at field sites, more is known about seasonal adaptations in terrestrial species than in aquatic species. Terrestrial snails and slugs are relatively inactive in winter, but the amount of activity depends on the severity of the climate. For example, in Europe the garden snail Helix aspersa hibernates for several months during the winter, but in southern California, if there is rainfall or irrigation, the same species is more active during winter than during summer. Slugs take shelter during cold periods by crawling into protected spaces between rocks or under ground litter. Here they contract their bodies and wait for more favorable temperatures. Snails carry their own physical protection in the form of a shell, and they use this to advantage in the winter by withdrawing into the shell and sealing the aperture with an epiphragm made of mucus and calcium. In this manner, snails can remain in a continuous state of hibernation for periods up to about 6 months. In the northern forests and southern mountains of Europe, the large snail Helix pomatia goes even further to prepare for winter (Lind, 1968). First, it searches for soft ground. Then it digs a hole using the method shown in Figure 10.1, which is the same method it uses to dig a nest for eggs in the early summer. The snail begins by moving slowly forward and downward. As it does so, soil sticks to the flattened sole of the snail's foot, and when the foot is then contracted, the soil is pulled back behind the snail. Thus, with successive extensions and contractions of the foot, the hole is excavated. Afterward, the snail turns itself over so that the opening of the shell (through which it breathes) is uppermost, and it then covers the roof of the hole with loose soil. Once that is done, the snail withdraws deep into its shell, and the shell aperture is sealed with two layers of thick epiphragm. The snail will only emerge from its nest with the warm temperatures and rainfall of spring. Remarkably, individual snails appear to return to the same hibernation site year after year. Although observed and commented upon by earlier authors, Lind (1989) has obtained the best evidence. Making observations every few days for five consecutive years, he recorded the summer excursions and winter hibernation sites of marked snails, Helix pomatia, in a wooded area near Copenhagen. The snails typically foraged in glades during the

The Temporal Organization of Behavior 251

Figure 10.1. Helix pomatia digs a hole to prepare for winter hibernation. (A) The snail locomotes forward and downward. (B) The sole is flattened against the soil. (C) A lifting motion loosens the soil. (D) As the snail contracts, the soil is lifted upwards and over the shell. (E) A new cycle of digging begins. After the hole is excavated, the snail turns over within the hole so that its shell aperture is uppermost. It then seals the aperture by secreting a calcareous epiphragm. Adapted from Lind (1968).

spring and summer, then returned to the woods to hibernate at the base of trees in late August (Lind, 1988). The hibernation sites of 454 marked snails were identified in two or more successive years. Overall, from 1157 hibernation events, 86.4% involved snails that returned to sites that were within 12m of the previous year's site. By comparison, the size of the home range for individual animals was 100 300m2, and there was no correlation between the size of the home range and the accuracy of homing. Although snails sometimes returned to within centimeters of a site used on previous occasions, other observations make it unlikely that the snails were attempting to return to the specific nest of previous years; rather, a local area seemed to be preferred. Lind also conducted a displacement experiment in which snails were removed from the site at which they had begun to hibernate and placed together at a new site located 10-24m from the original site. In the following winter, these snails returned to the original site, not to the (new) site from which they had emerged in the spring. The results of this experiment imply that, if the site preference is based on learning, which is not proven, the memory would have to persist for a full year (i.e., from the beginning of one hibernation to the beginning of the next). Even if site preferences are not learned but instead explained by sensory biases, the data still provide evidence of a surprisingly sophisticated sensory ability. From other studies that have looked at the mechanisms responsible for homing to daily rest sites (see Section 10.2), it is probable that snails use chemical cues to find hibernation sites, whether by responding to innate preferences or by

252 Behavior and Neurons in Gastropods

remembering learned odors. Another possibility is the use of the magnetic sense (see Section 6.4.5). The specific influence of temperature on hibernation is evident in the fact that, in Switzerland, snails living at an elevation of 580m hibernate 1 month earlier than snails living at 375m (see Lind, 1988). Nevertheless, it is interesting that although hibernation behavior is clearly an adaption to cold weather, and the exposure to low temperatures can induce hibernation, temperature alone is not responsible for initiating hibernation in nature. Field observations by Lind (1988) and other workers have revealed that hibernation often begins even before temperatures drop. Decreasing day length seems to be the critical factor. Bailey (1981) observed a group of snails in his laboratory throughout the year. He found that they reduced their activities when the photoperiod was shortened in step with the arrival of winter, even though the temperature was kept constant. Bailey also found evidence for an endogenous timing mechanism because snails that were maintained at constant temperature and with a constant 12-hour photoperiod nonetheless exhibited a circannual cycle that included seasonal changes of locomotion, feeding, and reproduction. The suggestion of internal cues for seasonal behavioral changes needs to be confirmed, however, because Bailey's observations covered only a single year. To summarize, the weight of evidence suggests that hibernation behavior is possibly readied by an internal circannual clock, it is begun in response to the declining photoperiod, and it is accelerated by falling temperatures. Dormancy during the summer months is called estivation. The main function of estivation is to limit evaporative water loss (see Section 5.5). Desert snails (e.g., Otala) can spend more than half their lives in estivation. To prepare for estivation, snails withdraw into their shells and seal the aperture with an epiphragm, much like the preparation for hibernation. However, there are interesting differences between hibernation and estivation. For one, hibernating snails remain either in a dug burrow or near the ground surface, whereas aestivating snails are typically found at heights above the ground, presumably to reach cooler air. Also, the epiphragms of estivating snails are thinner than those of hibernating snails, and the animals can more easily be roused from estivation than from hibernation. Finally, estivation is readily induced by hot dry conditions, whereas hibernation requires a convergence of temperature, photoperiod, and, perhaps, internal cues. Common to both types of dormancy, however, is a profound reduction of metabolism and protein synthesis. Given that the behavior of many gastropods changes drastically with season, one expects parallel changes in the nervous system, and various approaches have been taken to look for such changes. Several studies, including those by Kaufmann et al. (1995) and Bernocchi et al. (1998), have reported that CNS neurons in hibernating snails contain higher concentrations of peptides than do comparable neurons in active snails, based on immunohistochemical methods. Also, the levels of dopamine and serotonin in cell bodies, but not in the neuropil, are greater in hibernation than in activity.

The Temporal Organization of Behavior 253

These results imply that secretory material accumulates in cells during hibernation. Because the immunohistochemical signal is strong in hibernating snails, Bernocchi et al. (1998) point out that investigations of peptide function could take advantage of the opportunity to better visualize patterns of peptide expression. For example, these authors found more evidence for the co-localization of peptides in hibernating snails than in active snails. When Gainer (1972) studied the North African snail Otala lactea, he found that an identified neuron, cell 11, is typically silent when taken from dormant (hibernating) animals, but typically bursting or showing pacemaker activity when taken from active animals. These observations are consistent with the histochemical picture described above because, presumably, when cell 11 is not bursting it is not releasing its neurosecretory product, so the product should accumulate in the cell. Changes in the intrinsic electrical properties of cell 11 were also noted after long periods of dormancy. Principal among these were the loss of membrane rectification and the substantial reduction in the input resistance. Experiments indicated that both of these effects could be accounted for by a specific increase in membrane permeability to K+. Besides these electrophysiological effects, Gainer found that protein synthesis in cell 11 was about 50% less during dormancy than during the active period. A change in synaptic transmission was noted by Copping et al. (2000) who examined the connections made by the neuron RPeDl in Lymnaea stagnalis. As described in Section 5.2, RPeDl is a member of the respiratory central pattern generator and it is believed to be responsible for initiating respiration. Copping et al. discovered that the connections between RPeDl and its followers (e.g., VD2/3 and the VH, VI, VJ, and VK cells) were not apparent in 10 of 11 preparations studied in late March, whereas they were reliably present at other times of the year. Moreover, the cells that did not respond to spikes in RPeDl also did not respond to dopamine, the transmitter normally used by RPeDl. Since Lymnaea hibernates in the mud during the winter months and does not breathe air at this time, the data suggest that there is a down-regulation of postsynaptic dopamine receptors in the winter months. The seasonal modulation of electrical activity in neurons could be mediated by a rise in extracellular ion concentrations, particularly calcium. Prolonged dormancy, with the resultant loss of water, causes snail hemolymph to become hypertonic, and electrophysiological studies conducted in vitro indicate that hypertonic solutions cause a reduction in nervous activity (see Takeda and Ozaki, 1986). However, there are few data on the activity levels of CNS neurons in intact dormant snails. Although one imagines it to be low, unpublished work by Viktor lerusalimsky and myself in Helix aspersa indicated that at least a fair amount of neural activity persists. We used implanted fine wires to record action potentials in three peripheral nerves. While the general level of activity did generally decline during the course of estivation, strong bursts of activity continued to be recorded at irregular intervals for as long as 30 days after the start of estivation. The bursts occurred in the absence of any apparent sensory events or any overt motor acts.

254 Behavior and Neurons in Gastropods

Some of the results from Gainer's (1972) experiments, described above, actually implicate mechanisms other than modified blood composition as the root cause of changes in neuronal electrical activity. For example, since cells from aestivating animals were studied in vitro surrounded by normal snail saline, it is difficult to attribute their lack of bursting behavior to hypertonic blood. It is also interesting that the experiments in which Gainer compared cells from hibernating snails and active snails were performed only during the winter months. In a different set of experiments, when cell 11 was removed from estivating snails at different times of the year, there was a strong tendency for the cell to be silent in the winter months and bursting, or at least active, during the spring and summer months. In two successive years there was a striking increase in cell 11 's activity in the month of April, when the snails usually emerge from hibernation. These results—the inactivity of dormant cells in winter months, even when recorded in normal saline, and the relatively greater activity at other times of the year—suggest that the seasonality of electrical behavior is controlled by an endogenous mechanism within the nerve cells, similar to the mechanism for expression of the circadian clock, as described below. The idea of a circannual clock is consistent with the behavioral evidence, reviewed above, that likewise suggests an internal timekeeping mechanism operating on an annual cycle. A final point concerns whether the phenomena discussed above are the causes or consequences of seasonal behavioral change. Although this is a difficult issue to address experimentally, the idea that cell 11 might be involved is worth testing. This cell has apparent homologues in other terrestrial snails including Achatina (the PON cell) and Helix (neuron RPal, also called Fl and Br). According to Kerkhoven et al. (1993), RPD2 of Lymnaea and R15 of Aplysia are also homologous to cell 11. In none of these cases, however, has a function for the cell been clearly established. Studies in Lymnaea and Aplysia indicate that RPD2 and R15, respectively, are involved in respiration, cardiovascular function, water regulation, and other internal systems that are likely to be affected by hibernation (see Sections 5.5 and 8.6). The large size of cell 11, and its dorsal location (albeit in a subesophageal ganglion), make it possible to lesion the cell in order to assess the consequences, if any, on seasonal behaviors.

10.2.

Daily Cycles

The behavior of animals is strongly affected by the environmental changes that occur on a daily basis as a result of the earth's rotation about its axis. Depending on the species, activity is generally confined to either the night or the day. Although the daily control of activity patterns depends in part on external cues such as light and temperature, when these cues are kept constant under experimental conditions, many behaviors continue to be expressed in a pattern that approximates a 24-hour cycle (i.e., it is circadian). The persistence of a behavioral rhythm in the absence of relevant external cues

The Temporal Organization of Behavior 255

is due to the presence of an internal time-keeping mechanism, and the term "free-running" is used to describe the operation of the internal clock in situations where there are no external cues to entrain or reset the mechanism. Two marine gastropod species, Aplysia californica and Bulla gouldiana, have proven especially useful as model systems for studying the cellular basis of the internal clock, as we shall see in Section 10.4. Circadian rhythms have been demonstrated in all opisthobranch species and all pulmonate species that have so far been examined. Behavioral activity is usually recorded automatically using either timed video or switching devices that are triggered by the animal's movements. While the activity patterns of terrestrial gastropods tend to be strongly phase locked to the day-night cycle, those of opisthobranchs tend to be more loosely constrained. In particular, some sea hares are at least partially active in both the day and the night. For example, Aplysia parvula is continuously active during the daytime, but it is also active about 40% of the time at night (Carefoot, 1989). Aplysia fasciata, observed in Israel, shows a similar pattern of activity, but with reversed phase. It is mostly active at night, but 60% of the animals can be found copulating in the late afternoon (Susswein et al., 1983). The genus Aplysia is an interesting group because some species (e.g., A. californica) are diurnal whereas most others (e.g., A. dactylomela,

A. brasiliana, A. vaccaria, and A. Juliana are nocturnal. In Hawaii, where A. dactylomela and A. parvula are sympatric, the former is nocturnal but the latter is (mostly) diurnal (Carefoot, 1989). Although Carefoot looked for differences in ecological niches to explain the different activity patterns, he found none. Nor did he find that the two species interacted with one another, thus excluding competition as a factor dictating the out-of-synch circadian patterns. He suggests that the energy requirements of these tropical species requires them to divide their day into periods of rest and activity, but the assignment of the active period to either night or day is arbitrary. Further, when he compared his own observations of A. dactylomela and A. parvula with those of Susswein et al. (1983) on A. fasciata, he noticed that all three species appear to exhibit a similar circadian sequence of behaviors. In all three species, each behavioral cycle begins with a period dominated by feeding; next there is a period of rest (and digestion), and then a period of mating, after which the cycle repeats. Plaut (2000) has a different explanation for the nocturnalism of A. oculifera, which he studied in Israel. This animal lives in shallow rocky shores. Plaut found that the absence of daytime activity is dependent on the animals being exposed to strong sunlight. When put in a shaded environment, 60% of the animals fed during the day, compared to only about 16% that fed during the day under strong sunlight. Although these observations are sketchy, they suggest firstly that the circadian constraint on activity in A. oculifera is weak, and secondly that whatever nocturnalism is present in this species is determined by the avoidance of ultraviolet radiation from sunlight.

256 Behavior and Neurons in Gastropods

Terrestrial gastropod species have nocturnal habits. Slugs, especially, have activity patterns that are strongly phase dependent. This is explained by the fact that slugs are vulnerable to predation and desiccation, both of which are more likely to occur in the daytime. Temperature alone can entrain the circadian rhythm in slugs, as shown by a laboratory experiment conducted in constant darkness (Ford and Cook, 1987). In this experiment, the temperature-entrained rhythm was recognized and distinguished from the light-entrained rhythm because the former had a period length of 23 hours, while the latter had a period length >24 hours. When temperature changes and light changes are synchronized, as they are in nature, the combined effects of the two cues cause a slug's behavior to be almost entirely confined to periods of cool darkness. Experimental manipulations of these variables produce a rapid entrainment of activity patterns, as shown in Figure 10.2. The situation in regard to coastal "prosobranchs" is special because the activity of these animals is strongly influenced not only by light and temperature, but also by tides. To tease out the separate influences of circadian rhythms and tidal rhythms, Zann (1973) conducted laboratory experiments with the intertidal snail Melanerita atramentosa in New South Wales, Australia. He used a tide simulator together with controlled regimes of light and temperature. By setting external conditions constant for either the tidal cues or the circadian cues, he was able to show that the snails have two separate internal rhythms,

Figure 10.2. Entrainment of the activity rhythm of Limax pseudoflavus by light and temperature. Locomotion of a single slug was monitored by time-lapse photography in a controlled environment. The data are double plotted to show circadian patterns. Each line represents 48 hours; the second 24 hours in one line is repeated as the first 24 hours in the next line. For the first 13 days the slug's activity was strongly constrained by the 23-hour cycle of 7 hours light at 24°C and 16 hours dark at 13°C. On day 14, the photoperiod was changed to 12:12 light:dark with a constant 13°C temperature, and the slug's activity quickly became entrained to the new regime. From Ford and Cook (1987) with permission.

The Temporal Organization of Behavior

257

one running on a circadian basis and the other on a lunar basis. In a free-running condition, when light and temperature were kept constant while the tidal rhythm continued, the snails' activity was greater during the projected nocturnal high water than during the projected diurnal high water. There was little nocturnal activity unrelated to the tides. These experiments illustrate how tidal rhythms and circadian rhythms interact in intertidal snails. The timing of activity, mostly foraging, in relation to tides and sunlight is highly variable among intertidal species, and evidently dependent on a number of factors, including the possibility of desiccation and vulnerability to wave action and predation. Irrespective of tidal preferences, some species are primarily active at night, whereas other species are primarily active during the day. Similarly, tidal preferences can be independent of circadian preferences. For example, the high-shore species of pulmonate limpets, Siphonaria, are active mainly when exposed by low tides, whereas the low-shore species are mainly active when awash during ebb and/or flood tides (Hodgson, 1999). Even different populations of the same species can have different patterns of activity. Thus, "prosobranch" limpets, Patella vulgata, forage for food during daytime high tides at the Isle of Man, whereas at Alderney they forage during the night at low tides (Hartnoll and Wright, 1977). In both cases the animals avoid the daytime low tides, but why the pattern of behavior differs at the two locations is uncertain. A prominent feature of circadian behavior in limpets is their return to specific resting sites, or "homes," after foraging. This behavior, known since the time of Aristotle, is exhibited by both "prosobranch" species (Acmaea, Patella] and pulmonate species (Siphonaria). Individuals return to homes that have been partially excavated on rock surfaces. Because the small depression exactly fits the limpet's body, the animal is protected from desiccation when it sits tightly in its home. Obviously, the best protection is provided by sites that exactly fit an individual's own body, hence the adaptive function of homing. Site fidelity can be very high, as shown by the fact that some individual's can occupy the same home scar for 3 years (see Hodgson, 1999). In another study, with a different species, >90% of marked individuals remained on the same scar for a period of 7 months. To investigate the mechanism of homing in the limpet Patella, Cook et al. (1969) observed the excursions of 174 individuals as they traveled away from their homes. They found that 468 of 469 excursions were completed by successful homing. After systematically investigating all plausible mechanisms that could explain the homing ability, Cook et al. concluded that the animals probably rely on polarized chemical cues, particularly those contained in slime trails. However, since the limpets still homed even after they had been displaced from their outward trails, it is not the case that they simply return home by the same route taken when they left. Perhaps in such cases the limpets used persistent cues laid down during previous excursions, but the investigators were unable to remove all cues by scrubbing the rock with oven cleaner, which leaves in doubt exactly what cues are used.

258 Behavior and Neurons in Gastropods

Similar experiments have been performed with pulmonate limpets, again with equivocal results (Hodgson, 1999). Terrestrial gastropods return to homes at the end of every circadian cycle (i.e., in the late night). Although homing by terrestrial gastropods is not as precise as by limpets, its function is essentially the same, namely the avoidance of desiccation and the return to a sheltered place in which they survived the previous day. Also, chemical cues are again the sensory basis of homing behavior. Figure 10.3A shows the tracks made by six slugs, Limax pseudoflavus, traveling between their home brick and a feeding station. The use of slime trails as cues for homing is suggested by the fact that certain pathways are repeatedly followed. As a control for the possibility that the slugs in Figure 10.3A may have followed physical features in the test box, rather

Figure 10.3. Homing behavior in the slug Limax pseudoflavus. (A) The superimposed tracks of six slugs are shown relative to the feeding station and home, a hollowed-out brick. Dots represent the corners of tiles upon which the animals crawled. All the tracks shown at the extreme right are in the floor/wall angle; they are drawn separately for clarity. (B) After one slug had reached the food from its home (dotted line), the home was removed and the tiles nearest the right wall were rotated 180° so that the previously laid mucus tracks would no longer lie in the wall crease. The slug now attempted to reach home using the mucus tracks that lay at the left edges of the tiles. From Cook (1979) with permission.

The Temporal Organization of Behavior 259

than slime trails, the previously laid trails were moved in Figure 10.3B, but the slugs still spent most of their time crawling over the slime trails. On the other hand, if slugs are introduced into a clean arena containing a previously occupied home, they travel to the home quite directly even in the absence of established trails. It has also been noted that, when air currents are present, slugs typically leave home in a downwind direction and return home in an upwind direction, not necessarily following slime trails. Taken together, the foregoing observations indicate that slugs reach home using redundant chemical cues, that is, cues from distant sources (the home itself) as well as from proximal sources (usually slime trails) (Cook, 1979). Terrestrial snails also appear to home by using a combination of airborne chemical cues and chemical cues contacted in the substrate (Chase et al., 1978). Unlike limpets, snails and slugs do not home to a precise site, and several individuals typically occupy the same shelter. This raises the question whether they recognize their own individual chemical signal or only a species-specific signal. Although there has been some disagreement on this issue, most of the data point to the use of the nearest or best marked shelter, provided that it is marked by species-specific scents (Tomiyama, 1992).

10.3.

Endogenous Circadian Clocks

Internal time-keeping mechanisms, or circadian clocks, allow animals to organize the timing of their activities in relation to exogenous factors. With an internal clock the animal can anticipate the daily changes in light and temperature, and it can also program different kinds of activities, for example, feeding and mating, at different phases of the daily cycle. Different behaviors may even be expressed in rhythms of different lengths (Ford and Cook, 1987). Confirmation of an internal clock requires satisfaction of certain experimental tests. Principal among these is the demonstration of a freerunning rhythm in constant environmental conditions. Additional evidence for an internal rhythm is provided by proof that the behavior can be entrained by sensory cues ("Zeitgebers") that are phase shifted relative to the earth's day-night cycle. The most common zeitgebers used for entrainment are light and temperature (Fig. 10.2). A final test of the circadian clock is to show that it is temperature compensated. Free-running behavioral rhythms have been reported in several gastropod genera including Aplysia, Helix, Deroceras, Arion, and Bulinus. Entrainment (Fig. 10.2) has also been shown for many of these same species. However, only in the slug Limax maximus have all three criteria for the endogenous control of a behavioral rhythm been demonstrated. The temperature compensation of the circadian rhythm in Limax is striking, especially since it is known that temperature is a major factor controlling the slug's activity. In the experiments of Sokolove et al. (1977), locomotion and temperature compensation were measured in free-run conditions (i.e., in constant darkness). After 18 days at 11.5 °C, the slugs had clearly established a free-running

260 Behavior and Neurons in Gastropods

rhythm. When the temperature was then raised to 21.5°C, the slugs became less active, but the timing of their activity periods remained virtually unchanged; the mean g10 calculated for 16 snails was 0.99 ±0.01 (SEM). The authors contrast this result with the gio for pedal wave frequency in Limax, which other workers had reported as 2.1 over the same range of temperatures. Any internal oscillator that might be capable of controlling the behavioral circadian rhythm must possess the same essential features as are evident in the behavior itself, namely free-running periodicity, entrainment by phase offset or phase-reversed Zeitgebers, and temperature compensation. By all such criteria, the opisthobranch eye is a circadian clock. This was discovered by Jacklet (1969), who first reported that the isolated eye expresses a circadian rhythm of neuronal activity when maintained in culture. Most studies of the ocular clock have been conducted in two species, the sea hare Aplysia californica and the "bubble" snail Bulla gouldiana (Jacklet, 1989; Block et al., 1993), in which the clock's output can be recorded electrophysiologically using a suction electrode attached to the optic nerve. The signal is a compound action potentials (CAPs) that results when axons in the nerve fire in synchrony. On a scale of seconds, the CAPs may occur as an irregular pattern (Bulla), or as a pacemaking or bursting pattern (Aplysia). Regardless of the small-scale pattern, when the total number of spikes is measured over days under constant conditions of light and temperature, there is a strong circadian rhythmicity. Even when the eye and its nerve are surgically removed from the animal and maintained in culture, the same rhythm persists for 1 week or more. Figure 10.4 illustrates the remarkable steadiness of the rhythms that can be recorded in vitro in constant darkness. If animals are exposed in vivo to a light-dark cycle in which the two phases are of equal length (L: D 12:12), the spike activity recorded in vitro accelerates just before the projected dawn and peaks just after the projected dawn, as shown in Figure 10.4. This pattern indicates an effective entrainment by the light-dark cycle prior to dissection. The exact period length of the free-running rhythm depends on the composition of the culture medium, but it is generally about 24 hours. If a light-dark regime different from that experienced by the animal prior to dissection is imposed on the eye in vitro, the phase of the rhythm is shifted (i.e., the rhythm can be entrained in vitro). As for temperature compensation, the Q\0 of the ocular circadian rhythm in Aplysia is 1.02-1.07 over the range 12-22 °C, and the Qlo for Bulla is 0.98 over the range 12-25 °C. Thus, the eye meets all necessary tests as a bona fide circadian clock. Is the ocular pacemaker responsible for the circadian rhythm of locomotion? A prerequisite for such control is that the optic nerve fibers interact either directly or indirectly with central neurons controlling locomotion, and that increases in CAP frequency bias the neural control system towards expression of the behavior. An anatomical substrate consistent with this scenario has been demonstrated in Aplysia by Olson and Jacklet (1985), who visualized the optic nerve projection into the CNS by labeling its

The Temporal Organization of Behavior

261

Figure 10.4. Free-running rhythms of compound action potentials recorded in vitro from the optic nerves of isolated eyes, in two opisthobranch species. Spontaneous discharges were recorded in constant darkness at 15°C. Prior to dissection the animals were exposed to a cycle of 12 hours light, 12 hours dark. Here the 24-hour lines mark the time of the projected dawn. Note that an increase in activity precedes the dawn even in the first cycle. Adapted from Block et al. (1993).

fibers with either horseradish peroxidase or [3H]leucine. They found that afferents from the eye are widely distributed in the cerebral ganglia, the pedal ganglia and pleural ganglia, which happen to be the same ganglia that contain the trigger neurons and the central pattern generator for locomotion (see Section 6.2). Whether there are also projections into the abdominal ganglion could not be determined with the methods used. However, the projections from each eye were shown to be strongly bilateral, suggesting that the projections are overlapping and that together the two eyes exercise a redundant influence over CNS circuits. While the foregoing results indicate the presence of anatomical pathways by which the ocular pacemaker could influence locomotor circuits, no physiological evidence for such an interaction has been reported. Furthermore, if there is an interaction, it must be different in Aplysia and Bulla because, while the ocular rhythms are very similar in the two animals (Fig. 10.4), the behavioral rhythms are phase shifted; Aplysia is diurnal but Bulla is nocturnal. Thus, one would expect the optic nerve discharge to excite the locomotor circuit in Aplysia, but inhibit it in Bulla. An important test of the postulated behavioral role of the ocular pacemaker is to look for changes in behavior after removal of the eyes. Surprisingly, eyeless Aplysia still respond to the onset of light by increasing their behavioral activity (Lickey et al., 1977). In contrast to intact animals, eyeless animals do not anticipate the dawn, and their circadian rhythms are

262 Behavior and Neurons in Gastropods

blunted, but many animals nonetheless exhibit a rhythm. Some eyeless animals even have free-running behavioral rhythms in constant darkness. Overall, 39% of the eyeless animals in the study of Lickey et al. were judged to have recognizable free-running rhythms. The authors dismissed the possible influence of uncontrolled exogenous cues in triggering the freerunning rhythms because sometimes two animals living in the same aquarium exhibited free-running rhythms of different periods. Taken together, these results indicate the existence of an endogenous oscillator other than the eye. Extraocular photoreceptors are present in the rhinophores and tentacles of Aplysia (Chase, 1979), and perhaps in other tissues, but no free-running rhythm has yet been detected anywhere but in the eyes. Some early experiments led investigators to suspect that the bursting cell R15, and its putative homologue in Otala, cell 11, were endogenous oscillators (see Gainer, 1972). However, subsequent studies showed that the circadian rhythm of the spontaneous activity recorded in vitro is weak at best, and that whatever rhythm exists is probably caused by the trauma of dissection, not by environmental Zeitgebers. In any case, R15 is silent in vivo (Alevizos et al., 199la). Neuronal correlates of the circadian rhythm have been reported in land snails, but no pacemaker has been identified. In one study, using Helix aspersa maxima, the neurosecretory material contained within neurons was measured after staining with aldehyde fuchsin (Attia et al., 1998). The intensity of staining in one cell cluster, located near the cerebral commissure ("cerebral green cells"), was found to depend on the time of day at which the staining was done, and it was generally correlated with the amount of behavioral activity recorded at the same times. Thus, both neurosecretory staining and behavioral activity were greatest at the beginning of the nocturnal phase, but whether this indicates an accumulation of the secretory products during the daytime, or an increase in their production at night, was not determined. In another investigation, in Helix pomatia, the chemical sensitivity of the lip region was tested at different times of the day by means of electrical recordings from peripheral nerves (Voss et al., 1997). The snails were maintained on a cycle of L: D 12:12 until they were dissected and prepared for continuous electrophysiological recordings with constant lighting. At 2-hour intervals, either sucrose or salt was applied to the lips and the neural responses were measured. It was found that the spike responses to constant strength stimuli were greater during the projected dark phase than during the projected light phase. This result indicates an endogenous regulatory process that is consistent with the nocturnal habits of Helix. One interpretation is that the chemosensory system is down-regulated during "off' hours to allow cellular restoration in receptor cells. A similar process might be present in Aplysia, where nerve fibers from the ocular pacemakers project into anterior peripheral nerves that are homologous to the lip nerves of Helix (Olson and Jacklet, 1985). This anatomical projection could form part of a mechanism to regulate chemosensory organs in the head and tentacles.

The Temporal Organization of Behavior 263 10.4.

Mechanisms of Oscillation and Entrainment in the Eye

Considerable progress has been made in understanding the mechanisms responsible for the circadian oscillation of CAPs in opisthobranch eyes. The work has been facilitated by the availability of two model species, Aplysia and Bulla, each with its own experimental advantages, and the spur of competition provided by alternative models cannot be ignored. Although details of the clock mechanism differ in the two species, the present account emphasizes the common elements as discussed by Jacklet (1989) and Block et al. (1993). The pacemaker neurons, called basal retinal neurons (BRNs) in Bulla but simply pacemakers in Aplysia, have already been described in Section 3.4 and illustrated in Figure 3.5. They were first identified as the source of the CAP rhythm from experiments in which other cell types were surgically removed. While the cell bodies of the pacemaker neurons undergo changes in membrane potential, impulses are initiated only distally, probably in the initial portion of the optic nerve. Intracellular recordings show that cells situated within small isolated fragments of the whole pacemaker population still retain a circadian rhythm. The ability of individual pacemaker cells to generate a circadian rhythm has also been demonstrated by recording from dissociated cells in primary culture. In the initial study (Michel et al., 1993), membrane conductance was recorded throughout two circadian cycles in constant dim light. It was found that conductance varied spontaneously, with the highest conductances recorded in the projected late night. Conductances decreased approximately two-fold near the projected dawn. A voltage-sensitive potassium channel is probably involved in these conductance changes since a similar circadian variation in a voltage-gated potassium current (!KV) was observed by Barnes and Jacklet (1997), again in dissociated Aplysia retinal pacemaker cells. Currents were significantly greater in the pre-dawn period than in the post-dawn period. Three other ion currents, including the IA current, showed no circadian variation. These results indicate that hyperpolarization during the night, hence less spiking, could be caused by a clock-driven increase in potassium current. Even so, IKv is probably not the only current regulated by the circadian oscillator, at least in Aplysia, because it is mainly activated at membrane potentials more positive than -20 mV (Barnes and Jacklet, 1997). Block and colleagues believe that changes in membrane conductance are not part of the clock mechanism per se, even though they may be important for its expression (see Block et al., 1993). Their conclusion is largely based on the observation that stabilizing individual ionic conductances does not affect the underlying circadian rhythmicity. On the other hand, treatments with inhibitors of protein transcription and protein translation do affect the clock in a phase-dependent manner (Khalsa et al., 1996). Block and colleagues therefore propose a two-loop model for the BRN pacemaker. An intracellular loop, represented in Figure 10.5 by the reciprocal oscillation of unidentified components D and E, is responsible for the rhythmic synthesis of critical proteins.

264 Behavior and Neurons in Gastropods

Figure 10.5. A cellular model of the ocular pacemaker. An intracellular loop (D/E) is coupled to a transmembrane loop. According to the model, the circadian periodicity is generated by a rhythm of transcription and translation. It is then "read out" by changes in membrane conductance that cause changes in membrane potential and the production of action potentials. C, D, E, and F represent unidentified components. Light is able to entrain the rhythm through its depolarizing effect; the resulting increase in intracellular calcium resets the intracellular rhythm. Adapted from Block et al. (1993).

Cyclical variations in the levels of these proteins are translated into variations in potassium conductance by another unidentified process, F. The changes in conductance are part of the model's second loop, the transmembrane loop. While the changes in potassium conductance drive the rhythm in membrane potential and are ultimately responsible for the frequency of action potentials, they also initiate a feedback to the intracellular loop. The elegance of the two-loop model (Fig. 10.5) lies in its ability to explain a group of experimental results that are typically plotted as phase-response curves. For example, the curves shown in Figure 10.6 are derived from experiments that use pulse manipulations to mimic the effects of environmental stimuli that entrain circadian rhythms. It is apparent that the phase in which the manipulation is delivered is an important determinant of its effectiveness as a phase shifter. Also implicit in Figure 10.6 is the conclusion that membrane depolarization is central to the mechanism producing phase shifts because injections of depolarizing and hyperpolarizing currents produce opposite effects, and manipulations that affect membrane potential indirectly produce results consistent with the direction of the inferred change in membrane potential. Also, phase shifts from light pulses can be blocked by blocking membrane depolarization. These results can be explained by

The Temporal Organization of Behavior 265

Figure 10.6. Phase-response curves drawn from multiple studies in Bulla and Aplysia. These results indicate that membrane potential changes are crucial for the phase shifts that accompany entrainment to new light cycles (see Fig. 10.2). Hyperpolarization, or treatments that cause hyperpolarization, produce phase shifts opposite to those induced by depolarization or treatments that cause depolarization. Adapted from Block et al. (1993).

assuming that the transmembrane loop (Fig. 10.5) serves both as an output pathway for the oscillation of protein synthesis and as an input pathway for the perturbing pulse. Calcium is seen as an important mediator in the transmembrane loop because manipulations that prevent calcium fluxes also prevent phase shifts. While the rhythm of the ocular pacemaker is clearly entrained by light and other environmental stimuli, the rhythm may be modulated by internally

266 Behavior and Neurons in Gastropods

generated signals. The types of internal influence appear to differ in Aplysia and Bulla. In Bulla, the activities of the two eyes are coupled. When the CAP rhythms were recorded in vivo from an animal kept in constant darkness, the phases of the two eyes remained in almost exact register for at least 10 weeks (Roberts and Block, 1985). However, if the cerebral commissure was severed, thus disconnecting the eyes, the phases gradually drifted apart. In another type of experiment, phase shifts were selectively induced in one eye, for example, by a light pulse; a short while later it was noticed that the second eye had also shifted, even though the second eye had not received any light pulse. From physiological experiments, it was learned that electrical activity in efferent fibers of one eye causes depolarization of the BRNs in the second eye, but no other retinal cell type seems to be affected. This result is consistent with idea that the membrane potential of BRN cells plays a critical role in mediating rhythm generation, as postulated by the model summarized above. In contrast to Bulla, the ocular pacemakers of Aplysia are only weakly coupled, if at all. In Aplysia, a different type of internal influence on the ocular pacemaker has been described. In this case, the modulatory signals originate in the cerebral ganglion, not in the eyes. The influence of the brain is best demonstrated by phase resetting experiments conducted in vitro. When the eye is still attached to the brain, the ocular rhythm is reset by light pulses of 12 hours or longer that delay the onset of darkness. However, when the eye is disconnected from the brain, the resetting effect is greatly attenuated. For example, a light pulse of 18 hours ordinarily resets the clock by about 6 hours, but if the eye is denervated, the effect is negligible. Serotonin is implicated as the agent of modulation in these experiments because serotonin is present in the centrifugal pathway and applications of exogenous serotonin mimic the resetting effects of neural activity (Nadakavukaren et al., 1986). Also, when applied in brief pulses serotonin is a potent agent for phaseshifting the CAP rhythm. While serotonin and light have opposite effects when used separately (Fig. 10.6), when used together their effects interact (Nadakavukaren et al., 1986). One idea is that endogenous serotonin modulates the effects of light, and that perhaps one function is to accommodate the circadian rhythm to the seasonal rhythm (Block et al., 1993). Just as the interocular signal in Bulla impinges directly on the pacemaker cells, so too in Aplysia, it appears that serotonin directly modulates I K v in pacemaker cells (Barnes and Jacklet, 1997).

10.5.

Multiple Influences on the Selection of Behaviors

The circadian clock is not alone in influencing the level of behavioral activity that is expressed in any given hour. Some of the non-clock influences are external, mainly having to do with weather, while others are internal, involving changes in motivation. Together, these factors affect not only the general level of activity, but also the specific types of behaviors that are expressed.

The Temporal Organization of Behavior

267

Ultimately, information from the clock, from external elements and from the internal state must be integrated to yield an observable sequence of behaviors. Terrestrial species are particularly affected by environmental conditions. Even during spring and summer, when activity is greatest, daily fluctuations in weather strongly influence the level of activity in snails and slugs. Activity occurs at different temperatures in different locations, even for the same species, suggesting that the animals are capable of acclimatizing to local conditions. Changes in temperature, in either direction, often accelerates activity, and very bright sunlight inhibits activity regardless of the temperature. However, the single most important variable influencing the activity of terrestrial gastropods is undoubtedly moisture. Although they have nocturnal habits largely to avoid daytime evaporative water loss, even the expression of a circadian rhythm is dependent on a minimal level of humidity. For example, the African snail Achatina fulica ceases to be active when humidity falls below 50%, whereas high levels of humidity facilitate activity (Takeda and Ozaki, 1986). Although Achatina normally rests during the day, rainfall initiates activity, and sustained levels of environmental moisture can cause the snails to become almost constantly active with little circadian rhythmicity. Similar effects are observed in other species. Since many neurons are excited by hypotonic saline solutions (Hasegawa and Takeda, 1986), it is attractive to hypothesize that the dilution of blood by the uptake of external water is the direct cause of increased behavioral activity (Takeda and Ozaki, 1986). However, the electrophysiolgical data are not entirely consistent; while some neurons are excited by hyposmotic solutions, others are inhibited. Whether key neurons necessary for behavioral arousal are excited by blood hydration and whether excitation through sensory pathways also plays a role, are questions that remain to be answered. Besides water and temperature, other stimuli have also been discovered to activate gastropods. Kupfermann and Weiss (1981) found, in Aplysia, that animals become more responsive to food stimuli after they have been handled or after they have had their tails pinched. A similar effect has been observed in Helix, but not in Lymnaea (Adamo and Chase, 1991b). Kupfermann and Weiss (1981) supposed that handling arouses the animal and has a non-specific effect in facilitating the expression of behaviors. Since it was known that serotonin plays a role in facilitating certain individual behaviors such as feeding, defensive withdrawals, and cardiovascular responses, Kupfermann, Weiss, and colleagues proposed that a group of serotoninsecreting neurons could constitute a central arousal system capable of facilitating multiple behaviors. Palovcik et al. (1982) tested this idea in Aplysia and Pleurobranchaea by examining the effects of serotonin injections on the occurrence of spontaneous behaviors. They found that serotonin increased the frequency of many types of behaviors ranging from movements of the lips or the mantle to increases in locomotion. In general, the injected animals became more alert and more active. Similar effects of serotonin have been reported in other species (Clione, Sakharov, 1991; Helix, Adamo and Chase, 1991b).

268 Behavior and Neurons in Gastropods

It is tempting to imagine that a circadian clock, perhaps the ocular clock, regulates serotonin levels in the body and that the daily transition from the inactive state to the active state is caused by serotonergic excitation. One version of this hypothesis would have serotonin acting as a hormone. Indeed, measurements of blood serotonin levels in Aplysia californica reveal a circadian variation in serotonin concentrations (Levenson et al., 1999). However, contrary to predictions, blood levels are lowest during the day, when activity levels are highest. Also, animals kept in constant darkness show little or no circadian variation of serotonin blood levels, whereas behavioral activity levels continue to vary according to the projected day-night cycle. These results make it unlikely that the circadian rhythm of activity in Aplysia is mediated by humoral serotonin, but it remains possible that there is a circadian variation in the synaptic release of serotonin. The idea of arousal has been interpreted somewhat differently, and more controversially, by some authors who postulate that all goal-directed behaviors, but especially feeding and mating, are affected by a common motivational system. Evidence for this is seen in the apparent ability of certain stimuli simultaneously to affect both feeding and mating. Common arousal could explain why, at times, an animal can be equally likely to express either one of two goal-oriented behaviors. However, serotonin is not likely involved in any such common arousal system because, in Helix, serotonin injections increase the general level of activity, but they do not increase either feeding or mating (Adamo and Chase, 1991b). In Aplysia, injections of serotonin actually decrease feeding latencies (Levenson et al., 1999). Other evidence, not linked to any particular mechanism, supports at least a limited version of the common arousal hypothesis, in so far as it indicates interactions between motivated behaviors. For example, Susswein et al. (1984) observed that the head waving that characterizes the appetitive phase of feeding in Aplysia is indistinguishable from the head waving that precedes mating. It has also been observed that animals in pursuit of prospective mates, and hence judged to be sexually motivated, will readily take a piece of food when it is presented to them, thus manifesting a motivation to feed (Susswein et al., 1983, 1984). Subsequently, it was discovered that a pheromone, released by Aplysia during mating, stimulates feeding in conspecifics (Botzer et al., 1991; Blumberg and Susswein, 1998). Certain observations in Helix aspersa suggest that arousal in respect to one goal can readily transfer to another goal. It can be observed, for example, that exposure to food stimuli arouses snails to mate. Thus, snails that are fed after a period of starvation mate much more readily than do snails that have been given unlimited access to food (Adamo and Chase, 199la). A related phenomenon can be seen in snails that have been subjected to both starvation and social isolation. If the snails are taken from isolation and placed in a group together with several pieces of carrot, the animals quickly start feeding on the carrot, but their genitals become everted as they feed, just as if they were courting. Several animals may be seen feeding on the same piece of

The Temporal Organization of Behavior 269

carrot, each with everted genitals, yet none is showing any apparent interest in the other snails. Sexual arousal can become so great in these animals that we have seen them shoot their darts, not at other snails, but into the carrot. Only after the animals have eaten sufficiently and the carrot is removed, will mating pairs form. Because food is so effective in stimulating sex in Helix, we routinely use this protocol to facilitate mating in experiments on reproduction (Rogers and Chase, 2001). The interaction between food stimuli and sexual behavior is quite different in Aplysia. In this species, mating is actually inhibited when an animal is exposed to food stimuli, even if it does not ingest anything (Nedvetski et al., 1998). Taking into account a number of observations on the interactions between stimuli and goal-oriented behaviors (feeding and mating), it is apparent that the interactions are opposite in Aplysia and Helix, as shown in Figure 10.7. The reasons for these mirrored effects are not clear, but it is noteworthy that the two species have very different time budgets, with mating occupying much more of the animal's total time in Aplysia than in Helix. In any case, the summary of effects presented in Figure 10.7 provides little support for the idea of a common arousal system because in both species there are stimuli that increase the motivation for one behavior while

Figure 10.7. Contrasting interactions between stimuli and goal-directed behaviors in Aplysia and Helix. Goal-directed behaviors are usually initiated (downward arrows) when an object specific to one goal is present by itself. Such goal-directed behaviors may be either facilitated (diagonal arrows) or inhibited (diagonal circles) by the presence of a second goal-specific object. Differences between the species may be related to the fact that Aplysia spends much more time in mating activities than does Helix. Based on Adamo and Chase (1991a,b), Blumberg and Susswein (1998), Nedvetzki et al. (1998), and unpublished observations.

270 Behavior and Neurons in Gastropods

decreasing the motivation for another. When Susswein's group found that food stimuli and conspecific stimuli have dissimilar effects on sexual motivation (Fig. 10.7; Nedvetzki et al., 1998), even they abandoned the idea of common arousal in Aplysia, despite having championed it for many years. Direct interactions between arousal mechanisms intrinsic to specific neural circuits explains the data of behavioral interactions at least as well as does the ill-defined notion of common arousal. For example, in Pleurobranchaea californica, Jing and Gillette (2000) found that certain serotonergic neurons in the circuit for escape swimming induce arousal not only for escape swimming but also for feeding after escape. The diagram in Figure 10.8 shows that certain serotonergic neurons within the escape swim system synapse on other serotonergic neurons in the feeding circuit, in the latter case including the metacerebral giant cell (labeled as MCG in Fig. 10.8). This causes a mild excitation of feeding behavior for several minutes after escape swimming.

Figure 10.8. Synaptic interactions in Pleurobranchaea that are responsible for the immediate suppression of feeding during escape swimming, and the prolonged but mild excitation of feeding that follows escape swimming. The acute suppression of feeding is caused mainly by the connection from A-cil, an output neuron of the swim central pattern generator, to II, an inhibitory interneuron in the feeding network. The delayed arousal of feeding is attributed to activity in four serotonergic cells, As \-4. Escape swimming also facilitates crawling by exciting the G neurons, which are themselves serotonergic modulators in premotor circuits. From Jing and Gillette (2000) with permission.

The Temporal Organization of Behavior

271

Ethologists and psychologists have provided us with the concept of drive to explain the regulation of goal-directed behaviors. Essentially, the level of the drive for a particular behavior determines the likelihood that the behavior will be performed. Therefore, the behavior that an animal engages in at any particular time depends upon the relative levels of its various drives, which in turn depends largely on the degrees of satiation. Leonard and Lukowiak (1986) posited six drives for Aplysia, namely feeding, escape, female sex, male sex, and egg laying. Similar drives might be found in other gastropods, but the sex component must be different in genera other than Aplysia because some, for example Lymnaea, never have female sex drive (De Boer et al., 1996). Also, for species that mate as simultaneous reciprocal hermaphrodites, like Helix, the male and female drives are probably united. A good illustration of how drive state can influence behavioral expression is provided by experiments with Pleurobranchaea performed by Gillette et al. (2000). When tested with high concentrations of the feeding stimulant betaine (10~ 4 M or greater), most animals responded by everting their buccal mass and biting, but at lower concentrations the animals responded by turning away. Similarly, taurine was a potent aversive stimulus at high concentrations, but feeding responses could be elicited with low concentrations. Significantly, the response that any particular betaine or taurine stimulus elicited, whether feeding or avoidance, was largely dependent upon the degree of feeding satiation. Thus, hungry animals, which had lower feeding thresholds, showed more feeding responses to both betaine and taurine than did sated animals, whereas sated animals showed more avoidance responses, again regardless of which chemical was presented. The regulation of goal-directed behaviors has also been investigated by studying what happens when goal objects are eliminated from the animal's environment. In several such experiments performed by Susswein and colleagues, the time that Aplysia devoted to feeding and mating was measured under conditions in which there was either no food available or no conspecifics available. When food was removed, the animals spent a lot more time mating and laying eggs. On average, the total time devoted to all reproductive activities doubled, with the time devoted to mating increasing from 6.1 hours to 11.5 hours per day (Ziv et al., 1991b). Also, circadian patterns of activity were affected. Egg laying, which is ordinarily nocturnal, became equally frequent during the day as during the night, while the pattern of daytime mating became even more pronounced than usual. A different effect was observed after removal of potential mates because now, unexpectedly, the time spent in feeding actually decreased. Ziv et al. (199la) attributed this result to the loss of a pheromonal signal from mating conspecifics (Fig. 10.7; Botzer et al., 1991; Blumberg and Susswein, 1998). Similar to the case when food was removed, when mates were removed, the circadian pattern of activities was again disrupted, with feeding behavior occurring at a relatively constant rate during all hours of the day. These experiments need to be interpreted with caution because the removal of a goal object will have many consequences including making more time available for other behaviors, reducing sensory

272 Behavior and Neurons in Gastropods

stimulation and increasing drive for the unobtainable goal. Nevertheless, the results suggest that motivational states interact with circadian oscillators to determine the timing of behaviors, and the drive to mate has a greater influence on the temporal patterning of behavior than does the drive to feed.

10.6.

Singleness of Action

Early in the 20th century Charles Sherrington discovered that neurophysiological mechanisms protect animals from attempting to perform two contradictory behaviors at the same time (see Kovac and Davis, 1980a). He referred to this constraint as the "singleness of action." For the spinal reflexes studied by Sherrington it is obviously true that the same limb cannot be simultaneously flexed and extended. On the other hand, it is clearly possible to chew gum and walk at the same time. Similarly, there is no physical constraint that prevents Aplysia from simultaneously feeding and mating as a female, but the interesting fact is that Aplysia rarely does so (Susswein et al., 1983; Ziv et al., 1991a). Thus, there must be factors other than simple physical constraints that cause animals not to perform certain behaviors simultaneously. In the broader context, the singleness of action is part of the animal's organization of behavior. To understand why two behaviors are not performed at the same time is to gain insight into how the animal "decides" what sequence of behaviors it will perform. Gastropod molluscs have proven to be good models for identifying some of the neural mechanisms that operate to enforce the singleness of action, even when it involves complex behaviors. An attractive way to conceptualize the organization of behavior is to view it as a hierarchy, which can be defined as "the organization of unrelated acts of behavior with a 'priority sequence' in which a dominant behavior takes precedence over (i.e., partially or completely suppresses) a subordinate behavior" (Kovac and Davis, 1980a, p. 85). To this end, Davis and colleagues conducted a series of experiments with Pleurobranchaea californica to ascertain which behaviors are dominant over which others. A choice paradigm was used in which animals were simultaneously presented with two stimuli of "comparable" strengths, each of which was the trigger for a different behavioral act. For example, food would be presented to an animal that was lying on its back, thus simultaneously stimulating both feeding and righting. The results from these experiments can be interpreted to represent the animal's "decisions" in the face of competing stimuli. However, there are several caveats. For one, stimulus strengths are not easily equated, even when they are calibrated against response strengths, as generally done by Davis and colleagues. This is important because the dominance of one behavior over another is seldom absolute; if a stimulus is sufficiently strong, it can usually elicit a response even in a subordinate behavior. Also, some of the stimuli used by Davis et al. were not natural; for example, they used injections of CNS extract to elicit egg laying. In other experiments (e.g., with food and

The Temporal Organization of Behavior

273

potential mates), the stimuli were not presented strictly simultaneously. It must also be borne in mind that the dominance of one behavior over another is not stable, but instead it is variable depending on learning experiences and motivational states (see Adamo and Chase, 199la). Despite the qualifications noted above, the priority sequence that was constructed for Pleurobranchaea (Kovac and Davis, 1980a) is interesting in itself and even more useful for comparison with other gastropods. Escape swimming is the most dominant of all behaviors, and feeding is dominant over mating and righting. However, feeding and withdrawal are mutually suppressive, with the degree of dominance in each case dependent on relative stimulus strengths. Finally, egg laying is mutually compatible with both righting and mating. These relationships gain significance when compared to the behavioral hierarchies for other gastropods, even though somewhat different experimental protocols apply in each case. Contrasts in the relationships between feeding and mating are particularly interesting. Whereas feeding dominates mating in Pleurobranchaea, mating dominates feeding in Helix, Ercolania, and Aplysia (see Adamo and Chase, 199la). This pattern of preferences indicates that feeding dominates in carnivorous species (Pleurobranchaea, Tritonia), which may infrequently find food, while mating dominates in herbivorous species, which more often lack mates. The same reasoning applies to the relationship between feeding and withdrawal behaviors, where feeding is more dominant in carnivores (Pleurobranchaea, Clione) than in herbivores (Lymnaea, Helix}, probably because carnivores cannot afford to pass up a meal (see Inoue et al., 1996). No priority sequence has been constructed for Aplysia, but suppressive interactions between behaviors have been described (Leonard and Lukowiak, 1986). For example, egg laying suppresses feeding, as in Pleurobranchaea and many other species (see Section 8.6). Feeding, on the other hand, suppresses defensive behaviors including gill/siphon withdrawal, locomotion, and inking. Copulation in either the male or the female role also suppresses withdrawal reflexes, and copulation as a female suppresses feeding, as already noted. The mutual exclusion of feeding and mating as a male is probably dictated by physical constraints, but otherwise the interactions involve increases in the thresholds for the suppressed behavior. Animals mating as males do not simultaneously lay eggs, although animals mating as females often do; again, this difference may be due to physical constraints in the former case. In principle, the interactions between competing behaviors could be mediated either hormonally or neurally. The weight of current evidence suggests that neural pathways are mainly responsible for the types of behavioral interactions under consideration here, but hormonal influences may play a modulatory role. The suppression of feeding behavior during egg laying is an example of an interaction that was first thought to be mediated hormonally (i.e., by the egg laying hormone), but which was later shown to be mediated by neural feedback signaling the passage of eggs down the reproductive tract (Ter Maat and Ferguson, 1996; Section 8.6).

274 Behavior and Neurons in Gastropods

Some authors have suggested that serotonin could function as a hormone to select behaviors. While there is no doubt that serotonin is present in many neuronal circuits where it serves to either mediate or modulate individual behaviors, the idea that it might function as a hormone has only recently been seriously investigated. Levenson et al. (1999) found that blood levels of serotonin are indeed modulated in Aplysia, but the behavioral consequences of these changes are uncertain. It is curious, for example, that when blood serotonin levels were high, the latencies to feed were long, and conversely, when serotonin levels were low, latencies were short. This finding is at odds with earlier data showing that serotonergic neurons in the cerebral and buccal ganglia are capable of facilitating feeding behaviors (see Section 7.5). The results from Levenson et al. (1999) therefore imply that blood serotonin is either unable to reach postsynaptic targets in the feeding circuitry at sufficient concentrations to have a behavioral effect, or it is otherwise less effective than locally released serotonin. The relationship between circulating serotonin and locomotion is also perplexing. Whereas some authors have reported that injections of serotonin can induce locomotion (see below), Levenson et al. (1999) found that locomotion was negatively correlated with blood serotonin levels. In summary, while serotonin may act as a hormone to mediate the general state of behavioral activity, as discussed earlier, there is little evidence to suggest that it determines which specific behavior is performed at a given moment, or that it mediates interactions between competing behaviors. The multiplicity of synaptic transmitters has led some workers to suggest that each behavior has its own chemical signature. The putative role of APGWamide in the control of male mating behavior (see Section 8.4) is a good example. The peptide is present within the neurons that are known to control the behavior, and injections of APGWamide are specifically effective in eliciting penile eversion. In this case, APGWamide is thought to act as a neurotransmitter. A more radical view of how chemicals might mediate gastropod behavior is presented by Sakharov, mostly in Russian language publications (but see Sakharov, 1991). His idea is that neurotransmitters flow into an extracellular pool that is embedded within the neuropil. From there, each transmitter triggers not just one type of behavioral act, but several different ones, depending on its concentration. Mixtures of transmitters can produce still other behaviors. In this scheme, chemicals become more important than synaptic connections in determining behavioral output. It allows for the activation of some behaviors and the suppression of other behaviors by the same chemical message. To test his hypothesis, Sakharov injected either dopamine or serotonin into intact specimens of Clione; in some cases, precursors to these transmitters were used instead. He and his colleagues found that injections of serotonin caused animals to switch from a slow mode of swimming to a fast mode, caused them to extend their feeding cones and suppressed their withdrawal reflexes. Contrastingly, injections of dopamine inhibited slow swimming, facilitated withdrawal reflexes, and put the animals in a state of rest. These results are consistent with other evidence indicating that serotonin transmission mediates the initiation of locomotion

The Temporal Organization of Behavior

275

(see Chapter 6). They also suggest that serotonin and dopamine may have humoral roles in causing major switches of behavioral state. However, they are not sufficient to confirm Sakharov's general hypothesis about how chemicals select behaviors. The major role of synaptic inhibition in mediating the singleness of behavior is evident in the studies listed in Table 10.1. The table includes examples of different kinds of dominant and subordinate behaviors taken from several gastropod species. In all cases, it is shown that there is an inhibitory pathway from the neural circuit controlling the dominant behavior to the neural circuit controlling the subordinate behavior. A typical example is shown in Figure 10.8, illustrating the interaction between circuits controlling escape swimming and feeding in Pleurobranchaea. In this case, outputs from the circuit for escape swimming inhibit neurons in the circuit for feeding. It is not easy to demonstrate that cellular interactions of this sort are actually responsible for the singleness of behavior, but it is reasonable to think that when the dominant circuit is active, the subordinate circuit will be inhibited, with the result that the threshold for expression of the subordinate behavior will be increased. In Clione, where several inhibitory interactions between neuronal circuits have been discovered, the pattern of neural inhibition corresponds to the pattern of behavioral inhibition, as shown in Figure 10.9. The connections referred to in Table 10.1 apparently function solely to prevent the expression of incompatible behaviors. Often these connections are made between cells in different ganglia. In Pleurobranchaea, for example, there is a pathway from the feeding circuit, in the buccal ganglion, to the circuit that controls withdrawal of the oral veil, in the cerebral ganglion. A pair of identified buccal cells is responsible for at least 75% of the inhibition exerted through this pathway (Kovac and Davis, 1980b). These cells, named CDs for "corollary discharge," are thought to be both necessary and sufficient to produce inhibition of the withdrawal response during feeding. Intracellular stimulation of the cells in the absence of feeding causes inhibition of the withdrawal response (monitored in a peripheral nerve), whereas hyperpolarization of the cells during generation of the feeding motor pattern prevents inhibition. The pathway in this case is probably polysynaptic, but the inhibitory connections are monosynaptic in other cases (Inoue et al., 1996; Norekian and Satterlie, 1996). A striking example of inhibition is evident during operation of the feeding pattern generator in Lymnaea. Figure 10.10 shows that certain cells in the pedal ganglion are inhibited in a precise phase-dependent manner while the feeding CPG is active. Specifically, the pedal cells are strongly inhibited during the N2 phase, when rasping normally occurs in the intact animal (see Section 7.3). Cells that are affected in this way have been found in three distinct clusters of the pedal ganglion, namely the E, F, and M clusters. Cells in the F cluster, such as the one illustrated in Figure 10.10, innervate the body wall, the mantle column and the foot, and they are believed to be part of the neural system responsible for locomotion (Kyriadides and McCrohan, 1988). Thus, one interpretation of the physiological inhibition seen in these

Table 10.1. Central inhibitory connections between dominant systems and subordinate systems ensure that incompatible behaviors are not executed simultaneously Animal

Dominant Behavior > Subordinate Behavior

Pleurobranchaea

Feeding > whole-body withdrawal

Clione Aplysia Lymnaea

Feeding > whole-body withdrawal Feeding > withdrawal, secretion Whole-body withdrawal > respiration

Clione Pleurobranchaea Clione Clione

Whole-body withdrawal > swimming Escape swimming > feeding Escape swimming > feeding Startle (escape) > swimming

Helix

Mating > withdrawal

Level of Inhibition in Subordinate System

Reference

Command interneurons, motoneurons (presynaptic)? Interneurons, motoneurons Interneurons, motoneurons Central pattern generator, motoneurons, muscles? Interneurons motoneurons Command interneurons Interneurons, motoneurons Central pattern generator, motoneurons, modulatory neurons Command interneurons, modulatory neurons

Kovac and Davis (1980b) Norekian and Satterlie (1996) Teyke et al. (1990b) Inoue et al. (1996) Norekian and Satterlie (1996) Jing and Gillette (2000) Alania et al. (1999) Norekian (1997) Balaban and Chase (1990)

The Temporal Organization of Behavior 277

Figure 10.9. Physiological interactions between neural networks are consistent with behavioral observations of hierarchical relationships. Circles at the top represent identified neurons in Clione that elicit the behaviors shown at the bottom. Physiological interactions are indicated as excitatory (open triangles) or inhibitory (filled circles), monosynaptic (solid lines) or polysynaptic (dotted lines). Note that the pattern of inhibitory connections from left to right corresponds to the order of behavioral dominance, as indicated by the arrows left to right. From Norekian and Satterlie (1996) with permission.

neurons during feeding is that it prevents inappropriate body movements while the radula is in contact with the food. If such movements were allowed to occur, they could dislodge food from the mouth or otherwise disrupt the feeding action. The inhibitory signals typically originate in cells that are placed relatively "high" in the dominant circuit. A good example is the suppression of swimming by startle in Clione (Norekian, 1997). Startle responses occur in swimming animals whenever a looming threat is detected. Although the same muscles are used for swimming and startle, the motor pathways are independent. The startle response consists of just one or two wing power strokes which propel the animal about three times faster than even the fastest swimming rate. The neural control of startle resides, at the motoneuronal level, in two pairs of large motoneurons, each of which is responsible for flexing the wings either dorsally or ventrally. At a higher level, a pair of cerebral interneurons, Cr-St, is responsible for integrating sensory information and exciting the motoneurons with fast EPSPs of unusually high amplitude (up to 50 mV). While the Cr-St cells drive the motoneurons to generate the power strokes of the startle response, they simultaneously inhibit at least four identified cell types involved in the control of normal swimming, including motoneurons, interneurons, and modulatory neurons. Thus, the Cr-St cells simultaneously excite the startle motoneurons and inhibit the swim neurons, and both actions are mediated by acetylcholine. The circuitry allows for swimming to be temporarily suppressed so that a startle response can be rapidly inserted. Similarly, in Helix, the inhibition of withdrawal responses during mating can be traced to high-level command cells in the mesocerebrum. Although the

278 Behavior and Neurons in Gastropods

Figure 10.10. Cyclical inhibition of a neuron in the pedal ganglion of Lymnaea simultaneously with the excitation of two feeding motoneurons in the buccal ganglion. (A) The inhibited neuron is a member of the F cluster, which innervates the body wall, the mantle column and the foot. (B) The motoneurons Bl and B2 receive excitation during two phases of the feeding cycle, Nl and N2. Note that inhibition in PeF is delayed relative to the start of Nl, but rapid after the start of N2. Because inhibition is greatest during N2, when rasping normally occurs, it may serve to suppress body movements that would otherwise dislodge food from the mouth or interfere with effective feeding. Adapted from Kyriakides and McCrohan (1988) with permission of Company of Biologists, Ltd.

principal function of the mesocerebrum is to control motor activity in sexual organs (see Section 8.4.2), one output signal is directed to the parietal ganglion to inhibit withdrawal responses that might otherwise interfere with mating (Balaban and Chase, 1990). This action may be especially important for reducing the withdrawal responses that are evoked by receipt of the dart. A similar inhibition of defensive responses is seen in Aplysia during the appetitive phase of feeding. In this case, the inhibition is mediated by the cerebral neuron C-PR, and it appears to be part of the cell's overall function to command a state of feeding arousal (see Section 7.5.4;

The Temporal Organization of Behavior 279

Teyke et al., 1990b). A final example is illustrated in Figure 10.8, where it can be seen that activity in the swim circuit of Pleurobranchaea causes excitation of cell II in the feeding circuit. Cell II then inhibits the cells PCP and PSE, both of which are command cells for feeding (Jing and Gillette, 2000). One apparent exception to the rule that inhibitory interactions originate at high levels in neural circuits is the suppression of swimming, by withdrawal, in Clione (Norekian and Satterlie, 1996). Here, inhibition originates in the Pl-W cells, which are thought to be motoneurons (Fig. 10.9). However, the Pl-W cells are unusual in that they have central connections as well as extensive axonal projections to the periphery (see Section 9.3.1). These connections, plus the fact that the activation of a single Pl-W neuron can initiate the wholebody withdrawal response, endow the Pl-W cells with properties similar to those of interneuronal command cells. Subordinate behaviors are often suppressed at multiple sites involving several levels of control. The redundant nature of inhibition in subordinate systems is perhaps best illustrated by the inhibition of neurons that modulate, rather than directly control, the behavior. Two examples are listed in Table 10.1. One involves serotonergic neurons in the pedal ganglion that facilitate whole-body withdrawal behavior in Helix (Zakharov et al., 1995). It has been found that electrical stimulation of the mesocerebrum not only inhibits those neurons in the parietal ganglion that command withdrawal, it also inhibits certain serotonergic neurons in the pedal ganglion that facilitate withdrawal (Balaban and Chase, 1990). A second example concerns the facilitation of wing contractions in Clione by a group of serotonergic cells in the pedal ganglion. Here again, when the dominate circuit is active, in this case the circuit responsible for the startle response, the modulatory cells are inhibited (Norekian, 1997). One particular neuron is of special interest because it apparently serves a common functional role in numerous pulmonate and opisthobranch species. Alania et al. (1999) have called attention to the unique features of a pair of neurons, the P1B cells, that have somata in the pleural ganglia and that project axons to the buccal ganglia. When studied in seven species, the cells were found to have similar properties in all cases based on their morphology, FMRFamide-like immunoreactivity, electrotonic coupling and, especially, their ability to inhibit buccal neurons participating in the motor program for feeding. The P1B cells seem to be responsible for suppressing feeding during defensive responses. This is interesting because the suggested homology of the P1B cells extends to both herbivorous species (Helisoma, Planorbis, Lymnaed) and carnivorous species (Clione). Thus, all these species may use homologues of the same cell to suppress feeding during defensive responses, even though they have different modes of defense (i.e., the herbivores use withdrawal whereas the carnivores use escape locomotion). The suppression of subordinate behaviors by dominant behaviors can be likened to the modifications of behavior produced by learning and motivation in that all three processes involve changes in the "normal" sensory control of the behavior. Thus, regardless of whether feeding is suppressed by

280 Behavior and Neurons in Gastropods

dominance, avoidance learning or satiation, the end result is that food stimuli fail to elicit buccal movements. This raises the question whether the neural mechanisms of suppression are the same in each case. Only in the feeding system of Pleurobranchaea has this question been addressed in detail. While dominance, food avoidance learning, and feeding satiation all share one common mechanism in this system, which is the inhibition of the command cells PCP and PSE (Fig. 10.8), the suppression by swimming is greater than is the suppression by learning or satiation (Jing and Gillette, 2000). The difference may lie in the identified interneurons 12, which are silenced by inhibition during swimming but which continue to respond to food stimuli after avoidance learning or satiation. In fact, the tonic firing of the 12 cells in the latter instances cause the feeding oscillator to lock into the retraction phase, thus causing functional suppression by hyperexcitation rather than by synaptic inhibition.

References Adamo SA, Chase R (1988) Courtship and copulation in the terrestrial snail Helix aspersa. Can J Zool 66:1446-1453. Adamo SA, Chase R (199la) The interactions of courtship, feeding, and locomotion in the behavioral hierarchy of the snail Helix aspersa. Behav Neural Biol 55:1-18. Adamo SA, Chase R (1991b) "Central arousal" and sexual responsiveness in the snail, Helix aspersa. Behav Neural Biol 55:194-213. Alania MA, Panchin YV, Sakharov DA (1999) Pleural-buccal interneurons in the pteropod mollusc Clione limacina. J Comp Physiol A 185:267-275. Alevizos A, Bailey CH, Chen M, Koester J (1989a) Innervation of vascular and cardiac muscle of Aplysia by multimodal motoneuron L7. J Neurophysiol 61:1053-1063. Alevizos A, Weiss KR, Koester J (1989b) SCP-containing R20 neurons modulate respiratory pumping in Aplysia. J Neurosci 9:3058-3071. Alevizos A, Weiss KR, Koester J (199la) Synaptic actions of identified peptidergic neuron R15 in Aplysia. I. Activation of respiratory pumping. J Neurosci 11:1263-1274. Alevizos A, Weiss KR, Koester J (1991b) Synaptic actions of identified peptidergic neuron R15 in Aplysia. II. Activation of the large hermaphroditic duct. J Neurosci 11:1282-1290. Alkon DL (1980) Cellular analysis of a gastropod (Hermissenda crassicornis) model of associative learning. Biol Bull 159:505-560. Alkon DL, Sakakibara M, Forman R, Harrigan J, Lederhendler I, Farley J (1985) Reduction of two voltage-dependent K+ currents mediates retention of a learned association. Behav Neural Biol 44:278-300. Andrew RJ, Savage H (2000) Appetitive learning using visual conditioned stimuli in the pond snail, Lymnaea. Neurobiol Learn Mem 73:258-273. Angeloni L, Bradbury J (1999) Body size influences mating strategies in a simultaneously hermaphroditic sea slug, Aplysia vaccaria. Ethol Ecol Evolution 11:187-195. Antic S, Wuskell JP, Loew L, Zecevic D (2000) Functional profile of the giant metacerebral neuron of Helix aspersa: temporal and spatial dynamics of electrical activity in situ. J Physiol 527:55-69. Antonov I, Kandel ER, Hawkins RD (1999) The contribution of facilitation of monosynaptic PSPs to dishabituation and sensitization of the Aplysia siphon withdrawal reflex. J Neurosci 19:10438-10450. Antonov I, Antonova I, Kandel ER, Hawkins RD (2001) The contribution of activitydependent synaptic plasticity to classical conditioning in Aplysia. J Neurosci 21:6413-6422. 281

282

References

Arshavsky Yu.I, Beloozerova IN, Orlovsky GN, Panchin Yu.V, Pavlova GA (1985) Control of locomotion in marine mollusc Clione limacina. III. On the origin of locomotory rhythm. Exp Brain Res 58:273-282. Arshavsky Yu.I, Deliagina TG, Orlovsky GN, Panchin Yu.V, Pavlova GA, Popova LB (1986) Control of locomotion in marine mollusc Clione limacina. VI. Activity of isolated neurons of pedal ganglia. Exp Brain Res 63:106-112. Arshavsky Yu.I, Deliagina TG, Orlovsky GN, Panchin Yu.V (1988) Control of feeding movements in the freshwater snail Planar bis corneus. III. Organization of the feeding rhythm generator. Exp Brain Res 70:332-341. Arshavsky Yu.I, Orlovsky GN, Panchin Yu.V, Pavlova GA (1989) Control of locomotion in marine mollusc Clione limacina. VII. Reexamination of type 12 interneurons. Exp Brain Res 78:398^106. Arshavsky Yu.I, Deliagina TG, Gelfand IM, Orlovsky GN, Panchin Yu.V, Pavlova GA, Popova LB (1990) Neural control of heart beat in the pteropod mollusc Clione limacina: coordination of circulatory and locomotor systems. J Exp Biol 148:461-475. Arshavsky Yu.I, Deliagina TG, Orlovsky GN, Panchin Yu.V, Pavlova GA, Popova LB (1991) Locomotion of Clione limacina in relation to various types of behaviour. In Simpler Nervous Systems, Sakharov DA, Winlow W (eds), Manchester University Press, New York, pp. 290-315. Arshavsky Yu.I, Deliagina TG, Okshtein IL, Orlovksy GN, Panchin YV, Popova LB (1994) Defense reaction in the pond snail Planorbis corneus II. Central pattern generator. J Neurophysiol 71:891-897. Arvanitaki A, Chalazonitis N (1958) Configurations modales de 1'activite propre a different neurones d'un meme centre. J Physiol (Paris) 50:122-125. Asami T, Cowie RH, Ohbayashi K (1998) Evolution of mirror images by sexually asymmetric mating behavior in hermaphroditic snails. Am Nat 225-236. Aspey WP, Blankenship JE (1976) Aplysia behavioral biology: II. Induced burrowing in swimming A. brasiliana by burrowed conspecifics. Behav Biol 17:301-312. Attia J, Blanc A, Hassnaoui M, Rehailia M, Buisson B (1998) A neurosecretory phenomenon in some cerebral nervous cells of the land snail Helix aspersa maxima (Mollusca, Gastropoda), and its relationship to circadian locomotor and food-intake activities. Biol Rhythm Res 29:329-342. Audesirk TE (1977) Chemoreception in Aplysia californica. III. Evidence for pheromones influencing reproductive behavior. Behav Biol 20:235-243. Audesirk TE, Alexander JE Jr, Audesirk GJ, Moyer CM (1982) Rapid, nonaversive conditioning in a freshwater gastropod. I. Effects of age and motivation. Behav Neural Biol 36:379-390. Avila C (1998) Chemotaxis in the nudibranch Hermissenda crassicornis: does ingestive conditioning influence its behaviour in a Y-maze? J Moll Stud 64:215-222. Bablanian GM, Treistman SN (1985) The effect of hyperpolarization of cell R15 on the hemolymph concentration of intact Aplysia. J Comp Physiol B 155:297-303. Bailey CH, Castellucci VF, Koester J, Kandel ER (1979) Cellular studies of peripheral neurons in siphon skin of Aplysia californica. J Neurophysiol 42:530-557. Bailey SER (1981) Circannual and circadian rhythms in the snail Helix aspersa Miiller and the photoperiodic control of annual activity and reproduction. J Comp Physiol 142:89-94. Balaban PM (1979) A system of command neurons in snail's escape behavior. Acta Neurobiol Exp 39:97-107.

References

283

Balaban P (1993) Behavioral neurobiology of learning in terrestrial snails. Prog Neurobiol41:l-19. Balaban P, Chase R (1990) Stimulation of mesocerebrum in Helix aspersa inhibits the neural network underlying avoidance behavior. J Comp Physiol A 166:421-427. Banta PA, Welsford IG, Prior DJ (1990) Water-orientation behavior in the terrestrial gastropod Limax maximus: the effects of dehydration and arginine vasotocin. Physiol Zool 63:683-696. Barnes S, Jacklet JW (1997) Ionic currents of isolated retinal pacemaker neurons: projected daily phase differences and selective enhancement by a phase-shifting neurotransmitter. J Neurophysiol 77:3075-3084. Baur B (1998) Sperm competition in molluscs. In Sperm Competition and Sexual Selection, Birkhead TR, M011er AP (eds), Academic Press, Orlando, pp. 255-305. Benjamin PR (1983) Gastropod feeding: behavioural and neural analysis of a complex multicomponent system. In Neural Origin of Rhythmic Movements, Roberts A, Roberts B (eds), Cambridge University Press, Cambridge UK, pp. 159-193. Benjamin PR, Staras K, Kemenes G (2000) A systems approach to the cellular analysis of associative learning in the pond snail Lymnaea. Learn Mem 7:124-131. Bernheim SM, Mayeri E (1995) Complex behavior induced by egg-laying hormone in Aplysia. J Comp Physiol A 176:131-136. Bernocchi G, Vignola C, Scherini E, Necchi D, Pisu MB (1998) Bioactive peptides and serotonin immunocytochemistry in the cerebral ganglia of hibernating Helix aspersa. J Exp Zool 280:354-367. Bieler R (1992) Gastropod phylogeny and systematics. Ann Rev Ecol Syst 23:311-338. Binder E (1976) La pariade chez le genre Gymnarion Pilsbry 1919 (Pulmonata, Urocyclidae?); role de 1'organe frontal. Arch Moll 107:249-255. Blankenship JE, Rock MK, Robbins LC, Livingston CA, Lehman HK (1983). Aspects of copulatory behavior and peptide control of egg laying in Aplysia. Fed Proc Am Soc Exp Biol 42:96-100. Block GD, Khalsa SBS, McMahon DG, Michel S, Guesz M (1993) Biological clocks in the retina: cellular mechanisms of biological timekeeping. Int Rev Cytol 146:83-144. Blumberg S, Susswein AJ (1998) Consummatory feeding movements in Aplysia fasciata are facilitated by conspecifics with access to mates, by reproductive tract homogenates and by bag cell peptides. J Comp Physiol A 182:175-181. Boer HH, Montagne-Wajer C (1994) Functional morphology of the neuropeptidergic light-yellow-cell system in pulmonate snails. Cell Tiss Res 277:531-538. Borovikov D, Evans CG, Jing J, Rosen SC, Cropper EC (2000) A proprioceptive role for an exteroceptive mechanoafferent neuron in Aplysia. J Neurosci 20: 1990-2002. Bottazzi F, Enriques P (1900) Recherches physiologiques sur le systeme visceral des Aplysies et de quelques Cephalopodes. Arch Ital Biol 34:111-143. Botzer D, Blumberg S, Ziv I, Susswein AJ (1991) Common regulation of feeding and mating in Aplysia fasciata: pheromones released by mating and by egg cordons increase feeding behavior. Behav Neural Biol 56:251-261. Brace RC (1977) Anatomical changes in nervous and vascular systems during the transition from prosobranch to opisthobranch organization. Trans Zool Soc Lond34:l-25. Brezden BL, Yeoman MS, Gardner DR, Benjamin PR (1999) FMRFamide-activated Ca2+ channels in Lymnaea heart cells are modulated by "SEEPLY," a neuropeptide encoded on the same gene. J Neurophysiol 81:1818-1826.

284

References

Brezina V, Evans CG, Weiss KR (1994a) Characterization of the membrane ion currents of a model molluscan muscle, the accessory radula closer muscle of Aplysia californica. II. Depolarization-activated K currents. J Neurophysiol 71:2113-2125. Brezina V, Evans CG, Weiss KR (1994b) Enhancement of Ca current in the accessory radula closer muscle of Aplysia californica by neuromodulators that potentiate its contractions. J Neurosci 14:4393^4411. Brezina V, Orekhova IV, Weiss KR (1996) Functional uncoupling of linked neurotransmitter effects by combinatorial convergence. Science 273:806-810. Brezina V, Orekhova IV, Weiss KR (2000) Optimization of rhythmic behaviors by modulation of the neuromuscular transform. J Neurophysiol 83:260-279. Brierley MJ, Staras K, Benjamin PR (1997a) Behavioral function of glutamatergic interneurons in the feeding system of Lymnaea: plateauing properties and synaptic connections with motor neurons. J Neurophysiol 78:3386-3395. Brierley MJ, Yeoman MS, Benjamin PR (1997b) Glutamatergic N2v cells are central pattern generator interneurons of the Lymnaea feeding system: new model for rhythm generation. J Neurophysiol 78:3396-3407. Brown FA Jr, Webb HM, Brett WJ (1960) Magnetic response of an organism and its lunar relationships. Biol Bull 118:382-392. Brown FA Jr, Webb HM, Barn well FH (1964) Compass directional phenomenon in mud-snails and its relation to magnetism. Biol Bull 127:206-220. Brown RO, Mayeri E (1989) Positive feedback by autoexcitatory neuropeptides in neuroendocrine bag cells of Aplysia. J Neurosci 9:1443-1451. Brown RO, Pulst SM, Mayeri E (1989) Neuroendocrine bag cells of Aplysia are activated by bag cell peptide-containing neurons in the pleural ganglion. J Neurophysiol 61:1142-1152. Brussaard AB, Kits KS, Ter Maat A, Van Minnen J, Moed PJ (1988) Dual inhibitory action of FMRFamide on neurosecretory cells controlling egg laying behavior in the pond snail. Brain Res 447:35-51. Bryan PJ, Slattery M (1996) Reproductive behavior of the gymnosomatous pteropod Clione antarctica. Invert Reprod Dev 29:143-148. Buckett KJ, Peters GJ, Dockray J, van Minnen J, Benjamin PR (1990a) Regulation of heartbeat in Lymnaea by motoneurons containing FMRFamide-like peptides. J Neurophysiol 63:1426-1435. Buckett KJ, Peters M, Benjamin PR (1990b) Excitation and inhibition of the heart of the snail, Lymnaea, by non-FMRFamidergic motoneurons. J Neurophysiol 63:1436-1447. Bullock TH (1953) Predator recognition and escape responses of some intertidal gastropods in the presence of starfish. Behaviour 5:130-140. Bullock TH (1965) Mollusca: Gastropoda. In Structure and Function in the Nervous Systems of Invertebrates, vol II, Bullock TH, Horridge GA (eds), WH Freeman, San Francisco, pp. 1283-1386. Byrne JH, Kandel ER (1996) Presynaptic facilitation revisited: state and time dependence. J Neurosci 16:425^35. Byrne JH, Shapiro E, Dieringer N, Koester J (1979) Biophysical mechanisms contributing to inking behavior in Aplysia. J Neurophysiol 42:1233-1250. Carefoot T (1987) Aplysia: its biology and ecology. Oceanogr Mar Biol Ann Rev 25:167-284. Carefoot T (1989) A comparison of time/energy budgeting in two species of tropical sea hares Aplysia. J Exp Mar Biol Ecol 131:267-282.

References

285

Carefoot TJ, Pennings SC, Danko JP (1999) A test of novel function(s) for the ink of sea hares. J Exp Mar Biol Ecol 234:185-197. Carew TJ, Kandel ER (1977) Inking in Aplysia californica. I. Neural circuit of an all-or-none behavioral response. J Neurophysiol 40:692-707. Carew TJ, Castellucci VF, Byrne JH, Kandel ER (1979) Quantitative analysis of relative contribution of central and peripheral neurons to gill-withdrawal reflex in Aplysia californica. J Neurophysiol 42:497-509. Gate JM (1968) Mating behavior in Mitra idae Melvill, 1893. Veliger 10:247-252. Charnov EL (1979) Simultaneous hermaphroditism and sexual selection. Proc Natl Acad Sci USA 76:2480-2484. Chase R (1979) Photic sensitivity of the rhinophore in Aplysia. Can J Zool 57:698 701. Chase R (1982) The olfactory sensitivity of snails, Achatina fulica. J Comp Physiol 148:225-235. Chase R (1985) Lessons from snail tentacles. Chem Senses 11:411-426. Chase R (1986) Brain cells that command sexual behavior in the snail Helix aspersa. J Neurobiol 17:669-679. Chase R (2000) Structure and function in the cerebral ganglion. Microsc Res Tech 49:511 520. Chase R (2001) Sensory organs and the nervous system. In The Biology of Terrestrial Molluscs, Barker GM (ed), CAB International, Wallingford, UK, pp. 179-211. Chase R, Croll RP (1981) Tentacular function in snail olfactory orientation. J Comp Physiol 143:357-362. Chase R, Tolloczko B (1987) Evidence for differential DNA endoreplication during the development of a molluscan brain. J Neurobiol 18:395^106. Chase R, Tolloczko B (1989) Interganglionic dendrites constitute an output pathway from the procerebrum of the snail Achatina fulica. J Comp Neurol 283:143-152. Chase R, Tolloczko B (1992) Synaptic innervation of the giant cerebral neuron in sated and hungry snails. J Comp Neurol 318:93-102. Chase R, Tolloczko B (1993) Tracing neural pathways in snail olfaction: from the tip of the tentacles to the brain and beyond. Microsc Res Tech 24:214—230. Chase R, Pryer K, Baker R, Madison D (1978) Responses to conspecific chemical stimuli in the terrestrial snail Achatina fulica (Pulmonata: Sigmurethra). Behav Biol 22:302 315. Chiel HJ, Weiss KR, Kupfermann I (1986) An identified histaminergic neuron modulates feeding motor circuitry in Aplysia. J Neurosci 6:2427-2450. Church RJ, Lloyd PE (1991) Expression of diverse neuropeptide cotransmitters by identified motor neurons in Aplysia. J Neurosci 11:618-625. Cimino G, Fontana A, Gavagnin M (1999) Marine opisthobranch molluscs: chemistry and ecology in sacoglossans and dorids. Current Org Chem 3:327 372. Cleary LJ, Byrne JH (1993) Identification and characterization of a multifunctional neuron contributing to defensive arousal in Aplysia. J Neurophysiol 70:1767-1776. Cleary LJ, Byrne JH, Frost WN (1995) Role of interneurons in defensive withdrawal reflexes in Aplysia. Learn Mem 2:133-151. Cobbs JS, Pinsker HM (1982) Role of bag cells in egg deposition of Aplysia brasiliana. II. Contribution of the egg movement to elicited behaviors. J Comp Physiol A 147:537-546. Coggeshall RE (1967) A light and electron microscope study of the abdominal ganglion of Aplysia californica. J Neurophysiol 30:1263-1287.

286

References

Cohen JA, Weiss KR, Kupfermann I (1978) Motor control of buccal muscles in Aplysia. J Neurophysiol 41:157-180. Cohen TE, Kaplan SW, Kandel ER, Hawkins RD (1997) A simplified preparation for relating cellular events to behavior: mechanisms contributing to habituation, dishabituation, and sensitization of the Aplysia gill-withdrawal reflex. J Neurosci 17:2886-2899. Colgan DJ, Ponder WF, Eggler PE (2000) Gastropod evolutionary rates and phylogenetic relationships assessed using partial 28S rDNA and histone H3 sequences. Zoologica Scripta 29:29-63. Conn JP, Kaczmarek LK (1989) The bag cells of Aplysia. Molec Neurobiol 3:237-273. Cook A (1979) Homing by the slug Umax pseudoflavus. Anim Behav 27:545-552. Cook A (1985a) Functional aspects of trail following by the carnivorous snail Euglandina rosea. Malacologia 26:173-181. Cook A (1985b) The organisation of feeding in the carnivorous snail Euglandina rosea. Malacologia 26:183-189. Cook A, Bamford OS, Freeman JDB, Teideman DJ (1969) A study of the homing habit of the limpet. Anim Behav 17:330-339. Cook RT, Bailey SER, McCrohan CR, Nash B, Woodhouse RM (2000) The influence of nutritional status on the feeding behaviour of the field slug, Deroceras reticulatum (Miiller). Anim Behav 59:167-176. Cooke IRC, Gelperin A (2001) In vivo recordings of spontaneous and odor-modulated dynamics in the Limax olfactory lobe. J Neurobiol 46:126-141. Copping J, Syed NI, Winlow W (2000) Seasonal plasticity of synaptic connections between identified neurones in Lymnaea. Acta Biol Hung 51:205-210. Croll RP (1983) Gastropod chemoreception. Biol Rev 58:293-319. Croll RP, Chase R (1980) Plasticity of olfactory orientation to foods in the snail Achatina fulica. J Comp Physiol 136:267-277. Croll RP, Davis WJ (1982) Motor program switching in Pleurobranchaea: II. Ingestion and egestion in the reduced preparation. J Comp Physiol 147:143-154. Croll RP, Van Minnen J (1992) Distribution of the peptide Ala-Pro-Gly-Trp-NH2 (APGWamide) in the nervous system and periphery of the snail Lymnaea stagnalis as revealed by immunocytochemistry and in situ hybidization. J Comp Neurol 324:567-574. Croll RP, Davis WJ, Kovac MP (1985a) Neural mechanisms of motor program switching in the mollusc Pleurobranchaea. I. Central motor programs underlying ingestion, egestion, and the "neutral" rhythm(s). J Neurosci 5:48-55. Croll RP, Kovac MP, Davis WJ (1985b) Neural mechanisms of motor program switching in the mollusc Pleurobranchaea. II. Role of the ventral white cell, anterior ventral, and B3 buccal neurons. J Neurosci 5:56-63. Croll RP, Kovac MP, Davis WJ, Matera EM (1985c) Neural mechanisms of motor program switching in the mollusc Pleurobranchaea. III. Role of the paracerebral neurons and other identified brain neurons. J Neurosci 5:64-71. Crow T (1985) Conditioned modification of phototactic behavior in Hermissenda. II. Differential adaptation of B-photoreceptors. J Neurosci 5:215-223. Crow T, Tian L-M (2000) Monosynaptic connections between identified A and B photoreceptors and interneurons in Hermissenda: evidence for labeled-lines. J Neurophysiol 84:367-375. Dall WH (1881) Intelligence in a snail. Am Nat 15:976-978. Darwin C (1871) The Descent of Man, and Selection in Relation to Sex. Facsimile ed, Princeton University Press, Princeton, N.J.

References

287

De Boer PACM, Jansen RF, Ter Maat A (1996) Copulation in the hermaphroditic snail Lymnaea stagnalis: a review. Invert Reprod Dev 30:167-176. De Boer PACM, Jansen RF, Koene JM, Ter Maat A (1997a) Nervous control of male sexual drive in the hermaphroditic snail Lymnaea stagnalis. J Exp Biol 200:941-951. De Boer PACM, Ter Maat A, Pieneman AW, Croll RP, Kurokawa M, Jansen R (1997b) Functional role of peptidergic anterior lobe neurons in male sexual behavior of the snail Lymnaea stagnalis. J Neurophysiol 78:2823-2833. Delaney K, Gelperin A (1986) Post-ingestive food-aversion learning to amino acid deficient diets by the terrestrial slug Limax maximus. J Comp Physiol A 159:281-295. Delaney K, Gelperin A (1990) Cerebral interneurons controlling fictive feeding in Limax maximus. II. Initiation and modulation of fictive feeding. J Comp Physiol A 166:311-326. De Lange RPJ, Van Minnen J (1998) Localization of the neuropeptide APGWamide in gastropod molluscs by in situ hybridization and immunocytochemistry. Gen Comp Endocrinol 109:166-174. De Lange RPJ, Joosse J, Van Minnen J (1998a) Multi-messenger innervation of the male sexual system of Lymnaea stagnalis. J Comp Neurol 390:564-577. De Lange RPJ, De Boer PACM, Ter Maat A, Jansen CP, Van Minnen J (1998b) Transmitter identification in neurons involved in male copulation behavior in Lymnaea stagnalis. J Comp Neurol 395:440—449. Deliagina TG, Orlovsky GN (1990) Control of locomotion in the freshwater snail Planorbis corneus. II. Differential control of various zones of the ciliated epithelium. J Exp Biol 152:405-423. Deliagina TG, Arshavsky YI, Orlovsky GN (1998) Control of spatial orientation in a mollusc. Nature 393:172-175. Deliagina TG, Orlovsky GN, Selverston AI, Arshavsky YI. (2000) Neuronal mechanisms for the control of body orientation in Clione II. Modifications in the activity of postural control system. J Neurophysiol 83:367-373. Denny M (1980) The role of gastropod pedal mucus in locomotion. Nature: 285:160-161. Detwiler PB, Alkon DL (1973) Hair cell interactions in the statocyst of Hermissenda. J Gen Physiol 62:618-642. De With ND (1996) Oral water ingestion in the pulmonate freshwater snail, Lymnaea stagnalis. J Comp Physiol B 166:337-343. De Witt TJ (1996) Gender contests in a simultaneous hermaphrodite snail: sizeadvantage model for behavior. Anim Behav 51:345-351. Dickinson PS (1980) Neuronal control of gills in diverse Aplysia species: conservative evolution. J Comp Physiol 139:17-23. Diefenbach TJ, Koss R, Goldberg JI (1998) Early development of an identified serotonergic neuron in Helisoma trivolvis embryos: serotonin expression, de-expression, and uptake. J Neurobiol 34:361-376. Dieringer N, Koester J, Weiss KR (1978) Adaptive changes in heart rate of Aplysia californica. J Comp Physiol 123:11-21. Dorsett DA (1986) Brains to cells: the neuroanatomy of selected gastropod species. In The Mollusca, vol 9, part 2, Willows AOD (ed.), Academic Press, Orlando, pp. 101-187. Dorsett DA, Evans CG (1991) Potassium and calcium currents in dissociated muscle fibres of the mollusc Philine aperta. J Exp Biol 155:305-321.

288

References

Dubuc B, Castellucci VF (1991) Receptive fields and properties of a new cluster of mechanoreceptor neurons innervating the mantle region and the branchial cavity of the marine mollusk Aplysia californica. J Exp Biol 156:315-334. Dudek FE, Cobbs JS, Pinsker HM (1979) Bag cell electrical activity underlying spontaneous egg laying in freely behaving Aplysia brasiliana. J Neurophysiol 42:804-817. Eberhardt B, Wabnitz RW (1979) Morphological identification and functional analysis of central neurons innervating the penis retractor muscle of Helix pomatia. Comp Biochem Physiol 63A:599-613. Elliott CJH, Benjamin PR (1985a) Interactions of pattern-generating interneurons controlling feeding in Lymnaea stagnalis. J Neurophysiol 54:1396-1411. Elliott CJH, Benjamin PR (1985b) Interactions of the slow oscillator interneuron with feeding pattern generating interneurons in Lymnaea stagnalis. J Neurophysiol 54:1412-1421. Elliott CJH, Benjamin PR (1989) Esophageal mechanoreceptors in the feeding system of the pond snail, Lymnaea stagnalis. J Neurophysiol 61:727-736. Emberton KC (1994) Polygyrid land snail phylogeny: external sperm exchange, early North American biogeography, iterative shell evolution. Biol J Linn Soc 52:241-271. Emery DG (1992) Fine structure of olfactory epithelia of gastropod molluscs. Microsc Res Tech 22:307-324. Erlichman JS, Leiter JC (1997) Identification of CO2 chemoreceptors in Helix pomatia. Am Zool 37:54-64. Fan X, Croll RP, Wu B, Fang L, Shen Q, Painter SD, Nagle GT (1997) Molecular cloning of a cDNA encoding the neuropeptides APGWamide and cerebral peptide 1: Localization of APGWamide-like immunoreactivity in the central nervous system and male reproductive organs of Aplysia. J Comp Neurol 387:53-62. Fang X, Clark GA (1996) Differential responses of Aplysia siphon motor neurons and interneurons to tail and mantle stimuli: implications for behavioral response specificity. J Neurophysiol 76:3895-3909. Farley J, Han Y (1997) Ionic basis of learning-correlated excitability changes in Hermissenda type A photoreceptors. J Neurophysiol 77:1861-1888. Fenoglio C, Scherini E, Necchi D, Soldani C, Bernocchi G (1997) Perineuronal glial system in the cerebral ganglion of active and hibernating Helix aspersa. Tiss Cell 29:561-572. Ferguson GP, Benjamin PR (199 la) The whole body withdrawal response of Lymnaea stagnalis. I. Identification of central motoneurones and muscles. J Exp Biol 158:63-95. Ferguson GP, Benjamin PR (1991b) The whole body withdrawal response of Lymnaea stagnalis. II. Activation of central motoneurones and muscles by sensory input. J Exp Biol 158:97-116. Ferguson GP, Parsons DW, Ter Maat A, Pinsker HM (1986) Spontaneous and elicited bag cell discharges in gonadectomized Aplysia. J Exp Biol 123:159-173. Ferguson GP, Ter Maat A, Parsons DW, Pinsker HM (1989a) Egg laying in Aplysia. I. Behavioral patterns and muscle activity of freely behaving animals after selectively elicited bag cell discharges. J Comp Physiol A 164:835-847. Ferguson GP, Ter Maat A, Pinsker HM (1989b) Egg laying in Aplysia. II. Organization of central and peripheral pathways for initiating neurosecretory activity and behavioral patterns. J Comp Physiol A 164:849-857.

References

289

Ford DIG, Cook A (1987) The effects of temperature and light on the circadian activity of the pulmonate slug, Limax pseudoflavus

Evans. Anim Behav

35:1754-1765. Fraenkel G (1927) Beitrage zur Geotaxis und Phototaxis von Littorina. Z Vergl Physiol 5:586-597. Translation in The Organization of Action: A New Synthesis, Gallistel CR, Lawrence Erlbaum, Hillsdale, N.J. Fredman SM, Jahan-Parwar B (1975) Synaptic connections in the cerebral ganglion of Aplysia. Brain Res 100:209-214. Fredman SM, Jahan-Parwar B (1980) Processing of chemosensory and mechanosensory information in identifiable Aplysia neurons. Comp Biochem Physiol 66A:25-34. Fredman SM, Jahan-Parwar B (1983) Command neurons for locomotion in Aplysia. J Neurophysiol 49:1092-1117. Fretter V (1984) Prosobranchs. In The Mollusca, vol 7, Tompa AS, Verdonk NH, van den Biggelaar JAM (eds), Academic, Orlando, pp. 1-45. Fretter V, Graham A (1994) British Prosobranch Molluscs: their Functional Anatomy and Ecology, 2nd edition, Ray Society, London. Frost L, Kaplan SW, Cohen TE, Henzi V, Kandel ER, Hawkins RD (1997) A simplified preparation for relating cellular events to behavior: contribution of LE and unidentified siphon sensory neurons to mediation and habituation of the Aplysia gill- and siphon-withdrawal reflex. J Neurosci 127:2900-2913. Frost WN, Kandel ER (1995) Structure of the network mediating siphon-elicited withdrawal in Aplysia. J Neurophysiol 73:2413-2427. Frost WN, Katz PS (1996) Single neuron control over a complex motor program. Proc Natl Acad Sci USA 93:422-426. Frost WN, Clark GA, Kandel ER (1988) Parallel processing of short-term memory for sensitization in Aplysia. J Neurobiol 19:297-334. Frysztak RJ, Crow T (1994) Enhancement of type B and A photoreceptor inhibitory synaptic connections in conditioned Hermissenda. J Neurosci 14: 1245-1250. Furukawa Y, Kobayashi M (1987) Neural control of heart beat in the african giant snail, Achatina fulica Ferussac. II. Interconnections among the heart regulatory neurones. J Exp Biol 129:295-307. Gainer H (1972) Effects of experimentally induced diapause on the electrophysiology and protein synthesis patterns of identified molluscan neurons. Brain Res 39:387-402. Gamkrelidze GN, Laurienti PJ, Blankenship JE (1995) Identification and characterization of cerebral ganglion neurons that induce swimming and modulate swim-related pedal ganglion neurons in Aplysia brasiliana. J Neurophysiol 74:1444-1462. Gandhi CC, Matzel LD (2000) Modulation of presynaptic action potential kinetics underlies synaptic facilitation of type B photoreceptors after associative conditioning in Hermissenda. J Neurosci 20:2022-2035. Garton DW, Roller RA, Caprio J (1984) Fine structure and vital staining of osphradium of the southern oyster drill, Thais Haemastoma canaliculata (Gray) (Prosobranchia: Muricidae). Biol Bull 167:310-321. Gelperin A (1999) Oscillatory dynamics and information processing in olfactory systems. J Exp Biol 202:1855-1864. Gervais R, Kleinfeld D, Delaney DR, Gelperin A (1996) Central and reflex responses elicited by odor in a terrestrial mollusc. J Neurophysiol 76:1327-1339.

290

References

Getting PA (1989) Reconstruction of small neural networks. In Methods in Neuronal Modeling, Koch C, Segev I (eds), MIT, Cambridge, Mass, pp. 171-194. Getting PA, Dekin MS (1985) Tritonia swimming: a model system for integration within rhythmic motor systems. In Model Neural Networks and Behavior, Selverston AI (ed), Plenum, New York, pp. 3-20. Getting PA, Lennard PR, Hume RI (1980) Central pattern generator mediating swimming in Tritonia. I. Identification and synaptic interactions. J Neurophysiol 44:151-164. Giardino ND, Aloyz RS, Zollinger M, Miller MW, Desgroseillers L (1996) L5-67 and LUQ-1 peptide precursors of Aplysia californica—distribution and localization of immunoreactivity in the central nervous system and in peripheral tissues. J Comp Neurol 374:230-245. Gillette R (1991) On the significance of neuronal giantism in gastropods. Biol Bull 180:234-240. Gillette R, Huang R-C, Hatcher N, Moroz LL (2000) Cost-benefit analysis potential in feeding behavior of a predatory snail by integration of hunger, taste, and pain. Proc Natl Acad Sci USA 97:3585-3590. Gillette R, Saeki M, Huang R-C (1991) Defense mechanisms in notaspid snails: acid humor and evasiveness. J Exp Biol 156:335-347. Gommerat I, Gola M (1996) Glial potassium channels activated by neuronal firing or intracellular cyclic AMP in Helix. J Physiol 495:649-664. Graham A (1973) The anatomical basis of function in the buccal mass of prosobranch and amphineuran molluscs. J Zool (Lond) 169:317-348. Graham A (1985) Evolution within the Gastropoda: Prosobranchia. In The Mollusca, vol 10, Trueman ER, Clarke MR (eds), Academic Press, Orlando, pp. 151-186. Greeff JM, Michiels NK (1999) Sperm digestion and reciprocal sperm transfer can drive hermaphrodite sex allocation to equality. Am Nat 153:421-430. Hadfield MG, Switzer-Dunlap M (1984) Opisthobranchs. In The Mollusca, vol 7, Tompa AS, Verdonk NH, van den Biggelaar JAM (eds), Academic Press, Orlando, pp. 209-350. Hall JD, Lloyd PE (1990) Involvement of pedal peptide in locomotion in Aplysia: modulation of foot muscle contractions. J Neurobiol 21:858-868. Hamilton PV (1991) Variation in sense organ design and associated sensory capabilities among closely related molluscs. Am Malac Bull 9:89-98. Hamilton PV, Russell BJ (1982) Celestrial orientation by surface-swimming Aplysia brasiliana Rang (Mollusca:Gastropoda). J Exp Mar Biol Ecol 56:145-152. Hamilton PV, Winter MA (1984) Behavioural responses to visual stimuli by the snails Tectarius muricatus, Turbo castanea, and Helix aspersa. Anim Behav 32:51-57. Hartnoll RG, Wright JR (1977) Foraging movements and homing in the limpet Patella vulgata L. Anim Behav 25:806-810. Hasegawa K, Takeda N (1986) Responses of giant neurons to osmotic stress in the giant african snail, Achatina fulica. Comp Biochem Physiol 84A:289-294. Haszprunar G (1985) The fine morphology of the osphradial sense organs of the Mollusca. I. Gastropoda, Prosobranchia. Phil Trans R Soc Lond B 307: 457^96. Haszprunar G (1988) On the origin and evolution of major gastropod groups, with special reference to the Streptoneura. J Moll Stud 54:367^441. Haszprunar G, Huber G (1990) On the central nervous system of Smeagolidae and Rhodopidae, two families questionably allied with the Gymnomorpha (Gastropoda:Euthyneura). J Zool (Lond) 220:185-199.

References

291

Hawkins RD, Greene H, Kandel ER (1998) Classical conditioning, differential conditioning, and second-order conditioning of the Aplysia gill-withdrawal reflex in a simplified mantle organ preparation. Behav Neurosci 112:636-645. Hening WA, Walters ET, Carew TJ, Kandel ER (1979) Motorneuronal control of locomotion in Aplysia. Brain Res 179:231-253. Hermann PM, Ter Maat A, Jansen RF (1994) The neural control of egg-laying behaviour in the pond snail Lymnaea stagnalis: motor control of shell turning. J Exp Biol 197:79-99. Hermann PM, De Lange RPJ, Pieneman AW, Ter Maat A, Jansen RF (1997) Role of neuropeptides encoded on CDCH-1 gene in the organization of egglaying behavior in the pond snail, Lymnaea stagnalis. J Neurophysiol 78: 2859-2869. Hernadi L (1992) Relationships between the distribution of serotonergic cell bodies and the running of vascular elements in the central nervous system of the snail, Helix pomatia. Comp Biochem Physiol 103A:85-92. Hickie C, Walters ET (1995) Motor neuronal control of tail-directed and headdirected siphon responses in Aplysia californica. J Neurophysiol 74:307 321. Hodgson AN (1999) The biology of siphonariid limpets. Oceanogr Mar Biol Ann Rev 37:245-314. Horn CC, Koester J, Kupfermann I. (1998) Evidence that hemolymph glucose in Aplysia californica is regulated but does not affect feeding behavior. Behav Neurosci 112:1258-1265. Hoyle G (1983) Muscles and their Neural Control. John Wiley & Sons, New York. Hughes GM, Tauc L (1963) An electrophysiological study of the anatomical relations of two giant nerve cells in Aplysia depilans. J Exp Biol 40:469-486. Hume RI, Getting PA, Del Beccaro MA (1982) Motor organization of Tritonia swimming. I. Quantitative analysis of swim behavior and flexion neuron firing patterns. J Neurophysiol 47:60-74. Hurwitz I, Susswein AJ (1996) B64, a newly identified central pattern generator element producing a phase switch from protraction to retraction in buccal motor programs of Aplysia californica. J Neurophysiol 75:1327-1344. Hurwitz I, Kupfermann I, Susswein AJ (1997) Different roles of neurons B63 and B34 that are active during the protraction phase of buccal motor programs in Aplysia californica. J Neurophysiol 78:1305-1319. Hurwitz I, Perrins R, Xin Y, Weiss KR, Kupfermann I (1999) C-PR neuron of Aplysia has differential effects on "feeding" cerebral interneurons, including myomodulinpositive CBI-12. J Neurophysiol 81:521-534. Iglesias J, Castillejo J (1999) Field observations on feeding of the land snail Helix aspersa Miiller. J Moll Stud 65:411^123. Illich PA, Walters ET (1997) Mechanosensory neurons innervating Aplysia siphon encode noxious stimuli and display nociceptive sensitization. J Neurosci 17:45^69. Illich PA, Joynes RL, Walters ET (1994) Response-specific inhibition during general facilitation of defensive responses in Aplysia. Behav Neurosci 108:614-623. Inoue T, Haque Z, Lukowiak K, Syed NI (2001) Hypoxia-induced respiratory patterned activity in Lymnaea originates in the periphery. J Neurophysiol 86:156-163. Inoue T, Takasaki M, Lukowiak K, Syed NI (1996) Inhibition of the respiratory pattern-generating neurons by an identified whole-body withdrawal interneuron of Lymnaea stagnalis. J Exp Biol 199:1887-1898.

292

References

Ito I, Nakamura H, Kimura T, Suzuki H, Sekiguchi T, Kawabata K, Ito E (2000) Neuronal components of the superior and inferior tentacles in the terrestrial slug, Limax marginatus. Neurosci Res 37:191 200. Jacklet JW (1969) Circadian rhythm of optic nerve impulses recorded in darkness from isolated eye of Aplysia. Science 164:562-563. Jacklet JW (1989) Circadian neuronal oscillators. In Neuronal and Cellular Oscillators, JW Jacklet (ed.), Marcel Decker, New York, pp. 482-527. Jacklet JW, Rine FJ (1977) Facilitation at neuromuscular junctions: contribution to habituation and dishabituation of the Aplysia gill withdrawal reflex. Proc Natl Acad Sci USA 74:1267 1271. Jahan-Parwar B, Fredman SM (1979) Neural control of locomotion in Aplysia: role of the central ganglia. Behav Neural Biol 27:39 58. Janse C (1983) The function of the statocyst sensory cells in Aplysia limacina. J Comp Physiol 150:359-370. Janse C, Van der Wilt GJ, Van der Plas J, Van der Roest M (1985) Central and peripheral neurones involved in oxygen perception in the pulmonate snail Lymnaea stagnalis. Comp Biochem Physiol 82A:459^69. Janse C, Van der Wilt GJ, Van der Roest M, Pieneman AW (1988) Intracellularly recorded responses to tilt and efferent input of statocyst sensory cells in the pulmonate snail Lymnaea stagnalis. Comp Biochem Physiol 90A:269-278. Jansen RF, Pieneman AW, Ter Maat A (1999) Pattern generation in the buccal system of freely behaving Lymnaea stagnalis. J Neurophysiol 82:3378-3391. Jing J, Gillette R (1999) Central pattern generator for escape swimming in the notaspid sea slug Pleurobranchaea californica. J Neurophysiol 81:654—667. Jing J, Gillette R (2000) Escape swim network interneurons have diverse roles in behavioral switching and putative arousal in Pleurobranchaea. J Neurophysiol 83:1346-1355. Jing J, Weiss KR (2001) Neural mechanisms of motor program switching in Aplysia. J Neurosci 21:7349-7362. Johnson PM, Willows AOD (1999) Defense in sea hares (Gastropoda, Opisthobranchia, Anaspidea): multiple layers of protection from egg to adult. Mar Fresh Behav Physiol 32:147-180. Kabotyanski EA, Baxter DA, Cushman SJ, Byrne JH (2000) Modulation of fictive feeding by dopamine and serotonin in Aplysia. J Neurophysiol 83:374-392. Kandel ER (1976) Cellular Basis of Behavior. Freeman, San Francisco. Kandel ER (1979) Behavioral Biology of Aplysia. Freeman, San Francisco. Kandel ER (2001) The molecular biology of memory storage: A dialogue between genes and synapses. Science 294:1030-1038. Kandel ER, Schwarz JH (1982) Molecular biology of learning: modulation of transmitter release. Science 218:433^43. Kanz JE, Quast WD (1990) Respiratory pumping seizure: a newly discovered spontaneous stereotyped behavior pattern in the opisthobranch mollusc Aplysia californica. J Comp Physiol A 166:619-627. Karhunen T, Vilim FS, Alexeeva V, Weiss KR, Church PJ (2001) Targeting of peptidergic vesicles in cotransmitting terminals. J Neurosci 21:RC127 (1-5). Katz PS, Frost WN (1995) Intrinsic neuromodulation in the Tritonia swim CPG: The serotonergic dorsal swim interneurons act presynaptically to enhance transmitter release from interneuron C2. J Neurosci 15:6035-6045.

References

293

Katz PS, Frost WN (1997) Removal of spike frequency adaptation via neuromodulation intrinsic to the Tritonia escape swim central pattern generator. J Neurosci 17:7703-7713. Kaufmann W, Kerschbaum HH, Hauser-Kronberger C, Hacker GW, Hermann A (1995) Distribution and seasonal variation of vasoactive intestinal (VlP)-like peptides in the nervous system of Helix pomatia. Brain Res 695:125 136. Kavaliers M, Ossenkopp K.-P (1991) Opioid systems and magnetic field effects in the land snail, Cepaea nemoralis. Biol Bull 180:301-309. Keating C, Lloyd PE (1999) Differential modulation of motor neurons that innervate the same muscle but use different excitatory transmitters in Aplysia. J Neurophysiol 82:1759-1767. Kemenes G (1994) Processing of mechano- and chemosensory information in the lip nerve and cerebral ganglia of the snail Helix pomatia L. Neurosci Behav Physiol 24:77-87. Kemenes G, Staras, K, Benjamin PR (2001) Multiple types of control by identified interneurons in a sensory-activated rhythmic motor pattern. J Neurosci 21:2903-2911. Kerkhoven RM, Ramkema MD, van Minnen J, Croll RP, Pin T, Boer HH (1993) Neurons in a variety of molluscs react to antibodies raised against the VD1/RPD2 alpha-neuropeptide of the pond snail Lymnaea stagnalis. Cell Tiss Res 273:371-379. Khalsa SBS, Whitmore D, Bogart B, Block GD (1996) Evidence for a central role of transcription in the timing mechanism of a circadian clock. Am J Physiol 271:C1646-C1651. Khan HR, Price DA, Doble KE, Greenberg MJ, Saleuddin ASM (1999) Osmoregulation and FMRFamide-related peptides in the salt marsh snail Melampus bidentatus (say). Biol Bull 196:153-162. Klein M (1994) Synaptic augmentation by 5-HT at rested Aplysia sensorimotor synapses: independence of action potential prolongation. Neuron 13:159-166. Kleinfeld D, Delaney DR, Fee MS, Flores JA, Tank DW, Gelperin A (1994) Dynamics of propagating waves in the olfactory network of a terrestrial mollusc: an electrical and optical study. J Neurophysiol 72:1402-1419. Knight JB, Cox LR, Keen AM, Batten RL, Yochelson EL, Robertson R (1960) Systematic descriptions. In Treatise on Invertebrate Palenotology, Part I, Moore RC (ed.), University of Kansas, Lawrence, pp. 169 310. Koene JM, Chase R (1998) Changes in the reproductive system of the snail Helix aspersa caused by mucus from the love dart. J Exp Biol 201:2313-2319. Koene JM, Jansen RF, Ter Maat A, Chase R (2000) A conserved location for the central nervous system control of mating behaviour in gastropod molluscs: evidence from a terrestrial snail. J Exp Biol 203:1071-1080. Koester J (1989) Chemically and electrically coupled interneurons mediate respiratory pumping in Aplysia. J Neurophysiol 62:1113-1126. Koester J, Alevizos A (1989) Innervation of the kidney of Aplysia by L10, the LUQ cells, and an identified peripheral motoneuron. J Neurosci 9:4078^088. Koester J, Koch UT (1987) Neural control of the circulatory system of Aplysia. Experientia 43:972-980. Koester J, Mayeri E, Liebeswar G, Kandel ER (1974) Neural control of circulation in Aplysia. II. Interneurons. J Neurophysiol 37:476-496. Kohn AJ (1983) Feeding biology of gastropods. In The Mollusca, vol 5, part 2, Saleuddin ASM, Wilbur KM (eds), Academic Press, New York, pp. 1-63.

294

References

Kojima S, Nanakamura H, Nagayama S, Fujito Y, Ito E (1997) Enhancement of an inhibitory input to the feeding central pattern generator in Lymnaea stagnalis during conditioned taste-aversion learning. Neurosci Lett 230:179-182. Kovac MP, Davis WJ (1980a) Reciprocal inhibition between feeding and withdrawal behaviors in Pleurobranchaea. J Comp Physiol 139:77-86. Kovac MP, Davis WJ (1980b) Neural mechanisms underlying behavioral choice in Pleurobranchaea. J Neurophysiol 43:469^487. Kozak JA, Weiss KR, Brezina V (1996) Two ion currents activated by acetylcholine in the ARC muscle of Aplysia. J Neurophysiol 75:660-677. Kupfermann I (1970) Stimulation of egg laying by extracts of neuroendocrine cells (bag cells) of abdominal ganglion of Aplysia. J Neurophysiol 33:877-881. Kupfermann I (1974) Feeding behavior in Aplysia'. a simple system for the study of motivation. Behav Biol 10:1^26. Kupfermann I, Carew TJ (1974) Behavior patterns of Aplysia californica in its natural environment. Behav Biol 12:317-337. Kupfermann I, Carew TJ, Kandel ER (1974) Local, reflex and central commands controlling gill and siphon movements in Aplysia. J Neurophysical 37: 996-1019. Kupfermann I, Kandel ER (1969) Neuronal controls of a behavioral response mediated by the abdominal ganglion of Aplysia. Science 164:847-850. Kupfermann I, Weiss K (1976) Water regulation by a presumptive hormone contained in identified neurosecretory cell R15 of Aplysia. J Gen Physiol 67:113-123. Kupfermann I, Weiss K (1981) Tail pinch and handling facilitate feeding behavior in Aplysia. Behav Neural Biol 32:126-132. Kupfermann I, Teyke T, Rosen SC, Weiss KR (1991) Studies of behavioral state in Aplysia. Biol Bull 180:262-268. Kurokawa M, Ohsuga K, Kuwasawa K (1998) A reexamination of the synaptic connection between neuron L7 of the abdominal ganglion and neurons of the branchial ganglion in Aplysia californica, A. kurodai and A. Juliana. Neurosci Lett 241:49-52. Kyriakides MA, McCrohan CR (1988) Central coordination of buccal and pedal neuronal activity in the pond snail Lymnaea stagnalis. J Exp Biol 136: 103-123. Lalli CM (1972) Structure and function of the buccal apparatus of Clione limacina, with a review of feeding in gymnosomatous pteropods. J Exp Mar Biol Ecol 4:101-118. Lalli CM, Gilmer RW (1989) Pelagic Snails, Stanford University Press, Stanford, Calif. Land MF (1981) Optics and vision in invertebrates. In Handbook of Sensory Physiology, vol. VII/6B, Autrum H (ed), Springer, Berlin, pp. 471-594. Landolfa MA, Green DM, Chase R (2001) Dart shooting influences paternal reproductive success in the snail Helix aspersa (Pulmonata, Sytylommatophora). Behav Ecol 12:773-777. Langlois TH (1965) The conjugal behavior of the introduced European giant garden slug, Limax maximus L., as observed on South Bass Island, Lake Erie. Ohio J Sci 65:298-304. Laurienti PJ, Blankenship JE (1996) Parapodial swim muscle in Aplysia brasiliana. I. Voltage-gated membrane currents in isolated muscle fibers. J Neurophysiol 76:1517-1530. Laurienti PJ, Blankenship JE (1997) Serotonergic modulation of a voltage-gated calcium current in parapodial swim muscle from Aplysia brasiliana. J Neurophysiol 77:1496-1502.

References

295

Lechner HA, Baxter DA, Byrne JH (2000) Classical conditioning of feeding in Aplysia: II. Neurophysiological correlates. J Neurosci 20:3377-3386. Lehman HK, Greenberg MJ (1987) The actions of FMRFamide-like peptides on visceral and somatic muscles of the snail Helix aspersa. J Exp Biol 131:55-68. Lemaire M, Chase R (1998) Twitching and quivering of the tentacles during snail olfactory orientation. J Conip Physiol A 182:81-87. Lemmnitz G, Schuppe H, Wolff HG (1989) Neuromotor bases of the escape behaviour of Nassa mutabilis. J Exp Biol 143:493-507. Leonard JL (1990) The hermaphrodite's dilemma. J Theor Biol 147:361-372. Leonard JL, Lukowiak K (1986) The behavior of Aplysia californica Cooper (Gastropoda; Opisthobranchia): I. Ethogram. Behaviour 98:320-360. Leonard JL, Lukowiak K (1991) Sex and the simultaneous hermaphrodite: testing models of male-female conflict in a sea slug, Navanax inermis (Opisthobranchia). Anim Behav 41:255-266. Leonard JL, Edstrom J, Lukowiak K (1989) Reexamination of the gill withdrawal reflex of Aplysia californica Cooper (Gastropoda; Opisthobranchia). Behav Neurosci 103:585-604. Levenson J, Byrne JH, Eskin A (1999) Levels of serotonin in the hemolymph of Aplysia are modulated by light/dark cycles and sensitization training. J Neurosci 19:8094-8103. Levy M, Levy L, Susswein AJ (1997) Respiratory pumping in Aplysia fasciata in natural and artificial tide pools. J Comp Physiol A 180:81-89. Li G, Chase R (1995) Correlation of axon projections and peptide immunoreactivity in mesocerebral neurons of the snail Helix aspersa. J Comp Neurol 353:9-17. Li L, Garden RW, Floyd PD, Moroz TP, Gleeson JM, Sweedler JV, Pasa-Tolic L, Smith RD (1999) Egg-laying hormone peptides in the Aplysiidae family. J Exp Biol 202:2961-2973. Lickey ME, Wozniak JA, Block GD, Hudson DJ, Augter GK (1977) The consequences of eye removal for the circadian rhythm of behavioral activity in Aplysia. J Comp Physiol 118:121-143. Ligman SH, Brownell PH (1985) Differential hormonal activation of the bag cell neurons on the arterial system of Aplysia. J Comp Physiol 157:31-37. Lind H (1968) Hibernating behaviour of Helix pomatia L. (Gastropoda, Pulmonata). Vidensk Meddr Dansk Naturh Foren 131:129-151. Lind H (1988) The behaviour of Helix pomatia L. (Gastropoda, Pulmonata) in a natural habitat. Vidensk Meddr Dansk Naturh Foren 147:67-92. Lind H (1989) Homing to hibernating sites in Helix pomatia involving detailed longterm memory. Ethology 81:221-234. Linhart YB, Thompson JD (1995) Terpene-based selective herbivory by Helix aspersa (Mollusca) on Thymus vulgaris (Labiatae). Oecologia 102:126-132. Lloyd PE Mahon AC, Kupfermann I, Cohen JL, Scheller RH, Weiss KR (1985) Biochemical and immunocytological localization of molluscan small cardioactive peptides in the nervous system of Aplysia californica. J Neurosci 5:1851-1861. Lohmann KJ, Willows AOD (1987) Lunar-modulated geomagnetic orientation by a marine mollusk. Science 235:331-334. Lukowiak K, Jacklet JW (1972) Habituation and dishabituation: Interactions between peripheral and central nervous systems in Aplysia. Science 178:1306-1308. Lukowiak K, Ringseis E, Spencer G, Wildering W, Syed N (1996) Operant conditioning of aerial respiratory behaviour in Lymnaea stagnalis. J Exp Biol 199:683-691.

296

References

Lukowiak K, Cotter R, Westly J, Ringseis E, Spencer G, Syed N (1998). Long-term memory of an operantly conditioned respiratory behaviour pattern in Lymnaea stagnalis. J Exp Biol 201:877-882. MacGinitie GE (1934) The egg-laying activities of the sea hare, Tethys californicus (Cooper). Biol Bull 67:300-303. Mackey SL, Kandel ER, Hawkins RD (1989) Identified serotonergic neurons LCB1 and RGB 1 in the cerebral ganglia of Aplysia produce presynaptic facilitation of siphon sensory neurons. J Neurosci 9:4227-4235. Magoski SN, Knox RJ, Kaczmarek LK (2000) Activation of a Ca2+-permeable cation channel produces a prolonged attenuation of intracellular Ca2+ release in Aplysia bag cell neurones. J Physiol 522:271-283. Malyshev AY, Norekian TP, Willows AOD (1999) Differential effects of serotonergic and peptidergic cardioexcitatory neurons on the heart activity in the pteropod mollusc, Clione limacina. J Comp Physiol 185:551-560. Marin A, Alvarez LA, Cimino G, Spinella A (1999) Chemical defence in cephalaspidean gastropods: origin, anatomical location and ecological roles. J Moll Stud 65:121-131. McCarragher G, Chase R (1985) Quantification of ultrastructural symmetry at molluscan chemical synapses. J Neurobiol 16: 69-74. McClellan AD, Brown GD, Getting PA (1994) Modulation of swimming in Tritonia: excitatory and inhibitory effects of serotonin. J Comp Physiol A 174:257-266. McCrohan CR (1984) Initiation of feeding motor output by an identified interneurone in the snail Lymnaea stagnalis. J Exp Biol 113:351-366. Mclntosh MJ, Olivera BM, Cruz LJ (1999) Conus peptides as probes for ion channels. Meth Enzymol 294:605-624. McKenzie JD, Caunce M, Hetherington MS, Winlow W (1998) Serotonergic innervation of the foot of the pond snail Lymnaea stagnalis (L.). J Neurocytol 27:459-470. McPherson DR, Blankenship JE (199la) Neural control of swimming in Aplysia brasiliana. I. Innervation of parapodial muscle by pedal ganglion motoneurons. J Neurophysiol 66:1338-1351. McPherson DR, Blankenship, JE (1991b) Neural control of swimming in Aplysia brasiliana. III. Serotonergic modulatory neurons. J Neurophysiol 66:1366-1379. McPherson DR, Blankenship JE (1992) Neuronal modulation of foot and body-wall contractions in Aplysia californica. J Neurophysiol 67:23-28. Mead AR (1979) Economic Malacology, Academic Press, London. Michel S, Geusz ME, Zaritsky JJ, Block GD (1993) Circadian rhythm in membrane conductance expressed in isolated neurons. Science 259:239-241. Michiels NK (1998) Mating conflicts and sperm competition in simultaneous hermaphrodites. In Sperm Competition and Sexual Selection, Birkhead TR, M011er AP (eds), Academic Press, Orlando, pp. 219-254. Mirolli M, Talbott SR (1972) The geometrical factors determining the electrotonic properties of a molluscan neurone. J Physiol 227:19-34. Morgan PT, Perrins R, Lloyd PE, Weiss KR (2000) Intrinsic and extrinsic modulation of a single central pattern generating circuit. J Neurophysiol 84:1186-1193. Moroz LL (2000) Giant identified NO-releasing neurons and comparative histochemistry of putative nitrergic systems in gastropod molluscs. Microsc Res Tech 49:557-569.

References

297

Morton DW, Chiel HJ (1993) In vivo buccal nerve activity that distinguishes ingestion from rejection can be used to predict behavioral transitions in Aplysia. J Comp Physiol A 172:17-32. Morton JE (1979) Molluscs, 5th edn, Hutchinson, London. Mpitsos GP, Lukowiak K (1985) Learning in gastropod molluscs. In The Mollusca, vol 9, part 1, Willows AOD (ed.), Academic Press, Orlando, pp. 95-267. Munoz DP, Pawson PA, Chase R (1983) Symmetrical giant neurones in asymmetrical ganglia: implications for evolution of the nervous system in pulmonate molluscs. JExpBiol 107:147-161. Murphy AD (2001) The neuronal basis of feeding in the snail, Helisoma, with comparisons to selected gastropods. Prog Neurobiol 63:383^108. Murphy BF, Hadfield MG (1997) Chemoreception in the nudibranch gastropod Phestilla sibogae. Comp Biochem Physiol 118:727-735. Murphy GG, Glanzman DL (1997) Mediation of classical conditioning in Aplysia californica by long-term potentiation of sensorimotor synapses. Science 278:461-471. Murray JA, Willows AOD (1996) Function of identified nerves in orientation to water flow in Tritonia diomedea. J Comp Physiol A 178:201-209. Nadakavukaren JJ, Lickey ME, Jordan WP (1986) Regulation of the circadian clock in the Aplysia eye: mimicry of neural activity by serotonin. J Neurosci 6:14-21. Nakamura H, Kojima S, Kobayashi S, Ito I, Fujito Y, Suzuki H, Ito E (1999) Physiological characterization of lip and tentacle nerves in Lymnaea stagnalis. Neurosci Res 33:291-298. Nargeot R, Baxter DA, Byrne JH (1997) Contingent-dependent enhancement of rhythmic motor patterns: an in vitro analog of operant conditioning. J Neurosci 17:8093-8105. Nargeot R, Baxter DA, Byrne JH (1999) In vitro analog of operant conditioning in Aplysia. II. Modifications of the functional dynamics of an identified neuron contribute to motor pattern selection. J Neurosci 19:2261-2272. Nedvetzki Y, Markovich S, Susswein AJ (1998) Inhibition of mating in Aplysia fasciata by food stimuli in the environment, but not by post-ingestion stimuli causing satiation. J Comp Physiol 182:183-190. Nolen TG, Johnson PM (2001) Defensive inking in Aplysia spp.: Multiple episodes of ink secretion and the adaptive use of a limited chemical resource. J Exp Biol 204:1257-1268. Nolen TG, Johnson PM, Kicklighter CE, Capo T (1995) Ink secretion by the marine snail Aplysia californica enhances its ability to escape from a natural predator. J Comp Physiol A 176:239-254. Norekian TP (1997) Coordination of startle and swimming neural systems in the pteropod mollusk Clione limacina: role of the cerebral cholinergic interneuron. J Neurophysiol 78:308-320. Norekian TP, Satterlie RA (1996) Whole body withdrawal circuit and its involvement in the behavioral hierarchy of the mollusk Clione limacina. J Neurophysiol 75:529-537. Olivera BM, Walker C, Cartier GE, Hooper D, Santos AD, Schoenfeld R, Shetty R, Watkins M, Bandyopadhyay P, Hillyard DR (1999) Speciation of cone snails and interspecific hyperdivergence of their venom peptides. Ann NY Acad Sci 870:223-237. Olson LM, Jacklet JW (1985) The circadian pacemaker in the Aplysia eye sends axons throughout the central nervous system. J Neurosci 5:3214-3227.

298

References

Ono JK, McCaman RE (1980) Identification of additional histaminergic neurons in Aplysia: improvement of single cell isolation techniques for in tandem physiological and chemical studies. Neurosci 5:835-840. Oppliger A, Hosken DJ, Ribi G (1998) Snail sperm production characteristics vary with sperm competition risk. Proc R Soc Lond B 265:1527-1534. Page LR (1992) New interpretation of a nudibranch central nervous system based on ultrastructural analysis of neurodevelopment in Melibe leonina. II. Pedal, pleural, and labial ganglia. Biol Bull 182:366-381. Painter SD, Chong MG, Wong MA, Gray A, Cormier JG, Nagle GT (1991) Relative contributions of the egg layer and egg cordon to pheromonal attraction and the induction of mating and egg-laying behavior in Aplysia. Biol Bull 181:81-94. Painter SD, Clough B, Garden RW, Sweedler JV, Nagle GT (1998) Characterization of Aplysia attractin, the first water-borne peptide pheromone in invertebrates. Biol Bull 194:120-131. Palovcik RA, Basberg BA, Ram JL (1982) Behavioral state changes induced in Pleurobranchaea and Aplysia by serotonin. Behav Neural Biol 35:383-394. Panchin YV, Popova LB, Deliagina TG, Orlovsky GN, Arshavsky YI (1995a) Control of locomotion in marine mollusk Clione limacina VIII. Cerebropedal neurons. J Neurophysiol 73:1912-1923. Panchin YV, Arshavsky YI, Deliagina TG, Popova LB, Orlovsky GN (1995b) Control of locomotion in marine mollusk Clione limacina IX. Neuronal mechanisms of spatial orientation. J Neurophysiol 73:1924-1937. Pearce TA, Gaertner A (1996) Optimal foraging and mucus-trail following in the carnivorous land snail Haplotrema concavum (Gastropoda:Pulmonata). Malacol Rev 29:85-99. Pennings SC (1991) Reproductive behavior of Aplysia californica Cooper: diel patterns, sexual roles and mating aggregations. J Exp Mar Biol Ecol 149:249-266. Pentreath VW, Berry MS, Osborne NN (1982) The serotonergic cerebral cells in gastropods. In Biology of Serotonergic Transmission, Osborne NN (ed.), Wiley, New York, pp. 457-513. Peretz B (1969) Central neuron initiation of periodic gill movements. Science 166: 1167-1172. Peretz B (1970) Habituation and dishabituation in the absence of a central nervous system. Science 169:379-381. Peretz B, Jacklet JW, Lukowiak K (1976) Habituation of reflexes in Aplysia: contribution of the peripheral and central nervous systems. Science 191:396-399. Perlman AJ (1979) Central and peripheral control of siphon-withdrawal reflex in Aplysia californica. J Neurophysiol 42:510-529. Perrins R, Weiss KR (1996) A cerebral central pattern generator in Aplysia and its connections with buccal feeding circuitry. J Neurosci 16:7030-7045. Peschel M, Straub V, Teyke T (1996) Consequences of food-attraction conditioning in Helix: a behavioral and electrophysiological study. J Comp Physiol 178:317-327. Phares GA, Lloyd PE (1996) Immunocytological and biochemical localization and biological activity of the newly sequenced cerebral peptide 2 in Aplysia. J Neurosci 16:7841-7852. Pin T, Gola M (1984) Axonal mapping of neurosecretory Helix bursting cells. Functional aspects of peripheral mulitbranched axons. Comp Biochem Physiol 79A:611-620.

References

299

Pinsker HM, Dudek FE (1977) Bag cell control of egg laying in freely behaving Aplysia. Science 197:490-493. Pinsker HM, Parsons DW (1985) Temperature dependence of egg laying in Aplysia brasiliana and A. californica. J Comp Physiol B 156:21-27. Plaut (2000) Nocturnalism in Aplysia oculifera (Adams & Reeve, 1850): an avoidance behavior minimizing exposure to ultraviolet radiation? The Veliger 43:105-109. Ponder WF, Lindberg DR (1997) Towards a phylogeny of gastropod molluscs: an analysis using morphological characters. Zool J Linn Soc 119:83-265. Popescu IR, Willows AOD (1999) Sources of magnetic sensory input to identified neurons active during crawling in the marine mollusc Tritonia diomedea. J Exp Biol 202:3029-3036. Poteryaev DA, Zakharov IS, Balaban PM, Belyavsky AV (1998) A novel neuropeptide precursor gene is expressed in the terrestrial snail central nervous system by a group of neurons that control mating behavior. J Neurobiol 35:183-197. Prato FS, Kavaliers M, Thomas AW, Ossenkopp, KP (1998) Modulatory actions of light on the behavioural responses to magnetic fields by land snails probably occur at the magnetic field detection stage. Proc R Soc Lond B 265:367-373. Prescott S, Chase R (1996) Two types of plasticity in the tentacle withdrawal reflex of Helix aspersa are dissociated by tissue location and response measure. J Comp Physiol A 179:407^14. Prescott SA, Chase R (1999) Sites of plasticity in the neural circuit mediating tentacle withdrawal in the snail Helix aspersa: implications for behavioral change and learning kinetics. Learn Mem 6:363-380. Prescott SA, Gill N, Chase R (1997) The neural circuit mediating tentacle withdrawal in Helix aspersa, with specific reference to the competence of the motor neuron C3. J Neurophysiol 78:2951-2965. Price DA, Lesser W, Lee TD, Doble KE, Greenberg MJ (1990) Seven FMRFamiderelated and two SCP-related cardioactive peptides from Helix. J Exp Biol 154:421^37. Prince J, Nolen TG, Coelho L (1998) Defensive ink pigment processing and secretion in Aplysia californica: Concentration and storage of phycoerythrobilin in the ink gland. J Exp Biol 201:1595-1613. Prior DJ, Hume M, Varga D, Hess SD (1983) Physiological and behavioural aspects of water balance and respiratory function in the terrestrial slug, Limax maximus. J Exp Biol 104:111-127. Prior DJ, Welsford IG, Banta PA (1989) Ingestion of substrate fluid by Helix aspersa Muller: a feeding response induced by low molecular weight chemical stimuli. Comp Biochem Physiol 94A:73-74. Quinlan EM, Murphy AD (1996) Plasticity in the multifunctional buccal central pattern generator of Helisoma illuminated by the identification of phase 3 interneurons. J Neurophysiol 75:561-574. Ram JL, Gallardo CS, Ram ML, Croll RP (1998) Reproduction-associated immunoreactive peptides in the nervous systems of prosobranch gastropods. Biol Bull 195:308-318. Ratte S, Chase R (2000) Synapse distribution of olfactory interneurons in the procerebrum of the snail Helix aspersa. J Comp Neurol 417:366-384. Rayport SG, Ambron RT, Babiarz J (1983) Identified cholinergic neurons R2 and LPlj control mucus release in Aplysia. J Neurophysiol 49:864-876. Redman RS, Berry RW (1991) Temperature-dependent peptidergic feedback: potential role in seasonal egg-laying in Aplysia. J Neurosci 11:1780-1785.

300

References

Rittenhouse AR, Price CH (1986) Anatomical and electrophysiological study of multitransmitter neuron R14 of Aplysia. J Comp Neurol 247:447^56. Roberts MH, Block GD (1985) Analysis of mutual circadian pacemaker coupling between the two eyes of Bulla. J Biol Rhythms 1:55-75. Rogers D, Chase R (2001) Dart receipt promotes sperm storage in the garden snail Helix aspersa. Behav Ecol Sociobiol 50:122-127. Rollo CD, Wellington WG (1979) Intra- and inter-specific agonistic behavior among terrestrial slugs (Pulmonata:Stylommatophora). Can J Zool 57:846-855. Rose RM, Benjamin PR (1979) The relationship of the central motor pattern to the feeding cycle of Lymanea stagnalis. J Exp Biol 80:137-163. Rosen SC, Weiss KR, Goldstein RS, Kupfermann I (1989) The role of a modulatory neuron in feeding and satiation in Aplysia: effects of lesioning of the serotonergic metacerebral cells. J Neurosci 9:1562-1578. Rosen SC, Teyke T, Miller MW, Weiss KR, Kupfermann I (1991) Identification and characterization of cerebral-to-buccal interneurons implicated in the control of motor programs associated with feeding in Aplysia. J Neurosci 11:3630-3655. Rosen SC, Miller MW, Evans CG, Cropper EC, Kupfermann I (2000) Diverse synaptic connections between peptidergic radula mechanoafferent neurons and neurons in the feeding system of Aplysia. J Neurophysiol 83:1605-1620. Roubos EW, Van de Ven AMH (1987) Morphology of neurosecretory cells in basommatophoran snails homologous with egg-laying and growth hormoneproducing cells of Lymnaea stagnalis. Gen Comp Endocrinol 67:7-23. Royer S, Coulson RL, Klein M (2000) Switching off and on of synaptic sites at Aplysia sensorimotor synapses. J Neurosci 20:626-638. Runnegar B, and Pojeta J (1985) Origin and diversification of the Mollusca. In The Mollusca, vol 10, Trueman ER, Clarke MR (eds), Academic Press, Orlando, pp. 1-57. Sadreyev RI, Panchin YV (2000) The role of putative glutamatergic neurons and their connections in the locomotor central pattern generator of the mollusk, Clione limacina. Comp Biochem Physiol A 126:193-202. Sahley C, Gelperin A, Rudy JW (1981) One-trial associative learning modifies food odor preferences of a terrestrial mollusc. Proc Natl Acad Sci USA 78:640-642. Sakharov DA (1991) Use of transmitter precursors in gastropod neuroethology. In Molluscan Neurobiology, Kits KS, Boer HH, Joosse J (eds), North-Holland, Amsterdam, pp. 236-242. Sakharov DA, Milosevic I, Salimova N (1989) Drug-induced locomotor stereotypies in Aplysia. Comp Biochem Physiol 93C:161-166. Satterlie RA (1985) Reciprocal inhibition and postinhibitory rebound produce reverberation in a locomotor pattern generator. Science 229:402^04. Satterlie RA (1991) Neural control of speed changes in an opisthobranch locomotory system. Biol Bull 180:228-233. Satterlie RA (1993) Neuromuscular organization in the swimming system of the pteropod mollusc Clione limacina. J Exp Biol 181:119-140. Satterlie RA, Norekian TP (1996) Modulation of swimming speed in the pteropod mollusc, Clione limacina: role of a compartmental serotonergic system. Invert Neurosci 2:157-165. Satterlie RA, LaBarbera M, Spencer AN (1985) Swimming in the pteropod mollusc, Clione limacina. I. Behaviour and morphology. J Exp Biol 116:189-204.

References

301

Satterlie RA, Norekian TP, Pirtle TJ (2000) Serotonin-induced spike narrowing in a locomotor pattern generator permits increases in cycle frequency during accelerations. J Neurophysiol 83:2163-2170. Sawada M, Blankenship JE, McAdoo DJ (1981) Neural control of a molluscan blood vessel, anterior aorta of Aplysia. J Neurophysiol 46:967-986. Scheller RH, Jackson JF, McAllister LB, Rothman BS, Mayeri E, Axel R (1983) A single gene encodes multiple neuropeptides mediating a stereotyped behavior. Cell 35:7-22. Schmalz E (1914) Zur Morphologic des Nervensystems von Helix pomatia L. Z Wiss Zool 111:506-568. Schmekel L (1985) Aspects of evolution within the opisthobranchs. In The Mollusca, vol 10, Trueman ER, Clarke MR (eds), Academic Press, Orlando, pp. 221-267. Schutt A, Basar E, Bullock TH (1999) Power spectra of ongoing activity of the snail brain can discriminate odorants. Comp Biochem Physiol A 123:95-110. Schwarz M, Susswein AJ (1986) Identification of the neural pathway for reinforcement of feeding when Aplysia learn that food is inedible. J Neurosci 6:1528-1536. Scott ML, Brezina V, Weiss KR (1997) Ion currents and mechanisms of modulation in the radula opener muscles of Aplysia. J Neurophysiol 78:2372-2387. Shope SB, McPherson D, Rock MK, Blankenship JE (1991) Functional and morphological evidence for the existence of neurites from abdominal ganglion bag cell neurons in the head-ring ganglia of Aplysia. J Comp Physiol A 168:539-552. Sigvardt KA, Rothman BS, Brown RO, Mayeri E (1986) The bag cells of Aplysia as a multitransmitter system: identification of alpha bag cell peptide as a second neurotransmitter. J Neurosci 6:803-813. Skelton M, Alevizos A, Koester J (1992) Control of the cardiovascular system of Aplysia by identified neurons. Experientia 48:809-817. Smelik T (1995) The natural firing pattern of peptidergic neurones in the pond snail Lymnaea stagnalis. PhD thesis, Vrije Universiteit, Amsterdam. Smit AB, Syed NI, Schaap D, van Minnen J, Klumperman J, Kits KS, Lodder H, van der Schors, RC, van Elk R, Sorgedrager B, Brejc K, Sixma TK, Geraerts PM (2001) A glia-derived acetylcholine-binding protein that modulates synaptic transmission. Nature 411:261-268. Snyder NFR (1967) An alarm reaction of aquatic gastropods to intraspecific extract. Cornell Univ Agr Sta Mem 403. Soffe SR, Benjamin PR, Slade CT (1978) Effects of environmental osmolarity on blood composition and light microscope appearance of neurosecretory neurones in the snail, Lymnaea stagnalis (L). Comp Biochem Physiol 61A:577-584. Sokolove PG, Beiswanger CM, Prior DJ, Gelperin A (1977) A circadian rhythm in the locomotor behaviour of the giant garden slug Limax maximus. J Exp Biol 66:47-64. Solem A (1974) The Shell Makers, Wiley, New York. Sossin WS, Kirk MD, Scheller RH (1987) Peptidergic modulation of neuronal circuitry controlling feeding in Aplysia. J Neurosci 7:671-681. Spengel JW (1881) Die Geruchsorgane und das Nervensystem der Mollusken. Z Wiss Zool 35:333-383. Staras K, Kemenes G, Benjamin PR (1998) Pattern-generating role for motoneurons in a rhythmically active neuronal network. J Neurosci 18:3669-3688. Staras K, Kemenes G, Benjamin PR (1999a) Electrophysiological and behavioral analysis of lip touch as a component of the food stimulus in the snail Lymnaea. J Neurophysiol 81:1261-1273.

302

References

Staras K, Kemenes G, Benjamin PR (1999b) Cellular traces of behavioral classical conditioning can be recorded at several specific sites in a simple nervous system. J Neurosci 19:347-357. Stinnakre J, Tauc L (1969) Central neuronal response to the activation of the osmoreceptors in the osphradium of Aplysia. J Exp Biol 51:347-361. Stoll CJ (1976) Extraocular photoreception in Lymnaea stagnalis (L.). In Neurobiology of Invertebrates: Gastropoda Brain, Salanki J (ed.), Akademia Kiado, Budapest, pp. 487^95. Stommel EW Stephens RE, Alkon DL (1980) Motile statocyst cilia transmit rather than directly transduce mechanical stimuli. J Cell Biol 87:652-662. Stopfer M, Carew TJ (1996) Heterosynaptic facilitation of tail sensory neuron synaptic transmission during habituation in tail-induced tail and siphon withdrawal reflexes of Aplysia. J Neurosci 16:4933-4948. Strong JA, Fox AP, Tsien RW, Kaczmarek LK (1987) Stimulation of protein kinase C recruits covert calcium channels in Aplysia bag cell neurons. Nature 325:714-717. Susswein AJ, Kupfermann I (1975) Bulk as a stimulus for satiation in Aplysia. Behav Biol 13:203-209. Susswein AJ, Gev S, Feldman E, Markovich S (1983) Activity patterns and time budgeting of Aplysia fasciata under field and laboratory conditions. Behav Neural Biol 39:203-220. Susswein AJ, Gev S, Achituv Y, Markovich S (1984) Behavioral patterns of Aplysia fasciata along the Mediterranean coast of Israel. Behav Neural Biol 41:7-22. Susswein AJ, Schwarz M, Feldman E (1986) Learned changes of feeding behavior in Aplysia in response to edible and inedible foods. J Neurosci 6:1513-1527. Suzuki H, Watanabe M, Tsukahara Y, Tasaki K (1979) Duplex system in the simple retina of a gastropod mollusc, Limaxflavus L. J Comp Physiol A 133:125-130. Syed NI, Bulloch AGM, Lukowiak K (1990) In vitro reconstruction of the respiratory central pattern generator of the mollusk Lymnaea. Science 250:282-285. Syed NI, Harrison D, Winlow W (1991) Respiratory behavior in the pond snail Lymnaea stagnalis. I. Behavioral analysis and the identification of motor neurons. J Comp Physiol A 169:541-555. Syed NI, Winlow W (199la) Respiratory behavior in the pond snail Lymnaea stagnalis. II. Neural elements of the central pattern generator (CPG). J Comp Physiol A 169:557-568. Syed NI, Winlow W (1991b) Coordination of locomotor and cardiorespiratory networks of Lymnaea stagnalis by a pair of identified interneurones. J Exp Biol 158:37-62. Takeda N, Ozaki T (1986) Induction of locomotor behavior in the giant african snail, Achatina fulica. Comp Biochem Physiol 83A:77-82. Takeda N, Tsuruoka H (1979) A sex pheromone secreting gland in the terrestrial snail, Euhadra peliomphala. J Exp Zool 207:17-26. Ter Maat A, Ferguson GP (1996) Neuronal input contributes to sequence of Aplysia egg laying behaviors. J Comp Physiol A 179:775-783. Teyke T, Kupfermann I (1989) A subpopulation of cerebral B cluster neurones of Aplysia californica is involved in defensive head withdrawal but not appetitive head movements. J Exp Biol 147:1 20. Teyke T, Weiss KR, Kupfermann I (1990a) Appetitive feeding behavior of Aplysia: behavioral and neural analysis of directed head turning. J Neurosci 10:3922-3934. Teyke T, Weiss KR, Kupfermann I (1990b) An identified neuron (CPR) evokes neuronal responses reflecting food arousal in Aplysia. Science 247:85-87.

References

303

Thiele J (1929-1931) Handbuch der Systematishchen Weichtierkunde, vol. 1, Jena. Thomas JD, Assefa F, Cowley C, Ofosu-Barko J (1980) Behavioural responses to amino acids and related compounds, including proprionic acid, by adult Biomphalaria glabrata (Say), a snail host of Schistosoma mansoni. Comp Biochem Physiol66C: 17-27. Thomaz D, Guiller A, Clarke B (1996) Extreme divergence of mitochondrial DNA within species of pulmonate land snails. Proc R Soc Lond B 263:363-368. Thompson JT, Lowe AD, Kier WM (1998) The columellar muscle of prosobranch gastropods: morphological zonation and its functional implications. Invert Biol 117:45-56. Tomiyama K (1992) Homing behaviour of the giant african snail, Achatina fulica (Ferussac) (Gastropoda; Pulmonata). J Ethol 10:139-147. Tompa AS (1984) Land snails (Stylommatophora). In The Mollusca, vol 7, Tompa AS, Verdonk NH, van den Biggelaar JAM (eds), Academic Press, Orlando, pp. 45-140. Trappman W (1916) Die Muskulatur von Helix pomatia L. Zugleich ein Beitrag zur Kenntnis der Locomotion unserere einheimishchen Pulmonaten. Z Wiss Zool 115:489-585. Tremblay JP, Colonnier M, McLennan H (1979) An electron microscope study of synaptic contacts in the abdominal ganglion of Aplysia californica. J Comp Neurol 188:367-390. Trudeau L-E, Castellucci VF (1992) Contribution of polysynaptic pathways in the mediation and plasticity of Aplysia gill and siphon withdrawal reflex: evidence for differential modulation. J Neurosci 12:3838-3848. Trudeau L-E, Castellucci VF (1993) Sensitization of the gill and siphon withdrawal reflex of Aplysia: multiple sites of change in the neuronal network. J Neurophysiol 70:1210-1220. Trueman ER (1983) Locomotion in molluscs. In The Mollusca, vol 4, part 1, Saleuddin ASM, Wilbur KM (eds), Academic Press, New York, pp. 155-198. Tuersley MD (1989) How is food arousal manifested in the pond snail Lymnaea stagnalis! An overview. J Moll Stud 55:209-216. Ungless MA (1998) A Pavlovian analysis of food-attraction conditioning in the snail Helix aspersa. Anim Learn Behav 26:15-19. Van Minnen J, Smit AB, Joosse J (1989) Central and peripheral expression of genes coding for egg-laying inducing and insulin-related peptides in a snail. Arch Histol Cytol 52:241-252. Van Minnen J, Schallig HDFH, Ramkema MD (1992) Identification of putative egg-laying hormone containing neuronal systems in gastropod molluscs. Gen Comp Endocrinol 86:96-102. Van Mol J-J (1974) Evolution phylogenetique du ganglion cerebroi'de des Gasteropodes Pulmones. Haliotis 4:77-86. Vehovszky A, Elliott CJH (2001) Activation and reconfiguration of fictive feeding by the octopamine-containing modulatiory OC interneurons in the snail Lymnaea. J Neurophysiol 86:792-808. Villim FS, Cropper EC, Price DA, Kupfermann I, Weiss KR (2000) Peptide cotransmitter release from motorneuron B16 in Aplysia californica: costorage, corelease, and functional implications. J Neurosci 20:2036-2042. Voss M, Kottowski K, Wiinnenberg W (1997) Neurophysiological indication of a diurnal rhythmicity in chemosensitivity of the snail Helix pomatia. Comp Biochem Physiol 116A:37^2.

304

References

Vreugdenhil E, Jackson JF, Bouwmeester T, Smit AB, Van Minnen J, Van Heerikhuizen H, Klootwijk J, Joosse J (1988) Isolation, characterization, and evolutionary aspects of a cDNA clone encoding multiple neuropeptides involved in the stereotyped egg-laying behavior of the freshwater snail Lymnaea stagnalis. JNeurosci 8:4184-4191. Wade CM, Mordan PB, Clarke B (2001) A phylogeny of the land snails (Gastropoda: Pulmonata). Proc R Soc Lond B 268:413-422. Walters ET (1987) Site-specific sensitization of defensive reflexes in Aplysia: a simple model of long-term hyperalgesia. J Neurosci 7:400^07. Walters ET (1991) A functional, cellular, and evolutionary model of nociceptive plasticity in Aplysia. Biol Bull 180:241-251. Walters ET, Erickson MT (1986) Directional control and the functional organization of defensive responses in Aplysia. J Comp Physiol A 139:339-351. Walters ET, Carew TJ, Kandel ER (1981) Associative learning in Aplysia: evidence for conditioned fear in an invertebrate. Science 211:504-506. Walters ET, Byrne JH, Carew TJ, Kandel ER (1983) Mechanoafferent neurons innervating tail of Aplysia. I. Response properties and synaptic connections. J Neurophysiol 50:1522-1542. Wareing DR (1986) Directional trail following in Deroceras reticulatum (Muller). J Moll Stud 52:256-258. Watanabe W, Kawahara S, Kirino Y (1998) Morphological characterization of the bursting and nonbursting neurones in the olfactory centre of the terrestrial slug Limax marginatus. J Exp Biol 201:925-930. Wayne NL, Nick T, Block GD (1996) Effects of temperature on reproductive neuroendocrine function in Aplysia californica. Gen Comp Endocrinol 102:351-359. Wedemeyer H, Schild D (1995) Chemosensitivity of the osphradium of the pond snail Lymnaea stagnalis. J Exp Biol 198:1743-1754. Weiss KR, Cohen JL, Kupfermann I (1978) Modulatory control of buccal musculature by a serotonergic neuron (metacerebral cell) in Aplysia. J Neurophysiol 41:181-203. Weiss KR, Bayley H, Lloyd PE, Tenenbaum R, Kolks MA, Buck L, Cropper EC, Rosen SC, Kupfermann I (1989) Purification and sequencing of neuropeptides contained in neuron R15 of Aplysia californica. Proc Natl Acad Sci USA 86:2913-2917. Whelan HA, McCrohan CR (1996) Food-related conditioning and neuronal correlates in the freshwater snail Lymnaea stagnalis. J Moll Stud 62:483-494. Wiens BL, Brownell PH (1995) Neurotransmitter regulation of the heart in the nudibranch Archidoris montereyensis. J Neurophysiol 74:1639-1651. Willows AOD (1967) Behavioral acts elicited by stimulation of single, identifiable brain cells. Science 157:570-574. Willows AOD (1971) Giant brain cells in mollusks. Sci Am 224:69-75. Willows AOD (1999) Shoreward orientation involving geomagnetic cues in the nudibranch mollusc Tritonia diomedea. Mar Fresh Behav Physiol 32:181-192. Willows AOD, Pavlova GA, Phillips NE (1997) Modulation of ciliary beat frequency by neuropeptides from identified molluscan neurons. J Exp Biol 200:1433-1439. Wilson GF, Kaczmarek LK (1993) Mode-switching of a voltage-gated cation channel is mediated by a protein kinase A-regulated tyrosine phosphatase. Nature 366:433^38. Wilson GF, Richardson FC, Fisher TE, Olivera BM, Kaczmarek LK (1996) Identification and characterization of a Ca2+-sensitive nonspecific cation

References

305

channel underlying prolonged repetitive firing in Aplysia neurons. J Neurosci 16:3661-3671. Wolff HG (1975) Statocysts and geotactic behaviour in gastropod molluscs. Fortschr Zool 23:63-84. Wright WG, Kirschman D, Rozen D, Maynard B (1996) Phylogenetic analysis of learning-related neuromodulation in molluscan mechanosensory neurons. Evolution 50:2248-2263. Xin Y, Weiss KR, Kupfermann I (1995) Distribution in the central nervous system of Aplysia of afferent fibers arising from cell bodies located in the periphery. J Comp Neurol 359:627-643. Xin Y, Weiss KR, Kupfermann I (1996) A pair of identified interneurons in Aplysia that are involved in multiple behaviors are necessary and sufficient for the arterialshortening component of a local withdrawal reflex. J Neurosci 16:4518^-528. Xin Y, Koester J, Jing J, Weiss KR, Kupfermann I (2001) Cerebral-abdominal interganglionic coordinating neurons in Aplysia. J Neurophysiol 85:174-186. Yeoman MS, Kemenes G, Benjamin PR, Elliott CJH (1994a) Modulatory role for the serotonergic cerebral giant cells in the feeding system of the snail, Lymnaea. II. Photoinactivation. J Neurophysiol 72:1372-1382. Yeoman MS, Pieneman AW, Ferguson GP, Ter Maat A, Benjamin PR (1994b) Modulatory role for the serotonergic cerebral giant cells in the feeding system of the snail, Lymnaea. I. Fine wire recording in the intact animal and pharmacology. J Neurophysiol 72:1357-1371. Yeoman MS, Brierley MJ, Benjamin PR (1996) Central pattern generator interneurons are targets for the modulatory serotonergic cerebral giant cells in the feeding system of Lymnaea. J Neurophysiol 75:11-25. Yusa Y (1996) The effects of body size on mating features in a field population of the hermaphroditic sea hare Aplysia kurodai Baba, 1937 (Gastropoda: Opisthobranchia). J Moll Stud 62:381-386. Zack S (1974) The effects of food deprivation on agonistic behavior in an opisthobranch mollusc, Hermissenda crassicornis. Behav Biol 12:223-232. Zaitseva OV, Bocharova LS (1981) Sensory cells in the head skin of pond snails. Cell Tiss Res 220:797-807. Zakharov IS (1994) Avoidance behavior of the snail. Neurosci Behav Physiol 24:63-69. Zakharov IS, lerusalimsky VN, Balaban PM (1995) Pedal serotonergic neurons modulate the synaptic input of withdrawal interneurons of Helix. Invert Neurosci 1:41-52. Zann LP (1973) Interactions of the circadian and circatidal rhythms of the littoral gastropod Melanerita atramentosa Reeve. J Exp Mar Biol Ecol 11:249-261. Zhuravlev V, Bugaj V, Kodirov S, Safonova T, Staruschenko A (2001) Giant multimodal heart motoneurons of Achatina fulica: a new cardioregulatory input in pulmonates. Comp Biochem Physiol A 130:183-196. Zhuravlev VL, Bugaj V, Safaonova T, Senkov, D, Kodirov S (1999) The chronoinotropic effects of new regulatory input to the heart of land pulmonates. Acta Biol Hung 50:309-318. Ziv I, Markovich S, Lustig C, Susswein AJ (1991a) Effects of food and mates on time budget in Aplysia fasciata: integration of feeding, reproduction, and locomotion. Behav Neural Biol 55:68-85. Ziv I, Lustig C, Ben-Zion M, Susswein AJ (1991b) Daily variation of multiple behaviors in Aplysia fasciata: integration of feeding, reproduction, and locomotion. Behav Neural Biol 55:86-107.

This page intentionally left blank

Taxonomic Index Accra, 25 Achatina, 16, 20, 27, 32, 80-82, 91, 114-115, 130-134, 137, 157, 160, 173, 183, 214-215, 217, 254, 267 Acmaea, 126, 216-217, 257 Acteonidae, 12 Akera, 99, 241 Anaspidea, 13, 24, 99, 128, 215-216, 224, 240, 247-248

Aplysia, vii, ix-x, 13, 20, 23, 27-29, 36, 42-43, 46, 48, 50, 55-56, 60-63, 66, 71-72, 74-78, 82-83, 87-91, 94, 96-97, 99, 108, 111, 114, 122-125, 128, 135, 145, 147, 149-151, 153-154, 156-158, 160-161, 166-167, 169, 172-174, 176, 181-182, 184, 189, 193-195, 197-198, 200-209, 212, 214-217, 222, 224-226, 228-229, 231, 233-235, 237-238, 240-248, 254-255, 259-263, 266-274, 278 brasiliana, 99-100, 107-108, 110, 112-114, 118,205, 216, 255 dactylomela, 128, 245, 255 depilans, 110 californica, 26, 58, 61, 100, 108, 110, 112-113, 118, 128, 132, 135, 151, 173, 176, 181, 184, 194, 197, 201-202, 205, 240, 246-247, 255, 260, 268, 270 fasciata, 99-100, 110, 167, 173, 176, 194, 246, 255 Juliana, 126, 247, 255 kurodai, 174 oculifera, 255 parvula, 203, 255 saltator, 99 vaccaria, 4, 172, 202, 247, 255 Aplysiidae, 12 Archaeogastropoda, 10, 12. See also Archaeogastropoda-Vetigastropoda Archaeogastropoda-Vetigastropoda, 11, 22, 126, 195 Archidoris, 76-77 Ariolimax, 179

Avion, 189, 215, 259 Arionidae, 183 Basommatophora, 15, 25, 86, 140, 146, 196, 199 Bellerophontacea, 8 Biomphalaria, 37, 131-132 Buccinum, 12, 127, 215 Bulinus, 189, 259 Bulla, 44-46, 240-241, 255, 260-261, 263, 266 Busy con, 12, 127, 201, 215 Caenogastropoda, 10-11, 22, 170-171, 196 Cavoliniidae, 101 Cepaea, 15-16, 52, 173, 176, 202, 238 Cephalaspidea, 13, 21, 25, 129, 216 Clione, 13, 50, 72, 77, 93, 103, 106, 107, 110-113, 116-118, 128-129, 146, 215, 221-222, 267, 273-275, 277, 279 antarctica, 181 limacina, 80, 101, 104, 113, 129, 136, 181 Concholepas, 201 Conus, 12, 127, 195, 207-208, 215-216 Coryphella, 25 Cypraea, 95 Deroceras, 163, 183, 259 Dolabrifera, 241-242 Ercolania, 273 Euglandina, 131, 134-136, 214 Euhadra, 183 Euthyneura, 10, 21-22 Glaucus, 178 Gonaxis, 131, 214 Gymnarion, 183 Gymnosomata, 13, 101, 129 Haliotis, 126, 202 Haplotrema, 135

307

308 Taxonomic Index Helisoma, 37, 94, 125, 130, 144-145, 147, 156, 279 Helix, 16, 27, 46, 60, 62-63, 65, 74-75, 81-82, 91,97, 118, 130, 140, 146, 164, 173, 194-196, 219, 221-223, 244, 254, 259, 262, 267-269, 271, 273, 277, 279 aperta, 183 aspersa, 75, 115, 137, 179, 189-193, 201-202, 214, 222, 250, 253, 268 aspersa maxima, 262 lucorum, 218 pomatia, viii, 26, 53, 67, 71, 116, 175, 191, 214, 250, 262 Hermissenda, 14, 43, 45^6, 48, 50, 120, 122, 132-133, 215 Heteropoda, 43^4, 100-101, 127, 131, 134 Heterostropha, 12 Hexabranchus, 99 Indiana, 158 Limacina, 101, 128-129 Umax, 16, 46, 97, 130, 154, 164, 189 corsicus, 179 maximus, 95, 132-133, 137-138, 163, 179, 201, 259-260 pseudoflavus, 258 redii, 179 Littorina, 11, 119-120, 126, 189 irrorata, 44 obtusata, 171 pintado, 171 Lymnaea, 15, 29, 36, 40, 44, 50, 67-68, 70-71, 75-78, 81, 86, 89-94, 116, 124-125, 130, 140-145, 147-148, 151-152, 154, 156-157, 159-160, 164-166, 169, 173-174, 184-189, 190, 193-196, 198-199, 201-204, 208-209, 211-212, 218-221, 245, 253-254, 267, 271, 273, 275, 279 Melampus, 88 Melanerita, 256 Melibe, 21 Mesogastropoda, 10-12, 22, 126 Mitra, 182 Monodonta, 126 Nassa, 217, 219 Nassarius, 51 Natica, 134 Navanax, 129, 134-135, 174-175, 215-216, 246

Neogastropoda, 10-12, 22, 126-127, 215 Nerita, 126 Notaspidea, 13-14, 99, 104, 224, 245 Nudibranchia, 13-14, 24-25, 43, 45, 93, 99, 104, 133, 172, 216-217, 223-224 Otala, 91, 119,252,262 lactea, 253 vermiculata, 183 Partula, 174 Patella, 126, 257 Pentaganglionata, 22 Phestilla, 132 Philine, 58, 129 Phylliroidae, 99 Physa, 173, 176 Planorbis, 93-94, 145, 147, 149, 219, 221, 279 Pleurobranchaea, 14, 99, 103-105, 112, 125, 129, 146, 148, 150, 153-154, 156, 164, 172, 201, 215, 245, 267, 270-273, 275, 279-280 Pleurotomaria, 11 Pteropoda, 13, 80, 101, 128, 221 Rumina, 173 Sacoglossa, 14, 24, 128, 172, 177 Siphonaria, 257 Streptoneura, 21 Strombus, 11, 45, 196 gibberulus, 216 gigas, 214 Strophocheilus, 196 Stylommatophora, 15-16, 29, 136, 222 Succinea, 136 Systellommatophora, 15, 23, 25 Tectibranchia, 13, 20 Tegula, 201 Thecosomata, 13, 101, 128-129 Tritonia, 14, 45, 51-52, 94, 99, 103-105, 107-108, 110-112, 114, 122-123, 125, 129, 145, 223-224, 273 Turbo, 44 Umbraculum, 14, 128 Urosalpinx, 126 Vallonia, 196 Vetigastropoda, 10. See also Archaeogastropoda-Vetigastropoda Viviparus, 111

Neuron Index L9, 230, 236 L10 (interneuron I), 78, 80, 82, 88-89, 157 LI4, 247 L16, 231 L21, 231 L24 (interneuron XI), 78 L29, 231-232, 237, 239-240 L30, 231,240 L31, 247 L33a/b, 85, 233 L34, 239 L35, 231 LBS, 85, 231, 233 LBVC, 73-74, 76, 157, 231 LDG1, 227, 237, 240 LDHI, 76 LDS1, 228, 231 LE cluster, 229-230, 233, 237, 240, 243, 247 LFS cluster, 83, 228, 230, 239, 243 LFSA/SB, 228, 230 LP1, 20, 245-246 LP117, 233-234 LUQ (L2-L6), 88-89 M cluster, 157 MGC (MCC), 155-158, 160, 165 PAS, 74, 78 Pll, 157 POP, 112-113 R2, 20, 29-30, 157, 245-246 R3-R14, 76 R15, 37, 74, 84-85, 90-92, 210-211, 254, 262 R19, 247 R20, 84-85 R25/L25 (interneuron II), 78, 84-85, 226 RB HE , 76, 78 RDS, 228, 230-231 RE cluster, 230, 247 RF cluster, 230, 247 VC cluster, 233, 247

Achatina d-RPLN, 80, 82 d-VLN, 80, 82 GCN. See Aplysia, MGC PON, 91, 254. See also Aplysia, R15 Aplysia A cluster, 139, 140 B cluster, 139, 140 B3, 161, 165-166 B4/5, 149, 154-155, 166 B8, 145 B9, 161 B15, 161, 163 B16, 161, 163 B20, 154-155 B21,158-159 B22, 158 B31/B32, 145, 153, 166 B34, 145, 147, 149 B51, 149, 167, 169 B63, 145, 147, 153 B64, 145, 147 B65, 147, 149, 153 bag cells, 27, 197-201, 203, 205-209, 211, 213 Bn, 157 C cluster, 108 C2, 158 CBI-1, 150, 158 CBI-2, 150, 154, 156, 158 CBI-3, 154-155, 158 CBI-5/6, 158 CBI-8/9, 158 CC2, 89 CC3, 78, 89 CC5, 74, 78 CC7, 73, 78 C-PR, 157-158, 278 F cluster, 108, 205 H cluster, 194 L7, 27, 63, 78, 82, 85, 157, 227, 230, 233, 237, 240, 248

309

310

Neuron Index

Bulla BRN, 263, 266 H, 46 R, 46 Clione CPB2, 116 CPB3, 116, 118 Cr-St, 277 P1B, 279 Pl-W, 221-222, 279 type 7/8, 105 type 12, 111 Helisoma B2, 147 BCN1, 147 Helix Cl (GSC, SCC). See Aplysia, MGC C3, 222-223 cerebral green, 262 Pa3, 82 RPal (Fl, Br), 91, 254. See also Aplysia, R15 Hermissenda C, 47 photoreceptor type A, 46^7, 120-122 photoreceptor type B, 46, 120-122 Lymnaea A cluster, 219, 221 B3, 144 B4, 141 B7a, 141 CDCs, 199-203, 207, 209, 212 CGC. See Aplysia, MGC CV1, 165 E cluster, 275 Ehe, 75 F cluster, 275 Hhe, 76, 81 Ip3 (Ip3I), 68, 70, 81, 89 K hi , 76, 78, 81 L/RPeDl 1,81, 221 M cluster, 275 NIL, 143-144, 147 N1M, 143-144, 153, 165 N2d, 144 N2v, 144-145, 157

N3p/N3t, 144, 147 OC, 144, 147, 153 Pelb cluster, 187 RPD2, 81, 89, 91-92, 254 RPeDl, 68-70, 81, 253 RPeN cluster, 211 rAL, 187-189 rVL, 187 She, 76-77 SO, 141-144, 152-153, 156 VD1, 81, 91 VD2/3, 253 VD4, 70, 81 VH cluster, 253 VI cluster, 253 VJ cluster, 68, 81, 253 VK cluster, 81, 253 yellow/yellow-green/dark green (Alcian stained), 76, 90 Otala cell 11, 91, 253-254, 262. See also Aplysia, R15 Planorbis Al, 94 DRN1, 221 Pleurobranchaea Al, 104 A10, 104 anterior ventral cells, 150 Asl^, 104 B3, 150 CDs, 275 11, 279 12, 164, 280 IVs, 104 PCp, 279-280 PSE, 164, 279-280 ventral white cell, 150, 153 Tritonia C2, 103-104, 109-110 DRI, 104, 110 DSIs, 103-104, 108, 110 Pd5, 94, 122-123 Pd6, 94 Pd21, 94 VSIs, 103-104, 110

Subject Index Command cells, 78, 80, 82, 104, 108, 150, 152-155, 157-158, 164, 198, 205, 211, 221, 223, 240, 277-280 Body wall, 53, 76, 82, 108, 113, 145, 177, 187, Conditioning of aerial breathing, 67 218, 221-222, 245, 247, 275 appetitive, 132-133, 165-166 Buccal mass, 53-54, 56, 58, 72-73, 124-126, classical (pavlovian), 132-134, 164, 169, 128-130, 140, 144-145, 148-149, 155, 158, 243-244 218-219, 271 of fear response, 244 Burrowing, 12-13, 215-216, 252 of feeding, 163-169 food aversion (food avoidance), 133-134, Cardiovascular system, 71-82 164-165, 244, 280 blood vessels, 14, 26-27, 67-68, 83, 186 of gill withdrawal, 234-240, 243 heart, 6, 68, 83, 87-88, 91, 151, 157, 193, non-associative. See Habituation, 246 Sensitization Cavity operant, 67, 167-169 buccal, 124-125, 127-129, 140 of phototaxis, 120-122 mantle (pallial), 5-6, 8, 10, 12-15, 20, of siphon withdrawal, 235, 237-243 35-37, 67-68, 83-84, 126, 171, 226, of tail withdrawal, 238 247-248 of tentacle withdrawal, 223, 224 Central nervous system (CNS), vii, 17-33, 37, Copulation, 170-174, 176, 179, 181-182, 40-41, 50, 54, 56, 59-65, 68, 70-71, 80, 88, 184-195 91, 93, 97, 103, 123, 136-137, 147, 153, 187, Courtship, 176, 178-181, 183-184, 190-194 189, 194-195, 201, 203, 205, 212, 218-219, Crawling, 93-99, 108, 113, 122-123 222, 226-228, 230, 233, 236, 245, 252-253, 260-261, 272 Dart, 179-181, 190-193, 278 Central pattern generator (CPG) feeding, 140-147, 149-150, 153-154, 157, 160, Detorsion, 6, 19-23 164, 166-167, 169, 275 Egg laying, 195-213 respiratory, 68-70, 253 Eggs, 171, 174-177, 182, 195-198 swimming, 103-107, 110-111, 116, 261 Epiphragm, 86, 242, 250 whole-body withdrawal, 221 Escape response, 7, 34, 72, 83, 94, 96, 99, 101, Chemoreception, 34-38, 40-42, 131-135, 139, 104, 136, 216-217, 220, 226, 244, 270-271, 183-184, 216-217, 252 273, 275, 279. See also Reflexes, Startle Cilia, 13, 42, 45^9, 93-94, 99, 122-123, 126, response 128, 171 Estivation, 86, 252-253 Circadian clock, 45^6, 259-266 Euthyneury, 7, 12, 20-23 Circadian rhythm, 42-43, 51, 250, 254-259, Excretion, 85-92 268, 271-272 Classification (of the Gastropoda), 5, 9-12, 16, Eyes, 4, 15, 34, 38, 43-47, 62, 118, 120-122, 127, 131, 221, 260-263, 266 21, 126

Activity-dependent extrinsic modulation (ADEM), 242-244

311

312 Subject Index Food finding, 37-38, 100, 113, 131-140 Foods, 72, 102, 120-123, 126-131, 147, 163-167, 246 Foot, 3, 7, 13, 52-54, 59, 63, 87, 93-101, 122, 139, 145, 183, 187, 190, 194, 196, 217-219, 221-222, 250, 275

Habituation, 61, 64-65, 218, 234-236 Hair cells, 48-51, 116, 118, 120 Hermaphroditism, 170-175, 177, 189, 271 Hierarchy, behavioral, 272-273 Hydrostatic pressure, 53-55, 71, 75, 87, 127, 129, 186, 190

Ganglia, 4, 14, 17-29 abdominal, 20, 27, 38, 73, 76, 78, 88, 91, 206, 210-211, 234, 245, 247, 261 (see also Ganglia, parietovisceral) branchial, 17, 62-63, 85 buccal, 17, 75-76, 140-150, 152-155, 158, 160, 266, 275, 279 cerebral, 17, 21-22, 24, 27, 44, 48, 73-74, 89, 93, 97, 103, 105, 108, 110, 114, 136, 140, 145, 150, 153-155, 157-158, 189, 191-192, 194-195, 199, 201, 203, 205-206, 209, 211, 219, 222, 237, 261, 266, 275 anterior lobe, 187 mesocerebrum, 192-194, 277-279 metacerebrum, 155, 222 procerebrum, 22, 25, 29, 136-139 ventral lobe, 187 cerebropleural, 47, 105 epiathroid, 22, 25 hypoathroid, 22, 25 intestinal, 6, 13, 18, 20, 24, 80 optic, 46-47, 122 osphradial, 17, 37, 69 pallial, 20, 22 parietal, 17, 20, 22, 66, 68, 71, 77, 82, 90-91, 187, 201, 219, 278-279 parietovisceral, 20, 60-61, 198 (see also Ganglia, abdominal) pedal, 17, 21-22, 24, 48, 68, 77, 80-81, 93-94, 97, 105, 107-108, 111-113, 116, 121-122, 136, 145, 157, 187, 191-192, 195, 201, 211, 219, 222, 224, 233, 240, 261, 275, 279 peripheral, 59, 62-63, 65-66, 114, 136 pleural, 18, 20-22, 24-25, 48, 77, 90, 93, 97, 99, 109-111, 136, 187, 201, 203, 205-206, 209, 211, 219, 221, 223-224, 233, 245, 261, 279 subesophageal, 17, 20, 75, 82 supraesophageal, 17, 19-20 tentacle, 32, 62 visceral, 13, 20-21, 24, 66, 68, 71, 75, 77, 82, 90, 201 Gigantism, neuronal, 29-30 Gill(s), 6-7, 10-14, 53, 62, 66, 71, 82-84, 247. See also Reflexes, withdrawal, gill Glia, 26-29

Ink, vii, x, 83, 226, 232-234, 237, 244, 246-248 Ion currents (ion channels) IA, 104, 207-208, 248, 263 IKV, 263, 266 Ca++, 52, 56, 75, 112, 127, 157, 206-208, 242, 265 K+, 29, 43, 58, 111, 121-122, 127, 161, 206, 238-239, 253, 263-264 Na+, 49, 56, 127, 206 nonspecific cation, 207-209 Kandel ER, ix-x, 60-61, 78, 151, 224, 234-235 Learning. See Conditioning, Habituation, Sensitization Lung, 7, 66-71, 82, 86 Magnetoreception, 51-52, 122-123 Mantle, 194, 196, 221, 226, 230-231, 233, 247, 267, 275. See also Cavity, mantle Mating. See Copulation, Courtship Mechanoreception, 34, 38-42, 131, 135, 222 Mechanosensory neurons, 40-42, 59, 63, 134, 154, 158-159, 228-230, 233-235, 240 Mesocerebrum. See Ganglia, cerebral Metacerebrum. See Ganglia, cerebral Modulation. See Neuromodulation Motivation, 152 appetitive, 134-135, 151, 156-158, 186, 197, 210, 267-268, 278 arousal, 151, 157-158, 184, 190, 267-270, 278 consummatory, 134-135, 140, 151-152, 156, 158, 163, 197-198, 210 hunger, 151 sexual, 186-187, 269 Motor programs, 108, 110, 116, 140-150, 152-154, 157-160, 164-167, 279. See also Central pattern generator Mucus, 86, 95-96, 101, 126, 128, 178-180, 182, 233-234, 245-246, 250. See also Trail following Muscles, 53-58 accessory radula closer (ARC), 55, 57-58, 145, 156, 160-161

Subject Index accessory radula opener, 57, 160-161 body wall, 53-54, 113, 124, 186, 189, 221, 245, 247 buccal protractor (12), 145 retractors, buccal mass, 53-54, 58, 218 columellar, 53, 55, 136, 214, 218-219, 221 dorsal longitudinal, 219 pedal, 218 penis, 53, 186, 195, 218 tentacle, 53, 218, 222 vasoconstrictor, 55, 73, 76, 78, 80, 82 Nerves acoustic (static). See Nerves, statocyst buccal, 149, 155, 167 commissural, 17, 200 connective cerebrobuccal, 153-155, 165 cerebropedal, 98, 108, 113, 192 cerebropleural, 22, 98, 113, 205 pleurovisceral (pleuroabdominal), 20, 82, 198, 200, 205, 206, 209, 237, 239-240 ctenidial, 65 esophageal, 167, 169 lip, 135-136, 140, 155, 166, 194, 262 nervus cutaneus pedalis primus dexter (NCPPD or NCPD), 191-192 olfactory. See Nerves, tentacle optic, 44, 46, 221, 260-261, 263 osphradial, 37 pedal, 93-94, 97, 107-108, 113, 117-118, 219 penis, 187, 189, 191-192 peripheral, 21, 27, 31, 59, 62, 140, 193, 206, 222 salivary, 149 siphon, 63 statocyst, 48, 50-51 tentacle, 38, 132-135 vas deferens, 187 visceral, 88 Neuromodulation, 29, 51-52, 58, 75-76, 82, 84-85, 91, 93, 94, 99, 108-114, 122-123, 140, 144, 152-157, 160-163, 166, 210, 223, 236-242, 266, 273-274, 277, 279 Neurotransmitters. See also Peptides acetylcholine, 29, 31, 56, 58, 76, 111, 149, 160-161, 223, 247, 277 dopamine, 31, 56, 58, 68, 76, 94, 149-150, 252-253, 274-275 glutamate, 31, 56, 132, 147, 160-161 glycine, 31, 76 histamine, 31, 158 nitric oxide, 31, 158

313

serotonin (5-HT), 27, 31, 58, 76, 80, 88, 94-95, 110-113, 149-150, 155-156, 160-161, 186, 237-238, 240, 242, 252, 266-268, 274-275 Nociception, 38, 42, 222, 230, 234, 236-238 Opaline, 83, 226, 233, 246, 248 Osphradium, 12, 35-37, 40, 62, 90 Parapodia, 13, 53, 74, 78, 83, 96-97, 99-100, 105, 107-108, 111-112, 114, 116, 139, 194, 222, 225-226, 247 Peptides APGWamide, 155, 186-189, 192-195, 274 atrial gland peptides A/B, 203, 206 attractin, 182 bag cell, AP, 203 a-BCP, 203, 205-208, 210-211 j6-BCP, 203, 207 X-BCP, 203, 207 egg-laying hormone (ELH), 149, 198-203, 205, 209-211, 273 buccalin, 160-162 cardioactive (LCPs, MCPs, SCPs), 33, 75-76, 84, 149, 160-162 CDCH-I/ II, 202-203, 211-212 a/£3-CDCP, 211-212 cerebral peptide 2 (CP2), 84 conopressin, 186-188 DEILSR, 186 FLRFamide, 75 FMRFamide, 56, 74-75, 78, 88, 149, 160-161, 186, 192-193, 223, 238, 279 GFAD, 193 Lymnaea inhibitory peptide, 186 myomodulin, 76, 160-162, 186 octopamine, 144, 149 opioid, 52, 238 pedal (Pep), 80, 94, 113 pQDPFLRFamide, 75 R15cd, 85,90, 211 R15£2, 76 SEEPLY, 75, 186 Peripheral nervous system (PNS), 58-65, 97, 218, 226-227, 236, 245-246 Photoreceptors, 42-46, 49, 118, 120-122, 262 Pneumostome, 14, 66-67, 69-70, 81, 86, 221, 244 Postcerebrum. See Ganglia, cerebral Predation, 7, 14, 36, 41-42, 99-101, 116, 127-129, 131, 134-136, 196, 214-217, 222, 242, 244-246, 248, 250, 256-257 Procerebrum. See Ganglia, cerebral

314 Subject Index Radula, 4, 12, 14, 124-130, 147, 155, 158, 162, 171, 277 Reflexes local versus remote, 58-64, 218, 226, 233 withdrawal defensive. See Reflexes, withdrawal, gill and siphon gill, 60-62, 65, 223-240, 243, 247, 273 shadow, 218, 220 siphon, 59-62, 65, 85, 223-233, 235, 237-240, 242-243, 273 tail, 233-234 tentacle, 60, 62-63, 65, 74, 217-218, 221-223, 234, 244 whole-body, 218-222, 233, 279 Respiration, 13-14, 65-70, 80-82, 91, 114, 226, 253-254 Respiratory pumping, 82-85, 226, 233, 247 Rhinophore, 12-14, 36, 42, 61, 114-115, 124, 262 Sensitization, 65, 223, 235-244 Sheath, nerve, 24-26, 47, 62, 82, 198 Siphon, 35, 39, 43, 63, 82-83, 234, 247. See also Reflexes, withdrawal, siphon Sperm competition, 175-177, 180-181, 203 Startle response, 102, 277, 279

Statocyst, 47-50, 67, 115-118, 120 Streptoneury, 6-7, 17, 19-22 Swimming, 52, 71, 83, 93, 99-113, 116-118, 127-128, 221, 270, 273-275, 277, 279-280 Taxes anemotaxis, 114-115, 134 chemotaxis, 114-115, 133-134 geotaxis, 115-120 klinotaxis, 114-115, 134 magnetic, 50, 122-123, 252 phototaxis, 43, 119-122 rheotaxis, 114, 134 tropotaxis, 115, 134 Tentacles, 4, 12-13, 15, 37-40, 43, 50, 53, 55, 114-115, 129, 132-134, 136, 139, 171, 182-184, 190, 193-194, 196-197, 246, 262 Torsion, 4-8, 9, 19. See also Detorsion Trail following, viii, 38, 86, 129, 134-135, 183 Trophospongium, 27 Visceral nerve loop, 6-7, 12-13, 17-21, 24 Water regulation, 85-92 Withdrawal. See Reflexes