Echinoderm nutrition [1st edition] 9781000123609, 100012360X, 9781000139372, 1000139379, 9781000162301, 1000162303, 9781003078920, 1003078923

Table of contents :
Cover......Page 1
Half Title......Page 2
Title Page......Page 4
Copyright Page......Page 5
Table of Contents......Page 6
Foreword......Page 14
Preface......Page 16
Part One: Food and Feeding Mechanisms......Page 18
1. Chemoreception in echinoderms related to their structure and life style......Page 20
2. Physical perception of food and feeding......Page 22
3. Chemical perception of food and feeding......Page 25
4. Organs of chemoreception......Page 37
5. Variability of response to chemical stimuli......Page 39
6. Conclusions......Page 40
1. Trophic classification......Page 42
2.1 Tube-feet......Page 44
2.2 Feeding postures of pinnulate crinoids......Page 48
2.3 Functional morphology of feeding appendages in fossil crinoids......Page 52
2.4 Feeding in non-crinoid crinozoans......Page 56
3. Food composition......Page 57
4. Conclusions......Page 58
1. Deposit-feeding holothuroids......Page 60
2. Suspension-feeding holothuroids......Page 66
3. Discussion and conclusion......Page 70
1.1.1 Food of regular echinoids......Page 74
1.2.1 The Aristotle's lantern......Page 91
1.2.2 The tube-feet......Page 105
1.2.3 Capture of drift material......Page 106
1.2.5 Conclusions......Page 108
Feeding mechanisms......Page 109
Food......Page 121
2.2.1 Spatangoida......Page 122
Food......Page 124
Feeding mechanisms......Page 126
Feeding mechanisms......Page 131
2.3 Conclusion......Page 132
1.1 Food resources of asteroids......Page 134
1.2 Food of asteroids......Page 135
1.3 Trophic groups......Page 155
1.4 Breadth of diet and food preferences......Page 156
2.1 Intraoral feeders......Page 157
2.2 Extraoral feeders......Page 158
2.2.2 Capture of slow-moving, sedentary or attached unprotected prey......Page 160
2.2.3 Capture of slow-moving, sedentary or attached protected prey......Page 162
2.3 Ciliary-feeding and suspension-feeding......Page 165
3. Digestion time, feeding rate and feeding rhythm......Page 169
4. General conclusions......Page 175
Chapter 6 Food and feeding mechanisms: Ophiuroidea......Page 178
1. Carnivorous feeding......Page 182
2. Microphagous feeding......Page 191
4. Discussion......Page 200
5. Conclusions......Page 201
Part Two: Digestive Systems......Page 204
Chapter 7 Digestive systems: General considerations......Page 206
1.1 Recent crinoids......Page 208
3. Conclusion......Page 211
1. Anatomy of the gut......Page 212
2. Histology and cytology......Page 220
3. Digestive organs......Page 234
4. Mechanics of the gut......Page 237
5. Conclusions......Page 238
1.1 Regular echinoids......Page 240
1.2 Irregular echinoids......Page 248
2.1 Regular echinoids......Page 249
2.3 The hemal digestive system......Page 254
3.1 Buccal cavity and pharynx......Page 256
3.2 Esophagus......Page 257
3.3 Stomach......Page 258
3.5 Intestine......Page 259
4.1 Regular echinoids......Page 260
4.2 Irregular echinoids......Page 262
5. Conclusions......Page 263
1. General and comparative anatomy of the digestive system......Page 264
1.1 Survey of asteroid families......Page 265
1.2 The stomach retractor system......Page 291
1.3 General conclusions on anatomy......Page 293
2.1 The digestive epithelium......Page 295
2.2 The basi-epithelial (epineural) nerve plexus......Page 297
2.4 The tissues external to the connective tissue layer......Page 298
3.1 Stomach......Page 299
3.2 Brachial digestive organs......Page 301
3.3 Aboral digestive organs......Page 303
4.1 Asteriidae......Page 304
5. Conclusions......Page 306
1.1 Anatomy, histology and cytology of the digestive system......Page 308
3. Phrynophiurida......Page 313
4. Conclusions......Page 314
Part Three: Physiology and Biochemistry......Page 316
1.1 Intracellular digestion......Page 318
1.2 Extracellular digestion......Page 319
1.4 Digestive enzymes......Page 323
2. Larval echinoderms......Page 350
3. Conclusions......Page 351
1.1 The involvement of coelomocytes......Page 352
1.2 Absorption via epithelial cells......Page 354
1.3 Sugar absorption from the alimentary canal......Page 355
1.4 Alimentary absorption of amino acids......Page 357
2.1 General evidence for a parenteral route of nutrient absorption......Page 360
2.2 Mechanisms of absorption of monosaccharides......Page 361
2.3 Mechanisms of absorption of amino acids......Page 362
3. General considerations......Page 363
1. Respiration......Page 366
2. Somatic growth......Page 383
3. Gonadal growth......Page 396
4. Bioenergetics......Page 403
5. Conclusions......Page 406
Chapter 16 Nutrient translocation......Page 408
1.1 The hemal system......Page 409
1.2 Coelomocytic transport......Page 418
1.3 Perivisceral coelomic transport......Page 422
1.4 The perihemal coelomic system......Page 426
1.6 Integumental translocation......Page 427
2.1 The model......Page 428
1.1 Carbohydrate degradation......Page 432
1.2 Mitochondrial oxidations......Page 438
1.3 Nitrogen catabolism......Page 442
2.1 Glycogen synthesis......Page 444
2.2 Amino acid biosynthesis......Page 445
2.4 Lipid biosynthesis......Page 447
3.1 Control of phosphorylase activity in echinoid eggs......Page 449
3.3 Feedback inhibition of pyrimidine biosynthesis in echinoid embryos......Page 450
4. Conclusions......Page 451
Chapter 18 Steroid metabolism......Page 454
1. Sterols in echinoderms......Page 455
2. The origin of sterols in echinoderms......Page 458
3. The effect of dietary sterols on the sterol composition of echinoderms......Page 464
4. Elimination of sterols......Page 466
5. Conclusions......Page 473
1. Nitrogenous wastes of echinoderms......Page 474
3. Excretion and the coelomocytes......Page 480
4. The supposed renal organs of echinoderms......Page 482
5. Excretion in embryos and larvae......Page 483
6. Conclusions......Page 484
Part Four: Nutrition During Development......Page 486
Chapter 20 Nutrition of gametes......Page 488
1.1 Somatic structure of gonads in the Echinodermata......Page 490
1.2.1 Outer sac (found only in Asteroidea, Echinoidea and Ophiuroidea)......Page 491
1.2.2 Genital coelomic (perihemal) sinus......Page 492
2.1 Predictable generation of functional subdivisions of the germinal epithelium during gametogenesis......Page 495
2.2 Nutrient input to the gametogenic microenvironment and utilization by germinal cells......Page 507
2.2.1 Asteroidea......Page 508
2.2.2 Echinoidea......Page 511
3. Conclusions......Page 513
1.1 General considerations......Page 516
1.2 The yolk......Page 517
1.3 Formation of the yolk platelets......Page 518
1.4 Utilization of stored macromolecules......Page 520
2. Other echinoderms......Page 522
3. Conclusions......Page 523
1. Trophic categories in echinoderm larvae......Page 526
2. Morphology of the gut in planktotrophic larvae......Page 528
3.1 Planktotrophic larvae......Page 532
3.3 Incubated embryos......Page 534
5. Larval energy budget......Page 535
Part Five: Effects of Feeding on The Environment......Page 540
1. Deposit feeders......Page 542
3. Conclusions......Page 545
1. The physical effect of feeding on the substratum......Page 548
2.1 The effects on plants......Page 554
2.2 The effects on epibenthic animals......Page 560
2.3 Effects of limitations on echinoid distribution, abundance and mobility......Page 564
2.4 Stability in echinoid populations......Page 566
2.5 Conclusion......Page 568
1. Definitions and biases......Page 570
2.1 Temperate shores: Pacific coast of North America......Page 572
2.2 Temperate shores: Atlantic coast of North America......Page 576
2.3 Temperate shores: Atlantic coast of Europe......Page 577
2.4 Temperate shores: Pacific coast of Asia (Japan)......Page 578
2.5 Temperate south Pacific shores......Page 579
3. Role of asteroids in rocky subtidal regions......Page 580
3.1 Polar subtidal habitats: Antarctica......Page 581
3.2 Temperate subtidal habitats......Page 582
3.3 Temperate southern hemisphere......Page 586
3.4 Tropical subtidal habitats......Page 588
4.1 Asteroids and the organization of benthic marine communities......Page 592
4.2 Why are asteroids so important?......Page 597
4.3 Comparison to other communities......Page 599
4.4 Conclusions and future directions......Page 600
Bibliography......Page 602
List of contributors......Page 686
Species index......Page 688
Subject index......Page 698

Citation preview

ECHINODERM NUTRITION

luvenileAsterias vulgaris about 2 days old devouring a clam (from Mead 1901).

ECHINODERM NUTRITION Edited by

MICHEL JANGOUX

Universite libre de Bruxelles, Belgium

JOHN M.LAWRENCE University 0/ South Florida, Tampa, USA

A.A.BALKEMA/ROTTERDAM/1982

ISBN 90 6191 0803 @1982, A.A.Balkema, P.O.Box 1675,3000 BR Rotterdam, Netherlands Distributed in USA &. Canada by MBS, 99 Main Street, Salem, NH 03079, USA Printed in the Netherlands

TABLE OF CONTENTS

Foreword

xlli

Preface

xv

PART ONE: FOOD AND FEEDING MECHANISMS

Chapter 1 Perception ollood N.A.Sloan & Andrew C.Campbell

1. Chemoreception in echinoderrns related to their structure and life style 2. Physical perception of food and feeding 3. Chemical perception of food and feeding 4. Organs of chemoreception S. Variability of response to chemical stimuli 6. Conclusions

Chapter 2 Food and leeding mechanisms: Crinozoa David L.Meyer 1. Trophic classification 1.1 Stratification of feeding positions 1.2 Feeding periodicity 2. Functional morphology of feeding appendages 2.1 Tube-feet 2.2 Feeding postures of pinnulate crinoids 2.3 Functional morphology of feeding appendages in fossil crinoids 2.4 Feeding in non-crinoid crinozoans 3. Food composition 4. Conclusions

Chapter 3 Food and leeding mechanisms: Holothuroidea Claude Massin

1. Deposit-feeding holothuroids 2. Suspension-feeding holothuroids 3. Discussion and conclusion

3

3

5

8

20

22

23

2S 2S

27

27

27

27

31

35

39

40 41

43

43

49

S3

v

VI

Table o{ contents

Chapter 4 Food and feeding mechanisms: Echinoidea Chantal De Ridder & lohn M. Lawrence 1. Regularia 1.1 Food 1.1.1 Food of regular echinoids 1.1.2 Feeding rates 1.2 Feeding mechanisms 1.2.1 The Aristotle'slantern 1.2.2 The tube-feet 1.2.3 Capture of drift material 1.2.4 Ingestion of soft substrata 1.2.5 Conclusions 2. Irregularia 2.1 Irregularia Gnathostomata 2.1.1 Clypeasteroida Food, p.92; Feeding mechanisms, p.92; Food selection, p.104

2.1.2 Holectypoida Food, p.104; Feeding mechanism, p.105

2.2 Irregularia Atelostomata 2.2.1 Spatangoida Food, p.107; Feeding mechanisms, p.l09; Food selection, p.114

2.2.2 Cassiduloida Food, p.114; Feeding mechanisms, p.114

2.2.3 Holasteroida Food, p.115; Feeding mechanisms, p.115

2.3 Conclusion Chapter 5 Food and feeding mechanisms: Asteroidea

Michellangoux

1. Food and trophic groups 1.1 Food resources of asteroids 1.2 Food of asteroids 1.3 Trophic groups 1.4 Breadth of diet and food preferences 2. Feeding mechanisms 2.1 Intraoral feeders 2.2 Extraoral feeders 2.2.1 Capture of motile prey 2.2.2 Capture of slow-moving, sedentary or attached unprotected prey 2.2.3 Capture of slow-moving, sedentary or attached protected prey 2.2.4 Feeding on corpses and plants 2.3 Ciliary-feeding and suspension-feeding 3. Digestion time, feeding rate and feeding rhythm 4. General conclusions

57 57

57

57

74

74

74

88

89

91

91

92

92

92

104

105

105

114

115

115

117

117

. 117

118

138

139

140

140

141

143

143

145

148

148

152

158

Table o{ contents

Chapter 6 Food and feeding mechanisms: Ophiuroidea George Warn er 1. Carnivorous feeding 2. Microphagous feeding 3. Other feeding methods 4. Discussion 5. Conclusions

VII

161 165 172 179 179 180

PART TWO: DIGESTIVE SYSTEMS

Chapter 7 Digestive systems: General considerations Michel Jangoux

185

Chapter 8 Digestive systems: Crinozoa Michel Jangoux

187

1. Crinoid Crinozoa 1.1 Recent crinoids 1.2 Fossil crinoids 2. Non-crinoid Crinozoa 3. Conclusion

187 187 190 190 190

Chapter 9 Digestive systems: Holothuroidea Jean-Pierre Feral & C1aude Massin

191

1. Anatomy of the gut 2. Histology and cytology 3. Digestive organs 4. Mechanies of the gut 5. Conclusions

Chapter 10 Digestive systems: Echinoidea Chan tal De Ridder & Michel Jangoux

1. General and comparative anatomy of the digestive system 1.1 Regular·echinoids 1.2 Irregular echinoids 2. Histology and cytology of the digestive system 2.1 Regular echinoids 2.2 Irregular echinoids 2.3 The hemal digestive system 3. Digestive organs 3.1 Buccal cavity and pharynx 3.2 Esophagus 3.3 Stomach 3.4 Siphon 3.5 Intestine 3.6 Rectum

191 199

207 210 211

213 213 213 221 222 222 225 225 227 227 228 229 230 230 231

VIII

Table 01 contents

4. Digestive mechanics 4.1 Regular echinoids 4.2 Irregular echinoids 5. Conc1usions

231

231

233

234

Chapter 11 Digestive systems: Asteroidea Michellangoux

235

1. General and comparative anatomy of the digestive system 1.1 Survey of asteroid families 1.2 The stomach retractor system 1.3 General conc1usions on anatomy 2. Histology and cytology of the digestive system 2.1 The digestive epithelium 2.2 The basi-epithelial (epineural) nerve plexus 2.3 The connective tissue layer 2.4 The tissues external to the connective tissue layer 2.5 Note on the hemal system 3. Digestive organs 3.1 Stomach 3.2 Brachial digestive organs 3.3 Aboral digestive organs 4. Digestive mechanics 4.1 Asterüdae 4.2 Other families 5. Conc1usions

235

236

256

258

260

260

262

263

263

264

264

264

266

268

269

269

271

271

Chapter 12 Digestive systems: Ophiuroidea Michellangoux

273

1. Ophiurida 1.1 Anatomy, histology and cytology of the digestive system 1.2 Functions and mechanics of the gut 2. Oegophiurida 3. Phrynophiurida 4. Conc1usions

273

273

278

278

278

279

PART THREE: PHYSIOLOGY AND BIOCHEMISTRY

Chapter 13 Digestion lohn M. Lawrence

1. Post-metamorphic echinoderms 1.1 Intracellular digestion 1.2 Extracellular digestion 1.3 pU levels in extracellular digestion 1.4 Digestive enzymes 2. Larval echinoderms 3. Conc1usions

283

283

283

284

288

288

315

316

Table 0/ contents IX

Chapter 14 Epithelial absorption David Bam/ord

317

1. Absorption of organic nutrients from the alimentary lumen 1.1 The involvement of coelomocytes 1.2 Absorption via epithelial cells 1.3 Sugar absorption from the alimentary canal 1.4 Alimentary absorption of amino acids 2. Epidermal absorption 2.1 General evidence for a parenteral route of nutrient absorption 2.2 Mechanisms of absorption of monosaccharides 2.3 Mechanisms of absorption of amino acids 3. General considerations

317

317

319

320

322

325

325

326

327

328

Chapter 15 The utilization 0/ nu trients by postmetamo.rphic echinoderms

lohn M.Lawrence & I.M. Lllne .

331

1. Respiration 2. Somatic growth 3. Gonadal growth 4. Bioenergetics 5. Conclusions

331

348

361

368

371

Chapter 16 Nutrient translocation lohn C.Ferguson

373

1. Translocation systems 1.1 The hemal system 1.2 Coelomocytic transport 1.3 Perivisceral coelomic transport 1.4 The perihemal coelomic system 1.5 The water vascular system 1.6 Integurnental translocation 2. Conclusions: a model for nutrient translocation 2.1 The model

374

374

383

385

389

390

390

391

391

Chapter 17 Intermediary metabolism W.Ross Ellington

395

1. Catabolism 1.1 Carbohydrate degradation 1.2 Mitochondrial oxidations 1.3 Nitrogen catabolism 2. Biosynthesis 2.1 Glycogen synthesis 2.2 Amino acid biosynthesis 2.3 Pyrimidine and purine biosynthesis 2.4 lipid biosynthesis 3. Regulation of intermediary rnetabolism

395

395

401

405

407

407

408

410

410

412

x Table 01 contents 3.1 Control of phosphorylase activity in echinoid eggs 3.2 Control ofthe activity ofthe pyruvate dehydrogenase complex in echinoid eggs

3.3 Feedback inhibition of pyrimidine biosynthesis in echinoid embryos 4. Conclusions

412

413

Chapter 18 Steroid metabolism Peter A. Voogt

417

1. Sterols in echinoderms 2. The origin of sterols in echinoderms 3. The effect of dietary sterols on the sterol composition of echinoderms 4. Elimination of sterols 5. Conclusions

418

421

427

429

436

Chapter 19 Excretion Michel Jangoux

437

1. Nitrogenous wastes of echinoderms 2. Excretory sites of metabolie end products 3. Excretion and the coelomocytes 4. The supposed renal organs of echinoderms 5. Excretion in embryos and larvae 6. Conclusions

437

441

441

443

444

445

413

414

PART FOUR: NUTRITION DURING DEVELOPMENT

Chapter 20 Nutrition 01 gametes Charles W. Walker

1. Potential mechanisms of nutrient distribution and storage by somatic tissues of the gonad (excluding those in the germinal epithelium)

1.1 Somatic structure of gonads in the Echinodermata 1.2 Predictable changes in the sornatic structures of the gonad during gametogenesis that may be related to gamete nutrition 1.2.1 Outer sac (found only in Asteroidea, Echinoidea and Ophiuroidea) 1.2.2 Genital coelomic (perihemal) sinus 2. Potential mechanisms of nutrient distribution, storage, and utilization by the germinal epithelium

2.1 Predictable generation of functional subdivisions of the germinal epithelium during gametogenesis

2.2 Nutrient input to the gametogenic microenvironment and utilization by germinal ceUs

2.2.1 Asteroidea 2.2.2 Echinoidea 3. Conclusions

449 451

451

452 452

453

456

456

460

461

464

466

Table of contents

XI

Chapter 21 Nutrition o[ embryos Go[[redo Cognetti

469

1. Echinoids 1.1 General considerations 1.2 The yolk 1.3 Formation of the yolk platelets 1.4 Utilization of stored macromolecules 2. Other echinoderms 3. Conclusions

469

469

470

Chapter 22 Nutrition ollarvae Lucienne Fenaux

1. Trophic categories in echinoderm larvae 2. Morphology of the gut in planktotrophic larvae 3. Feeding oflarvae 3.1 Planktotrophic larvae 3.2 Lecithotrophic larvae: pelagic and demersal 3.3 Incubated embryos 4. Effects of dissolved organic matter on larval development 5. Larval energy budget

471

473

475

476

479

479

481

485

485

487

487

488

488

PART FIVE: EFFECTS OF FEEDING ON TUE ENVIRONMENT

Chapter 23 E[[ects o[[eeding on the environment: Holothuroidea Claude Massin

493

1. Deposit feeders 2. Suspension feeders 3. Conclusions

493

496

496

Chapter 24 Effects o[[eeding on the environment: Echinoidea lohn M.Lawrence & Paul W.Sammarco

499

1. The physical effect of feeding on the substratum 2. The effect of echinoid feeding on biological communities 2.1 The effects on plants 2.2 The effects on epibenthic animals 2.3 Effects of limitations on echinoid distribution, abundance and mobility 2.4 Stability in echinoid populations 2.5 Conclusion

499

505

505

511

515

517

519

Chapter 25 E[[ects o[ [eeding on the environment: Asteroidea Bruce A.Menge

521

1. Definitions and biases 2. Role of asteroids in the rocky intertidal region 2.1 Temperate shores: Paeific coast of North America 2.2 Temperate shores: Atlantic coast of North America 2.3 Temperate shores: Atlantic coast of Europe

521

523

523

527

528

XII

Table 01 contents

2.4 Temperate shores: Paeifie eoast of Asia (Japan) 2.5 Temperate south Paeifie shores 2.6 Tropical shores: Paeifie eoast of Panama 3. Role of asteroids in rocky subtidal regions 3.1 Polar subtidal habitats: Antaretiea 3.2 Temperate subtidal habitats 3.3 Temperate southern hemisphere 3.4 Tropieal subtidal habitats 4. Synthesis 4.1 Asteroids and the organization of benthie marine eommunities 4.2 Why are asteroids so important? 4.3 Comparison to other eommunities 4.4 Conelusions and future direetions

529 530 531 531 532 533 537 539 543 543 548 550 551

Bibliography

553

List of contributors

637

Species index

639

Subject index

649

FOREWORD

Living systems are open ones thermodynamically, and must obtain energy and matter to maintain themselves through time. This is a basic principle ofbiology. Understanding the ways in which living systems obtain and utilize nutrients is thus of fundamental importance. We have taken a broad view of nutrition deliberately, not confming ourselves to the requirements for, and utilization of nutrients. We begin with the perception of food and all feeding, digestive, and absorptive processes which lead to the uptake of nutrients, and con­ tinue through the ecological effects of feeding. A necessary consequence of feeding is the removal of matter and energy from the system which often involves the predation of other animals and plants. The removal of prey individuals from the system and the change in the physical and chemical state of biological material as a result of nutritional action both have environmental consequences. The purpose of this book is to present the state of knowledge concerning nutrition and point out directions for future work for the Echinodermata, an ancient group which shows great diversity in form and function, and whose feeding activities can have great environ­ mental impact. The echinoderm classification proposed in Moore's Treatise on Invertebrate Paleontology has been foUowed throughout this book. We sincerely thank all the contributors for their efforts. Thanks are also due to those who helped during the editorial work, especially N.Biot, J.Harray, and M.Klinkert (Labora­ toire de Zoologie, Universite Libre de Bruxelles). We are grateful to the 'Fondation Roi Uopold 111' and to the 'Ministere de I'Education nationale de Belgique' for financial sup­ port. The editors

XIII

PREFACE

It is tirnely that a book on echinoderm nutrition should appear. In the present volume echinoderm metabolism and many other phases of echinoderm nutrition are summarized and evaluated, many for the fust time. A much more meaningful concept of the overall nutrition of echinoderms will emerge from a knowledge of perception of food, environ­ mental effects of feeding, structure of the digestive organs, digestion, absorption, trans­ location and assimilation of nutrients, excretion and metabolism. The present volume should be awaited with anticipation not only by those interested in echinoderms, but also by those wishing to compare the nutrition of echinoderms with the nutrition of anirnals in other phyla. The primitive nature of the echinoderms makes them of special interest in this regard as a background for comparative studies on the nutrition of anirnals. In addi­ tion, a summary of this type will point out the deficiencies in our knowledge and thereby bring to the attention of younger scientists some of the problems they rnight fruitfully attack. There are no fmal answers in a treatise and part of its value is the stimulus it pro­ vides for another generation of researchers. It is my hope that this treatise will entice others to experience the joy of collecting from the ocean, working with the organisms at the seaside, and solving some aspect of the nutrition of echinoderms.

Arthur C. Giese Professor Emeritus of 8i%gy, Stanford University

xv

1. FOOD AND FEEDING MECHANISMS

N. A. SLOAN & ANDREW C. CAMPBELL

1

PERCEPTION OF FOOD

The turbidity of inshore waters and the irregularity of much of the sea-bed limit the effect­ iveness of sight as a major sensory modality for many benthic marine invertebrates (Mackie 1975). Vision is, however, used extensively by a few specialised groups such as cephalopods (Wells 1978) and crustaceans (Waterman 1961, Hazlett 1972). Nicol (1967) did not con­ sider that the ability to hear was widespread among marine invertebrates although some exceptions occur amongst the crustaceans (Salmon & Horch 1972, Meyer-Rochow & Pen­ rose 1976). By contrast it seems that the transfer of information chemically either by con­ tact or by water is particularly suitable for aquatic animals. Carthy (1958) pointed out that the usefulness of chemoreception is greatly enhanced when given directionality by current induced gradients. A further advantage is that only minute quantities of a chemical need be used to release a behavioural sequence (Lenhoff & Lindstedt 1974). Mackie & Grant (1974) among others, believed that chemoreception amongst marine invertebrates was a subject the importance of which 'cannot be overemphasised'. Such awareness had not escaped earlier workers, for example Pearl (1903), who wrote 'one of the most important factors in the sum total of activities of any aquatic organism is its reactions to chemical substances' . Modern techniques, particularly in the realm of electrophysiology, have led to a more accurate analysis of the processes of transduction and integration of sensory activity in nervous systems, but as Ramsay (1968) pointed out, chemoreception appears to lag behind on research effort by comparison with other sensory modalities, e.g. vision. It is encouraging that more recently Lenhoff & Lindstedt (1974) indicated some improvement of the situation concerning chemoreception research methods and objectives so that the paradox of inten­ sity of investigation on the one hand and biological significance on the other many be resolving (Kittredge et al. 1974). Concerning topics of inter-specific chemoreception in marine invertebrates generally, food detection has the best coverage (Passano 1957, Kohn 1961, Laverack 1963, Lindstedt 1971, Lenhoff & Lindstedt 1974). In a review offeeding, digestion and nutrition of all echinoderms, however, Ferguson (1969a) devoted only a very rninor portion to 'Chemosensitivity to food', a measure of how poorly understood the field of echinoderm chemoreception has been.

1. CHEMORECEPTION IN ECHINODERMS RELATED TO THEIR STRUCTURE AND LIFE STYLE Echinoderms have unusual anatomical arrangements linked with peculiar physiological and behavioural ones which have considerable effects upon food perception and feeding. The form of symmetry, especially in relation to the lack of a head, the distribution of sen­

3

4 N.A.Sloan & Andrew C. Campbell sory and nervous elements, and the distribution of locomotor units are of paramount importance. There is little doubt that the first echinoderms were sedentary, if not sessile, suspension feeders (Nichols 1969). The water vascular system with its protrusible hydraulic tube-feet probably arose as arespiratory system which, because of its arrangement with respect to the mouth, allowed it to fulfll a food collecting role too (Nichols 1969), as in the crinoids. In such animals the pentameric plan is ideal and the lack of a head no disadvantage. Most remaining extant echinoderms have undergone major evolutionary advances. These include the inversion of their bodies with respect to the substrate and the assumption, in some cases, of a superficial bilateral symmetry. Such developments, however, have not led to the universal abandonment of suspension fee ding. The inversion of the body allowed the water vascular system to take on an additional function, that of locomotion. Such slow moving or sedentary radially symmetrical animals can receive chemical stimuli emanating from all directions equally weIl. Moreover, a rounded or disc-like body widely covered with receptor units should provide an ideal mechanism for gross sensory perception and analysis by simultaneous monitoring of stimulus intensity at different positions on its surface. Adult echinoderms are generally sluggish and creep slowly over or through the substrate. Some of these remain for a great part of the time in one place, and only a few species, aIl holothuroids, have become pelagic. Judging by the success of some modern free-living echinoderms, whose populations may dominate the environment (Nichols 1975), the penta­ meric plan has not been a serious limitation on free life. With the water vascular system providing locomotion, free life in a number of styles (e.g. creeping, burrowing, crevice dwelling) became possible and with this possibility the chances of fllling a variety of niches as grazing herbivores, omnivores, predators and carrion feeders. All these types of feeding method rely to some extent on chemical or physical perception and the means to orientate to and then take in the food. Errant echinoderms must move to get their food. It is important that they do so effi­ ciently. This is the essence of evolutionary pressure which has produced the various habits and life styles displayed by the errant echinoderms. The most striking feature of the echinoderm nervous system is the lack of a brain. Sense organs are rare in this group and transduction of environmental stimuli is generaIly a func­ tion of single unspecified cells widely distributed over most of the body surface. For example, Cobb (1968a) suggests that most epithelial cells of echinoids have neuronal con­ nections and act as transducers, but this suggestion awaits experinlental proof. So far, no specific region of the nerYous system with responsibility for integration of primary sensory modalities has been designated. The echinoderm nervous system has been reviewed by J .E.Smith (1950, 1965) and by Pentreath & Cobb (1972). There are three essential parts, although these may not be pre­ sent in all echinoderms. They are: 1. The ectoneural system, mainly sensory, and occurring in all classes except crinoids. 2. The hyponeural (Lange's nerve in asteroids) lacking in echinoids and holothuroids where its motor function is taken over by part of the ectoneural system. 3. The apical system, important in crinoids in which it is motor and innervates the vis­ cera, but poorly developed in asteroids. This system lacks in the other classes. Prime responsibility for chemoreception is taken by the ectoneuronal system. Structurally it is divided into two parts. Associated with each ambulacrum is a well-developed radial nerve cord. It provides the major afferent pathways from the receptors of the test and am­

Perception of food 5 bulacrum, and in groups where the hyponeural system is lacking it provides motor pathways to the tube feet too. Each radial cord is connected with the others via a circum-esophageal nerve ring which provides a route for inter-radial integrative activities. In echinoids and asteroids, where there is a good covering of epithelial cells overlying the skeleton, a well­ developed basi-epithelial nerve plexus constitutes the second part of the ectoneural system (Bullock 1965, CampbeIl1973). The principal role ofthis plexus is to receive nervous impulses arising from the multiplicity of sensory cells and to pass them where applicable to the radial nerve cords, and to co-ordinate the activities of the test appendages. In the remaining three classes the basi-epithelial nerve plexus is less well-developed; there are fewer test appendages and, with the exception of holothuroids, the epithelium is often scanty and abraded. Analysis of sensory and motor phenomena in the plexus has been hindered by technical difficulties. The small diameter and diffuse nature of the component axons prevent intra­ cellular recording. Also the presence of free crystals of calcite, even away from the skeleton, makes probing with rme instruments difficult. Within the radial cords, however, some pro­ gress has been made using gross techniques such as wick electrodes (Sandeman 1965). Electrophysiological proof for the hypothesis of chemoreception in echinoderms, therefore, is in astate of conjecture. The hyponeural system is found in asteroids and ophiuroids where it co-ordinates motor activities of the tube-feet and ampullae and is important in locomotion. The locomotory tube-feet are hydraulic organs which possess muscles allowing stepping movements and retraction. In most asteroids, echinoids and holothuroids they are suckered which assists with adhesion, but in ophiuroids they are suckerless and not important for movement. A fuller account is given in Nichols (1966,1969). Tube-feet can be intimately involved in the feeding process. In some cases they are well-known as food gatherers. Their implication in food detection, mentioned later in the organs of chemoreception section, is not surprising because as extensions of the body wall they are cloaked with a basi-epithelial plexus. The development of the anatomical resources of the various classes (compared with their habitat, life style, and source of food in table 1) suggests the potential for food detec­ tion in various echinoderm groups. Those groups which lack a well-developed sensory component in the basi-epithelial nerve plexus may be restricted to a semi-sedentary life like the filter-feeding crinoids or the deposit-feeding holothuroids. Interestingly, although ophiuroids, with their relatively massive skeleton and subsequently reduced basi-epithelial plexus, are mostly suspension-feeders, some are polyphagous (Fontaine 1965, Chartock 1972, Dearborn 1977) and responsive to chemical stimuli. The increase in the sensory com­ ponent and motor sophistication has allowed for the development first of relatively unselec­ tive omnivorous feeding as in regular echinoids, and secondly a more selective predatory life as in some asteroids, where strong dietary preferences may be encountered. 2. PHYSICAL PERCEPTION OF FOOD AND FEEDING Large scale movements or aggregations of asteroids and echinoids have been related to food availability (Dana et al. 1972, Lawrence 1975a, Grassie et al. 1975, Glynn 1976, Sloan 1977, Mattison et al. 1977, Garnick 1978), whereas the presence of suspension feeders is more strongly related to current conditions (Warner 1971, Fedra et al. 1976, Meyer & Lane 1976, Meyer & Macurda 1977) which bring the food to the awaiting masses. There are a number of instances where purely physical characteristics of the environ­

Holothuroidea

Sedentary: many burrow

Sedentary burrowers

Holectypoida Oypeasteroida

Holasteroida Nucleolitoida Cassiduloida Spatangoida Dendrochirotida Dactylochirotida Aspidochirotida Elasipodida Molpadida Apodida

Symmetry

Diverse: suspension feeders; surface fIlm feeders; detrital feeders; carrion feeders; carni­ vores Most graze algae or encrusting anirnals: some catch suspended particles Ingest food partic1es occurring in or on substrate Ingest su bstra te and associated organic material Radial or bilateral

Bilateral

Bilateral

Radial

Radial

Suspension feeders Radial using tube-feet and mucus Voracious carniRadial vores; some omnivores and detrital feeders

Feeding mechanism

Almost all benthic Hard and soft Suspension or and sedentary: ground credeposit feeders a few species vice dwellers, using modified are pelagic surface tube-feet dwellers and pehigic

Sands and gravels

Sands and gravels

Benthic, usuallyon hardground

Echinoidea

Some sedentary; some errant

Oegophiurida Phrynophiurida Ophiurida

Ophiuroidea

Cidaroida Diadematoida Echinoida

Many errant; few sedentary

Platyasterida Paxillosida Valvatida Spinulosida Forcipulatida

Asteroidea

Benthic on or in soft ground; on hard ground

Benthic, usuallyon hard ground Benthic on or in soft ground; on hard ground

Sedentary or sessile

Articulata

Crinoidea

Some sedentary; some errant

Habitat

lifestyle

Order

Oass

Table 1. Relationship of anatomical and ecological characteristics of extant adult Echinodermata

Poorly developed

Weil developed but virtually no sense organs

WeIl developed but virtually no Sense organs

Weil developed sense organs in pedicellariae

Poorly developed with no sense organs as such WeIl developed in many cases; some simple sense organs, e.g.optic cushion Poorly developed in most cases

Basi-epithelial nerve plexus and sense organs

Very specialised in so me groups; employed in burrow-maintenance, respiration and various other activities Employed in locomotion, feeding and respiration, and various other activities

Employed in burrowing, respiration and probably sensory

Employed in locomotion, food handling, covering and various other activities

Not employed in locomotion, but used in fee ding and respiration

Not employed in locomotion, but used in feeding and respira tion WeIl developed for locomotion, burrowing, respiration and' sensory perccption

Water-vascular system

(\)

:::::

~1::7'

~

n

~

(\)

::t.. ::s E}

~

::s

2

~

::t..

:rgensen (1966) claimed that crinoids are not passive suspension feeders. However, passive suspension feeders are defined as those relying on exogenous currents (Wainwright et al. 1976) so that crinoids should properly be placed in this trophic category. Crinoids share this trophic category with other sessile inverte­ brates such as basket stars, alcyonarians, and corals. Observations of living comatulid and stalked crinoids in the natural environment (H.L. Clark 1915, Magnus 1963, Rutman & Fishelson 1969, Meyer 1973a, Fishelson 1974, Macurda & Meyer 1974, LaTouche 1978) support the classification of crinoids as passive rheophilic suspension feeders. Several possibilities for alternative trophic roles among living and ancient crinoids must also be considered. Before llving crinoids were observed in nature, it was gene rally supposed that crinoids extended the arms horizontally or with an upward inclination to capture food particles falling by gravitative settling. This behaviour, termed rheophobic by Breimer (1969), now appears to be restricted to deep-water crinoids expe­ riencing periods of slack water (peres 1959). Breimer (1969, 1978) has also interpreted several fossil forms having rigid or partly flexible stalks as rheophobic suspension feeders. Although some shallow-water comatulids prefer cryptic or semi-cryptic microhabitats where water movements are very gentle, they are most abundant where there is persistent water movement over the surrounding reef (Meyer 1973a). This suggests that these species also rely on very low velocity flow which circulates through the reef infrastructure. Thus they cannot be considered rheophobic in the strict sense. While crinoids are generally considered to derive their food supply from particles sus­ pended above the substratum, the utilization of benthic food materials by deposit feeding rnay be adopted occasionally. Observations from a submersible by Reyss & Soyer (1965) showed that the comatulld Leptometra phalangium rests directly on a soft substratum with the oral side upward and the arms held horizontally. They suggested that this may indicate

25

26 David L.Meyer

Figure 1. Schematic diagram of stratified community of Middle Paleozoic (Mississippian) crinoids and other epifaunal invertebrates, based on reconstructions of height of living position above sea floor (after Ausich 1980).

deposit feeding habits, although the same species also holds the arms in a vertical fIltration fan during periods of current flow. They did not directly observe transfer ofbenthic par­ ticles to the mouth, but the use of the tube-feet for this mode of fee ding does not seem unreasonable. In Fiji I have observed the five-armed comatulid Eudiocrinus sp. attached by the cirri to the vertical forereef wall, with the arms spread out in contact with the substra­ tum. It is possible that in this situation also the tube-feet may collect particulate food material from the surface of the rock substratum. It is well known that some ophiuroids utilize suspension feeding or deposit feeding depending on the presence or absence of cur­ rents (Magnus 1964, Woodley 1975). Deposit feeding has been suggested for certain stem­ less Paleozoic crinoids that rested directly on the sea floor (Ettensohn 1976, see also Kirk 1911). In living comatulid crinoids of the family Comasteridae the proximal pinnules are spe­ cialized in the possession ofteeth on the distal segments which form a terminal combo The entire pinnule is highly flexible and lashes inward and outward from the oral disk, coiling at the tip. These proximal pinnules appear to function as do the pedicellariae of echinoids to rid the oral surface of foreign particles (Meyer 1973a). However, Gislen (1924) suggested that the comb teeth perform a secondary fee ding function by pinching off particles of algae or bryozoans and transferring them to the food grooves. In many hours of observing comasterids in situ, I have not seen the comb teeth perform this function. The comb­ bearing pinnules are gene rally not long enough to reach the substratum. However, combs of Comaster sp. are present on pinnules far out on the arm. Because some species of this genus cling to the substratum with the arms, it seems possible that these distal comb­ bearing pinnules could perform the 'grazing' function suggested by Gislen. Field documen­ tation is required before deposit feeding and grazing can be accepted as alternative müdes of fee ding in crinoids.

Food and feeding mechanisms: Crinozoa. 27 1.1. Stratification of feeding positions

The elevation above the bottom at which epifaunal organisms feed has been recognized as an important parameter of trophic classification (Turpaeva 1957, Walker & Bambach 1974). Studies of fossil and living crinoids have demonstrated that crinoids can by no means be lumped within a single category as 'high-level' suspension feeders. In the first place, several groups of ancient crinoids lived a pelagic existence at least during part of their life. These include the Devonian Scyphocrinites which was suspended from a terminal float on the stern (Haude 1972), and the gigantic Seirocrinus of the Jurassic which is believed to have attached to floating logs (Seilacher et al. 1968; although an alternative interpretation was given by Rasmussen 1977). Better known examples are the stemless Cretaceous forms, Uintacrinus and Saccocoma (see Breimer 1978). Although many Recent comatulids have swimming ability, none are known to be pelagic. Analysis of stern lengths in species-rich assemblages of mid-Paleozoic stalked crinoids has led to the concept of 'tiered' or 'stratified' epifaunallevel-bottom communities of cri­ noids and other benthos (Lane 1963, 1973, Ausich 1980, Ausich et al. 1979). Interspecific differences in adult stern length form the basis for a three-tiered stratification of these cri­ noid comrnunities (fig.l, Ausich 1980). A different type of stratification has been recog­ nized within Recent communities of reef-dwelling comatulid crinoids. Some comatulids preferentially cling to alcyonarian whips and fans or sponges, providing elevation above the sea floor sirnilar to that of a stalked crinoid. Others gain elevation above the surrounding, reef by perching on prominent coral heads or rocks. These comatulids can be regarded as functional stalked crinoids (Meyer & Macurda 1977). Other comatulid species occurring on the same reefs are semi-cryptic, attaching within hidden crevices and extending the arms at levels closer to the floor of the reef and some are totally cryptic in small caves within the reef infrastructure (Meyer 1973a, b). 1.2. Feeding periodicity

Direct observation of Recent comatulids on coral reefs has revealed that many species are entirely cryptic by day but emerge for feeding by night (Magnus 1963, Meyer 1973b, Rut­ man & Fishelson 1969, Fishelson 1974, Meyer & Macurda 1980). Species feeding noctur­ nally occupy elevated as weil as semi-cryptic feeding sites, and thus contribute to the complex daily cycle of changing resource utilization on a reef. The ultimate biological sig­ nificance of this nocturnal behaviour is not yet understood. One possibility is that noctur­ nal feeders specialize on nocturnally emergent reef plankton. It has also been suggested that disturbance or predation by diurnal fishes has made nocturnal fee ding advantageous (Magnus 1963, Meyer & Macurda 1977). 2. FUNCTIONAL MORPHOLOGY OF FEEDING APPENDAGES 2.1. Tube-feet The morphology of the tube-feet of crinoids has been studied in detail in only a few species (Hamann 1889, Reichensperger 1908, Nichols 1960, 1962b, 1966, 1972, Holland 1969). Available information suggests overall similarity in structure and functior. among aliliving crinoids, stalked and unstalked. Nichols' (1960) light microscopic study of the tube-feet of

28 David L.Meyer

Figure 2A. Arrangementof the tube-feet along a crinoid pinnule; a) Plan view of a pinnule showing relationship between tube-feet (black) and lappets (stippled) ; b) Perspective view of part of a pinnule showing the angles at which tube-feet are held during feeding (lappets omitted); c) Transverse seetion across line ce in A, showing angles at which tube-feet are held in relation to lappets during feeding (after Nichols 1960). Figure 2B. Morphologyof the tube-feet and radial water vascular system in the comatulid crinoidAnte­ don; a) Transverse section of ambulacral part of a pinnule ; b) Enlarged transverse seetion ofa papilla; c) Transverse section of ambulacral part of an arm (after Nichols 1966a).

Food and feeding mechanisms: Crinozoa 29 the comatulid Antedon bifida is the most comprehensive work available and is the basis for the following account. The tube4'eet are arranged in groups of three (triads) along the ambulacral groove of each pinnule (fig.2a) and arm except near the mouth where they are single. These triads alternate along each side of the groove; each triad consists of a long, medium, and short tube-foot. Close_observation of the tube-feet of 15 other comatulid spe­ eies representing four families revealed a similar triad arrangement which is most likely common to allliving crinoids (Meyer 1979). The long or primary tube-foot extends between the flap-like lappets, which are fused to the proximal parts of the medium and short tube­ feet of each triad. Contraction of the medium and short tube-feet causes closure of the lappets over the ambulacral groove. In extended position, the primary tube-feet project laterally at a slight angle to the plane of the groove, while the medium tube-feet project at about 45 degrees to the groove and the short tube-feet stand roughiy perpendicular to the groove (fig.2a). The tube-feet are armed with papillae bearing terminal sensory hairs and containing mucus glands (fig.2b). Nichols (1960) reported a single muscle fiber within each papilla, but Holland's (1969) electron microseopie examination of the papillae revealed that the supposed muscle fiber in the comatulid Nemaster rubiginosa is a bundle of microtubules of uncertain function. According to Nichols (1960), contact of food particles with the sensory papillae stimu­ lates forcible ejection of mucus strands and lashing of the tube-feet inward toward the food groove. Fooiparticles ensnared in mucus are passed from the long to the medium to the short tube-feet which pack the material into a mucus string within the groove. This string is transported along the groove of the pinnules and arms to the mouth by ciliary cur­ rents. I have frequently observed this string travelling along the ambulacral groove of cri­ noids feeding in situ. Thus, the characterization of crinoids as 'ciliary-mucus feeders' by Hyman (1955) seems justified, although it neglects the role of the tube-feet. Nichols (1960, 1966, 1972) used the terms 'food net' and 'mucus net' to describe the means by which food particles are trapped by the tube:feet. Subsequent authors have employed the term 'mucus-net fee ding' to describe the crinoid feeding mechanism. This implies a weblike network of mucus strands extending over and between the extended tu~e-feet. While I have seen strands ofmucus clinging to the arms and pinnules (see also Magnus 1963), there is no evidence of a weblike net. It is clear from Nichols' work that ejection of mucus strands is restricted to those tube-feet stimulated by particle contact, and that particles are not trapped by a preformed web of mucus. Other invertebrates such as tunicates, poly­ chaetes, echiurids, and possibly pelecypods utilize a true mucus net or sheet in feeding (MacGinitie 1941, 1945, Monniot 1979), and thus application ofthe term 'mucus net' to crinoids should be abandoned because of its incorrect connotation for the feeding mecha­ nism.

2.1.1. Tube-feet as efficient filters Significant differences in spacing and length of the primary tube-feet in several sympatric comatulid species on Western Pacific reefs are correlated with feeding position and feeding posture (fig.3, Meyer 1979). Species having longer, more widely spaced tube-feet are gene­ rally those living semi-cryptically within the reef infrastructure, utilizing a multi-directional feeding posture. Crinoids with relatively shorter, more closely spaced tube-feet occur on elevated perches and utilize a ftltration fan feeding posture (see seetion 2.2.). Similar studies by Liddell (1980) in the Carribean a~e in general agreement with these results.

F igure 3. Relat ionship between spacing (abscissa) a nd length 01' tube-fee t in co m a tu lid crinoids . Pl o tt ed poi nt s are mea n spaci ng a nd le ngth with 95 % co nfi de n ce interva ls for eac h species . A. Crin o ids from Pa lau l sla nds, Weste rn Pac ifi c; B. Crin oi d s fr om Liza rd l sland, Great Barr ier R eef. Comas t eri dae : CPA R - Coman t hus parvicirrus, CGRC - Comast er gracilis, CSCH - Comanthina schlege li, CAP - Cap illa ster multiradiatu s, CMLT - Co mas t er multi· [ id u s, CN IG - Comatella nigra , CBE N - Comanthus benn e tti. Co lobometridae: COLO - Co lobometra persp inasa , CE NO - Cenome t ra be lla . PON T ­ Po nt iometra ander soni, O LlG - Oligometra serripinna. Tropiometrid ae: TRO P - Tropiome t ra afra. Maria me t ridae : LAMP - Lamprometra palma ta. Him ero m etridae : HROB - Himerometra robustipinna, HMAG - Him erometra magnipinna (after Me yer 19 7 9) .

w

"i: ~

~

t-



~

;:s

~

~

~

t-­

~

:=­ ;:s

~

0\

IN

1. 2. 3. 4. 5. 6. 7. 8.

2.5 2.5 ± 0.8 2.5 ± 2

55 ±5

1.5

15 24 14 12

20

14 4.1 ± 0.2

10 17

Wet gonad weight/wet body weight Dry gonad weight/wet body weight Organie gonad weight/organic body weight Dry gonad weight/dry body weight Wet gonad weight/wet eviscerated body weight Gonad volume/wet body weight Wet gonad weight/wet body tissue weight Gonad volume/wet eviscerated body weight

Sclerodactyla briareus

Id. Id.

Cucumaria lubrica Cucumaria pseudocurata Holothuria scabra Stichopus japonicus

HOLOTHUROIDEA

Id.

Tripneustes ventricosus

Id.

Tripneustes gratilla

Id. Id. Id. Id.

Strongylocentrotus purpuratus

Id. Id.

Strongylocentrotus intermedius

Id.

Strongylocentrotus nudus

Id. Id.

0.1 0 1.5 ± 0.5

5±2

1 1 2 5

1 2

3 0.6 ±0.4

0.04 0

20

2 ±0.5 0.3

2 3.5

3.5 ± 1 5±2

lO ±3

Wet gonad weight/wet disc weight Dry gonad weight/wet disc weight Dry gonad weight/body diameter Wet gonad weight/test diameter Gonad volume/test volume Gonad volume/(test diameter)3 Wet gonad weight/test volume

0.35 18

0.1 12

9. 10. 11. 12. 13. 14. 15.

35

15.5 ± 1.5 1.2

23 16.6

28.3 ± 2.6 30 ± 2

30 ± 1

52

14.5 ± 1.5 1.2

24 17.5

25.1 ± 2.8 30 ± 2

30 ± 1

3 8

0.04 0

24

2 ±0.5 0.3

1.8

4.3 ±0.6 6±2

10 ± 1

1 1 1

1 4 1

14

13

13 13

6 6 1 1 4

15 15

6 1 6

Engstrom 1974 Rutherford 1973 KrisIman 1967 Choe 1962 Tanaka 1958a Tanaka 1958b Boolootian 1966

Greenfield et al. 1958 Baker 1973 Fuji 1960c Yakolev 1976 Yakolev et al. 1976 Fuji 1960a Fuji 1960c Yakolev 1976 Bennett & Giese 1955 Giese et al. 1958 Holland & Giese 1965 Lawrence 1966 Gonor 1973a Kobayashi 1969 O'Connor et al. 1976 Moore, Jutare, Jones & McPherson 1963 Lewis 1966

Vt

W 0\

""

~

~

(;)

S·

(l)

g.

r:;'

-ti::s­

(;)

~ ~

(l)

~

""';'"

~ ~

~

;:s

('i;'

:;­

I::

~ ;:s

;:s

~

...6'

E:: N'

I::

~ (l)

366 lohn M.Lawrence & l.M.Lane vanbrunti was equal, although the size at which it occurred was greater in the latter species (Lessios 1979). Evechinus chloroticus became mature at three to four years of age in two localities different in the avililability of food, although individuals were stunted in size in the locality with less food (Dix 1970c). There are few reports of the size at sexual maturity in asteroids. The brooding Asterina gibbosa becomes sexually mature at 5 mm diameter (Emson & Crump 1976). Sexual maturity was related to species body size by Menge (1974, 1975) who reported that the brooding Leptasterias hexactis matured at 2 g wet body weight and two years of age, while Pisaster ochraceus matured at 70-90 g wet body weight and five years of age. It is not known whether these differences are genetically or nutrition­ ally controlled. The absolute weight of gonads increases with body size, the slope may be positive ini­ tially, but subsequently decreases with size in echinoids (Moore, Jutare, Bauer & Jones 1963, Moore, Jutare, Jones, McPherson & Roper 1963, Ebert 1966, Moore & Lopez 1966, Fuji 1967, McPherson 1968c, Dix 1970c, Gonor 1972, Miller & Mann 1973, Lilly 1975, Greenway 1977) and asteroids (Menge 1970, Menge & Menge 1974). A direct relationship has been noted between body size and the number of eggs brooded by a holothuroid (Eng­ strom 1974), the number and size of ova of an ophiuroid (Hendler 1975), and the number of eggs produced by an asteroid (Emson & Crump 1976), a holothuroid (Rutherford 1973, 1977), and an echinoid (Moore 1934, Timko 1975). The decrease in relative reproductive effort with an increase in body size is indicative of an allometric physiological relationship sirnilar to that of other size functions (see seetions 2.1 and 3.4) and perhaps related to them. It probably exists in all echinoderms. Whatever the basis of the relationship, the nutritional allocation to reproduction reaches a maximum and does not increase further with the slow increase in size. The relationship should be documented for other species and classes. 3.3. Sex and gonadal growth As the sexual products of males and females differ, it rnight be expected that the nutrition of the two sexes would be different. One indication of the difference is the biochemical composition of eggs and sperm. A primary distinction is that the primary lipid component of eggs is neutral lipid (Allen & Giese 1966, Lawrence et al.1966, Lawrence 1973, Hayashi & Takagi 1977, Oudejans & Van der Sluis 1979a,b), while that of sperm is phospholipid (Allen & Giese 1966, Lawrence et al. 1966, Lawrence 1973). It should be noted that with few exceptions (e.g. Turner & Lawrehce 1979), analyses have been made on whole gonads and not gametes. Even at the maximal gonad index, when mIed with gametes, the gonad has non-gametic tissue which comprornises to some extent the values obtained. It is strongly recommended that analyses of gametes be made rather than of whole gonads. The compo­ sition of the eggs themselves can vary with locality (Turner & Lawrence 1979) and from year to year (Öhman 1945, Turner & Lawrence 1979). There seems to be no difference in the biochernical components of the pyloric caeca of male and fernale asteroids (Allen 1965, Lawrence 1973, Ferguson 1975a, 1979) to the degree to which they have been analyzed. This indicates that there is no sexual differences involved in the nutrition of reproduction as far as the type of nutrient reserves is concerned. In addition to the difference in type of lipid present, the level of lipid is frequently greater in the ovaries than in the testes (table 8). Even in equal-size gonads, more energy would be deposited in the ovaries than in the testes in these species as a consequence. The ovaries also are slightly larger than the testes in many asteroids and echinoids (table 7),

The utilization 01 nu trients by post-metamorphic echinoderms 367 Table 8. Ratio of the level (%) of lipid in the ovaries at maturity to that of the testes at maturity in echinoderms Species

ASTEROIDEA

Asterias vulgaris Luidia clathrata Odontaster validus Oreaster herdnuznni Pisaster giganteus Pisaster ochraceus

ECHINOIDEA

Allocentrotus {ragi/is Arbacia /ixula Psammechinus miliaris Sterechinus neunuzyeri Stomopneustes variolaris Strongylocentrotus {ranciscanus Strongylocentrotus purpuratus

Ratio

References

1.6 2.0 1.5 1.9 1.9 2.5

Lowe 1978 Lawrence 1973 Pearse 1965 Rao 1965 Greenfield et al. 1958 Greenfield et al. 1958

1.1 1.8 1.3 1.3 1.3 1.0 1.3

Greenfield et al. 1958 Fenaux et al. 1977 Sukarno 1979 Pearse & Giese 1966a Giese et al. 1964 Greenfield et al. 1958 Giese et al. 1958

again indieating a greater nutrient drain on the female than on the male. Either the females have a higher nutrient intake, are more effieient in utilizing the nutrients abosrbed, or they have less nutrients available for growth or maintenanee. There have been no reports of any of these differenees between males and females. The maximal pyloric eaeea is equal in male and female Asterias vulgaris (Lowe 1978) and Oreaster reticulatus (Seheibling 1979), and only slightly higher in the fe male in Pisaster giganteus (Greenfield et al. 1958). If attaining the maximal gonad index is dependent on reserves in the pylorie eaeea (see seetion 3.5), one would expeet a sexual differenee in the maximal pylorie eaeea index. Potential sexual differenees in the pylorie eaeea related to sexual dimorphism in gonadal size and eomposi­ tion should be investigated further. The testes are mueh larger than the ovaries in two asteroids, Echinaster sp. (Ferguson 1974, 1975a) and Leptasterias hexactis (Menge 1970, 1975) (although the ovaries are slightly larger th~ the testes in Leptasterias pusilla, another brooding species ([R.H.Srnith 1971]). Menge (1970) eoncluded that the greater size of the testes in L.hexactis is due to the brooding behavior of females in this species (Le. brooding is energetieally eostly to fe males and the larger number of eggs eannot be produeed). The testes are not larger in Asterina gibbosa, another brooding asteroid (Emson & Crump 1976), but this speeies is protandrie. The greater energy investment in oogenesis whieh seems indieated for eehinoderms may be involved in protandry. Although most eehinoderms are gonoehorie, some are protandrie (Delavault 1966). Protandry plaees in a single individual the sequenee in the size differen­ tial at whieh testes and ovaries appear in separate individuals as has been noted in some gonoehorie speeies (see seetion 4.2). Brooding and protandry both have been related to the level offood supply inA.gibbosa (Emson & Crump 1976, 1979). Individuals beeome mature females and brood at a larger size where there is a greater variety of food and more animal prey. The appearanee of testes in smaller individuals results in the produetion of a greater number of funetional gametes than if an ovary were produeed. Menge (1975) sug­ gested that brooding in L.hexactis is a eonsequenee of small size whieh restriets the number of ova whieh ean be produeed. If these are seleetive pressures, they have not been adequate

368 lohn M.Lawrence & l.MLane to produce protandry and/or brooding in many small echinoderms which have poor food supplies. Brooding or internal development can be an additional energy drain on females. Fifteen species of echinoderm develop young within the ovary or coelom (Turner & Dearborn 1979). These eggs, like those of planktotrophic species, apparently do not contain adequate nutrients for development, implying nutrient transfer from the mother. The amount of transfer has been estimated only for the intraovarian development of the ophiuroid, Ophio­ notus hexactis. The organic content was calculated to increase as much as 100,000-fold during the incubation period (Turner & Dearborn 1979). This phenomenon is a very inte­ resting aspect of the nutrition of reproduction and further study should be rewarding. Brooding can also have a nutrition al impact by interfering with fee ding. Oral brooding by the asteroid L.hexactis results in utilization of nutrient reserves in the pyloric caeca of fe males (Menge 1975). Gladfelter (1978) reported that growth in non-brooding Cassidulus cariboearum was greater than in brooding individuals, but a mechanism was not suggested. Strathmann (1979) concluded that few hypotheses accounting for the high frequency of brooding with small adult size apply to aIl taxa, and that more than one hypothesis apply to most.

3.4. Food and gonadal growth An effect of quantity of food on gonadal growth has been postulated for asteroids (Galts­ hoff & Loosanoff 1939, Srnith 1940, Vevers 1949, Pearse 1965b, Allen 1965, Mauzey 1966, 1967, Crump 1971, Crump & Emson 1978, Lawrence & Dehn 1979) and echinoids (Moore 1934,1935,1937, Boolootian et al. 1959, Fuji 1960a,c, Kawamura 1964, 1965, 1966, 1973, Holland 1965a, Kawamura& Taki 1965, Buchanan 1966, Ebert 1966, Pearse & Giese 1966b, Leighton 1968, 1971, Vadas 1968, 1977, McPherson 1969, Pearse 1969b, Pearse et al. 1970, Gonor 1972, 1973a,b, Sumich & McCauley 1973, Fletcher et al. 1974, Kobayashi & Tokioka 1976, O'Connor et al. 1976, Kasyanov et al. 1977, Masuda & Dan 1977, Yako­ lev 1977, Regis 1979b) and a holothuroid (Rutherford 1977). Experimental demonstration of the effect of type of food on gonadal growth has been demonstrated for asteroids (Smith 1940, Hancock 1958, Crump 1971) and for echinoids (Fuji 1967, Leighton 1968, 1971, Vadas 1968, 1977, Lilly 1975). As with somatic growth (see section 3.3), it is difficult to separate the effects of the different characteristics of food which affect 1) rate of consump­ tion, 2) digestibility, 3) absorption, and 4) composition. The work on echinoids indicates that all characteristics are important and that, in general, food which supports somatic growth weIl also supports gonadal growth. This implies considerable versatility of the meta­ bolic pathways. Data on the effect of composition of food on gonadal growth are essentiaIly non-existent, e.g. it is known that brown algae support gonadal growth in echinoids, but it is not known what components are important. An important area for future work is the identification of the nutrient composition of foods so that an evaluation of the necessary factors can be made.

4. BIOENERGETICS Arecent interest in describing energy flow through natural systems, from sub-cellular to community levels, has provided insight into the energetic importance and internal function­

The utilization o[ nutrients by post-metamorphic echinoderms 369 ing of those systems (Odum 1962, Slobodkin 1962, Phillipson 1966, Odum 1968, Weigert 1968). A standard equation for measuring energy flows and a common unit, the calorie, have been adopted so that these often large studies may be comparable and comprehend­ able. The balanced energy equation is given as C = R + P + F + U (Ricker 1968); where C is energy consumed, R is energy for maintenance or respiration, P is energy of production or growth, Fis unabsorbed energy or energy in feces and U is energy lost as excreta (urine, mucus, etc.). Terms of the equation may be measured with varying degrees of accuracy; often one or more terms are estimated or found by subtraction. Energy studies on echinoderms are few. Energy studies for Strongylocentrotus droe­ bachiensis (Miller & Mann 1973, Propp 1977a), a regular .northern echinoid, Parechinus angulosus (Greenwood 1980), a regular South African echinoid, and Mellita quinquiesper­ [orata (Lane 1977), an irregular sub-tropical echinoid, are the four studies concentrating on energy aspects and using the energy equation. A less detailed energy budget (giving only P and R terms) was presented for Brisaster latifrons, a spatangoid, in Puget Sound, Washing­ ton (F.H.Nichols 1972, 1975). Other authors have dealt with the energy question, especially in feeding studies, but have not presented a straight-forward energy budget using the energy equation and calories as the unit of measurement (Moore, Jutare, Bauer & Jones 1963, Moore, Jutare, Jones, McPherson & Roper 1963, Moore & McPherson 1965, Fuji 1967, Leighton 1968, McPherson 1968a,b). In spite of the problems encountered in constructing energy budgets, the energy budgets of the echinoids, S.droebachiensis (Miller & Mann 1973), Parechinus angulosus (Greenwood 1980) and M.quinquiesper[orata (Lane 1977) were markedly similar and pointed out some interesting aspects about echinoid physiology which had not been considered. The energy budget presented by Propp(1977a) forS.droebachiensis gave slightly different results than the other three studies. These three studies either demonstrated (in the case of M.quinquies­ per[orata) or postulated by finding an imbalance in the energy equation (in the case of S. droebachiensis and P.angulosus) that dissolved organic carbon (DOC) constituted a large loss of absorbed energy (40 to 99 % in M.quinquiesper[orata, 24 to 53 % in P.angulosus, and 40 to 80 % in S.droebachiensis). Ferguson (1980) calculated that uptake of dissolved arnino acids by the asteroid Echinaster exceeds release by a ratio of nearly 4 to 1 at normal concentrations: release exceeds uptake at ambient about 25 % amino acid concentrations. Loss of dissolved organic carbon may be wide-spread among echinoids as Miller & Mann (1973) pointed out from examination of the data of Moore & McPherson (1965), Fuji (1967), and Leighton (1968). Direct measurement of DOC is needed to quantify the amount of release. The origin and fate of DOC release are presently unknown. Miller & Mann (1973) postulated that organie materialleaked or washed from the gut may be the source of DOC. Field (1972) observed that S.droebachiensis fed Laminaria longicruris released more DOC than unfed individuals. Field acknowledged that some of the released DOC could have come from feces, but did not consider the possibility of release of DOC in water flushed from the gut. Field (in Ebert 1975) suggested that regular echinoids may eat large quantities of plant food in order to satisfy their protein need, defecating most of the non-digestible carbo­ hydrate. Lilly (1975) proposed that the regular echinoid Tripneustes ventricosus improved the calorie:protein ratio by selectively absorbing protein from most foods, using this mecha­ nism to overcome the problem of protein-deficient foods. Lilly did not distinguish between soluble and structural carbohydrates in his study. Lawrence (1976b) found that four species of regular echinoids including T. ventricosus, absorbed soluble carbohydrates at the

370 lohn MLawrence & I.MLane same efficiency as protein. The levels of protein and soluble carbohydrate were similar in the food used in his study (Thalassia testudinum) which would mean that similar amounts would be absorbed. Propp (1977 a) hypothesized that excessive amounts of organic material are absorbed in order to gain sufficient nitrogen and phosphorus for nutritional use, the excessive amounts of carbohydrate being released as DOC. However, this implies no control over which compounds are absorbed. With the known role of active transport of the pro­ ducts of digestion, it is unlikely that the products of carbohydrates would be actively trans­ ported only to be released without use. The energy budget of Mquinquiesper[orata demonstrated that uptake of amino acids (as measured by glycine) was a relatively insignificant source (-0.5 %) of the total energy absorbed (Lane 1977). Ferguson (1980) measured rates of amino acid uptake using ambient concentrations and estimated that uptake could only provide approximately 5 % of the respiratory energy in an asteroid. Several authors (Clark 1969, Pearse & Pearse 1973, Stephens etal. 1978) have been concerned whether there is a net influx or amino acids into echino­ derms or not. This problem wou1d seem insignificant, at least in terms of energy require­ ments. However, more work dealing with energy provided by direct uptake of amino acids as weH as other dissolved carbon sources are needed. . Energy characteristics of populations of Pangulosus, S.droebachiensis (as given by Miller & Mann 1973, but not Propp 1977a), andMquinquiesper[orata were similar to each other and to those of other marine invertebrates. These echinoid populations had production esti­ mates that were moderate in comparison to other marine invertebrates. Production (body and gonad growth) comprised approximately 25 % and respiration comprised 75 % of assi­ milated energy (P + R). One and two year old individuals accounted for approximately half the production estimates in S.droebachiensis (Miller & Mann 1973) andMquinquiesper[o­ rata populations and one population of Pangulosus, due to their greater numbers and faster rate of growth. The regression equation given by McNeill & Lawton (1970) describing the relationship between respiration and production fit the data of these echinoid populations fairly weH. The production/consumption ratio for the three echinoid populations was low (-5 %) when compared to other marine invertebrates, indicating low rates of food absorption for these echinoids. In contrast, the population of S.droebachiensis studied by Propp (1977a) had a production/consumption ratio of 25 to 30 %. Miller & Mann (1973) and Lane (1977) found that food absorption rates were variable. Greenwood (1980) found decreasing absorption with increasing age but the results were based on limited data. Production esti­ mates for these populations did not include the loss of dissolved organic carbon. IncIuding dissolved organic carbon in the production term would have made these populations seem very productive when compared to other invertebrates, and aH efficiencies calculated using the production term would have greatly increased. The net growth efficiency for the popu­ lation of Mquinquiesper[orata in 1973 would have changed from 10 to 71 %. Therefore, describing the productivity of these echinoid populations depends on whether the energetic contribution of dissolved organic carbon is incIuded or not. Energy characteristics of popu­ lations are affected by the number of individuals and the average age of the individuals in that population as demonstrated by Greenwood (1980). In populations where recruitment and mortality vary from year to year, energy characteristics would also be expected to vary. Population energy budgets for Mquinquiesper[orata were slightly different in the two years studied, due to the dorninance of one year cIass in both years and its changing energy characteristics.

The utilization o[nutrients by post-metamorphic echinoderms 371 Unlike population energy budgets, energy budgets for individuals of different age or size are not affected by density and age dependent variables. Hence, they may better reflect the energy characteristics of that species. In most echinoid species studied, relative production or net growth efficiency decreases with increasing age. This was true for P.angulosus (Green­ wood 1980), 8.droebachiensis (Miller & Mann 1973, Propp 1977a), Strongylocentrotus inter­ medius (Fuji 1967), and Lytechinus variegatus (Moore, Jutare, Jones, McPherson & Roper 1963). For Strongylocentrotus purpuratus (Leighton 1968), growth efficiencies remained nearly unchanged with increasing age. In M.quinquiesper[orata, body growth declined with age (Lane 1977); however, a relatively great increase in the productivity of the gonad caused total production and net growth efficiency to increase from 0 through 3 years of age (7 % at 1 year and 17 % at 3 years). As has been noted in other echinoids (Moore & McPherson 1965, Gonor 1972), the relative amount of gonad produced increases with increasing age and then stabilizes. This increased gonad production is usually not sufficiently great to off­ set the declining productivity of body growth. Hence in the echinoids studied, ex cept M quinquiesper[orata, total productivity declines with age. Growth efficiencies may change also with season of the year (Lane 1977), diet (Fuji 1967, Leighton 1968), state of nutri­ tion and injury (Ebert 1968). From their energy studies, Miller & Mann (1973) estimated that energy requirements of populations of S.droebachiensis were only 7 % of algal production. An ecological efficiency for Mquinquiesper[orata was not calculated due to the difficulty in measuring production of food ingested (detritus). Energy budgets provide a framework for an evaluation of the various ways in which nutrients are utilized and for comparisons of taxa with different characteristics. Among those of interest are the allocation of different amounts of energy to the various body com­ ponents, the utilization of energy during prolonged exposure to different levels of food, the utilization of energy with pe rio die change in level of food and the difference in energy requirements with sex. Energy budgets also provide a method for evaluating the trophic role of echinoderms in terms of energy which goes into populations, maintained in popula­ tions, and released into the environment for utilization.

5. CONCLUSIONS Nutrition is a pervasive aspect of the biology of echinoderms and its study provides insight into the ways in which individuals and populations function. Although various aspects of the nutrition of post-metamorphic echinoderms have been investigated extensively, the species for which information is available is lirnited. Studies must be extended to other species and taxa and to other environments to fu1ly understand and evaluate the character­ istics of echinoderm nutrition.

16

JOHN C. FERGUSON

NUTRIENT TRANSLOCATION

The problem of distribution of nutritive materials to the various body parts of echinoderms has puzzled investigators for over 150 years. Indeed, it was in response to a prize offered by the French Institute to study circulation in asteroids, echinoids and holothuroids that Friedrich Tiedemann (l816) produced as his dissertation what was one of the first impor­ tant studies of these forms. The illustrative plates of this work are unmatched for their artistry, beauty and richness of detail. However, students had to await the development of histological techniques three quarters of a centwy later before the full complexity of the circulatory (or 'hemal') systems of echinoderms could be delineated. This was done, with great care, by Perrier (1875), Hamann (1884-1889) and Cuenot (1887-1948). The electron rnicroscope has begun to provide still deeper insight into the structure of these systems. As supposed circulatory systems, the hemal tissues of echinoderms clearly have major deflciencies. While remarkedly weIl developed in a few holothuroids, they are mostly incon­ spicuous in other types. They generaIly do not form clear cut vessels, but consist instead of spongy strands of tissue, partially or completely blocked by flbers and cells, and containing a gelatinous, PAS-positive ground substance. They are limited in their distribution in the body, are unidirectional in their orientation, and generaIly lack effective means of propul­ sion (although pulsations are described in some parts). For these reasons there has always been a reluctance to accept these tissues as being functional in the role their name implies. Consequently, the process of nutrient translocation has been repeatedly ascribed to other mechanisms. Primary among these are transport by coelomocytes, and circubttion via other body cavities, especiaIly the perivisceral coelornic cavity, the perihemal coelomic cavities, and the water vascular system. In addition, direct translocation of exogenous nutrients through the integument has been proposed. While knowledge continues to be amassed at a rapid pace, there is as yet no real consen­ sus on the relative signiflcance of these possible translocation mechanisms. In large part this is because the bulk of the information available is morphological in nature, and the morphology is unique, very difficult to understand, and has few accurate paralieis in other animal groups. Thus, it is from this confusing morphology that functional interpretations have had to be developed with little corroboration by experimental procedures. These inter­ pretations, then, have been excessive on speculation and short on evidence. Furthermore, morphology, especiaIly when it is as difflcult as this is, often tends to be itself subjective and interpretive - one tends to see what one expects to see. Fortunately, thanks to the great skill of many of the workers in the fleld, much accurate knowledge has been accumu­ lated in spite of the speculative conclusions often drawn from it. The following review will try to emphasize these sound observations and provide a compilation of the factual evidence, whlle largely deemphasizing, for the time being, the various conflicting interpretations that have been made of this evidence by those who have acquired it.

373

374 lohn C.Ferguson 1. TRANSLOCA TION SYSTEMS

1.1. The hemal system 1.1.1. Morphology The general form and development of the hemal tissues in various echinoderm groups has been reviewed by Hyman (1955), and was largely delineated by the work of Perrier (1875), Hamann (1884-1889), and Cuenot (1887-1948), with significant contributions by Ludwig (1889-1907). It usually consists of radial strands of tissue, running beneath the water vas­ cular canals, which connect with a pharyngeal hemal ring. This in turn may connect acen­ trally to a complex axial organ, attached aborally to a genital hemal ring. Associated strands of hemal tissue may be found in connection with the digestive organs and the gonads. While the axial organ is usually grossly visible in most species, the other tissues must often be studied through the use of histological sections. The major exception is in holo­ thuroids, which interestingly lack an axial organ, but often have a well-developed hemal plexus (rete mirabile) along the 100ps of the gut and respiratory trees. This plexus may share some histological characteristics with the axial organs of other echinoderms.

Figure 1. The complex hemal system of Stichopus moebi (after Herreid et al. 1976).

1 - esophagus; 2 - stomaeh; 3 - 'hearts'; 4 - primary dorsal vessel; 5 - dorsal transverse vessel; 6 ­ small intestine; 7 - ventral transverse vessel; 8 - large intestine; 9 - secondary dorsal vessel; 10 - ven­

tral collecting vessel; 11 - primary ventral vessel; 12 - vascular 'follicles'.

Nutrient translocation 375 The level of development of the hemal system can vary considerably between different species, but reaches its highest form in the holothuroid, Stichopus moebii (Herreid et al. 1976, 1977). Here it reportedly consists of a true, closed system of vessels generally lined with a non-ciliated endothelium, and covered on the outer surface by a typical peritoneum underlain by a thick connective tissue layer (fig.!). Colorless fluid (which can be visualized with methylene blue) is described as being pumped via 120-150 muscular, single chambered 'hearts' from a large dorsal vessel, through aseries of 'intestinal plates' and a plexus of ves­ sels into a ventral vessel. From here the fluid is conducted to the region of the small intes­ tine where it can pass up through another plexus of vessels associated with the left respira­ tory tree, or through other intestinal vessels, to the dorsal vessel. Anteriorly, the vessels seemingly diminish and do not appear to connect to a hemal ring (as may occur in other forms). The direction of flow through much of the system is reversible. The connective tis­ sue associated with the vessels of the respiratory plexus appears rich in neutral and acid mucopolysaccharides, and may possess lacunar spaces mied with many granular and agra­ nular coelomocytes. Especially in the respiratory plexus, but also in the rest of the septum, are found collections of necrotic cells containing much lipid and many iron granules. In Stichopus californicus, Prosser & Judson (1952) have described two large hemal ves­ sels along the intestine, one internal and one external. The external vessel consists of villus epithelium, circular muscles, and connective tissue with regions of hemopoiesis. It beats spontaneously at 4-5.5 beats per minute. The contractions are inhibited by acetylcholine and other drugs. Jensen (1975) has provided a more detailed description of the 'dorsal' hemal vessel of Parastichopus tremulus. It contains a circular muscle layer and nerve fibers beneath a vacuo­ lated peritoneum, and an inner connective tissue layer that nearly fills the lumen. While the elaborate forms of the hemal system of Stichopus spp. may be somewhat unique, related species also have well-developed hemal structures. The complex system of Holothuria tubulosa was beautifully illustrated by Tiedemann (1816) and again by Enriques (1902). Holothuria forskali was studied by Stott (1957). In his perhaps sirnplified illustra­ tion (fig.2b), a system ofvessels converges from the anterior loop (stomaeh) and leads into a single vessel, which in turn leads to an area of extensive branching over the middle gut (intestine ) in which the respiratory tree is interlaced. The vessels of this species do not have epitheliallinings, but do possess scattered muscular tissue and an outer peritoneum. Within the lumen are many granular and agranular coelomocytes. The thickened peritoneum of the intestinal part was seen to be heavily loaded with black, brown, yellow, or clear gra­ nules. When fed with iron saccharate, the iron appeared in coelomocytes in the left respira­ tory tree, the gut, and in trace amounts in the dorsal hemal canal of the gut, but not in the more complex channels. In Cucumaria elongata the hemal system is less weil developed and a complex hemal plexus (rete mirabile) is absent (Fish 1967a). Transverse hemal connections between the ves­ sels of different parts of the gut are seen, however, and the hemal vessels connect with a connective tissue-fluid compartment found beneath villus-like projections into the intesti­ nallumen (fig.2a). Coelomocytes are abundant in these parts. A pharyngeal hemal ring could not be found. The vessels are covered with a variable peritoneum containing strands of connective tissue underlain by a thin layer of circular and longitudinal muscle fibers, and a lining of indistinct connective tissue. Doyle (1967b) described an endothelium of cells 'like fibroblasts' in Cucumaria frondosa. A surprisingly extensive system of vessels, including a rete mirabile and a pulsatile affe­

376 lohn CFerguson

Figure 2. Holothuroid hemal systems. A. Cucumaria elongata (after Fish 1967a), B. Holothuria jorskali (after Stott 1957). 1 - ventral sinus; 2 - pharyngeal hemal ring (not seen) ; 3 - dorsal sinus; 4 - brown or yellow rete mirabile; 5 - unpigmented rete; 6 - ventral vessel; 7 - ventral connective; 8 - gonadic vessel; 9 - dorsal vessel; 10 - black rete; 11 - intestine.

rent vessel exists in Caudina chilensis (Kawamoto 1927). Flow of the hemal fluid is illus­ trated as progressing in a compiete circulation. In many of the smaller holothuroids, however, the hemal system is Iess complicated, and exists primarily äs a network of vessels along the gut, interconnecting between the Ioops. Bell & Farmanfarmaian (1967) briefly reported electron microscopically visible con­ nections between the hemal vessels and intestinaIIacunar spaces in Sclerodactyla bria­ reus, apparentIy similar to those reported by Fish (1967a) and earlier workers. These were described more compietely by Farmanfarmaian (1969a,b). He indicated that the connective tissue Iacunae contain a c1ear fluid that has the same staining properties as that in the me sen­ teric hemal channeis with which they connect. The contents of the he mal channels are moved in an ebb and flow pattern by weak Iocal contractions, but there is no net circuIa­ tion. Farmanfarmaian further reported that the gut in Leptosynpata (inhaerens?) produces a vigorous churning motion in the mesenteric hemal vessels which contain a 'viscous yellow fluid'. Again, there is no evidence of net circulation. When the vessels of both of these spe­

Nutrient translocation 377 cies were cut, there was only a slow leakage of viscous yellow fluid. The total capacity of the hemal system of S. briareus was estimated as 0.02 % of the total body weight. Histo­ chemically, the fluid appears rich in neutral polysaccharides or possibly non-sulphated mucopolysaccharides or glycoproteins. The lining üf the sinuses contain acid mucopoly­ saccharides. The fluid also contains lipofuscin spherules and coelomocytes. The basic form of the hemal system of echinoids (as reviewed by Cuenot 1948, and Hyman 1955) is somewhat sirnilar to that ofthe other classes, except that it is generally more fully developed. A ring of hema1 tissue encircles the esophagus, on the top of the lan­ tem, and gives off fine branches to each ambulacrum. These are located in the perihemal canal between the radial nerve and the water vascular canal. Branches are given off to the tube feet and, perhaps, other parts. Two major hemal channels are also found along much of the gut, giving off fine branches into the wall. They are apparently connected to the hemal ring by a single vessel. A conspicuous, spongy axial organ accompanies the stone canal through the perivisceral coelom from the hemal ring to the aboral surface, where it connects to an aboral he mal ring that may give off branches to the gonads (fig.3). A modern description of the system of Strongylocentrotus purpuratus has been pro­ vided by Campbell (1966). Much of the more recent work on this class has centered on the

Figure 3. Hemal system of Echinus sp. (after CUI!not 1948).

1 - genital hemal sinus; 2 - terminal sinus; 3 - aboral hemal ring; 4 - radial sinus; 5 - collateral sinus;

6 - external marginal sinus; 7 - intern al marginal sinus; 8 - oral hemal ring; 9 - axial organ.

378 lohn CFerguson

nature of the axial organ. This interest was in large part generated by the provocative claim of Boolootian & Campbell (1964, 1966), largely based on 'motion picture analysis', that S.purpuratus contains, within its axial organ, pulsatile vessels that form a true heart which circulates fluid throughout a closed hemal system. Spontaneous contractions of various parts of the hemal system have been noted by all workers back to Tiedemann (1816). Burton (1964) reported that while she also could see contractions in Echinus esculentus and Psammechinus miliaris, the hemal tissue next to the axial organ was spongy and often occluded. Movement of the contents was largely ebb and flow. When injected with dyes, some net flow in both directions could be detected. Millott & Vevers (1964,1968), Millott (1966,1967,1969), and Vevers (1967), working with Arbacia lixula, Diadema antillarum, E. esculentus, and Paracentrotus lividus, concurred with Burton and described the axial organ as composed of connective tissue and muscle ceIls, and being glandular by the transformation of cells, predominantly coelomocytes. They noted a number of direct connections between the perivisceral coelom, the major hemal vessels, and the lumen of the axial organ, as weIl as connections with the water vas­ cular system. By histochemical analysis, they found abundant acid mucopolysaccharides, lipid and protein-bound reducing groups, indole derivatives, pigments, and various other compounds. Boolootian & Campbell (1966) chemically analyzed fluid from the hemal sinus of s.purpuratus and found it to contain 131.6-168 mg% total nigroten and 48.3­ 63.0 mg% reducing sugar. The ultrastructure of the axial organ of P.miliaris has been studied by Jangoux & Schal­ tin (1977), who were unable to detect endogenous glandular tissue. They indicated that the organ possesses, on its external surface, numerous openings to dead-end canals mIed with coelornic fluid. Several ingrowths into the interior of the 'axial sinus' form a muscular tube (pulsatile vessel), which penetrates dorsally to the madreporic ampulla. The organ does not possess its own rnitotic ceIls, but only coelomocytes which enter by diapedesis. Most of these cells are degenerate phagocytes containing exogenous material. A thick nervous layer, with possible neurosecretory elements, exists near the stone canal. Bachmann & Goldschrnid (1978) observed a very similar structure in the axial complex of Sphaerechinus granularis. In asteroids there are again found radial hemal tissue in each ambulacrum, embedded in a membrane partially dividing the perihemal coelornic sinus. It gives off diffuse branches to the tube-feet and connects to an oral hemal ring. The hemal tissue is composed of connec­ tive tissue channels completely or partially obscured by coelomocytes and a viscous fluid. (Unger's 1962 diagrammatic representation of the hemal channels of Marthasterias glacialis as tubular vessels is exceedingly misleading.) The hemal ring attaches to an axial organ (less weIl developed than that of echinoids and isolated from the perivisceral coelom). On the aboral surfaces this organ connects with one or more aboral rings with genital branches, a 'head process' within a contractile 'dorsal sac' and usually a pair of'gastric hemal tuffs' to the pyloric stornach (fig.4). Cuenot (1901,1948), whose work was extraordinarily detailed and meticulous, described hema1 extensions into. the mesenteries supporting the pyloric caeca and from the tube-feet complex into the body wall. Most other ob servers (including myself) have had great difficulty seeing these, at least on a consistent basis. No hema1 struc­ tures are evident in the electron micrographs of the pyloric caeca of Asterias rubens by Bargmann & Behrens (1968) (although they may not have been looked for). Certainly, any direct hemal connection between the pyloric caeca and the other body regions is at best very tenuous.

Nutrient translocation 379

Figure 4. Asteroid hemal system. Arrows show interpreted direction of net movcment of hemal sub­

stance.

1 - aboral hemal ring; 2 - head process; 3 - genital hemal sinus; 4 - axial organ; 5 - transverse hemal

extension; 6 - radial hemal sinus; 7 - gastric hemal tuffs; 8 - oral hemal ring; 9 - tube-foot hemal

sinus.

Tangapregassom & Delavault (1967), working with Asterina gibbosa and Echinaster sepositus, and Walker (197 4a, 1979), working with Asterias vulgaris, Ctenodiscus crispatus, and Hippasteria phrygiana, have described the ultrastructure of an elaborate hemal sinus in the gonads. This is located between the germinal epithelium and a perihemal coelomic sinus. It is filled with what appears to be mucopolysaccharide containing fluid, which is more abund~t when gametes are maturing and after they have been shed. Connections of the geni­ tal hemal sinus can be traced along the gonoducts to a common aboral hemal ring. While the axial organ of asteroids is not as well developed as that of echinoids, it is still a distinctive structure. Its ultrastructure has been described by Bargmann & von Hehn

380 lohn CFerguson (1968) and Leclerc (1974). The axial organ is a spongy body covered with a network of coe­ lornic epithelial elements, lying on a collagen-containing basement membrane, and muscle cells. Both cell types contain a variety of inclusions and have cilia extending into the sur­ rounding axial (coelomic) sinus. The hemal channels are unlined and are crowded with large free coelomocytes rich in inclusion bodies. There mayaiso be some glandular cells and perhaps migrating germinal cells. The head process of the axial organ has a similar structure, with the addition of nerve cells containing electron-dense granules, bounded by distinct membranes. Both the axial organ and the head process can undergo pulsations. Ferguson (1966) studied cell production in the hemal and other tissues of Asterias for­ besi using tritiated thymidine and autoradiography. Occasional newly formed fIXed or free cells were detected in the axial organ and other hemal structures, but in a proportion not greater than, and in some cases less than, other regions of the body. An unusually high level of background radioactivity was no ted in the hemal structures, especially the axial organ. The hemal system of ophiuroids is much like that of asteroids, with only minor diffe­ rences. The axial organ was extensively studied by Fedotov (1924, 1926a), in a number of species. It, again, is a fibrous network supporting a variety of cells and a dense fluid. It lies in a divided axial sinus and is itself divided into two parts, a darker aboral portion and a lighter, more slender, oral portion. The latter may be homologous with the asteroid head. process (Smith 1940). Schechter & Lucero (1968) noted in their investigation of Ophioder­ ma panamensis that the perivisceral space is bridged by numerous collagenous trabeculae containing small hemal vessels. The hemal system ofvarious crinoids has been described by Hamann (1889) and Chad­ wiek (1907). In these forms the hemal system is a more diffuse system than seen in some of the other classes. It includes a number of hemal elements embedded within strands of connective tissue that largely fill the interior of the body. The axial organ is somewhat sirni­ lar to that of asteroids, but contains a number of follicle-like tubules (Tuzet & Manier 1960). The ultrastructure of these in Nemaster rubiginosa has been studied by Holland (1970). They are notable in their content ofrough endoplasmic reticulum and their appa­ rent production of a proteinaceous hemal fluid. As in the other classes, a few muscle cells, nerve cells and coelomocytes are present in the fibrous meshwork of the axial organ. Mucus­ like secretions are used in large quantities by crinoids to entrap food (Nichols 1960). Per­ haps the hemal system evolved in these or other primitive echinoderms as a mechanism associated with the production of these substances. 1.1.2. Experimental studies

While many workers have injected methylene blue and other dyes into hemal vessels to

help delineate their extent, there have been several more concerned attempts to use experi­

mental methods to clarify the role of the systems.

Verchowskaja (1931) removed the axial organs from Asterias rubens and found that indi­ viduals could suiVive for at least six months. The axial organs did not regenerate, and the only significant effect appeared to be an enlargement of the Tiedemann bodies on the ring canal of the water vascular system. Implantation of an additional axial organ perhaps in­ creased hardiness of the individuals and produced some enlargement of the recipient's own organ. The axial organs of Psammechinus miliaris were removed by Schinke (1950) with good survival of the individuals for up to 18 days. When Farmanfarmaian (1968) removed the

Nutrient translocation 381 axial organs from Arbacia punctulata, he recorded survival of the individuals for nearly nine months. With care, he achieved 90 % survival of the individuals after 1~ months. There was some tentative indication of partial regeneration of the organ. He also found that exci­ sion of the axial complex of Strongylocentrotus droebachiensis and A.punctulata, or sever­ ing the radial ambulacral structures of S.purpuratus, had !ittle effect on respiratory rates of the individual. Millott & Vevers (1964) reported that removal of the whole oral portion of the hemal system and adjacent parts of the water vascular system from Paracentrotus lividus still allows for survival of individuals for up to five months, although there is degeneration of part ofthe gut. Farmanfarmaian & Phillips (1962) severed the gonadal hemal connections in s.purpura­ tus and fed them C14-labelled algae. In specimens sacrificed 1-7 days after fee ding, the affected gonads and their normal neighbors were all found to have incorporated equal, low amounts of activity. This experiment was performed on sexually ripe individuals, however, when little nutrient transport to the gonads would be expected. Several workers have tried to examine hemal functions by using markers. Cuenot (1901) injected dyes into the perihemal coelomic canals of Asterias rubens and found them con­ centrated in the axial sinus the next day and in the adjacent stone canal subsequently. Oomen (1926b) injected methylene blue into the gut of Holothuria tubulosa and Holothuria stellati and saw it appear in the hemal network. Stott (1955, 1957) orally injected iron saccharate suspensions into Echinus esculentus and Holothuria forskali (it is questionable whether iron saccharate should be considered a nutrient or a foreign substance). No iron was found in the intestinal wall and hemal spaces of E.esculentus, although a great deal was seen in the larger hemal canals, and in the esophageal and pharyngeal regions, appa­ rently contained in granular coelomocytes. The axial organ became very rich in iron, and iron granules were also detected within the radial hemal strands. Little was seen in the gonad. The iron was again found in hemal tissues of H.forskali, held by clumps of coelo­ mocytes or acellular debris. Iron also appeared in abundance in the respiratory tree. Rosati (1970) injected ferritin and iron-dextran into H. tubulosa. The protein and carbohydrate was seen to cross the intestinal epithelium and to be taken up by phagocytes. Millott (1966) injected foreign cells, protozoans, and bacteria into the coelomic fluids of S.droebachiensis and A.punctulata. He found these taken up by the axial organ. They were encased in clot-like cyst structures, presumably wound spirally by continued rotation. These were especially abundant in the peripheral follicles. In some cases the axial organ became very swollen with these structures. Clotted fluid reinjected into the same animals produced sirnilar results. Removal of coelomic fluid, or naturallesions induces proliferation of cells from the peritoneal outer lining of the axial organ into the lumen, and the discharge of cells, secretion, and cell debris from the lacunae into the contractile vessel (Millott 1969). In individuals from a food-impoverished natural population of S.purpuratus, Johnson (1971) found black and brown clumpy or viscous material completely filling the hemal canals and axial organ, largely delineating the system. Similar material was found in the ampullae and tube-feet. It appeared to be composed largely of coelomocytes, necrotic cells, and amorphous granular material. Ferguson (1963a) fed small clams injected with algal protein hydrolysate-C 14, glycine-C 14, glucose-C1 4, or palmitic acid-C 14 to Asterias forbesi. Most of the ingested radioactivity was retained in the pyloric caeca for over a week. The tissue of the hemal septum reacted in a rather erratic manner to the substances, but was frequently found to have taken up fairly

382 lohn CFerguson high levels compared to other tissues. The highest concentration developed soon after feed­ ing, except for palmitic acid, where the maximum level was reached after three days. In order to get a more complete picture oftranslocation, Ferguson (1970) mimiced the release of dissolved nutrients from pyloric caeca into the coelom by injecting a C14-labelled amino acid mixture into Echinaster echinophorus (= E.modestus). All tissues lining the coelom, especially those of the pyloric caeca, rapidly incorporated the tracer. Within 24 hours some of the tracer could be seen heavily darkening portions of the radial hemal canal. This darkening could further be traced to the connective tissue layer of the tube-feet and into the terminal sucker (fig.5). The hemal incorporation did not occur centrifugally from the disc, but rather, sporadically along the arms. Where incorporation took place, heavily labelIed transverse branches of the radial hemal tissue could be traced in lateral extensions of the perihemal coelomic canals, around the tube-feet bases, to areas of elose juxtaposition with the perivisceral coelom. Many labelIed and unlabelled coelomocytes were also found in these regions.

1.1.3. Interpretation 01 hemal function In light of the foregoing summary of our knowledge of the hemal system, what conelusions may be made? First, it is evident that the system is ubiquitous in echinoderms. While a number of dis­ tinct differences exist in its form between elasses, and species within elasses, there is also a great deal of commonality. Second, it appears to be an active system serving a presumably useful function in the animals. It is not vestigial - there is no known architype of which it is the vestige. It has a high development in a number of very specialized species of the phy­ lum. Third, it is probable that the system may not have a single function, but several, as is the case for most organ systems throughout the animal kingdom. These could inelude in addition to nutrient translocation, neurosecretory transport, germinal cell transport, respi­ ration, microbial defense, antigen production, and excretion. Fourth, it is elear that no really comparable organ system exists among the more familiar animal forms, and hence, we are at a loss for a more appropriate term than 'hemal' to describe it. It is a unique sys­ tem, and it must be understood for itself, and not by comparison to others. And fifth, it is most obvious that far too little experimental work has been done to elarify its functions. Can the developing gonads accumulate nutrients if their hemal connection is severed? Can marker or tracer studies delineate the direction, velocity , and net transport (if any) of the substances contained in the vessels? Can tracer studies show in what cells these substances are synthesized? Can the uptake of these substances into developing cells be demonstrated? Or can it be shown that they are removed from the body as components of mucus and other types of secretions? It should be possible to answer experimentally these and other questions. These experiments must be completed before a definitive statement can be made on hemal system functions. In the meantime, the following highly speculative judgement is offered on the role of the system in nutrient translocation (other functions are beyond the scope of this treatise). It seems significant that the hemal system is often richly developed where there appears to be high nutritive demands, especially for mucoid-type substances. Particularly included are the ambulacral areas, which produce large amounts of secretion for feeding and tube­ foot attachment (Srnith 1937, Buchanan 1962, Chaet & Philpott 1964), and the gonads, where the mucoid substances are found in the nutritive reserves built up for the gametes (Chia 1968, Walker 1974, 1979, Nimitz 1976). The unique development of the gut-asso­

Nutrient translocation 383 ciated hemal systems of holothuroids and echinoids correlates with the production of large amounts of mucus and other secretions to encase coarse ingested material and protect the lining cells (see Holland & Nimitz 1964, Holland & Ghiselin 1970, Massin & Jangoux 1976). In some cases these elaborate systems may be dominated by a respiratory function (Her­ reid et al. 1976). They also appear to serve as a major conduit for coelomocytes in forms with a spacious perivisceral coelom. Thus, a major function of the hemal system appears to be to contribute forms of elabo­ rated nutritive substances to these regions. These may inc1ude, especially, mucopolysaccha­ rides, glycoprotein, and perhaps, glycolipid. The actual secretion may be from mucoid cells which consolidate or further modify the material or release it directly. The hemal system of most species, however, c1early does not have the structure of a conduit from the gut to the other tissues, or the motive facility to produce circulation. The mucoid material (or its precursors) seems to be extracted from the circulating perivisceral and perihemal coelomic fluids, or they are brought to the hemal tissue by degenerating phagocytic coelomocytes and other necrotic cells completing their growth cycle. In any case, there is little centrifugal flow in the hemal channels, but rather, the mucoid precursors seem to be produced or collected in the hemal tissue in the same general region in which they are utilized. If the hemal substance is ultimately eliminated from the body as mucoid-type secretion from the tube-feet, gut, respiratory trees, or other parts, the hemal system can also ulti­ mately serve as an excretory mechanism. For the material being passed to the gonad, how­ ever, a special problem exists. The axial organ appears to be ideally situated to collect and purify the hemal substances, destined especially for the gonads. There can be no question that this organ readily collects debris, some of which may pass out of the body via its con­ nection with the stone canal. Evidence for prolonged survival without this organ is signifi­ cant. The axial organ is not a heart. The observed contractions may assist in keeping its channels free, flltering, local respiration, or even expulsion of accumulated waste. The absence of an axial organ in holothuroids may be correlated with the unique form of the gonad in these forms. 1.2. Coelomocytic transport Coelomocytes have long been suspected of having a role in nutrient transport. Cuenot (1887,1891), Chapeaux (1893), and Cohnheim (1901) develop this idea after observing these cells accumulate dyes and fatty materials fed to asteroids. The ubiquitous presence of coelomocytes that take up dyes and other partic1es from the gut contents, and the appa­ rent impermeability of the holothuroid intestine, led Enriques (1902), Oomen (1926b), and Schreiber (1929, 1932c) to further promote the concept. Additional support was provided by Van der Heyde's (1922,1923) demonstration that coelomocytes in vitro could absorb large amounts of glucose from coelomic fluid. Kindred (1924) reported that migrating coe­ lomocytes of Arbacia punctulata take up nutrients from the digestive organs and liberate protein and lipid inc1usions throughout the body. Stott's (1955, 1957) experiments (pre­ viously described) showing coelomocytic uptake of ingested iron saccharate in Echinus esculentus and Holothuria forskali seem to be in line with these earlier conc1usions. The form and diversity of coelomocytes has been reviewed by Boolootian & Giese (1958), Boolootian (1962), Endean (1966), and Jangoux & Vanden Bossche (1975). In

384 lohn CFerguson some species there are a variety of types which may perform different functions, or repre­ sent different stages of development. Clearly, they serve as defensive phagocytes, and as agents of clotting. Undoubtedly, the nutritional role of these cells has been exaggerated. Certainly in many echinoids and holothuroids they enter the gut in great numbers, apparently conveyed across the spacious perivisceral coelom by the well-developed hemal system of these forms. In the gut they are active scavengers, taking up various types of particulate matter from the coarse material ingested by these forms. As much of this is non-nutritive in nature, it seems likely that this uptake is largely a defensive reaction. It may serve mainly to clear the absorptive mucosal epithelium of absorbed particles so that its efficiency in transport will not be impaired. Much of the material taken up by the phagocytes appears to be destined for elimination. The hemal system may again assist in conducting the cells to removal sites, or they may pass directly into the perivisceral coelom. Phagocytes have frequently been reported as leaving the body via the dermal papulae of asteroids, madreporites, echinoid gills, holothu­ roid respiratory trees, as weIl as in the feces of these forms (see Hyman 1955, Endean 1966). Perhaps the inert particles are elirninated, while phagocytized substances of nutri­ tive value (including debris from degenerate ceIls) are salvaged by the axial organ or other elements of the hemal system. There is experimental evidence that the nutritive transport role of coelomocytes must be largely limited to such scavenging, or the movement of preformed, elaborated elements. Following feeding Cl4-labelled algae to Strongylocentrotus purpuratus, Farmanfarmaian & Phillips (1962) found an initial peak in plasma radioactivity compared to the coelomo­ cytes. By the next day this peak had markedly declined and the coelomocytes possessed only slightly more label than the low levels of the cell-free plasma. Boolootian & Lasker (1964) observed slightly above average increases in the specific activity of red coelomocytes in this same species after the animals were fed C l4-labelled algae. When radioactive red coelomocytes were injected into fresh animals, 16 % of the label was subsequently found in the gut, 5 % in the plasma, and 1 % in the gonads. While these workers interpreted the experiments as supporting the hypothesis of coelomocytic transport, this conclusion does not seem justified. In somewhat similar fee ding experiments on the asteroid, Asterias forbesi, involving C14_ labelIed amino acids, glucose, and palmitic acid, Ferguson (1964a) also found that it took almost a day before radioactivity in coelomocytes surpassed that of cell-free fluid. The coe­ lomocytes took up amino acids to a greater degree than the other substances. Even so, the level of activity in these cells only increased with the level of other tissues found adjacent to the coelomic cavity, and after 48 hours, all the activity levels seemed to remain constant. In both these studies additional evidence (to be cited later) implicates the cell-free fluid (plasma) as the primary vehicle of transport. Thus, there is very little concrete evidence that coelomocytes are significant in trans­ locating nutrients from absorptive and storage sites in the gut to other areas of utilization. Their relationship with the hemal system does, however, indicate that they may be impor­ tant in the production of the hemal fluid, and in that sense they may have a local nutritive role. They are likely involved with protecting the mucosallining, and 'cleaning up' after normal cell replacement, and minor injury, but they probably acquire most of their own nutrients from the plasma portion of the coelomic fluid.

Figure 5. A, B, D. Autoradiographs of nearly adjacent sections of the ambulacrum of an asteroid (Echi­

naster sp.) in which a C l 4.labelled amino acid mixture has been injected into the periviseeral coelom 24

hours earlier. Note the rather specific retention of the label in the hemal tissue and associated connec­ live tissue of the tube-feet. Unlabelled amoeboeytes are located in the lumen of the left tube-foot and radial water eanal. C. Stained section of a tube-foot to show eonnective tissue layers (et) (from Ferguson 1970).

Figure 6. Oliary transport in arm of an asteroid (after Budington 1942). 1 - mesentery ; 2 - pyloric caecum ; 3 - ampulla; 4 - tube·foot.

Nutrient translocation 385 1.3. Pe1:ivisceral coelomic transport As a conspicuous, fluid-filled body cavity containing the gut, gonads, ampullae, and other structures, the perivisceral coelom must be considered a major candidate for being the prime system of nutrient translocation. Doubts of such a function, however, developed with the earlier workers, who, as has been indicated, championed the hemal system or coe­ lomocytes for this role. The perivisceral cavity is found in all forms, but is partially obliterated in crinoids and ophiuroids, and incompletely divided by mesenteries in other forms. It is gene rally lined with a distinct cuboidal (sometimes columnar) peritoneum, underlain by nerve, muscle, and loose collagenous connective tissue. Ultra-structure studies indicate that the lining epithelium is composed of choanocyte-like cells possessing a central flagellum surrounded by a number of microvilli (Bargmann & von Hehn 1968, Schechter & Lucero 1968, Walker 1974a, 1979, Jensen 1975). These cells may be phagocytic and capable of desquamating into coelomocytes (Vanden Bossche & Jangoux 1976). Gemmill (1915), Irving (1924), and Budington (1942) have carefully mapped out the ciliary currents of the perivisceral coelom of asteroids, and shown that these animals main­ tain the coelomic fluid in .an efficient directed circulation (fig.6). Yazaki (1930) has demon­ strated a similar circulation in Caudina chilensis, Austin (1966) in Ophiothrbc spiculata

Table 1. Coelomic fluid metabolites (mg %) Species

HOLOTHUROIDEA

Holothuria tubulosa Id. Stichopus japonicus Parastichopus californicus

ECHINOIDEA

Allocentrotus fragilis Paracentrotus lividus Sphaerechinus granularis Strongylocentrotus franciscanus Id. Strongylocentrotus purpuratus Id. Id. Id. Id.

ASTE ROI DEA

Patiria miniata Asterias forbesi Asterias rubens Pisaster brevispinus Pisaster giganteus Pisaster ochraceus Id. Id. Id. Pycnopodia helianthoides

NPN

Amino-N

2.5 1.1 2.8

0.4

2.6 3.75 5.7 8.6 2.04-6.31 traces 0.8-8.2 5.43 6.1 3.74 3.0-4.8 1.95 1.5 4.4

2

3.04-6.2 2-6 0.8-4

2.4

2.71

0.8 1.67

Reducing sugar

1.2-6.4 0.2 1.54 6.1 0.8-5.3 0 0-6.2 1.0-4.7 0.85-1.34 0.2 0.69 traces 1-5 0.25-0.29 29

References

Cohnheim 1901 Delaunay 1931 TanaJca 1958 Giese 1966a Giese 1966a Delaunay 1931 Cohnheim 1901 Myers 1920 Bennett &. Giese 1955 Hilts &. Giese 1949 Giordano et aI. 1950 LasJcer &. Giese 1954 Bennett &. Giese 1955 Giese 1966a Giese 1966a Ferguson 1964a Delaunay 1931 Giordano et al. 1950 Greenfield et al. 1958 Myers 1920 Greenfield et al. 1958 Giese 1966a Vasu &. Giese 1966 Myers 1920

386 lohn c.Perguson and Ophiura luetkeni, and Chesher (1969a) inMeoma ventricosa. Thus, in functional terms, the coelom is the dosest echinoderm analogue to the circulatory systems of other animals. Doubtless, the major benefit of this circulation is in respiration, but it also would assist in nutrient translocation. The coelomic fluid has an inorganic composition very similar to that of sea water, and the animals are very poikilosmotic (see reviews of Binyon 1966, 1972a). The cell-free fluid contains free amino acids, especially glycine and taurine (Giordano et al. 1950, Lasker & Giese 1954, Jeuniaux et al. 1962, Awapara 1962, Ferguson 1975a). These have been com­ !TIonly reported as a component of 'non-protein nitrogen' (table 1). It also contains some glucose and other reducing sugars (table 1), as weIl as low levels of proteins and probably complexes of protein with carbohydrate and lipid (Messina 1957, Giese 1966, Vasu & Giese 1966a, Holland 1964. Holland et al. 1967). These may vary with the state of the repro­ ductive cyde and nutrition (Holland 1964). Certainly the gut, or its appendages, is the major initial store of absorbed nutrients in most echinoderms, although gonads, body wall, and other tissue can subsequently contri­ bute reserves in starvation (see review of Ferguson 1969a). Digestion of food, then; would not be expected to lead to rapid translocation outside the digestive system as would be the case in vertebrates. Misunderstanding this and failing to observe transport, earlier workers conduded that the intestine was impermeable and that coelomocytes had to translocate the absorbed food materials. In contrast, D' Agostino & Farmanfarmaian (1960) seemed to expect rapid transmural translocation. They interpreted data from experimental work on Leptosynapta inhaerens to show active glucose movement across the intestine. Farmanfarmaian (1963) and Rundies & Farmanfarmaian (1964) also interpreted studies on Strongylocentrotus purpuratus, L. inhaerens, Holothuria tubulosa, and Sc/erodactyla briareus as showing active absorption into the coelomic fluid. Significant transmural transport at reasonable substrate concentra­ tions, however, was not achieved. Their data, nevertheless, are compatible with the trans­ location model that will be developed shortly, as are the data of Lawrence et al. (1967) illustrating uptake at normal substrate levels of amino acids and sugars into both surfaces of the gut of Stich opus parvimensis. Attempts by Farmanfarmaian & Phillips (1962) to demonstrate meaningful transloca­ tion of ingested nutrients by S.purpuratus have met with somewhat more success. They fed C14-labelled algae to this echinoid and observed a dearly detectable peak, mainly due to galactose, in cell-free coelomic fluid radioactivity within six hours of the start of the feed­ ing period (which apparently did not last more than five hours). After the peak there was a decline to a constant level that lasted at least several days. Boolootian & Lasker (1964) per­ formed a sirnilar study on the same species, but fed their specimens a different species of algae. They also found a modest initial peak of activity in the coelomic fluid a few hours after feeding, but noted that it was composed mainly of manitol (which was a major com­ ponent of their algae). Most of the ingested activity was retained in the gut, and only appeared slowly in coelomocytes and other tissues. The results of these two studies unequi­ vocally show that ingested nutrients are reaching the coelomic compartment, where pos­ sibly they could be translocated to other body regions. It may be significant, however, that in both cases the observed peaks were largely due to compounds found in the food in high concentrations, but not representing usual metabolites of the animals. The peaks could be due to diffusion and a somewhat sluggish ability of the digestive tissues to metabolize these substances.

Nutrient translocation 387 Undoubtedly, however, echinoids can readily remove nutrients from their coelomic fluids. This was shown when Lasker & Giese (1954) injected relatively small quantities of glucose into S.purpuratus, and measured rapid diminishment of fluid levels of this sub· stance. Active uptake of ingested amino acid by an asteroid has been demonstrated by Ferguson (I 979). Echinaster modestus was permitted to take up a liquid amino acid mixture. These compounds were found concentrated in the pyloric caeca at 20·50 times the initial medium level. The elevated tissue levels dirninished only slowly, and were still significantly elevated after 48 hours. In an attempt to follow the route of translocation of ingested material, Ferguson (1964a) fed small clams injected with C14·labelled amino acids, glucose, and palmitic acid to Asterias forbesi. The distribution of the tracer in the tissues was followed with counting methods. Most of the ingested nutrients were found to be stored in the pyloric caeca. Some was appa· rently released to the coelornic fluid, where low and relatively constant levels of the tracers were maintained for at least a week. Generally , only a very small initial peak could be detected in the fluid with the first 24 hours. Some progressive transport to the body wall was also demonstrated. In autoradiographs of similarly treated animals, the peritoneum and other tissues in the vicinity of the coelom were found to have become radioactive from the ingested tracers (Ferguson 1963a). To further clarify the translocative process, micromolar concentrations of labelled arnino acids and glucose were injected directly into the coelom of A.[orbesi (Ferguson 1964a). (In 1922 Van der Heyde had injected enormous amounts (75·700 mg) of glucose into A.[orbesi and reported its rapid disappearance.) Most of the injected nutrients were removed by the tissues in the arms injected, although some were translocated to all regions of the animals. Interestingly, over half was taken up by the pyloric caeca. If translocation involved simple mucosal to serosal transmural transport, this should not have occurred. These results were confirmed in a later autoradiographic study on E.echinophorus (= E. modestus) (Ferguson 1970). Injected amino acids were taken up by the digestive glands and all the tissues lining the coelornic cavity, as weIl as certain other areas. In vitro studies have provided additional information. Ligatured pyloric caeca, gonads, rectal (intestinal) caeca, and portions of cardiac stornach of A.[orbesi were removed and incubated with labelled nutrients (Ferguson 1964b). In every case, an apparently active transport through the serosal coelomic surface could be demonstrated. The rate of uptake was related to media concentration, but subject to competitive effects. (The nature of the cellular transport mechanism has been more fully described by Ferguson 1968a). While this observed uptake by gonads and other tissues seems to confirm the coelomic translocation hypo thesis, the serosal uptake by pyloric caeca would appear to be in the wrong direction. As already outlined, previous workers had sought with little or no success to demonstrate significant mucosal to serosal translocation in echinoderms (although muco· sal uptake was readily demonstrable). This paradox is resolved if one considers that all sur· faces of the cells (or organs) not only actively take up the nutrients, but also release them. Thus, we are dealing with, on each cell surface, two rates - a flux in and a flux out. Tracer methods generally only show one of these rates. In the experiments just described, what was presumably measured was the serosal flux in, which was matched by an unseen (for lack oflabel), comparable flux out. The data available (Ferguson 1964b, 1968a, 1971) seem to indicate that both rates for a given membrane are concentration dependent. That is, flux out increases as cytoplasrnic pool concentrations increase, and flux in increases as

388 lohn C.Ferguson Table 2. Non-protein nitrogen in digestive tissues (mg/g dry tissue) Species

CRINOIDEA

F1orometra perplexa

HOLOTHUROIDEA

Parastichopus caIifornicus

ECHINOIDEA

Stomopneustes variolaris Allocentrotus fragiIis Strongylocentrotus franciscanus Strongylocentrotus purpuratus Dendraster excentricus

ASTEROIDEA

Patiria minÜlta Asterias rubens Pisaster brevispinus Pisaster giganteus Id. Pisaster ochraceus Id.

OPHIUROIDEA

Ophioderma panamensis

NPN

References

12.1

Giese 1966a

34.5 ± 9.1

Giese 1966a

65.0 ±4.5 19.0 ± 8.0 42.1 ± 7.0 41.9 ± 5.3 14.6 - 20.4

Giese Giese Giese Giese Giese

26.3 ± 17 4.9 (wet) 24.5 ± l3 10-19 14.5-14.9 9-27 12.1-33.6

Giese 1966a Jeuniaux et al. 1962 Giese 1966a Greenfield et al. 1958 Giese 1966a Greenfield et al. 1958 Giese 1966a

34.7

Giese 1966a

1966a 1966a 1966a 1966a 1966a

extracellular pool (coelomic) concentrations increase. The net effect, under any given con­ dition, is to achieve approximately steady state equilibrium concentrations in both the cellular and extracellular compartments. It is the magnitude of these concentrations that is indicated in tables 1 and 2. The normal coelomic fluid nutrient levels are very low and the tissue levels are gene rally high. While measurement of flux into the cellular compartment is rather easily accomplished using radiotracer methods, accurate determination of flux out is considerably more difficult. Ferguson (1964b) attempted to provide such data by placing ligatured pyloric caeca from A.forbesi in small vessels of sea water and chemically measuring the appearance of released nitrogenous compounds. The results closely fitted the mathematical model of the system, providing an equilibrium level in the extern al compartment (38 Ilg N per ml fluid) compara­ ble to that found in natural coelomic fluid. Release rates from the pyloric caeca appeared to be more than adequate to meet the needs of the animals. Recent preliminary studies with E.modestus, employing much more sensitive fluorometric methods and highly purified media, appear to confirm this model (Ferguson unpublished). These indicate a gross release rate from the pyloric caeca of about 0.75 mM amino acid per g tissue (wet weight) per day. The experiments of Farmanfarmaian (1969b) in which transmural translocation was only observed at high substrate concentrations may now be interpreted in the same manner. When gut of S.briareus was incubated with glucose-C1 4 at 0.5 mM concentration, this was transported as a net flux into the gut tissue and became incorporated into the probably rather large pool of amino acids maintained there. Only a small amount of the labelIed sugar passed out into the surrounding fluid because, after dilution in the cellular pool, it represented only a small portion of the total flux out into the coelomic compartment. The

Nutrient translocation 389 glucose that did move out was partially reabsorbed back into the gut (as well as other tis­ sues) by the serosal flux in until a steady state was established. When much higher concen· trations of glucose were incubated, the same process led to a temporary elevation of the intracellular pool promoting serosal flux out. As the intracellular pool diminished by meta­ bolic activity within the cells of the flux out returned to normal levels and the serosal flux in, elevated by the accumulation of glucose in the medium, began to exceed flux out, leading to a peak in the observed concentration. The same reasoning would explain the results of Farmanfarmaian & Phillips (1962) on S.purpuratus. Apparently there have been no comparable experimental studies to assess the role of the perivisceral fluid in ophiuroids and crinoids. In summary then, the perivisceral coelom represents a fluid compartment in intimate contact with the digestive system, the gonads, and many other vital body parts. The con­ tained fluid is maintained in efficient circulation by cilia. Nutrients (amino acids and glu­ cose) enter the fluid (plasma) from digestive and storage structures, and are efficiently removed from the fluid by exposed tissues. This is accomplished by what appears to be a ubiquitous cell process of uptake and release. The perivisceral coelom, then, must be con­ sidered a major avenue of translocation for these substances. There is little evidence as yet, however, that it translocates or does not translocate elaborated substances, such as protein, glycoprotein or glycolipid.

1.4. The perihemal coelomic system The perihemal coelomic system is a name applied to a variety of usually paired coelomic spaces found in the ambulacral area, around the mouth, around the anus, and in the wall of the gonads. While parts of it may be delineated by injection of dyes, much of it must be traced in histological sections. As the name implies, it usually encases the hemal structures, ineluding the axial organ. It also lies adjacent to many parts of the water vascular system and the nervous system. In asteroids, at least, extensions of the perihemal spaces may be traced into the connective tissue of the body wall where they appear to join with unlined lacunae. Most of the major museles of the body wall are located in conjunction with either the perihemal channels or the lacunae. There are a number of places of elose conjunction between the perihemal system and the perivisceral coelom. While undoubtedly there are some direct communications between these compartments, they for the most part are sepa­ rated by membranes. Like the perivisceral coelom, the perivisceral system is lined with a gene rally cuboidal peritoneum containing flagellated cells with microvilli (Walker 1974a, 1979). These presumably keep the fluid in active movement and local circulation. Generally, this system has received Httle experimental attention. Ferguson's (1970) autoradiographic study on Echinaster sp. did illustrate its probable involvement in nutrient translocation. Following injection of C14-labelled nutrients into the perivisceral coelom, highly radioactive portions of the elose junctions between the perivisceral and perihemal peritoneum could be seen. It was in approximately these same locations that hemal ele­ ments, contained within the perihemal space, became radioactive, and these could be traced to the connective tissue layer of the tube-feet (fig.S). By inference from the foregoing discussion of the function of the perivisceral coelom, it is probable that the perihemal spaces act as extensions of the perivisceral coelom in sup­ porting translocation of nutrients to those critical areas with which they associate. The extensive relationship between the perihemal spaces and the nervous and hemal systems

390 John C.Ferguson may be indicative of some unique nutrition al functions. Clearly, a great deal more experi­ mental work is needed to elucidate these. 1.5. The water vascular system As the water vascular system is weIl known, and as it is obviously dominated by other func­ tions, only a few things need to be mentioned of its role in nutrient translocation. The system is, again, a lined coelomic compartment which, especially in the ampuIlae, has a elose junction with the perivisceral coelom. The system is also elosely associated with the peri­ hemal spaces. Extensions of the hemal system join with the connective tissue layer of the tube-feet, where they appear to contribute the mucoid substance (or precursors) released in the terminal sucker (fig.5) (Ferguson 1970). There also is seemingly a connection between the madreporite vesiele and the axial sinus (Binyon 1964). Except in holothuroids, the fluid of the water vascular system is in communication with the exterior via the madreporite. Nevertheless, there appears to be little flow through the vessels, although local circulation probably exists and is important in respiration in some forms. The fluid maintains a slightly elevated concentration of potassium ion and also con­ . tains a small amount of protein (Binyon 1964), but has no properties indicative that it is an important translocative medium. The Tiedemann bodies found in some asteroids are defmitely not cytopoietic organs, but appear to function to purify the fluids. They have a number of similarities to axial organs (Tuzet & Manier 1960, Ferguson 1966). Certainly, the tube-feet-ampullae complex, with its continual activity and secretion, must have one of the highest metabolic demands of any tissue in the animals. As seen, some of this need, especially for the mucoid secretion, may possibly be met by the contri­ bution of elaborated nutritive elements from the hemal system. Presumably, much could also come through the same translocative process described for the perivisceral coelom as weIl as direct uptake from the environment. Translocation would have to occur across the ampullae. Considering their extreme thinness (Nichols 1959c, Kawaguti 1965b) and affffiity for labelled nutrients (Ferguson 1970), this is entirely reasonable. 1.6. Integumental translocation There are in the bodies of echinoderms a number of tissues which are not located in pro­ ximity to any of the compartments or structures that have been discussed as possibly having a role in nutrient translocation. Principal among these are the osteocytes in the ossieles, spines, and pedicellaria, the large exocrine glands which occur in some species, and the epidermis. These are active tissues. They produce a number of secretions, and when damaged, they can readily regenerate. This is true even for those parts attached to the body by the most tenuous of connections (Ebert 1967a, Chia 1969a, 1970a). One route through which nutrients might reach these locations is by diffusion through the dermal connective tissue. This is a loose collagenous tissue containing a mucopolysaccha­ ride ground substance (Ferguson 1960, Kawaguti 1966, Menton & Eisen 1970). Open spaces (lacunae) may exist in it, which possibly could assist in translocation, although there is no evidence of this. Rather, as was shown in the autoradiographic study on Echinaster . sp. by Ferguson (1970), nutritive substances seem to move from the vicinity of the somatic peritoneum outward through the connective tissue. Additional data by Ferguson (unpub­ lished) suggest that the somatic peritoneum is very porous, facilitating exchanges between the coelomic fluid and the connective tissue ground substance.

Nutrient translocation 391 Ferguson's (1970) autoradiographic study also showed that the chains of osteocytes within the ossicles readily became radioactive with the injected tracers, which presumably were passed on from cell to cello Large exocrine glands located in the outer regions of the dermis also became quite radioactive indicating that they had no difficulty obtaining nutrients through the dermal connective tissue. The same may be said for the cells found in abundance beneath the epidermis, but the epidermis itself, however, remained sub­ stantially free of the label. Thus, the fluidity of the connective tissue ground substance, together with cell to cell transport by chains of osteocytes, seems adequate to translocate nutrients throughout the dermis, but another mechanism must normally sustain the epidermis. This may be direct uptake of nutrients from the environment. First demonstrated by Stephens & Schinske (1961), and shown to be significant to echinoderm epidermal nutrition by Ferguson (1963b, 1967b, 1968b), the process has been weIl documented in a variety of echinoderms by a num­ ber of workers (see Bamford, chapter 14). It does not seem to be an essential function in asteroids (Ferguson, unpublished). At greatly reduced sea water levels of dissolved nutrients, the epidermis can still be maintained, probably by translocation through the dermis. The fact that a number of echinoderms can illustrate negative growth with starvation (Ebert 1967a, Jangoux & Van Impe 1977; see also Lawrence & Lane, chapter 15) strongly in­ dicates that body maintenance is a dynamic function. Chan&.es in the size of organs with the annual cycle also appears to be largely due to variations in cell numbers (Lawrence 1973, Fergu­ son 197 Sb). Cells and their constituents are constantly being resorbed and replaced, and tissues exist in a kind of steady state balance between these process at any given time. This dyna­ rnic nature makes the presence of efficient translocative mechanisms for both the supply and recovery of nutrients all the more essential. A very large quantity of material must be translocated just to sustain a constant state, and only marginal incremental amounts pro­ vide for growth, and seasonal fluctuations.

2. CONCLUSIONS: A MODEL FOR NUTRIENT TRANSLOCA TION The foregoing review has attempted to elucidate the known facts of nutrient translocation in echinoderms. Obviously, there are many gaps in the information, and some of the obser­ vations may fail the test of confirrnation. Nevertheless, there now appears to be sufficient knowledge to develop a theoretical model of the total process. The value of such a model is not only to formalize and clarify our thinking on these matters, but also, if it is success­ ful, to point to various areas which are examinable by observation and experimentation. Thus, it is hoped that the presentation of such a model will stimulate the acquisition of new data by which it may be refined and corrected. The form of the model presented will be an asteroid, as it is with this group that the most pertinent data have already been acquired. With slight modification, however, it should serve for most echinoderms.

2.1. The model The model conceives of the individual in dynamic terms. All substances, ceIls, tissues, and organs are, over the short term, in astate of balance between gain and loss, synthesis and degradation, uptake and release. Furthermore, the various compartments are maintained in a weIl-mixed condition by cilia, muscular movement, or other forces. As the individuals

392 lohn C.Ferguson go through their annual cycles this balance may shift slightly to provide for growth or reduction of various parts. Ingested nutrients are seen as being actively taken up and concentrated in the gut tissue. They are also released back into the lumenal spaces, both as specific secretions and passive release. The balance, however, is greatly in favor of the gut tissue, where the soluble nu tri­ ents are initially accumulated in cytoplasmic pools. Within these pools they are in exchange relationships with many intracellular 'compartments' or changes in state. These would in­ clude the use of nutrients for the synthetic and metabolie needs of the cells themselves, and the exocrine products they release. Some nutrients may re-enter the cytoplasmic pool through catabolic reactions.

Figure 7. Model of transport of soluble nutrients in the arm of an asteroid. AIrows show comparative fluxes (not to scale) between tissues and fluid compartments.

Nutrient translocation 393 The nutrients are also in an exchange relationship, through the serosal membrane, with the perivisceral coelomic fluid. They are released into it and actively removed from it. Again, however, the balance of the flux in and the flux out favors a high concentration in the tissue and a low one in the coelomic fluid. Usually a gradual net loss occurs to the coe­ lomic fluid to replace net removal from it by other tissues of the body. The low levels in the fluid are probably desirable in that they reduce the tendency to lose nutrients from the body to the surrounding sea water, and it simplifies osmotic problems. Sirnilarly, the nutrients in the coelomic fluid pool are exchanged by other cells of the body in balance with their cytoplasmic pools. They, in turn, may be partially in balance with other compartments of the body, such as the perihemal system and the water vascular system. Thus, most cells have access to the general pool of dissolved nutrients in the body, and maintain a constant exchange of their own nutrient pools with the general pool. Loss of nutrients to the exterior is further reduced by a basal lamina beneath the epi­ dermis which has limited permeability to organic substances. Under normal circumstances the epidermis, then, obtains most of its nutrients from the ubiquitous pool found in natural sea water. If, however, this is inadequate, it can maintain a minimal subsistence from what can pass through this barrier, or possibly, the barrier under these circumstances may be reduced. The foregoing would apply mainly to the simpler nutrient elements, especially amino acids and sugars. These exchanges are illustrated in figure 7. More complex substances are elaborated either by specific synthesis in the digestive structures, peritoneum, axial organ, or elsewhere, or by the degradation of moribund cells, including phagocytic coelomocytes and other cells released in dynamic replacement. These elements are mostly in the form of a variety of mucopolysaccharides, perhaps with glycoprotein and glycolipid. They are col­ lected by the various parts of the hemal system, possibly with the assistance of coelomo­ cytes. They also seem to be closely related to the general ground substance of the connec­ tive tissue to which the hemal system connects. The axial organ, especially, is a place of collection and purification of these substances, possibly to support gonad development (fig.4). In asteroids there is little facility to move substances directly from the pyloric caeca to the other tissues via the hemal system, but in the holothuroids and echinoids more highly developed channels possibly may perrnit some movement. The elaborated nutritive substances are moved locally through the spongy channels of the hemal system to adjacent places of utilization where they perhaps are collected and secreted by mucoid cells. These areas include the body surface (especially tube-feet) to form protective mucus coverings or food trapping nets, the gonads to supply materials for the developing gametes, and (in some forms) the gut to form protective mucus sleeves to con­ tain coarse food. While much of the data that have contributed to the development of this model have been discussed in the preceding pages, it is clear that many of the other aspects of it are readily testable. For example, methods exist to evaluate most of the indicated fluxes and, thus, develop a mathematical description of the translocation process between the various compartments. Ukewise, the dynamics of local material movement in the hemal channels should be measurable with tracer or stop-flow procedures. It is hoped that the application of such quantitative approaches will not only soon provide answers to the major uncertain­ ties of nutrient translocation in echinoderms, but will also lead to better understanding of ,their cyclical reproduction, starvation responses and trophic interactions. 1

17

W. ROSS ELLINGTON

INTERMEDIARY METABOLISM

Atkinson (1977) divided intermediary metabolism in heterotrophs into three interdepen­ dent processes (fig.l): catabolism via glycolysis, the Krebs cycle and hexose monophos­ phate shunt yielding ATP* ,NADPH and small molecular weight precursors for biosynthesis; biosynthesis (anabolism) producing the major building blocks (amino acids, nucleotides, fatty acids) of complex macromolecules; and growth using building blocks synthesized or absorbed. All three components are tightly linked. For instance, protein synthesis is required to replace catabolic enzymes lost in turnover while energy generated in catabolism provides the driving force for all biosynthetic processes. Knowledge of intermediary metabolism in echinoderms is limited especially in terms of metabolism in adult organisms. Considerable work has been done with gametes and embryos, due to the utility of these systems in developmental physiology. Considerable amounts of work has been done since earlier reviews on carbohydrate metabolism (Doezema 1969) and lipid metabolism (Fagerlund 1969) in echinoderms. There are recent reviews on metabo­ lism in echinoid eggs and embryos (Isono & Isono 1975, Y~nagisawa 1975a,b). This chapter is designed to review intermediary metabolism in echinoderms in the context of both adult and embryonic systems. Catabolism and biosynthesis of small molecules will be reviewed in detail. As very little information exists concerning macromolecular biosynthesis in adult echinoderms and as recent reviews on this subject dealing with embryonic systems have appeared (Guidice 1973, Stearns 1974), discussion of macromolecular biosynthesis (growth) will be lirnited to those points which relate directly to catabolism and biosynthesis. The overall goal of this review is to outline the major features of intermediary metabolism in echinoderms, pointing out areas which are in need of further research.

1. CATABOLlSM 1.1. Carbohydrate degradation 1.1.1. Glycogen breakdown

Glycogen or 'animal starch' consists of a branched chain of D-glucose residues which are linked by a-l , 4-glycosidic and a-l, 6-glycosidic bonds. Glycogen from the gut of the echi­ *Abbreviations: ATP- adenosine-5'-triphosphate; ADP - adenosine-5'-diphosphate; AMP - adenosine-5'­ monophosphate; NAD'" - nicotinamide adenine dinucleotide; NADH - reduced NAD+; NADP+ - nicotina­ mide adenine dinucleotide phosphate; NADPH - reduced NADP+; THF - tetrahydrofolic acid; cAMP ­ 3,5-cyclic adenosine monophosphate; UDPG - uridine diphosphate glucose; UMP - uridine-5'-mono­ phosphate; CMP - cytidine-5'-monophosphate; Pi - inorganic phosphate; IDP - inosine-5'-diphosphate; ITP - inosine-5'-triphosphate.

395

396 W.Ross Ellington

Figure 1. Sehematie block diagram of metabolism of a heterotrophie aerobic cell (from Atkinson 1977).

noid, Strongylocentrotus purpuratus, is similar in structure to glycogens found from other animals (Doezema 1969). This compound is widely distributed in echinoderms and is pre­ sent in especially high levels in reproductive organs where it is thought to be used in syn­ thesis of components of gametes and as a nutrient reserve (Doezema 1969). Glycogen break­ down is catalyzed bythe enzyme phosphorylase in the following reaction: glycogen(h-glueose) + Pi""* glucose-I-phosphate + glycogen(n - 1 glucose) In view of the ubiquitous distribution of glycogen in echinoderms, it is not surprising to find that phosphorylase is present in the tissues of these organisms (table 1). The lack of uniformity of enzyme units in table 1 makes comparison of relative enzyme activity diffi­ cult. Phosphorylase activity in the lantem musc1e of Echinus esculentus was observed to be among the lowest in a survey of enzyme activities in 30 species of marine invertebrates (Zammit & Newsholme 1976). Glycogen phosphorylase generally is found in two catalytic forms (a and b) which vary in terms of sensitivities to allosteric modifiers. This type of sys­ tem appears to be present in the tissues of echinoderms (Doezema 1967, Shoger et al. 1973). Regulation of glycogen breakdown will be discussed in section 3.2.

1.1.2. Glycolytic enzymes

Glycolysis involves the breakdown of hexose sugars and sugar phosphates and the produc­

tion of the three carbon compound, pyruvate. The glycolytic pathway pro duces ATP by

substrate level phosphorylations and yields a number of intermediates used in biosynthesis.

Intermediary metabolism 397 Table 1. Glycogen phosphorylase activity in the tissue of echinoderms Species

Tissue

Activity

References

Strongylocentrotus purpuratus Id. Echinus esculentus Sphaerechinus granularis

lantern retractor museie gut lantem retractor museie lantern retractor muscle fertilized eggs pluteus fertilized eggs fertilized eggs

0.4* 0.3* 1.2** 44t

Doezema 1967 Doezema 1967 Zammit & Newsholme 1976 Bergami et al. 1968 Bergami et al. 1968 Bergami et al. 1968 Bergami et al. 1968 Tazawa et al. 1977

Id.

Id. Paracentrotus lividus Anthocidaris crassispina

* nmoles/min/mg protein; ** Jl.moles/min/g wet weight at 2SoC; tt nmoles/min/mg protein at 20°C

lOt 6t 12t

O.4tt

t nmoles/min/mg nitrogen at 30°C;

The initial step of glucose degradation involves its phosphorylation to glucose-6-phosphate by the enzyme hexokinase. Hexokinase activity has been estimated in tissues of several echinoderms (table 2). Non-quantitative or electrophoretic surveys have demonstrated hexo­ kinase activity in the eggs of Arbacia punctulata and Echinarachnius parma (Krahl et al. 1953) and in the tissues of 11 species of asteroids (Schopf & Murphy 1973, Ayala et al. 1975, Manchenko et al.1977), and in the ophiuroid, Ophiomusium lymani (Ayala & Valen­ tine 1974). Hexokinase in egg and embryo extracts of Arbacia punctulata also phosphory­ lates deoxyglucose, mannose and fructose but the apparent Km for D-glucose is much lower than that of the other substrates (Krahl et al. 1953). In general, it is thought that D-glucose is the physiological substrate for this enzyme. Carbon from glycogen enters the glycolytic pathway in the form of glucose-I-phosphate. This compound must be converted to glucose-6-phosphate to enter the mainstream of gly­ colysis. Phosphoglucomutase accomplishes this conversion. This activity of this enzyme in echinoid tissues and eggs (table 2) is substantially higher than hexokinase. Phosphogluco­ mutase has also been demonstrated in Ophiomusium lymani (Ayala & Valentine 1974) and Nearchaster aciculosus (Ayala et al. 1975). In addition to the above enzymes, electropho­ retic surveys have revealed a number of other glycolY,tic enzymes in echinoderm tissues including phosphoglucose isomerase (Krahl et al. 1953, Schopf & Murphy 1973, Ayala & Valentine 1974, Ayala et al. 1975), triosephosphate isomerase (Schopf & Murphy 1973, Ayala & Valentine 1974, Ayala et al. 1975), aldolase, triosephosphate dehydrogenase and enolase (Ycas 1954). Lactate dehydrogenase performs the physiological function of regenerating oxidized coenzyme (NAD') which is utilized in the triose phosphate dehydrogenase reaction. Lac­ tate formation occurs generally under conditions of reduced oxygen availability when the capacity of mitochondrial oxidations to provide NAD+ is insufficient to support glycolytic flux. Lactate dehydrogenase from the tissues of echinoderms is L-specific, as is the' case with the chordates (Long 1976). In general, the muscle tissues of adult echinoderms have higher activities oflactate dehydrogenase than other tissues (table 2), indicating greater glycolytic potential. This is especially true of the longitudinal muscle of holothuroids such as Sclerodactyla briareus (Ellington 1976a,b, Ellington & Hammen 1977) and Thyone sp. (Zammit & Newsholme 1976). Activities ofthis enzyme in the direction oflactate oxida­ tion are generally much lower, as much as one order of magnitude (Ellington & Hammen 1977), indicating diminished capacity to utilize L-Iactate. This diminished capacity for lac­

398 W.Ross Ellington Table 2. Activities of enzymes of glycolysis as assayed in crude extracts of echinoderrn tissues Enzyme

Species and tissue

Hexokinase (Glucose + ATP -+ Glucose-6-phosphate + ADP)

Activity References Echinus esculentus: lantem l.6 a Zammit & Newsholme retractor muscle 1976 Thyone sp.: pharyngeal retractor O.6 a muscle Strongylocentrotus purpuratus: O.6 b Doezerna 1967 gut l.Ob S.purpuratus: lantem retractor muscle 3-6 b

Phosphoglucomu tase (Glucose-I-phosphate -+ Glucose-6-phosphate)

S.purpuratus: lantem retractor muscle S.purpuratus: gut Anthocidaris crassispina: eggs

Phosphofructokinase (Fructose-6-phosphate ATP -+ Fructose-l.6­ diphosphate + ADP)

Echinus esculentus: lantem O.6 a retractor muscle Thyone sp.: pharyngeal retractor o.sa muscle

+

Lactate dehydrogenase (pyruvate + NADH + If -+ L-Iactate + NAD+)

Sc1erodactyla briareus: longitudinal muscle Thyone sp.: pharyngeal retractor muscle E.esculentus: lantern retractor muscle Mellita quinquiesperjorata: gut Encope abe"ans: gut Arbacia punctulata: gut Lytechinus variegatus: gut Asterias jorbesi: pyloric caeca Astropecten brasiliensis: digestive tract Patiria miniata: digestive tract Cucumaria rubra: digestive tract Lytechinus pictus: digestive tract L.pictus: gonads Luidia clathrata: pyloric caeca

4-8 b S.6 c

Tazawa et al. 1977 Zammit & Newsholme 1976

3.8 d

Ellington 1976a

48.1 a

Zammit & Newsholme 1976

2.S a 3.1 e 3.l e 3.l e O.8 e OAe S09 b

Ellington & Lawrence 1973

Ellington unpublished Scheid & Awapara 1972

18 b 6b 2b 4b O.3 a

Durako et al. 1978

a. p.moles/min/g wet weight at 2SoC; b. nmoles/min/mg protein; c. nmoles/rnin/mg protein at 20°C; d. p.moles/min/g wet weight at ISoC; e. nmoles/min/mg protein at 2SoC

tate oxidation is clearly demonstrated by the high apparent Km for L-Iactate of 35.1 mM, of the lactate dehydrogenase ofthe longitudinal muscle of S.briareus (Ellington 1976a)_ The apparent Km's for pyruvate reduction by lactate dehydrogenase of the longitudinal muscle of S.briareus (Ellington 1976a), of pyloric caeca of Asterias forbesi (Ellington, un­ published) and longitudinal muscle of the holothuroid, Parastichopus parvimensis (Soe, unpublished) were 2.2, 0.6 and 0.4 mM, respectively. These values fall within the upper range of apparent Krn's for pyruvate which have been observed for vertebrate skeletal muscle L-Iactate dehydrogenases (Pesce et al. 1964). All three of these eclllnoderm lactate dehydrogenases were observed to be relatively insensitive to substrate inhibition by pyru­ vate, indicating a lack of regulation.

Intermediary metabolism 399 The general conclusion that can be drawn from glycolytic enzyme profIles of echino­ derm tissues is that the enzymes are, indeed, present in these organisms. Comparisons of relative importance of individual enzymes again are difficult due to the lack of uniformity of enzyme units. The most obvious gap in our knowledge can be found in the absence of information on two of the key regulatory enzymes of glycolysis, phosphofructokinase and pyruvate kinase. Analysis of the concentrations of glycolytic intermediates in the eggs and embryos of Anthocidaris crassispina and Pseudocentrotus depressus clearly demonstrated that these two enzymes represent potential sites of metabolic regulation (Yasumasu et al. 1973). Both pyruvate kinase and phosphofructokinase are present in the longitudinal muscle of Holothuria grisea (Avelar et al. 1978). The activities of phosphofructokinase in muscle of Echinus esculentus and Thyone sp. have been estimated (table 2). Pyruvate kinase from the asteroid, Asterias rubens, exhibits some regulatory properties which may be related to the importance of this enzyme in controlling anaerobic glycolysis in this species (Hoffman 1976, Hoffman & Radeke 1978). Phosphocreatine is a potent inhibitor of pyruvate kinases from the echinoid, Heliocidaris crassispina, and the holothruoid, Pro­ tankyra bidentata (Wu et al. 1978). This inhibition constitutes a regulatory mechanism by which glycolysis can be activated under conditions of high energy demand, as evidenced by the decline in the phosphocreatine pool.

1.1.3. Anaerobic glycolysis Given the presence of glycolytic enzymes as well as high activities of L-Iactate dehydroge­ nases in some of the tissues of echinoderms, it is not surprising that lactate is produced as a result of carbohydrate breakdown. When fertilized eggs of Arbacia punctulata were placed in a KCN solution, there was an increase in lactate content of the eggs (Perlzweig 1928). Lac­ tate production has also been demonstrated in the eggs of the echinoids, Echinus esculentus (Cleland & Rothschild 19 52a,b), Arbacia lixula, Paracentrotus lividus, Psammechinus mili­ aris (Aketa 1964), Hemicentrotus pulche"inus, Heliocidaris crassispina and Pseudocentrotus depressus (Aketa 1957). Anaerobic glycolysis in the longitudinal muscle of holothuroids has been investigated in some detail. The longituqinal muscle of Holothuria nigra contained 0.08 % lactate; this level was doubled in fatigued muscle and accounted for the disappearance of glycogen (Boyland 1928). In vitro preparations of the longitudinal muscle of Stichopus mollis showed muscle contracture when incubated in high potassium sea water and there was a direct relation between lactate content and external potassium concentration (Gay & Simon 1964). Incu­ bation of S. mollis longitudinal muscle in the presence of D-glucose with or without high potassium resulted in substantial amounts of lactate production (Gay & Simon 1964). There was nearly a 20-fold increase in lactate content of the longitudinal muscle (0.52 -+ 9.27 Ilmolesjg wet weight)when specimens of S.briareus were maintained under hypoxic condi­ tions for 24 hours (Ellington & Hammen 1977). In addition, in vitro preparations of the longitudinal muscle of S.briareus readily converted [U_1 4C]-D-glucose to labelled lactate under aerobic conditions; incorporation was quadrupled under hypoxic incubation condi­ tions (Ellington& Hammen 1977). The studies on anaerobic glycolysis in holothuroid muscle indicate that lactate is an important glycolytic end product; this is undoubtedly the case with other echinoderm tis­ sues having high lactate dehydrogenase activities. A variety of echinoderms have tissues which show relatively low activities of lactate dehydrogenase. This is particularly true of the digestive tracts of echinoids (Ellington & Lawrence 1973) and the pyloric caeca of

400 W.Ross Ellington asteroids (Ellington, unpublished; Durako et al. 1978). It is becoming increasingly evident that many marine invertebrates utilize alternate pathways of anaerobic glycolysis and accumulate such compounds as alanine, succinate, propionate and D-octopine (Hochachka 1975). Incubation of pyloric caeca of A.[orbesi under hypoxic conditions did not result in a significant conversion of [U- 14C]-D-glucose to lactate \.Ellington 1975). The longitudinal muscle of Thyonella gemmata also produced lactate but prolonged incubation resulted in a significant production of succinate (Ellington, unpublished). Thus, it is apparent that gly­ colytic pathways producing end products other than L-Iactate may be present in echino­ derms. Characterization of these pathways would be most revealing especially in those species which are highly tolerant to exposure to anoxic conditions or have internal tissues which are continually bathed in hypoxic coelomic fluid (see section 2.2.3).

1.1.4. Aerobic glycolysis The tissues of echinoderms have the capacity to oxidize glucose completely. Production of 14C02 from the degradation of [14C]-D-glucose has beeri demonstrated in the eggs of the echinoids Strongylocentrotus droebachiensis (Zotin & Vorobeva 1968), Pseudocentrotus depressus (Isono & Yasumasu 1968), Psammechinus miliaris (Bäckström et al. 1960) and Atbacia punctulata (KrahlI956). Homogenates of the gut and testes of Strongylocentrotus purpuratus produced 14C02 as a result ofthe degradation of [6- 14C]-D-glucose (Ulbricht 1974). In addition, in vitro preparations of longitudinal muscle of Sclerodactyla briareus catabolized [l_14C]-pyruvate to 14C02 (Ellington 1976a,b). Iodoacetic acid, a potent inhi­ bitor of glycolysis, markedly inhibited the oxygen consumption of intestine homogenates from the echinoid, S.droebachiensis (Percy 1971). These studies clearly establish the pre­ sence of aerobic glycolysis.

1.1.5. Hexose monophosphate shunt The hexose monophosphate shunt or pentose phosphate pathway is an alternate route of glucose utilization by which NADPH is generated for biosynthetic reactions and pentoses are synthesized for biosynthesis of nucleosides and nucleotides. The existence of significant shunt activity is generally documented by providing the tissue with [l_ 14C]-D-glucose or [6- 14C]-D-glucose and then measuring 14C02 evolution. The hexose monophosphate shunt produces 14C02 by decarboxylation of the glucose skeleton at the I-position by the action of the NADP-requiring enzyme, 6-phosphogluconate dehydrogenase. Radioactive CO 2 pro­ duced by glycolysisjKrebs cycle is derived from both [l- 14C]-D-glucose and [6_ 14C]-D­ glucose. Thus, one can determine the relative activities of the shunt and aerobic glycolysis by measuring the ratios of 14C02 as derived from the degradation of [l_ 14C]-D-glucose and [6_ 14C]-D-glucose. A ratio greater than 1.0 indicates participation ofthe hexose monophos­ phate shunt in glucose utilization. Isotopic studies with specifically labelIed [l4C] -D-glucose have revealed significant hexose monophosphate shunt activity in the eggs and embryos of a number of echinoids including Psammechinus miliaris (Bäckström et al. 1960), Arbacia punctulata (KrahlI956) and Pseudocentrotus depressus (Isono & Yasumasu 1968). Shunt activity increased substan­ tially after fertilization in the latter two species. The activity of the hexose monophosphate shunt has also been investigated in homogenates of the gut and testes of the echinoid Strongy­ locentrotus purpuratus (Ulbricht 1974). In general, cold acclimation produced enhanced shunt activity in the gut and decreased activity in testes (Ulbricht 1974). Within acclimation groups, acute temperature change altered the ratio of 14C02 production from the two sub­

Intermediary metabolism 401 Table 3. Activities of glucose-6-phosphate dchydrogenase and 6-phosphogluconate dehydrogenase in the tissues of echinoderms Enzyme

Species and tissue

Activity

References

Glucosc-6-phosphate dehydro­ genase (Glucose-6-phosphate + NADP+ ~ D-6-P-Glucono­ lactonel + NADPH)

Arbacia punctulata: eggs

14 a,2 b

Strongylocentrotus purpuratus: gut S.purpuratus: lantern retractor muscle

13-23 c

Broyles & Stritmatter 1971,Krahletal.19SS Doezema 1967

6-Phosphogluconate dehydro­ Arbacia punctulata: eggs genase (6-Phosphogluconate + NADP+ ~ Ribulose-S-phosphate S.purpuratus: gut S.purpuratus: lantern retractor + C02 + NADPH) muscle

2-4 c 7-7.S a 9c 1-2 c

Broyles & Stritmatter 1971 Doezema 1967

a. J,Lmoles/min/embryo x 10 7 ; b. J,Lmoles/min/g wet weight at 19°C; c. J,Lmoles/min/mg protein

strates. The physiological significance of these temperature dependent changes in pathway flux is uncertain. In view of the critical importance of the two NADP-linked dehydrogenases in the initial stages of the hexose monophosphate shunt, it is not surprising to fmd studies on glucose-6­ phosphate dehydrogenase and 6-phosphogluconate dehydrogenase in the tissues of a num­ ber of species (table 3). Species-specific changes in the activities of these enzymes have been observed in the developing eggs of the echinoids P. miliaris (Bäckström 1959), Hemicentro­ tus pulcherrimus (lsono 1962) and Paracentrotus lividus (Bäckström 1963). The pH opti­ mum of glucose-6-phosphate dehydrogenase from the gut and testes of S.purpuratus was 8.0 to 8.1, and apparent Km's for glucose-6-phosphate varied depending on the tissue, pre­ vious thermal history of the animal and assay temperature (Ulbricht 1974). Glucose-6­ phosphate dehydrogenase from the intestine of Strongylocentrotus droebachiensis was not inhibited by 5-bromouracil, a weIl established inhibitor of most glucose-6-phosphate dehydrogenases (Percy 1971). Hexose-6-phosphate dehydrogenases, which have broad spe­ cificity for sugar phosphates (glucose-6-phosphate, galactose-6-phosphate, deoxyglucose­ 6-phosphate and glucose) and can use either NAD+ or NADP+, have been found in rela­ tively high activities in the tissues of 18 species of asteroids and three holothuroids (Mochi­ zuk1 & Hori 1976). These enzymes have been extensively studied in terms of physico­ chernical and catalytic properties, but the physiological role of hexose-6-phosphate dehy­ drogenases relative to the hexose monophosphate shunt is uncertain.

1.2. Mitochondrial oxidations 1.2.1. Krebs cycle Krebs cyde activity in the tissues of echinoderms is demonstrated from a number of lines of evidence induding the presence of substrate-derived radioactivity in Krebs cyde inter­ mediates, stimulation of oxygen consumption upon addition of exogenous substrates and the presence of Krebs cyde enzymes. Fixation of 14C02 into organic acids by tissues of the holothuroid, Leptosynapta inhaerens, resulted in a relatively uniform distribution of radioactivity in succinate, fumarate, citrate, isocitrate, 2-oxoglutarate and malate (Harn­

402 W.Ross Ellington men & Osborne 1959). In vitro preparations of the longitudinal muscle of ScIerodactyla briareus degraded [U_ 14C]-D-glucose into Krebs cycle intermediates such as citrate, 2-oxo­ glutarate, fumarate and malate (Ellington & Hammen 1977). Appearance ofradioactivity in Krebs cycle intermediates was also observed in similar experiments with the pyloric caeca of Asterias forbesi (Ellington, unpublished). The oxygen consumption of sperm suspensions of Anthocidaris crassispina was stimulated when Krebs cycle intermediates (Mohri 1962) or octanoate (Mohri 1957) were added to the medium. Echinoid eggs also showed increased oxygen consumption in the presence of citrate, 2-oxoglutarate, succinate, fumarate, malate and glutamate (Cleland & Rothschild 1952a,b). Addition of acetate caused an increase in oxygen consumption of in vitro preparations of the pyloric caeca of the asteroid, Pisaster ochraceus (Allen 1965). Quantitative studies of the activities of Krebs cycle enzymes in echinoderms are limited (table 4). The pyruvate dehydrogenase enzyme complex in the eggs of Pseudocentrotus depres­ sus is under tight regulation, primarily by intracellular calcium concentrations (Yasumasu 1976). The muscles of Echinus esculentus and Thyone sp. contain both citrate synthase and NAD+-linked isocitrate dehydrogenase (table 4) but the activities are so low as to place in doubt the quantitative importance of aerobic energy metabolism in these tissues (Alp et al. 1976). Citrate synthase has been partially purified from the eggs of the echinoid, A.cras­ sispina (Okabayashi & Nakana 1979). This enzyme is strongly inhibited by ATP which might be a means of regulating Krebs cycle activity when ATP levels are high. Malate dehy­ drogenase is found in both mitochondrial (mMDH) and cytoplasmic (sMDH) forms. The

Table 4. Activities of enzymes associated or directly involved in the Krebs cycle as assayed using cmde extracts of echinoderm tissues Enzyme Pyruvate dehydrogenase com­ plex ~pyruvate + CoA + NAD -+ Acetyl CoA + CO 2 + NADH) Citrate synthase (Acetyl-CoA + Oxaloacetate -+ Citrate + CoASH) NAD+-Isocitrate dehydroge­ nase (Isocitrate + NAD+ -+ 2-0xoglutarate+ NADH = Co 2 )

Activity References Yasumasu 1976 Pseudocentrotus depressus: eggs 20.4 a Species and tissue

Echinus esculentus: lantern retractor musde Thyone sp.: pharyngeal retractor muscle E.esculentus: lantern retractor musde Thyone sp.: pharyngeal retractor muscle

Mellita quinquiesperforata: gut Malate dehydrogenase (both cytoplasmic and mitochondrial) Encope abe"ans: gut Arbacia punctulata: gut Luidia clathrata: pyloric caeca Lytechinus variegatus: gut Sc1erodactyla briareus: longitu­ dinal musde

2.6 b

Alp et al. 1976

O.Sb 0.2 b 0.1 b 241 c 313 c S67 c 1,120c 424 c 6.S d

Ellington & Lawrence 1973 Durako et al. 1978 Ellington & Lawrence 1973 Ellington & Hammen 1977

a. nmole/min/mg protein; b. Ilmoles/min/g wet weight at 25°C; c. Ilmoles/min/mg protein at 25°C; d.llmoles/min/g wet weight at 15°C (only cytoplasmic form)

Intermediary metabolism 403 combined activities of sMDH and mMDH in the gut tissue of four species of temperate echinoids indicate that these enzymes are highly active in echinoderm tissues (table 4). Malate dehydrogenase activity has also been demonstrated in the asteroids Nearchaster aciculosus (Ayala et al. 1975), A.forbesi and Asterias vulgaris (Schopf & Murphy 1973) and in a holothuroid, Holothuria sp. (Mendes 1954). Mitochondrial and cytoplasmic malate dehydrogenases have been purified from the plutei of Strongylocentrotus purpura­ tus (Ozaki & Whiteley 1970). The two forms differ in catalytic properties with sMDH being better suited for oxaloacetate reduction and mMDH for malate oxidation. It was sug­ gested that these two isoenzymes constitute a shuttle mechanism by which reducing equi­ valents can be shuttled from the cytoplasm to the mitochondrion (Ozaki & Whiteley 1970). Mitochondrial and cytoplasmic forms of malate dehydrogenase have also been observed in the gut of the sand dollar, Mellita quinquiesperforata (Ellington & Lawrence 1974b), the longitudinal muscle of S.briareus (Ellington 1976b) and in the asteroid,Patiria pectinifera (Manchenko & Serov 1977). The presence of succinate dehydrogenase in echinoderms has been inferred by the observation that malonate, a potent competitive inhibitor of this enzyme, caused decreased oxygen consumption in echinoid eggs (Cleland & Rothschild 1952a,b, Ycas 1954). Succinate dehydrogenase activity is clearly present in the pyloric caeca of A.forbesi. as shown by in vitro enzyme assays (Ellington, unpublished). It has been observed, on the other hand, that malonate caused an increase in respiration of spermatozoa from A. crassispina (Mohri 1957). In addition, attempts at demonstration of succinate dehydrogenase in the longitudinal muscle of Holothuria sp. were not successful (Mendes 1954). The overwhelming evidence from radioactive tracer, exogenous substrate and enzyme activity studies is that echinoderms, in general, possess fully functional Krebs cycles, as is the case with most non-parasitic animals. The basic questions that remain center around the quantitative significance of aerobic metabolism and the role and nature of fatty acid oxidation in Krebs cycle activity.

1.2.2. Cytochromes and oxidative phosphorylation Echinoderms appear to have fully functional cytochrome systems and coupled oxidative phosphorylation for the generation of ATP. The eggs ofthe echinoids,Paracentrotus lividus and Sphaerechinus granularis. have been shown, by means of difference spectra determina­ tions, to contain cytochromes a, a3' band c (Maggio & Ghiretti 1958). Cytochrome oxidase activity in the unfertilized eggs of these two species was 280 and 340 J.d 02/hr/mg protein, respectively. Upon fertilization, cytochrome oxidase activity increased approximately 30 % in P.lividus (Maggio 1959). Cytochrome c oxidase activity was measured in gut, testes and ovaries of the echinoid Strongylocentrotus purpuratus (Ulbricht 1974). Organ differences in enzyme activity were observed; the pattern was changed in response to temperature accli­ mation of whole animals. The organ differences closely paralIed changes in the oxygen consumption of isolated tissues following temperature acclimation. Succinate-cytochrome c reductase and NADH-cytochrome c reductase activities have been demonstrated in the eggs of P.lividus (Rappoport et al. 1958). The above enzymatic machinery is undoubtedly involved in the generation of ATP. Cell­ free preparations of unfertilized eggs of Arbacia punctulata showed high rates of inorganic phosphate disappearance and oxygen uptake in the presence of exogenous 2-oxoglutarate, oxaloacetate and succinate (Keltch et al. 1950). The fertilized eggs of Psammechinus milia­ ris rapidly incorporated 32p inorganic phosphate into ATP (Lindberg 1949). The role of

404 W.Ross Ellington

-

the cytochromes in the generation of ATP has been shown by inhibitor studies. ATP levels in developing eggs of S.purpuratus were reduced by the presence of carbon monoxide (Epel 1963), and by sodium azide and sodium cyanide in Strongylocentrotus droebachiensis (Zotin et al. 1965). Sodium cyanide produced a complete exhaustion of the phosphagen reserve, arginine phosphate, in developing eggs of Pseudocentrotus depressus (Yanagisawa 1968). Uncoupling of oxidative phosphorylation from electron transport is accomplished via 2,4 dinitrophenol (DNP) which increases oxygen consumption and decreases ATP syn­ thesis. Addition of DNP to the developing eggs of P.miliaris anlilP.lividus produced a great increase in respiration (Immers & Runnstrom 1960). The ATP content of S.droebachiensis eggs was reduced by DNP (Zotin et al. 1965). ATP in echinoid eggs appears to be main­ tained by aerobic metabolism as is shown by the great changes in energy status when elec­ tron transport is blocked or uncoupled from oxidative phosphorylation. The situation in adult echinoderms is uncertain. This is especially true of the internal tissues which are sepa­ rated by large diffusion distances from external oxygen in the sea water.

1.2.3. The problem o[internal tissues o[ echinoderms The basic structure of the body walls of most echinoderms, coupled with ineffieient inter­ nal transport systems, greatly limits the amount of oxygen available to internal tissues. As a consequence, oxygen tensions in the coelomic fluid of the echinoid Strongylocentrotus purpuratus, have been reported as low as 40 mm Hg (Webst er & Giese 1975, Johansen & Vadas 1967), even when individuals were placed in oxygenated sea water (P0 2 = 130 ­ 157 mm Hg). Comparable measurements with asteroids indicate that oxygen tensions in the coelomic fluid range from 50-60 mm Hg (J.A.Petersen, personal communication). Needle oxygen electrode measurement of P0 2 in gonads of S.purpuratus gave values rang­ ing from 0.0 to 10.0 mm Hg, indicating that this tissue is essentially operating under anoxic conditions throughout the life cycle of the organism (Webster & Giese 1975). It is becom­ ing increasingly evident that many adult echinoderms can survive for extended time periods (>4 days) in the absence of oxygen. This phenomenon has been documented in asteroids (Theede 1973, Shick 1976), an ophiuroid (Theede 1973) and in a number of holothuroids (Ellington 1976b). The low oxygen tensions in the coelomic fluids of some echinoderms as weIl as the high degree of anoxia tolerance, may suggest a general trend toward decreased reliance on aero­ bic energy metabolism in the intern al tissues of echinoderms. The internal tissues of holo­ thuroids show this phenomenon. The longitudinal muscle of Parastichopus tremulus had the lowest rate of oxygen consumption observed in a survey of tissue respiration in 11 spe­ eies from four invertebrate phyla (Mattison 1959). Similar low rates oflongitudinal muscle oxygen consumption have been observed with tissues from Sclerodactyla briareus (Elling­ ton & Hammen 1977) and Stichopus mollis (Gay & Simmon 1964). The activity of cyto­ chrome oxidase (Mattison 1959) and the presence of cytochromes b, c + c, and a + a3 (Mat­ tison 1962) could not be demonstrated in extracts of longitudinal muscle of Parastichopus tremulus. Electron micrographs of longitudinal muscle of Isostichopus badionotus reveal that this muscle is devoid of a sarcoplasmic reticulum and has an extremely low mitochon­ drial density (R.B.Hill, personal communication). Isolated mitochondria oflongitudinal muscle of S.briareus have few cristae, implying poor capaeity for aerobic metabolism (Ellington 1976b). Thus, the failure to detect cytochromes and cytochrome oxidase is probably not due to the absence of these proteins but due to concentrations below the level for unequivocal detection.

Intermediary metabolism 405 It is clear that the longitudinal muscle of holothuroids is characterized by reduced capa­ city for aerobic energy metabolism and it is probable that this is true for other types of echinoderm internal tissues. The presence of impaired aerobic energy metabolism implies that these tissues have relatively low energy demands or that they rely appreciably on anaerobic pathways of energy production under normal conditions. This latter possibility has not been adequately tested and poses a promising area of future research. 1.2.4. Phosphagens Phosphagens are compounds by which energy can be stored during periods of energy sur­ plus and released du ring periods of high energy demand such as intense muscular activity, This process of energy storage and release is catalyzed by a phosphokinase using the follow­ ing generalized reaction: • storage ' Phosphagen P + ADP Phosphagen + ATP , release Two phosphagens, arginine phosphate and creatine phosphate, have been found in the tis­ sues and gametes of echinoderms (Moreland et al. 1967). The arginine phosphokinase from the longitudinal muscle of Holothuria [orskali has been purified and characterized in terms of a variety of physico-chernical and catalytic properties (Anosike et al. 1975, Anosike & Watts 1975). The arginine phosphokinase from the echinoid, Centrostephanus rodgersii, and arginine and creatine phosphokinases from another echinoid, Heliocidaris erythrogram­ ma. have been partially characterized (Griffiths et al. 1957). The quantitative significance of phosphagens in the energy metabolism of echinoderms remains to be determined.

1.3. Nitrogen catabolism 1.3.1. Amino acid degradation Amino acid catabolism has not been studied extensively in the echjnoderms. The apodous holothuroid, Chiridota rigida, takes up [U)4C] -glycine through the integument and cata­ bolizes the breakdown of this compound to 14C02 (Ahearn & Townsley 1975). The enzy­ matic steps involved in glycine degradation are not known but the appearance of 14C02 suggests the possibility of the operation of a glycine cleavage system (Kikuchi 1973), which catalyzes the following reaction: Glycine + THF + NAD+ -+ 5,lO-methylene THF + CO 2 + NH 3 + NADH Table 5. The aerobic utilization of [1)4 C]-alanine by a mince of the longitudinal muscle of the sea cucumber. Sclerodactyla briareus. Incubation was 60 min at 15°C. Adapted from Ellington (1976b) Fraction

Radioactivity (DPM)

Percent of total radioactivity

Pro tein CO 2 Alanine Glutamate Aspartate Asparagine Lactate Unknown carboxylic acid (#1) Unknown carboxylic acid (#2)

72 823 64250 6984700 12800 18450 44450 126453 42031 37002

0.98 0.87 94.36 0.17 0.25 0.50 1.71 0.57 0.36

Total

7402959

99.77

406 W.Ross Ellington

This system is widely distributed in the vertebrates. Minees of the longitudinal musc1e of Sclerodactyla briareus readily metabolize [1.1 4C]-alanine into other amino acids, some ear­ boxylie acids ,lind 14C02 (table 5). The pattern of labelling of the various intermediates c1early indicates the presenee of glutamate-pyruvate transaminase sinee the transamination produet in the tissue would be [l_14C]-pyruvate whieh eould then be shunted to the Krebs eyde via the pyruvate dehydrogenase reaetion. The enzymatie eapacity for the degradation of alanine, aspartate and glutamate in the tissues of eehinoderms has been demonstrated by means of direet assay of three key enzymes - glutamate-pyruvate transaminase, glutamate-oxaloaeetate transaminase and glutamate dehydrogenase. Both transaminases are present in the eehinoderms as eytoplas­ rnie and rnitoehondrial farms. Cytoplasrnie glutamate-pyruvate tl'ansaminase aetivity in the longitudinal musele of S.briareus is substantially higher than the eorresponding mito­ ehondrial form (table 6). In the ease of glutamate-oxaloaeetate transaminase from three tis­ sues of the holothuroid, Molpadia arenicola, the mitoehondrial and eytoplasmie aetivities are very similar (Dhanani & Kitto 1970). In those instanees where both transaminases have been measured in the same tissue of adult eehinoderms (Ellington 1976b, Okabe & Noma

Table 6. Activity of enzymes involved in amino acid metabolism in echinoderms. The letters in paren­ theses refer to mitochondrial (M) or cytoplasmic (C) forms Enzyme

Species and tissue

Activity

References

Glutamate-pyruvate trans­ aminase (pyruvate + Gluta­ mate ~ Alanine + 2-0xo­ glutarate)

Sclerodactyla briareus: longitu­ dinal musc1e Lytechinus variegatus: pluteus Asterina pectinifera: pyloric caeca cardiac stomach gonads

10.8 (C)a 0.1 (M)a 0.76 b

Ellington i976b

4.7 c 2.9 c 3.8 c

Okabe & Noma 1974b

Glutamate-oxoacetate transaminase (Oxaloacetate + Glutamate ~ Aspartate + 2-0xoglutarate)

Molpadia arenicola: water ring, tentac1es

1.12 (C)d 1.44 (M)d 1.83 (C)d 1.20 (M)d 1.15 (C)d 3.06 (M)d

Dhanani & Kitto 1970

0.39 (c)a

Ellington 1976b

2.4 c 1.4 c 1.3 c 2.12 b

Okabe & Noma 1974b

longitudinal musc1e respiratory trees Sclerodactyla briareus: longitudinal musc1e Asterina pectinifera: pyloric caeca cardiac stomach gonads Lytechinus variegatus: pluteus

Black 1964

l31ack 1964

Glutamate dehydrogenase (Glutamate + NAo+ -e­ 2-0xoglutarate + NH3 + NADH +H")

Sclerodactyla briareus: longitudinal musc1e

O.12 a

Ellington 1976b

Arginase (Arginine + H2 0 ~Ornithine + Urea)

Asterias forbesi: pyloric caeca

1.0a

Hanlon 1975

a.l1moles/min/g wet weight at 15°C; b.l1moles/min/106Iarvae at 25°C; c. units/mg protein at 37°C; d. A340 min/g wet weight at 23°C

Intermediary metabolism 407 1974a), there is a general trend suggesting that glutamate-pyruvate transaminase is more active (table 6). Glutamate-oxaloacetate transaminase activity has been demonstrated in a number of other echinoderms induding Pisaster brevispinus, Dermasterias imbricata, Ophioderma panamensis, Ophioplocus esmarki, Strongylocentrotus purpuratus and Arbacia punctulata (Dhannani & Kitto 1970). The eggs of Strongylocentrotus nudus have also been shown to contain this enzyme (Botrinnik & Neyfakh 1969). Glutamate dehydrogenase seems to be widely distri­ buted in the echinoderms, occurring in A.punctulata, Strongylocentrotus droebachiensis, Echinarachnius parma and S.briareus (Villee 1960). Arginase is present in the pyloric caeca of the asteroid, Asterias forbesi (Hanlon 1975). Loest (1979) observed urease activity in the echinoid, Lytechinus variegatus, and purified the enzyme 300-fold. It was suggested that this enzyme might be involved in calcification processes in this species (Loest 1979). As is apparent from the above limited survey of enzymes involved in amino acid catabo­ lism, a considerable amount of work remains to be done in this area. Of special interest would be studies on the breakdown of branched chain amino acids as weIl as the aromatic amino acids.

1.3.2. Nitrogenous wastes Arginase in echinoderms may function not only in arginine catabolism but also as the ter­ minal enzyme of the urea cyde of ammonia detoxification. High concentrations of urea have been found in the pyloric caeca, gonads and cardiac stomach of the asteroid, Asterina pectinifera (Okabe & Noma 1974b). The coelomic fluids of echinoids such as Arbacia lixula, Sphaerechinus granularis, Psammechinus microtuberculatus (Sanzo 1907) and Para­ centrotus lividus (Delaunay 1931, Fechter 1971) contain significant amounts of urea. These high concentrations of urea probably could not have arisen solely from the break­ down of dietary arginine. It seems likely that urea production may be through de novo synthesis of arginine in echinoderms, although no evidence has been put forward to support this hypothesis. Only one species, P'lividus, has been demonstrated to be ureotelic, as 91 % of nitrogenous wastes passed to the external environment is in the form of urea nitrogen (Fechter 1971). The echinoid Diadema antillarum, on the other hand, excretes 60 % of nitrogenous wastes as ammonia and the remainder as free amino acids (Lewis 1967). Other echinoderrns excrete significant amounts of ammonia induding the asteroids, Asterias rubens (Delaunay 1926) and Luidia clathrata (Ellington & Lawrence 1974a), the echinoid Strongylocentrotus droebachiensis and the holothuroid Eupentacta quinquesemita (Emer­ son 1969). The physiological aspects of excretion are reviewed by Jangoux (chapter 19). 2. BIOSYNTHESIS

2.1. Glycogen synthesis Sugar intraconversion has been extensively studied in the echinoid, Strongylocentrotus purpuratus (Doezema 1967, Ulbricht 1974, Doezema & Phillips 1970). In vitro prepara­ tions of the intestine of S.purpuratus readily incorporated [6_ 14C)_glucose into glycogen (Doezema 1967, Doezema & Phillips 1970). Those portions of the intestine having the highest level of glycogen, showed the highest rate of glycogen synthesis. Homogenates of testes, ovaries and gut of S.purpuratus also showed high rates of glycogen synthesis from [l_14C)-D-glucose and [6)4C)-D-glucose (Ulbricht 1974). Cold acclimation of specirnens

408 W.Ross Ellington resulted in increased rates of glycogen synthesis in testicular and ovarian homogenates, but equivocal results were observed for gut homogenates (Ulbricht 1974). A variety of sugars can be used as precursors of glycogen in the gut of S.purpuratus. [l_14C]-D-mannose, [l_ 14C]-D-mannitol and [U)4C]-D-galactose were readily converted to glycogen in in vitro intestine preparations. In all cases, when the glycogen was degraded, all radioactivity was found to be in D-glucose (Doezema 1967, Doezema & Phillips 1970). Thus, D-mannose, D-mannitol and D-galactose are glucogenic precursors in S.purpuratus. Addition of streptomycin and penicillin G to the in vitro intestine preparation did not block the conversion of [l_ 14C]-D-mannitol to glycogen, indicating that this process is not mediated by bacterial activity (Doezema 1967, Doezema & Phillips 1970). [l)4C]-Fucose was not converted to glycogen in the gut of S.purpuratus. Very little is known about the terminal enzyme of glycogen synthesis, glycogen synthe­ tase, but studies on the previous two enzymatic steps have been done with preparations from the eggs of Anthocidaris crassispina (Tazawa et al. 1977). Phosphoglucomutase seems to be more kinetically suited for conversion of glucose-6-P to glucose-l-P than the opposite reaction. However, metabolite measurements revealed that glucose-6-P is present in a con­ centration nearly three orders of a magnitude greater than glucose-l-P. Thus, glucose-l-P must be rapidly converted to UDPG by UDPG-pyrophosphorylase. Activities of UDPG­ pyrophosphorylase and pyrophosphorylase in A. crassispina eggs are 9.7 and 205 nmoles/ min/mg protein, respectively (Tazawa et al. 1977). The high pyrophosphorylase activity makes the normally reversible UDPG-pyrophosphorylase reaction irreversible by keeping the pyrophosphate concentrations extremely low. The UDPG is then avaiIable for glycogen synthesis via glycogen synthetase. The synthesis of glycogen from precursors other than glucose (for instance, amino acids) involves the process of gluconeogenesis. That is, this process involves conversion of these compounds in the direction which is the reverse of gly­ colysis. Certain specialized enzymes are involved in this process. One of these, fructose-I, 6-diphosphatase (fructose-I, 6-diphosphate + H 2 0 ~ fructose-6-P + Pi) is present in the eggs of Hemicentrotus pulcherrimus and Pseudocentrotus depressus (Yasumasu et al. 1975) and in the tissues of the echinoid Echinopsis sp. (Aiazzi & Bozzi 1963). Phosphoenolpyruvate carboxykinase (oxaloacetate + ITP ~ phosphoenolpyruvate + IDP+ CO 2 ) is another enzyme involved in gluconeogenesis and activity of this enzyme has been demonstrated in the musde of Sclerodactyla briareus (Ellington & Hammen 1977), Echinus esculentus and Thyone sp. (Zammit & Newsholme 1976).

2.2. Amino acid biosynthesis The tissues of echinoderms gene rally have relatively large pools of free amino acids. In general, glycine, alanine, serine and taurine are the most abundant free amino acids as is the case with the eggs of Lytechinus variegatus (Silver & Comb 1966), the asteroids, Echi­ naster sp. (Ferguson 1975) and Luidia clathrata (Simpson et al. 1959) and the holothu­ roids Thyone sp. (SimpsQn et al. 1959) and Stich opus japonicus (Severin et al. 1972). An annual cyde of changes in the relative concentrations of free amino acids has been observed in the asteroid, Echinaster sp. (Ferguson 1975). Free amino acids probably function, in part, during the process of cell volume regulation, as has been demonstrated in Luidia cla­ thrata (Ellington & Lawrence 1974a). . Little information is available on whether all the amino acids in this pool are synthesized by cellular biosynthetic machinery. The eggs of Paracentrotus lividus, incubated with [U­

Intermediary metabolism 409 f4C] -D-glucose, showed radioactivity in alanine, glutamate, aspartate and serine in that order of decreasing radioactivity (Monroy & Vittorelli 1962). The longitudinal musc1e of Sclerodactyla briareus rapidly converted [l_14C]-pyruvate to radioactive alanine (Ellington & Hammen 1977). Incubation of longitudinal musc1e of the same species with [U_ 14C]_D_ glucose under aerobic conditions produced labelling of alanine, predominantly, with very low levels of labelling of glutamate and aspartate (Ellington & Hammen 1977). Similar experiments with the longitudinal musc1e of Thyonella gemmata resulted in labelling of alanine, glutamate, aspartate, serine, glutamine and asparagine (Ellington, unpublished).

Figure 2. The orotic acid pathway for the de novo biosynthesis of uridine monophosphate (UMP). Abbreviations are as folIows: carbamoylphosphate synthetase (CP Sase), aspartate carbamoyl transfe­ rase (AC Tase), dihydro-oratase (DHOase), orotateribosyltransferase (OPR Tase) and orotidine-S'-phos­ phate decarboxylase (OMP decarboxylase) (from Crandall & Tremblay 1976).

410 W,Ross Ellington These tracer studies confirm the presence of glutamate dehydrogenase, glutamate-oxalo­ acetate transaminase and glutamate-pyruvate transaminase in the tissues of echinoderms (see section on amino acid catabolism). These experiments also indicate the presence of the serine biosynthetic pathway as weIl as the presence of glutamine and asparagine syn­ thetases.

2.3. Pyrimidine and purine biosynthesis The generalized orotic acid pathway for the de novo biosynthesis of the pyrimidine uridine­ 5'-monophosphate is shown in figure 2. Developing eggs of Arbacia punctulata incorporated p 4C]-NaHC0 3 into uridine-5'-monophosphate (Crandall & Tremblay 1976). Incubation of the above system with either p 4C]-carbamoylaspartate or [6_ 14C]-orotic acid produced labelled UMP and CMP, with the bulk of the radioactivity in uridine-5'-monophosphate. The P 4C]-HC0 3 was incorporated into the C-2 position of the uracil ring, which is consis­ tent with the orotic acid pathway (Crandall & Tremblay 1976). Further evidence support­ ing the presence of the orotic acid pathway in eggs of A.punctulata was provided by the demonstration of activity of carbamoylphosphate synthetase, aspartate carbomoyltransfe­ rase, dihydro-orotase, dihydro-orotate dehydrogenase, orotate phosphoribosyltransferase and orotidine-5'-monophosphate decarboxylase. It is clear that developing eggs of A.punc­ tulata are capable of the de novo biosynthesis of pyrimidines by the orotic acid pathway. This biosynthesis pathway in A.punctulata may be regulated by end product inhibition (discussed in section 4.4). Incubation of the developing eggs ofPsammechinus miliaris with p4C] -formate resulted in incorporation of radioactivity into hypoxanthine and guanine (Hultin 1957). In similar experiments with Paracentrotus lividus, guanine was labelled in trichloroacetic extracts while both adenine and guanine were labelled in the RNA fraction (Hultin 1957). It is likely that P 4C]-formate is incorporated into the developing purine skeleton in the form of 5,10-methenyl-THF. Incubation of developing eggs of A.punctulata with 4C]_ HC0 3 resulted in labelled GMP and AMP (Crandall & Tremblay 1976). The observations of incorporation of [14C]-formate and [14C]-HC0 3 into purine skeletons strongly support the suggestion that purine biosynthesis occurs in echinoderms.

P

2.4. Lipid biosynthesis 2.4.1. Lipogenesis As storage materials, lipids are extremely important in the energy economy of echinoderms (Lawrence 1976). As a consequence, the biosynthesis oflipids in echinoderms from carbo­ hydrate substrates has been studied extensively. Most of these studies have involved provid­ ing the intact organism or in vitro tissue preparation with [14C]-D-glucose or [l.l4C]-ace­ tate. After incubation, lipids are extracted using chloroform-methanol (Folch et al. 1957) and radioactivity is counted in the crude lipid fraction. In some cases, lipids are separated into the various lipid classes (neutral lipids, phospholipids, glycerol ethers and sterols). In vitro preparations of the pyloric caeca of Asterias [orbesi incorporated p 4C]-acetate and P4C]-glycerol into both neutral and phospholipids (Karnovsky & Brumm 1955). In the case of both substrates, phospholipids contained twice the radioactivity of neutral lipids and p4C]-acetate was incorporated only into the fatty acid moity. The developing eggs of Paracentrotus lividus also incorporated [l_14C]-acetate and [l_14C]-glycerol into

Intermediary metabolism 411 lipids (Mohri 1964). Incorporation of [l_14C]-acetate was much higher in the phospholipids and revealed that ceph1l1in (phosphatidylethanolamine) was much more highly labelled than lecithin (phosphatidylcholine) even though lecithin was quantitatively the most abun­ dant phospholipid. [1_ 14C] -glycerol was a relatively poor substrate for lipogenesis in the embryos of P'lividus. This poor incorporation may be due to the fact that glycerol is not a direct substrate for lipogenesis but must be converted to L-3-glycerol-phosphate by the . glycerokinase reaction (Mohri 1964). Synthesis of phospholipids has been further investi­ gated in the eggs and developing embryos of Arbacia punctulata (Schmell & Lennarz 1974), Strongylocentrotus purpuratus and Lytechinus pictus (Pasternak 1973). In vitro preparations of the pyloric caeca ofthe asteroid, Pisaster ochraceus incorporated [1-14C]-acetate and [u.l 4C]-D-glucose into crude lipid fractions (Allen 1965, Allen & Giese 1966). Acetate appeared to be the best substrate for lipogenesis. Fasted animals showed a much higher rate of lipogenesis than unfasted. In addition, the pyloric caeca of fasted ani­ mals incorporated radioactivity into phospholipid and neutral lipid in a 2: 1 ra,ti0. The quan­ titative significance of de novo lipid biosynthesis in P.ochraceus could not be calculated due to dilution of the tracer in the endogenous acetate pool (Allen 1965). The high lipid content of the pyloric caeca (10-12 % dry weight) vs. the low lipid content of the prime food item of this asteroid (Mytilus edulis) coupled with the qualitative differences in fatty acid content suggest that even though differential uptake and utilization occur, de novo biosynthesis of lipids is occurring in this species (Allen 1965). The capacity for lipid biosynthesis varies from tissue to tissui in the echinoderms. In vitro incorporation of [l_14C] -acetate was investigated in various tissues of the asteroid Asterlas rubem, the echinoid, Echinus esculentus and the holothuroid, Holothurla forskali (Allen 1968). Tissues of both A.rubens and E. esculentus have high rates of lipid synthesis, especially in the pyloric caeca and gut. The highest rate found in Holothurla forskali was in the gut, an order of magnitude lower than that of the two most active tissues in the other species. Again, some caution must be placed in interpretation of absolute values of rates due to between-tissue-type and between-species variation in isotope dilution. The rates of lipid synthesis from [1_ 14C] -acetate have been investigated in gut and testes homo­ genates of tissues from specimens of S.purpuratus which had been acclimated to various temperatures (Ulbricht 1974). Cold acclimation resulted in increased incorporation into lipids in homogenates ofboth tissues at all incubation temperatures. Voogt and co-workers have demonstrated in vivo incorporation of [l_14C] -acetate into neutral, phospholipids and sterols in a wide range of echinoderms including the following: asteroids A.rubens, Marthas­ terlas glacialis, Astropeeten aranciacus and Echinaster sepositus (Voogt 1973a); holothu­ roids Cucumarla planci, Holothuria tubulosa and Stich opus regalis (Voogt & Over 1973); ophiuroids Ophiura albida and Ophioderma longicauda (Voogt 1973b); echinoidsParacen­ trotus lividus, Echinus acutus and Psammechinus miliaris (Voogt 1972). In all cases, the bulk of the radioactivity was found in the saponifiable lipid fraction. Ether lipids differ from triglycerides and phospholipids in that they contain an alkyl (-OR) oralkenyl (-O-CH =CH-R)group in the place of one acyl group(-O-C-R). The similarity in structure of these compounds to triglycerides may indicate that they serve prirnarily in energy storage (Allen 1976). In vitro preparations of the pyloric caeca of A.[or­ besi incorporated P4C] -acetate and P4C] -glycerol into glycerol ethers (Karnovsky & Brumm 1955). Further studies with [14C] -acetate revealed that bulk of incorporation was in alkyl glycerol ethers rather than alkenyl glycerol ethers (Ellingsboe & Karnovsky 1967). [l_14C]-stearic acid and [l_14C]-stearaldehyde were both incorporated into alkyl glycerol

q

412 W.Ross Ellington ethers in A.[orbesi pyloric caeca, but [1_ 14C] -stearalhyde was a much better precursor of alkenyl glycerol ethers. An alkyl glycerol ether biosynthetic system has been found in the rnicrosomal fraction of the pyloric caeca and gonads of A.[orbesi (Snyder et al. 1969). This biosynthetic system utilizes long chain alcohols, gylceraldehyde-3-phosphate, ATP, coenzyme A and Mg+ 2 to produce the alkyl ethers. The rnicrosomal enzyme system will not produce alkenyl ethers (Snyder et al. 1969). The actual details of fatty acid biosynthesis in echinoderms are unknown. Presumably, biosynthesis of saturated fatty acids is catalyzed by a multi-enzyme fatty acid synthetase. The reducing power, in the form of NADPH, is provided by glucose-6-phosphate and 6­ phosphogluconate degydrogenases (see section 2.1.5). The incorporation of [1_ 14C] -acetate into the free fatty acids of the pyloric caeca A.rubens, gut and gonad of E. esculentus and ovary and gut of Hforskali was primarily in the saturated fatty acids (no C = C bonds) with monoenes, dienes and polyyenes containing lesser amounts of radioactivity (Allen 1968). The above pattern was also characteristic of the fatty acids found in the alkylglycerol ethers, triglycerides and phospholipids labelled as a result of [l.1 4C] -acetate incorporation by the pyloric caeca of A.rubens (Allen 1968).

2.4.2. Sterol biosynthesis and metabolism This topic is reviewed by Voogt (chapter 18) .

.

3. REGULATION OF INTERMEDIARY METABOLISM The process of intermediary metabolism is a tightly-linked, complex process with numerous pathways often competing for common substrates. Regulatory controls must act on meta­ bolism to maintain proper flux of substrate and to be able to respond to changes in the physiological demands of the cell/tissue system. Knowledge of the mechanisms of regula­ tion of metabolism in echinoderms is extremely limited, especially in adult echinoderms. This review will treat three echinoderm systems in which control mechanisms have been documented in detail.

3.1. Control ofphosphorylase activity in echinoid eggs Phosphorylase exists in two enzymatic forms in most animal tissues. Phosphorylase a is termed the active enzyme because of its high activity in the absence of AMP. The activity of phosphorylase b is dependent on the modulator, AMP. Phosphorylase b can be converted to form a by the action of phosphorylase kinase which phosphorylates and activates the enzyme at the expense of ATP. The developing eggs of Pseudocentrotus depressus and Hemicentrotus pulcherrimus showed increased activity of both AMP independent and total phosphorylase activity (Shoger et al. 1973), suggesting that there was an increase in phos­ phorylase a activity associated with fertilization and development. Since the total phos­ phorylase activity increased, some other re action in addition to the conversion of b ~ a, may be occurring. In rabbit cardiac muscle, phosphorylase kinase is activated in the pre­ sence of the hormone second messenger , cyclic 3',5'-AMP, and trace amounts of Ca+ 2 (Ozawa & Ebashi 1967). Intracellular calcium concentrations increase in echinoid eggs due to the release of bound calcium upon fertilization (Mazia 1937, Nakamura & Yasumasu 1974). In addition, adenyl cyclase activity (ATP ~ cyclic 3',5'-AMP) in the membrane

Intermediary metabolism 413 fragments of the eggs of Lytechinus pictus increased after fertilization (Castenada & Tyler 1968). The above observations suggest that fertilization in the eggs of P.depressus and Hpul­ che"imus results in the following sequence of events in terms of phosphorylase activity: Fertilization

/increased adenyl cyclase activity ~ increased cAMP "increased intracellular Ca+ 2 " \ phosphorylase conversion (b ~ a) - activation of phosphorylase kinase

This hypothesis was tested using the eggs of P.depressus (Shoger et al. 1973). Preincuba­ tion of unfertilized eggs in the presence of ATP and Mg+ 2 resulted in no pronounced increase in AMP - independent phosphorylase activity. When either calcium or cyclic 3', S'-AMP were added alone there was activation but maximal activation occurred in the pre­ sence ofboth Ca+ 2 and cyclic 3', S'-AMP. Under these conditions, the activity ofphospho­ rylase a in unfertilized eggs was comparable to that of fertilized eggs. Thus, fertilization results in the initiation of a complex regulatory system allowing for a qualitative change in phosphorylase as mediated by changes in intracellular concentrations of Ca+ 2 and cyclic 3', S'-AMP. It must be emphasized that this mechanism does not account for the increase in total phosphorylase in P.depressus and Hpulche"imus eggs upon fertilization, only for tue conversion of some b form to phosphorylase a (Shoger et al. 1973). Phosphorylase in the gut of the sea urchin, Strongylocentrotus purpuratus, appears to be primarily in the b form, as additional AMP greatly activates phosphorylase activity in in vitro enzyme assays (Doezema 1967). Intraconversion of b ~ a may also be controlled by a cyclic nucleotide mediated system.

3.2. Control 01 the activity 01 the pyruvate dehydrogenase complex in echinoid eggs Activity of the pyruvate dehydrogenase complex was not detectable in mitochondria from unfertilized eggs of Pseudocentrotus depressus, but this activity increased and reached a maximum 20 rninutes after fertilization (Yasumasu 1976). Activity in unfertilized eggs could be stimulated by addition of Ca+ 2 at a physiological concentration but pyruvate de­ hydrogenase activity from fertilized eggs was only slightly stimulated by Ca+ 2 • In other animal systems pyruvate dehydrogenase activity is controlled by phosphorylation and de­ phosphorylation according to the following scheme: Kinase Pyruvate dehydrogenase-P (inactive)

/'

/'

Phosphatase

Pyruvate dehydrogenase (active)

The pyruvate dehydrogenase phosphatase is activated by calcium. The calcium activation of pyruvate dehydrogenase in unfertilized P.depressus eggs may be due to activation of a speci­ fic phosphatase and dephosphorylation of the enzyme complex (Yasumasu 1976).

3.3. Feedback inhibition olpyrimidine biosynthesis in echinoid embryos As discussed in section 3.3, the developing embryos of Arbacia punctulata are capable of the de novo biosynthesis of pyrimidines by the orotic acid pathway (Crandall & Tremblay

414

W.Ross Ellington

Figure 3. The effect of pyrimidine and purine nucleosides on the incorporation of [I4C]-Na HC0 3 into orotic acid by embryos of Arbacia punctulata (from Crandall & Tremblay 1976).

1976). Addition of the nuc1eoside uridine resulted in a substantial decrease in the incorpora­ tion of [14C] -HC0 3 into orotic acid (fig.3). This inhibition appears to be true end product inhibition, as the purines guanosine and adenosine were generally ineffective as inhibitors (fig.3). This end product inhibition signals decreased biosynthesic output when pyrimidine levels are high (Crandall & Tremblay 1976). 4. CONCLUSIONS The existing data on intermediary metabolism in echinoderms support the conc1usion that these organisms possess the same basic pathways present in other metazoans. The catabo­

Intermediary metabolism 415 lism of carbohydrates is mediated by glycolysis, the hexose monophosphate shunt and the Krebs cycle. Echinoderms catabolyze certain amino acids and fatty acids, but the details of these degradative pathways are poorly known, especially in terms of fatty acid break­ down. Glycogen synthesis has been studied in some detail in the echinoderms, as is the case with the biosynthesis of lipids. The regulation of intermediary metabolism in adult echino­ derms remains a fruitful area of future research. In light of the diverse physical environments occupied by echinoderms, studies of regulation of metabolism in response to changes in ertvironmental variables should be most revealing.

ACKNOWLEDGEMENTS I thank Drs W.V.Allen, C.S.Hammen and J.M.Lawrence for criticism of the chapter.

18

PETER A. VOOGT

STEROID METABOLISM

All eukaryotic organisms investigated thus far contain sterols. Only in some rare cases other - but structurally closely related - compounds are found instead of sterols. The universal presence of sterols is understandable from their function, for it is generally accepted that they are essential structural elements stabilizing cell membranes and involved in regulating membrane viscosity and permeability (Bloch 1979). A wide diversity of sterols is known and each year their number still increases due to the discovery of new sterols. They are dif- J ferent from each other in the number of carbon atoms, the number and the site of double bonds, the configuration at C-24, and the degree of hydroxylation. Based on the sterol composition known at that time, Bergmann (1962) stated that the sterol mixtures of 'higher' animals are less complex than those of 'lower' ones. This could be due to a tendency to dirninish the diversity of sterols used resulting in an increase in the relative amount of cholesterol. Since cholesterol is nearly the sole sterol present in the ver­ tebrates, he supposed that cholesterol had survived by being the 'fittest' sterol. In this res­ pect the sterol composition of echinoderms is highly interesting, because the Echinoder­ mata - belonging to the Deuterostornia - are gene rally considered to be closely related to the Vertebrata. From this evolutionary point of view one might expect the sterol mixtures of echinoderms to be relatively simple with a predominance of cholesterol. Bergmann's (1962) ideas imply that organisms achieve and maintain their own specific sterol composition. Generally speaking, two sources are available to organisms to meet their need of sterols. The first is de novo biosynthesis. It is easy to understand that in prin­ ciple a characteristic sterol pattern can be obtained via this mechanism. However, not all organisms are able to synthesize sterols; others have only limited capacities or make only little use of them. This endogenous supply may be replaced or supplemented by exogenous sterols from the diet, the second source. This implies that animals at the end of a food chain would accumulate the sterols of all preceding links, unless they are able to convert them into their own type of sterols, or possess mechanisms for regulating sterol composi­ tion. Thus, the sterol pattern of an organism is the result of contributions from the diet and from sterol metabolism. Besides the general function in membranes, sterols have a special function as precursors for steroid hormones in vertebrates. This aspect has been little studied in invertebrates. A third function of sterols in vertebrates is that they are precursors for the synthesis of cholic acids. These surface-active compounds play an important role in the emulsification and subsequent resorption of dietary lipids; further they are involved in the complex regu­ lation of blood-cholesterollevels. In two classes of the echinoderms (viz. Asteroidea and Holothuroidea) compounds are found, which in several aspects resemble the cholic acids of the vertebrates. These com­ 417

418 Peter Voogt pounds, called saponins, are steroid derivatives and are highly surface-active . Mackie et al. (1977) suggested that the saponins present in the stornach of asteroids might solubilize cholesterol from the prey and facilitate its absorption . Saponins of asteroids are structurally related to sterol sulphates, which are present in rather high concentrations. Goodfellow & Goad (1973) suggested that sulphurylation may be a means to eliminate excess sterols, and especially cholesterol, obtained from the diet.

1. STEROLS IN ECHINODERMS As mentioned before, sterols of echinoderms are, from a theoretical point of view , of great interest because it rnight be expected that they will show strong resemblance with those of vertebrates. A short time after the discovery of cholesterol Doree (1909) undertook what may be called the first study on the comparative biochemistry of sterols. Most animals investigated contained cholesterol, but he found that asteroids contained another type of stero!. Bergmann & Stansbury recognized already in 1944 that asteroid sterols consisted of 7 .6. -sterols. From that time Bergmann and his coworkers started aseries of investigations into echinoderm sterols. In his classic paper of 1962, which can be considered to be his testimony, Bergmann showed that, on the basis of the sterol type present, echinoderms can be divided into two groups. The first group comprising the Crinoidea, Echinoidea and Ophiuroidea contains ,0,.5-sterols, whereas the second group, comprising the Asteroidea and Holothuroidea, contains .6.7 -sterols. This discovery was considered by Bergmann to be a further indication to the phylogenetic relationships within the phylum of Echinodermata.

Figure 1. The mean percentage composition, according to carbon level, of the sterol mixtures obtained from representatives of the various echinoderm classes.

Steroid metabolism 419 On the ground of embryological data, Hyman (1955) had related echinoids and ophiuroids to one another and the asteroids to the holothuroids. These relationships are in full agree­ ment with those which can be deduced from the distribution of sterol types. Otherwise there is a considerable amount of comparative biochemical data which support the relation­ ships given (Bolker 1967, Grossert 1972). However, according to Fell (Fell & Pawson 1966) these data are of litde value, because these similarities do not reflect phylogenetic relation­ ships but result from convergent evolutions. Based on palaeontological data he thinks aste­ roids to be more closely related to ophiuroids, and the echinoids to the holothuroids. Studies on the sterol composition of echinoderms have been seriously hampered for a long time by the lack of adequate analytical techniques. Bergmann (1962) already recog­ nized that the sterols of echinoderms consist of complex mixtures and that progress in this field could be possible only if suitable separation and identification techniques were avail­ able. In true prevision he expected much of vapour chromatography, which had been introduced shortly before. This came true when Gupta & Scheuer (1968) used gas-liquid chromatography and mass spectrometry to examine the sterol mixtures from one represen­ tative of each echinoderm class. Their results confirmed the class-wise distribution of 1::,5_ and 1::,7 -sterols and the complexity of the sterol rnixtures, consisting of C27 -, C28 -, C29 - and C30 -sterols. Since that time similar studies have been made on several species of all five classes, but the picture has remained unchanged. From these analyses some generalizations can be drawn. In all species studied C26 -sterols are present in small concentrations, generally less than 3 % (Voogt 1972a,b, 1973a,b, Voogt & Over 1973, Voogt & Schoenmakers 1973, Goad et al. 1972). Somewhat higher concen­ trations (6-9 %) were found by Goad etal. (1972) for the ophiuroids Ophiocomina nigra and Ophiura albida. Incidentally, very small amounts of C3Q-sterols are present. C2Tsterois are predominant in all classes (with the exception of crinoids) and are particularly abun­ dant in echinoids, making up 75-90 % of the total sterols. Generally the sequence of the relative amounts of the sterol types is: C27 > C28 > C29 . However , in the holothuroids the sequence is: C27 > C29 > C28 , whereas in the crinoids this is: C28 > C27 > C29 . The propor­ tional composition according to their carbon content of the sterol mixtures in the echino­ derm classes, and based on data from Goad et al. (1972), Voogt (1972a, 1973a,b) and Voogt & Over (1973) are summarized in figure 1. The main sterol in echinoids is cholesterol (fig.2), making up 70-80 % of the total (Voogt 1972a,b). Cholesterol is present also in crinoids and ophiuroids but in much lower concen-

Figure 2. Structure of cholesterol with numbering of carbon atoms.

420 Peter Voogt

trations. Inophiuroids cholesterol constitutes about 25 % ofthe total (Voogt 1973a) and 20.5 %in the crinoid Antedon bifida (Gupta & Scheuer 1968). The main sterol in most asteroids is cholest-7-en-3ß-ol. Voogt (1973b) reported values ranging from 25-60 %, while those of Goad et al. (1972) were somewhat higher (35-70 %). In some asteroids species cholest-7-en;Jß-ol is only a minor component; it constitutes 7.5 % in Acanthaster planci (Gupta & Scheuer 1968) and 5 % in Echinaster sepositus (Voogt 1973b). With reference to holothuroids Gupta & Scheuer (1968) reported that cholest-7-en-3ß-ol makes up 43.5 % of the sterol rnixture in Holothuria atra, being the main sterol. This sterol is also predominant in Stichopus japonicus and Holothuria tubulosa (Nomura et al. 1969a) and in Cucumaria hyndmani and Cucumari elongata (Goad et al. 1972). In contrast, Voogt & Over (1973) reported that cholest-7-en-3ß-ol made up only 3-12 % of the sterols of the Mediterranean species Cucumaria planci, Holothuria tubulosa and Stichopus regalis. Since the pioneering work of Gupta & Scheuer (1968) the sterol composition of a great number of echinoderms has been described in detail (see reviews by Austin 1970 and Goad 1976). In the last few years the yearly number of papers on echinoderm sterol composition has somewhat decreased and, perhaps due to the availability of mass spectrometry, interest seems to be concentrated more on searching for novel sterols. Indeed, the echinoderms have contributed considerably in this respect. In the asteroid Asterias amurensis, Kobayashi et al. (1972, 1973) found a C26 -sterol, called asterosterol, which was characterized as 24­ nor-5a-cholesta-7,22-dien-3ß-ol. At the same time, this sterol was tentatively identified by Matsuno et al. (1972) in several Japanese asteroids and by Goad et al. (1972) in holothu­ roids. Kobayashi et al. (1974) and Boll (1974) proved the proposed structure to be correct by chernical synthesis. Rubinstein (1973) found a new component, identified as 24-nor-cholest-5-en-3ß-ol, among the C26 -sterols of ophiuroids and crinoids. Goad et al. (1972) showed 24-nor-cholest­ 22-en-3ß-ol to be present in the asteroid Henricia sanguinolenta. A new C27-sterol, amuresterol, with the structure (22E,24S)-27-nor-24-methyl-5a­ cholesta-7,22-dien-3ß-ol was isolated by Kobayashi & Mitsuhashi (1974) from the asteroid A.amurensis and since then found to be present in several other asteroid species. A new C2s-sterol, 24~-methyl-5a-cholesta-7 ,22,25-trien-3ß-ol was isolated from the aste­ roid Leiaster leachii (Teshima et al. 1974). Gupta & Scheuer (1968) isolated a novel C3Q­ sterol, acanthasterol, from A.planci. Sheikh et al. (1971) found the structure of acanthas­ terol to be (22R,23R,24R)-22,23-methylene-23,24-dimethyl-5a-cholest-7-en-3ß-ol. The corresponding stanol, gorgostanol, was isolated from this species by Kanazawa et al. (1974). A new impetus was given by the development of high resolution nuclear magnetic reso­ nance (NMR) techniques, with the possibility of ascertaining the configuration of methyl and ethyl groups at C-24. Smith et al. (1973) showed that the C28 -.6.7 -sterol of Asterias rubens has the 24a-configuration, which was found also by Rubinstein (1973) for the C28 ­ .6.5 -sterol from the ophiuroid O.nigra. Although no recent information is available on the C-24-configuration in C29 -sterols of echinoderms, some older data (Toyama & Takagi 1955, Matsumoto & Wainai 1955) make it plausible that they also have the 24a-configuration. In this respect the C24-alkyl sterols of echinoderms are different from those in algae, which generally have a 24ß-configuration. Since many herbivorous echinoderms feed on algae, these studies may be very helpful in establishing the way in which the sterol pattern in echinoderrns is achieved and particularly the contribution from the diet.

Steroid metabolism 421 2. THE ORIGIN OF STEROLS IN ECHINODERMS

2.1. De novo sterol synthesis After the report of Salaque et al. (1966) that no radioactivity was incorporated from 11,2­ 14CI-acetate into the sterols ofthe echinoidParacentrotus lividus, Voogt (1972b) showed that this species as well as Echinus acutus and Psammechinus mi/iorls synthesized sterols from 11.14CI-acetate. Smith & Goad (1974) found that 12)4CI-mevalonate was readily incor­ porated into the intermediates squalene and lanosterol in Echinus esculentus. Desmosterol wa! highly radioactive, whereas cholesterol was only poorly labelled. Injected 126)4CI-desmos­ terol apparently was not converted into cholesterol, whereas dihydrolanosterol in which the ~24-bond is absent was converted into cholesterol. From these results Smith & Goad (1974) concluded that E.esculentus is able to synthesize cholesterol but that the ~24-reduc­ tase is rate limiting. Rubinstein (1973) reported that P.miliarls is able to synthesize C27 ­ sterols both from 11- 14CI-acetate and from 12- 14CI-mevalonate. The ophiuroids Ophiura albida and Ophioderma longicauda were able to synthesize sterols from injected 11- 14CI-acetate (Voogt 1973a). The observation ofRubinstein (1973) that both 11- 14CI-acetate and 12)4CI-mevalonate were incorporated into squalene and cho­ lesterol, whereas C2S - and C29 -sterols were not labelled, in Ophiocomina nigra is important. In the crinoid Antedon bifida 12-14CI-mevalonate was incorporated into squalene and un­ identified sterols, but labelling was very low (Rubinstein 1973). Thus, in those echinoderm classes in which the sterols are of the ~S_type, sterol biosyn­ thesis from acetate and mevalonate has been observed in several representatives. Though the picture is still very incomplete there is some evidence that the biosynthesis de novo is restricted to C27 -sterols. Nomura et al. (1969b) studied the sterol biosynthesis in the holothuroid Stichopus japonicus and observed that 11 ,2- 14CI-acetate was incorporated into squalene but that lanos­ terol and thus also the sterols, which were mainly of the ~ 7 -type, were unlabelled. They sug­ gested that possibly S.japonicus is incapable of cyclizing squalene to lanosterol and con­ cluded that the sterols and dimethylsterols of this species are totally obtained via the food. In contrast, Voogt & Over (1973), using 11-14CI-acetate, showed that Cucumaria pfand, Holothuria tubulosa and Stich opus regalis synthesize unidentified sterols. Incorporation of radioactivity into the sterols of H tubulosa and S.regalis was very poor, however. This was confirmed and extended by Rubinstein (1973), who showed in Cucumaria lactea the incor­ po ration of 12- 14CI-mevalonate into squalene and into C27 -sterols, which included cholest-7­ en-3ß-ol. Thus far the data obtained suggest that, as in the three classes mentioned before, the biosynthesis of sterols in holothuroids is also restricted to those with 27 carbon atoms. Sheikh & Djerassi (1977) studied sterol biosynthesis in Stichopus californicus by using ICH 3-3 HI-acetate and 13-3HI-Ianosterol. They found nearly the same specific radioactivities for a1l the individual sterols isolated, that of the C29 -sterols being somewhat higher in the acetate incubation and that of cholesterol together with cholestanol being somewhat higher in the lanosterol incubation. Sheikh & Djerassi (1977) concluded that 'Stichopus californi-' cus can biosynthesize sterols de novo from 3H 3C-COO-r and can transform lanosterol to ~ 5 - and ~ 7 -unsaturated sterols'. The incorporation of radioactivity from acetate into sterols was very low (only 0.067 %). They also concluded that this 'sea cucumber can alkylate the cholesterol side chain at posi­ tion 24 to furnish both ~s_ and ~7-24-methyl and ethyl sterols, thus excluding a purely dietary origin'.

422 Peter Voogt

HO lanosterol

HO

!

HO 4 A - dimethylcholesta - 8,24 - dien -3ß-ol

dihydrolanosterol 1f

HO

.,r-••

4A-dimethylcholest-8-en-3ß-ol

1f

!

HO

HO

4A-dimethylcholesta-7.24-dien-3ß-ol

4A-dimethylcholest -7- en-3ß-ol

HO

HO

40'- methylcholesta-7.24 - dien -3ß-ol

4O'-methylcholest-7-en-3ß-ol

HO cholesta-7.24-dien-3ß-ol

.

HO

cholest-7-en-3ß-ol

Figure 3. Probable routes for the biosynthesis of cholest-7-en-3ß-ol in asteroids. The main pathway is indicated by bold arrows. Possible additional routes, depending on the stage in which,",.24 saturation occurs, are indicated by thin arrows. All the intermediates except for those marked * have been iden­ tified in Asterias rubens (modified with permission from Smith & Goad 1975).

Steroid metabolism 423 In the conversion of lanosterol to C27 -sterols, the C4-attached methyl groups must be eliminated. The intermediacy of a 3-keto-4-carboxylic acid in this process is generally accepted. Lanosterol is labelled at C-3 and yet tritium is retained in cholesterol and its homologues. According to Sheikh & Djerassi (1977) this is suggestive of an alternative mechanism in holothuroids. However, since the paper does not give proof that the label still is present at C-3, it may have been removed at some stage and reincorporated at a later one. Moreover, several authors have pointed out the risk of using tritium with respect to the chance of label exchange. Since the label distribution in these experiments is quite equal in the sterols isolated, it would be very desirable to know the location of the label. In conclusion, there is evidence that holothuroids can synthesize cholest-7-en-3ß-ol and possibly cholest-5-en-3ß-ol de novo, and that at least one holothuroid also can alkylate the C27 -sterols to give C2S - and C29 -ones. Much more data are known about the sterol synthesizing capacity in asteroids. Smith & Goad (1971) showed that Asterias rubens and Henricia sanguinolenta incorporated 12_14 CI_ mevalonate into squalene, 4,4-dimethyl sterols and cholest-7-en-3ß-ol. Incorporation of 11 or 2- 14CI-acetate into the sterols of A.rubens, Astropecten aranciacus and Echinaster sepositus was also observed by Voogt (1973b), though incorporation was rather poor. Smith & Goad repeated their experiments (Goad et al. 1972, Smith & Goad 1975) and showed that after 17 hours the radioactivity was about equally distributed between squa­ lene and the 4,4-dimethyl sterols, and that labelling in the 4-desmethyl sterols remained low even after prolonged incubation times up till 41 hours. They also showed that the radioactivity in the 4-desmethyl sterols was fully associated with the C27-sterols and was particularly present in cholest-7-en-3ß-ol. On the basis of a careful analysis of the radio­ labelled 4,4-dimethyl and 4-monomethyl sterols they proposed a probable pathway for the biosynthesis of cholest-7-en-3ß-ol in asteroids (fig.3). Teshima & Kanazawa (1975) studied the sterol biosynthesis in the asteroid Leiaster leachii, which had been injected with 12- 14CI-mevalonate ten days before being sacrificed. The highest radioactivity was present in cholest-7-en-3ß-ol, but two other sterols, with a proposed structure of C27'6.2,24 and C27~7,25 were radioactive too. On the basis of these and other results, they concluded that L.leachii is able to synthesize ~7-cholestenol from me va­ lonate, but that it seems not to possess the ability for either alkylation at C~24 or the intro­ duction of double bond at C-22 of ~7-cholestenol. Thus, C26 -, C2S - and C29 -sterols in L. leachii should exclusively originate from the diet. Teshima & Goad (in: Goad 1975) incubated A.rubens and Porania pulvillus with 12_14 CI­ mevalonate for several weeks. The main sterollabelled was cholest-7-en-3ß-ol, but no incor­ poration into cholesterol had taken place. Voogt & van Rheenen (1976b) incubatedA.rubens with 126.14CI-cholesterol for 18 hours. The sterols were hydrogenated, and their acetates purified in thin-Iayer chromatography, and then subjected to preparative gas-liquid chromatography. The levels of radioactivity of the C27-, C2S- and C29 -sterol fractions allowed the conclusion that A.rubens cannot syn­ thesize C2S - and C29-sterols from cholesterol. Since alkylation might occur at a stage in the sterol biosynthesis prior to cholesterol, they also used 12-14CI-mevalonate as the precursor, but again radioactivity was restricted to C27 -sterols. Incubation for 72 hours with L-Imethyl­ 14CI-methionine, the methyl donor in the alkylation process, resulted in the complete absence of radioactivity in the sterols of A.rubens. After incubation with 12-14CI-mevalonate a slight incorporation of radioactivity into 50:­ cholesta-7,22-dien-3ß-ol was observed. This might indicate thatA.rubens is able to intro­

424 Peter Voogt duce a double bond at C-22. To confirm this, incubations were also done with 126-14CI­ cholesterol, with the result that both cholesta-5,22-dien-3ß-ol and 5a-cholesta-7,22-dien­ 3ß-ol were radiolabelled. This contrasts with what has been found for L.leachi by Teshima & Kanazawa (1975). The capacity of introducing a double bond at the C22-position pro­ bably strongly extends the possibilities for A.rubens to vary its sterol composition. Furthermore, some evidence indicates that A.rubens can synthesize cholest-5-en-3ß-ol de novo, but only. by a minor pathway (Voogt & van Rheenen 1976b). This point deserves further attention because it is gene rally accepted that asteroids lack the enzymes catalyzing the fmal two steps in sterol biosynthesis, and which lead from a A7 -sterol via aAs, 7 -one to aA S -sterol. In conclusion, sterol biosynthesis has been demonstrated in representatives of all five echinoderm classes, but very likely in all of them the biosynthesis is restricted to C27-sterols. Only Sheikh & Djerassi (1977) reported alkylation in a holothuroid. In those classes in which the sterols are of the A5 -type, cholesterol is the end product, whereas in asteroids and holothuroids it is cholest-7-en-3ß-ol. According to Sheikh & Djerassi (1977) Stichopus cali[ornicus synthesized cholest-5-en-3ß-ol in addition. Even if the capacity of introducing double bond at C-22 were a common property of echinoderms, the sterol synthesizing capacity still would be of limited extent and by no means sufficient to provide all the dif­ ferent sterols encountered in echinoderms. This means that sterols other than the C27 -ones are of dietary origin.

2.2. Utilization and modi[ication o[ dietary sterols Food provides echinoderms nearly exclusively with sterols of the A5 -type. Differences can be expected in the fate of these sterols depending on the consumer's own sterol type. Little is known on this subject in echinoderms with regard to A5 -sterols. The only report seems to be that of Smith (1971), who showed that Echinus esculentus was unable to transform [4- 14C] -cholesterol into other sterols or two metabolize [4.l 4C] -stigmast-5-en-3ß-ol. No evidence was obtained of dealkylation at C-24 to give C27-sterols. These observations need to be extended to other species of echinoids, and also to ophiuroids and crinoids. For the moment it seems that in species with A5 -sterols, the dietary sterols are used as they are and are added to the sterol pool. Whether the resorption of dietary sterols occurs at random or selectively is unknown, but obviously a selective resorption would extend greatly the pos­ sibilities of an animal to produce its own specific sterol pattern. Since the sterol pool is also supplied with C27-sterols synthesized de novo, the predominance of C27-sterols in echi­ noids and ophiuroids is easily understood. The fate of dietary sterols in asteroids has been investigated rather extensively. Fager­ lund & Idler (1960) initiated these investigations and showed that Pisaster ochraceus con­ verted ingested [4.l 4C]-cholesterol into cholest-7-en-3ß-ol. This conversion was so efficient that Fagerlund (1969) suggested that this so-called bioconversion would be the sole source of sterols in asteroids, and assumed that they convert the bulk of exogenous A5 -sterols into the corresponding A7 -sterols. Bioconversion was confirmed for Asterias rubens and Solaster papposus by Smith & Goad (1971). They showed that 41 hours after the injection of [4­ 14C]-cholesterol inA.rubens some cholest-7-en-3ß-ol had been formed, but that the bulk of radioactivity was present in 5a-cholestan-3ß-ol. In an experiment lasting for seven days the major radioactive component was still 5a-cholestan-3ß-ol, but cholest-7-en-3ß-ol also had been formed in appreciable amounts. Incubations of S.papposus led to similar results (Smith & Goad 1971).1t seemed possible that 5a-cholestanol is an intermediate in the bio­

Steroid metabolism 425

Figure 4. Probable route(s) in asteroids for the conversion of dietary 6 5 -sterols into the corresponding

67 -sterols.

conversion of cholesterol. This was confirmed by incubating A.rubens with [4_ 14C] -cholesta­ nol for three days: 18 % of the radioactivity was recovered in cholest-7-en-3ß-ol. Thus, bio­ conversion may follow the route 6 5 -+ 60 -+ 6 7. In mammals, the reduction of the 6 5 -bond in 3ß-hydroxysteroids is performed stepwise, in that the 6 5 -3ß-hydroxysteroid is first oxidized to a 6 5 -3-oxosteroid, which is isomerized to the 6 4 -3-oxosteroid followed by reduction to the 6 0 -3ß-hydroxysteroid. Therefore, Smith et al. (1972) investigated the intermediacy of 3-oxosteroids in the bioconversion of choles­ terol in A. rubens and Porania pulvillus. They first demonstrated that [4)4C] "cholest-4-en­ 3-one was converted into 5a-cholest-7-en-3ß-ol. The obligatory intermediacy was shown by the injection of [4- 14C,3a- 3 H] -cholesterol into A.rubens. From the greatly decreased 3H/ 14C ratio in the isolated cholestanol they concluded that the formation of cholestanol from cholesterol proceeds predominantly through a 3-oxosteroid intermediate, though a direct reduction of cholesterol to cholestanol could not be excluded (figA). Theoretically, bioconversion rnight also proceed by areversal of the final steps in choles­ terol biosynthesis, thus 6 5 -+ 6 5,7 -+ 6 7. However, Frantz et al. (1964) reported that the sys­ tem 6 7 -+ 6 5,7 -+ 6 5 is irreversible in rat liver. With respect to the mechanism of bioconversion, Sheikh & Djerassi (1977) made a num­ ber of interesting observations with the holothuroid Stichopus cali[ornicus. From their experiments with 11 ,2- 3H2 f-cholesterol and 13-3 HI-cholest-7-en-3ß-ol they concluded that 'the sea cucumber can transform cholesterol to 6 7-cholestenol, and vice versa. The latter process proceeds with retention of tritium at position 3'. If indeed the label is still at C-3 and has not been removed in the meantime, this is an argument against the mechanism as proposed by Srnith et al. (1972) for asteroids. After Smith et al. (1972), Sheikh & Djerassi (1977) supposed that the conversion has proceeded via the 6 5,7-diene. Since in both incu­ bations cholestanol was radioactive, they concluded that this excludes the transformation sequence via the 3-oxosteroids. They consider a direct bio-reduction of the double bond to be a conceivable alternative.

426 Peter Voogt For comments on the use of tritium see the discussion of biosynthesis of sterols de novo (section 3.1). Voogt & Van Rheenen(1976a)injected 126- 14CI-cholest-S-en-3ß-ol intoA.rubens and proved that the label was stil1located at C-26 in the radioactive Sa-cholest-7-en-3ß-ol iso­ lated. This implies that cholesterol had not been broken down to smaller units which had been reused for the synthesis of 6. 7 -cholestenol. Smith (1971) reported thatI4- 14CI-stigmast-S-en-3ß-ol injected into A.rubens was effi­ ciently converted into Sa-stigmastan-3ß-ol and Sa-stigmast-7-en-3ß-ol. This was confirmed by Voogt & van Rheenen (1976a), who further stated that neither in well-fed nor in starved animals was any trace of dealkylation observed. Although experimental evidence for the bioconversion of 6. 5 -sterols into the correspond­ ing 6. 7 -ones is only known for C27 6.5 and C29 6. 5 , it is likely that asteroids can carry out bio­ conversion on all dietary sterols. The absence of dealkylation was not expected, because this process is well-known from

Figure 5. Sterol pattern of Asterfas rubens (upper) and Mytilus edulis (below) ascertained by gas­ chromatographie analysis; c - internal reference cholestane ; cv - internal refercnce cholesteryl valerate. la. C 26 6.5,22. 2a. C 27 6.5,22 ; 3a. C 27 6.5 ; 4a. C286.5,22; 5a. C 28 6.S oj. C286.5,24(28~; 6a. C296.5,22; 7a. C.96.5 + C296.5,24 (28); The corresponding 6.7 -sterols in Asterias rubens are indicated with b.

Steroid metabolism 427 insects and because it was reported to occur in the mollusc Patella vulgata (Collignon­ Thiennot et al. 1973). Its absence limits the possibilities of the asteroids to adjust the ste­ rols offered in the diet to its particular need at the moment. As to the use of sterols from the diet by echinoderms we should be aware that C28 ­ sterols of echinoderms characteristically have the 24a-configuration. Carnivorous species of course will be provided with this type of sterol, but herbivorous ones are confronted with the fact that algae have 24ß-alkyl sterols. Little is known on this particular subject. Rubinstein & Goad (1974) showed a 24a-methyl sterol in the diatom,Phaeodactylum tricornutum. According to Goad (1976) 'this indicates that invertebrate sterols may origi­ nate to a large extent from diatoms but obviously a rigorous examination of the sterols of more diatom species is required to validate this suggestion'.

3. THE EFFECT OF DIETARY STEROLS ON THE STEROL COMPOSITION OF ECHINODERMS An interesting example of the effect of dietary sterols on sterol composition and of applica­ tion of the ideas about bioconversion is the following: HaIe et al. (1970) elucidated the structure of gorgosterol, a sterol isolated from a gorgonian, and reported it to be a unique C30-sterol with a cyclopropane ring at C22-C23. Sheikh et al. (1971) demonstrated that acanthasterol is the b> 7 -analogue of gorgosterol. Based on the ideas of Smith & Goad (1971) about the mechanism ofbioconversion, Kanazawa et al. (1974) started a search in Acanthaster planci for the occurrence of gorgostanol, the presumed intermediate in the conversion of gorgosterol into acanthasterol and succeeded in isolating the predicted sub­ stance. As dietary sterols seem to be neither alkylated nor dealkylated by the asteroid Asterias rubens and only seem to be converted into the corresponding b>7 -sterols, we may expect a strong resemblance between the sterol pattern of this asteroid and that of its main prey the sea mussei Mytilus edulis (fig.5). The sterol composition of A.rubens fed on the usual prey Medulis, differs from that of Medulis in that its level of C27 -sterols is much higher and that of C28 -ones is much lower (Voogt, unpublished, table 1). There are no great differences in the sterol composition of the body wall and the pyloric caeca. The sterol composition of the periwinkle Littorina littorea is very different from that of Medulis; the level of C27 -sterols being about twice as high and that of C28 -sterols being only about one-fifth. When A.rubens is fed on L.littorea, the sterol composition of the body wall shows only minimal changes, whereas that of the pyloric caeca shows a continuous increase in C27-sterols and a continuous decrease in C28 ­ sterols. Thus, changes in the composition of dietary sterols are reflected much stronger in the sterol composition of the pyloric caeca than in that of the body wall. These data sug­ gest that the level of C28 -sterols in the body wall has been kept rather constant at the expense of that in the pyloric caeca. The decrease of the C26 -sterols in the pyloric caeca exceeds that of these sterols in the body wall, but this also might be explained by assuming a rather stable pool of sterols with a low turnover in the body wall and a much more dyna­ mic pool with high turnover rates in the pyloric caeca. The sterols in the body wall will be located mainly in membranes, and will be involved in regulating the microviscosity of these membranes, and will be involved in regulating the microviscosity of these membranes. Therefore, it may be important to keep their composition as constant as possible. On the

5.0 42.6 50.3 2.2

Mytilus edulis

Diet

1.0 83.2 10.6 5.2

Littorina Iittorea

1.7 75.6 20.5 2.3

2.0 81.0 14.9 2.1

1 week 1.2 78.6 17.4 3.0

2 weeks

Fed on Littorina Iittorea 0.9 82.6 14.6 1.8

3 weeks 1.6 66.7 29.6 2.2

Fed on Mytilus edulis 1.2 78.6 17.4 2.8

1 week 0.4 81.4 14.3 3.7

2 weeks

Fed on Littorina Iittorea 0.4 89.3 8.2 2.1

3 weeks

~ ~ ....

Fedon Mytilus edulis

C28 L:.5, 111 (28 \

C29 l:.5,24 (28)

C29 l:.S,22 C29 L:.S

1.9 0.8

10.0

39.5

3.0 2.0

7.4

66.3 15.3 4.6

C 26 L:.7,22 C 27 l:.7,22 C 27 L:.7

4.4 12.2 31.3

0.7

C26'~S,22

C27 l:.5,22 C27 L:.S C27 L:.5,24 C28 l:.5,22 C28 L:.s+ 10.9

C 28 l:.7,22 14.5 C 28 L:.7+ C 28L:.7,24(28) 6.0 C 29 l:.1,22 1.1 C 29 l:.7 1.2 0.9 1.2

4.0

2.0 18.1 62.9

1.7 13.8 61.3

1 week

1.0 1.8

6.0

11.4

1.2 14.1 64.5

2 weeks

3 weeks

1.2 0.2

4.2

10.4

0.9 14.5 68.1

1.1 1.3

8.1

21.4

1.5 13.0 54.0

Fedon Mytilus edulis

Fed on Mytilus edulis Fed on Littorina Iittorea

Pyloric caeca

Type of sterol Body wall

Asterias rubens

Littorina Iittorea

Diet

1.0 1.8

4.8

12.6

1.0 15.6 63.0

1 week

3.0 0.7

5.8

8.5

0.4 9.2 72.2

2 weeks

Fed on Littorina Iittorea

1.0 1.1

3.0

5.2

0.4 7.8 81.5

3 weeks

~

Body wall Pyloric caeca

~ ....

Type of sterol

Mytilus edulis

~

00

N

Asterias rubens

oe Asterias rubens, fed on Mytilus edulis or Littorina Iittorea

Table 2. Sterol composition (in percent of total sterols) in the pyloric caeca and the body wall of Asterias rubens, fed on Mytilus edulis or Littorina Iittorea (Voogt unpublished)

C26 C27 C28 C29

Type of sterol

Table 1. Relative sterol composition, according to carbon level, in the pyloric caeca and the body wall (Voogt unpublished)

Steroid metabolism 429 other hand, the pyloric caeca may store dietary sterols and thus reflect rather strongly changes in the diet. This is even more pronounced when the individual sterols are deter­ mined (Voogt, unpublished, table 2). The sterol composition in the body wall of A.rubens shows a decrease of C27 -sterols, particularly of C27 .67,22, and an increase of the C28 - and C29 -sterols with starvation (Voogt unpublished, table 3). This increase can only be a relative one, caused by an absolute de­ crease of C27 -sterols. Because the ratios of increase are different for each sterol, a different turnover rate for each sterol is indicated. In the pyloric caeca, there is a decrease of C27 .67,22 too, but the level of C27 .67 is somewhat higher. The total of C27 -sterols in the pyloric caeca shows a slight decrease leading to an increase of the other sterols. However, C28 .67,22 is nearly constant, which may imply that this sterol is metabolized rapidly too or is transferred from the pyloric caeca to the body wall. The sterol composition of the body wall and pylo­ ric caeca of A. rubens fed on haddock (of which the sterols are supposed to consist nearly exclusively of cholesterol) strongly resembles that of starved ones. After five weeks of star­ vation there seems to be areturn to more normal values in the body wall in that C27 -sterols increase and the C28 - and C29 -sterols decrease (Voogt, unpublished, table 4). This tendency is continued after six weeks. The same tendency can be observed in the pyloric caeca after three weeks of starvation. This means that there is a continuous increase of C27-sterols and a decrease of C28 - and C29 -sterols which leads to a sterol composition in this organ quite different from that in fed animals. However, the sterol in this organ is quite different from that in fed animals. However, the sterol composition in the pyloric caeca after five weeks of starvation strongly resembles that of the normal composition in the body wall. This might be explained by assuming that the pyloric caeca have metabolized or released all its stored reserve sterols and that only the sterols incorporated in the pyloric caeca membranes are left, which of course will resemble those in the membranes of the body wall. The foregoing studies have shown that the diet may have a profound influence on the sterol composition of asteroids. This composition will be the result of the contributions from de novo synthesis, the diet, and the removal of sterols from the sterol pool. 4. ELIMINATION OF STEROLS In principle, three ways can be suggested for the removal of sterols from the sterol pool, viz. excretion of sterols as such, breakdown of sterols, and derivative formation. No perti­ nent data are available with reference to the excretion of sterols from echinoderms, but it is well-known that coelomocytes, which have a relatively high level of cholest-5-en-3ß-ol, leave the animal via the tube-feet. This may lead to considerable but as yet non-quantifiable losses of sterols. Little is known about the catabolism of sterols in echinoderms. Voogt & Schoenmakers (1973) found some indications that cholest-5-en-3ß-ol and stigmast-5-en-3ß-ol (sitosterol) injected into Asterias rubens were broken down to a considerable extent. Voogt & van Rheenen (1976a) confirmed this, reporting that after the injection of 126.1 4CI-cholest-5-en­ 3ß-ol and 14- 14CI-stigmast-5-en-3ß-ol, respectively 9,43 % and 0.19 % of the radioactivity recovered was in the form of carbon dioxide. This suggests that the side chain particularly is subject to catabolism, which, however, does not necessarily lead to a total breakdown of the molecule, but may be necessary to produce steroids of the pregnane type. There was incorporation of radioactivity from both injected precursors into nearly all lipid classes, particularly into phospolipids (14.52 % and 8.74 % respectively), but also into fatty acids

1.1

1.6

0.2 12.8 58.5 I7.7 7.0 0.2 4.4 53.0 21.2 12.9 2.3 6.2 6.4 47.3 22.4 15.8 2.3 5.3

2.0 14.4 56.7 15.6 6.4 0.8 2.6 0.5 4.9 60.6 14.9 12.7 1.7 4.9

5.2 60.4 16.5 12.7 2.1 3.8

Fedon haddock

C26 67,22 C27 67,22 C27 67 C28 67,22 C2867'1:. C28 6 7,24(281 C29 67,22 C 29 6 7

Type of sterol

1.7 13.8 61.3 14.5 6.0 1.1 1.2

0.4 11.9 41.7 29.9 11.2 2.3 2.6 0.9 14.5 56.0 19.0 7.8 0.9 1.2

0.7 13.5 66.8 10.5 5.4 1.6 1.5

1.5 13.0 54.0 21.4 8.1 1.1 1.3

0.2 9.9 40.7 30.5 13.4 3.7 1.6

3 weeks

Time of starvation

o weeks

6 weeks

Time of starvation

o weeks 3 weeks

Pyloric caeca

Body wall 5 weeks

0.3 12.7 61.7 17.8 6.2 0.5 0.7

5 weeks

0.3 13.5 73.2 8.3 3.7 0.6 0.4

6 weeks

Table 4. Sterol composition (in percent of total sterols) in the pyloric caeca and the body wall of Asterias rubens after different times of starvation (Voogt unpub­ lished)

C26 67,22 C27 6 ?,22 C27 67 C28 67,22 C286 7+ C 2s C17,24(281 C29 67,22 C29 67

Fed on Mytilus edulis

Fed on Mytilus edulis Starved

...

~

~

~

~

Pyloric caeca

Body wall Fed on haddock

Type of sterol Starved

~ 0

Table 3. Sterol composition (in percent of total sterols) in the pyloric caeca and the body wall of Asterias rubens fed on Mytilus edulis, compared with that of asteroids fed either on haddock or starved for three weeks (Voogt unpublished)

Steroid metabolism 431 and acylglycerols. It was calculated that only 28.3 % of the cholesterol administered was recovered in the sterols of A.rubens; in the case of ß-sitosterol the corresponding value was 60.6 %. These fmdings indicate that sterols are intensively metabolized and that this pro­ cess may contribute considerably to the removal of excess sterols. However, more data are necessary to quantify this contribution. Derivative formation may playa role in the removal of sterols from the sterol pool and, if derivatization occurs selectively, can bring about changes in the sterol pattern. In this respect, the fmding of Björkman et al. (1972) that in A.rubens cholesterol sulphate was present in the rather high concentration of about 1.3 mgj g dry tissue is important. The pre­ sence ofthis high amount of a AS-sterol in a species believed to containonly A7-sterols was surprising. Goodfellow (1974) found that 30-40 % of the total sterol was in the sulphated form in A.rubens. The gonads particularly appeared to be rich in sterol sulphate. The fatty acid sterylesters made up only a few percent of the total sterol. Goodfellow & Goad (1973) found that C27-sterols constituted ab out 95 % of the sterol sulphates in A.rubens, with cholesterol constituting 45 %, and cholestanol about 11 %. This composition strongly dif­ fers from that of the free sterols in which cholestanol and cholesterol together make up about 2-3 %. The fattyacyl sterylesters showed a sterol composition similar to that of the free sterols, but the amount of cholest-7-en-3ß-ol was somewhat higher. These different sterol compositions indicate that, in fact, there are at least two separate sterol pools. Goodfellow & Goad (1973) reported that all three sterol fractions were radioactive 48 hours after the injection of 14.1 4 CI-cholesterol into A.rubens. In the free sterol fraction, most activity was still present in cholesterol (60 %), but 5a-cholestanol (33 %) and cholest­ 7-en-3ß-ol (7 %) were also radioactive. Most of the radioactivity (> 90 %) in the fattyacyl sterylesters was present in 5a-cholestanol, which only constitutes for about 1 % of the este­ rified sterols. 5a-cholestanol (55 %), was the most highly labelIed sterol in the sterol sul­ phates, followed by cholesterol (44 %), with litde activity present in cholest-7-en-3ß-ol. After incubation of A.rubens for a month with 12- 14CI-mevalonate, no radioactivity was found in cholesterol or 5a-cholestanol of the sterol sulphates, but a small amount of incor- . poration into cholest-7-en-3ß-ol had taken place. Goad (1976) therefore suggested that sterol sulphates arise from two sources: 'A proportion of any dietary cholesterol is sul­ phated unchanged by the animal, while a further portion may be reduced to 5a-cholesta­ nol, which can then be either sulphated or converted to a lesser extent into cholest-7-en­ 3ß-ol. On the other hand, sulphated cholest-7-en-3ß-ol is probably mainly derived from a ' pool of the sterol synthesized de novo'. The function of sterol sulphates, which are present in all five echinoderm classes (Good­ fellow 1974), is still unknown. In mammals, cholesterol sulphate may be aprecursor for pregmenolone sulphate, and sterol sulphates also have been suggested as excretory material (Winter & Bongiovanni 1968). It would be very interesting to know whether they have sirnilar roles in echinoderms. Interesting sterol derivatives were reported by Nomura et al. (1969a), who demonstrated the presence of A5-C 27 -, C2S - and C29 -sterol xylosides in the holothuroids Stich opus japoni­ cus and Holothuria tubulosa. Unfortunately nothing is known about their origin, but the place of the double bond and the alkyl groups at C-24 suggest that they originate from the diet. Elyakov et al. (1980) reported in Stichopus japonicus the presence of A7 -C 26 -, C27 -, C2S ­ and C29 -sterol xylosides, some of them containing in addition A22 or A24 (28) double bonds. These data suggest that all sterols present in this species can be glycosilated. They found

432 Peter Voogt

these compounds to be present also in Isostichopus badionotus and Synapta lappa. In recent years the steroid metabolism in asteroids and particularly inA.rubens has received considerable attention (e.g. Schoenmakers 1979). In contrast, hardly any atten­ tion has been paid to the other echinoderm classes in this respect. It was shown that cho­ lesterol can be metabolized to give pregnenolone, which in its turn is converted into pro­ gesterone. Progesterone can undergo a variety of reactions, all belonging to two types, viz.: hydroxylation and reduction. In this way 20a-hydroxyprogesterone, 21-hydroxyprogeste­ rone (11-desoxycorticosterone), and 17a-hydroxyprogesterone are produced. By the action of a Sa-reductase, Sa-pregnane-3,20-dione is formed. Progesterone is also metabolized, pro­ bably via 17a-hydroxyprogesterone into androstenedione, testosterone and Sa-androstane­ 3,17-dione. No evidence has been obtained as yet for the biosynthesis of estrogens, though the interconversion ofestradiol-17(3 and estrone is weIl documented. Conversion percen­ tages are generally low and it seems that only small amounts of sterols will be metabolized in this way. For this reason steroid metabolism in echinoderms is somewhat beyond the scope of this chapter, which deals with the sterol composition in echinoderms and the ways in which this composition is brought about. Therefore, the extensive literature on steroid metabolism will not be reviewed, nor will the interesting possibilities of steroids for repro­ duction physiology in echinoderms be discussed. Thus far, no other ways for the removal of excess sterols are known in the echinoderm classes with L).5-sterols. How these animals manage to achieve and maintain their characteris­ tic sterol pattern is stilliargely unknown and there is a strong need for investigations on the turnover of sterols in echinoderms. The holothuroids and asteroids are unique within the echinoderms in that they contain saponins called holothurins and asterosaponins, respectively. These saponins are alike in that they give a so-called aglycone (a triterpene in holothurins and asteroid in asterosapo­ nins), a varying number of sugars, and usually sulphate upon hydrolysis (for review, see Grossert 1972). A variety of holothurins has been isolated, to which gene rally trivial names have been assigned. There is still much uncertainty about their chemical structure. The point of attachment of the sulphate to the alycone has not been demonstrated with certainty. It has been found that sulphate is completely absent in the holothurins of Thelonota ananas (Kelecom et al. 1976a) and Stichopus japonicus (Kitagawa et al. 1978b). Chanley et al. (1960) reported that the sugar moiety is attached at the 3(3-position of the aglycone of holothurin A isolated from Actinopyga agassizi, which was confirmed by Kitagawa et al. (1978b) for the holothurins A and B from S.japonicus. Chanley et al. (1960) found the sugar moiety of holothurin A from A.agassizi to be a tetrasaccharide with the sequence quinovose, 3-0-methyl-glucose, glucose and xylose, the xylose being attached to the aglycone. The sugar sequence in other holothurins is stilllargely unknown. Yasumoto et al. (1967) isolated holothurin B, which upon hydrolysis gave only D-quinovose and D­ xylose, from two species of Holothuria. In contrast Kitagawa et al. (1978b) found the holothurins A and B from S.japonicus to differ only in the composition of the sugar moieties, which in both cases are hexasaccharides. Several structures have been given fOI the aglycones in holothurins. They are all chemi­ cally related with lanosterol (Roller et al. 1969), but there is still some uncertainty about the stereochemistry at C-20 (Grossert 1972). The uncertainty about the proposed struc­ tures has increased, because there is some evidence that the aglycone(s) isolated after methanolic acid hydro lysis may differ from that of the triterpenic moiety in the intact

Steroid metabolism 433 holothurin (Premuzic 1971), because artifacts (e.g. the introduction of double bonds or formation of methoxyl groups) may arise during the hydrolysis (Kelecom et al. 1976b). Kelecom et al. (1976a) stated that they had isolated two genuine aglycones from the holo­ thurins of Thelonota ananas and assigned them the structures 23~-acetoxy [',.8-holostene­ 3ß-ol and 23~-acetoxy [',.8,25_holostadiene-3ß-ol. Kitagawa et al. (1978b) claimed to have isolated the genuine aglycones, called holotoxigenols, from the holothurins A and B of S. japonicus. Kitagawa et al. (1978c) also elucidated the structure of holothurin B from Holo­ thuria leucospilota, although uncertainty remained about the configuration at C-22. little data are available with respect to the biosynthesis of holothurins. Elyakov et al. (1975) showed that S.japonicus is able to synthesize saponins from 11,2- 14CI-acetate. The radioactivity was primarily (80-90 %) present in the aglycones, which were lanostane-like; over 80 % of the latter radioactivity being associated with one aglycone, viz. stichopogenin A4 • Similar data were obtained by Kelecom et al. (1976b). They observed slight incorpora­ tion of radioactivity (5 x 10-2 %) into the holothurins of T.ananas three days after injection with 11.1 4CI-acetate. The holothurins A and B were nearly exclusively labelIed in the agly­ cone moiety. Because the sugars did not contribute to the radioactivity of the saponins, they suggested that these may be of exogenous origin. This is rather obvious, for sugars in animals in general originate from the diet, and glyco(neo )genetic capacity is usually re­ stricted to a single organ. Sheikh & Djerassi (1976) observed that 13-3HI-Ianosterol was incorporated into the holo­ thurogenin (= holotoxinogenin) of Stichopus californicus 300 times better than 3H-acetate. They showed that the radiolabel, in the holothurogenins synthesized from 13- 3HI-Ianosterol was still at C-3. Therefore they suggested that 'it is likely that in nature holotoxin originates in part from dietary lanosterol or a closely related precursor derived from lanosterol'. In this respect it is important that the presence of lanosterol in holothuroids has been shown (Nomura et al. 1969a). However, this triterpenoid is also an intermediate in the biosynthe­ sis of sterols, which may imply a competition for lanosterol between the synthesis of sterols and that of holothurogenins. It is gene rally supposed that holothurins act as repellants and have a protective function. This is partly based on the non-uniform distribution of these substances within the animal. High concentrations of holothurins are present in the Cuvierian tubules, which are extruded when those holothuroids which possess them are disturbed. The holothurins are highly toxic and even low concentrations may be very effective deterents. Matsuno & Ishida (1969) found holothurins to be present in all the tissues of H.leuco­ spilota. High concentrations, which showed seasonal fluctuations, were present in the Cuvierian tubules and ovaries. The increase of holothurin concentration in the ovaries prior to the spawning and the low concentration in the testes at the same time suggest that holo­ thurins are involved in reproduction (see also the discussion below on asterosaponins). The asterosaponins of asteroids have many properties in common with the holothurins. They are toxic to a variety of animals, among which fishes, lethal at high concentrations (Hashimoto & Yasumoto 1960, Rio et al. 1965) and invoking intensive avoidance reactions even at low concentration in the whelk Buccinum undatum (Mackie et al. 1968). The che­ mical structure of asterosaponins is analogous to, but is better defmed than, that of the holothurins. On acid hydrolysis, they give a steroidal aglycone, sugars, and sulphate. The sulphate group is always at the 3ß-position, while the sugar moiety is always at the 00­ position of the aglycone. The sugar moiety in asterosaponin A of Asterias amurensis is a tetrasaccharide too, containing two moleeules each of D-quinovose and D-fucose (Yasu­

434 Peter Voogt moto & Hashimoto 1965). In asterosaponin B from the same species, the sugar moiety is a pentasaccharide consisting of two molecules D-quinovose and one molecule each of D­ fucose, D-xylose and D-galactose (Yasumoto & Hashimoto 1967). The saponin of Marthas­ terias glacialis contained a tetrasaccharide consisting of two molecules D-quinovose and one molecule each of D-glucose and D-fucose (Mackie & Turner 1970). Recently, Kitagawa & Kobayashi (1977, 1978) showed that the sugar moiety of thornasterol A, the major saponin of Acanthaster plan ci, is a pentasaccharide consisting of two molecules D-quino­ vose and one molecule each of D-fQcose, D-xylose and D-galactose. This sugar composition is identical with that of asterosaponin B (Yasumoto & Hashimoto 1967), mentioned above. Ikegami et al. (1979) isolated from A.amurensis a saponin, called B2 , differing from saponin Bin that another molecule of D-quinovose is present in place of the molecule D-fucose. Because little data on the sugar composition in saponins are available, it is impossible to draw conclusions, but data suggest that there may be only little variation in this composi­ tion. A great variety of aglycones has been identified (for review, see Goad 1976). They gene­ rally are based on the cholestane or the pregnane skeleton and most of them have a double band at C-9( 11). Many of the structures proposed have become uncertain, since in recent years some evidence was obtained that the aglycones isolated after acid hydro lysis might not represent the true structure of the steroid in the intact saponin. Kitagawa et al. (1975) isolated thornasterol A and B from A.planci, the former being a C2,-aglycone, whereas the latter possessed an additional methyl group at C-24. Their structures are 3ß,6a,20-jrihydroxy-5a-cholesta-9(1 1)-en-23-one and 3ß,6a-20-trihydroxy­ 24-methyl-5a-cholesta-9(I 1)-en-23-one respectively. Acid treatment of thornasterol pro­ duced 3ß,6a-dihydroxy-5a-pregn-9(ll )-en-20-one and 3ß-hydroxy-5a-cholesta-9(11 ),20(22)­ dien-23-one. Therefore Kitagawaetal. (1975) considered the 20-hydroxy steroids to be the true sapogenols, whereas 3ß,6a-dihydroxy-5a-pregn-9(1l)-en-20-one may be an artifact produced by the hydrolysis as had been supposed be fore by Shimuzu (1972). This was proved (Kitagawa et al. 1978a) in an elegant way by subjecting thornasterol to the enzymes (glycosidases) of Charonia lampas, the natural enemy of A.planci. This yielded the genuine sapogenols with the structures given before, whereas the pregnene-like type was completely absent. Ikegami et al. (1979) elucidated the structure of the aglycone of saponin B2 from A.amurensis, which turned out to be identical with that of thornasterol A (Kitagawa et al. 1975). Minale et al. (1978, 1979) and Oe Simone et al.(1979a,b) determined the structure of several aglycones from Echinaster sepositus and Astropecten aranciacus and M.glacialis, respectively. The proposed structures are different from those found by Kitagawa et al. (1978a) and Ikegami et al. (1979). Due to the uncertainty of the natural occurrence of some of them, they will nbt be discussed further. Kitagawa & Kobayashi (1978) proposed the structure of thornasterol A to be: 20~­ hydroxy -6a-O-[ß- D-fucopyranosyl( 1~2 )-ß-D-galactopyranosyl( 1~4)- [ß-D-quinovopyranosyl (I ~2)]-ß-D-xylopyranosyl(1 ~3)-ß-D-quinovopyranosyl]-5a-cholest-9(11 )-en-23-one-3ß-yl sodium sulphate. Ikegami et al. (1979) determined the total structure of saponin B2 of A. amurensis to be: 20~-hydroxy-6a-0[ß-D-quinovopyranosyl-(1 ~2)-ß-D-galactopyranosyl (1~4)-[ß.-D-quinovopyranosyl(I~2)]-ß-D-xylopyranosyl(I~3)-ß-D-quinovopyranosyl]-5a­

cholest-9(1 l)-en-23-one-3ß-yl sulphate. The sirnilarity between the two structures proposed is obvious. There is a great need for studies on the biosynthesis of asterosaponins as weIl with res­ pect to the site of production as to the pathway followed and the precursor(s) used. The

Steroid metabolism 435 presence of a methyl group at C-24 in thomasterol B suggests that the aglycones at least partly may be of dietary origin. Teshima et al. (1977) injected 14-14CI-cho\esterol into the body cavity of A.rubens to study the cholesterol metabolism, and isolated 5a-cholestane­ 3ß,oo-diol both in free and in 3ß-sulphated form among many other products. These com­ ponents are probably intermediates in the asterosaponin biosynthesis, on the basis of their structure, which confirms the use of dietary sterols for aglycone synthesis. However, Mackie et al. (1977) found a low incorporation of radioactivity into saponins from 14- 14CI-choles­ terol and a higher one from 12- 14CI-mevalonate in Marthasterias glacialis. This finding sug­ gests that aglycone may originate both from dietary sterols and from sterols synthesized de novo, the latter source perhaps being the most prevalent. This is in good agreement with results of Goad et al. (1972), who postulated that the saponin synthesis might branch off from the sterol synthesis at the 4,4-dimethyl sterollevel. This recalls the precursor supposed for the synthesis of holothurogenin. Whichever the case, synthesis of asterosapogenin will either interfere with sterol synthesis or withdraw sterols from the sterol pool and, in both cases, may influence the sterol pattern. Asterosaponins are generally considered to be offensive or defensive weapons and there is some evidence that saponins may be released more or less continuously from the tube­ feet, with the result that the presence or approach of an asteroid is noticed at some distance by several marine animals. This release may lead to considerable losses of saponins, which implies that the biosynthesis may make high demands upon the asteroid. Mackie et al. (1977) recently suggested that the asterosaponins present in the digestive system might serve to solubilize cholesterol from the prey, thus facilitating its absorption. Yasumoto et al. (1966) found saponin in all the tissues of A.amurensis. High concentra­ tions were found in the stornach and the gonads. Unfortunately, no distinction was made between testes and ovaries, which makes comparison with the holothurin distribution im­ possible. The concentration of asterosaponins was lowest in winter, the spawning season. This is the reverse of that in holothuroids as reported by Matsuno & Ishida (1969). It was supposed by Ikegami et al. (1972) that high concentrations of saponins may prevent spawn­ ing. More experimental evidence was given by Ikegami (1976). In contrast to the findings of Yasumoto, Mackie et al. (1977) observed high concentra­ tions of asterosaponins in the spawning season in Mglacialis. Voogt & Huiskamp (1979) found asterosaponins to be hardly detectable in the testes of A.rubens, whereas they were always present in the ovaries. The concentration in the ovaries was low during summer and increased until spawning. At that time concentrations up to 25 mg/g fresh weight were reached. Voogt & Van Rheenen (1979) demonstrated that ovarian saponins are located within the oocytes, and that they were still present even in the shedded oocytes. The con­ centration did not decrease significantly in the subsequent stages following fertilization, and asterosaponins were present at high concentrations even in the free swimming larvae. Inhibition of spawning by saponins in A.rubens therefore seems highly improbable. This also holds for the inhibiting action of asterosaponins on the meiotic maturation of asteroid oocytes as reported by Ikegami et al. (1976). Voogt & Van Rheenen (1979) suggested that saponins are stored in the oocytes only to provide the embryos with these substances as protection against predation. Lucas et al. (1979) also reported that asterosaponins are pre­ sent in the eggs and larvae of A.p/anci in about the same concentrations. They gave experi­ mental evidence that a number of planktivorous fish avoid the larvae of this asteroid. Three species of fish were able to discriminate significantly food particles containing saponins at concentrations of 1 x 10- 7 parts per fresh weight. The real concentration in eggs and larvae

436 Peter Voogt of A.planci is about four magnitudes higher, which is in the same order as that observed by Voogt & Van ~eenen(1979) in A.rubens. Great amounts of saponins (for an animal of 100 gabout 800 mg) are lost with spawn­ ing and will have to be replaced in the next reproductive cycle. This cyclic event will have consequences for the asterosaponin metabolism in female asteroids. Therefore, studies are badly needed which compare asterosaponin metabolism in female and male asteroids with respect to the.implications on sterol metabolism.

5. CONCLUSIONS The sterol pattern of an animal is the result of an equilibrium between the processes which provide sterols, viz. de novo synthesis and absorption of dietary sterols, and the processes which eliminate sterols. The sterol pattern is adynamie state influenced by several factors. These factors have been reviewed primarily in a qualitative way as quantitative data are largely absent. Thus we know which factors affect sterol composition, but we do not know to what extent and in what way they interact to produce a specific sterol pattern. Studies to quantify the effects are needed to get a better insight and are strongly recommended.

19

MICHEL JANGOUX

EXCRETION

Echinoderms differ from most macro-invertebrates by lacking morphologically differen­ tiated excretory organs, although several internal organs have been considered renal organs as the result of morphological and experimental studies (Burian 1924). Echinoderms are gene rally considered ammonotelic, but they also produce a small amount of urea. Soluble nitrogenous wastes simply diffuse into the external medium through the respiratory sur­ faces (Delaunay 1931). Excretion is the elimination of metabolie wastes, essentially nitrogenous products which_ originate from the utilization of dietary proteins and from the breakdown and turnover of body cell constituents (pro teins and nuc1eic acids). Some authors believe that the reaction of coelomocytes of echinoderms to foreign particles or bodies (phagocytosis or encyst­ ment which may be followed by release of the cells into the external medium) proves the excretory function of these cells. However, this cannot be interpreted as true excretion, but only as a defensive mechanism against foreign material. 1. NITROGENOUS WASTES OF ECHINODERMS Various nitrogenous wastes have been found in extracts of organs and of whole animais, in the coelomic fluid, or in the surrounding sea-water.

1.1. Amino acids The occurrence of relatively high levels of free amino acids in the internal organs or the coe­ lomic fluid is not of excretory significance, but loss ofamino acids into the surrounding sea­ water does occur. According to Delaunay (1926a, 1931). 17-28 % of the total nitrogen excreted by Asterias rubens, Paracentrotus lividus and Holothuria tubulosa is in the amine form. Amino nitrogen represents 26-29 % of the excreted nitrogen in Diadema antillarum (Lewis 1967). In contrast, Diehl & Lawrence (1979) were unable to detect amino acid ex­ cretion in Luidia clathrata.

1.2. Ammonia Ammonia excretion predominates among echinoderms (Delaunay 1928 through 1934). Propp (1977b) calculated that ammonia constitutes 80 to 85 % of the nitrogen excreted by Strongylocentrotus intermedius, Strongylocentrotus nudus, Patiria pectini[era and Aphelasterias japonica. The highest concentration of ammonia in internaiorgans occurs in the digestive organs (Delaunay 1926a, 1931: Asterias rubens, Paracentrotus lividus and Holothuria tubulosa,- Lewis 1967: Diadema antillarum; Fechter 1973a: P.lividus). Levels of ammonia in the coelomic fluid of echinoderms are less than 1 mg % (table 1). 437

...

-

---_ .. _.

(-)

0.40 1.00 0.50-2.00

(-)

x x

(+)

(-)

0.09

0.16

x x

(-) 0.13 x

x 0.12 x

x x

0.27

x x

traces

x x

0.03

(-) x x

x

0.21

x x x x

0.07 0.03

x

(+)

(-)

x (-)

X

Creatinine-N

1.0-1.3 0.12 0.16 1.98-2.51 3.27-3.61 0.92

(+) Present but unquantified; (-) Undetected; x Not tested

ASTEROIDEA

Asterias jorbesi Asterios rubens Pisaster ochraceus Pycnopodia helianthoides

0.08

x x

0.24 0.65

PSammechinus microtuberculatus Sphaerechinus granularis Strongylocentrotus jranciscanus

Id. Id.

x

3.10-4.28

(-)

x

(-)

(-)

x x

(-)

ECHINOIDEA 0.42-1.48

(+)

(-)

Diademo antillarum Arbacia Iixula Arbacia punctulata Paracentrotus lividus

0.14

Creatine-N

x x x

Uric acid-N

x (-)

0.783-1.051 0.07

Urea-:-I

(-)

x

Ammonia-N

Sclerodactyla briareus

Id.

Holothuria tubulosa

HOLOTHUROIDEA

-_._~

Species

Table 1. Nitrogenous wastes in coelomic fluid (mg/IOO ml)

Van der Heyde 1922, 1923b Delaunay 1931 Myers 1920 Myers 1920

Lewis 1967 Sanzo 1907 Van der Heyde 1922, 1923b Mourson & Schlagdenhauffen 1882 Delaunay 1931 Fechter 1973a Sanzo 1907 Sanzo 1907 Myers 1920

Sanzo 1907 Delaunay 1931 Van der Heyde 1922, 1923b

References

~

w

~

~

­ ~ ~

~

11>

;:r.

00

Excretion 439 The rate of ammonia excretion has been estimated for only a few echinoderms (table 2). The excretion rate changes with nutrition. Ammonia nitrogen level of the fluid of D.antil­ larum markedly increase when fed on a protein diet (Lewis 1967). Diehl & Lawrence (1979) reported an increase of ammonia excretion in Luidia clathrata with starvation and with maintenance level diets of prey with a high protein level (50-80 % protein by dry weight). They suggested that the increase of excretion rate was probably due to tissue catabolism. Although total nitrogen was measured rather than ammonia, Fuji (1967) showed that S. intermedius fed on a non-protein diet (starch and agar-agar) excreted from 0.2-1.3 mg nitrogen/day depending on the size of the animal (30-80 mm test H.D.). As indicated, the largest part of this nitrogenous material was undoubtedly ammonia. The release of ammonia changes with the process of intracellular isosmotic regulation. Changes in the intracellular levels of free amino acids associated with isosmotic intracellular regulation have been reported for Strongylocentrotus droebachiensis, A.rubens andL. clathrata (Lange 1964, Jeuniaux et al. 1962, Binyon 1972b, Ellington & Lawrence 1974a, respectively). An increase in ammonia excretion associated with hyposmotic stress has been reported for Eupentacta quinquesemita, S.droebachiensis and L.clathrata (Emerson 1969 and Ellington & Lawrence 1974a, respectively). The increase in the rate of ammonia excre­ tion of hyposmotically stressed animals has been interpreted by these workers as catabo­ lism of amino acids resulting in ammonia excretion during intracellular isosmotic regula­ tion.

1.3. Urea Mourson & Schladgdenhauffen (1882) were the first to detect urea in the coelomic fluid (table 1). Excretion of urea by Asterias rubens into the surrounding sea water was observed by Fosse (1913). Urea has been detected in the axial organ and gut of Paracentrotus Zivi­ dus (Fechter 1973a) and the pyloric caeca and gonads of Asterina pectinifera (Okabe & Noma 1974a). Urea is excreted in very low levels (Delaunay 1931), and sometimes cannot be detected (Lewis 1967: Diadema setosum). Webb et al. (1977) calculated that 17 % of the nitrogen excreted by Holothuria atra is urea. Diehl & Lawrence (1979) found that the excretion rate ofurea by Luidia clathrata was not affected by nutrition and always remained low. They suggested that urea synthesis may result from nutritional pathways, not excretory ones. In contrast, Fechter (1973a) reported that P.lividus is mainly ureotelic, more than 90 % of its nitrogenous wastes being urea. It is unlikely that nitrogenous excretion in this echinoid should be so different from that of other echinoderms.

1.4. Uric acid Griffiths (1888: Asterias rubens) and Van der Heyde (l922, 1923b: Asterias forbesi, Arba­ cia lixula and SclerodtIctyla briareus) claimed that uric acid is the main njtrogenous waste eliminated by echinoderms. According to them, uric acid is produced primarily by the digestive organs. It is now established that uric acid does not occur in echinoderms or occurs only in very low levels. Sulima (l913) was unable to fmd uric acid in total extracts of Holothuria tubulosa or Stichopus regaZis. Przylecki (1926) detected traces of uric acid in tissues of Echinus sp., Marthasterias glaciaZis, Ophiura sp. and Antedon bifida. He stated that the uric acid could only originate from the degradation of purine bases. Delaunay (l926a, 1931) reported that the fluids of A.rubens and H. tubulosa are free of uric acid, and that Paracentrotuslividus contains only traces (table 1).

440 Michel Jangoux Table 2. Rates of ammonia-nitrogen excretion by echinoderms Species

HOLOTHUROIDEA

Eupentacta quinquesemita HolothurÜl atra

ECHINOIDEA

DÜldema antillarum Strongylocentrotus droebachiensis Id.

ASTE ROIDEA

Luidia clathraJa Id.

1. ng NH4+ ­ N/g dry wt per hr 2. ng NH 4+ - N/g wet wt per hr

Excretion rates

References

11.5 1 90 2

Emerson 1969 Webb et al. 1977

99 3 2.9 1 0.43-0.76 4

Lewis 1967 Emerson 1969 Propp 1977a

7.5-12 1 5.2-10.3 1

Ellington & Lawrence 1974a Diehl & Lawrence 1979

3. ng NH4+ ­ N/animal per hr 4. J.Lg NH4+ - N/g dry wt per hr

Bordas (1899) and VanderH~yde (1922) reported the presence of crystalline forma­ tions (sl!pposedly uric acid) in the respiratory trees of H.tubulosa and S.briareus. Lauga & Lecal (1967) described crystals of uric acid in the coelomic fluid offour species of holo­ thuroids. These crystals are of various shapes and supposedly would be excreted through the respiratory trees. It would be undoubtedly interesting to investigate more thoroughly the details of the chemical nature of these coelomic fluid components.

1.5. Other compounds Creatine and creatinine were found in the coelomic fluid of Pisaster ochraceus (Myers 1920) andParacentrotus lividus (Fechter 1973a) (table 1). Phosphorous excretion was measured in a few species of echinoids and asteroids, and used as an indication of catabolic activity by Moore & McPherson (1965) and Propp (1977a), but there is no real information on the physiological significance of phosphorus excretion in echinoderms.

1.6. Enzymes involved in excretory processes

Pyloric caeca of Asterias forbesi show considerable but rather variable arginase activity

(Hanlon 1975). Urease activity was reported in pyloric caeca of Asterias forbesi (Hanlon

1975) and in total extracts of Lytechinus variegatus (Loest 1979). The presence of xan­

thine oxidase was demonstrated in total extracts of Echinus sp. (Przylecki 1926) as weil as

in the gut of Anthocidaris sp. (Nagase 1962). According to Przylecki (1926), Echinus sp.

and Marthasterias glacialis possess uricase and are able to transform uric acid to allantoin.

Fosse & Brunel (1929) reported that allantoinase occurs in the tissues of Ophiura tex tu­

rata and of two unidentified species of asteroid and echinoid.

1.7. Conclusions

Data related to nitrogenous excretion are still very scattered among the echinoderms.

There is only partial information on a very few species. It is quite evident that echinoderms

are ammonotelic. With few exceptions (see Fechter 1973a), most authors agree that urea

does not result from excretory pathways. The excretion of amino acids is a rather peculiar

phenomenon which still must be investigated carefuily. The .excretory mechanisms and

Figure I. Anal excretion of chlorophenal red following intracoelomic injection of the dy e.

Excretion 441 metabolie pathways are almost completely unknown. The levels of activity of the various enzymes involved in catabolism of nitrogenous compounds needs to be investigated in the various organs of more groups of echinoderms, and their possible nutrition al function must be considered. Physiological or ecophysiological works are greatly needed to ascertain the effect of nutritional and reproductive state, age, and of environmental variables such as temperature and salinity. 2. EXCRETORY SITES OF METABOLIC END PRODUCTS According to Delaunay (1931) ammonia simply diffuses into the external medium through the respiratory surfaces (respiratory trees of holothuroids, gills of echinoids, papulae of asteroids, bursae of ophiuroids). While studying water exchanges between the external medium and body cavities in Echinus esculentus, Fechter (1972) observed an active water outflux from the general coelom to the exterior through the hind gut, and claimed that excreted water could carry away nitrogenous wastes. Lawrence (unpublished) observed active water expulsion from the anus of Diadema setosum that had been placed in hypoto­ nie sea-water. Active water expulsion from the anus of asteroids (e.g. Oreaster reticulatus, Porania pulvillus and Asterias rubens) has also been observed (Tennent & Keiller 1911, Gemmill1915 and Jangoux 1976, respectively). Fechter(1973b) suggested that analcone rotation in diadematid echinoids would improve evacuation of soluble wastes by pro­ moting their elimination by the hind gut. Cuenot (1901) injected a diffusible dye, methyl green, into the coelom of several echi­ noderm species, and observed selective accumulation of the dye in the digestive cells of the pyloric caeca of Marthasterias glacialis or in those of the second loop of the gut of E. esculentus and Paracentrotus lividus. Dyes were gradually released into the gut lumen. Working with A.rubens with chlorophenol red, Jangoux et al. (1975) and Jangoux (1976b) confirmed these observations. They also showed that chlorophenol red is actively concen­ trated by the rectal caeca, but not in the pyloric caeca. Excretion of chlorophenol red by the rectal caeca is very rapid. The caeca contract rather strongly and drive the stain into the lumen and then into the external medium through the anus (fig.l). The ability of the vertebrate kidney to actively excrete diffusible dyes of the sulfophe­ nolphthalein group (phenol red, chlorophenol red) is weil known (Gerard & Cordier 1934, Sottiurai & Longley 1970). The physiological meaning of this mechanism in echinoderms is still unknown, but it clearly shows that A.rubens can excrete through its rectal caeca some rather complex soluble compounds present in the coelornic fluid. The identification of com­ pounds naturally excreted by this mechanism must still be made. As ammonia and urea are highly diffusible, it is improbable that they are excreted by this process other than being eliminated along with the gut fluid. Excretion by the gut of echinoids of highly soluble compounds such as estradiol-3-sulfate has been suggested by Creange & Szego (1967: Strongylocentrotus [ranciscanus). They suggested that such a mechanism could be of some importance in excretion of pheromones into the surrounding medium. This is certainly worth investigating with asteroids and echinoids. 3. EXCRETION AND THE COELOMOCYTES One of the main functions of the coelomocytes (at least of hyaline coelomocytes, some­ times also called ameobocytes, leucocytes, lymphocytes or phagocytes) is to maintain

442 Michel Jangoux animal integrity by neutralizing by phagocytosis or encysting unwanted material in the body cavities (Hilgard & Phillips 1968, Johnson 1969, Jenkin 1976). It has never been demonstrated that coelomocytes play any role in the elimination of nitrogenous wastes by echinoderms (Delaunay 1931, Endean 1966).

3.1. Intracoelomic injection olloreign particles The phagocytosis of particulate substances (e.g. carmine particles, colloidal carbon, Trypan blue) by coelomocytes frequently has been interpreted as proving that these cells are able to accumulate and translocate 'insoluble wastes' and thus have a renal function (e.g. Kowa­ levsky 1889, Durharn 1891, Chapeaux 1893, Schultz 1895, Cuenot 1897, 1901,Van der Heyde 1922). It should be noted that coelomic epithelial cells also are able to phagocytize foreign particles (Cuenot 1901,vanden Bossehe & Jangoux 1976). The reaction of coelo­ mocytes towards injected particles is undoubtedly the manifestation of a defensive mecha­ nism and cannot be considered a true excretory process. 3.2. Brown bodies Brown bodies are clots of degenerating coelomocytes laden with brown-yellow pigmented granules (Hetzel 1963). They are abundant in tissues and coelomic fluid of echinoids and holothuroids, but less so in ophiuroids (He~ouard 1889, Cuenot 1891b, Awerinzew 1911, Hetze11963, 1965, Massin 1978, Höbaus 1979). Brown bodies have never been observed in asteroids. According to Cuenot (1891b) and Höbaus (1979) the pigmented material repre­ sents solid wastes. Hetzel (1965) considered it 'a product, if not truly excretory, at least of no use to the holothurian'. Briot (1906) and Arvy (1957) showed that brown bodies con­ tain debris of parasitic origin. Awerinzew (1911) believed that they are partly excretory products and partly remains of plant pigments of dietary origin. It is rather improbable that brown bodies contain true metabolie wastes. If they do, the mechanism of excretion of echinoids and holothuroids, where brown bodies occur in great number, differs completely from that of asteroids, where brown bodies apparently never occur. The dietary origin of that material, as supposed by Awerinzew (1911), seems more plausible. The fact that most asteroids are not plant-eating could explain the absence of brown bodies in the species studied. We may suppose that plant material, like pigments or pigment by-products, are absorbed and transformed only slightly or not at an by digestive cells of echinoids and holothuroids. These poorly digestible substances then could be rejected and retained by coelomocytes. If this is correct, the reaction of coelomocytes here is also one towards unwanted 'substances of exogenous origin. 3.3. Elimination 01 used coelomocytes Laden coelomocytes either can be accumulated in internal organs (e.g. axial organ of echinoids, Polian vesicles of holothuroids or Tiedemann's bodies of asteroids), remain in the coelom, or be eliminated into the surrounding sea-water. Elimination of single cells supposedly occurs by cellular diapedesis through the external epithelium, the tube-feet, or the respira­ tory surfaces. Clots of laden coelomocytes are evacuated in different ways, such as autotomy of the papular tip in asteroids or of the extremity of the gills in echinoids (Cuenot 1948, Binyon 1972a), constriction of the external epithelium of tube-feet of echinoids (Höbaus 1979), and passage through the wall of respiratory trees of holothuroids (Herouard 1889, Massin 1978).

Excretion 443 4. THE SUPPOSED RENAL ORGANS OF ECHINODERMS The lack of morphologically differentiated renal organs resulted in investigators attributing a renal function to various organs. These are primarily organs which accumulate used coe­ lomocytes, and sometimes organs with structural similarities to known vertebrate or inver­ tebrate kidneys. 4.1. Axial organ of echinoids An excretory function was assigned to the echinoid axial organ by Perrier (1875), Koehler (1883) and Sarasin & Sarasin (1887,1888). Kowalevsky (1889) and Cuenot (1897,1901) injected colored particles into echinoids and observed invasion of the axial organ by laden coelomocytes. They concluded that this indicated an excretory function of the organ. There is now a general consensus that the echinoid axial organ functions as a mechanism of defense against invasion by foreign organisms (Millott 1967). More recent ultrastructural investigations confrrm that laden coelomocytes are stocked and degraded within the axial organ (Jangoux & Schaltin 1977, Bachmann & Goldschmid 1978). The fate of degraded coelomocytes remains undescribed. A high level of urea has been found in the axial organ of Paracentrotus lividus (Fechter 1973a). Whether this results from degradation of coelo­ mocytes is not known. 4.2. Axial organ of asteroids Cuenot (1901) and Pietschmann (1906) assigned a renal function to this organ. According to Bargmann & von Hehn (1968) the structural similarity of the axial organ of Asterias rubens with the glomeruli of the vertebrate kidney suggests an excretory function. This hypothesis has never received experimental confirmation. It should be noted that the axial organ of asteroids, in contrast to that of echinoids, never stocks used coelomocytes (Van­ den Bossche & Jangoux 1976).

4.3. Gills of echinoids Echinoid gills are the site of elimination of coelomocytes (Durham 1891, Binyon 1972a). Cobb & Sneddon (1977) observed laden coelomocytes accumulate in the waU of the gills and concluded this indicated an excretory function.

4.4. Tiederrumn 's bodies of asteroids and echinoids They were mainly studied in asteroids in which they were considered to be kidneys by Kowalesky (1889) as they stock carmine-Iaden coelomocytes. Svetlov (1916) described 'excretory products' inside the ambulacral epithelium of Tiedemann's bodies. Bergmann & Behrens (1964) confirmed the observations of Kowalevsky and stated that these organs are involved in a process of 'cleaning' of the ambulacral fluid. According to Bachmann & Goldschmid (1980) the Tiedemann's bodies of echinoids serve a similar function to the axial organ of echinoids.

4.5. Polian vesicles of holothuroids The internal cavity of these vesicles are frequently invaded by numerous brown bodies of various sizes (see seetion 3.2). For that reason they were considered as excretory by Jour­ dan (1883) and Baccetti & Rosati (1968).

444 Michel Jangoux 4.6. Respiratory trees 01 holothuroids These organs were supposed to be kidneys for two reasons: (a) they contain crystalline for­ mations that look like uric acid crystals (Bordas 1899a, Van der Heyde 1922), (b) they are the site of elimination of coelomocytes (Herouard 1889, 1895, Schultz 1895, Barthels 1895, Cuenot 1891b, Bertolini 1933b). 4.7. Coelomic ums olsynaptids These structures clearly accumulate laden coelomocytes and were thus considered excre­ tory (Schultz 1895, Cuenot 1902, H.L.Clark 1899a, 1907, Bertolini 1936). 4.8. Sacculi 01 crinoids Sacculi are Iittle spherical bodies which line up along the ambulacral grooves of calice, arms and pinules (Holland 1967b). As they accumulate laden coelomocytes, they were considered as excretory organs (Chadwick 1907, Reichensperger 1912, Prennant 1928). 4.9. Stone canal Hartog (1887) stated that the stone canal ('madreporic system') is morphologically and

ontogenetically a nephridium.

4.10. Gonads

According to Giard (1877), 'litde brownish concretions' appearing in the gonads of Psam­

mechinus miliaris during gonadal inactivity are excretory products. Russo (1901) and Ber­

toIini (1936) claimed that the gonad of Holothuria tubulosa has a renal function during

gonadal inactivity as it accumulates laden coelomocytes.

4.11. Digestive tract This was considered as excretory for three reasons: (a) it secretes uric acid (see section 1.4), (b) some portions are able to excrete organic soluble substances occurring in the coelornic fluid (see section 2), (c) laden coelomocytes can cross the gut, at least in holothuroids (Ber­ tolini 1934b, 1936). 4.12. Conclusions

With the exception of a few digestive organs where excretion is known to occur, the sup­ posed kidneys of echinoderms seem to have no true excretory function. Excretory func­ tion has been assigned to them from speculative interpretations or because they are sites of passage or accumulation of laden coelomocytes. The invasion of some internaIorgans by coelomocytes is still a poorly understood phenomenon. We are completely ignorant of the fate of invading cells. Undoubtedly this phenomenon offers a large and very interesting open field for future investigations.

5. EXCRETION IN EMBRYOS AND LARV AE Very few authors have investigated excretion in embryos and larvae of echinoderms. Am­ monia excretion by fertilized eggs of echinoids has been reported by Ashbel (1931 : Para­ centrotus Iividus and Sphaerechinus granularis), Örström (1941: P.lividus) and Hutchens et aL (1942: Arbacia punctulata). Embryonic excretion seems to be transitory, occurring

Excretion 445 only just after fertilization (Örström 1941, Hutehens et al. 1942). Brookbank & Whiteley (1954) were unable to deteet ammonia exeretion from early embryo to early pluteus in Strongylocentrotus purpuratus. They stated that there is no nitrogenous exeretion during the morphogenetie period leading to larval stage. Urease and several enzymes involved in the formation of urea from purines (xanthine oxidase, uriease, allantoinase, allantoiease) have been found in embryos of S.purpuratus. These enzymes supposedly funetion to add nitrogen from purine eatabolism to the meta­ bolie ammonia pool during the pre-feeding stages of development (Brookbank & Whiteley 1954). Aeeording to Runnström(l912a: P'lividus), the intestine of the pluteus eould have a renal funetion. Fenaux et al. (1980) ealeulated that 75% of the nitrogen exereted by plu­ tei of P'lividus is ammonia. Rate of ammonia exeretion in fed and unfed eulture of plutei varies from 1.1 to 1.4 natgJmg proteinjhour.

6. CONCLUSIONS The eoneept of exeretion in eehinoderms is rather eonfused, mainly as two different phy­ siologieal funetions, exeretion and eoelomoeytie defensive mechanism, have been frequently eonfused. Exeretion is the elimination of metabolie wastes, previously detoxified or not, into the external medium. Metabolie wastes in eehinoderms are ehiefly ammonia. In most species urea is present in relatively low levels. Nitrogenous exeretion is still poorly known in eehinoderms and it is presently impossible to propose a general exeretory scheme. As indieated earlier, physiologieal and eeophysiological works on these topies are greatly needed. The reaetion of eoelomoeytes towards unwanted material indieates their immunologieal eapability. It has never been proven that these eells eliminate metabolie wastes. Yet the fact that eoelomoeytes laden with unwanted material ean invade some internal organs (e.g. the eehinoid axial organ) and are degraded indicates that these organs are of great physiologieal interest. Here also more elaborate studies would be of great help in the understanding of general physiology of eehinoderms.

4. NUTRITION DURING DEVELOPMENT

20

CHARLES W. WALKER

NUTRITION OF GAMETES

Many marine organisms which employ external fertilization carry out gametogenesis within relatively simple reproductive organs. This is especially t~ue in the phylum EChinodermata where millions or hundreds of millions of gametes can be differentiated using various schedules based ultimately on an an nu al gametogenic cycle and where gonads are simple sacs of tissue without elaborate or permanent internal structural specializations. Numerous reports document prolific gametogenesis in all five classes of the echinoderms involving single annual elaboration of spermatozoa and either single annual, biennial, or possibly, in some species, biannual or continuous formation of primary oocytes. (For all classes prior to 1966, see Boolootian 1966. For Asteroidea also see Mauzey 1966, Chia 1968, 1970b, Kim 1968, Crump 1971, R.H.Smith 1971, Jangoux & Vloebergh 1973, Achituv 1973, Walker 1973, 1979, 1980, Nimitz 1976, Worley et al. 1977. For Echinoidea, prior to 1975, see Piatigorsky 1975; also see Cochran & Engelmann 1975, Masuda & Dan 1977, Lane & Lawrence 1979b. For Crinoidea, see Holland et al. 1975, Holland & Kubota 1975. For Ophiuroidea, see Patent 1968, 1969, 1976, Fenaux 1968a, 1970, 1972, Tyler 1976, 1977. For Holothuroidea, see Engstrom 1970, Menker 1970, Rutherford 1973, Atwood 1974, Krishnan & Dale 1975, Green 1976, 1978). Many ofthese studies describe a signifi· cant increase in size of the gonad and accumulation of enormous numbers of spermatozoa or primary oocytes (occasionally only a few very large primary oocytes) in temporary storage in the gonadallumen. In many cases, the prodigious amounts of nu trients utilized du ring such intensive gametogenesis are retrieved from storage in specialized extra· or intragonadal organs or ceIis (e.g. pyloric caeca in the Asteroidea [Delaunay 1926b, Cuenot 1948, Farmanfarmaian et al. 1958, Cognetti & Delavault 1962, Mauzey 1966, Anderson 1966, Pearse 1965a,b, Ferguson 1975a,b, Nimitz 1976 and Jangoux & Van Impe 1977], ovarian and testicular nutritive phagocytes in the Echinoidea [e.g. Fuji 1960a, Giese 1961, Farmanfarmaian & Phillips 1962, Holland 1965b, 1967a, Holland & Giese 1965, Lawrence et al. 1965, Takashima & Takashima 1965, Lawrence et al. 1966, Lawrence 1967, Taka· shima 1968b, Holland & Holland 1969, Pearse 1969a,b, 1970, Chatlynne 1969, 1972], gonadal accessory and body wall cells in the Crinoidea [Chubb 1906, Harvey 1930, Hol· land & Kubota 1975], and gut wall and visceral peritoneum in the Holothuroidea [Krish. nan 1968, Krishnan & Dale 1975]). Abortive oocytes often contribute nutrients to other more successful primary oocytes in asteroids, echinoids and crinoids and also in ophiuroids (Mortensen 1936, where no other regions of extra· or intragonadal storage have yet been identified). AlternativeIy, nutrients may simply pass directly and more or less continuously to the gonad from food ingested by the gut (Delavault 1960, Cognetti & Delavault 1962, Pearse 1965a,b, in press). Mechanisms of nutrient transfer between organs of ingestion or extragonadal storage and the gonad have been carefully investigated only in the Asteroidea 449

450 Charles W. Walker and Echinoidea. Even in these classes, significant questions remain concerning the relative importance of the perivisceral coelomic fluid and the haemal system in nutrient transport (see Ferguson, chapter 16). While somewhat more information exists on nutrient distribu­ tion, storage and utilization by somatic and germinal cells forming spermatogenic and ooge­ nie epithelia, only preliminary observations have been made on nutrient distribution and storage by somatie tissues composing the remainder of the gonad. Since Ferguson has considered (see chapter 16) an currently recognized me ans of trans­ port which might possibly bring nutrients into the gonads, the present chapter will be limited to a discussion of potential mechanisms of intragonadal nutrient receipt, storage, distribution and utilization in all five classes of Echinodermata. Areport of this kind depends heavily on studies which describe seasonal changes in somatic structure of the gonads and also changes in the structure of the germinal epithelium. Almost every author has divided gametogenesis into numerous stages of various lengths. For the sake of simpli­ city, four phases have been used in the present study. They are distinguished on the basis of the principal activities of germinal cells at the time. Usually included is an agametogenic phase (occurring between successive gametogenic events and absent, for one reason or another, in those species which have overlapping or continuous cycles of gametogenesis), and always included are a proliferative phase (recognized by the occurrence of gonial mito­ sis), a differentiative phase (du ring which gametes are differentiated and stored and which often overlaps extensively with the preceding one), and finally, a short evacuative phase (during which gametes are released from the gonads). With the exception of oocytes, the cells of echinoderms are small and the boundaries between them are often invisible with traditionallight mieroscopy. Light and electron microscopic observations of tissues prepared with sophisticated flxation and plastic embed­ ding techniques must be used to provide adequate preservation and resolution of cellular interrelationships; the recognition of mitotic and meiotie division figures which are cru­ cially important in studies of gametogenesis can be achieved in this way only. At present, for the entire phylum, there are only eight studies of the structure of the gonads and ten studies of gametogenesis available which utilize such techniques. A number of other works, based on electron microscopie observations, deal with limited aspects of the structure of somatic and germinal cells of the gonad; these also provide valuable information. Many workers have used positive results with the periodic acid Schiff (PAS) routine and mercurie bromphenol blue (MPB) (also eosinophilia) as indicative of the presence of neutral muco­ polysaccharide complexes and protein. When these substances are present in signiflcant amounts in any tissue, they are interpreted as nutrient reserves (Holland & Giese 1965, Takashima & Takashima 1965, Mauzey 1966, Chia 1968, 1970b, Takash~a 1968b, Walker 1973, 1976a, 1979, 1980, Holland & Kubota 1975, Krishnan & Dale 1975, and Masuda & Dan 1977). In the sections that follow, pertinent literature will be reviewed and combined with my own observations concerning: 1) possible gametogenesis-related mechanisms of nutrient distribution and storage by somatic portions of echinoderm gonads and 2) possible gameto­ genesis-related mechanisms of nutrient receipt, storage, distribution, and utilization within gametogenic epithelia. The following summary of conclusions may be drawn from this work: 1) Many echinoderms may temporarily store nutrients intragonadally, for short or long periods of time, either in the genital hemal sinus of the gonad (in asteroids, crinoids, and holothuroids), in cells ofthe visceral peritoneum ofthe gonad (in h010thuroids), or in somatic or germinal cells of the germinal epithelium (in echinoids, asteroids and crinoids).

Nutrition o[ gametes 451 2) During gametogenesis in many species, germinal and somatic cells together form struc­ tural and fun~tional sub divisions of the germinal epithelium on a predictable schedule. 3) Such sub divisions provide the structural basis for organization of the microenvironment of small groups of germinal cells where interrelated activities such as nutrient storage, distri­ bution and utilization, hormone synthesis, and often simultaneous phases of gametogenesis can occur in an orderly way. The latter sub divisions might be considered as sites for the transduction of environmental and extra- or intragonadal forms of informational and nutritional input to the gametes.

1. POTENTIAL MECHANISMS OF NUTRIENT DISTRIBUTION AND STORAGE BY SOMA TIC TISSUES OF THE GONAD (excluding those in the germinal epithelium)

1.1. Somatic structure o[gonads in the Echinodermata The variable arrangement of reproductive systems in the different classes of echinoderms was carefully described by Hyman (l955), who also summarized much early literature con­ cerning structure of the gonads. Many accounts have appeared within the past 70 years, which specifically mention and add significantly to our understanding of the structure of the wall of the gonad. These include, for asteroids, Gemmill (l911, 1912, 1914), Tangapre­ gassom & Delavault (1967), Brusle (1969), Davis (l971), R.H.Smith (l971), Walker (l974a,b, 1976a,b, 1979, in press) and Chia (l980); for echinoids, Palmer (l937), Wilson (1940), Fuji (l960a), Holland (1965b), Kawaguti (1965a), Campbell (1966), Davis (1971) Chatlynne (1972) and Pearse (in press); for crinoids, Davis (1971), Holland (1971), Holland et al. (1975), Holland & Kubota (1975), and Holland (in press); for ophiuroids, Fedotov (l926b), Smith (l940), Austin (l966), Patent (l968, 1976), Davis (l971), and Hendler (in press); and for holothuroids, Davis (1971), Atwood (1973), and Krishnan & Dale (1975). Considered together, such studies demonstrate several structural features which are universally charac­ teristic of the phylum. Echinoderm gonads are sac-like organs, single or multiple, with large lumina that do not contain elaborate or permanent internal structures as seen, for instance, in mammals (figs.I-5). They are located in the perivisceral coelom or the genital (perihe­ mal) coelom. They are intimately associated with major aboral or oral hemal and some­ times perihemal channels which interconnect all gonads and also lie in contact with all major organ systems within each anima!. Gonads usually open to the exterior of the body by variously placed single or multiple gonopores (the existence of which is debated in cri­ noids and many ophiuroids and the position of which is often correlated with the mode of reproduction, e.g. gonoducts are directed to release eggs orally in brooding asteroids). In asteroids, echinoids, and ophiuroids, the wall of the gonad includes two sacs (out er and inner), each composed of several characteristic layers (figs.l b-3b). Throughout the gonad the outer sac is separated from the inner sac by the genital coelomic (perihemal) sinus (figs.l b-3b). The outer sac carries out day-to-day activities important in its own main­ tenance, it determines the overall shape of the gonad, its circular muscles contract during gamete release and its flagellated epithelial cells generate strong currents in the perivisceral coelom and weak local currents in the genital coelomic sinus (Walker 1979). In all echino­ derms, the inner sac is an expanded genital hemal ~inus which be ars the germinal epithe· lium on its inner wall (figs.lb-3b, 4c, 5b).lts major functions include longitudinal muscular

452 Charles W. Walker

contraction during gamete release, current generation in the genital coelornic sinus (aste­ roids and ophiuroids), perivisceral coelom (holothuroids) or genital coelom (crinoids), daily maintenance, intragonadal nutrient transport and storage (in many species), and, principally, gametogenesis. Gonads of crinoids lack a developmental homolog of the outer sac; here, the inner sac is surrounded by a genital (perihemal) coelom (figs.4b,c, 23) and both of these are encased by the body wall of the arms and genital pinnules. Because the gonads of holothuroids lack an outer sac altogether and so do not inelude a perihaemal component, the external surface of the hemal sinus in holothuroids is directly bathed by perivisceral coelornic fluid in which the entire gonad is suspended (fig.5a,b). 1.2. Predictable changes in the somatic structure 01 the gonad during gametogenesis that may be related to gamete nutrition

In most echinoderms, germinal epithelia annually, biennially or continuously produce tre­ mendous numbers of gametes which are temporarily stored in and usually released rather abroptly from gonads (extremely small adult echinoderms may produce limited numbers of oocytes, e.g. certain interstitial holothuroids). During radical changes in dimension, progressive adjustments in the appearance and characteristics of somatic portions of the gonads result which relate to the delivery and storage of nutrients, to formation, storage and subsequent voiding of gametes and to recovery after gamete release. The following section is concerned with tissues of the somatic portions of the gonad which seem to change in a predictable manner during gametogenesis. Consideration will be given to the possible nutrient transport or storage capacities of each tissue in the gonadal wall. 1.2.1. Outer sac (Iound only in Asteroidea. Echinoidea and Ophiuroidea) 1.2.1.1. The visceral peritoneum. The major constituent of this epithelium is a layer of interconnected flagellated-collar cells (Walker 1979) underlain by granule-containing peri­ karya and nerve tracts; musele fibers which are thought to be specializations of the basal portions of some flagellated-collar cells have also been seen in the asteroids (Hamann 1885, Walker 1976a,b, 1979) (fig.6). During gametogenesis flagellated-collar cells change drasti­ cally in shape and size. During the agametogenic phase, when gonads are smalI, flagellated­ collar cells appear columnar or cubical and are set elose together; toward the end of gametogenesis, when gonads have reached maximum dimensions, the cells are flattened and interconnected by extremely attenuated cellular processes. Walker (1976a, 1979) has shown that such cells in asteroids remain in contact with their neighbors throughout game­ togenesis by adhering zonules that ron completely around their free surfaces (figs.6,7). Except for occasional regions where nerve cells emerge (fig.6, inset), it is assumed that this layer forms a continuous, unbroken and relatively impenetrable covering for the surface of the gonad. As a result, the proposition that large amounts of nutrients reach the germinal epithelium simply by passage from the coelomic fluid through or between flagellated­ collar cells and then through most of the wall of the gonad does not seem to be realistic for asteroids. Moreover in asteroids, echinoids and ophiuroids, flagellated-collar cells do not possess organelles or external architecture normally associated with transporting epi­ thelia (Berridge & Oschman 1972), nor have they been shown to store nutrients prior to or during gametogenesis. It is likely that the major function of flagellated-collar cells in all three elasses is, like that of the peritoneum throughout the body, the generation of currents

Nutrition of gametes 453 of coelomic fluid to facilitate the diffusion of respiratory gases. Nutrients acquired from the coelomic fluid are probably used locally for cellular maintenance (Walker 1979). 1.2.1.2. Connective tissue layer. This substantiallayer of collagenous and putative elastic tissue determines the overall external shape of gonads in asteroids, echinoids, and ophiu­ roids and it fluctuates greatly in dimensions during gametogenesis (Walker 1979). Although thick and robust during the agametogenic phase of gametogenesis, it becomes increasingly attenuated during growth and accumulation of gametes in the lumen of the gonad. After the evacuation of gametes, the connective tissue layer returns to its thickened agametogenic state. During these radical changes in dimension, fibers of connective tissue are damaged and presumably require nutrients for repair (Walker 1976a, 1979).1t has never been postu­ lated that nu trients are stored in the connective tissue of the outer sac and, on the basis of available evidence, it is likely that the function of the connective tissue layer is primarily structural. 1.2.1.3. Epithelial cells associated with the inner basal lamina of the connective tissue layer of the outer sac. In asteroids, an especially well-developed, discontinuous layer of circularly arranged flagellated-muscle cells is attached to the inner basal lamina of the connective tissue layer of the outer sac. Such epithelio-muscular elements are either not present or were not well-developed at the time of observation in echinoids and ophiuroids (Davis 1971, Pearse in press). Although non-muscular flagellated epithelial cells are present in a sirnilar location in ophiuroids, flagellated cells have not been recognized in echinoids. Fla­ gella on both kinds of cells project into the genital coelomic (perihemal) sinus and pro­ duce localized currents of coelomic fluid sirnilar to those observed elsewhere by Gemmill (1915). Cells in this position in asteroids, echinoids, and ophiuroids have never been shown to contain significant amounts of potentially nutritive substances. Prior to gametogenesis, in both males and females of some crinoids, cells situated in an analogous place on the inner basal lamina of the body wall in the arms and genital pinnules contain eosinophilic, PAS- and MPB-positive spherules (diameter 6 11m) (Carpenter 1884, Holland & Kubota 1975). During gametogenesis, these cells lose their spherules and become smaller and PAS­ and MPB-negative. It has been suggested that, in these species, such cells of the body wall are an important source of nutrients utilized in gametogenesis (Holland & Kubota 1975). 1.2.2. Genital coelomic (perihemal) sinus The genital coelomic sinus is a fluid-mIed space that totally separates the outer and inner sacs from each other in asteroids, ophiuroids, ;md probably in echinoids (figs.l b-3b, 8). In all three classes it appears to be homologous and presumably provides the same interconnec­ tion between the gonads and most other internal organs (Cuenot 1887, 1891b, 1948). From studies on members of each of these classes, it is apparent that the dimensions of the genital coelomic sinus vary from wide and expansive prior to gametogenesis to nearly occluded after the enlargement of the germinal epithelium. In those species of echinoids where mas­ sive, nutrient-containing nutritive phagocytes fill the lumen of the gonad prior to gamete proliferation and differentiation, the dimensions of the genital coelomic sinus are reduced during most of the year. As Ferguson notes (see chapter 16), ' ... it is probable that the perihemal spaces act as extensions of the perivisceral coelom in supporting translocation of nutrients to those criti­ cal areas with which they associate'. In some species, Pearse (unpublished) using echinoids

454 Charles W. Walker and Walker (unpublished) using the asteroid, Pteraster militaris, have both observed direct connections between branches of the perihemal coelomic channelleading to the gonads and the perivisceral coelomic cavity. It is altogether possible that nutrients are brought through the genital coelomic sinus into the gonads where they are distributed for imme­ diate use or where they are stored by various intragonadal means for subsequent use during gametogenesis. Obviously this important portion of the gonads in asteroids, echinoids, and ophiuroids, and the analogous space in crinoids (the genital coelom, fig.4b,c), should be thoroughly scrutinized so that its likely role in intragonadal nutrient transport may be understood. 1.2.3. Inner sac (present in all echinoderms) 1.2.3.1. Epithelium associated with the outer wall o[ the genital hemal sinus ([acing the genital coelomic sinus in asteroids, echinoids, and ophiuroids; the genital coelom in cri­ noids; and the perivisceral coelom in holothuroids). In asteroids, the principal components

of this discontinuous epithelium are longitudinally arranged, flagellated-muscle cells iden­ tical in all ways except orientation and position to those attached to the connective tissue layer of the outer sac (Walker 1974a,b, 1976a,b, 1979). Flagellated epithelial cells, muscle fibers and nerve tracts have been seen in this location in the gonads of all other echino­ derms. These cells are stretched and flattened during growth of gonads. With the exception of certain holothuroids, such epithelial cells have never been implicated in intragonadal nutrient storage. The layer serves as the visceral peritoneum in the gonads ofholothuroids (figs.9,10). In different species, the epithelium varies in condition; it may be a layer of squamous, non­ vacuolated cells (Hyman 1955, Atwood 1973) (fig.9), or it may consist of tall, columnar cells, packed with membrane-bound globules (Hyman 1955, Krishnan & Dale 1975) (fig. 10). In certain species, the flagellated cells composing it apparently do store nutrients in such globules which are PAS- and MPB-positive and occasionally contain crystalline pro­ teins (Krishnan 1968, Krishnan & Dale 1975). mtrastructural observations provide evidence that globule formation occurs in cytoplasm nearest the perivisceral coelomic fluid. The authors suggest that raw nutrient materials used in constructing globules are derived from coelomic fluid by endocytosis. Krishnan & Dale (1975) noted that the epithelium does not show fluctuations in size and nutrient content that can be correlated with gametogenesis. On the basis of such limited information, it is impossible to determine whether presump­ tive nutrients located in the cells of the visceral peritoneum are utilized by cells of the ger­ minal epithelium. 1.2.3.2., Genital hemal sinus. This space is well-developed and consistently encountered in all echinoderms. It is elaborated to varying degrees in different classes and, depending upon the species considered, it may or may not show changes in dimensions during gametogene­ sis. In all classes, it is usually composed of two fibro-granular laminae, the outer and inner walls, which enclose aspace often mIed with amoeboid or granule-containing cells, collagen fibers, and (in histological preparations) precipitated remnants of hemal fluid (fig.12). In those species that have been observed ultrastructurally, the genital hemal sinus is a continu­ ous space not lined internally by epithelial cells. Branches of the hemal system interconnect the genital hemal sinuses of all gonads in an individual with each other and with hemal components of other major organ systems. Throughout the gonad, the inner wall of the

Nutrition o[gametes 455 genital hemal sinus directly underlies somatic and germinal cells of the germinal epithelium. Because of its oscillations in size and content during gametogenesis in some echinoderms, the genital hemal sinus deserves to be distinguished as a unique component of the hemal system and not simply equated, as it so often is, with random channels in what is otherwise considered to be a connective tissue layer. This is simply not a valid description of the genital hemal sinus when ii is storing potentially nutritive hemal fluid. In asteroids, the genital hemal sinus is greatest in volume each year prior to and during gametogenesis and, at this time, is filled with granular P AS- and MPB-positive hemal fluid (Walker 1973, 1974a, 1976a, 1979, 1980). As gametogenesis continues, the contents of the genital he mal sinus diminish progressively and predictably in volume, in granularity, and in PAS- and MPB-positivity. In fact, Mauzey (1966), Chia (1968, 1970b), and Walker (1973, 1976a, 1979, 1980) noted that the PAS-positive contents ofthe genital hemal sinus in ovaries appear to decrease in volume in direct proportion to the increase in PAS­ positivitiy of vitellogenic primary oocytes attached to the inner wall of the genital hemal sinus (figs.l1, 13). As noted by Walker (1979,1980), the 'nutrient' contents of the genital hemal sinus in testes decrease in amount while proliferation and differentiation of sperma­ tocytes occur (fig.14). The works of Farmanfarmaian et al. (1958), Ferguson (1975a,b), Nimitz (1976), and Jangoux & Van Impe (1977) suggest that, in some asteroids, there is an annual period of exchange of substances between the pyloric caeca and the gonads (see Lawrence & Lane, chapter 15). They show that, while large amounts of amino acids, fatty acids and carbohydrates originally present in storage cells of the caeca are decreasing, a simul­ taneous increase in the nutrient content of the 'gonad' as a whole begins (Oudejans & Van der Sluis 1979a,b, Oudejans et al. 1979). Microscopic observations of the structure of the gonads during this presumed exchange shows that the only corresponding observable gross difference in size and condition of the tissues of the gonad is seen in the genital hemal sinus, where the volume, granularity and PAS- and MPB-positivity of its hemal fluid fluc­ tuate. On the basis of these points of evidence, it is assumed that nutrients are distributed within the gonad and accumulate extracellulady in the genital hemal sinus for subsequent use during gametogenesis. Invaginations of the genital hemal sinus into the lumen of the ' gonad are common in all echinoderms; they appear triangular in cross-section, run longitu­ dinally along the gonad for many micrometers and are filled with nutrient-Iaden hemal fluid. These are presumably important in distributing nutrients evenly to developing ger­ minal cells. In echinoids, there are no reports indicating similar storage of potentially useful nutrient reserves in the genital haemal sinus of either sex. In fact, there is no comprehensive des­ cription of the seasonal condition of the genital haemal sinus for any echinoid. It is there­ fore impossible to suggest the role played by this part of the echinoid gonad during gameto­ genesis. By inference from the foregoing discussion of the activities of the genital haemal sinus in asteroids, it is likely that this space in echinoids may, at least, provide for an even distribution of nutrients throughout the gonad where, prior to gametogenesis, they are taken up and stored by nutritive phagocytes located in the germinal epithelium. Since nutrients are stored in nutritive phagocytes in both the testes and ovaries· of echinoids, other forms of intragonadal nutrient sequestration may be unnecessary. It is perhaps im­ portant to mention the work of Okada (1979) which indicates that gonads atrophy when hemal channels leading to them are severed. In some crinoids, presumed nutrient substances are localized in enlarged gonadal acces­ sory cells situated in the genital hemal sinus (Holland & Kubota 1975). Prior to gameto­

456 Charles W. Walker

genesis in both sexes, proteinaceous and carbohydrate substances are found in membrane­ bound globules in these cells. During gametogenesis, gonadal accessory cells lose their obvious nutrient contents and shrink in size (Holland & Kubota 1975). Although these authors did not mention how gonadal accessory cells transfer their contents to germinal cells, they do clearly show a correlation between reduction in the size and nutrient content of globules in gonadal acce.ssory cells and the increase in size and numbers of germinal cells. Holland (personal communication) indicates that the hemal fluid in crinoid gonads is PAS· and MPB-positive and granular during gametogenesis. In ophiuroids and holothuroids, there is little information available on fluctuations in the dimensions or possible nutrient content of the genital hemal sinus during gameto~ne­ sis. The micrographs ofPatent (1968) and my own observations indicate the presence of coa~ulated fluid in the genital hemal sinus (fig.15) of GorgonocepluJlus sp. be fore gameto­ genesis begins. In Cucumaria curata, Krishnan & Dale (1975) found the he mal fluid to be mildly PAS- and MPB-positive in no apparent relationship to gametogenesis. At present we have no unequivocal evidence demonstrating intragonadal nutrient storage in either ophiu­ roids or holothuroids. Because of the probable significance of the genital hemal sinus in nutrient transport and storage in asteroids and crinoids and its intirnate contact with ger­ niinal cells in all classes, it undoubtedly would be valuable to observe the changing status of this portion of the gonad more closely during the gametogenic cycle in echinoids, ophiroids, and holothuroids. 2. POTENTIAL MECHANISMS OF NUTRIENT DISTRIBUTION, STORAGE, AND UTILIZATION BY THE GERMINAL EPITHELIUM 2.1. Predictable generation o[[unctional subdivisions o[ the germinal epithelium during gametogenesis

In most echinoderms, both spermatogenesis and oogenesis involve the simultaneous and orderly progression of enormous numbers of germinal cells through apreeise sequence of cytological changes. These culminate in the differentiation and temporary storage of hund­ reds of millions of spermatozoa and in the differentiation and often the storage of miIlions of primary oocytes. The fact that such extensive gametogenesis is accomplished efficiently and briefly in what are consistently described as structurally simple gonads suggests that the gametogenic mechanism in both testes and ovaries is under precise control and is per­ haps more complex than has yet been appreciated. As Holland & Dan (1975) and Roosen­ Runge (1977) pointed out, studies of seasonal cycles of reproduction in echinoderms are concemed with gross changes in the form of the germinal epithelium and rarely provide detailed descriptions of the interactions of cells involved. Very few ultrastructural studies of oogenesis have been published (Verhey & Moyer 1967, Tsukahara & Sugiyama 1969, Cruz-Landim & Beig 1973, Takashima 1972, 1976, and Chatlynne 1972), and as Roosen-Runge (1977, p.98) noted 'the spermatogenic process in its continuity has not been thoroughly investigated in even a single species (of echino­ derm)'. Where observations have been made, it is clear that germinal epithelia are not simple continuous layers composed exclusively of germinal cells. Instead, in all classes, both somatic and germinal cells are found. Furthermore, germinal cells are gene rally not arranged randomly with respect to somatic cells. Rather, somatic cells usually deterrnine

Legends to figures 1 to 13: Figure la. Diagrammatic representation of a vertical section through the disc and basal portions of two rays in an asteroid showing three possible arrangements of the rcproductive system seen in different species. In the ray on the right, the gonad opening aborally is single sac-like structure typieal of most asteroids; the gonad opening orally is typieaI of asteroids Iike Leptasterias and Henrieia whieh brood their young. In the ray to the Ieft, the serial form of gonadal distribution is shown; this is typieal of genera like Astropeeten, Luidia and Aeanthaster. Figure 1b. A section of the wall of a testis taken from the asteroid, Asterias vulgaris. showing the arrangement of tissues forming both outer and inner sacs. Cells lining the genital coelomic sinus (GCS) are predominantly flagellated and muscular. Bar represents 10 Jlm. Figure 2a. Diagrammatic representation of a vertical section through the disc of an ophiuroid showing three pos­ sible arrangements of the reproductive system. To the right of the gut notice the multiple. small gonads that open into the genital bursa (GB) as in the Euryalae; to the leH of the gut the smgle gonad (G) opens either into the genital bursa or near thc genital slit (GS). Figure 2b. A section of the wall of a testis taken from the ophiuroid. Ophiopholus aeuleata. showing the tissues forming both outer and inner sacs. Cells associated with both sacs and lining the genital coelomic sinus (GCS) are flagellated; only those associated with the inner sac have been shown to be muscular. The lumen of the tes­ tis is filled with stored, differentiated spermatozoa. Bar represents 30 Jlm. Figure 3a. Diagrammatic representation of a vertical section through the test of a regular echinoid, showing the arrangement of elements of the reproductive system. Figure 3b. A section of thc wall of a testis taken from the echinoid, Strongylocentrotus droebaehiensis. showing the tissues forming both outer and inner sacs. Cclls are clearly associated with the outer wall of the genital hemal sinus (GHS) facing the genital coelomic sinus (GCS); such cells are muscular and flagellated. Similar cells have not been identified in association with the inner basal lamina of the connective tissue layer of the outer sac (CTL) and are not obvious in the micrograph as weil (compare with Ib and 2b where cells are typically found in this position). Large granules are present in nutritive phagocytes of the germinal epithelium (GE); spermato­ gonia can be recognized between nutritive phagocytes. Bar represents 10 Jlm. Figure 4a. Diagrammatic representation of a vertical section through the calyx of a crinoid; the hemal strand (HS) interconnects all gonads which are located in the arms (AR) or genital pinnules (P) or both. Figure 4b. Cross section of a genital pinnule from the crinoid, F1orometra perplexa. showing an enlarged testis (Ts) surrounded by the genital coelom (GC) and body wall (BW) of the pinnule. Bar reprcsents 80 Jlm. Figurc 4c. Enlarged view of thc testis shown in 4b, demonstrating that it includes the genital hemal sinus (arrows) and germinal epithelium characteristic of the inner sac in other echinoderms; the testis is surrounded by the genital coelom (GC) (also see figure 23). Bar represents 40 Jlm. Figure 5a. Diagrammatic representation of a seetion along the longitudinal axis of a holothuroid, showing one possible arrangement of tubules of the single gonad. Figure Sb. Cross section through one testis tUbule from the holothuroid, Sclerodaetyla briareus. showing the genital hemal sinus and its associated epithelia (internally .- the germinal epithelium (GE) and externally - the visceral peritoneum (VP). In this case, the cells of the visceral peritoneum are expanded with storage granulcs and are flagellated and muscular (arrow). Bar represents 20 Jlm. Figure 6. Several interconnected flagellated-collar cells forming the visceral peritoneum of the ovary in the aste­ roid. Ctenodiscus erispatus. Notice the granule-filled nerve tracts that pass between flagellated-collar cells near their attachment to the undcrlying connective tissue layer (CTL). The inset shows a portion of a perikaryon emerging into the perivisceral coelom (C) between surrounding flagellated-collar cells. Bar represents 3 Jlm. Figure 7. A tangential section of the surface of an ovary in Asterias vulgaris. Above, the seetion passes through the junctions between several cells; below, the section passes nearer the coelom. The junctions seen in this figure are of the same type seen in figure 6. Bar represents 2 Jlm. Figure 8. Panoramic view of the wall of the testis of Ctenodiscus erispatus showing thc tissues composing both outer and inner sacs and emphasizing the variable size of the genital coelomic sinus (GCS) (compare with figures la-3a, 11-14). In this species the genital haemal sinus (GHS) is filled with few cells and no collagen fibers unlike the situation in Asterias vulgaris (see figure 12). Bar represents 5 Jlm. Figure 9. Longitudinal section along a single ovarian tubule from the holothuroid, Chirodota laevis. The con­ struction of the ovary from thc inner sac only is clearly demonstrated. In this animal, which was collected in the summer, the visceral peritoneum is composed simply of small columnar cells without globular inclusions. Bar represents 40 Jlm. Figure 10. Cross section of a testis tubule from Sclerodaetyla bria.reus showing the enormous columnar globule­ containing cells of thc visceral peritoneum. Bar represents 10 Jlm. Figure 11. Panoramic view of two lobes of the ovary in Asterias vulgaris during the agametogenic phase of the year. The intense PAS-positivity of the coagulated contents of the genital hemal sinus is obvious. Bar represcnts 60Jlm. Figure 12. View of the genital hemal sinus (GHS) at the base of two spermatogenic columns in Asterias vulgaris (see numerous cellular processes near the inner wall of the genital hemal sinus (lW) Gompare with figure 14). The hemal sinus space is filled with amoeboid cells and granular hemal fluid. Bar represents 4 Jlm. Figurc 13. Several tear-drop shaped primary oocytes in the early growth phase seen in the ovary of Asterias vul­ garis. These oocytes are larger in size than other non-gametogenic oocytes found between them, are intensely PAS-positive and are associated with invaginations of the genital hemal sinus which are also PAS-positive. Inset shows follicle cells which surround oocytes and are more obvious in plastic than in paraffin sections. Bar repre­ sents'40 Jlm.

Legends to jigures 14 to 26: Figure 14. Several spermatogenic columns vicwed in the testis of Asterias vulgaris during the proliferative phase of spermatogenesis and before the differentiation of the primary spermatocytes. Arrows point out the lighter nuclei of somatic cells. The circle designates an area similar to that seen in figure 12. Inset shows a view of columns in a living testis taken with interference optics. Bar represents 40 pm. Figure 15. Section through several ovaries from the ophiuroid, Gorgonocephalus arcticus. The ovarian lumen (L) is filled with cells while the genital hemal sinus (GHS) contains coagulated hemal fluid. Bar represents 20 pm. Figure 16. Longitudinal section of two spermatogenic columns from Asterias vulgaris showing the axial somatic cell and processes from spermatocytes. Bar represents 20 pm. Figures 17-19 are diagrammatic representations of annual modifications in the condition of the inner sac of the testis (a) and ovary (b), including both the genital hemal sinus (GHS) and the germinal epithelium (GE) for asteroids, echinoids, and crinoids. One reproductive cycle is represented from left to right, beginning after the evacuation of spermatozoa or oocytes from the previous year's gametogenesis (vertical black line on left) and progressing to a similar phase in the current year (vertical black line on the right). Changes may include cyclic variation in the quality and volume of hemal fluid and in the architecture of the germinal epithelium. Somatic cells are depicted with black nuclei; germinal cells have clear nuclei with a dark nucleolus. Figure 17a. Asteroid testis. Following evacuation of most spermatozoa, phagocytic somatic cells in the early agametogenic epithelium engulf residual spermatozoa (extreme left). SUbsequently, during the late agametogenic phase, vacuolarized derivatives of phagocytic cells (now in the gonadallumen) may synthesizc steroids (second panel in 17a). Hemal fluid (presumably nutritive in character) begins to accumulate in the genital hemal sinus (GHS) and spermatogonia initiate intensive mitosis. Resulting primary spermatocytes and somatic cells together generate thousands of subdivisions of the germinal epithelium, the spermatogenic columns_ After columns attain a characteristic height (75-150 pm) cells at their tips begin meiosis. Proliferation of spermatocytes occurs simultaneously with differentiation until presumed 'nutrient' substances in the GHS are depleted; spermatogo­ nial mitoses occur at a significantly reduced rate. Finally columns degrade as those spermatocytes forming them enter meiosis. The sornatic cells originally in columns directly give rise to the phagocytic cells that engulf residual spermatozoa. Figure 17b. Asteroid ovary. Following evacuation of most ova, follicle cells or other somatic eells in the ovary phagocytize residual ova. Later 'vesicular' somatic cells are found in the ovarian lumen. Just prior to the initia­ tion of oogonial mitoses, steroid sccreting cells are recognizable in the gonadallumen. Follicles develop in the early growth phase of vitellogenesis. Such oocytes extend into the ovarian lumen in the typical tear-drop shape and are attached to the inner wall of the genital hemal sinus (GHS) (which is filled with presumed 'nutrient' contents). Finally, oocytes enter the late growth phase of vitellogenesis as follicles are transferred into the ova­ rian lumen. Substances in the GHS diminish in quality and volume concomitantly with vitellogenesis. Figure 18a, b. Echinoid testis and ovary. Immediately after evacuation of spermatozoa and ova, remnants of nutritive phagocytes engulf residual gametes in testes and ovaries. Subsequently, nutritive phagocytes enlarge with nutrients in the form of glycogen and !arge PAS-positive globules. Spermatogonial and oogonial mitoses begin and resulting germinal cells are forced luminally between the enlarged nutritive phagocytes. Ouring this time, nutrients are presumably passed to germinal cells in both testis and ovaries. Towards the lumen, primary gametocytes undergo meiosis. In most cases, both proliferation and differentiation of germinal cells continues until reserves in the nutritive phagocytes are depleted and the cells have shrunk. Oifferentiated spermatozoa and primary oocytes are stored in the lumen until evacuation. Figure 19a. Crinoid testis. Very little is known about spermatogenesis based on da ta collected throughout the year. When spermatozoa are being stored in the lumen of the testis, invaginations of the genital hemal sinus are present. Somatic cells do not seem to form well-defined relationships with germinal cells. Figure 19b. Crinoid ovary. Ouring the agametogenic phase of the year, clumps of somatic and germinal cells are present in the germinal epithelium. The genital hemal sinus (GHS) contains hemal fluid and globule-filled cells. Oogonial mitoses begin; subsequently single oocytes sink into the genital hemal sinus in an invagination of the inner wall of the sinus. Ouring vitellogenesis, nutrients may be passed from the genital hemal sinus across its inner wall and into the invaginations, each of which contains a single prirnary oocyte. Figure 20. Section of the testicular wall of Strongylocentrotus droebachiensis showing three distinct nutritive phagocytes (see three major lumps toward lumen) containing 'nutritive' globules. Lumen contains differentiated spermatozoa; pockets of spermatocytes can be seen between phagocytes. Inset shows an entire tubule of testis with numerous spermatozoa (dark radiallines) between nutritive phagocytes. Bar represents 20 pm. Figure 21. Section of the ovarian wall in Strongylocentrotus droebachiensis showing many distinct nutritive phagocytes and surrounding primary oocytes (near the genital hemal sinus (GHS». Inset shows cross-sections of two ovarian tubules with marginal oogonia. Bar represents 20 pm. Figure 22. Section of the ovarian wall in the asteroid, Hippasteria phrygiana, showing enlarged vitellogenic pri­ mary oocytes and surrounding 'nursc cells' (arrows). Bar represents 100 pm. Figure 23. Testis of the crinoid F7orometra serratissima showing folds of the inner wall of the genital hemal sinus (GHS) and other features of the gonadal wall (courtesy of Louise Bickell and Or F.S. Chia). Bar represents "" 15 pm. Figure 24. Testis of the ophiuroid, Ophiura albida. (courtesy of P.A.Tyler), showing subdivisions of the germi­ nal epithelium into structures which are similar to the spermatogenic columns of asteroids. Bar represents 30pm. Figure 25. Coagulated, PAS-positive material is obvious luminally between the viteUogenic oocytcs in Asterias vulgaris. Bar represents 30 pm. Figure 26. Section through lobes of the ovary of the asteroid, Luidia clathrata. showing two size c1asses of 00­ cytes (one marginal, one lumina!). Bar represents 100 pm.

A - anus; AG - axial gland; AL - Aristotle's lantern; AS - axial sinus; BW ~ body wall; C - perivisceral coelom; CTL - connective tissue layer (outer sac of gonad); G - gonad; GB - genital bursae; GCB ­ genital coelomic branch; GCS - genital coelomic (perihemal) sinus; GD - gonoduct; GE - germinal epi­ thelium; GHB - genital hemal branch; GHS - genital hemal sinus; GP - gonopore; GR - genital rachis; GS - genital slit; Gt - gut; IS - interradial septum; M -mouth; Md - madreporite; R - ray; SC - stone canal; T - test; VP - visceral peritoneum.

A - anus; AG - axial gland; Ar - arm; BW - body wall; C - perivisceral coelom; CT - conncctive tissue; CTL - conncctive tissue layer (outer sac of gonad); FT - fee ding tentacles; G - gonad; GC - genital coelom of pinnulc; GD - gonoduct; GE - germinal epithelium; GHS - genital hemal sinus; GP - gono· pore; HS - hemal strand; L - lumen of gonad; M - mouth; P - pinnules; S - stalk; TF - tube·feet of pinnule; Ts - testis; VP - visceral peritoneum.

C - perivisceral coelom; CTL - connectivc tissue layer (outer sac of the gonad); GCS - genital coelomic (pcrihemal) sinus; GE - germinal epithelium; GHS - genital hemal sin us; L - lumen of gonad; VP - vis­ ceral peritoneum,

C - perivisceral coelom; GCS - genital coelomic (perihemal) sinus; GHS - genital hemal sinus; IW ­ inner wall of the genital hemal sinus; L - lumen of gonad; OW - outer wall of genital hemal sinus.

GC - genital coelom of pinnule; GCS - genital coelomic (perihemal) sinus; GE - germinal epithelium; GHS - genital hemal sinus.

C - perivisceral coelom; GC - genital coelom of pinnule; GHS - genital hemal sinus; L - lumen ofgonad.

Nutrition o[ gametes 457 the distribution of germinal cells in the lumen of the gonad by forming specific associations with one or several of them. Such associations organize the germinal epithelium lining the sac-like gonads of echinoderms into numerous structural and functional sub divisions (e.g. ovarian follieles and spermatogenic columns in asteroids) where single or small groups of gametogenic cells may be dealt with at any time. With sufficient knowledge of a specific species, one finds that the degree of elaboration of these sub divisions varies predictably du ring the year. In both sexes, marked seasonal changes in the gross appearance of the ger­ minal epithelium noted in the literature usually reflect interaction between germinal and somatic cellular elements of its sub divisions. Only the ultrastructural studies of Tsukahara & Sugiyama (1969), Chatlynne (1972), and Takashima (1968b, 1972, 197~) using female echinoids and those of Walker (1980) using male asteroids document changing conditions of the germinal epithelium in terms of interactions between somatic and germinal cells. Other studies merely confirm the presence of somatic cellular elements in the lumen of the gonad of both sexes and point out the possible functional significance of such cells during gametogenesis (Cognetti & Delavault 1962, Schroeder 1971, Nimitz 1976, Schroeder et al. 1977,1979, Kubota et al. 1977, Schoenmakers et al. 1977, Worley et al. 1977, Dehn & Hinsch 1978, Hinsch & Dehn 1979). In the following section, information available in the literature and my own observations are combined in describing those structural subdivisions of the germinal epithelium which have been recognized or may be proposed for the various elasses of the Echinodermata. An appreciation of the possible existence and nature of these cellular associations is an essential prerequisite for an understanding of the nutrition of ger­ minal cells. During spermatogenesis in the Asteroidea, the spermatogenic epithelium is subdivided into thousands ofvisibly distinct columnar entities, the spermatogenic columns (these structures were referred to as 'colonettes' by Cognetti & Delavault 1962) (figs.14,16,17a). Ultrastructural observations show that each fully developed column is composed of at least one axial somatic cell which is surrounded by ~400 spermatocytes (Walker 1980). Incipient columns form during the early spermatogenic phase after spermatogonial mitoses begin in the testicular germinal epithelium. Such mitoses continue to occur at the bases of columns even after meiosis and spermiogenesis begin at their tips (Smith 1971, Walker 1980). Sper­ matocytes are apparently displaced from the bases of columns along their axes towards the tips. The elongated format of columns provides loose spatial separation between diffe­ rent generations of germinal cells involved simultaneously in proliferative and differentiative phases of spermatogenesis (Walker 1980). Finally, formation of primary spermatocytes ceases basally and columns degrade completely as those germinal cells still composing them complete their differentiation. During the subsequent agametogenic phase of the sperma­ togenie cycle, columns do not exist, and 'vesicular cells' (Cognetti & Delavault 1962) phago­ cytize residual spermatozoa within the lumen of the testis. These phagocytic cells are derived directly from the axial somatic cells found earlier in the columns; along with the so-called interstitial cells mentioned by Kubota et al. (1977), they are the only somatic cells that have been identified in the testicular lumen of asteroids (Walker 1980). During the differentiative phase of oogenesis (fig.17b), ovarian follieles segregate primary oocytes from each other (Delavault 1960, Cognetti & Delavault 1962, Mauzey 1966, Kim 1968, Walker 1980). Follieles form in the lumen of the ovary in elose association with the inner wall of the genital haemal sinus. Pockets of oogonia and small, newly formed primary oocytes are surrounded by a mass of follicular cells. Following growth, the oocytes separate from each other, extend into the lumen, and become encireled by a layer of folliele cells

458 Chorles W. Walker (fig.13, inset). At this time, oocytes are pear-shaped and are closely associated by astalk of cytoplasm to an indentation of the inner wall of the genital hemal sinus (figs.ll,13,17b). As vitellogenesis begins, the oocytes become PAS-positive. Finally larger and more mature follicles are encountered in the ovarian lumen where a PAS-positive substance similar in appearance to hemal fluid surrounds them (Walker 1973). The origin of this 'hemal fluid­ like' substance is unknown. Until they are evacuated from the ovary, oocytes are sheathed by follicular cells which are finally stretched into very thin sheets. 'Nurse cells', phagocytic cells and steroid-secreting cells have also been described from the lumen of the ovary; these somatic cells may or may not be derived from follicular-cells (Schoenmakers 1979). 'Nurse cells' are commonly found in large clumps attached to the inner wall of the genital hemal sinus and are particularly numerous and well-developed in species which produce large macrolecithal primary oocytes (fig.22). In no case has cytoplasmic continuity between 'nurse cells' and primary oocytes been reported. In both testes and ovaries of the Echinoidea, sub divisions of the germinal epithelium appear less discrete than those seen in asteroids (figs.20,21). Upon closer examination, one can see that subdivisions are present and are based on the existence and dimensions of fla­ gellated nutritive phagocytes (Holland & Giese 1965, Takashima & Takashima 1965, Taka­ shima 1968b,Holland & Holland 1969,Chatlynne 1972,Masuda & Dan 1975,Pearse in press). In both sexes, sub divisions of the gametogenic epithelia are structurally analogous to sper­ matogenic columns seen in the testes of asteroids; in echinoids, follicular cells never sur­ round developing oocytes (Holland 1965b, Holland & Holland 1969, Chatlynne 1972). Early in gametogenesis, germinal cells are positioned among the bases of the enlarged nutritive phagocytes which fill the lumen of the gonads. Such nutritive phagocytes are packed with membrane-bound eosinophilic and PAS-positive globules (indicating protein and neutral muco-polysaccharide contents) (figs.3b,18a,b,20,21). As gametogenesis commences, ger­ minal cells associated with the outer surface of nutritive phagocytes begin to move toward the lumen where resulting gametes are stored for varying periods of time. Proliferation of germinal cells occurs where nutritive phagocytes are attached to the inner wall of the geni­ tal hemal sinus, and differentiation occurs toward their luminal ends. As gametes become numerous in the testes and numerous and larger in the ovaries, the globules within the cytoplasm of the nutritive phagocytes decrease in number and PAS-positivity, until the nutritive phagocytes containing them shrink, leaving the gonadallumen filled with stored spermatozoa or primary oocytes. Shrunken nutritive phagocytes are seen peripherally near the inner wall of the genital hemal sinus. Aside from their size and obvious nutrient storage capabilities, the relationship of nutritive phagocytes and germinal cells in both testes and ovaries of echinoids is structurally very similar to that ofaxial somatic cells and spermato­ genie cells seen in spermatogenic columns of asteroids. In the Crinoidea, a careful study of gametogenesis has been made only in Comanthus japonica using light rnicroscopy (Holland et al. 1975, Holland & Kubota 1975) and in Flo­ rometra serratissima using electron microscope (Bickell et al. 1980) (figs.19a,b, 23). Un­ described somatic cells are present along with germinal cells on the inner wall of the genital he mal sinus. In testes, no specific structural relationship has been reported between sper­ matogenic and flagellated somatic cells and it is impossible to state whether sub divisions of the spermatogenic epithelium exist (Holland in press). Numerous infoldings of the inner wall of the genital hemal sinus carry both somatic and germinal cells into the lumen of the gonad and increase the surface area of the germinal epithelium (figs.4b,23). The relationship between somatic and germinal cells in the ovaries of crinoids is unlike

Nutrition o[ gametes 459 that encountered elsewhere in the echinoderms already mentioned; it is most clearly ex­ plained by Holland et al. (1975) and Holland (in press)(fig.l9b). Prior to vitellogenesis, primary oocytes are formed among somatic cells of the oogenic epithelium. As oocytes enlarge, each forces the development of an invagination of the inner wall of the genital hemal sinus into the sinus space. When the invagination completely surrounds it, the oocyte appears to reside within the genital hemal sinus (Hamann 1885, Holland 1971, in press). Actually each oocyte is surrounded by the acellular inner wall of the genital hemal sinus which forms the invagination and so is still associated with the germinal epi­ thelium and the lumen of the ovary (fig.19b). This relationship between primary oocytes and the germinal epithelium is obvious when one can clearly discern the aperture of the invagination and recognize that somatic cells of the germinal epithelium are intimately associated with each oocyte at this point. While oocytes are within these invaginations, they are separated by a single acellular basal lamina from the hemal fluid which in some species of crinoids contains nutrient-Iaden gonadal accessory cells. As accessory cells lose their PAS- and MPB-positivity, vitellogenesis occurs in the cytoplasm of invaginated oocytes. Ouring ovulation, oocytes are squeezed through the tiny aperture of this invagination (see Holland & Dan 1975) and accumulate in the ovarian lumen prior to evacuation (fig.l9b). Numerous studies of game.togenesis in both sexes of members of the Ophiuroidea have been made using light microscopy (Taylor 1958, Fenaux 1968a, 1970, 1972, Singletary 1970, Hendler 1973, Tyler 1977). Only one ultrastructural investigation deals with oogene­ sis (Kessel 1968) and none deal with spermatogenesis. Although they have not been dis­ cussed elsewhere, subdivisions of the spermatogenic epithelium may extend from the inner wall of the genital haemal sinus toward the testicular lumen. Seen with the light microscope, these are similar in appearance to spermatogenic columns found in asteroids (fig.24). These structures are also recognizable in micrographs provided by Patent (1968,1969), Fenaux (1970) and Tyler (1977). As is the case with light microscopic observations of the sperma­ togenic columns in asteroids, only germinal cells are apparent in these column-like struc­ tures. With the information currently available, there is no way to tell whether somatic cells are also present. Resolution of this point must await ultrastructural studies of the sper­ matogenic epithelium in this class. Although Davis (1971) did recognize an oogenic epithelium in Ophiothrix spicuIata, Ophioderma panamensis, and Ophiopteris papillosa, Patent (I 969, 1976) did not in Gor­ gonocephalus caryi. Patent stated that oocytes initially proliferate and differentiate in the genital rachis (equivalent in position in the gonad to the genital hemal sinus). Oocytes eventually come to lie in the ovarian lumen where they are closely surrounded by an incom­ plete layer of follicular cells (fig.l 5). When young, such cells in follicles are attached to the wall of the genital haemal sinus by nurse or attachment cells and seem to be contained in the sinus space. The similarity of the position and relationship of attachment cells and oocytes in ophiuroids to that of somatic and germinal cells in the ovaries of crinoids is striking. Electron microscopic observations of the structure and interrelationships of ger­ minal and somatic cells in this class are essential before firm statements can be made regarding the presence of subdivisions of the germinal epithelium. Finally, in Holothuroidea, subdivisions of the testicular epithelium have not been recog­ nized. Although both somatic and germinal cells are found, most studies emphasize the disorganized nature of the epithelium (Menker 1970, Atwood 1974, Green 1978) (fig.5b). Only by reference to the scanning electron micrographs of Krishnan & Dale (1975) are we reminded of the structural format of spermatogenic columns seen in asteroids.

460 Char/es W. Walker The progress of primary oocytes through gametogenesis in the synaptid, Rhabdomolgus ruber, seems to involve considerable shifting in their position (Menker 1970). Presumably, young oocytes are found in the genital hemal sinus. As vitellogenesis begins and continues, growing oocytes are said to migrate through the inner wall of the genital hemal sinus into the ovarian lumen. Menker's drawings slow elose association with oocytes and the inner wall of the genital hemal sinus during all stages of development (also see fig.9). Hamann (1884) provided a drawing of the gonad in Synapta digitata and showed oocytes in the genital he­ mal sinus. He also mentioned connections between oocytes and the wall of the gonad through a 'strang~ Micrographs in Hansen's (1975) study of the ovaries from several elasi­ pods and in Green's (1978) study of Leptosynapta tenuis showa similar relationship of oocytes to the inner wall of the genital he mal sinus. Green mentioned follicular cells that surround young oocytes and, following ovulation; he also identified a follicular membrane still associated with the gonad in the region of the inner wall of the genital haemal sinus. It would be most interesting to determine whether invaginations of the inner wall of the geni­ tal haemal sinus are present in holothuroids as they are in crinoids, and whether Hamann's 'strang' is equivalent to the aperture and somatic cells seen by Holland & Dan (1975) in C. japonica. Under the latter circumstances, the oocytes of holothuroids might seem to be within the genital hemal sinus although they would actually still be associated at one edge with non-germinal cells of the germinal epithelium in the gonadallumen.

2.2. Nutrient input to the gametogenic microenvironment and utilization by germinal cells In animals gene rally , gametogenesis is regulated by the interaction of factors derived from outside the gonad (extrinsic factors: e.g. environmental stimuli and nutrient and hormonal input) and those possibly originating locally from somatic and germinal cells themselves (intrinsic factors: e.g. substructure of the germinal epithelium and hormonal input or endo­ genous cytological rhythms); together, these constitute the structural, functional and bio­ chemical microenvironment influencing germinal cells at any given time (Roosen-Runge 1977). Even though considerable miscellaneous data are available on various aspects of reproduction in many echinoderms, no attempt has been made to place this information in the context of modern ideas concerning the microenvironmental organization of the ger­ minal epithelium (Roosen-Runge 1977). Such a task is essential to a thorough understanding of nutrient receipt and utilization by germinal cells. Undoubtedly, sub divisions of the germinal epithelium which exist during gametogenesis and which have been described or proposed for many echinoderms in section 3.1, provide the structural basis for organization of the microenvironment of single (primary oocytes) or small groups of differentiating germinal cells (primary oocytes or spermatogenic cells). Sub divisions form soon after spermatogonial or oogonial mitosis begins and throughout gametogenesis, change in structure depending on the association and interactioIl. of the ger­ minal and somatic cells which compose them (figs.13-21). It is logical to conelude that the generation and maintenance of sub divisions is a major prerequisite for successful gameto­ genesis in many echinoderms. Presumably, inter action between achanging suite of extrinsic and intrinsic microenvironmental factors occurs within such sub divisions to yield the maxi­ mum output of viable gametes. Several consecutive events must occur if nutrients are to be processed from ingested food-stuffs and ultimately delivered in usable form to germinal cells. In terms of the pre­ sent discussion, the most important of these events inelude accumulation of appropriate

Nutrition ofgametes 461 nutrients in extra- or intragonadal sites of storage (while active feeding occurs), subsequent mobilization of these nutrients from storage prior to or during gametogenesis, transport of nutrients to the gonad and transfer to the gametogenic microenvironment (if nutrients originate from extragonadal sites), and receipt and utilization of nutrients by germinal cells. The entire sequence of events involved in nutrition of gametes has not been determined for any echinoderm. Based on the structural and physiological evidence available, the following seetions will suggest possible mechanisms of nutrient transfer to the gametogenic micro­ environment and also possible mechanisms of nutrient uptake by germinal cells located there. Preliminary interpretations are given for asteroids and echinoids only; discussion of such mechanisms for crinoids, ophiuroids and holothuroids is premature. 2.2.1.rtsteroidea 2.2.1.1. Spermatogenesis. The concept of spermatogenesis presented here is based on the works of Pearse (1965a,b), Boolootian (l966), Mauzey (l966), Brusle (1968), Chia (l968, 1970), Delavault & Brusle (1968), Kim (1968), Crump (1971), Smith (1971), Jangoux & Vloebergh (1973), Walker (1973,1979,1980), and Jangoux & Van Impe (1977). The sper­ matogenic cycle in asteroids gene rally is completed in one year and includes a long aspermato­ genie phase (not seen in either Odontaster validus or Patiria miniata, Pearse 1965a, or pos­ sibly Ctenodiscus crispatus, Walker 1976a). During proliferative and differentiative phases, spermatogenic columns furnish a structural basis for organization of the spermatogenic microenvironment (Walker 1979, 1980). In each column, small numbers of germinal cells pass through successive phases of spermatogenesis in an orderly fashion. Since thousands of adjacent columns are similarly elaborated, nutrients and hormones might be delivered simultaneously to a large number of spermatogenic cells which are all at approximately the same spermatogenic stage. Remarkably precise and predictable coordination of successive steps of spermatogenesis might result. Extrinsic nutrient transfer to each spermatogenic column depends initially on the ingestion of food, which in different species is brought about through a variety of food gathering techniques and is often dependent upon seasonal cycles of active fee ding. Nutrients so derived are presumably assimilated into storage in the pyloric caeca or, in some cases, are delivered more directly to the gonads. In the former more common circumstance, nutrients are mobilized from storage be fore gametogenesis begins and they reach the testes where they accumulate in the genital hemal sinus. Finally, nu trients pass across the inner wall of the genital hemal sinus and are distributed to germinal cells composing the spermatogenic columns. The total volume of nutrients delivered annually to spermatogenic columns in all gonads is large and becomes available only as the result of radical changes in the overall physiolo­ gical condition of the individual. Concerning the release of nutrients from storage in aste­ roids and echinoids, Binyon (1972a, p.l70) stated 'profound changes in the metabolism of the animal from a predominantly synthetic one to one involved with the mobilization of such reserves (nutrients) and their translocation to other organs (gonads) is reminiscent of those changes found in higher animals resulting from the interplay of extrinsic and hormonal factors'. Ferguson (1975b), Jangoux& Van Impe (1977), and Walker (1980) have made sirnilar observations. Since numerous recent studies have demonstrated the presence of steroids and steroidogenic enzymes in the gonads of asteroids (Schoenmakers 1979, Schoenmakers et al. 1976, 1977) it is reasonable to accept Binyon's suggestion and to pos­ tulate the existence of an endocrine mechanism controlling the release of nutrients from

462 Charles W. Walker storage in the pyloric caeca or elsewhere. Such a mechanism might employ somatic cells of the germinal epithelium as a steroidogenic tissue. These cells are involved in the structure of spermatogenic columns (where they are called axial somatic cells) and during the degra­ dation of columns they actively phagocytize spermatozoa, spermatids, and even spermato­ cytes (Walker 1980) (fig.17a). Since no other cellular candidates are present in the testicu­ lar lumen during the aspermatogenic phase and since no cells migrate into the testis from other sourees, somatic cells mayaiso be equivalent to or give rise to the 'vesicular cells' seen by Cognetti & Delavault (1962) and by Brusle et al. (1970). During the aspermatogenic phase, a predictable series of interactions occurs between somatic and germinal cells which convert the testes from its phagocytic post-spawning condition through a secretory stage (when 'vesicular cells' are present) to its pre-spermatogenic status (Walker 1980). In the mammalian testis, moleeules salvaged during phagocytosis of residual spermatozoa may be utilized by Sertoli cells in the synthesis of steroids that are subsequently active during sper­ matogenesis. The same kind of efficient utilization of recycled substances may result from the intense phagocytic activities of functionally analogous somatic cells found in the aste­ roid testis. 'Vesicular cells', presumably derived from phagocytic somatic cells are secretory in character. In frozen seetions of unfIXed testes prepared at this time and assayed histo­ chemically for ~S-3ß-hydroxysteroid dehydrogenase (HSD) activity, a strong blue diforma­ zan deposition is localizable within the cytoplasm of many of the luminal secretory 'vesi­ cular cells' and sirnilar cells in the genital haemal sinus (Walker 1980). Testicular steroido­ genesis occurs when nutrients are passing between pyloric caeca and the testes (Schoen­ makers et al. 1976,1977, Walker 1980). Under stimulus from an appropriate extrinsie fac­ tor (perhaps decreasing ambient temperature), steroids might be released from 'vesicular' somatic cells in the testis, travel to the pyloric caeca, and there initiate and maintain mobi­ lization of stored nutrients. Nutrients derived in this fashion cou1d enter the gonad and accumulate in the genital hemal sinus. A sirnilar mechanism has been proposed to explain the predictable release of stored nutrients from the pyloric caeca in female asteroids (Schoenmakers 1979) and may even be important in mobilizing nutrients from nutritive phagocytes in echinoids, from the gonadal and body wall accessory cells in crinoids, and from the gut wall and visceral peritoneum in holothuroids. In structural terms, spermatogenic columns in the testes of asteroids are composed of germinal cells that slJrround one or more axial somatic cells. The central core of each column includes elongated cytoplasmic processes from germinal cells that reach down to­ ward the inner wall of the genital haemal sinus at the base of each column (figs.12,14,16, 17a). All processes are filled with profiles of smooth endoplasmic reticulum, microfilaments, microtubules, lipid droplets and elongated mitochondria; near the genital haemal sinus, they also appear to form endocytotic vesicles. While no experimental proof is available, it is likely that the central core of each column functions as a temporary intragonadal distribution sys­ tem transporting nutrients from the genital hemal sinus to developing germinal cells. located at various heights along the columns. Furthermore, when such nutrients are no longer present, columns apparently become irrelevant and degrade as those germinal cells still forming them complete their differentiation (Walker 1980). There is no specific infor­ mation available on the uptake of nutrients by spermatogenic cells.

2.2.1.2.0ogenesis. The following interpretation of oogenesis in asteroids is based on the studies of Cognetti & Delavault (1962); Pearse (196Sa,b); Kim (1968); Smith (1971); Achi­ tuv (1973); Jangoux & Vloebergh (1973); and Walker (1973, 1974a, 1976a, 1980). Depend­

Nutrition 01 gametes 463 ing on the species considered and on the means of reproduction employed, asteroids utilize one of at least three oogenic strategies (Pearse 1965a). The first of these is seen in species of Asterias and Pisaster and is sirnilar to the situation just described for the testes of most asteroids (fig.17b). Oogenesis is completed in one year, and a long agametogenic phase intervenes between annual gametogenic events. Nutrients ultimately utilized in oogenesis are acquired through feeding and accumulate as reserves stored in the pyloric caeca prior to and during gametogenesis. As differentiation of oocytes commences, ovarian follicles presumably serve as the structural basis for organization of the oogenic microenvironment (fig.17b). During gametogenesis, the structural relationship of somatic and germinal cells changes according to the degree of differentiation of primary oocytes. Within follicles, primary oocytes experience two major periods of growth. During early or initial growth, oocytes increase slightly in size, do not accumulate appreciable yolk reserves and remain transparent when viewed in squash preparations from living ovaries. Toward the conclusion of this period, enlarging oocytes protrude into the ovarian lumen but maintain contact with the indented inner wall of the genital hemal sinus by a cytoplasmic stalk (fig.l3). Ultrastructural observations on Asterias rubens by Schoenmakers (1979) indicate that the plasma membrane of the stalk near the genital haemal sinus is capable of endocytosis. Membrane-bound, PAS-positive vesicles are more abundant in cytoplasm nearest the inner wall of the genital haemal sinus. Eventually, oocytes enter a vitellogenic growth period during which ovarian follicles are detached from the inner wall of the genital haemal sinus and come to lie in increasing numbers in the ovarian lumen (figs.17b,26). Oocytes in the follicles enlarge rapidly, become opaque and intensely PAS-positive. In their histological preparations, Walker (1973,1979) and Schoenmakers (1979) recognized a coagulated, PAS-positive substance, perhaps derived from the genital hemal sinus or from the lysis of abortive oocytes, that is distributed randornly among vitellogenic oocytes external to their follicular coverings. This substance contains amoeboid cells, called 'nurse cells' by Schoen­ makers (1979) (fig.25). These cells are never connected to oocytes or follicle cells by cyto­ plasmic bridges. With the exception of these points, there is no information available on the uptake of nutrients by the oocytes of asteroids at various stages of differentiation. As a result, it is impossible to suggest the relative amounts of time each oocyte spends in auto­ synthetic or heterosynthetic yolk formation or how these processes correlate with the initial and vitellogenic periods of oocyte growth. The numerous oocytes which result from oogenesis are equal in size, are relatively small and oligolecithal, and are neariy all evacuated each year. While conspicuous amounts of nutrients are not stored in the germinal epithe­ lium, substances resulting from the phagocytosis of residual oocytes might be available for use during the ensuing oogenic effort. Schoenmakers et al. (1978) and Schoenmakers (1979) described putative steroid secreting cells in the ovarian lumen in A.rubens, and also demonstrated varying estrogen levels and steroid metabolism in the ovaries and the pyloric caeca. They concluded that an estrogen is synthesized in the ovary (although such synthesis has not yet been demonstrated, Voogt personal communication) and may be im­ portant during vitellogenesis and in promoting the mobilization of nutrients from storage in the pyloric caeca. An alternative oogenic strategy is seen in some members of the Echinasteridae, Asterini­ dae, and Astropectinidae and probably occurs in several other farnilies (fig.26). Here, nutrients are not obviously stored in the pyloric caeca. Instead, nutritive substances utilized in gametogenesis are delivered more directly to the ovaries and then stored in the germinal epithelium in somatic 'vesicular' phagocytic cells and, indirectly in oocytes that are des­

464 Charles W. Walker

tined to abort (Cognetti & DelavauIt 1962, Pearse 1965a,b, Smith 1971, Achituv 1973, and Worley et al. 1977). Finally, a third strategy is encountered in Odontaster validus (pearse 1965a). In the absence of 'vesicular' storage cells, nutrients are derived from abortive oocytes and also continuously from the pyloric caeca. Species employing either of the latter two strategies lack an agametogenic phase between successive oogenic events, and oocytes require at least two years to differentiate fully. As a result, ovaries in these animals contain two major size classes of oocytes during most of the year, one representing new cells in the initial phase of growth (first year) and the other representing cells one year old and in the vitel­ logenic phase of growth (second year). üften differentiated second year oocytes are very large, macrolecithal and opaque, and they are suspended in a coagulated PAS-positive sub­ stance external to their follicle cells (fig.26). From studies of several species in the latter two categories, it is obvious that not all cells present after the first year of growth ultimately become fully differentiated oocytes. In­ stead, many vitellogenic oocytes abort (e.g. about two-thirds in O.validus, Pearse 1965a) and presumably yield nu trients which are then utilized by other more successful oocytes which continue differentiating. In some species and not in others, it appears that phagocytic cells in the lumen of the ovary (derived from follicular cells?) (Smith 1971) may participate in the degradation of abortive oocytes. Variations in detail of all three strategies occur. In his study of Leptasterias pusilla, Smith (1971) estimated that more than three-quarters of all oocytes formed are aborted and that their nutrients are utilized du ring the differentiation of successful oocytes. He proposed (p.127) that, 'since primary oocytes are produced ini­ tially almost two years prior to being released as ripe oocytes, there is a mechanism for adjusting the balance between exogenous resources and the numbers of developing oocytes such that a maximum number of oocytes can complete matura ti on' . This mechanism must be precisely controlled. Abortion of vitellogenic oocytes occurs while cells in their first year of growth and other vitellogenic oocytes are also present and remain healthy. More­ over, cells in their first year of growth do not begin vitellogenesis prematurely when the process commences normally in second year oocytes. Intact ovarian follicular cells surround all oocytes that continue successfully through differentiation, while follicular cells are always lost from and may actually phagocytize unsuccessful oocytes. It is clear that success­ ful growth and differentiation of oocytes in asteroids depend on the existence and mainte­ nance of ovarian follicles and that this, in turn, is indicative of a healthy microenvironment. It is possible that the mechanism of oocyte abortion and digestion is regulated by substances like I-methyl adenine which are released from surrounding intact ovarian follicles in res­ ponse to low levels of gonad stimulating hormone (GSS) (Harai & Kanatani 1971, Harai et al. 1973, Cochran & Kanatani 1975). 2.2.2. Echinoidea In both sexes of most echinoids, nutritive phagocytes usually grow in size and store nutri­ ents before gametogenesis can occur (figs.18a,b, 20, 21). During gametogenesis, the degree of development of nutritive phagocytes varies inversely with the numbers (testes) and size and numbers (ovaries) of differentiating gametes (Holland & Holland 1969). Nutritive pha­ gocytes clearly form a structural basis for organization of the microenvironment of germi­ nal cells in both sexes of the echinoids. In either case, nutrient input to the microenviron­ ment is derived directly from ingested food. Resulting nutrients are assirnilated into storage in nutritive phagocytes both prior to and during gametogenesis. Although there are extra­

Nutrition o[ gametes 465 gonadal sites of nutrient storage in echinoids, these are not obviously involved with supply­ ing nutrients to developing germinal cells (Farmanfarmaian & Phillips 1962, Lawrence et al. 1966). For example, in Strongylocentrotus purpuratus, the gut probably acts as a storage organ, 'responsive to immediate needs of a starving animal and as a region for immediate deposit of excess nutrient' (Lawrence et al. 1966, p.289). Many investigators endorse the assumption that nutritive phagocytes store nutrients for use du ring gametogenesis in PA8-positive, eosinophilic globules and as cytoplasrnic glycogen (Russo 1907, Caullery 1925, Cuenot 1948, Fuji 1960a, Holland 1965b, Holland & Giese 1965, Takashima & Takashima 1965, Vehrey & Moyer 1967, Millonig et al. 1968, Nif>rre­ vang 1968, Anderson 1968, Takashima 1968b, Longo & Anderson 1969, Holland & Holland 1969, Chatlynne 1969, 1972, Pearse 1969a,b, 1970, Bal1970, Kobayashu & Konaka 1971, Tsukahara 1969, Gonor 1973a,b, Guidice 1973, Cruz-Landim & Beig 1973, Piatigorski 1975, Fenaux et al. 1977, Masuda & Dan 1977, Lane & Lawrence 1979, Pearse in press). Nutrient input to these specially adapted storage cells occurs prior to and also during game­ togenesis (Takashima 1976). Studies by Bennett & Giese (1955) on Strongylocentrotus [ranciscanus and by Moss & Lawrence (1977) on Mellita quinquesper[orata indicate peaks in biochemical constituents of the coelornic fluid and gonads, respectively, preceding game­ togenesis. Such results can be interpreted as evidence of nutrient supply to and storage by the nutritive phagocytes. Furthermore, three days after intracoelomic injection of tritiated d-glucose, Takashima (1976) found labelled nutritive phagocytes indicating that they were acquiring nutrients during gametogenesis. Such cells clearly ingest deteriorating oocytes during oogenesis (Chatlynne 1971) and residual spermatozoa and oocytes after the evacua­ tion of gametes (Holland 1965b, Holland & Giese 1965, Beig & Cruz-Landim 1975, Fenaux et al. 1977, Masuda & Dan 1977). There are no ultrastructural studies available which des­ cribe the cytology of nutritive phagocytes during the pre-gametogenic sequestering of nutri­ tive reserves. Both estrogen and progesterone have been found in the gonads of echinoids (Botticelli et al. 1961); these may be involved in vitellogenesis or in promoting the mobili­ zation of nutrients.

2.2.2.1. Spermatogenesis. Other than the preliminary ultrastructural efforts of Longo & Anderson (1969) and Beig & Cruz-Landim (1975), studies on spermatogenesis in echinoids are based entirely on light rnicroscopy. There is no information on intercellular transfer of nutrients from storage in nutritive phagocytes to spermatogenic cells. At most, it can be said that structural interdependency exists between these cells. In most echinoids, sperma­ togenesis is completed annually. The autoradiographic studies of Holland & Giese (1965) which are based on observations of spermatogenic cells of Strongylocentrotus purpuratus treated with tritiated thyrnidine, show that primary spermatocytes differentiate into sper­ matids in six days and that spermatozoa result through subsequent sperrniogenesis in about eight more days. Since spermatozoa are produced throughout this period at the same pace between the plonths of August and March, it is clear that in this species, differentiation of one generation of spermatogenic cells must occur simultaneously with the production of new cells (this is not the case in Stylocidaris a[finis, Holland 1967). This situation is similar to that seen in the testes of Asterias vulgaris (Walker 1980) and obviously requires signifi­ cant nutrient input. 2.2.2.2. Oogenesis. In most echinoids, annual cycles of oogenesis occur which yield rela­ tively small (100-200 pm) ova. An agametogenic phase often intervenes between major

466 Charles W. Walker oogenic efforts, during which, among other things, residual oocytes are digested by nutri­ tive phagocytes. In each ovary, the initiation of oogenesis may be simultaneous among most of the oocytes or may be asynchronous so that oocytes act independently. Fortu­ nately, many authors have utilized the electron microscope in investigating the process of nutrient transfer from storage in the nutritive phagocytes to uptake by developing primary oocytes (Takashima & Takashima 1965, Vehrey & Moyer 1967, Millonig et al. 1968, Taka­ shima 1968b, Ba11970, Kobayashu & Konaka 1971, Tsukahara 1971, Chatlynne 1972, Cruz­ Landim & Beig 1973, Masuda & Dan 1978). Two growth phases are recogniZable during oogenesis in most echinoids. These are early growth (premicrovillous stage of Chatlynne 1971) and vitellogenesis (microvillous stage ofChatlynne 1971). Before the early growth phase begins, nutritive phagocytes are maximally developed and are filled with glycogen granules (40-60 nm); PAS- and MPB-positive, eosinophilic, membrane-bound globules; mitochondria; a single nuc1eus; and rough endoplasmic reticulum. At this time, they are encirc1ed basally by tightly packed c1umps of oogonia. During early growth, oocytes en­ large tremendously and are surrounded by processes of the nutritive phagocytes. The cyto­ plasmic volume of the oocytes increases as ribosomes and rough endoplasmic reticulum pro­ liferate extensively. It has been suggested that the considerable development of such orga­ nelles is essential for autosynthetic (intra-oocyte) yolk synthesis. Granular yolk is developed when oocytes are so constructed. There is no direct evidence as to the source of nutrients used during early cytoplasmic enhancement of oocytes. Nutritive phagocytes release glyco­ gen granules through holes(?) that develop through~ut gametogenesis in the plasma mem­ brane. These granules pass into the spaces between oocytes and nutritive phagocytes. Many authors show the intake of granules by endocytotic vesic1es formed by the plasma mem­ brane of the oocyte. Vitellogenic growth follows, with extensive formation of microvilli on the surface of the oocytes. Endocytotic (pinocytotic) vesic1es are obvious in association with these cytoplasmic extensions. No one has discussed the transfer of substances from the PAS-positive globules to germinal cells. Numerous Golgi bodies are recognizable in the cytoplasm of vitellogenic oocytes when yolk platelets begin to appear. Throughout these changes, the nutritive phagocytes shrink and lose their nutrient reserves.

3. CONCLUSIONS Most studies of reproduction in the echinoderms are only peripherally related to gameto­ genesis and very few were specifically designed to investigate the nutrition of germinal cells. Understandably, no attempt has been made to consider the means of nutrient supply to germinal cells as part of an overall mechanism governing gametogenesis. The intentions of the present work have been to provide a concise statement of the structural background necessary for a meaningful discussion of gamete nutrition and to propose strategies of gametogenesis-related nutrient distribution, storage, and utilization based on available infor­ mation. Conc1usions drawn about dynamic processes must be regarded as tentative until they can be tested experimentally. Several facts have emerged from this comparative study which suggest that the overall mechanism of gamete nutrition is ultimately based on the sirnilar morphology of relevant somatic and germinal portions of the gonad but is variable in terms of which tissues in dif­ ferent groups of echinoderms are involved in nutrient storage. In asteroids, echinoids, and ophiuroids, the gonads are composed of both outer and inner sacs. The outer sac has a non­

Nutrition

0/ gametes 467

nutritive, structural role during the annual reproductive cycle. Gonads in all classes include an inner sac where gametogenesis-related activities occur. The inner sac is universally com­ posed of the intirnately associated genital hemal sinus and germinal epithelium, which together form the only components of gonads in crinoids and holothuroids. The genital hemal sinus is interconnected with the same space in other gonads and with similar spaces in most major body organs. Such a characteristic structural relationship suggests that the haemal system in echinoderms may be involved in synchronizing the gametogenic activities of different gonads or of different parts of the same gonad, perhaps by providing a common route for the transfer of substances like nutrients and hormones. The inner wall of the genital he mal sinus lies in direct contact with somatic and germinal cells of the germinal epithelium throughout the gonad, and the sinus is often mIed with PAS- and MPB-positive substances which are presumably nutritive in character. It is concluded that the genital hemal sinus probably plays a significant role in intragonadal nutrient distribution and storage in many echinoderms. In preparation for prolific and rapid gametogenesis, echino­ derms temporarily store nu trients intragonadally and in a variety of other locations as well (e.g. ovarian 'vesicular cells' in asteroids; ovarian and testicular nutritive phagocytes in echi­ noids; gonadal and body wall accessory cells in crinoids; and, possibly, thevisceral peritoneum of the inner sac of the gonad in holothuroids). Nutrient reserves are laid down when the meta­ bolism of the animal is geared to synthesis of storage products. Except in asteroids, where nutrients are sequestered in the pyloric caeca, there is little evidence for extensive nutrient storage in extragonadal sites in other classes of the phylum. Apparently, once nutrients are positioned intragonadally, the gonads of many echinoderms are relatively self-sufficient in terms of their major nutritional requirements for gametogenesis (in all cases, some level of continuous, direct input from the gut may be involved). Within the sac-like gonads of most echinodeims, millions of germinal cells are elaborated, differentiated and stored on an annual basis. It is at first difficult to imagine how prolific and orderly gametogenesis can occur in these structurally simple entities. In a11 recognized examples, subdivision of the germinal epithelium into small groups of interrelated somatic and germinal cells occurs soon after gonial proliferation commences. It would appear that sub divisions are essential for successful gametogenesis in many echinoderms. Sub divisions are a means of increasing the surface area in the interior of an otherwise structureless gonadallumen and they provide a variety of orderly methods for bringing nutrients to relatively small numbers of germinal cells. Sub divisions vary in structure between sexes and, in females, between classes of the phylum. In the testicular epithelium, they extend from the inner wall of the genital he mal sinus into the testicular lumen. Where adequate information is available (for asteroids and echinoids), it is obvious that spermatogenic cells surround somatic ce11s and are intimately associated with processes from them (axial somatic cells in asteroids and nutritive phago­ cytes in echinoids). In the ovaries of echinoids, a sirnilar relationship between nutritive phagocytes and oocytes is seen. Asteroids employ somatic cells in the construction of ovarian follicles, each of which contains a single oocyte. These subdivisions also extend into the lumen of the gonad. An entirely different situation is encountered in crinoids, ophiu­ roids, and holothuroids. Here, oocytes apparently bulge into the genital hemal sinus, while., remaining in contact with somatic cells in the ovarian lumen. The position of differentiating oocytes lumina11y in asteroids and echinoids and in association with the genital hemal sinus in crinoids, ophiuroids and holothuroids might indicate that the mechanism of nutrient transfer to oocytes is fundamentally different in these two groups of classes.

468 Charles W. Walker Sub divisions of the germinal epithelium may have significance greater than that in gamete nutrition. Based on the changing structure of subdivisions, it is likely that single or small groups of germinal cells within them are either directly or indirectly affected by a changing suite of factors, like environmental and hormonal stimuli, that may promote, maintain or curtail specific gametogenic events. In other words, transduction of microen­ vironmental input to the gonad seems to occur at the structurallevel of sub divisions and probably results in the organized, tempo rally separated events of gametogenesis. In the asteroids and possibly in the echinoids, somatic cellular components of sub divisions seem to be pluripotent and, at various times of the year, are intimately involved in the structure and activities of sub divisions, in carrying out extensive phagocytosis, and in mediating extrinsic (e.g. gamete shedding substance and nutrients) and in contributing intrinsic (e.g. I-methyl adenine and steroids) microenvironmental factors influential during gametogene­ sis. Gonads of echinoderms should provide favorable preparations for use in the investiga­ tion of gametogenesis more generally because they are numerous, structurally uncompli­ cated, and easily cultured organs in which enormous numbers of germinal cells pass syn­ chronously through successive steps in the process. Recent fmdings indicate that few of the structural or operational features associated with specialized kinds of reproduction (such as internal fertilization) are present and the gonads of echinoderms provide an excellent model system for observing events such as mitosis, meiosis, and cellular differentiation that are basic to gametogenesis. The preceding discussion has identified several major gaps which persist in our current understanding of gamete nutrition in the echinoderms. It is hoped that the structural and functional relationships suggested here will provide a framework for future critical studies of the sequence of events involved in supplying nutrients to gametes du ring their prolifera­ ton and differentiation and also of the overall mechanisms governing gametogenesis in echinoderms.

ACKNOWLEDGEMENTS I thank Drs H.Kanatani, D.M.Patent, J.S.Pearse, H.Schoenmakers, P.Schroeder and P.Voogt for discussion of the subject covered in this chapter; Drs J .M.Anderson, N.D.Holland, M. Jangoux and J .M.Lawrence for criticism of the chapter; Drs P.A.Taylor, J .D.Green, F .S.Chia and Ms L.Bickell for providing material and photographs; N.Jendrock, V.Perron, J.Douillet and E.Wong for their assistance; Central University Research Fund and Hubbard Fund (Uni­ versity of New Hampshire) for financial support in preparation of the chapter.

Notes on the plates between pages 456 and 457

Figures 9, 11, 13, 15, 22, 25 and 26 are light micrographs of 10 lJlTI paraffin sections taken of ovaries and testes from a variety of echinoderms and stained with Mallory's phosphotungstic acid hematoxylin (PT AH) (figs. 11 and 13 have also been treated with the PAS routine). Figures 1b-3b, 4b,c, Sb, 10, 14, 20, 21 and 24 are light micrographs of 11JlT1 plastic sections taken of the walls of ovaries and testis from a variety of echi­ noderms and stained with buffered azure B. Figures 6-8,12,16 and 23 are electron micrographs taken from thin sections showing the germinal and somatic components of the wall of the testis in Asterills vulgaris, Ctenodiscus crispatus and F1orometra serritissima and 'contrasted' with uranyl acetate and lead citrate.

21

GOFFREDO COGNETTI

NUTRITION OF EMBRYOS

1. ECHINOIDS

1.1. General considerations The early period of echinoid development is characterized by a very high rate of cell divi­ sion. In Arbacia punctulata, at 23°C, the fIrst division takes place 50 minutes after fertili­ zation, and subsequent cleavages every 22-32 minutes (Harvey 1956). From the zygote through the pluteus stage, which is reached about 24 hours after fertilization, about 12 c1eavages occur, but the change in mass is negligible. After the completion of the intestine, plutei start feeding and an increase in weight begins. It therefore seems conceivable that some macromolecules which are to be utilized later are synthesized during oogenesis. For years the general idea has been that the synthesis and storage of nutrient macromolecules during oogenesis, and their subsequent breakdown during embryogenesis, occur to main­ tain the internal pool of the embryos with precursors as the synthetic activity of macro­ molecules by the embryo proceeds. It has been shown however that proteins such as his­ tones (Cognetti et al. 1974, 1977a), tubulin (Cognetti et al. 1977b) and ribosomal proteins (Cognetti et al. unpublished) are synthesized by the oocyte. Therefore the synthetic activity during the oogenesis involves the accumulation of proteins which will be directly utilized, together with the storage of nutritional substances which will be degraded. Beside the utilization of stored macromolecules, many precursors (such as amino acids, nuc1eotides, sugars) can be taken up by the embryo and then enter the metabolic pathways. The question is whether this phenomenon, connected with the membrane permeability, can be considered a nutritional event. In effect, the membrane permeability changes during development, and the lowest degree of permeability is found with the eggs (Nakano & Mon­ roy 1957, 1958a,b, Monroy & Nakano 1959) where no nutritional events are required. The permeability sharply increases after fertilization. In addition, there is much evidence that the movement of molecules through the membrane is an active transport. Membrane per­ meability is also a function of uptake, in the sense that certain internal concentrations of a precursor selectively inhibit its further uptake (Piatigorsky & Whiteley 1965, Mano et al. 1977). Nevertheless, this manner of nutrition is c1early subordinate to that of recycling pre­ cursors obtained from the breakdown of accumulated macromolecules. It must be conc1uded, in fact, that the uptake of precursors from the medium is not I'\ecessary for the survival of the embryos, which develop normally and at anormal c1eavage rate when grown in artificial sea water containing no organic molecules. From an evolutionary point of view, the low con­ centration in sea water of molecules such as free amino acids, sugars or nuc1eotides, would have adversely affected the survival if the nutritional mechanism were only, or mainly, based on the uptake of these rare precursors from the surrounding medium. Uptake and sub se­ 469

470 Goffredo Cognetti quent digestion of macromolecules have not been described before the completetion of the intestine, and therefore is not considered here. 1.2. The yolk It is generally accepted that the site of storage of the nutritive macromolecules synthesized during oogenesis is the yolk. 1.2.1. Oil Droplets (or fatty yolk) are found in the cytoplasm of oocytes, eggs and embryos, and were estimated by Afzelius (1957) in egg of Psammechinus miliaris to be on the order of several thousand. They are irregular bodies, with a diameter between 0.4-1Ilm and lack a surrounding membrane. Many oll droplets may coalesce, forrning a larger droplet. Pasteels et al. (1958), in Paracentrotus lividus and Gross et al. (1960), in Arbacia punctulata reported that oll droplets are associated with mitochondria in the egg, and later scatter throughout the cytoplasm. The main component of oll droplets of eggs of A.punctulata are triglycerides (Harvey 1956). Chelomin & Svetashev (1978) reported the composition of the oll droplets of Strongylocentrotus intermedius to be 71.8 % triglycerides; 14.0 % phos­ pholipids, 3.3 % cholesterol. tittle is known about the fate of oll droplets during the deve­ lopment. Pelluet (1938) reports a decrease in the number of oil droplets from egg to four blastomere stage, but a slight increase at blastula. This might be explained as a turnover of lipid. 1.2.2. Yolk platelets (or yolk spherules, yolk particles, yolk granules): are electron-dense granules, distributed more or less uniforrnly throughout the cytoplasm (Afzelius 1956a, Monroy & Maggio 1963). They are present in large amounts, ranging between 27.2 % and 38 % of the total egg weight (Harvey 1932, Costello 1939). Intact yolk particles, isolated from eggs of Arbacia punctulata by means of differential centrifugation of the total homo­ genate, appear as spherical or ellipsoidal structures, 1-1.7 Ilm long and yellow to brown in color (Gross et al. 1960). The internal ultrastructure is gene rally granular; sometimes, espe­ cially in large size egg species, with lipid-like inclusions (Lönning 1976). The yolk platelets are surrounded by a thin granular peripheral membrane, about 5 nm across the dense line in profJ.le, which is often lost during isolation (Gross et al. 1960, Monroy & Maggio 1960). A heterogeneity in yolk particles population has been described: in eggs of A.punctulata large particles with a rather granular structure occur, together with others that are less dense and more uniform in internal structure (Gross et al. 1960). A difference in sedimen­ tation rate between two classes of yolk granules is also present in eggs and embryos of Lythechinus variegatus (Krischer & Chambers 1970) and in eggs of Strongylocentrotus, purpuratus (Schuel et al. 1975). These authors also reported a class-specific distribution of some enzymes. The composition of the yolk platelets has been analyzed primarlly by histochemical techniques. Monne & SlautJerback (1950) reported that the yolk of eggs of Paracentrotus lividus was composed of amino-polysaccharides, associated in a protein-arninosugar-lipid complex. Polysaccharides also occur in Hemicentrotus pulche"imus (Sakai 1956, Taka­ shima 1971) and in Heliocidaris crassispina (Takashima 1971). Mucopolysaccharides and sialic acid occur in Arbacia lixula, Echinocardium cordatum, Echinus esculentus, Paracen­ trotus lividus, and Psammechinus miliaris (Immers 1960). Unfortunately no serious frac­

Nutrition oi embryos 471 tionation of the carbohydrates has been done from isolated yolk particles, so little is known about the nature of these polysaccharides. Glycogen has been detected by Takashima (1968a, 1970). The lipid composition in yolk platelets from eggs of Strongylocentrotus intermedius is 51.3 % triglycerides; 22.8 % phospholipids; 7.2 % cholesterol. The predominant phospho­ lipid is diacylphosphatidylcholine (58.3 %) (Chelomin & Svetashev 1978). The lipid/protein ratio (w/w) is 0.77. The protein level in yolk granules of eggs of S.purpuratus is about 30 % of the total proteins (Schuel et al. '1975). Fry & Gross (1970b) reported that 50-80 % of the total egg proteins of A.punctulata arecontained in the yolk; the authors themselves, however, pointed out the coarseness of the data. The presence of glycoproteins in the yolk has been reported for several echinoid species (Immers 1960, Kondo 1972, Schuel et al. 1975). Malkin et al. (1965) isolated a high molecular weight (894,000) protein with a cristalline structure from the yolk granules of eggs and embryos of S.purpuratus, A.punctulata and L.pictus. A very similar protein (24 S) occurs in the yolk platelets of Hpu/che"imus eggs (Ii et al. 1978). Associated with it in the yolk are three water-soluble lipoproteins, rich in lipids and carbohydrates. This finding is consistent with the hypothesis that these proteins are utilized for energy during development (Ii et al. 1978). The relationship between yolk granules and lysosomes is not clear. If they are distinct organelles, they may interact to digest stored 'food'. In fact, the presence of hydrolytic enzymes has been detected in the yolk platelets of echinoids. Their presence can be inter­ preted in two ways: either yolk granules and lysosomes fuse together, as in other organisms, or both enzymes and substrates coexist, from the beginning, in the same organelle. In this case no distinction would exist between lysosomes and yolk particles. At our present stage of knowledge, it is impossible to prefer one hypothesis to another.

1.3. Formation oi the yolk platelets Echinoid oocytes are commonly divided into three classes: primary, growing, and terminal (Cowden 1962). The beginning of vitellogenesis occurs relatively late, during the 'growing' phase (Cowden 1962, Esper 1965). In oocytes of Hemicentrotus pulchemmus, the yolk platelets appear to consist of a light and a dense part, at the beginning of the vitellogenesis (Takashima 1971). The dense part is an aggregate of coarse electron-dense granules, 20-30 nm in diameter. The platelets undergo an evolution, since in the mature eggs the granules appear finer, being 15-20 nm in diameter. The mechanism of formation of the yolk par­ ticles is quite unclear. N41rrevang (1968) who classified the mode of formation of the yolk, assigned the echinoids to the class of organisms which form yolk by accumulation of mate­ rial within pre-existing vacuoles. He did not exclude, however, the possibility of a derivation of the vesicles from the plasma membrane, through a micropinocytic mechanism. Although Verhey & Moyer (1967) claimed that no visible transfer of material occurred from accessory cells to the oocyte by pinocytosis, Takashima & Takashima (1966) found a pinosome in early oocytes of H.pulche"imus and of Pseudocentrotus depressus: this pinosome, however, was no longer observable at later stages. Two types ofpinocytic vesicles were successively detected in oocytes of H.pulchemmus: aclass of small pinosomes, 100 mm in diameter (a-pinosomes) and a class oflarge ones, 500-700 mm in diameter (ß-pinosomes) (Tsukahara & Sugiyama 1969). The ß-pinosomes seem to be involved in the formation of the cortical granules. The a-pinosomes (Tsukahara 1970) instead, move into the inner

472 Goffredo Cognetti cytoplasm, and their coating material disappears. Some a-pinosomes then surround an immature yolk platelet, attach to its surface and fuse with it. This might explain their dis­ appearance, observed by Takashima & Takashima (1966) and the lack of observation of pinosomes by Verhey & Moyer (1967). No a-pinosomes are subsequently found around mature yolk platelets (Tsukahara 1970). Concerning the nature of the material absorbed by the oocyte by pinocytosis, evidence has been presented that particles of glycogen are discharged by the nurse ceIls into the interceIlular space and incorporated by pinocytosis into the oocytes (Takashima & Takashima 1965, Takashima 1968, 1970, 1976, Takashima & Torninaga 1975). This of course does not exclude the possibility that other macromole­ cules, such as proteins or lipids, rnight be incorporated as weIl. The origin of the immature yolk platelets is totally unclear because of the lack of defi­ nitive experiments. One possibility is that yolk platelets are formed in the Balbiani body, as supposed by Millonig (1957), and therefore would originate from the rough endoplasrnic reticulum, whose polyribosomes rnight be the source of the yolk proteins. According to another hypothesis, the immature yolk platelets originate in the cisternae of the Golgi apparatus, since they are always observed to be near them (Tsukahara 1970); early work by Afzelius (1956b, 1957), indeed, reported the fusion of Golgi vesicles with endocytic vacuoles as the origin of the yolk particles. An interesting aspect of the proposed derivation of the yolk platelets from the Golgi vesicles is the possibility that the accumulation of nutritional matter would occur within a lysosome: the lysosomes are, of course, of Golgi derivation. The presence of hydrolytic enzymes within yolk spherules supports this idea. According to this hypothesis, the immature yolk platelet could be a lysosome. Its enrich­ ment with nutritional substances would transform it into a yolk platelet, which would be­ come a lysosome again after the completion of the digestive process. To accept this theory, however, one must postulate some type of internal compartmentalization of the lysosome, as weIl as a regulation of the mechanism of interaction between the enzymes and the stored matter. At the present stage of our knowledge, no definitive proof exists for or against this point of view (for a general review on the relationships between yolk and lysosomes, see Pasteeis 1973). In addition, the existence of at least two populations of yolk spherules, dif­ ferent in size and in enzymatic content, indicates that many more points have to be clari­ fied on the origin of the yolk platelets in echinoid oocytes. A second problem, but directly connected to the preceding one, regards the synthesis of the macromolecules stored in the yolk particles. If, as seems indicated by the above data, pinocytosis is directly involved in yolk formation, it is·conceivable that nurse cells are the site of synthesis of the matter contained in the yolk. On the other hand, because of the broad variety of substances existing within a yolk platelet, the possibility of a synthesis at least of some of them by the oocyte should be considered, The above observations, done by electron rnicroscopy or histochernical staining, do not give sufficient information. Defi­ nitive biochernical experiments have not been performed. A reason for this lack of informa­ tion rnight be due to the fact that a technique which yielded a purified culture of echinoid oocytes in vitro was not available in the past. Methods have been described (Sconzo et al. 1972, Cognetti et al. 1977a) but no new experiments on this problem have been performed. Data presented by Esper (1962) on the incorporation of 14C-glucose into the yolk of oocytes and eggs of Arbacia punctulata do not resolve the question, because the radioactive precursor was administered to live females and sections of the gonads were examined at a later time. Incorporation into the yolk found under these conditions may indicate either transfer from the gonadal ceIl or endogenous synthesis by the oocyte.

Nutrition o[ embryos 473 1.4. Utilization o[ stored macromolecules The general lack of precise information about the origin and the chemical nature of the yolk in echinoderms leads to uncertainty about yolk utilization. The observation that the yolk components of eggs of other organisms are progressively degraded and the degradation products re-utilized as food by the developing embryos (Gross 1954) applies to echinoids mostly by analogy. Nevertheless some proof ofyolk lysis exists, but for the most part no direct evidence has been presented that the nutritional molecules, whose degradation has been demonstrated, were originally stored in the yolk platelets, or elsewhere. 1.4.1. Evidence o[ utilization o[ pro teins by embryos The disappearance or the decrease in level of proteins during development generally can be considered as a nutrition al process, since no free or bound nitrogen leaves the cell during echinoid development (Gustafson & Hasselberg 1951). Nevertheless, other interpreta­ tions can be made in some instances: e.g. certain acidic non-histone proteins decrease, or even disappear during echinoid development (Hill et al. 1971, Cognetti et al. 1972). This is probably due to the modulation of the gene regulation, and should not be considered as a nutritional event, at least in a strict sense, even if some of these proteins are probably degraded and their amino acids recycled. The first indication of proteic yolk degradation was that of Gustafson, who was cited by Kavanau (1954) as observing an intense degradation of the yolk platelets during echi­ noid embryogenesis. Kavanau (1953, 1954), using a technique of separation of free amino acids from proteins in Strongylocentrotus purpuratus and in Paracentrotus lividus, claimed to have found four periods of intense yolk protein breakdown: during early cleavage, in late blastulae, in early gastrulae and in early plutei. This was based on the observation that almost all the 20 free amino acids showed peaks of concentration at these stages. However, no evidence was shown that this increase of concentration of amino acids was due to the breakdown of the yolk. In addition, Kavanau (1958) partially invalidated his results show­ ing that proteolytic artifacts might have arisen from his technique. Perlmann & Kaltenbach (1957) prepared rabbit serum antibodies from whole egg extract of P'lividus. When these antibodies were tested on total extracts from eggs or embryos, two antigens were selected, whose concentration decreased during the development, especially after gastrulation. The authors interpreted the data as yolk protein breakdown. The evidence is supported by the fact that cell fractionation localized the two antigens mainly (but not exclusively) in the yolk platelets. The authors unfortunately do not give details of the rela­ tive concentrations of these proteins in the yolk and elsewhere, nor details of the cell frac­ tionation process, so the possibility of contamination cannot be excluded. However, one must conclude that, with the limitations already mentioned, the yolk is the main candidate as the site of storage of macromolecules to be broken down. An important, although indirect, evidence of a breakdown of reserve proteins arises from the findings of Fry & Gross (1970a,b). These authors found no change in the amino acid pool occurring during echinoid embryogenesis: this means that this pool must be re­ plenished as protein synthesis proceeds. Since several new syntheses are switched on at gas­ trulation (for a complete review on this aspect see Giudice 1973), the sharp decrease of the antigens observed by Perlmann & Kaltenbach (1957) at this stage is consistent with the evidence of an increased demand of amino acids from the pool to provide for increased protein synthesis. Other important data arise from the calculations of absolute rates of

474 Golfredo Cognetti

protein synthesis in embryos of Arbacia punctulata (Fry & Gross 1970b). Embryos synthe­ size proteins at a rate of 12 x 10-6 IJ.g of protein per embryo per minute. This means that in one day a turnover of 45 % of the total proteins would occur, provided that this rate of synthesis is constant. This evaluation (45 %), although very approximate, is in agreement with the estimations of total yolk proteins contained in an egg (Fry & Gross 1970b, Schuel et al. 1975) already mentioned. Other evidence of the processing of the proteins contained in the yolk platelets results from the isolation and characterization of lytic enzymes associated with them. A problem with such kind of experiments arises from the cell fractionation techniques that do not guarantee 100 % purity of the yolk platelets and yield preparations more or less contarni­ nated by other particles, such as cortical granules, mitochondria or 'heavy bodies'. The data shown, however, are gene rally acceptable, because higher percentages of the lytic enzymes in question are present in the yolk platelet fraction, compared to fractions en­ riched in other organelles. Two proteolytic activities, an acid and a basic one, have been found in embryos of Lythechinus variegatus by Krischer & Chambers (1970). Two classes of yolk platelets were separated by sucrose gradient centrifugation. The acid protease is contained in the class which sediments faster, whereas the alkaline one is contained in the class which sediments less readily. The degrees of association of these enzymes with the yolk particles are also different: the alkaline protease is bound tighter than the acid one, which is easily released in the supernatant. The acid protease has been identified as a cathepsin of D type. The finding is of interest since this enzyme is frequently contained in the lysosome of adult tissues (De Duve 1959, Sawant et al. 1964), and the relationship between yolk platelets and lysosomes which has been discussed. The basic protease, on the other hand, is tightly bound to an insoluble core, or to the membrane of the platelet, and its behavior resembles that of a chymotrypsin. 1.4.2. Evidences 01 utilization 01 stored polysaccharides The presence of glycogen in echinoid eggs of several species has been detected and evaluated in the past (Ephrussi & Rapkine 1928, Perlzweig & Barron 1928, Stott 1931, Ephrussi 1933, Chaigne 1934, Blanchard 1935, Zielinski 1939, Örström & Lindberg 1940, Hutchens et al. 1942, Lindberg 1943, Cleland & Rothschild 1952, 1957, Aketa 1952a). These authors, however, do not indicate its localization within the egg. Takashima (1968,1970) documented the presence of glycogen granules in the yolk platelets in oocytes and eggs of Heliocidaris crassispina, Hemicentrotus pulche"imus and Strongylocentrotus purpuratus. Unfortunately no quantitative analysis has been performed on the amount of glycogen in the yolk, com­ pared to the rest of the egg. The utilization of the glycogen starts immediately after fertili­ zation: a conspicuous drop in level (about 12 %) has been reported for Psammechinus mi/iaris and Echinocardium cordatum (Lindberg 1943) and even more for Paracentrotus lividus (Örström & Lindberg 1940). This high rate of degradation, however, is not constant through subsequent development: in general, after the first step, glycogen level remains almost constant, dropping again after gastrulation. No more glycogen is found at the plu­ teus stage (Ephrussi & Rapkine 1928, Ephrussi 1933, Zielinski 1939). Attempts to calcu­ late the turnover of total carbohydrates gave contradictory results. This is quite under­ standable if one considers the high amount of various species of carbohydrates present in the embryo, free or complexed to proteins or to lipids and bound in structures such as membranes and organelles. Data regarding glycogen alone, instead, are a certain indication of nutritional events.

Nutrition 01 embryos 475

Other indications of utilization of the carbohydrate of the yolk platelets arise from the isolation and characterization of enzymes involved in the carbohydrate metabolism: Acid phosphatase was found associated with the yolk in eggs and embryos of Arbacia punctulata (Cousineau & Gross 1960, Don~ & Cousine au 1967) and glycosidases in the yolk of eggs of s.purpuratus (Schuel et al. 1975). In particular these authors have found an acid nitrophe­ nyl phosphatase and an Q-L-fucosidase evenly distributed within both populations of yolk platelets, while an N-acetyl glucosaminidase and an N-acetyl galactosaminidase are preferen­ tially associated to the slower sedimenting yolk particles. In addition, almost all of these enzymes are usually found in the lysosomes, thus providing additional arguments to the problem of the lysosome-yolk platelet relationship. 1.4.3. Embryo respiration and lipid utilization Another important phenomenon which gives indications on the carbohydrate and/or lipid utilization following fertilization is the oxygen uptake, and carbon dioxide, release occurring at fertilization (complete reviews }bout this subject can be found in Giudice 1973 and in Yanagisawa 1975a). As is wellknown, a sharp increase in oxygen consumption takes place at fertilization in practically all the echinoid species. Two minutes after fertilization, the respiratory increase is 15 times the level ofunfertilized eggs (Ohnishi & Sugiyama 1963). A sharp decrease follows and the oxygen consumption reaches a plateau at a level of four times that of unfertilized eggs. Another increase is noted between 16-cell stage and blastula. A third increase then occurs, to a level which is maintained up to pluteus stage (Giudice 1973). In order to correlate the oxygen uptake with the carbohydrate and lipid metabolism, parallel evaluation of carbon dioxide displacement is important. When calculating produc­ tion of carbon dioxide one must be very careful because of the errors arising from the release of carbon dioxide by reaction of the acids discharged by the eggs at fertilization with carbonates contained in the sea water. The respiratory quotient (RQ), obtained from the ratio of carbon dioxide produced/oxygen consumed gives indications of the substrates for the oxidative metabolism. Isono (1963) found that the RQ of eggs and embryos of Anthocidaris (Heliocidaris) crassispina remained constant and near unity from fertilization to the hatching blastula stage. The RQ decreases after the swimming blastula stage. The data are interpreted as utilization of carbohydrates during the ftrst phase, and oflipids in the second phase. An interesting conftrmation arises from the observation (Iso no 1963, Mohri 1964, Isono & Yasumasu 1968) that the total lipids contained in an egg decrease only after the mesenchyme blastula stage. Correlating these data with those already men­ tioned, of a sharp decrease in glycogen content following fertilization, it has been suggested (Mohri 1964, Isono & Isono 1975) that in early stages of echinoid development respiration proceeds mainly through the oxidation of carbohydrates and successively, after the blastula stage, through the oxidation of the lipids.

2. OTHER ECHINODERMS 2.1. The yolk

The structure of the yolk platelets is similar to that of echinoids. There is a relationship between the size of the eggs and the structure of the yolk. Echinoderm eggs range in size

476 Goffredo Cognetti

between 50 pm, as in some Ophiuroidea, up to 0.5-1.0 mm as in some Holothuroidea and Asteroidea. According to Lönning (1976), who examined 32 different echinoderm species of Asteroidea, Echinoidea, Ophiuroidea and Holothuroidea, small eggs have yolk platelets with a simple granular ultrastructure, whereas yolk platelets from large eggs are larger, with a more complicated ultrastructure, with large lipid-like inclusions embedded in the granular matrix. The above generalization applies to almost all the observed species. However, in three asteroids (Solaster endeca, Solaster papposus, and Psi/aster andromeda) which have large (about 1 mm) eggs, the yolk seems to be constituted mainly by oll droplets (Lönning 1976). It is interesting to compare these results to the theory ofMortensen (1927), who suggested that echinoderms with large eggs, and therefore richer in yolk, would develop directly without areallarval stage, whereas echinoderms with small eggs would develop through a pelagic larval stage, in order to feed themselves before the metamorphosis (see Fenaux, chapter 22). As regards Crinoldea, yolk particles and oll droplets have been found by Holland (1978) by electron rnicroscopy in eggs and embryos of Comanthus japonica. 2.2. Yolk formation

Electron microseopie studies on oocytes of the holothuroid Sc/erodactyla briareus led Kes­ sel (1966) to suggest an interaction between Golgi vesicles and the endoplasrnic reticulum as the origin of the yolk platelets. No definitive evidence, however, was reported. 2.3. Utilization of stored matter

The presence of glycogen or glycogen-like carbohydrates in the yolk granules of asteroid oocytes (Pisaster ochraceus and Patiria miniata) was detected by Nirnitz (1976) by histolo­ gical techniques. A decrease of the total carbohydrate content has been detected in em­ bryos of the holothuroid Cucumaria curata (Turner & Rutherford 1976). The value found was 0.82 pg/egg during embryogenesis. This is equivalent to an oxygen consumption of 0.614 pI 02/embryo/28 days, sufficient to account for embryonie energy requirement. The oxygen consumption in the asteroid Marthasterias glacialis was measured by Borei (1948), who showed that there is not an immediate, sharp rise in respiration after fertlliza­ tion as observed with echinoids, but rather a gradual rising, resembling the profile of oxygen consumption of echinoids after the sudden increase has passed. No interpretation however was given to these data.

3. CONCLUSIONS Much more work has to be done, especially on echinoderms other than echinoids, to clarify the mechanism and the pathways of echinoderm embryonie nutrition. The lack of precise information too often results in either inductive or deductive conclusions, which would be bett er justified on an experimental basis. A large part of the experiments reported in this article are based onhistochemical rather than biochemical methods; in addition many experiments were performed long ago, when techniques were not sophisticated. The em­ ployment of radioisotopes along with the new cell fractionation techniques might give

Nutrition of embryos 477 definitive answers to many problems. Deflnitive and unambiguous evidence especially is needed: 1) On the origin of the yolk granules. It is necessary to clarify whether they derive from the Golgi, or from the endoplasrnic reticulum, and solve the question of their relation­ ship with the lysosomes. It is also important to know the site(s) of synthesis of the stored matter. 2) On the fate of the accumulated nutritional macromolecules. It is necessary to enlighten the entire process of macromolecular utilization during embryogenesis, specifying which molecules are involved and identifying a1l the steps of the yolk degradation.

ACKNOWLEDGEMENTS I thank Dr G.Giudice for discussion of the subject covered in this chapter; Drs G.Giudice and B.Ramsay Shaw for criticism of the chapter; CNR contract 77 .00339 .95 (research pro­ ject in the biology of reproduction) for flnancial support in preparation of the chapter.

22

LUCIENNE FENAUX

NUTRITION OF LARV AE

1. TROPHIC CATEGORIES OF ECHINODERM LARVAE Most echinoderms possess larvae with a more or less long pelagic life, although all classes have incubating species which gene rally inhabit the polar seas (Hyman 1955). This pelagic life of the larvae involves nutrient uptake from the surrounding environment when the embryo reaches the stage when all yolk reserves are completely used. Therefore it would seem at first sight that larvallife would be directly related to the amount of yolk reserves. Thorson (1946) adopted this hypothesis after analyzing reproduction of marine inverte­ brate and gave a classification in which larval development and nutrition of the embryo is correlated with the diameter of the oocyte. There exist rnany exceptions to this classification, some authors believing that the dia­ meter of the oocyte is not the only criterion enabling one to predict the type of larval deve­ lopment. Strathmann & Vedder (1977) noted that even if the quantity of organic matter increases with the diameter or the volume of an oocyte, it is not proportional to the volume; srnall oocytes can have more highly concentrated matter than larger sizes. According to Turner & Lawrence (1978) one cannot deduce larvallife habits from the egg diameter, as a clear relation does not exist between the size of the oocyte and its chemie al composition. Chia (1974) classified marine invertebrate larvae into three categories: - Those which do not feed: lecithotrophic larvae in which all requirements for growth and metamorphosis are met by yolk reserves. They either do not have a gut or one which is, in any case, never used. - Those which have facultative feeding: lecithotrophic larvae which do or do not use their functional gut. - Those which feed: larvae which are planktotrophs, detrital feeders, adelphophages and parasites. Among the echinoderms, the classical planktotrophic larvae (auricularia of holothuroids, bipinnaria of asteroids, pluteus of ophiuroids and echinoids) have a functional gut and processes or arms edged by a well-developed ciliated band which, as we shall see later, plays an important part in food collection. These larvae normally live three to eight weeks after hatching in the temperate seas. During larval development the processes or arms grow in size and number. For this larval growth and for the subsequent development of the juvenile, the larva must collect food from the surrounding environment. Pelagic larvae with lecithotrophic development often have either a reduced number of appendages (plu­ teus of many ophiuroids) or a 'barrel shape'. In the latter the digestive tract is sometimes incomplete. Pelagic life until metamorphosis usually lasts from one to two weeks. This shortness and the development of a reduced number of processes, indicate less need for 479

480 Lucienne Fenaux nutrition than in the pelagic planktotrophic larvae, whose needs seem to be completely covered by the yolk reserves . Derbes (1847) and Dufosse (1847) were the first to study the development of the edible echinoid. Since then, most authors have reared echinoderm larvae with nanoplankton from sea water. The method consists in the transfer of the larvae every one or two days to a freshly prepared medium. Allen & Nelson (1910) described an artificial culture medium for marine planktonic organisms. This medium has been used by workers to rear larvae with cultures of diatoms and unicellular algae. Onoda (1936) used a special method: he reared some J apimese plutei in a bowl in which the sea water was never changed but to which was added a little filtered sea water in which sea weeds had been washed . Grave (1902b) placed echinoderm larvae in fresh sea water to which were added some surface sand from an aquarium containing a cul­ ture of diatoms. With Hinegardner (1969) a new era began; he tried to calculate methodi­ cally the number of larva per rn1 and the number of diatoms or other algae for each larva to obtain a successful development. Now larval development followed by successful meta­ morphoses has been obtained by many authors in laboratory with cultures of unicellular algae or diatoms as food source (see Lawrence et al. 1977). The stornach contents of many different planktotrophic larvae collected in the sea have been described (see Mortensen 1921, Thorson 1946). Bougis (1964) demonstrated that the arms of echinopluteus decrease with starvation for a few days and grew again when the lar­ vae were placed in non-filtered sea water. The rate of regeneration of the appendages was related to 'effective phytoplankton', i.e. the portion usable for feeding by the echinoplu­ teus (fig.l). The bacteria of sea water can also be used (Deveze 1953), as larval growth was less in the absence of bacteria.

Figure 1. Change in the length evolution of the skeleton of post-oral arms of pluteus (Arbacia lixula) reared in different sea waters: water taken in the Bay of Villefranche-sur-Mer (continuous line), water taken at the bay entrance (broken line), the same water fIltered on a millipore filter of 0.45 J.l.m (dotted line) (one micrometric division equals about 10 J.l.m) (from Bougis 1964).

Nutrition o[ larvae 481 In Iecithotrophic pelagic larvae the absence of a mouth, a c10sed anus or the development of an anterior intestine without any communication with the rest of the gut makes plankto­ trophy impossible. One cannot help wondering if the absorption of dissolved organic sub­ stances as was observed in embryo of a viviparous species, Axiognathus squamata (= Amphi­ pholis squamata) (Fontaine & Chia 1968) is not a more generalized phenomenon than lar­ vae with abbreviated development would use.

2. MORPHOLOGY OF THE GUT IN PLANKTOTROPHIC LARVAE The development of echinoderm larvae has been described in many different species. For a c1assicallarva (auricularia, bipinnaria and pluteus) the pattern is practically the same. The egg, after segmentation, produces a blastula covered with cilia and begins to lead a pelagic life. After formation of the archenteron, a prismatic-looking gastrula appears. This under­ goes interior and exterior modification. The archenteron bends and divides into three parts by two constrictions: esophagus, stomach and intestine. The blastopore becomes the anus

Figure 2. Change of typicallarval forms in echinoderms from the dipleurula. Row A - transformation of the dipleurula in echinopluteus; row B - ophiopluteus; row C - auricularia; row D - bipinnaria. The circumoral ciliated band is indicated by asolid line, the circumoral field is indicated by a dashed line. (from J.Müller, modified by Dawidoff 1928).

482 Lucienne Fenaux

Figure 3. Feeding mechanisms in four types of echinoderm larvae A, B, C. Schematic diagrams for larvae. A. Ventral view; B. Sagittal section (in the places offigure A) ; C. Cross section (from figure A) ; D. Ventral view of an ophiopluteus. ad.b. - adoral band; a.l.a. - antero-lateral arm ; a.s. - anal sphincter; c.s. - cardiac sphincter; cil.b. ­ ciliated band ; es. - esophagus; int. - intestine ; l.ad.b. - lower adoral band ; l.st. - lower stomaeh; par.b. - paroral band; p.d.a. - postero-dorsal arm; p.l.a. - postero-lateral arm ; p.o.a. - post-oral arm ; p.o. ­ pyloric sphincter ; p. t. b. - post-oral transverse band ; pr. t.b. - pre-oral transverse band; s.o. b. - suboral pocket ; u.ad.b. - upper adoral band; U.st. - upper stornach (open arrows: water currents ; arrows drawn with a single line: paths of particles) (from Strathmann 1971).

Nutrition o[ larvae 483 and the mouth, newly formed, appears at the bottom of a transverse groove. The anus moved towards this side (Dawydoff 1928). The cilia are concentrated and form a crown of long, strong cilia surrounding the oral groove. The circumoral field is between this ciliated band and the mouth; the aboral field, where the anus is, is at the outside of this ciliated crown. This aboral field is not as rich in cilia as the circumoral field. All echinoderm larvae reach this stage of development, the dipleurula. The circumoral band grows bigger and forms two lobes. The form and the development of these lobes differ in holothuroids, asteroids, echinoids, and ophiuroids (fig.2) The band forms arms supported by a ca1careous skeleton in the pluteus, and processes, with or without ca1careous concretions, in the auricularia and bipinnaria. In the crinoids, the larva resembles an older stage of development of auricularia, the doliolaria, having a barrel shape with strong cilia crowns. The cilia of the circumoral field and the appendages,

Figure 4. Seetion through the stoma eh (st) and intestine (I) of a braehiolaria. Note that the pyloric and anal orifiees are open (from a photomierograph of Cobut 1976).

484 Lucienne Fenaux by their beating, collect suspended particles (GemmilI1916, Runnström 1918, Tattersall & Sheppard 1934, Strathmann 1971). Their position is briefly summarized below. The larval mouth is situated at the bottom of a groove which separates the larval body into two parts: preoral and anal fields. The ciliated band which follows the edge of this groove involves two transversal parts, pre- and postoral. The mouth is also surrounded by another ciliated band, the adoral band, whose upper part adjoins the transverse preoral ciliated band (fig.3b). There also exists a bap-d of cilia which follows the edges of the transverse groove between the pre- and post oral bands: these are parorals bands (fig.3d). Each cell has a cilium with an average length of 25 11m (Strathmann 1971). The cilia are not only concentrated on the ciliated bands: the circumoral field is ciliated in the bipinnaria, echinopluteus and some auricularia (Garstang 1939, Strathmann 1971). In the aboral field they are rare or concentrated at the posterior end of the larva (pluteus). Most authors recognize three sphincters in the gut: cardiac (separating the esophagus and stornach), pyloric (between the esophagus and intestine) and anal. Although the cardiac sphincter is clearly visible in the living larvae and in sections of the gut, the pyloric and anal sphincters are not. According to Cobut CI 976) the intestine separates itself from the stornach and moves in a tangential direction towards the surface of the latter in the young larval stages of Asterias rubens (fig.4). In the older bipinnaria the intestine sticks to the wall of the stornach giving the impression of a clearer rupture between the two portions of the gut. We have also observed this in the bipinnaria of Marthasterias glacialis and in an undetermined auricularian. In one larva, the echinopluteus of Colobocentrotus atratus, Mortensen CI 921) described a contraction separating the intestine and the anus.

Figure 5. Beating of cilia in a pluteus and path of particles through the ciliated band. A. Larva which does not feed; B. Larva wh ich feeds cf - cireumoral field; eb - ciliated band ; af - anal field (from Strathmann et al.- 1972).

Nutrition 01 larvae 485 3. FEEDING OF LARV AE 3.1. Planktotrophic larvae 3.1.1. Feeding mechanisms

The planktotfophic larva of the echinoderms is a suspension-feeder which feeds on small diatoms, phytoplankton, small flagellates and detritus. It develops easily in a laboratory if given a food complement at the end of the endotrophic phase. Strathmann (1971), in his remarkable works on the nutritive behavior of planktotrophic larvae of echinoderms, showed that they require organisms whose length and width do not exceed 100 to 200 /J.m and 65 to 85 /J.m respectively. The collection of food particles has been described by Strathmann (1971). The beating of the cilia and the ciliated band which surround the circumoral field produces a current directed towards the exterior of the field and at right angles to the band (arrows A and B of figs.3a and b). Water enters the circum­ oral field from the sides and the anterior end (arrows C of fig.3a). The disposition of the ciliated band in alllarvae is such that its largest part produces a current with posterior or transverse components; only some parts create an anterior component to the current. The currents play an important role in feeding and in propulsion of the larvae. Water.and particles in suspension are brought into the circumoral field by a flow corning from the anterior end or the sides. Two things can happen: the larva feeds and in this case the particles are retained at the inner side of the ciliated band while the water is passed over the band (distance X-V fig.3c); or the larva does not feed and particles and water go over the ciliated band (distance X-Z fig.3c). Food intake is made possible by areverse beat­ ing of the cilia in contact with the particle (Strathmann et al. 1972) (fig.5); The cilia of the circumoral field, when they exist, and those of about one-fifth of the esophagus, function in food transport towards the mouth. These cilia beat in the direction of the stomach and concentrate the collected particles in the lower part of the esophagus (Strathmann 1971). Larvae feeding in a suspension too rich in particles can stop feeding and even reject the contents of the buccal cavity or the upper part of the esophagus, by an inversion in the beating of the cilia of the pre- and post-oral transverse bands and the adoral band. The peristaltic contractions of the esophagus can carry this process even farther and remove the particles from the lower part of the esophagus; the sphincter between the stomach and the esophagus stays closed during the rejection period (Strathmann 1971). Gemmill (1916) described the closing of the buccal cavity by a flexion behind the preorallobe on the rest of the body in the bipinnaria of Porania pulvillus. Mackie et al. (1969) observed during each ciliary reversal, monophasic potentials not exceeding 20 /J.V. These reversal potentials appeared only during the ciliary reversal. Nervous structures have been described in echinoderm larvae (see Mackie et al. 1969). In the pluteus of Strongylocentrotus droe­ bachiensis, Mackie et al. identified histologically nerve cells and suggested that a neural conduction in the covering epithelium or in the mesenchymental elements may play a role in this phenomenon. Particles accumulated in the esophagus are pushed into the stomach by a wave of con­ traction of circular muscles followed by the opening of the cardiac sphincter (fig.6). A vigo­ rous mixing of the food takes place in the stomach by the action of cilia. The food accu­ mulates in the lower part of the stomach and subsequently is pushed into the intestine. Sorting ofthe particles occurs in the stomaeh. Strathmann (1971) noted that, after inges­

486 Lucienne Fenaux tion of a mixed suspension of approximately the same sized particles of algae (Cricosphaera and Phaeodactylum) and carmine or calcium carbonate crystals, the non-living particles passed into the intestine and were defecated before the living particles. The cilia of the intestine ceUs, inactive except du ring defecation, rapidly transfer the waste material to the anus. In the larvae studied by Strathmann (1971) the time of passage of food through the gut is generally 15 minutes, but it varies according to the concentration of particles in the environment. It depends above all on the time of the passage between the upper and the lower parts of the stornach. In a medium in which four-armed plutei of Paracentrotus IM­ dus and Arbacia lixula ate a mixture of phytoplankton, Rassoulzadegan & Fenaux (1979) observed that the type which is more concentrated is the fIrst to be ingested. When this concentration decreases, the other types became more concentrated in turn and then are eaten.

3.1.2. Food treatment in the gut The digestive tract consists of an epithelium supported by a fille basal membrane. The bor­ dering cells vary in aspect according to the section studied. They are flattened in the buccal -cavity and esophagus. The stornach and the intestine are ciliated. A mucus fIlm containing large amounts of mucopolysaccharies protects the esophagus (Ryberg & Lundgren 1975). It seems to be secreted by some ceUs of the adoral band with a ciliated and glandular structure. This fIlm does not show any more metachromasia in the stornach and intestine. Ryberg & Lundgren suggested that the material was formed by par­ tial hydrolysis of sulphomucopolysaccharide when it reaches the stornach. A good supply of food stimulated mucus secretion and swallowing. Gustafson et al. (1972) believed that a significant part of this food, made up of small organisms, is covered with this mucus secre­ tion. This would avoid abrasion of the digestive wall. Cells with large vacuoles (nutritive vacuoles of Runnström 1912a) with proteic grains - probably digestive enzymes - in their apical part can be seen in the stornach. The number of these supposed zymogen ceUs in­ creases with the larva's age. According to Runnström (1912a,b), food is absorbed in the stornach (rnid-gut). Runn­ ström found a network of 'mesenchymatic ceUs' particularly weU-developed in the distal part of the gut with which it is in contact by cytoplasrnic extensions. Runnström presumed that the last unit has an excretory role because: 1) the degree ofvacuolisation increases during the larvallife; 2) atrophy of the ceUs of the posterior part of the gut occurs when larvae are starved. It would be interesting to repeat the study of food metabolism in echi­ noderm larvae as the data available are old and fragmentary. Hydrolysis of food is brought about by digestive enzymes. Some have been found in the larva. Two enzymes of the carbohydrate metabolism are known in the echinopluteus: ß­ 1,3 glucanase and a:-amylase. After partial purifIcation of the guts of exogastrulating em­ bryos of Dentraster excentricus, Vacquier (1971) found the activity of ß-l,3 glucanase (an enzyme which liberates glucose from laminarin) is more concentrated in the gut than in other embryonic tissues, suggesting a digestive role. Vacquier et al. (1971) found a maltase activity which appears during the differentiation of the gut for this same species. A pepti­ dase, Q-Ieucylaminopeptidase, has been found in the pluteus of Echinarachnius parma (Doyle 1956). Non-specifIc esterases found in the esophagus of the larva of Psammechinus miliaris (Ryberg 1973) hydrolyze lipids. Evola-Maltese (1957) noted that the presence of alkaline phosphatase in the larva of Paracentrotus lividus is more active during the pluteus stage. This enzyme seerns to accumulate in the intestinal lumen more than within the cells

Nutrition of larvae 487

(Okazaki 1956 quoted by Giudice 1973). Acetylcholinesterase, although not a digestive enzyme, plays a role in larval nutrition by regulating of intestinal movements (Augustinnson & Gustafson 1949). 3.1.3. Role played by planktotrophic larva in energetic transfer The environmental role of planktotrophic larvae of echinoderms in the transfer of sub­ stances formed by the phytoplankton can be especially important. Strathmann (1971) studied the clearance rates of particles in suspension for many species of planktotrophic echinoderm larvae. The elearance rate is proportional to the length of the ciliated band and depends on the concentration of partieles. Rassoulzadegan & Fenaux (1979) noted that the elearance rate also depends on the temperature. For Arbacia lixula, living in the Medi­ terranean Sea and with summer-autumn breeding season, the maximal elearance rate occurs at a temperature of 22°C, elose to ambient temperatures observed in mid-summer to rnid­ autumn. The rate of daily ingestion was estimated to be 50 to 458.8 IJg of carbon per m 3 when the larvae are abundant in the sea at Villefranche-sur-Mer (France). According to the rates observed in those experiments, a larva can ingest a quantity of organic carbon repre­ senting 2 to 72 % of organic carbon of its body in a day. 3.2. Lecithotrophic larvae: pelagic and demersal

Echinoderms whose larvae do not feed or have facultative nutrition are gene rally deep-sea species or live in an environment where physicochemical conditions are not favorable to the embryo. This is the case for 95 % of marine invertebrates in the polar seas (Thorson 1946) and 83 % of echinoids and ophiuroids at Cedar Key in Florida (Stancyk 1973). Lecithotrophic larvae, pelagic or demersal, often have a different morphology to those of the elassic planktotrophic larvae. There are weIl known examples in the ophiuroids whose principal changes are in the reduced number of arms and, as a result, in the reduced length of the ciliated band. Thus Amphiura chiajei and most probably A.mediterranea have pelagic lecithotrophic larvae with only one pair of arms as opposed to classical planktotrophic lar­ vae of ophiuroids which possess four pairs (Fenaux 1963, 1968). The exterior form of some larvae is sometimes so changed that the arms do not exist and the skeleton is reduced to incomplete rods (Ophiopluteus claparedei). The pluteus form can completely disappear and be replaced by a modified doliolaria (Ophioderma brevispina, Grave 1916 and Ophio­ derma longicauda, Fenaux 1969). In some larvae with these exterior changes, the gut is also modified: no larval mouth (Heliocidaris erythrogramma, Mortensen 1921 ; Mediaster aequalis, Birkeland et a/. 1971), no opening of the pharynx into the stornach (Peronella /esueuri, Mortensen 1921), closed blastopore. These larvae with or without a weIl-formed . gut do not need any external nutritive support to reach metamorphosis as many rearing . experiments in laboratory have proved. It would be interesting to see if complete develop­ ment occurs in sea-water lacking dissolved organic substances. At least, Amphioplus abditus undergoes larval development and metamorphosis within a demersal fertilization membrane (Hendler 1975). 3.3. Incubated embryos It is necessary to distinguish between the viviparous species and those which protect their eggs. Parents which supply material and energy to their offspring which have exhausted

488 Lucienne Fenaux their vitellogenic nutrients are viviparous. The transfer of material can take place through the use of nurse eggs or the ingestion of parental body fluids. In echinoderms where vivipa­ rity is known, there is no placenta (Turner 1977). There are two classical examples of embryonic nutrition by the mother in viviparous species of Ophiuroidea: Ophionotus hexactis (Mortensen 1921; Turner & Dearborn 1979) and Axiognathus squamata (= Amphipholis squamata) (Quatrefages 1842, Fell 1940, 1946). Ophionotus hexactis, a six-armed subantarctic brittlestar, is an intraovarian brooder. There are observations showing that ingestion of nurse eggs or absorption of parental fluids occurs. However, it has been observed that each ovary produces 4 000 to 100 000 ovocytes and only one embryo develops in an ovary. This embryo forms an ophiopluteus and then a post-metamorphic juvenile which grows inside the ovary. Fell (1946) described the deve­ lopment of one or two oocytes of 100 ~m diameter inside the genital bursae of A.squama­ tus. The larva is a vestigial pluteus with a skeleton reduced to two calcareous fenestrated plates. The gut is incomplete, lacking an anus and a mouth. The rounded posterior part of the larva is due to an excrescence of the bursal wall. This excrescence is composed of undif­ ferentiated tissue without vascular structures. Fell believed that the absence of sinuses indicates that this tissue has no nutritive function. The yolk reserves are not sufficient to assure complete larval development as excised embryos were unable to survive more than five days if certain substances contained in Erdschreiber culture medium were not added in the culture (Fell 1940). Sinuses appear at the beginning of incubation in the genital bursal wall. According to Fell, this wall secretes substances which, released into the lumen of the bursae, were directly absorbed by the embryonic tissues. Fontaine & Chia (1968) reported the assimilation of dissolved organic molecules (glucose and glycine) in embryos of A. squamata in different stages of development. Assimilation by the embryos, which are brooded in the bursae of this viviparous species, is greater than by adult tissues. Absorption and assimilation of dissolved organic molecules therefore occurs in incubated embryos as in juvenile Echinoidea (Pearse & Pearse 1970). 4. E FFECTS OF DISSOLVED ORGANIC MATTER ON LARVAL DEVELOP­ MENT The uptake of dissolved organic mattl:r from sea water by the epithelium of echinoderm larvae has been described by few authors (Dixit 1973, Pavillon 1976). Dixit observed changes in the glycine transport system during embryonic and also larval development of the echinoid Strongylocentrotus purpuratus. According to him, there is a logarithmic in­ crease in Kt of the glycine transport system during the development. The actions of a vitamin (riboflavin) and of two amino-acids (glutamic acid and glycine) were tested on the growth of two echinoid larvae, Arbacia lixula and Paracentrotus lividus (Pavillon 1976). Presumably these substances (ectocrine substances) are released by living organisms. Riboflavin has no significant effect on the larval growth. On the other hand, the size of the pluteus of both species increases significantly when solutions of those amino­ acids (around 0.1 ~g/l to 51 ~g/l) are added to natural or artificial sea water. 5. LARVAL ENERGY BUDGET Since the rniddle of the 20th century, scientists have studied nutritional relations between

Figure 6, Succcssivc stages of wave contraction of esophageal muscles and opening of the cardiac sphinc­ ter in an auricularia (photographs by Fenaux), A. Lateral view, A - anus; BC - buccal cavity; CS - car­ diac sphincter; CB - ciliated band; ES - esophagus; IN - intestine; ST - stornach. B, C, D. Three stages of wave contraction of circular muscles.

Nutrition of larvae 489 echinoderm larvae and other marine organisms. The larvae move in swarms and, in the sea water layers in which they live, they find more or less propitious conditions for their mor­ phogenesis and later for their metamorphosis. Hinegardner (1969) stated that proper food is an important factor for larval development during experimental cultures of echinoplutei. Bonar (1978) observed regeneration of amputed arms which is dependent on the food supply and the state of development of echinoplutei rudiment. Some physic or trophic sea water conditions have been simulated in aquaria in which larvae have been reared. Thus, the clearance rate and the uptake of dissolved organic matter have been quantified. It is possible to diagram the energy budget of a larva:

Sources of nutrients

Lost energy (fecal pellets, dissolved organic material)

- yolk reserves

t Energy used for metabolism (respiration) t t t

\,

- dissolved orga­ --+ nie substanees f - food particles

LARVA -+ METAMORPHIC -+ Larvallife pro- -+ JUVENILE "', LARVA longed until exo­ '" genous factors larval morphodevelopment of are present genesis adult structures

Experimental data have furnished some terms of the energy budget given above: energetic sources, and energy used for larval morphogenesis and metabolism (respiration). Other terms of this equation are still unknown especially the lost energy (dissolved organie material). The study of starved larval metabolism mayaiso add to our knowledge. We know that in a starved echinopluteus (Runnström 1912b), the morphology is modified in two ways. Larval arms are greatly shortened, which involves phagocytosis of the skeletogen cells. This observation has also noted by Metschnikoff (1884, reported by Runnström 1912b) for Bipinnaria asterigera, the larva of Luidia sarsi. The other process shown by starved larva involves the gut, the lumen of which shrinks. In both cases, larval movements are reduced. These experiments about larval nutrition should be completed by other studies, namely: larval digestion and evolution ofbiochemical components (especially the lipid reserves) during larval growth.

ACKNOWLEDGEMENT I thank Dr John M.Lawrence for discussion of the subjeet covered in this ehapter.

5.EFFECTS OF FEEDING ON THE ENVIRONMENT

23

CLAUDE MASSIN

EFFECTS OF FEEDING ON THE ENVIRONMENT: HOLOTHUROIDEA

Holothuroids can be divided by their feeding habits more or less into two principal groups, suspension feeders or deposit feeders. The environmental effect of the feeding activity by holothuroids will differ according to the group to which they belong. 1. DEPOSIT FEEDERS The deposit feeding holothuroids can have two environment al effects: they can change the size ofthe ingested partieles, and they can turn over the sediment (bioturbation).

1.1. Change in average sediment diameter A change in average sediment diameter could result from mechanical and/or chemical action when passing through the digestive tract. Townsley & Townsley (1972) observed an increase in the frequency of smaller particles through the digestive tract of Holothuria ieucospilota and Stichopus japonicus, and suggested a third possibility: oral egestion of the large partieles. This seems unlikely because of the absence of any sorting mechanism in the oral cavity of holothuroids. The parts of the holothuroid's gut capable of a mechanical grinding action are the pha­ ryngeal bulb, the posterior intestine, and above all the stornach (for defmition of these seg­ ments, see Feral & Massin, chapter 9). These are the regions of the gut which possess weIl developed museles. In the posterior intestine, only some aspidochirotes possess powerful museles. As sediment is embedded in mucus in this segment, attrition of particles is unlikely. A stornach, in deposit feeders, is found only in the apodids. Only the stornach and, to a lesser extent, the pharyngeal bulb mix, compress and push the sediment into the next diges­ tive part. They could act mechanically on the alimentary bolus. Any effect of the pharyn­ geal bulb is probably minimal because of the shortness of these regions. The effect of the stornach on the ingested sediment has never been studied. The available information comes only from species without stomachs. In these species mechanical action (attrition and grind­ ing) on the alimentary bolus is negligible or absent (Crozier 1918, Bertram 1936, Yama­ nouchi 1939, Trefz 1958, Yingst 1974). Any chemical action on ingested calcareous deposits would result from an action of the intestinal juice. The average pH of the intestinal juice is usually between 5 and 7 (Yama­ nouchi 1939, Bakus 1973). Trefz (1956 in: Bakus 1973) and Bakus (1973) elaimed that dis­ solution does not occur as the intestinal pH is elose to neutrality when the digestive tract is fuH, sea water has a buffering action, and the sediment is in the gut for only a short time. No dissolution of calcareous deposits fed to holothuroids was detected after periods of 5 493

494 Claude Massin to 8 weeks. Yingst (1974) came to the same conclusion for Parastichopus parvimensis. Bertram (1936) and Yamanouchi (1939) concluded that the chemical effect of digestion on calcareous material by holothuroids is weak and without ecological consequence. Barth et al. (1968), observing the intestinal content of Holothuria sp., assumed that there is no pH effect on food as calcareous skeletons in the intestinal contents were unaffected. Khri­ pounoff (1979) found no variation in carbonate between the mouth and anus in the diges­ tive tract of some abyssal holothuroids. However, Webb et al. (1977) stated that possible carbonate dissolution in Holothuria atra between the pharynx and the posterior intestine would be between 0 and 8.5 %, but this is only a theoretical calculation. In the peculiar case of Stichopus moebi (now Isostichopus badionotus), Crozier (1918) believed that chemical action of intestinal juice on calcareous deposits is weak. However, con­ sidering the large amount of sediment passing through the digestive tract, he concluded that holothuroids are involved in deepening lagoons and in formation of mud. Mayer (1917) and Mayor (1918, 1924) went further and asserted thai Isostichopus badinotus is able to dissolve large quantities of sediment. Their experiments, as Bakus (1973) said, are unreli­ able. Nevertheless, one must recognize that the intestinal juice of Lbadionotus has a rather low pH compared to other holothuroids, especially when the digestive tract is fuH of sand (4,8 to 5,5). As Lbadionotus eats discontinuously, with long periods of sand retention, there may be slight dissolution of calcareous deposits in this species. In general, it seems that chemical action on a calcareous sediment passing through the digestive tract of deposit feeding holothuroids is very weak. However it is possible that, for some species and after a long period of time, there is a significant effect on the substratum. The abyssal holothuroid Psychropotes longicauda is able to modify the sediment granulo­ metry by dissolving the organo-mineral aggregates (Khripounoff 1979). But this phenome­ non corresponds to a chemical dissolution of the organic matter bin ding particles together, not of the inorganic particles themselves. 1.2. Turnover in the sediment stratification and its consequences

The importance of holothuroids in reworking muddy and sandy bottom has been empha:. sized by Crozier (1918). This reworking has two effects: destruction of initial stratification (Frizzel et al. 1966) and modification of the stability of the sediment. According to the fee ding manner of the holothuroids, there are different possibilities for reworking. The aspidochirote holothuroids living on the sediment gene rally are not selective for particle size (Yamanouchi 1939, Bonharn & Held 1963, Bakus 1968, Yingst 1974, Sloan & von Bodungen 1980). Only Stichopus tremulus is able to select coarser par­ ticles (Hauksson 1979). These particles would be more concentrated in the fecal pellets and thus at the water-sediment interface. Some holothuroids live with the mouth down in the sediment (Caudina spp., Molpadia spp.,Holothuria spp.) and bring up deep deposits to the surface (Yamanouchi 1926, 1929, Hatanaka 1939, MacNae & Kalk 1962, Mosher 1980). Others, e.g. Leptosynapta tenuis, defecate approximately 40 % of the ingested sediment below the surface and transfer deposits from the water-sediment interface into the substra­ tum (PoweIl 1977). The general influence of such bioturbation is difficult to estimate because there are few data on the quantification of sediment reworking (Bakus 1973) and because bioturbation by holothuroids is mixed with that of other animals. Normally the higher the density of holothuroids, the higher the influence on the destruction of initial sedimentation. But this

Ellects olleeding on the environment: Holothuroidea 495 is not always true and on deep sea bottoms where invertebrate assemblages are depauperate, the activity of fauna (e.g. holothuroids) seems to have accentuated effects (Rowe 1974). The effects are accentuated because of the very low sedimentation rate and because the scarcity of organic matter might mean that the deposit feeders such as holothuroids must ingest more sediment to meet their metabolic needs. Bioturbation by deposit feeders, keep­ ing the upper layer of sediment mixed, prevents calcium carbonate from being rapidly buried and lost to the system (Paul 1977). In abyssal communities, above the dissolution depth of calcium carbonate (lysocline), this role can be ascribed mainly to holothuroids because they are the predominant group (Zenkevitch 1963 in: Paw~on 1966b). This mixing of calcium carbonate and thus its perpetual contact with sea water is very important in the carbon dioxide cycle (PauI1977). Feces (fecal pellets or fecal mounds) can bury plants or animals. This is particularly true with burrowing holothuroids. Mosher (1980) reports that the feces of Holothuria arenicola constantly bury the .seagrass Thalassia testudinum around the burrow, inhibiting the deve­ lopment ofthe grass bed. Generally the seagrass is not found within a radius of 15 to 20 cm surrounding the burrow site. The number, shape, andmechanical resistance of feces in shallow water protected areas, and especially in the deep sea, are indicators of holothuroid density and diversity and of bioturbation (Kitchell et al. 1978). The reworking of the first upper centimeters of the sediment sometimes leads to an alteration in the stability of the sediment. For the majority of deposit feeding holothuroids, and especially on muddy or sandy bottoms, the reworking of the sediment produces a fluid fecal-rich surface that is easily resuspended by low-velocity currents (Rhoads & Young 1970). This phenomenon is not peculiar to holothuroids as other invertebrates such as crus­ taceans (Callianassa sp.) have a sirnilar effect (Aller & Dodge 1974). The physical instability of the fecal surface 1) clogs the fIltering structures of suspension fee ding organisms, 2) buries newly settled larvae or discourages the settling of some larvae, and 3) prevents sessile epifauna from attaching to an unstable mud bottom. Rhoads & Young (1971) claimed that unstable sediment between the fecal mounds of Molpadia oolitica is free of suspension feeders (polychaetes, amphipods and bivalves). Renaud-Mornant & Helleouet (1977) stated that the abundance of Holothuria atra in a lagoon is unpropitious to meiofauna. This restric­ tion must be mainly ascribed to the perpetual reworking by H.atra on the sediment surface rather than to predation, as meiofauna are scarce in the intestinal contents of holothuroids. However, Dayton & Hessler (1972) suggested that deep sea croppers, such as holothuroids, prey on smaller deposit feeders (meiofauna) and are largely responsible for the maintenance of high species diversity of small deposit feeders by reducing the probability of competitive exclusion. This is only one hypo thesis to explain faunal diversity in deep sea and, as dis­ cussed by Carney (1977) there are not enough data on the abyssal ecosystem to confirm or refute it. The destabilization action of holothuroids has other aspects. There is a transfer of the fluid fecal material rich in organic matter, and of interstitial water rich in dissolved material from the sediment to the water colurnn. These elements can promote development and growth of some molluscs such as oysters (Rhoads 1973). Not all holothuroids destabilize the top sediment. According to Myers (1977a,b) and Powell (1977), Leptosynapta tenuis increases the stability of the upper 3 cm of the sediment but decreases the stability below (3 to 10 cm of depth). Both workers reached the same conclusions through very different methods: direct measurement of viscosity of the sediment (Myers) and sediment reworking budget (PoweIl).

496 Claude Massin

Therefore holothuroids (particularly the conveyor belt feeders), mainly on sandy or muddy bottoms, can modify the structure of an ecosystem by preventing or restricting the development of populations of suspension feeders as weIl as of the meiofauna. An exception must be made in the case of Leptosynapta tenuis and probably in a wider sense of aIl holo­ thuroids which are funnel-feeders. On the whole, environmental effects of fee ding activities of holothuroids act mainlyon the faunal composition of the surrounding substratum. The biological effect of deposit fee ding holothuroids involves amensalism as defined by Rhoads & Young (1970): inhibition by one trophic group (deposit feeders) on any.other trophic group (suspension feeders).

2. SUSPENSION FEEDERS The suspension fee ding holothuroids, Le. in the majority of the dendrochirotes, seem to have little effect on the sediment. They take little or no sediment when eating. Hence it is reasonable to suppose that they have no effect on the sediment particle size. However, like deposit feeders, they are able to turn over the sediment. Many dendrochirotes live in the sediment and modify the initial stratification in the top few centimeters. Some bury themselves verticaIly with anus down (e.g. some Cucumaria spp.) and transfer to the sediment particles obtained from water or sediment-water inter­ face. Others, such as Sclerodactyla briareus (Pearse 1908), Trachythyone tergestina, Pseu­ docnus kollikeri, Colochirus cucumis (Mortensen 1927) or Neopentadactyla mixta (Kön­ necker & Keegan 1973) are typicaIly disposed in a U-shaped form within the sediment. N mixta defecates by extending its anus free of the deposit. Hs feces (particles mixed with mucus) are resuspended in water, are not filtered out by the holothuroids and rise to the water surface (Könnecker & Keegan 1973). We have no information on the influence (e.g. amensalism) that very dense populations of dendrochirotes (as described by Könnecker & Keegan 1973 and Keegan & Könnecker 1980) may have on the ecosystem in which they live.

3. CONCLUSIONS Deposit fee ding holothuroids have no significant action on the granulometry of sediment particles. However some abyssal holothuroids, by dissolving organic matter binding particles together (generaIly very smaIl particles), indirectly modify sediment granulometry. As renewal of organic matter is low in the deep sea, active feeding of holothuroids could change the mechanical properties of the sediment and promote resuspension of particles or turbidity currents. By turning over the sediment in muddy and sandy areas, holothuroids playa role in bioturbation. In shaIlow water, this role must be limited because the quantity of sediment moved is low in comparison with that moved by other animals such as polychaetes, entero­ pneusts or decapods (Thorson 1966). On the deep sea bottom, where they are often the dominant group of macrofauna, holothuroids could play an important role in the -carbonate cycle and thus on the carbon dioxide cycle and the pH of sea water. More information is needed on their rate of feeding and on their role in bioturbation. Deposit feeding holothuroids have direct (predation) and indirect (amensalism) effects

E//ects 0/ /eeding on the environment: Holothuroidea 497 on meiofauna and macrofauna. However meiofauna are known to be scarce in intestinal contents. As most of the meiofauna move very slowly in the sediment and are unable to escape deposit feeders, it is possible that holothuroids have a negative selection with respect to meiofauna or that apart of the meiofauna is rapidly destroyed in the digestive tract. This question is still to be solved. Suspension fee ding holothuroids are sedentary and passive in their feeding, but may be extremely dense. There is no information available on the effect of their feeding activity on the planktonic communities, either on adult or larval forms. According to Levinton (1972) there would be no competitive exclusion by exploitation of food resources with regard to other suspension feeders but a competitive exclusion for feeding space is possible. It would be interesting to make cage experiments like those of Renaud-Mornant & Helle­ ouet (1977) or Reise (1977) to ascertain the influence of holothuroids on other trophic groups. The role of suspension fee ding holothuroids in removing material from the water column and transferring it to the bottom community through incorporation into their body components has not been considered. Holothuroids are conspicuous components of many marine communities, from the tropics to the poles, and from the intertidal to the deep sea. The environmental effect of their feeding activities should be considerable, especially in deep sea, but lütle work has been done thus far to document it.

JOHN M. LAWRENCE & PAUL W. SAMMARCO

24

EFFECTS OF FEEDING ON THE ENVIRONMENT: ECHINOIDEA

The effect of fee ding by echinoids on the environment can be great as a result of several of their characteristics. Echinoids are frequently found in dense populations (Moore 1966) and can be long-lived (Ebert 1975). They can have rapid rates of fee ding , but can persist although food levels may be low (Lawrence 1975a). Feeding can have a direct physical effect through the ingestion or manipulation of the substratum, and a biological effect on the organisms consumed. Either of these effects can have indirect consequences.

1. THE PHYSICAL EFFECT OF FEEDING ON THE SUBSTRATUM

1.1. Soft substratum Although some groups of regular echinoids such as the echinothurüds live on soft substrata and ingest it (Lawrence 1975a), they are few and their biology has been little studied. Most regular echinoids with this way oflife dwell in the deep sea where there is little water move­ ment. They may have considerable impact on the soft bottom, however, as Kitchell et al. (1978) reported that echinoid plowmarks resulting from deep burrowing during sediment ingestion are the most common trace in the Antarctic. Echinoid plowmarks were found in 52 % of the observations of the Bellinghausen Basin. Characteristics and frequency of the traces were used to relate behavioural foraging efficiency to nutrient supply, and led to the conclusion that the level of nutrients in the Antarctic was much greater than in the Arctic. The great majority of echinoid inhabitants of soft substrata which feed by ingesting par­ ticulate matter are the irregular echinoids. Although they are found extensively in shallow water, the effect of their feeding activities on the environment has not been well studied. The principal physical effect of feeding by the sand dollar Mellita quinquiesperforata upon the substratum is size-sorting of detrital grains (Bell & Frey 1969). Finer particles are ingested, incorporated into feces, and subsequently adrnixed with sand passing through the posterior lunule. This results in a concentration of finer grains along the axis of the trail and partial depletion of fine grains in sediment bordering the trail. A similar change in dis­ tribution of particles result from the movement and feeding of Echinarachnius parma (R.D.Srnith 1978). Stanley & James (1971) concluded that the feeding and burrowing activities of Eparma were the most important biological agents reworking the superficial sediments of the Sable Island Bank. The consequences of these activities have not been evaluated. Inclined attitudes of the body, in which the anterior end of the animal is buried in the sand and the oral surface forms an acute angle with the substratum have been reported for 499

500 lohn M.Lawrence & Paul W.Sammarco Dendraster excentricus (H.L.Clark 1901, 1935, Mortensen 1921, Durharn 1955, Raup 1956, Chia 1969, Merrill & Hobson 1970, Birkeland & Chia 1971, Timko 1975, O'NeillI978),

Dendraster laevis (Merrill, in Merrill & Hobson 1970) and the west African rotulids (D'arte­ veIle 1935). The test acts as a lifting body in such aposture (O'Neill 1978, see Oe Ridder & Lawrence, chapter 4). Beds of animals in this posture should act as a large sediment trap and remove quantities of material that consequently would not be transported furtheT. Spatangoids (he art urchins ) bury in sand or mud to varying degrees depending on the species and the type of substratum (Nichols 1959b, d, Buchanan 1966, A.B.Smith 1978a). Water is circulated over the surface of the spantangoids by ciliated spines (clavules) (Gislen 1924, Chesher 1969, Nichols 1959b,d). Whether these respiratory currents carry particu­ late matter for feeding is not known, but food is obtained through the respiratory funnel of Echinocardium cordatum when the echinoid is buried (Buchanan 1966). In this case (as with Echinocardium pennatifidum and Spatangus purpureus), the introduction of oxygenated water into the substratum via the funnel results in local oxidation and the precipitation of ferric iron (Gage 1966). Locomotion of E.cordatum is disruptive to the substratum; the degree of this disruption may be to be related to availability of food, as Buchanan (I 966) reported more rapid movement in clean, as opposed to silty, sand. Kier & Grant (I965) reported that the locomotion of Meoma ventricosa also resulted in a sorting of sand. Ingestion of the substratum in feeding by spatangoids is frequently selective (Nichols 1959d, Lawrence & F erber 1971). Chesher (1969a) reported that the spatangoid M. ventricosa has a profound effect on the particle size distribution tending to shift the larger particles to the surface where the echinoids congregated, the surface of the sand was composed of coarse particles instead of being fine and poody sorted. He calculated that a single individual could turn over 330 ce of sand per hour. At a typical density of three individuals of M ven­ tricosalm 2 , 990 ce of sand could be overturned each houT. This could provide fresh surfaces for bacteria and algal growth, helping maintain the availability of food for the echinoids. Cassiduloids also ingest large quantities of sub strata. Both Apatopygus recens and Cassi­ dulus cariboearum have beenreported to feed continuously (Higgins 1974, Gladfelter 1978, respectively). Cassidulus cariboearum of 8 g wet weight contained approximately 2.5 g dry weight of sand and tumed over 7-17 % of the total gut sand per houT. As with the spatan­ goids, C cariboearum is a somewhat selective deposit feeder . Unfortunately, potential amensal effec-ts of echinoid fee ding in soft substrata, similar to that known for holothuroids (see Massin, chapter 23), have not been investigated.

1.2. Hard substratum The ingestion of hard substrat um by regular echinoids during the fee ding process has been termed bioerosion (Bromley 1975, Neumann 1966). Species of regular echinoids which have been reported to cause bioerosion are listed in table 1. Many echinoids are known to erode the intertidal or subtidal regions as the result of activity of both the spines and teeth (Otter 1932). Otter concluded that the correlation of burrows with the absence of other protection in areas of high water energy suggests that burrows result from activity induced by the environment and not from fee ding. In this way, they would serve predominantly a protective function; Yonge (1951) agreed with this conclusion. Although the burrows undoubtedly provide protection, they are not essential in aIl situations. Echinometra mathaei were usually able to survive after being removed from their burrows (Otter 1937). Fewkes (1889) discounted a protective function and basis for the burrows of Strongylocen­

E[[ects o[[eeding on the environment: Echinoidea 501 Table 1. Reports of bioerosion by regular echinoids Species

References

CIDARIDAE Eucidaris thouarsii Id. Id. Eucidaris tribuloides

Agassiz 1874

Fewkes 1890b

Glynn et al. 1979

McPherson 1968a, b, c

DIADEMATIDAE Centrostephanus rodgersi Diadema antiilarum Id. Id. Id. Id. Id. Id. Id. Diadema savignyi Diadema setosum Echinothrix calamaris Echinothrix diadema

Sinclair 1959

Hawkins 1979

Hunter 1977

Lewis 1964

Roberts 1972

Sammarco 1977

Sammarco et al. 1974

Stearn & Scoffin 1977

Storr 1964

VaIenciennes 1854

Lawrence,unpublished

Lawrence,unpublished

Cai1Iiaud 1857

STOMECHINIDAE Stomopneustes variolaris Id.

Stephenson & Stephenson 1972

Sloan et al. 1979

ARBACIIDAE Arbacia pustulosa

John 1889

TOXOPNEUSTIDAE Lytechinus variegatus Sphaerechinus granularis Id.

Neumann 1966

John 1889

Krumbach 1914

ECHINIDAE Echinus esculentus Paracentrotus lividus Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. hammechinus miliaris

Krumbein & Van der Pers 1974

Agassiz 1874

Bell 1892

Bennett 1825

Caillaud 1854, 1855, 1857

Deshayes 1855

Eales 1967

Fewkes 1890a

Fisher 1864

Forbes 1841

Gosse 1855

Hesse 1867

John 1889

Kempf 1962

Koehler 1883, 1927

Leach 1812

Lory 1855

Märkel & Maier 1967

Robert 1854

Romanes 1911

de Serres 1856, 1857a, b

Trevelyan 1849

Valenciennes 1855

Koehler 1927

502 JohnMLawrence &Paul W.Sammarco Table 1 (continued) Species

References

Psammechinus miliaris

Mortensen 1943a

ECHINOMETRIDAE Colobocentrotus pedifer Echinometra mathaei Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Echinometra lucunter Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Echinometra subangularis Id. Echinometra oblonga Id. Id. Echinometra vanbrunti Id. Id. Echinostrephus aciculatus

Morrison 1954

Agassiz 1899

Benyahu & Loya 1977

A.H.Clark 1949

H.L.Clark 1921

Döderlein 1885

Doty & Morrison 1954

Endean et al. 1956

Hodgkin 1959

James & Pearse 1969

Kelso 1970

Khamala 1971

Lawrence 1978

Morrison 1954

Morton & Challis 1969

Newell1954

Russo 1977, 1980

Sioan et al. 1979

Stephenson et al. 1931

Stephenson & Stephenson 1972

Taylor & Lewis 1970

Tokioka 1963

Umbgrove 1939, 1947

Yamazato et al. 1976

Abbott et al. 1974

Cailliaud 1857

H.L.Clark 1899a, 1933

Dong (in: Abbott 1974)

Ginsburg 1953

Grünbaum (in: Abbott 1974)

Grünbaum et al. 1978

Hunt 1969

Kaye 1959

Lawrence & Kafri 1979

Lewis 1960

McLean 1964, 1967

Ogden 1974

Rathbun 1879

Souri 1954

Stephenson & Stephenson 1950

Studer 1880

Teytaud 1971

Van Loenhoud & Van de Sande 1977

Verrill 1907

Voss & Voss 1955

Wilson (in: Abbott 1974)

Rathbun 1879

Studer 1880

Kelso 1970

Mortensen 1943b

Russo 1977

Agassiz 1874

Fewkes 1890b

Kier & Grant 1965

A.H.Clark 1952

Effects ojfeeding on the environment: Echinoidea 503 Table 1 (continued) Species

References

Echinostrephus aciculatus

ILL.Clark 1921 Lawrence 1978 Russo 1980 Campbell et al. 1973 Hughes 1977 Lawrence 1970b Sloan et al. 1979 Stephenson 1944 Stephenson et al. 1931 Stephenson & Stephenson 1972 Dix 1970a Farquhar 1894 Healy 1968 H.L.Clark 1938 Dakin 1952 McNeil & Musgrave 1926 Sinclair 1959 Stephenson & Stephenson 1972 Stephenson & Stephenson 1972 Morton 1973a, b Morton & Challis 1969

ld. ld.

Echinostrephus molaris ld. ld. ld. ld. ld. ld.

Evechinus chloroticus Id. ld.

Heliocidaris erythrogramma Id. ld. Id. Id.

Heliocidaris tuberculata Heterocentrotus mammillatus ld.

STRONGYLOCENTROTIDAE

Allocentrotus [ragilis Strongylocentrotus droebachiensis Id.

Strongylocentrotus franciscanus Strongylocentrotus purpuratus Id. Id. Id. ld.

Boolootian et al. 1959 Fewkes 1889, 1890a Stephenson & Stephenson 1972 Stephenson & Stephenson 1972 Agassiz 1874 hwin 1953 Johnson & Snook 1927 Ricketts & Calvin 1968 Stephenson & Stephenson 1972

trotus droebachiensis near Grand Manan as burrowing echinoids were found only in sub­ stratum ofhard slate. No excavations were found in nearby pools, under the same condi­ tions, where the substratum was quartzite. However, Goss-Custard et al. (1979) concluded that excavations of Paracentrotus lividus, where possible, provide protection against dis­ lodgement by waves and possibly protection against predators. Fewkes proposed that mechanical erosion resulted simply from the echinoid's activity. Even though some studies have reported the presence of substratum with encrusting algae in the gut of echinoids in such situations, this may be coincidental to the burrowing pro­ cess. The bottom of the cupshaped burrows of S.droebachiensis and Paracentrotus lividus are said to be free of encrusting coralline algae while the walls are not (Cailliaud 1854, 1857, Fewkes 1890a, Fisher 1864). This may be the result of intense feeding, lack of light, or a combination of both. Certainly the burrows of the tropical echinoids, E.mathaei, Echino­ metra lucunter, and Echinometra viridis are kept free of algal growth (Lawrence personal observation, Ogden 1974, and Sammarco personal observation, respectively). One would anticipate that the potential for echinoid movement in high water energy environments is low. It is entirely possible in such situations that reflexive fee ding activity may occur even though the food return is low or non-existent. Echinoids kept without food in the laboratory ingest any non-biological material within their capacity.

504 lohn M.Lawrence & Paul W.Sammarco

The fact remains that the production of burrows as a direct result of feeding has been demonstrated only for Elucunter in Barbados (McLean 1964, 1967). On the leeward shores where no macroscopic algae grow intertidally and where the water activity is too low to transport drift food, the guts of the echinoids found in burrows on the beachrock are filled with substratum which contains endolithic blue-green algae. McLean estimated that an individual could remove substrate at a rate of 14 cma/year. Grünbaum (in Abbott 1974) estimated that an individual E.lucunter could erode 190 g oflimestone/year. Adey (1975) considered Elucunter to be of prime importance in erosion of algal reefs in the Virgin Islands. Ginsberg (1953) considered this species to be responsible for much of the bioerosion of beachrock in the littoral and sub-littoral in the Caribbean. Kaye (1959) stated that exca­ vation by E.lucunter reduced entire tidal terrace surfaces in Puerto Rico, and the large inter­ tidal notches in East Indian coral rock have been attributed to E.mathaei (Umbgrove 1939). Russo (1980) calculated that E.mathaei and Echinostrephus acciculatus remove 0.1-0.2 and 0.2-0.4 g calcium carbonate/day/individual, respectively. He estimated that 80-325 g calcium carbonate/meter square/year was removed by both species on Enewetak Atoll. The difficulty of estimating feeding rates for burrowing species must be noted. The relationship between fee ding rates of echinoids which ingest drift food and the rate of erosion remains to be demonstrated. Burrows can serve as focal points which accelerate local erosion. Otter (1937) reported that empty echinoid burrows on coral reefs can collect sand and cause abrasion under the influence of water movement. The unlikely report that Strongylocentrotus purpuratus burrowed into steel pilings (Irvin 1953) has more recently been attributed to grazing which exposed the substratum to physical erosion (Ricketts & Calvin 1968). Intertidal burrows can also retain water and function as miniature tide pools, providing refuges from dessica­ tion for numerous intertidal species. The browsing of echinoids on superficial algae of coral reefs rasps away the topmost layer of rock (Otter 1937). Such grazing activity can result in bioerosion even though it is sufficiently non-Iocalized that burrows are not formed. Great quantities of substratum can be removed. The subtidal notches in limestone in Bermuda are considered to result partly from browsing by Lytechnius variegatus (Neumann 1966). Healy (1968) concluded that bioerosion by Evechinus chloroticus resulted in the undermining and breakup of the sea­ ward edge of shore platforms in Auckland. The grazing by Echinus esculentus on the poly­ chaete Polydora cilÜlta removes the top layer of rock to the bottom of the polychaete's burrow (Krumbein & van der Pers 1974). These authors estimated that at least a 1 cm layer of rock was removed annually as a result. A density of 1-7 E.esculentus/m 2 can apparently clear approximately 200-800 m 2 • Cleaned surfaces were quickly (34 days) overgrown with algae and other organisms, including polychaete larvae which preferentially settled near sur­ vivors. McPherson (1968c) found that the cidaroid Eucidaris tribuloides ingested 2-3 g wet weight of carbonate rock/day and suggested that it rnight result in slow erosion. J .B.Lewis (1974) also suggested that the browsing by Diadema antillarum was important in reef erosion. Diadema antillarum causes extensive pitting of the coral reefs of St Lucia (Roberts 1972). Sammarco et al. (1974) and Sammarco (1977) showed that feeding by D. antillarum altered the rnicro-morphology of Acropora palmata; 16 individuals/m 2 com­ pletely removed all distinguishable calical structure in aperiod of months. The amounts of the reef eroded by D.antillarum can be considerable, estimates ranging from ca.0.3-0.8 kg CaCO a/m2 /month (Hawkins 1979, Hunter_1977, Stearn & Scoffin 1977). The latter authors estimated that 86 % of all bioerosion of the Bellairs reef in Barbados results from the feeding

Effects of feeding on the environment: Echinoidea SOS activity of D.an tillarum, although Hawkins (1979) considered their values to be a maximal estimate. Stearn & Scoffm (1977) have calculated that bioerosion by D.antillarum exceeds that of all other organisms on the fringing reef at Barbados. 1.3. Conc/usions There seems to be no question that feeding by both irregular and regular echinoids can result in direct alteration of the environment. The evaluation of this effect thus far has been primarily descriptive, with quantitative analysis restricted to bioerosion by regular echinoids. The impact of fee ding by both irregular and regular echinoids on the environment must be investigated further, not only in terms of the effect on the substratum itself, but for poten­ tial amensal effects as weIl. 2. THE EFFECT OF ECHINOID FEEDING ON BIOLOGICAL COMMUNITIES Regular echinoids differ from irregular echinoids by having a protrusible Aristotle's lantern which allows them to feed on macroscopic organisms. Many species of regular echinoids have been noted to have an effect on biological communities as a result of their feeding activity (table 2). Questions of considerable interest are: a. the effects of feeding on the epibenthic plant components of the community; b. the effects of fee ding on the epibenthic animal components of the community; c. the limitations on echinoid distribution, abundance, and mobility which in turn dif­ ferentially distribute the effects of their feeding; and d. stability of echinoid populations which affects the long-term role of echinoids. 2.1. The effects on plants 2.1.1. Effect on biomass Numerous reports of increased epibenthic plant biomass or cover following a decrease in echinoid abundance in the field have implicated the echinoids as controlling agents (see Lawrence 1975a). Experimental manipulation of echinoid densities has shown the density at which the effects of various species can be observed. Densities of Paracentrotus lividus of 0, 1,3 and 6/2.75 m2 resulted in 100,33-50,30, and 0 % cover, respectively (Kitching & Ebling 1961, 1967). Twenty Strongylocentrotus polyacanthus/m2 reduced algal cover by 60-70 % within aperiod of nine months (Palmisano & Estes 1977). A biornass of 2 kg (wet weight) of Strongylocentrotus droebachiensis/m 2 resulted in kelp regression (Breen & Mann 1976a). Density manipulation experiments involving S.droe­ bachiensis, Strongylocentrotus franciscanus, and Strongylocentrotus purpuratus demonstra­ ted that these echinoids clearly control algal abundance (Irvine 1973a,b, Paine & Vadas 1969, Vadas 1968). Sixteen Diadema antillarum/yd 2 were unable to prevent the growth of algae (Randall1961). Sammarco (1977, 1980a) found a negative log-linear relationship between density of D.antillarum and algal cover. Experimental densities ranging from approximately 0-64/m2 decreased algal percent-cover, including corallines and rnicroalgae, from 100 % to as low as 22.5 % under both primary succession conditions (fresh, bare substratum) and post-primary conditions (natural recent reef substratum). It is important that the level of

506 lohn M.Lawrence & Paul W.Sammarco Table 2. Studies noting an effect of feeding by regular echinoids on the biological community Species CIDARIDAE

Eueidaris thouarsii Eueidaris tribuloides DIADEMATIDAE

Astropyga radiata Centrostephanus eoronatus Id.

Centrostephanus rodgersi Diaderruz antilfarum Id. Id. Id. Id. Id. Id.

Diaderruz mexieana

Diaderruz setosum

Id.

Diaderruz sp.

Eehinothrix cafarruzris

Eehinothrix diaderruz

Eehinothrix sp.

TOXOPNEUSTIDAE

Lyteehinus anamesus Id.

Lyteehinus variegatus Id. Id.

Lyteehinus williamsi Tripneustes ventrieosus Id.

ARBACIIDAE

Arbaeia punetulata Id.

Arbaeia lixufa Id. Id. Id.

ECHINIDAE

Echinus eseulentus Id. Id. Id. Id. Id.

Eveehinus ehlorotieus Id. Id. Id. Id.

Loxeehinus alba Paraeentrotus lividus Id. Id. Id.

References Glynn et al. 1979

Sammarco 1977

Bak & Nojima 1980

Vance 1980

Vance & Schmidt 1979

Shepherd 1973

Bak & Van Eys 1975

Ogden, Abbott & Abbott 1973

Ogden, Brown & Salesky 1973

Randall1961

Smith 1973, 1975

Sammarco 1977, 1980

Sammarco et al. 1974

Glynn et al. 1979

Dart 1972

Ormond & Campbell 1971

Doty 1971

Castro 1971

Ormond & Campbelll971

Doty 1971

Leighton 1971

Leighton et al. 1966

Camp et al. 1973

Keller 1976

Greenway 1977

Sammarco 1977

Keller 1976

Lilly 1975

Karlson 1975, 1978

Guida 1976

Kempf 1962

Larkum et al. 1967

Neill & Larkum 1966

Solazzi 1968

Drew et al. 1967

Forster 1959

Jones & Kain 1967a, b

Jorde & Klavestad 1963

Kain & Jones 1966

Lewis 1964

A.L.Ayling 1978

A.M.Ayling 1976

Don 1975

Droomgoole 1964

Lukens 1974

Dayton et al. 1973

Gamble 1967

Kempf 1962

Kitching & Ebling 1961, 1967

Larkum et al. 1967

Effects of feeding on the environment: Echinoidea 507 Table 2 (continued) Species

References

Paracentrotus lividus Id. Id. Id. Id.

L.R.Lewis 1964

Neill & Larkum 1966

Neill & Pastor 1973

Ott IX Mauer 1976

Solazzi 1968

ECHINOMETRIDAE Echinometra lucunter Echinometra viridis Heliocidaris erythrogramma Heterocentrotus mammillatus

Lawson 1966

Sammarco 1977

Shepherd_1974

Dart 1972

STRONGYLOCENTROTIDAE Strongylocentrotus droebachiensis Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Strongylocentrotus franciscanus Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Strongylocentrotus polyacanthus Id. Id. Id. Id. Id. Id. Id. Strongylocentrotus purpuratus Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id. Id.

Breen & Mann 1976a, b

Fewkes 1889, 1890

Himmelman & Steel1971

Irvine 1973a, b

Lang & Mann 1976

Lubchenco & Menge 1978

Mann 1972,1973,1977

Mann & Breen 1972

Miller et al. 1971

Newcombe 1935

Propp 1971, 1977

Vadas 1968, 1977

Birkeland 1971

Ebert 1977

lrvine 1973a, b

Lees 1970

Leighton 1971

Leighton et al. 1966

North 1964, 1965

North & Pearse 1970

Mattison et al. 1977

Paine & Vadas 1969

Vadas 1968, 1977

Barr 1971

Dayton 1975b

Estes 1974

Estes & Palmisano 1974

Estes et al. 1978

Palmisano 1975

Palmisano & Estes 1977

Simensted et al. 1978

Agegian 1978

Dayton 1970, 1975a, 1975b

Lees 1970

Leighton 1971

Leighton et al. 1966

Lowry & Pearse 1973

McLean 1962

North 1964, 1965

North & Pearse 1970

Paine & Vadas 1969

Pearse & Hines 1979

Sousa 1977,1980

Vadas 1977

Wilson et al. 1977

508 lohn MLawrence & Paul W.Sammarco resolution - macro- vs microalgae - be taken into consideration in comparing independent studies on effects of echinoid grazing or grazing by any other herbivore. Percent-cover data usually yield nominal information on changes in algal biomass. Den­ sities of D.antillarum over a range of 0-64/m2 reduced algal biomass from as high as 1.3 g (decalcified dry weight)/300 cm 2 to zero or negligible amounts under primary succession conditions over aperiod of ~ 10 months (Sammarco 1977). Changes in grazing pressure under post-primary succession conditions do not always result in predictable changes in algal biomass, depending upon the nature or, more appropriately, the species composition of the algae initially present, e.g. Peyssonelia sp., an encrusting brown alga, is not easily removed from the benthic substratum by grazing and withstands average densities of D. antillarum of 731m2 (Sammarco 1977). The comparative effects of grazing by sympatric species of echinoids on algal abundance are interesting. The impact of fee ding by echinoids on the benthic community is species­ specific, e.g. successful algal recolonization occurs in the northeastern Pacific at densities of 1 S./ranciscanus/m2 0r 10 S.purpuratus or 10 Lytechinus anamesuslm2 (Leighton et al. 1966). Arbacia lixula and P.lividus occur in sirnilar, overlapping habitats in the Mediterra­ nean, but the latter has less of an effect on the environment than the former, which rasps Of abrades the substratum more severely (Kempf 1962). Similarly A.lixula influences algal community structure by eliminating sporelings in feeding, while P.lividus feeds only on established macroalgae (Neill & Larkum 1965, Larkum et al. 1967). Diadema antillarum and Echinometra viridis are commonly associated in the subtidal of Jamaica and probably compete for food, yet they appear to have very different degrees of impact on algal abundance. On natural recent reef snbstratum, average densities of D. antillarum as high as 72.8/m 2 allow mean algal biomasses of 1.5-7.1 g (decalcified dry weight)/0.25 m 2 while E. viridis densities of ca.45/m 2 allow up to 14.7 g/0.25 m2 (Sam­ marco 1977). These high densities of D.antillarum reduce percent-cover to 37 %, while high densities of E. viridis reduce it to 53 %. These differential effects are probably due to species-specific fee ding and migratory behaviors in the echinoids. Diadema antiilarum generally feeds nocturnally, making extended excursions over reef areas and sand flats (Randall et al. 1964, Ogden et al. 1973a,b, Smith 1973,1975). Echinometra viridis remain dose to its refuge, fee ding only in the immediate vicinity of its microhabitat. Echinometra lucunter exhibits a similar lack of movement (Abbott et al. 1974, Grunbaum et al. 1978). Leighton et al. (1966) reached similar conclu­ sions concerning Lytechinus anamesus which has a more superficial, less intense effect on algal abundance and community structure than S./ranciscanus and S.purpuratus due to its more pronounced migratory behavior and propensity for climbing. 2.1.2. Effect on community structure Feeding preferences vary greatly between species of echinoids and influence the effects of feeding on benthic community structure. Forster (1959) observed areas browsed by Echi­ nus esculentus and noted that any food preference would affect the proportions of prey species. This concept is the major basis for interpretation of the effects of grazing by regular echinoids on community strUGture (e.g. Lubchenco 1978). Although echinoids have prefe­ rences, the actual diets in the field are often determined by food availability, as is the usual case with generalists (see Lawrence 1975a, Ogden 1976). Ayling (1978) has shown that Evechinus chloroticus exhibits distinct food preferences in the laboratory which remain consistent between populations from different habitats. Despite this, the echinoid's field

E[[ects o[ [eeding on the environment: Echinoidea 509 diet is determined solely by abundance of prey. Fuji (l967) and Kawamura (1973) have noted similar responses in Strongylocentrotus spp. Lees (1970) believed this extreme flexi­ bility in potential diet may be a result of past selection against strong preferences or specialist feeding behavior. Similarly, Vance & Schmitt (l979) contended that the dietary breadth of Centrostephanus coronatus is an indirect evolutionary response to regulation of its abun­ dance and distribution by a predator, the sheephead Pimelometopon pulchrum. Vadas (l977) also found that S.droebachiensis and S.[ranciscanus exhibited strong food preferences in the laboratory; however, in this case relative abundance of prey species in the field was not correlated with feeding frequency. Feeding rates were clearly higher on pre­ ferred algae. Here feeding activities in the field represented a compromise between preference and availability, since absorption efficiencies, growth and reproduction were maximal when the echinoids were fed on a diet of preferred food. Strongylocentrotus purpuratus exhibits less pronounced food preferences than its congeneric sympatrics. Differences in food preferences between echinoid species can sometimes account for differential species-specific effects oftheir grazing, e.g. S.droebachiensis and S.[ranciscanus have greatly different effects on subtidal algal community structure as a result of both species-specific feeding preferences and feeding rates (Irvine 1973a). In contrast, S.[rancis­ canus and S.purpuratus exhibit extremely similar food preferences but have different im­ pacts on algal community structure due to the higher feeding rate and degree of mobility of the former. Lytechinus variegatus and Tripneustes ventricosus commonly occur sympatri­ cally in Jamaica, but the latter has a greater effect on Thalassia testudinum. Tripneustes ventricosus in grass beds feeds almost exclusively on T. testudinum while L. varietagus has several modes of feeding: herbivory on T. testudinum when it is available and detritivory when it is not (Keller 1976). Echinoid grazing affects not only algal percent-cover, but algal species composition as weIl. Natural density shifts of echinoids in the field have been observed by Leighton (1960, 1971) in S.purpuratus, S.[ranciscanus, and L.anamesus. Invasions by these species result in rnarked changes in species composition, the effect being confounded by differences in pre­ ferential feeding. Experimental removal of these species from California coastal waters results in the reappearance of previously rare algal species (Leighton et al. 1966). The removal of Echinus esculentus yielded similar results along with a pronounced shift in domi­ nant species including the establishment and persistence of the perennial Laminaria sp. (Jones & Kain 1967b). Dayton (1975b) found that echinoid grazing is a limiting factor determining the lower depth limits of Laminaria, Agarum, and Alaria. Randali (1961) found that the experimental exclusion of Diadema antillarum resulted in the survival of the red alga Spyridia filamentosa and the brown Dictyota bartayresii. Sammarco (1977) demonstrated experimentally that an increase in the density of D.antil­ larum resulted in a shift in dominants from fleshy reds (e.g. Cordylecladia peasiae and Her­ posiphonia spp.) and greens (Qadophora spp.) to minute greens, corallines, and blue-greens under primary succession conditions. Under extreme grazing pressures, almost all algae are consumed except the most hardy species such as the corallines and some opportunistic algae including bangioides and colonial diatoms. This is similar to the fmdings of other investigators for other echinoid species (Forster 1959, Kempf 1962, Dromgoole 1964, Lawson 1966, Leighton etal. 1966, Vadas 1968, Propp 1971, 1975, Dayton etal. 1973, Luckens 1974, Dayton 1975a, Lubchenco & Menge 1978). Certain toxic or distasteful algae (Caulerpa spp. and Laurencia spp.) can persist despite extraordinarily high echinoid densities (Ogden 1976, Sammarco 1977), althoughL.variegatus eats Caulerpa sp. (Lowe &

510 John M.Lawrence & Paul W.Sammarco Lawrence 1976) and Echinothrix calamaris will feed upon Laurencia sp. in the laboratory (Castro 1971). Several investigators have concluded that echinoids playa greater role in controlling epibenthic community structure than herbivorous moIluscs, primarily as a result of the effectiveness of the Aristotle's lantern. The echinoid's size and their tube feet's ability to secure and hold macroscopic algae also provide an advantage (Luckens 1974, Dayton 1975a, Dayton et al. 1977, Sousa 1977). This is probably highly variable with respect to the spe­ cies in question, e.g. although when fishing activities reduced densities of Loxechinus albus to very low levels, the feeding activities of dense populations of herbivorous gastropods allowed only encrusting coralline algae to persist in certain areas (Dayton et al. 1977). Ayling (personal communication) found no basic change in community structure after Evechinus chloroticus was excluded from a subtidal area in New Zealand, although there was an increase in cover by ephemeral algae and coralline turf algae. His study demonstrated an impact of herbivorous gastropods along with that of the echinoids. Removal of the gas­ tropods resulted in an increase in algal cover despite continued echinoid grazing. Other groups of grazers and browsers should not be underestimated in their ability to control ben­ thic community structure. Leighton (1971) has listed numerous kelp predators and North (1978) has reported the control of kelp by crabs and fish as weIl as by echinoids. Studies investigating the effects of echinoid grazing have also yielded important infor­ mation about the relationship between grazing and algal diversity, which has broad ecologi­ cal implications. Vadas (1968) and Paine & Vadas (1969) proposed that algal diversity would be greatest at intermediate levels of grazing, supposedly due to reduced levels of competition between algae for resources such as space and light. This effect is particularly pronounced if the competitively dominant algae is the preferred prey of the grazer. This hypothesis was based predominantly on correlative data drawn from field measurements of natural algal species diversity (number of species) and echinoid density in subtidal benthic communities from various sites in Washington's coastal waters. Lubchenco (1978) experimentally manipulated the densities of grazing gastropods in the New England intertidal and found a gene rally negative functional relationship between grazing pressure and diversity when the herbivore preyed selectively upon a competitive inferior. If the herbivore preferred the competitive dominant, a curvilinear relationship resembling Paine & Vadas' prediction resulted, with a peak in diversity at intermediate grazing pressures. Experiments in St Croix testing the same hypothesis (Ogden et al. 1973, Sammarco et al. 1974) involved the removal of an entire population of D.antillarum from a lagoonal patch reef. Within three months, despite a significant increase in the number of algal species, the reef became monopolized by a partially calcified phaeophyte - Padina sanctae-crucis - which significantly reduced algal diversity and equitability. Other non-treatment reefs possessing intermediate densities of D.antillarum maintained relatively high levels of algal diversity. These results also supported Paine & Vadas' (1969) hypothesis. SirniIar but more complex and longer-term experiments in Jamaica gave very different results (Sammarco 1977). The removal of D.an tillarum from both small enclosed areas and a whole patch reef resulted in a significant increase in algal diversity (number of species and H'), and a negative log-linear functional relationship between density of D.antillarum and algal diversity. This did not support Paine & Vadas' hypothesis, or agree with the previous results with D.antillarum A statistical re-analysis of the data from which the original Paine & Vadas hypothesis was drawn did not demonstrate a significant peak in diversity at inter­

E//ects 0//eeding on the environment: Echinoidea 511 mediate grazing pressures but showed a significant negative log-linear relationship (Sam­ rnarco 1977). The ecological principles upon which their hypo thesis was based were none­ theless quite sound. It is possible that either a curvilinear or negative linear relationship can exist between these two variables, but the precise nature of the relationship for any given community may be dependent upon stochastic successful recruitment of the competitive dominant, its growth rate, and time (Sammarco 1977). This may be described through a three-part model where­ by: 1. If no competitive dominant exists in the community, a negative linear relationship may result, 2. If the competitive dominant recruits highly successfully, the curvilinear relationship rnay result, 3. Ifthe competitive dominant possesses a slow growth rate and/or recruits at a low rate, the negative linear relationship will initially emerge and gradually develop into a curvilinear one through time.

2.2. The e//ects on epibenthic animals Although most regular echinoids are herbivorous, they may be omnivorous or entirely car­ nivorous when animals are the primary food available. The primitive Perischoechinoidea are predominantly carnivorous (see Lawrence, 1975a). The effects of this predation are direct, but effects also may result from incidental, non-directed browsing or ingestion. This type of effect has been termed 'biological disturbance' by Dayton (1971).

2.2.1. Prey-predator relationships Although Allen (1899) reported that Psammechinus miliaris did great damage to oyster beds, Orton (1924) and Hancock (1955) concluded that the echinoid was feeding primarily on the epi- and endozoic fauna on the oyster shell which indirectly affected the oyster's viability. McNeill & Iivingstone (1926) cited anecdotes of the destruction of clams by Echinometra mathaei on reefs in the South Pacific. They also concluded that the burrowing of E.mathaei could dislodge Tridacna maxima and were thus responsible for the absence of the clam on the outer edge of the reef, although Mortensen (1943b) pointed out that there was no proof for this. Subtidal communities of Mytilus edulis in the Bay of Fundy have been reported to be controlled by grazing by Strongylocentrotus droebachiensis, so that the mussei existed only in the intertidal beyond the occurrence of the echinoid (Newcombe 1935). Bertram (1936) suggested that Diadema setosum controls the local distribution of sessile epibenthic reef fauna, by grazing on attached organisms. Grazing by Diadema antiilarum and other echinoids significantly deter settlement in the shallow epibenthic reef community (Sammarco 1977, Schumacher 1974). By experimentally varying densities of D.antillarum from 0 to 641m2 , Sammarco (1977,1980) found that coral settlement decreased in at least a genus-specific manner from 371 to 13 spatl0.25 m2 under primary succession conditions. Coral survival, growth, and competitive success, however, were greatest at intermediate densities (ca. 41m 2 ) at which both competition for space with algae and the intensity of biological disturbance from grazing were relatively low. Densities of Echinometra viridis of 451m 2 failed to produce similar effects, apparently due to their more limited movement and localized grazing. Karlson (1975, 1978) found a correlation between intensity of graz­

512 lohn MLawrence & Paul W.SammIJrco

ing by Arbacia punctulata on the structure of the epibenthic animal community which he attributed primarily to feeding preferences and predation on colonies of Hydractinia. The appe arance and fee ding of Echinus melo results in destruction of dense bryozoan stock (Strenger & Splechtna 1978). Guida (1976) suggested that predation by A.punctulata is a major factor in selection for the cryptic boring habitat of Cliona celata. lt has been experi­ mentally demonstrated that successful settlement in corals becomes more cryptic as grazing pressure by DiademIJ antillarum increases (Sammarco 1977). Vance (1980) found that grazing by Centrostephanus coronatus exerts great control over the flora and fauna of subtidal rock surfaces. The release of grazing pressure resulted in a shift in the composition of the community from coralline algae and encrusting ecto­ procts to fleshy algae, sponges, tunicates, and erect ectoprocts. Grazing tended to decrease local diversity, but Vance believed that it might actually increase diversity on a larger scale by creating patches in the environment (sensu Levin & Paine 1974). Sammarco (1977, 1980a) also found that the release of grazing pressure by D.antillarum resulted in an increase in coral spat diversity under primary succession conditions. In contrast, the opposite occurred under post-primary conditions due to a drop in species evenness in settlement (Sammarco 1977). The same result occurred if Echinometra viridis, Lytechinus williamsi and Eucidaris tribuloides were simultaneously eliminated.

Figure 1. Energy flow through the bottom community in the seaweed zone of St Margaret's Bay, Nova Scotia. Units are kcal/m 2 per year, except for biomass, which is in kcal/m 2 (from Miller et al. 1971).

Effects of feeding on the environment: Echinoidea 513

The effects of echinoid fee ding can extend throughout the various components of the community. An energetic analysis of trophic interrelationships involving S.droebachiensis in St Margaret's Bay, Nova Scotia (fig.l) clearly illustrates this point (Miller et al. 1971). This model has been used to describe the effects of changes in predator and echinoid bio­ mass (Breen & Mann 197 6a,b). There is a difference in total community structure based on trophic interrelationships involving Strongylocentrotus polyacanthus in the presence and absence of predation on the echinoids (fig.2; Estes & Palmisano 1974, Palmisano & Estes 1977, Estes et al. 1978, Simenstad et al. 1978). 2.2.2. Non-trophic relationships Non-trophic, higher order interactions also occur between the echinoid Heliocidans erythro­ gramma and the razor shell Pinna dolobrata which occur on the seagrasses of sublittoral platforms in the Spencer Gulf (Shepherd 1974). Here trre razor shell provides a suitable settling substratum for the echinoid. A limited positive feedback system was found; increases in density of Herythrogramma resulted in increased fee ding on the seagrasses which, in turn , significantly increased the density of P.dolobrata (fig.3). In contrast, tne shells of the bi­ valve Chama pellucida provide arefuge to epibionts from the grazing by Centrostephanus coronatus, suggesting a mutualistic effect: the epibionts causing less predation on the bi­ valve by asteroids (Vance 1978). Intertidal burrows formed by echinoids (see seetion 1.2) retain water and affect local bio ta by functioning as miniature tide pools. The burrows of Echinometra mathaei furnish shelter for numerous organisms (Morrison 1954). Morton & Challis (1969) found a wealth of gastropod molluscs within the depressions formed by E. mathaei and He tero cen tro tus

Figure 2. Diagram of the interactions within nearshore communities of the western Aleutian Islands with and without sea otters (from Palmisano & Estes 1977).

514 lohn M.Lawrence & Paul WSammarco

Figure 3. Postulated relationship between the density of seagrass, the razor shell Pinna dolobrata and the echinoid Heliocidaris erythrogramma (from Shepherd 1974).

mammillatus. The empty burrows of Strongylocentrotus purpuratus are known to shelter limpets and chitons (Ricketts & Calvin 1968). Thus, the mere presence of a population of any species of echinoid exhibiting this characteristic may increase environmental hetero­ geneity, providing refuge for other epibenthic members of the community. An unusual situation in which Psammechinus miliaris feeds on locally abundant accu­ mulations of detached Laminaria saccharina has been reported (Bedford 1978). The echi­ noid facilitates decomposition of the alga and increases the surface area available to sapro­ phagous micro-organisms and decreases the amounts of nutrients available to meiofauna. The feeding activities of irregular echinoids also have secondary biological effects. The movements of the clypeasteroid Dendraster excentricus affects the distribution and abundance of the seapen Renilla kollikeri, as the latter is uprooted by the echinoid's movement (Kas­ tendiek 1975). Similarly, the activity of D. excentricus reduces the density of tube-building tanaid crustaceans with the result that newly metamorphosed individuals of the sand dollar (which settle preferentially in the vicinity of the adults) have increased survival (Highsmith 1977). The inclined orientation of D.excentricus in soft sediment at moderate current velocities creates eddies. This might affect the feeding abilities of other suspension feeders (Merrill & Hobson 1970) and could be considered a form of interspecific interference com­

Effects of feeding on the environment: Echinoidea 515

petition. It is possible that the fee ding activities of soft-bottom echinoids could significantly affect meiofaunal species composition, but this remains to be demonstrated. Echinoids are a group of browsers or detritovores which may contribute to the control of soft-bottom community structure (see Rhoads & Young 1970).

2.2.3. Interphyletic competition Interphyletic competition between echinoids and other grazers or browsers have been noted, e.g. echinoids are capable of fee ding on both drift and attached algae, but abalone feed primarily on the former (l.eighton 1971, Lowry & Pearse 1973, Shepherd 1973). For this reason, Strongylocentrotus purpuratus has been considered a better competitor for food thanHaliotis referens (l.eighton 1971), and Centrostephanus rodgersi a better com­ petitor than Haliotis ruber (Shepherd 1973). Certainly the echinoids possess broader niches with respect to feeding than their associates, but their comparative competitive abilities remain to be experimentally defined. When food is not limiting or when common predators are abundant, competition for refuges predominates. In such cases, abalone may have an advantage over their echinoid associates (Lowry & Pearse 1973). Diadema antillarum also seems to represent a significant threat - possibly competitive - to Eupomacentrus planifrons, the common three-spot damselfish. The damselfish main­ tains active sustained a.ttacks against individual D.antillarum which enter its territory or algal turf du ring the day (Sammarco, unpublished). Williams (1977, 1978a,b, 1979) con­ sidered this interaction a result of competition for food. However, it might also be a result of the male damselfish protecting the eggs within its territory from predation by the omni­ vorous echinoid. In either case, coral settlement and survival may be affected as a result of localized absence or reduction of grazing by the echinoid (Sammarco 1977, 1980a,b).

2.3. Effects of limitations on echinoid distribution, abundance and mobility Since the effects of echinoid fee ding vary with density, any factor which influences echinoid distribution and abundance can also potentially affect benthic community structure. These limiting factors can be either physical or biological in nature.

2.3.1. Physical factors Wave action can significantly reduce or entirely eliminate echinoids locally in the intertidal or subtidal. Fewkes (1889) noted Strongylocentrotus droebachiensis at high densities in the upper subtidal at Grand Manan, while ridges immediately above were covered with kelp. In the Bay of Fundy, the functional upper limit of this echinoid corresponds to the lower limit of Mytilus edulis (Newcombe 1935). Himmelman & Steele (1971) found that S.droebachien­ sis moved upward during periods of high surge to feed and that the degree of fee ding was corre­ lated with the amount of exposure. The amount of energy correlates with the density of S. droebachiensis (Lubchenco & Menge 1978). In the northeastern United States, the high wave action apparently depresses its abundance, allowing the alga Chondrus crispus to dominate. In the Bay of Fundy, the opposite is the case and the alga is scarce. Wave action also affects the local distribution of a number of other species. In l.eigh, New Zealand, Evechinus chloroticus and Lunella smaragda graze intensively only up to a level which corresponds to the lower limit of the alga Cardophyllum spp. (Chapman 1966). Feeding activity of Eucidaris tribuloides is reduced in areas of high surge (McPherson 1968a,b). The vertical distribution of Diadema sp. and Echinothrix sp. in Hawaü is con­

516 lohn M.Lawrence & Paul W.Sammarco trolled by their inability to withstand a high degree of wave-energy (Doty 1971) and their upper limit corresponds to the lower limit of the alga Eucheuma sp. John & Price (1979) attributed the lack of algae in rocky areas, except where great wave energy occurred, to echinoid grazing. Goss-Custard et al. (1979) concluded that the hardness and unsuitability of the slates at Carriagathorna for burrowing restricted the presence of Paracentrotus lividus to a very few crevices. Schroeter (1978) found that Strongylocentrotus purpuratus dominated habitats where Strongylocentrotus franciscanus were absent because of the greater tolerance of the former to desiccation, heat stress, or the effects of wave and surge action. Physical pertur­ bations induced by man can also act as a limiting factor in echinoid grazing systems just as in numerous other communities. Dawson et al. (1960) documented an oll spill which caused mass mortality in echinoids which resulted in a significant increase in algal abun­ dance.

2.3.2. Biological factors Biotic factors also affect echinoid distribution, abundance, and movement. Drift algae are frequently a major food supply for echinoids (see De Ridder & Lawrence, chapter 4), and an adequate supply results in decreased movement (Leighton et al. 1966, Lees 1970, Mat­ tison et al. 1977). Caging, one of the major techniques utilized to investigate certain ecolo­ gical problems associated with echinoids, may also affect echinoid feeding by reducing or eliminating the possibility of predation on the echinoids and the availability of drift mate­ rial. Thus, caged echinoids may be expected to move and come into contact with attached food sources to a greater extent than usual. Drift algae can potentially be produced by echinoid grazing and storms, which can serve as a valuable food source for other species in deeper water (see Lawrence 1975a). Predation has an impact on echinoid distribution and it has been suggested that this can result in cryptic habits. This behavior may be exhibited either continuously (Lowry & Pearse 1973) or diurnally (Thornton 1956, Kitching & Ebling 1961, Lawrence & Hughes­ Games 1972, D.P.B.smith 1973, 1975, Nelson & Van~~ 1979). If refuges are not available, echinoids often aggregate, possibly as a defense against pre­ dation. This behavior has been noted for Diadema spp. (Valdez & Villalobos 1978, Pearse & Arch 1969). The latter believe that this is indicative of true social behavior in diadematids, because of recognizable patterns in spine contact which appear between individuals. Strongy­ locentrotus droebachiensis also exhibits aggregative behavior of two types (Garnick 1978). Due to their strong chemosensory abilities, they concentrate in open areas of high algal abundance for fee ding. When not feeding, they sense other conspecifics and aggregate in cryptic areas. The degree of crypticity in echinoids may alter both the intensity and the distribution of their effects on epibenthic community structure - both temporally and spatiaIly. Lees (1970) has postulated that selection for stationary cryptic behavior has occurred, and that any change in this behavior would be advantageous only under conditions of reduced food availability. Predation affects echinoid abundance as weIl as distribution, potentially reducing or even eliminating them from certain areas. This, in turn, depresses the effects of their grazing. If predation pressure is sufficiently high, echinoid movement can decrease, either because of the possibility of predation andj or because of increased food availability resulting from decreased grazing. This occurs in Strongylocentrotus purpuratus in the presence of high den­

Effects of feeding on the environment: Echinoidea 517 sities of the sea otter Enhydra lutris (Lowry & Pearse 1973, Estes 1974, Estes & Palmisano 1974, Palmisano 1975, Palmisano & Estes 1977, Estes et al. 1978) or Pycnopodia helian­ thoides and Anthopleura xanthogrammica (Dayton 197 5a). Eucidaris thouarsiiisknown to browse on coral in the tropical Pacific, and its effects are particularly pronounced on the Galapogos Islands due to the echinoid's high density there. This is not the case on the mainland coast where it is believed intensive fish predation depresses the echinoid's population density (Glynn et al. 1979). The possibility that competition controls echinoid distribution has received little atten­ tion. Schroeter (l978) suggested that S.purpuratus is not found in the sheltered low inter­ tidal and subtidal habitats, even though they are preferred, because of competition with Strongylocentrotus franciscanus which forces emigration.

2.4. Stabi/ity in echinoid populations There is some question about the stability of low- and high-density populations of echinoids. This refers to stability as either persistence of a particular state in time (persistence stability) or repeating, cylical fluctuations (adjustment stability; Margalef 1969). Scott (l902), considering Strongylocentrotus droebachiensis, believed that persistence stability was the case on the Canadian coast. Supposedly, any increase in echinoid density would result from an 'abnormal' increase in their food supply or a decrease in predator abundance, and an equilibrium between echinoids and seaweeds occurred due to their long interaction. Mann (l977) evaluated the stability of an ecosystem in St Margaret's Bay, Nova Scotia, involving S.droebachiensis. Mann and others (Miller et al. 1971, Mann 1972, 1973, Mann & Breen 1972, Breen & Mann 1976a,b) followed changes in community structure and the interrelationships of certain component species over aperiod ofyears. They des­ cribed an initially stable kelp ecosystem with a moderate population of herbivores, rnainly S.droebachiensis. The echinoids were preyed upon primarily by the crab Cancer irroratus and the lobster Homarus americanus. Beginning in 1968, echinoids became locally abundant and overgrazed the kelp beds, producing large areas of echinoid-dominated barren grounds. Since lobster biomass decreased during the same period, they concluded that the reduction in predation led to increased echinoid abundance, kelp bed destruction and the establish­ ment of at least a new quasi-stable community structure. Here low levels of primary and secondary production could be expected to persist over long periods of time. This persis­ tence in the apparent absence of food was due to their ability to subsist on low levels of drift material (see Lawrence 1975a; see chapter 15). The subtidal populations of S.droeba­ chiensis are not necessarily as stable. Himmelman & Steele (l971) have reported highly suc­ cessful settlement at certain times and places, but no long-term observation~ have been made. Similarly, work done with Strongylocentrotus polyacanthus in the Aleutian Islands (Estes 1974, Estes & Palmisano 1974, Palmisano 1975, Palmisano & Estes 1977) has led Simenstad et al. (l978) to propose the existence of alternate stable-state communities (sensu Sutherland 1974), defmed by the presence or absence of dense sea otter populations. Com­ munity structure supposedly be comes stabilized following a change in predator density. It was proposed that a macroalgal-dominated community developed from a herbivore­ dominated one in 20-30 years as the sea otter population became re-established and dense on certain islands. Shepherd (1974) also ascribed stability to the epibenthic community involving Heliocidaris erythrogramma (see section 2.2), and suggested that a disturbance might affect the stability of the system.

518 John MLawrence & Paul W.Sammarco The shift from macroalgal- to echinoid-dominated communities in southern California was similarly exemplified by the decline of the coastal kelp-beds (North 1964, 1967, Leighton et al. 1966, North & Pearse 1970, Pearse et al. 1970). Although the sea otter was approaching extinction by 1900, the kelp beds did not begin to decrease in extent imme­ diately, an important point made by Randali (1972). A combination of other factors (e.g. increase in water temperature) depressed kelp settlement and growth, and was later com­ pounded by an increase in echinoid density (particularly Strongylocentrotus [ranciscanus and Strongylocentrotus purpuratus). High densities of these echinoids were maintained by the discharge of organic material into some areas. This community persisted through con­ tinued recruitment until an intervening program of echinoid destruction and kelp propaga­ tion was instituted, leading to the re-establishment of the bed. Nevertheless, even after years of effort, the new macroalgal-dominated community is not yet stable, for continual efforts are required to maintain and extend it (North 1978). The absence of an immediate increase in echinoid abundance after elimination of the sea otter demonstrates the importance of stochastic recruitment processes and the result­ ing variability in time-lag related phenomena (see Sammarco 1978). Other echinoid preda­ tors may also playa significant causative role in delays of population growth (see Lawrence 1975). Recruitment or survival of juvenile echinoids is variable within kelp beds. Mann (1973) reported that settlement of S.droebachiensis is highest in the kelp understory and that the abundant detritus present provides food for the juveniles. He concluded that pre­ dation normally kept the numbers of survivors low. Although Dayton (197 5b) suggested that the haptera provide some refuge for juveniles of S.polyacanthus, North (1978) reported a strong inverse correlation between recruitment success of Strongylocentrotus spp. and algal biomass. Patchiness in echinoid settlement is not uncommon in nature (e.g. Ebert 1966, Moore & Lopez 1972). Cyclical fluctuations in echinoid populations with long-term adjustment stabilities have received only a minor amount of attention. This phenomenon clearly occurs in the southern California kelp-beds, where periodic destruction of kelp by echinoids and subsequent re­ establishment occurs with aperiod on the order of several years to a decade (North 1964, Pearse et al. 1970). The echinoids evidently do not persist in sufficient numbers to prevent completely the settlement and subsequent growth of sporelings. Pearse & Hines (1979) des­ cribed the expansion of a central California kelp forest following the mass mortality by disease of S.[ranciscanus. The best-documented study of the effects of cyclical fluctuations in echinoid populations involves S.droebachiensis in the Strait of Georgia as it contrasts with the situation in Nova Scotia. Foreman (1977) found evidence that populations of S.droebachiensis in the strait undergo periodic environmentally controlled outbreaks which are responsible for localised perturbations of the benthic macrophyte community. He proposed that the increase in den­ sity of the echinoid occurred in the marginal ranges of its habitat when specific physical factors provided favorable conditions for increased reproductive success. The echinoids, in response to conditions resulting from their overgrazing, eventually emigrated. Local densities again declined, allowing algal recolonization and growth. Migratory effects may be characteristic of such genera as Astropyga. Bak & Nojima (1980) reported an invasion by Astropyga radiata of a patch of the eelgrass Zostera marina. The patch was completely eliminated, with the echinoids eating the upper parts of the turions fust; subsequently, the entire turion was eaten. The removal of the eelgrass not only resulted in the disappearance of the usual fauna for the area, but erosion of the substratum as weIl.

E//ects 0//eeding on the environment: Echinoidea 519 Massive settlement and elimination of macrophytes by dense populations of juvenile echinoids has been reported for Lytechinus variegatus (Camp et al. 1973) and Tripenustes ventricosus (Lilly 1975). In both instances, the seagrass Thalassia testudinum was devoured to the surface of the substratum. The high density of L. variegatus persisted only a few months, undergoing a mass mortality - evidently the result of reduced salinities (Lawrence 1975c). The population of T.ventricosus declined gradually through time. Extraordinary periodic fluctuations in tidallevel can also result. in mass mortalities of echinoids (Glynn 1968). All of the species thus far investigated with respect to community stability have common biological characteristics: high growth rates and high fecundities (Ebert 1975). It would be of interest to compare echinoids exhibiting low growth and fecundity rates with those possessing high rates and their relative effects on benthic community structure.

2.5. Conclusion The work done thus far has demonstrated that fee ding by regular echinoids can significantly affect the composition and abundance of their prey and associated organisms. These effects can be related to food preferences, breadth of potential diet, and disturbance abilities. The ability to feed at a high rate on a variety of both plant and animal foods, to move over large areas while foraging, and to persist despite low levels of food make the regular echi­ noids a versatile and important biotic component of the marine environment. Many aspects of their functional role in community ecology remain to be investigated. Among them is the investigation of their impact in other significant marine environments, such as sea-grass beds, subtidal communities, and polar seas, although the analysis of temperate shallow­ water and intertidal communities or of coral reefs has not been exhausted. An important aspect of future evaluations of the role of echinoids as determinants in community structure will of necessity revolve around an investigation of the mechanisms controlling echinoid population dynamics. This may be a key to the understanding of the problem, as it leads into the other aspect of their role which needs further investigation: the control of stability of the marine communities of which echinoids are an integral part.

25

BRUCE A. MENGE

EFFECTS OF FEEDING ON THE ENVIRONMENT: ASTEROIDEA

During the past 15 years, ecologists have devoted increasing attention to the analysis of the organization of natural communities. Some of the more important and revealing of these studies have been those performed in benthic marine communities. An interesting and undoubtedly significant feature of many such studies is that asteroid predators are both a conspicuous and functionally important component of the community. Though early zoologists and naturalists knew that most asteroids were predaceous and some sus­ pected that they performed a significant role in their communities, it was not until the mid­ 1960's that the potential importance of asteroid predation was experimentally demonstrated (Paine 1966). Since that time, increasing numbers of studies on the community role of asteroids have been published. Taken as a whole, the view that has emerged from these studies is that (1) predation plays an unexpectedly great role in organizing benthic marine communities, (2) asteroids are among the most important predators in these systems, and (3) despite the different evolutionary asteroid lineages characterizing different geographic regions, asteroid-dominated communities seem typified by a great degree of evolutionary and ecological convergence. In tbis paper I review tbis rapidly expanding body of literature from an ecological­ evolutionary viewpoint. A more general, recent feview of the feeding ecology of asteroids may be found in Sloan (l980b). My goals are: (l) to systematically and critically ex amine those studies wbich provide at least some insight into the role of asteroids in organizing their community, (2) to extract from these studies a view of the general role played by asteroids in bentbic marine communities, (3) to suggest the adaptive features of asteroids wbich are the foundation of their community role, (4) to consider the possible bounds to their ubiquitousness, (5) to briefly compare these studies to systems dominated by non­ asteroid predators and (6) to suggest potential areas of further research.

1. DEFINITIONS AND BIASES The definitions of key terms such as community, competition, stability, etc., as used here are presented in table 1 (updated from table 1 in Menge 1977). In tbis paper, I adhere to the bias that the best possible test of hypotheses concerning the causes of natural patterns is a carefully conceived and controlled field experiment (see e.g. ConneIl1974a,b, 1975). Observational and comparative studies can suggest such causes, and can eliminate some, but not all, alternative explanations. Of course, not all systems lend themselves to experimentation and some ecological characteristics usually cannot be manipulated (e.g. aspects of the physical environment such as wave shock). Moreover, 521

522 Bruce A.Menge experimentation without proper controls or a good knowledge of the basic natural history of a system can lead one astray. A useful observational method of obtaining some insight into the effects a predator has on its prey is particularly applicable to asteroids. This is the calculation of a prey removal rate (based on estimates of predator density, proportions of the population fee ding over time, diet composition, feeding rates, and predator activity) and comparing this to prey availability, or ideally productivity of available prey. This method has been used in several studies (e.g. Paine 1969a, Menge 1972b), but as Peterson & Bradley (1978) pointed out, it must be used with caution. Prey digestion times in asteroids may take widely varying amounts of time, depending on prey size. Hence, large prey will be overrepresented and small prey underrepresented in point-in-time diet sampies. Peterson & Bradley (1978) offered a means of correcting for such bias.

Table 1. Definitions of ecological terms Term

Definition

Community

Directly or indirectly interacting assemblage of species occupying a particular habitat. Includes all trophic levels Collective expression referring to the 'appearance' of a community; determined by quantifying distribution, abundance, body size, trophic relationships, and diversity of species in the community Collective expression referring to the mechanistic dynamics of a community; determined by evaluating the roles of predation, competition, other biotic interactions, disturbances, colonization, and temporal and spatial heterogeneity in determining commu­ nity structure Mutual striving by two (or more) species for a resource or resources (usually food or space) in short supply. Both species are nega­ tively affected Disruption of habitat or organisms caused by the non-trophic acti­ vities of organisms which usually leads directly or indirectly to mortality of the affected organisms Disruption of habitat or organisms caused by the physical environ­ ment which usually leads directly or indirectly to mortality of the affected organisms Variation in the physical environment. Two components are con­ stancy (probability that the environment will remain the same over time) and contingency (probability that the environment will vary in the same manner over time; follows Colwelll974) fhysical conditions of environment which approach or exceed the tolerance limits of an organism General term including the two immediately preceding definitions Deliberate, controlled alteration or perturbation of a biotic or physical characteristic with subsequent monitoring of the res­ ponse of the relevant populations. Controls include both unmani­ pulated populations and experiments testing for effects of any artificial devices used in the experiment other than the intended effect Assemblage of species in a community of comparable trophic status or which utilize similar food or spatial resources (Root 1967) Consumption of one species by another. Can refer specifically to carnivory or more generally to include all consumer-prey inter­ actions (see Lubchenco 1979 for discussion) Number of species in a community Structural irregularity in habitat space Irregularity over time in the physical and biotic environment

Community structure Community organization

Interspecific competition Biotic disturbance Physical disturbance Environmental predictability

Environmental or physical stress Environmental rigor Experiment

Guild Predation Species diversity (= richness) Spatial heterogeneity Temporal heterogeneity

E[[ects o[[eeding on the environment: Asteroidea 523

2. ROLE OF ASTEROIDS IN THE ROCKY INTERTIDAL REGION For several reasons (reviewed in Conneil 1972, Paine 1977), the structure and organization of rocky intertidal cornrnunities are better known than any other cornrnunity on earth. It is perhaps significant that asteroids are an important component of many of these commu­ nities. Unfortunately, the predatory role of asteroids is weil known in only a relatively small number of regions. As a result, this review will necessarily be weighted towards the Pacific and Atlantic shores of North America as points of reference. Because sirnilar factors may have different effects in different systems or even different parts of the same system (e.g. Lubchenco 1978, Menge 19:76, 1978a, Menge & Sutherland 1976), extrapolations of the interpretations of these studies must be done with care. 2.1. Temperate shores: Pacific coast o[ North America

In this and foilowing sections, I first describe relevant patterns of community structure (Le. distribution, abundance, size structure, species diversity, trophic structure) in a given system and where possible, follow this with a discussion of the actual or suggested effect of asteroids in the system. In hard substratum systems, spatial structure is largely a function of how sessile species use the surface of the rock, and how far they extend above the sur­ face. Thus 'primary' space is the rock surface itself, 'canopy' space is the space occupied by the taU (> 5 cm) macroalgae, and 'understory' space is space occupied by short (> 5 cm) macroalgae. Though mussei byssus and algal holdfasts may harbor rich assemblages of orga­ nisms (see Suchanek 1979), I restrict my attention to relatively large epifaunal organisms. 2.1.1. Community structure Space utilization patterns of the major macroscopic species in the rocky intertidal regions ofWashington state have been quantified by Dayton (1971, 1975a) and Paine (1974) and are briefly summarized as follows: on the outer coast, numericaUy dominant occupiers of the high intertidal are barnacles (Chthamalus dalli and Balanus glandula) and musseis (both Mytilus cali[ornianus and Medu/is (Suchanek 1978). In general musseis tend to be most abundant at the sites most exposed to wave shock (e.g. Tatoosh & Waadah Island, Dayton 1971, Suchanek 1978), and barnacles are more common at somewhat more protected sites (e.g. Portage Head, Dayton 1971). These species te nd not to occur in solid (about 90-100 % cover) zones. Rather , barnacles gene rally cover 25 % of the primary space. Musseis may cover more space but tend to be patchy in distribution (Dayton 1971). On the outer coast in the upper part of the mid intertidal, musseis and gooseneck barna­ eIes tend to form a nearly solid zone which is often characterized by holes or patches formed by bashing logs and wave action (Dayton 1971, Levin & Paine 1974, Paine 1977). These patches of free space tend to be initially colonized by algae (Porphyra, Ulva), some barnacles, Medulis, and various species of algae, often including the large sea palm Postelsia palmae[ormis (Dayton 1973a, Paine 1977, 1979, Suchanek 1978). A persistent and impor­ tant feature of most sites is that the mussei bed ends abruptly at its lower edge. Below this stage, primary space is occupied by a large variety of mobile and sessile invertebrates and algae (see Ricketts et al. 1968, Kozloff 1973, for accounts of these species and their natural history). Solid or near-solid bands of organisms are replaced by eIumps or patches of algae and sessile animals and much of the substratum is free of sessile organisms (55-90 %; Day­ ton 1971) and available for colonization (Le. is 'free space'). This sharp lower edge of M

524 Bruce A.Menge cali[ornianus has been shown to be a remarkably constant feature of outer co ast commu­ nities, with virtually no changes observed in localities monitored for over a decade (Paine 1974). Organisms commonly occurring below the mussei zone include small numbers of most of those species found in the higher intertidal plus up to four species of anemones, four species of limpets, up to seven species of chiton, at least three species of herbivorous gastro­ pods, three to five species of asteroids, two to three species of predaceous gastropod, one to two species of grapsid crab, one to two species of cancrid crab, and a wide variety of algal genera and species and surfgrasses (e.g. brown algae: Hedophyllum, Fucus, Laminaria, Lessoniopsis, and Alaria; red algae: Endocladia, lridaea, Polysiphonia, Corallina, Gigartina, Odonthalia, Rhodomela, and Lithothamnium; green algae: Enteromorpha, Spongomorpha, Gadophora, Rhizoclonium, Ulva, Monostroma; surfgrasses: Phyllospadix spp.). This listing is incomplete but gives a notion of the species richness of the lower half of outer co ast rocky shore communities in the Pacific Northwest (see Dayton 1975a). A significant feature of lower shore areas is the presence of occasional very large M. cali[ornianus (Paine 197 6a). The major differences in community structure between the outer coast and the more protected shores of Puget Sound and regions north are (1) the near absence of M cali[or­ nianus and its replacement by Balanus cariosus as the zone forming species in the higher rnid intertidal, (2) a general reduction in canopy cover and abundance of ephemeral species, (3) increased densities of some herbivorous molluscan species, especially limpets, and (4) a greater [raction of unoccupied space at all tidallevels (e.g. Dayton 1971, 1975a). In summary, the salient features of community structure in this system are (1) zones of barnacles (high intertidal), musseis (upper mid intertidal), and algae (mid and low inter­ tidal), (2) a sharp and near constant lower edge of the mussei zone, (3) a relatively low species richness above and a high richness below this edge, and, correlated with this, (4) relatively small and large availabilities of unoccupied space above and below this edge, res­ pectively.

2.1.2. Trophic structure The trophic organization of this system has received considerable attention all along the west coast of North America. The two most abundant asteroids, the large Pisaster ochra­ ceus (mean wet weight ranges from 150 to 2450 g; Menge 1972b, Paine 1976a,b) and the small Leptasterias hexactis (mean wet weight ranges from 3 to 8 g; Menge 1972b) are the top benthic predators in this system. Though both species are dietary generalists, they, combined with the predaceous gastropods (Thais spp., Searlesia; Connell 1970, Dayton 1971, Louda 1979, Ricketts et al. 1968), prey most heavily on musseis and barnacles. Pre­ ferred prey of P.ochraceus are musseis (Mytilus cali[ornianus andMedulis; Paine 1969a, Landenberger 1968), and of L.hexactis are small gastropods (Littorina scutulata, L.sitkana, Lacuna variegata, Tonicella lineata, Cyanoplax dentiens; Menge 1972a) and chitons. More generally , musseis and barnacles are the most important numerical components of the diets of these asteroids throughout the west coast (Mauzey 1966, Mauzey et al. 1968, Menge 1972a,b, Menge & Menge 1974, Paine 1966, Dayton 1973b, Feder 1969, 1970).

2.1.3. Community organization Paine (1966) noted the sharp lower edge of mussei beds, the predilection of Pisaster ochra­ ceus for musseis and barnacles, and the conjunction of the upper limit of the foraging range of this asteroid with the lower edge of the mussei bed. He hypothesized that P.ochra­

Effects of feeding on the environment: Asteroidea 525 ceus determined this lower limit and tested his hypothesis by removal of this asteroid from portions of the shoreline at two sites (Makah Bay, Paine 1966; and Tatoosh !sland, Paine 1974). Comparisons of events at the manipulated sites to undisturbed control sites a few meters away showed thatMytilus californianus becomes the dominant space occupant within 2-3 yr. This occurs either by settlement, survival and growth of musseis below the normal limit (Paine 1966) or by a glacier-like 'cascade' of adult musseis downward from the usual high-mid intertidal band (Paine 1974). In either case, the musseis outcompete most other sessile speeies (e.g. barnac1es, anemones, algae), eliminate the habitats of many motile speeies (e.g. some limpets, chitons and other gastropods), and thus reduce speeies richness sharply (e.g. from 15 to 1 speeies; Paine 1974). Paine conc1uded from this study that P.ochraceus was a 'keystone' predator (1969b), Le. a predator which is not the most abundant speeies in the system but one which has an inordinately great effect on commu­ nity structure relative to the roles of other speeies in the community. Paine further sug­ gested that this phenomenon might be of widespread ecological significance (Paine 1966, 1977). The studies of Dayton (1971, 1973b) extended and strengthened this conc1usion. Day­ ton was able to establish that other invertebrate consumers had measurable and significant effects on patterns of community structure. However, he conc1uded from his elegant and now c1assic study (1971) that P.ochraceus was ultimately the most important consumer in the system. Several important prey speeies, mostly barnac1es and musseis, were able to escape gastropod predators or biological disturbers (e.g.limpet 'bulldozing' effects) in time, space or size. None of these could escape the large P.ochraceus with the exception of M californianus. Individuals of this mussei occasionally escape predation by P.ochraceus long enough to reach a size too large to be captured by this asteroid (Paine 197 6a). A noteworthy feature of this rocky intertidal community is its apparently high degree of persistence over time. Paine (1974) documented a remarkable level of constancy in the lower limit of Mcalifornianus. The abundances and sizes of P.ochraceus (Menge 1974, Paine 1976a) and Leptasterias hexactis (Menge 1974) also seem relatively constant over time. In part, these apparent low levels of variability are due to the intense biotic interac­ tions which occur in the community, particularly those centering around the asteroids. As noted above, predation by P.ochraceus appears primarily responsible for several persistent features of prey populations (e.g.lower limit of Mcalifornianus or Balanus cariosils, low abundance of barnac1es, musseis and other prey in the lower half of the intertidal zone). Another interaction potentially important to the stability of this community is competition between P.ochraceus and L.hexactis (Menge 1972b, 1974, Menge & Menge 1974). First, the diets of both asteroids are very similar. Further, the biomass of these two asteroid spe­ eies are inversely correlated from site to site. One consequence of this interaction is that when the abundance of P.ochraceus is low, L.hexactis becomes larger and more abundant and expands its diet to inc1ude larger prey individuals and speeies of larger mean body size (Menge 1972b). This suggests that the response of prey populations to sudden reductions in abundance of P.ochraceus would be damped by a compensatory increase in L.hexactis (and perhaps other, non-asteroid predators) abundance and/or size. Similarly, Paine (1976a) suggested that L.hexactis stabilizes mussei beds by preying on the small musseis which recruit to the byssal 'forest' beneath a mature mussei bed. Such predation may reduce the likelihood that the attachment of the larger musseis in the matrix of the mussei bed will be weakened by crowding from below by the growing small musseis. In the low intertidal region Pycnopodia helianthoides occurs with P.ochraceus and L.

526 Bruce A. Menge hexactis. This large asteroid (up to 1 m in diameter, Mauzey et al. 1968) preys on echinoids (e.g. Mauzey et al. 1968, Paul & Feder 1975), and has an important indirect effect on com­ munity structure in this region (Dayton 1975a). The echinoidStrongylocentrotus purpura­ tus has an important effect on species composition and patchiness of algae (Paine & Vadas 1969; see Lawrence & Sammarco, chapter 23). The effect of P.helianthoides is primarily to clear 1-2 m 2 areas of echinoids by either stampeding or eating them (Dayton 1973b, 1975a). This represents areal decrease in echinoid density and not a relocation since most succumb as prey of the large solitary anemone, Anthopleura xanthogrammica, or are washed away by waves (Dayton 1973b). These cleared areas are thus available for colonization by algae. As P'helianthoides move in and out of this region (these asteroids are primarily subtidal), a series of asynchronous patches are produced which thus appear to account for some of the observed patchiness and diversity observed (Dayton 1975a). 2.1.4. Defensive responses A large body of more indirect evidence exists which suggests many asteroids on the Pacific coast of North America are also important predators on numerous species of motile orga­ nisms. Many species of limpets, abalone, echinoids, holothuroids, gastropods and bivalves exhibit 'running or defensive responses' involving a sudden acceleration in their rate of movement in response to contact with, or at least detection of, certain asteroids (including Pisaster ochraceus, Leptasterias hexactis, and Pycnopodia helianthoides e.g. Feder 1963, 1972, Margolin 1964, Montgomery 1967, Mauzey et al. 1968, Menge 1972a, Phillips 1975). This change in behavior is usually quite conspicuous and may involve other changes, such as mushroorning, extension of mantle tentacles, twisting, jumping and even swimming. Another type of behavior involves covering the shell with the mantle (Diodora aspera; Mar­ golin 1964). Although the role of these behaviors was once unknown, mainly because some of the asteroids were not observed to prey on species with defensive responses, most inves­ tigators now interpret them as true defenses against potentially severe predation by aste­ roids (e.g. Margolin 1964, Mauzey et aL 1968, Menge 1972a, Feder 1972). That these res­ ponses are so widespread among the invertebrate fauna of the Pacific coast suggests that asteroid predation has been an intense and chronic selective force in this system over evo­ lutionary time. More specifically, several mobile prey species seem to coexist with P.ochraceus at least partly as a consequence of their running or defensive response. Primary among these are limpets (D.aspera, Notoacmaea scutum, Collisella pelta) and coiled gastropods (Tegula funebralis, Calliostoma ligatum). Although these species are occasionally eaten by P.ochra­ ceus and L.hexactis (sometimes to a significant extent; see Paine 1969a, Menge 1972a,b), they are often abundant or large, or both, despite the presence of their predators. These species thus seem to have effectively escaped control by their predators, at least to the extent that they can coexist with them and maintain breeding populations. Other forms of escapes from predation in this system include spatial refuges by barnacles and musseis (above the foraging range of the asteroids) and size refuges of M.californianus. Note that the size and defensive escapes allow prey coexistence with the predator while the spatial refuges of upper intertidal barnacle and mussei zones do not. I shall term the former as types of 'coexistence' escapes or refuges and the Iatter as 'non-coexistence' escapes or refuges. In summary, these studies of asteroid predators in the intertidal region of the Pacific coast of North America provide a strong and convincing case for the importance of asteroids

Effects of feeding on the environment: Asteroidea 527 as perhaps the dominant structuring force in this system. Patterns of diversity, abundance, distribution, size and behavior of the numerically dominant sessile and motile biota are regulated in large measure by the predatory activities of asteroids. In light of this, among the questions to be considered below are (1) do asteroids have comparable effects in other intertidal communities and (2) are asteroid species 'keystone' predators in such systems?

2.2. Temperate shores: Atlantic coast of North America 2.2.1. Community structure On the Atlantic shores of North America, rocky intertidal habitats occur primarily north

of Cape Cod. South of this cape, sandy or muddy shores prevail. North of this cape to the northern shore of Cape Breton !sland, the rocky intertidal region is like that on the Pacific coast in several ways. A high zone of barnacles (Balanus balanoides; 25-95 % cover) is sharply delineated from either a lower mussei zone (Mytilus edulis; 50-90 % cover; exposed shores) or a fucoid-dominated zone (Fucusdistichus, F.vesiculosus or Ascophyllum nodo­ sum; 30-95 % canopy cover; moderate to protected shores: Menge 1976). In the 10w zone at exposed shores, Medulis covers primary space (20-95 % cover) under a sparse canopy of the kelp Alaria esculenta. On moderate to protected shores, the low zone is dominated by the shrubby red alga, Chondrus crispus (Lubchenco & Menge 1978). The boundaries between these zones are sharply defmed as on the Pacific coast. However, species richness of both algae and animals is considerably lower on the east coast (Menge 1976, Lubchenco & Menge 1978). For example, if occasional subtidal 'intruders' into the intertidal are excluded, there is only one abundant intertidal acorn barnacle in New England (vs three on the west coast plus a gooseneck barnacle, Pollicipes polymerus), one common mussei (vs two), one limpet (vs four to six), and one whelk (vs four). Several west coast taxa either have no east coast analogue (e.g. there are no gooseneck barnacles or hermit crabs in the east coast rocky intertidal) or the analogues are very rare and primarily subtidal species (e.g. anemones, chitons). In contrast, there are similar numbers of herbivorous coiled gastropods (- 5) and asteroid species (3-5) in the intertidal region of each coast. In the latter case, however, the asteroids occur primarily in the low zone, whereas on the west coast, at least two of the asteroid species (Pisaster ochraceus, Leptasterias hexactis) extend weil into the mid zone (Menge 1976, Menge & Sutherland 1976, Lubchenco & Menge 1978).

2.2.2. Community organization Each summer and early autumn (July-October), the asteroids Asterias forbesi and A. vulga­ ris invade the lower intertidal region from subtidal regions in New England. Experimental studies (Lubchenco & Menge 1978, Menge unpublished) indicate that these species have a major impact on community structure of protected to moderately exposed shores. How­ ever, on exposed shores, neither asteroids or other predators influence community struc­ ture. Here musseis outcompete the alga Chondrus crispus and dominate the shore because they have escaped predation. This is a non-coexistence escape, since the asteroids are absent or nearly so, at exposed sites. Our experiments indicate that musseis would outcompete C crispus elsewhere, but for the predatory activities of the often dense populations of Asterias spp., Thais lapillus, and crabs. Even at protected and intermediate areas where predators are common, musseis occasionally escape in size (a coexistence escape). I have observed a few clumps or individuals of Mytilus edulis in the low zone of Grindstone Neck, Maine which seemed to have achieved immunity to predation by Asterias spp. because of their large size (- 10 cm in length). These individuals persisted at least through the years 1972­

528 Bruee A.Menge 1976, despite the frequent presence of dense aggregations of A.vulgaris on and around them. These asteroids were eating smaller Medulis. The large periwinkle Littorina littorea (up to ca. 3 cm long), mayaiso escape in size from Asterias spp. We have seen predation by these asteroids on this littorine only rarely (0.6 % of the diet at Grindstone Neck; Lubchenco & Menge 1978), despite a high frequency of contact between these abundant species. Another type of coexistence escape in the low zone is a local or patchy escape in space by musseis. Local patches of musseis develop in the low zone and persist through the warmer half of the year when Asterias spp. actively feed. These patches may cover 10-100 m 2 and have been observed at all relatively protected sites studied in New England. By persisting longer than one year, the musseis in these patches can reach reproductive maturity (Menge, unpublished). Eventually, these patches are eliminated by aggregations of predators or storms. We interpret such escapes as resulting from 'patchiness' or local variation in preda­ tion intensity (Lubchenco & Menge 1978). The apparent lack of behavioral escapes bylittoral prey species in this system is interest­ ing, and suggests that here predation intensity on mobile species is less chronic, or severe, or both than in the intertidal zone of the Pacific Northwest. In contrast to the west coast asteroids, which reside in the mid and low intertidal region, Asterias spp. do not range above the low zone (ca. + 2 ft or 0.71 m) and are not normally intertidal residents. This is puzzling, since Medulis, probably the preferred prey of Asterias spp. (Menge 1979, Lubchenco & Menge 1978), is often abundant in both low and mid zones while it is usually scarce in the subtidal region (Menge 1979). Moreover, other prey species such as barnacles are gene rally more abundant intertidally than subtidally (Menge 1976,1979, Lubchenco & Menge 1978). Though several hypotheses could be offered to explain this failure of Asterias spp. to invade the mid-intertidal region of New England, only one can be approached with the evi­ dence available. This hypothesis, that predation by seagulls keeps asteroids out of the mid zone, is suggested by Verbeek's (1977) observations of predation by seagulls (Larus argen­ tatus) on A. rubens (probably =A. vulgaris) in Britain. Two arguments mitigate against this hypothesis. First, dense populations of Asterias spp. occur in the low zone in New England for periods of about three months each summer. Although gulls (L.argentatus and L.mari­ nus) occasionally prey on asteroids, the asteroids are not eliminated from the low zone (personal observation). Second, asteroids reside in the intertidal zone on the west coast and are exposed to predators. The gaudily colored Pisaster oehraeeus is occasionally eaten by seagulls (L.glaueeseens; C.Marsh, personal communication), but clearly is not excluded from this region by seagull predation. On these bases, I conclude that seagull predation by itself cannot explain the absence of Asterias spp. from the intertidal zone in New England. In summary, asteroids seem to be an important structuring agent in the low, but not the mid, intertidal region of the rocky shores of New England. Observations along the shore of Nova Scotia (Lubchenco & Menge, personal observation) and the experiments of Peterson (1979) in New Jersey, suggest that the studies in New England are probably generally appli­ cable at least south of wintertime pack ice.

2.3. Temperate shores: Atlantie eoast 01 Europe 2.3.1. Community strueture The rocky shores of Europe evidently exhibit many similarities to those in the Northwest

E[[ects o[[eeding on the environment: Asteroidea 529

Atlantic (e.g. see Lewis 1964, 1976, 1977, Stephenson & Stephenson 1972, Hatton 1938). For example, the zonation patterns and species composition of European and Northwest Atlantic intertidal shores are quite similar (compare Lewis 1964 to Menge 1976, Lubchenco & Menge 1978). Perhaps the most noteworthy difference is that the European shores har­ bor a more diverse community. Thus, among the more common species, which are typically intertidal residents, there are more species of limpets (3 vs 1), barnacles (up to 5 vs 1 or 2), herbivorous gastropods (6-8 vs 3-5), anemones (1 vs 0), and algae (twice as many species) in European vs New England shores (approximate members; extracted from Lewis 1964, South 1976, Parke & Dixon 1976). Unfortunately, I know of no studies on the predatory effects of intertidal asteroids on European shores. Apparently, most asteroids in this region are largely subtidal. However, comments by Lewis (1964,1976, 1977) suggest that Asterias rubens, like A.vulgaris (= rubens?) and A,forbesi in New England, invades the low intertidal region seasonally (though in winter rather than summer). Lewis (1964, p.295, 1976, 1977) commented that these invading asteroids can devastate the musseis or barnacles in the low intertidal zone, though no data are presented. In addition, Connell (1961) mentioned that A.rubens entered one of his low zone predator exclusion cages and killed most of the barnacles therein. Though no one has pursued this hint of an important asteroid effect in this region, it seems likely that low intertidal asteroids play comparable roles in the temperate shores of Europe and New England. Asteroids are evidently not found in the mid intertidal region of Euro­ pean shores (e.g. Lewis 1964), though the explanation of this absence is unclear. 2.4. Temperate shores: Pacific coast o[ Asia (Japan) 2.4.1. Community structure and organization

Patterns of community structure and organization for temperate Asian shores are poorly known. Since to my knowledge, the only data available are for the shores of Japan, I will summarize this information as at least partly representative of Asian rocky intertidal com­ munities. Descriptions of zonation patterns, partial species composition, and aspects of the dynarnics of the shores in the vicinity of Asamushi, Japan are described by Hoshiai (1964) and Hoshiai et al. (1965). Seven zones of sessile biota are recognized, including the typical temperate zones ofbarnacles, musseis (two species in two zones), and algae (three separate zones). In addition, a serpulid zone is also evident (Pomatoleios and Hydroides). Though only pictures and numerical densities of sessile animals are given, the amount of free space available appears relatively low in this region (Hoshiai 1964, Hoshiai et al. 1965). Diversity of macroscopic species also seems relatively high (e.g. five barnacle species, four limpet species, one littorine species, five other herbivorous gastropod species, two mussei species, and four predaceous gastropods). Hoshiai (1964) noted that two asteroids, Asterias amuren­ sis and Aphelasterias japonica, occur in the lowest intertidal zone. Luckens (1970) reported that Asterina pectini[era also occurs in this region and that the former two asteroids actually range as high as the lower edge of the belt of Chthamalus challengeri, which puts them into the high rnid intertidal (Hoshiai 1958). Despite the fact thatA.amurensis andA.japonica are both generalized carnivores, Luckens (1970a,b) argued that they are scarce and have no significant impact on community structure in this region. The apparent high cover of pri­ mary space in the lower intertidal shore (e.g. see figs.2 and 3 in Hoshiai et al. 1965, fig.9 in Hoshiai 1964) supports this claim, but better data and field experimentation are needed.

530 Bruce A.Menge In summary, the limited geographic extent of the studies of Hoshiai & Luckens (ie. Asamushi, at the northern tip of Honshu, the main Japanese island), and the poor quantifi­ cation of predator effects make even a tentative conclusion on the role of asteroids in the organization of this community unwise.

2.5. Temperate south Pacific shores 2.5.1. New Zealand Relatively little work has been done on the community role of asteroids in the southern hemisphere. The one published experimental study available (Paine 1971) reveals some remarkable paralleis in asteroid effects between New Zealand shores and northern hemis­ phere temperate shores.

2.5.2. Community structure On wave swept rocky shores in New Zealand, the general patterns of community structure are very similar to those described in detail above for other temperate shores. There is a high barnacle zone (Chamaesipho spp.), amiddie mussei zone (perna canaIiculus), and a lower algal zone (Durvillea antarctica: Paine 1971, Morton & Miller 1968). Other typical organisms include limpets, chitons, predaceous gastropods and such alga genera as Corallina, Litho­ thamnium and Gigartina. Finally, wave swept shores have intertidal populations of the large, many-armed asteroid Stichaster austraIis (Paine 1971, Barker 1977). This asteroid preys primarily on musseIs and barnacles. To ascertain the role of S.australis in this community, Paine (1971) performed experi­ ments parallel to those done in the northeast shores of the Pacific. The response to asteroid removal was as predicted. The lower limit of P.canaliculus at experimental areas very quickly extended down ward. Within 15 months, the musseIs dominated primary space in the middle half of the shore, ranging down to the upper limit of D.antarctica. When both S.australis and D.antarctica were removed, P.canaliculus covered the entire lower two-thirds of the intertidal region; in neither case did similar changes occur in controls. A second response to the manipulation was a decline in species richness like that seen in north temperate experiments (from 20 to 14 species in nine months, Paine 1971). It is important to note that as in other such experiments, the response to manipulation of S. australis occurred even though other mussei predators (gastropods;Neothais scalaris) occurred in the exclusion areas. A description of the shores of the subantarctic Macquarie Island (54°36'S latitude, 158° 45'E longitude; Simpson 1976) suggests asteroids may be important there. The rocky inter­ tidal region is only 0.71 m wide and, though zones occur, barnacIes and musseis seem absent. From top to bottom, the zones are (1) lichens, (2)Porphyra sp., (3) 'bare' and (4) 'upper red and kelp' (D.antarctica). The dominant animals include limpets, chitons, litto­ rines, trochid (herbivorous) gastropods, and, in the low zone and subtidal, three asteroids (Anasterias directa, A.mawsoni, and Asterina hamiItoni). Simpson indicates that a limpet (Patinigera macquariensis) and the trochid Cantharidus coruscans both exhibit escape res­ ponses to the Anasterias spp. However, the role these asteroids play in this system is not clear. Simpson (1976) feIt that they were minor predators but devoted only a small part of his research effort toward this problem. Chile: The community structure of the temperate Chile an coast is much like that noted for other regions (Paine, personal communication). From high to low intertidallevels, there

Ellects olleeding on the environment: Asteroidea 531 occurs a barnacle zone (two species of Chthamalus); a mussei zone (Perumytilus purpura­ tus); a diverse zone with many species of limpets, chitons, coiled gastropods, barnacles, musseis, and algae; and an alga zone (the large Lessonia nigrescens, Durvillea antarctica, and a prominent band of encrusting coralline algae). Two species of echinoids (Tetrapygus niger and Loxechinus albus) occur in the algal zone. Asteroids occurring in this community are Heliaster helianthus, Stich aster striatus, Meyenaster gelatinosus (low zone only) and

Patiria chilensis.

The most abundant of these asteroids is H.helianthus (Viviani 1975, Edding 1978).lts diet is broad (31 prey species) but it preys most heavily on barnacles, musseis and limpets (Edding 1978). Such habits, its size (- 120 g wet weight; Edding 1978) and its abundance suggest that it may playa role similar to that observed with Pisaster ochraceus and S.aus­ traUs. In fact, removal of H.helianthus results in downward extension of the lower limit of P.purpuratus and areduction in diversity (R.T .Paine, J .C.Castilla and J .Cancino, personal communication). This result occurred within two years of initiation of the experiment and is exactly parallel to Paine's (1966, 1971, 1974) previous experiments in other regions. Note again that the strong response observed occurred whenH.helianthus alone was removed. Other intertidal predators were not manipulated and thus seem to have subordi­ nate effects to H.helianthus. In summary, studies in two southern temperate intertidal communities repeat the basic theme which has emerged from studies in northern temperate regions. Asteroid species in New Zealand and Chilean rocky intertidal communities appear to be the dominant carni­ vore, and through their foraging activities, are evidently responsible for several dominant features of community structure.

2.6. Tropical shores: Pacific coast 01 Panama The Pacific coast of Panama (Bay of Panama) is characterized by tides of 7.1 m. Intertidal community structure here is strikingly different from most temperate shores (Menge & Lubchenco 1981). Briefly, (1) zonation is much more diffuse and visually less apparent, (2) space occupancy by sessile biota at all tidallevels is exceedingly low (1-5 % cover), (3) erect macroalgae are all but absent, and (4) species richness of animals at all trophic levels is high. Moreover, a forcipulate asteroid, Heliaster microbrachius, occupies a relatively high trophic position in the food web, though it is probably not the top predator (Menge & Lubchenco 1981). Though some gastropod species have running responses to this asteroid, H. kubinjii is rare, of relatively small size « 100 g), and restricted in habitat. All individuals observed have been found in or near deep crevices and never in the open as are asteroids in most other systems. I tentatively conclude that asteroids playa minor role in this tropical intertidal community.

3. ROLE OF ASTEROIDS IN ROCKY SUBTIDAL REGIONS As many authors have noted, the subtidal region has several major physical and biological differences from the intertidal region. Physically, the subtidal is gene rally less variable than the intertidal in temperature, salinity, water motion, and light intensity. Biologically, diversity is usually greater. However, zonation patterns occur in many subtidal regions (see below) and asteroids are often a conspicuous component of subtidal communities. A major

532 Bruce A.Menge hindrance encountered in studies of subtidal communities is the great increase in effort required per unit of information gleaned. Both quantitative observational and experimental research are harder and oftentimes unrealistic in subtidal systems. However, with the advent of scuba and underwater 'habitats' as scientific tools, we have gained vastly in our under­ standing of natural subtidal communities, though the understanding is stilliargely based on observational, rather than experimental information. 3.1. Polar subtidal habitats: Antarctica

The subtidal fauna of the Antarctic includes at least 18 species of asteroids (Dearborn 1977). Though the diets of many of these, as represented by dredged specimens, are given by Dearborn, I know of only one study in this region which has attempted to evaluate the impact of asteroid predators in a specific benthic community (Dayton et al. 1974). 3.1.1. Community structure Dayton et al. (1974) have provided considerable insight into the structure and organization of a subtidal (33-60 m) benthic community in McMurdo Sound, Antarctica. The spatial structure of this system is dominated by sponges (ca. 18 spp; total primary cover of 55.5 %). Bryozoans, actinians, hydroids and other sessile biota cover 5.4 % of the primary space and the remaining space (39.1 %) is 'free' space (actually a mat of sponge spicules). The sponges can reach gigantic proportions (e.g. 2 m high, 1.5 m diameter). Mobile species associated with these sessile biota include nudibranchs, pycnogonids, gastropods, echinoids , nemer­ teans , amphipods and at least seven asteroid species. Of these, five species of asteroids are common and prey on sponges (fig.1). Attempts to directly evaluate the dynamics of this system initially proved largely unsuc­ cessful (Dayton et al. 1974). Extensive exclosure and enclosure experiments performed over 12 months revealed that rates of sponge growth and asteroid and nudibranch fee ding

Figure 1. Food web of the Antarctic subtidal sponge-dominated community. Numbers in parentheses are mean percent cover. Percentages of so me items in the diet are given by each line. Arrows point to the predator. Data taken from Dayton et al. 1974.

E//ects 0/ /eeding on the environment: Asteroidea 533

are extremely slow in this stable but very cold environment (x temperature at 75 m = -1.87 ± 0.11 (SD) oe). In many cases, feeding nudibranchs or asteroids were often in exactly the same position after one year. With one exception, sponge growth was so slow that it could not be measured (even in the absence of predators). The exception was Mycale acerata, some individuals of which exhibited an increase in biomass of 10 to 67 % over one year. As a result, Dayton et al. resorted to indirect means of measuring the impact of asteroid predation. Predator ingestion rates, diet composition, growth, respiration rates and predator and prey abundances were estimated from field and laboratory measurements. These, combined with literature-based estimates of assimilation and energy allocation to reproduction and growth, permitted the derivation of estimates of the potential impact of each predator on sponge abundance. The general conclusion reached using these methods was that most sponges had 'escaped' their predators in size. That is, at the observed rates of growth, feeding, etc., the available sponge biomass could supply the predators with food for up to 1400 years depending on the sponge species. Dayton (1979) presented data and observations from this system which cover a 10 year period (1967-77). These studies expand on several points. First, most sponge species did not grow over the 10 year period, supporting the earlier notion of extremely slow rates of growth. Second, Macerata remained the major exception to this latter trend by growing relatively rapidly at least for 9 years. However, this sponge appears to live only about 10­ 20 years, whereupon it dies for unknown reasons. Another species, Rossella racovitzae also grew significantly over this time interval. Third, several individuals of two large sponge species (Rossella nuda and Scolymastra joubini), many of which were thought to be hund­ reds of years old, died in the 10 year period. Mortality causes were not clear but appeared to be predation by Acodontaster conspicuus. Dayton (1979) suggested that these sponges cannot recover if they lose> 20-30 % of their volume to predatory attack. Thus, the esti­ mates of standing crop may be too high, though the basic suggestion, that available food could support the predators for many years, remains. Hence, Dayton and his coworkers suggest that two predators, the asteroids A.conspicuus and the nudibranch Austrodoris mcmurdoensis could rapidly eliminate sponge biomass if they were more dense. Why then are not these species denser? Figure 1 indicates that both adults and juveniles of the large A. conspicuus (wet weight ranges up to ca. 800 g) are preyed upon by the small but common dietary generalist Odontaster validus. Thus, Dayton et al. (1974) suggested that O. validus both controls adult A.conspicuus populations and provides a recruitment 'bottleneck' by feeding on detritus and therefore possibly on settled larvae. Actual or potential controlling factors on other predators are not as yet understood. They concluded that (1) the abundance of Macerata, the potentially dominant competitor, is controlled by asteroid predation. (2) Individuals of other sponges escape such control and grow large because of predation by O. validus on larvae and adult A.conspicuus also kills most sponge recruits and maintains a rather high level of free space in the system. However, this hypothesis would be difficult to test. The refuge in size achieved by these sponges is clearly a type of coexistence escape. 3.2. Temperate subtidal habitats 3.2.1. Pacific northeast The subtidal asteroid fauna of the northeast shores of the Pacific is highly diverse (e .g.

534 Bruce A.Menge

Mauzey et al. 1968, Carey 1972). Although the diets of many of these species are known to a greater or lesser extent (e.g. Mauzey et al. 1968, Carey 1972, Rosenthal1971, Rosen­ thai & Chess 1972), only two efforts have been made in this region to estimate the impact of subtidal asteroids on their prey (Birke land 1974, Birkeland et al. 1980). Birkeland (1974) focused on a key structural element of subtidal sandy habitats in Puget Sound, Washington, the pennatulacean anthozoan or sea pen Ptilosarcus gurneyi. P.gurneyi forms a zone or band ranging in depth from 10 to 25 m throughout most of Puget Sound. The animals are solitary colonies which lodge themselves in sand with a fleshy peduncle and extend the feather-like fee ding portion of the colony up into the water column. They are stiffened by an intern al style which displays annual growth rings. Birkeland (1974) found that P.gurneyi matures at 5 years and reaches at least 14 years of age. Densities of these large anthozoans average about 201m2 (Birkeland 1974). Extensive observations offeeding asteroids (and nudibranchs) revealed that this single species of sea pen is food for seven predator species. Of these, four are asteroids. Of these, one is a strict sea pen specialist (Hippasteria spinosa), one specializes on sea pens in this habitat but preys on other species elsewhere (Dermasterias imbricata), and two are genera­ lis~s eating other prey in addition to sea pens (Mediaster aequalis and Crossaster papposus). All of the asteroids in this food web except H.spinosa are preyed upon by a top carnivore asteroid, Solaster dawsoni (Birkeland 1974). As in many subtidal systems, field experimentation proved unfeasible, However, Birke­ land obtained rough estimates of the impact of these asteroids on populations of P.gurneyi by combining predator and prey densities and sizes, predator feeding rates, and prey growth and recruitment rates. These calculations suggested that P.gurneyi is not overexploited for several reasons. First, P.gurneyi has unpredictable recruitment in space and time, which Birkeland suggested impairs the recruitment of the sea pen specialist, H.spinosa. Second, this unpredictable sea pen recruitment is apparently characterized by local saturation of patches. This 'floods' the small predators (nudibranchs and juvenile asteroids) and many recruits escape small predators by growing too large to be consumed. Third, by contracting completely into the sand at unpredictable times, adult sea pens reduce the probability of their being encountered by a predator. Fourth, the abundance of generalist asteroids is pro­ bably controlled by S.dawsoni. Hence, these predators do not reach the high densities which would lead to overexploitation (Birkeland 1974). The specializedH.spinosa somehow avoids S.dawsoni predation, perhaps via the large pedicellariae on its aboral surface. Another study, in the San Juan Islands, Washington, appears quite similar to the case of P.gurneyi. Many subtidal habitats in the San Juan Islands have high densities of a small epi­ benthic holothurian Cucumaria lubrica (Birkeland et al. 1980). Like P.gurneyi, this species persists in the face of predation by seven species of asteroids. Calculations suggest that only 3 % of the Clubrica population is removed by predation each year. As in the prior study, S.dawsoni pteys on most of the predators of C lubrica. Calculations suggest that S. dawsoni prevents these asteroids from building populations dense enough to eliminate C lubrica. The two major predators of Clubrica (Solaster stimpsoni and D.imbricata) are sug­ gested to avoid elimination by S.dawsoni by escapes in behavior (S.stimpsoni), size (both asteroids), and space (vertical walls, S.stimpsoni; S.dawsoni is suggested to fall off walls in its efforts to capture S.stimpsoni, which it does by leaping onto the prey's aboral surface). The factor preventing dense populations of escaped S.stimpsoni and D.imbricata is not clear, but Birkeland et al. (1980) suggested that a parasitic green alga is responsible for mortality of large, escaped asteroids.

Effects of feeding on the environment: Asteroidea 535 3.2.2. New England No published studies are available on subtidal habitats in either the northwest Atlantic (New England and the maritime provinces) or the northeast Atlantic (northern Europe) in which the effect of asteroid predators is convincingly demonstrated. However, some information is available which suggests that asteroids have a major impact on the structure of some sub­ tidal communities in both regions. In New England, subtidal populations of the generalized predators Asterias forbesi and A. vulgaris range in depth from the low intertidal region to 50 m (e.g. Miner 1950, Menge 1979). The densest populations appear to cluster around mussei beds (or oyster beds further south; Loosanoff 1964). These asteroids occur in all types of substratum ranging from solid rock to mud (e.g. Menge 1979). Since they appeared to be major predators in the low inter­ tidal zone (Lubchenco & Menge 1978), they also may have an impact on subtidal commu­ nities. To test this hypothesis, I initiated an Asterias spp. removal experiment in June 1973, using the granite piers supporting the Long Island Bridge in Boston Harbor, Massachusetts (Menge 1979). The subtidal sessile epifauna on these piers consists of musseis, barnacles, sponges, anemones, colonial tunicates, hydroids, ectoprocts, and algae. Musseis are the most abundant sessile species on both the vertical walls (e.g. 26 % cover in June 1975) and horizontalledges (58 % cover in June 1975) of the piers. Besides Asterias spp., preda­ tors include crabs and lobsters. These removal areas were monitored monthly in summer but infrequently during the rest of the year. The asteroid populations built up rapidly after each removal, and predation was probably only partially reduced. Thus, although mussei cover on horizontal substrata (ledges) was greater on the experimental pier than on the control pier after two years (77 % vs 58 %; n = 10 quadrats 0.25 m2 in area), this difference is not statistically significant (1­ way ANOVA). Interestingly, mussei cover on ledges of both experimental and control piers increased significantly (24-77 % on the experimental; 25-58 % on the control pier). One explanation is that removals from the experimental pier led to a general dilution of asteroid abundance in the vicinity, including the control pier (the piers are about 20 m apart). If so, then both piers experienced removal of asteroids, though the effect was presumably greater on the experimental pier. Though this study needs to be repeated, these results suggest that Asterias spp. may affect community structure in the shallow subtidal region. Since Mytilus edulis can outcom­ pete other species when released from predation in the intertidal (Lubchenco & Menge 1978), there is no reason to suspect a different outcome in the subtidal. In support of this argument, Paine (1976b) suggested that certain subtidal populations of Mcalifornianus have indeed escaped their predators (a coexistence escape), and in these habitats (shallow subtidal pinnacles) are dominant space occupants. If M.edulis are dominant space competi­ tors in the subtidal, then Asterias spp. are likely to be important in preventing musseis from exerting this dorninance. However, the situation is unquestionably more complex than this, since Aldrich (1976), Ennis (1973) and Menge (1979) have noted that Asterias spp. are prey to crabs, lobsters, and cannibalistic attacks. Further, crabs and lobsters prey on musseis as weIl. Hence, the organization of subtidal communities in New England is not yet understood; more work is clearly needed. 3.2.3. European shores Although much work has been done on asteroid diets in subtidal habitats in northern

536 Bruce A.Menge

Figure 2. Generalized food web of the northern European subtidal region. Percentages of some items in diets are given by the appropriate lines. Data for Astropecten irregularis are from Christensen 1970. Those for Luidia ciliaris are from Brun 1972 and Hunt 1925 (parentheses). Data far Asterias rubens are from Anger et al. 1977. Other data taken from Verbeek 1977, Mayo & Mackie 1976, Feder & Christen­ sen 1966, and Gullikson & Skjaeveland 1973.

Europe, there are no data available which allow some estimation of the predatory impact of asteroids on their prey. Figure 2 represents a very generalized food web extracted from several studies (Hunt 1925, Feder 1967, Christen sen 1970, Brun 1972, Gullikson & Skjaeve­ land 1973, Mayo & Mackie 1976, Anger et al. 1977, Verbeek 1977). These data were obtained from the North Sea and adjacent waters. Substrata varied from cobble-gravel to sandy mud. The web is thus not necessarily representative of one particular site. The diet percentages shown are extracted from those studies which gave numerical diet composition (Anger et al. 1977, Christensen 1970, Brun 1972). In only one of these studies (Anger et al. 1977) were diets directly observed in the field. The others were based on gut contents of dredged specimens, which as many workers have noted, often lose or drop prey while being dredged and hence do not yield unbiased sampies of diet composition. The food web (fig.2) illustrates several features. First, it is complex. The number of aste­ roid predators is large and some are secondary or even tertiary carnivores. Second, most of these predators are quite generalized in their diets (e.g. Astropecten irregularis, Asterias rubens, Luidia ciliaris, Marthasterias glacialis). Third, at least one of these asteroids, A. rubens, seems very flexible in its ability to adapt to local prey availability (e.g. Anger et al. 1977). Further, A.rubens and A . irregularis would appear to be potentially important pr i­ mary predators on the benthos. However, although both Anger etal. (l977;A.rubens) and Christensen (1 970; A.irregularis) suggested that these asteroids have a major impact on ben­ thic community organization in this region, no data are yet available to permit even a rough estimate of their respective effects. Certain prey species in this food web have escape or avoidance responses to asteroids, inc1uding some of the intermediate-Ievel asteroid predators. For example, A.rubens runs in response to L.ciliaris, Crossaster papposus and Solaster endeca (Mayo & Mackie 1976)

Effects of feeding on the environment: Asteroidea 537 while Mglacialis runs in response to Cpapposus, individuals of which also avoid each other (Mayo & Mackie 1976). In addition, Feder (1967,1972) and Feder & Christensen (1966) have observed many species of limpets (Emarginula reticulata, Patella vulgata, Patina pel­ lucida, Acmaea tessulata), coiled gastropods (Gibbula cineraria, Cantharidus c1elandi, Cal­

liostoma zizyphinum, Natica intermedia, N.montagui, Nassarius reticulatus, N.incrassatus, Buccinum undatum, Haliotis tuberculata), bivalves (Cardium echinatum, Laevicardium crassa, Pecten spp., Chlamys spp., Spisula subtruncata) and an eehinoid (Strongylocentro­ tus droebachiensis) to run, twist, leap, raise their mantles (molluses) extend their tube feet and flee (echinoid), or otherwise respond in a defensive manner to contaet with several other asteroid speeies (includingA.rubens, Mglacialis, Cpapposus, S.endeca, Asterina gib­ bosa). This rieh array of responses suggests rather strongiy that (1) asteroid predation is a major selective agent operating on these benthic species, and (2) asteroids have a strong predatory impact on the subtidal benthie eommunities of this region. However, further work is needed to substantiate these suspieions.

3.3. Temperate southern hemisphere 3.3.1. Chile No experimental information on the eommunity role of subtidal asteroids is available for Chile. However, some insights into this system are provided by the work of Viviani (1978) and Dayton et al. (1977). Viviani (1978) examined diets, patterns of arm regeneration, aggregation tendencies, aggression, and escape responses in five species of subtidal asteroids. These are Meyenaster gelatinosus, Luidia magellanica, Stichaster striatus, Heliaster helianthus and Patina chilen­ sis. Figure 3 shows a food web extraeted from the work of Viviani (1978). Note that most of the species are rather generalized in diet. L.magellanica is the most specialized predator, feeding on eight species of echinoderms, and shares the role of top predator with Mgelati­ nosus. Hhelianthus, though most eommon in the intertidal, also oecurs in the subtidal where it displays the broadest diet of these asteroids (preying on 58 species). S. striatus is also generalized (preying on 28 species) but preys primarily on sessile species (fig.3). P. chilensis is suggested to be primarily a seavenger but also preys on various sessile species. The community roles of these predators is not known but the observations of Dayton et al. (1977) and Viviani (1978) suggest asteroid predation exerts severe effects on prey populations. Dayton et al. (1977) observed that many mobile prey of M.gelatinosus exhibit strong and very effective escape responses to this large, voracious asteroid. The data of both Viviani (1978) and Dayton et al. (1977) indicate that this speeies is very generalized in diet, preying on asteroids (including cannibalism), eehinoids, predaceous and herbivorous gastropods, chitons, musseis, barnacles and tunicates (fig.3, Dayton et al. 1977). Moreover, most of the mobile species have eseape responses and Mgelatinosus individuals have strong flight responses to other Mgelatinosus and to L. magellanica. L. magellanica also exhibits cannibalism and intra- and interspecifie aggression. None of the primary asteroid carnivores (Hhelianthus, S.striatus, P.chilensis) exhibits cannibalism or interspecifie predation and all have gregarious habits. Viviani (1978) also observed high rates of arm regeneration in all populations of these asteroids which he examined. Laboratory observations suggest that these arms were lost during attaeks by Mgelatinosus or L.magellanica. Finally, Viviani (1978) found that to escape death during such attaeks, the vietim asteroid must be greater than some minimum size (exeept P.chilensis, all of whieh are small). Otherwise, the prey

538 Bruce A.Menge

Figure 3. Generalized food web for the shallow subtidal region of the north and central coast of Chile (lquique region). Arrows point to the predator. Modified from Viviani 1978.

is completely devoured. Moreover, the juveniles of these species and all individuals of P. chilensis te nd to occur in cryptic habitats (e .g. S.striatus recruits are restricted to the hold­ fasts of the alga Lessonia nigrescens and the undersides of rocks). Such behavioral responses of predators and prey and solitary habits of the two top pre­ dators suggest that cannibalism and predation by the asteroids are important agents of mortality , and thus of natural selection, on many of the mobile species. The possible effects of these predators on the sessile prey are not suggested, but figure 3 suggests aste­ roid predation mayaiso have a significant impact on these species.

3.3.2. South Australia The community roles of four asteroids on subtidal (9 m below MLLW) epifaunal commu­ nities of pier pilings were investigated near Adelaide, Australia (Keough & Butler 1979). The community is a relatively complex one. Major sessile species include sponges, ascidians, bryozoans, a stony coral, and bivalves. Herbivorous gastropods and various detritus fee ding crustaceans occurred in trophically low positions in this web. Three of the asteroids (Tosia australis, Patiriella brevispina and Petricia vernicina) are primary carnivores and one (Cos­ cinasterias calamaria) is a secondary carnivore. However, C.calamaria is evidently preyed upon by tertiary carnivores (heterodontid sharks and majid crabs). The other asteroids seem not to have predators. Other predators of the sessile prey include fish, predaceous gastro­ pods and dorid nudibranchs. The authors estimated that ca. 5 % of the primary space was bare. They attempted to determine the role of these asteroids by enclosing individuals of each species into separate cages placed over the substratum on the pilings. Uncaged and caged controls were used to allow estimation of fee ding rates of each asteroid. The results of these experiments after six months suggest that the asteroids are not major predators on the sponges, tunicates and bryozoans that occupy most of the space on the pilings. Although percent cover of many sessile species changed during the experiment, few of the decreases could be attribu­ ted to asteroid predation.

E[[ects o[ [eeding on the environment: Asteroidea 539 In view of the intermediate level of the asteroid predators and the high cover of sessile biota, this prelirninary result is not surprising. Experiments on the role of other factors in organizing this community are not yet available. Keough & Butler noted that the asteroids are generally relatively scarce on these pilings and rarely crawl more than 0.5-1.5 m up the pilings. The nature of the surrounding substratum is not given, but this observation may suggest that these pilings are either not representative of the natural habitat of these aste­ roids or that the habitat is largely unavailable to them. At any rate, further work on the structure and organization of this interesting system is clearly needed.

3.4. Tropical subtidal habitats Though initial reports of the presumed short- and long-term effects of Acanthaster planci on coral reefs were weakly supported by data, work done over the past decade offers some insight into the true role of A.planci in coral reef communities. Despite these recent efforts, much remains to be learned.

3.4.1. Community structure Generally , coral reef morphology includes a shoreward reef flat, a reef crest, and a forereef slope (e.g. Wells 1957). Imposed upon this morphology are distinct zonation patterns (e.g. Goreau 1959, Loya 1972, Goreau & Goreau 1973, Randalll973, Glynn et al. 1972, Glynn 1976). Although details vary considerably, reef flats tend to have relatively little coral cover, a high cover of coralline algae, relatively low species richness and are relatively spa­ tially homogeneous. Relative to reef flats, reef crests tend to have high covers of coral, algae or both, high species diversity, and are spatially heterogeneous. Finally, reef slopes tend to have high covers of hard corals, sponges, algae and soft coral, very high diversity and are highly heterogeneous (see e.g. RandallI973). Some idea of the diversity and abun­ dance of species other than those mentioned above (e.g. polychaetes, sipunculids, forarni­ niferans, crustaceans, molluscs, echinoderms, fish) is given for a Caribbean reef in Glynn (1973a, table 2). However, as yet, detailed quantification of overall community structure (see table 1 for definition) on coral reefs has not been done. 3.4.2. Community organization Studies on A.planci have been centered in three main geographie regions: the east Pacific (Panama), central west Pacific (Hawaii, Mariana Islands and the Great Barrier Reet), and the Red Sea. Research at the latter site has focused primarily on the behavior and physio­ logy of feeding in A.planci. Studies in the former two regions have attempted to interpret the effects of A.planci on community structure. In the east Pacific, coral reefs are not ubiquitous features of the subtidalseascape, but do occur (Glynn et al. 1972). Hermatypic corals are much less diverse on these reefs than on reefs in other regions (about 15 spp. vs 50 to > 100 spp. in other tropical areas) (Glynn et al. 1972, Goreau & Goreau 1973, Lang 1973, Randall 1973). However, the general reef structural pattern outlined above is observed on these reefs (Glynn 1976). The reef flat has both low coral cover and species richness, the reef crest has high coral cover but rela­ tively low richness and the reef slope has relatively low coral cover and relatively high species richness. Porter (1972, 1974) compared coral abundance, diversity and density of A.planci on two of these reefs and suggested that A.planci played a community role comparable to

540 Bruce A.Menge that of Pisaster ochraceus in the Pacific Northwest (Paine 1966, 1974). Thus on the reef slope of one reef, high density of A.planci was correlated with high species richness and low cover of corals. On the other reef, low A.planci density was correlated with low rich· ness and high cover of corals. Porter (1972, 1974) interpreted these observations as follows: (1) A.planci preys preferentially on the coral Pocillopora damicornis, the dominant space occupant. (2) P.damicornis is competitively dominant in both overgrowth and aggressive competition interactions (as suggested by Glynn et al. 1972). (3) As a result of selective predation, competitively inferior and non·preferred species are maintained in the system. In light of more recent work, this interpretation now appears to have been premature. Glynn (1976, 1977) suggested that (1) Porter's estimates of apparently greatly different A. planci densities were an artifact of sampling techniques; densities based on more intensive censuses actually appear similar on the two reefs (Glynn 1976). (2)P.damicornis is not an aggressively dominant species (in terms of extracoelenetric feeding response). (3) Further, this coral is not usually preferred as prey because it harbors species of crustacean symbionts (a shrimp, Alpheus sp.; crabs of the genus Trapezia) which often prevent A.planci from mounting and fee ding on their host. Since most P.damicornis have these symbionts, they are partially immune from predation. As a result, A.planci prey more on broken branches 'of P.damicornis. These are small and have few or no symbionts. Glynn (1976) interpreted the structure and diversity patterns listed earlier as a consequence of several factors includ· ing (a) physical harshness and disturbance (caused by either occasional severe desiccation on the reef flat which lowers both diversity and abundance by affecting all species, or wave damage and talus slumps on the reef slope which lowers abundance but increases diversity by acting in a selective manner onP.damicornis) and (b) predation by many species (e.g. Glynn et al. 1972, Glynn 1973b) including A.planci. Though Glynn's interpretations need further confirmation, his work indicates that clear understanding of this system depends on careful observations and experiments. Observation of the defense of P.damicornis by symbionts clearly complicates the development of an understanding of the prey preferences of A.planci which is important to a broad under· standing of the community role of this asteroid. Research on the effect of A.planci in the Indo·West Pacific has centered around Guam and the Great Barrier Reef(e.g. Endean 1973, Randall1973, Chesher 1969, Weber & Wood· head 1970, Laxton 1974a,b, Laxton & Stablum 1974). There is little doubt that A.planci has had a great effect on some Pacific reefs (e.g. Endean 1973, Chesher 1969, Laxton 1974a, b, Goreau et al. 1972). However, these effects may be more restricted than previously believed (Weber & Woodhead 1970). For example, in a study in Hawaü (Branham et al. 1971), a large aggregation of A.planci was observed that did not move in over a year and evidently had a minor effect on the coral. Such apparently contradictory variations in the effect of A.planci have been suggested to result from a preference 'hierarchy' of this asteroid (Birkeland & Randall 1979). These workers suggested that A.planci prefers species of Acropora and will thus devastate areas dominated by this coral genus. They noted that feeding rates of A.planci are greatly reduced in areas dominated by other species, especially corals in the genera Porites and Pocillopora. The effect of an invasion of A.planci is suggested by Laxton (1974a,b) and Laxton & Stablum (1974). Photographic transects provided data on cover oflive and dead coral, soft coral, algae, hydrozoans, and sponges (though not species composition or number) on 14 reefs on the Great Barrier Reef complex (Laxton & Stablum 1974). Analysis of Laxton's (1974b) data reveals a significant correlation between overall cover of dead coral (Y) and

Effects of feeding on the environment: Asteroidea 541

density of A.planci (X) (Y = 10.4 + 12.8X, r =0.47, P =0.01). By examining his photo­ graphs for abundances of live coral, recently killed coral (skeleton still white and not over­ grown by mamentous algae), and coralline algae, Laxton (I 974b) attempted to evaluate the impact of A.planci on coral abundance. Further, by separating corals according to growth form, he attempted a cmde evaluation of prey selectivity in A.planci. Both of these procedures have come under attack. The former procedure may be limited because coloni­ zation studies suggest that coral skeletons are rapidly (ca. 1 week) overgrown by algae (Biggs & Eminson (I977). Thus, evidence of recent « 1 week) coral mortality may seriously underestimate the extent of actual recent « 1-2 months) mortality. The evaluation of prey selectivity has been criticized because prey selectivity of A.planci depends on nematocyst (Ormond et al. 1976, Moore & Huxley 1976) and symbiont defenses of the coral (Glynn 1976), as weIl as morphology and abundance. Perhaps the most critical assumption made in tbis analysis was that the observed cover of coralline algae equalled the cover of live coral eaten by A.planci some time prior to the sampling period. By comparing reefs attacked by A.planci to unattacked reefs, Laxton hoped to detect short- and long-term effects of this predator. The data are summarized in tables 2 and 3. Assuming that all dead hard coral on attacked reefs was living hard coral before attack and eventually became covered by coral­ line algae (= 28 ± 7 %; table 2), the cover oflive coral was reduced on the average to about 18 % of its original amount. Although tbis would appear to be devastating mortality, the consequence of tbis to overall community structure is unclear. Laxton (1974a; table 3) also suggested that brancbing staghorn and tabular corals are the primary targets of preda­ tion by A.planci. Such corals often recruit weIl and grow quickly (e.g. Loya 1976a,b, Con­ neIl1973, Glynn 1976) and may return to bigher abundances relatively soon. Further, thoughjuvenile A.planci feed on coralline algae (Yamagucbi 1974), it is not known if an increase in coralline algae will encourage A.planci recruitment. However, since much sur­ face area is covered by coralline algae on 'normal' reefs (e.g. 50-60 % table 2, Vine 1972),

Table 2. Cora! and coralIine a!gae cover on reefs with and without Aeanthaster plancia Reef characterization

Nb

Live hard cora!c

Dead hard cora!c

Coralline algae e

Aeanthaster planci absent or scarce

13

35 ±5

2±1

61 ± 2

13

34 ± 7d

0

54 ±6

6 ±2e

28 ± 7

54 ±6 f

6±2

0

82 ±2f

Aeanthaster plan eid abundant a) before attack (extra­ polated) b) upon departure of A.planci (observed) c) after departure (extra­ polated)

a. Raw data from Laxton 1974b b. Number of reefs observed c. Percent cover. Data are x ± lSE d. Observations made at point (b), abundances in (a) and (c) are based on the assumptions that aIl dead coral observed was killed by A.planei and that aIl this dead cora! would eventually be covered by coralline a!gae e. Decline in cora! cover is statistically significant: 1 way ANOVA; F = 13.8; 1, 24df; p 0.005 f. Presumed increase in coralline alga is statistically significant: 1 way ANOVA; F = 20.3; 1, 24df; p 0.001.