L ru a S J 0* s a m a A STUDENT'S TEXT. BOOK OF ZOOLOGY. By ADAM SEDGWICK, M.A., F.R.S., Professor of Zoology and Comparative Anatomy in the University of Cam bridge, Professor of Zoology (elect) at the Imperial College of Science and Technology. In Three Volumes. Profusely Illustrated. Vol. I. Protozoa to Chsetognatha xii., 620. 18/ (1898). Vol. II. Amphioxus and Vertebrata, xvi., 705. 21/- (1905). Vol. III. Tunicata, Enteropneusta,Echinodermata, and Arthropoda. The Introduction to Arthropoda,the Crustacea and Xiphosura by J. J. LISTER, M.A., F.R.S. The Insecta and Arachnida by A. E. SHIPLEY, M.A., Hon. D.Sc., Princeton, F.R.S. xii., 906. 24/- (1909). SOME OPINIONS OF THE PRESS ON VOLS. I AND II. ' Both as a successful teacher of many years' standing and as the translator of Glaus' " Text-Book of Zoology," a work which has for long held its own against younger rivals, the author is well A STUDENT'S TEXT-BOOK OF ZOOLOGY SOME OPINIONS OF THE PRESS continued. equipped for his task. He has produced, as was to be expected, a well-planned, well-illustrated, and trustworthy manual, a book which is so good, and so likely to exert a considerable influence by later editions, that our best plan will be rather to indicate the capabilities for improvement than to eulogize the successes at- tained." Athenceum (1898). " Zoological students owe a deep gratitude to Professor Sedgwick for this excellent and very comprehensive work." Spectator (1898). " As an exposition of the modern ideas of the morphological rela- tions of the coelom and the animals possessing it, Mr. Sedgwick's work will take a high rank." Literature (1898). " The book is a most valuable contribution to zoological literature and will be of special use to advanced students and teachers as well as to those working at special groups, who wish to keep in touch with general zoology." Lancet (1898). ' It makes an admirable and handy book of reference to others interested in natural history, who will here find the general nature and habits of a large number of animals described in a readable style. No efforts have been spared to make the path of the young zoologist as easy and pleasant as possible. The book is beyond question one of the best volumes on zoology at present available. "- Knowledge (1898). " Mr. Sedgwick is an original thinker with a clear, terse and incisive style." Guardian (1898). '' It is without doubt the completest work of its kind accessible to English readers." Pall Mall Gazette (1898). A STUDENT'S TEXT-BOOK OF ZOOLOGY SOME OPINIONS OF THE PRESS continued. " Mr. Sedgwick's book is a very good one, ably put together and likely to be extremely useful. It is, in fact, not only the last but the best zoological text-book in the language." Prof. Ray Lankester in Nature (1898). " Zoologists are to be congratulated on the appearance of the second volume of what promises to be a monumental work." Athenceum (1905). " There is all the information that the student, and even the advanced student, requires about the anatomy and the classification of vertebrated animals ; moreover all the salient facts are illus- trated by numerous figures in the text. We know of no English text-book that is so full." LanceJ (1905). " The student will find everything that a text-book of zoology should contain." British Medical Journal (1905). " This cautious and excellent text-book."- Zoologist (1905). "It is a fine piece of work which will fully uphold the reputa- tion that the first volume has gained."- Cambridge Review (1905). " The first volume has a well-deserved reputation for accuracy, clearness, terseness and independence, and in the crowd of text-books it has filled a definite place to the satisfaction of teachers as well as of students. " The scholarliness, clearness and carefulness of statement are obvious, but those who work with it will discover other virtues. A suggestive scepticism, a mature judgment and a more indefinable quality which we can only hint at in the phrase morphological insight." Nature (1905). A STUDENT'S TEXT-BOOK OF ZOOLOGY SOME OPINIONS OF THE PRESS continued. " It is marked by all the qualities of accuracy, lucidity and orderly method which were conspicuous in his first volume, and encourages us to hope that when the work is complete it will compare favourably with any of the books on a similar scale which have hitherto been produced more commonly in other countries. Mr. Sedgwick is not merely a biologist ; he is also a cosmical philosopher, who has ever kept in mind the relation of his particular subject to the vast scheme of things, and so is able to perceive the proper proportion of its details." Spectator (1905). " There is never any doubt what he means, and the lucidity of his exposition can only come from a thorough knowledge of the facts he has to lay before his readers." Guardian (1905). A STUDENT'S TEXT-BOOK OF ZOOLOGY CORRIGENDA "x Page27. Line 10-11 of paragraph 2, for " Rtiodooma " read " Rhodosoma." 167. Legend of Fig. 117, for " Parkinson " read " Parkinsonl." 177. Legend of Fig. 128, for " hedatic " read " hepatic." 183. Line 10 from bottom, for " Gymnasteria" read " Gymnasterias." 189. Fam. 1 delete the following genera Gnathaster Slad., Mimaster Slad., Leptogonaster Slad. 189. Line 8 of Fam. 1, for " tenuispiins " read " tenuis pints." 266. Line 10 from bottom, delete " Pseudostichopus Theel." ,. 269. Line 4 from bottom, for " Myriotroeha " read " Myriotrochus." 487. Line 10 from bottom, for " Armadillium " read " Armadillidium." 540. Last line, for " Coenobita " read " Cenobita," or alter the other mentions of the genus to Coenobita. ,, 573. Line 6 from bottom, for " Opisthoptaus " read " Opisthopatm." 716. From 25 line 12, 13. Taleporia and Solenobia are now classified with the Tineidae, which see. J2_- A STUDENT'S TEXT-BOOK OF ZOOLOGY BY ADAM SEDGWICK, M.A. F.R.S. FELLOW AND FORMERLY TUTOR OF TRINITY COLLEGE, CAMBRIDGE, AND PROFESSOR OF ZOOLOGY AND COMPARATIVE ANATOMY IN THE UNIVERSITY. PROFESSOR OF ZOOLOGY (ELECT) AT THE IMPERIAL COLLEGE OF SCIENCE AND TECHNOLOGY VOL. Ill THE INTRODUCTION TO ARTHROPODA, THE CRUSTACEA, AND XIPHOSURA, By J, J. LISTER, M.A. F.R.S. FELLOW OF ST. JOHN'S COLLEGE THE INSECTA AND ARACHNIDA By A. E. SHIPLEY, M.A. Hon. D.Sc. Princeton, F.R.S. FELLOW OF CHRIST'S COLLEGE AND READER OF ZOOLOGY IN THE UNIVERSITY LONDON SWAN SONNENSCHEIN AND CO. LTD NEW YORK: THE MACMILLAN CO. 1909 PREFACE IN parting with the present volume I have the satisfaction of feeling that I have accomplished the first part of the task which I set myself fourteen years ago. When I began I looked forward to executing the whole work, both Special and General Parts, myself in seven or eight years. That expectation has not been realized. Indeed had it not been for the assistance given me by my colleagues, Mr. J. J. Lister and Mr. A. E. Shipley, the present volume would have been far from completion. Thanks to their generous co-operation it is now finished, and I am in a position to turn my attention to the General Part. It is satisfactory to have finished it, but it is impossible to avoid a feeling of regret that owing to circumstances which were unforeseen when the work was commenced it has occupied so manv years and that the General Part, to write which the Special Part was undertaken, is not yet begun. The distribution of the work of this volume has been as follows : The chapters on Tunicata, Enteropneusta, Echinodermata. Onychophora and Myriapoda are by my- self ; the chapters on the Arthropoda in General and on the Crustacea, and the section on the Xiphosura are by Mr. Lister ; and those on the Insecta and the Arachnida by Mr. Shipley. Our thanks are especially due to Professor Herdman, Mr. Punnett, Dr. Bather, Professor MacBride, Mr. Sinclair, Dr. Caiman, Mr. L. A. Borradaile, Mr. David Sharp, Mr. Hugh Scott, Professor Imms, Mr. C. Warburton and Mr. C. C. Dobell for reading the proofs and for the advice and assist- ance which they have given in different parts of the volume. Our thanks are further due to those authors and pub- lishers who have allowed us to use illustrations which have VI PREFACE appeared in other works. The sources of these are acknow- ledged, but we must especially mention Claus's Lehrbuch, Shipley & MacBride's Zoology, Bronn's Thierreich, Kor- schelt & Heider's Textbook of Embryology, Delage & Herou- ard's Zoologie Concrete, Perrier's Traite de Zoologie, the Cambridge Natural History, Lang's Textbook of Comparative, Anatomy, Zittel's Grundziige der Palaeontologie , Lankester's Treatise on Zoology. I have also to state that the drawings from which figures 105 and 132 were engraved were made by Mr. J. C. Simpson under the direction of Professor MacBride. Now that the work is completed something must be said as to its architectural plan. This has been criticized (see, for example, the review in Nature, November 23, 1905) OIL the ground that the Arthropods are separated from the Annelids, and that the Tunicates and Enteropneusta are placed at the end of the Chordata. In the first volume the clue given by the coelom is mainly followed and this leads from the Annelids to the enterocoelic phyla the Phor- onidea, the Brachiopoda and the Chaetognatha. Having reached this point it was not possible to treat the Arthropoda until after the Enterocoela were finished. This accounts for the position of the Arthropoda at the end of the work. It may be objected that the Arthropoda should have come after the Annelida in volume I. But this would have involved the inconvenience of separating the Gephyrea Achaeta from the Gephyrea Armata, and the Brachiopoda and Polyzoa from the Chaetopoda. Moreover, the Arthro- poda differ so fundamentally from the Annelida in their coelomic arrangements and are such an enormous and self- contained group that it did not appear that any practical disadvantage would follow upon their separation from the latter which in some features of their organization they so closely follow. In fact we are here confronted with a difficulty which the systematic Zoologist meets at every turn and to which attention has often been called in the course of the work. When there is more than one clue, as there nearly always is, which shall we follow ? We are obliged to adopt a linear arrangement, whereas Nature PREFACE Vll is content with nothing less than arrangement in three dimensions. Following then the clue of the coelom and having reached so near the Chordata, it seemed natural and proper to take them after the Chaetognatha. I might of course have placed the Echinoderms here and treated the Enteropneusta after them. That was in fact the original intention, but after careful consideration there seemed no reason based on zoological affinity why this should be done and it would have been almost an outrage to have placed the Echinoderms next to Sagitta. It was accordingly decided to take the Chordata at this point, beginning with Amphioxus and re- serving the Enteropneusta to the end to lead on to the Echinoderms. Though I offer this explanation I quite recognize the validity of the criticism. All zoological arrangements are compromises, and none of them can be, now or ever, entirely natural. A. SEDGWTCK. CAMBRIDGE, November 25, 1908. TABLE OF CONTENTS CHAPTER I PAGE PHYLUM TUNICATA .... 1 Order 1. Ascidiacea. . . 25 Tribe 1. Ascidiae simplices (Monascidia) . . 26 2. Ascidiae composi- tae (Synascidia) 30 3. Ascidiae Salpae- fonnes (Asc. Luciae) ... 39 Order 2. Thaliacea ... 44 Sub-order 1. Hemimyaria (Safpida) . . 45 ,, 2. <>ydotnyaria (Doliolida}. . 54 Order 3. Appendiculariae . 60 CHAPTER II PHYLUM ENTEROPNEUSTA . 66 Order 1. Balanoglossida . 68 2. Cephalodiscida . 104 CHAPTER III PHYLUM ECHINODERMATA .115 Class I. ASTEROIDEA . . 166 Order 1. Encrinasteriae . . 188 2. Euasteriae. . .189 Sub-order 1. Phanerozonia 189 ,, 2. Cryptozonia . 191 Class II. OPHIUROIDEA . 194 Order 1. Lysophiurae . . 205 2. Zygophiurae (Ophiurae) . . .205 3. Streptophiurae . 206 , 4. Cladophiurae (Euryalae) . . . 207 Class III. ECHINOIDEA . Order 1. Palaechinoidea 2. Euechinoidea . Sub-order 1. Cidaroidu ,, 2. Diadematoida Section 1. Streptosoiuata 2. Stereosomata . Sub-order 3. Holectypoida . ,, 4. Clypeastroida 5. Spatangoida . Section 1. Asternata 2. Sternata Class IV. HOLOTHUROI- DEA . . Order 1. Actinopoda Sub-order 1. Aspidochiro- tae . 2. Dendrochirn- tae . Order 2. Paraetinopoda . Class V. CRIXOIDEA . Order 1. Larviformia 2. Fistulata . . 3. Camerata . . ,, 4. Flexibilia . . 5. Articulata . Class VI. CYSTIDEA. . Order 1. Amphorida 2. Rhombifera . ,, 3. Diploporida 4. Edrioasterida . Class VII. BLASTOIDEA CHAPTER IV PHYLUM ARTHROPODA PAGE 208 238 239 231? 239 . 240 . 240 242 243 244 24 245 247 265 266 268 269 270 293 293 295 297 298 303 309 309 309 310 310 . 314 60276 CONTENTS CHAPTER V PAGE Class I. CRUSTACEA . . 342 Sub-class 1. EXTOMOSTRACA 359 Order 1. Trilobita . . .361 2. Branchiopoda . . 3(57 Sub-order 1. PhyUopoda . 375 Tribe 1. Anostraca . . 376 2. Xotostraca . . 376 3. Conchostraca . . 377 Sub-order 2. Cladocera. . 377 Order 3. Ostracoda . . . 382 Sub- order 1. Myodocopa . 389 2. Podocopa. . 391 Order 4. Copepoda . . . 392 Sub-order 1. Eucopepoda . 404 2. Branchiura . 410 Order 5. Cirripedia . . . 413 Sub-order 1. Cirripedia Genuina . . 423 Tribe 1. Pedunculata . . 423 2. Operculata . . 425 Sub-order 2. Acrothoracica 426 3. Apoda . . 428 4. Rhizocephala 428 5. Ascotkoracica 434 Sub-class 2. MAI.ACOSTRACA 435 Order 1. Leptostraca . . 455 2. Syncarida . . . 459 3. Schizopoda . . 463 Tribe 1. Hemitropha . . 470 2. Holotropha . . 471 Order 4. Cumacea . . . 472 5. Tanaidacea . . 477 6. Isopoda . . . 479 Section 1. Isopoda genuina 487 Tribe 1. Oniscoidea . . 487 2. Asellota ... 487 3. Phreatoieidea . . 488 4. Valvifera ... 488 5. Flabellifera . . 488 6. Epiearidea . . 489 Section 2. Isopoda anomala 491 Order 7. Amphipoda. . . 492 Sub-order 1. Amphipoda genuina Division 1. Gam.mari.na Tribe 1. Corophiina ,, 2. Gammarina Genuina 499 500 500 500 PAGE Division 2. Hyperina . . 502 Tribe 1. Hyperina anomala 502 2. Hyperina normalia 503 Sub-order 2. Laemodipoda 503 Order 8. Stomalopoda . . 504 9. Decapoda . . .511 Section 1. Macrura Natantia 527 Tribe 1. Penaeidea . . 528 2. Stenopidea . . 531 3. Caridea ... 531 Section 2. Macrura Reptantia . . 534 3. Anomura . . 536 Tribe 1. Thalassinidea . . 537 2. Paguridea . . 539 3. Galatheidea . . 541 4. Hippidea . . . 542 Section 4. Brachyura . . 542 Tribe 1. Brachyura Anomala . . 543 2. Oxystomata . . 544 3. Oxyrhyncha . . 544 4. Catometopa . . 546 5. Cyclometopa . . 548 CHAPTER VI Class II. ONYCHOPHORA . 550 CHAPTER VII Class III. MYRIAPODA . . 578 Order 1. Pauropoda . . 580 2. Diplopoda . . 582 Sub-order 1. Psdapliog- nntha . . . 591 2. Chilo. IH't:}. p. 11!). t De Selys-Long^huinps. Arch. liiol., 17, I'.tol. p. 499. RENAL AND REPRODUCTIVE ORGANS. 17 beneath the endostyle, giving off near its origin a branch to the test ( Fig. 12). The other after giving off a branch to the test, which accompanies the branch from the endostylar vessel, is distribute:! to the digestive viscera, gonads and body wall. The endo- stylar vessel communicates with the sinuses in the pharyngeal wall and these again with a dorsal vessel running along the dorsal lamina. This posteriorly is distributed to the viscera and body wall, whence the blood is returned to the posterior end of the heart. It is a peculiarity very generally observable in Tunicata that the heart contracts a certain number of times in one direction and then a similar number of times in the opposite direction. Thus for a certain number of beats it acts as a respiratory heart driving the blood to the respiratory organs and thence to the system. It then reverses its action and becomes a systemic heart driving the blood first to the system and thence to the respiratory organs.* The blood is colourless and contains nucleated corpuscles. These are generally colourless and amoeboid, but some of them generally contain pigment, either yellow, red, brown or blue. In the Botryllidae the terminal branches of the test-vessels are dilated into ampulla-like sacs which have rhythmically contractile walls and assist in the circulation of the colony. The renal organs f are but little understood and appear to have no relation with the coelom. The only structures to which a renal function has been ascribed are some vesicular bodies containing concretions of uric acid and other substances, and placed in the walls of the intestine, in the mantle and sometimes in other places. These structures which have not been found in all Tunicates are without a duct ; so that the excretory matters cannot escape. In the Molgulidae there is a large saccular body of this nature on the right side of the body. It has been suggested that the neural gland plays a part in renal excretion. Reproductive Organs. The Tunicata are with very few exceptions (e.g. Oiko pleura dioica) hermaphrodite, and as a general rule the female organ ripens first. So far as can be ascertained the gonads have no relations with the coelom, either developmentally or otherwise. The glands are continuous with * Schultze, JcnaZeitsch., 35, 1901, p. 2'Z1 f Dahlgriin. Arch. f. mik: Anat. 58. 1901, p 60S z in c 18 PHYLUM TUNICATA (UROCHORDA). their ducts, the terminal parts of which are usually single and open into the atrial cavity. The ovaries and testes are generally placed near one another and contained either amongst or near the digestive viscera or in the mantle wall. In the Potyclinidae they are in the post-abdomen. In dona (Fig. 13) the testis is a branched gland ramifying on the wall of the intestine between the pylorus and the rectum. Its tubules gradually collect into one main duct, which accompanies the rectum and opens into the atrium in front of the anus by several openings, the walls of which contain a red pigment, consisting of red renal vesicles of the kind referred to above. The ovary is a rounded mass placed in the intestinal loop, and the oviduct accompanies the vas deferens to open close by it, far forward into the atrium. The gonads arise from a common mesodermal rudiment. In the synas- cidians they do not as a rule appear in the zooid which develops from the egg, and they may be absent from the first-formed generation of budded zooids (BotryUus. etc.). Reproduction by budding occurs in a large number of Tunicata. Sometimes, as in the synascidians. the budded individuals remain em- bedded in a common tunic with the parent ; sometimes, as in Tinliacea, they become eventually separate and lead for a time an independent existence. In the latter case the life history is complicated by the phenomenon of alternation of generations of the variety known as metagenesis ; for the indi- viduals which bud proceed from the egg and do not develop sexual organs. Budding does not occur in the Appcwliculnriae and monascidians. In all cases excepting BotryUus the budding is effected by the division of a ventrally placed process of the body called the stolon. The stolon contains a diverticulum of the pharynx known as the spicardium and given off between the endostyle and the oesophageal opening (p 15). The stolon also con- tains an extension of the mesodermal and vascular tissues of the parent. For a more detailed account of the phenomenon and of the origin of the ortmii- ach: / intestinal limp: ij n'ctiini. DEVELOPMENT. 19 to the accounts given below under each family or order. Here we will only remark that there is considerable variability- in the mode of formation of the organs. The pharynx, atrium, diges- tive organs and pericardium are usually derived from the epi- cardial process of the pharynx, while the ganglion in the synascidians usually develops from the endoderm, but it may arise from the mesoderm (Pyrosoma) or from the ectoderm (Thaliacea). In the Botnjllidac the endoderm of the parent does not participate in the budding process. Development. In most monascidians (except Ci/nl/il't, etc.) the eggs are fertilised in the sea or in the atrium and undergo their whole development outside the body of the parent. This is also the case in Doliolum and in the Appendiculariae. In the synascidians on the other hand the early development usually takes place in the atrial cavity or in incubatory pouches of it. In the salps the egg undergoes its early development in the ovary ; in the later stages it emerges into the atrial cavity but remains connected with the parent by the placenta The eggs are frequently laid in their follicle, which is somewhat com- plicated. It is formed of two layers (Fig. 14), the outer of which consists of vacuolated cells ; these are prolonged into papillae, and help to float the egg in the sea. The inner layer consists of follicle-cells which have migrated inwards, and are called the test-cells because formerly they were supposed to give rise to the test of the adult. The two layers are separ- ated by a structureless chorion. The development generally leads to the formation of a free-swimming tailed larva, the tadpole larva, by means of which the species is distributed over a wider area. The tad- pole larva is nearly always formed in the monascidians (it is absent in some species of Molgula) and in the synas- cidians (absent in Pyrosoma). Tt is akn fnnnrl in Dnlinhim FIG. 14. Mature egg from the oviduct of Ciona in JJOllOlUm, i,i>,-xtinalis (after Kupffer). < i<>llir!r cells but not in thp aln< im,,m-like cells); d chorion; . test-cells; _ baipb. / ovum ; z gelatinous substance. The eggs usually have but little yolk. In the synascidians they are however richer in yolk. and in Pyrosoma the cleavage is actually meroblastic 20 PHYLUM TUNICATA (UROCHORDA). Development of the tadpok larva. The segmentation is complete and leads to the formation of a blastosphere, from which the gastrula arises by invagination. The gastrula FIG. 15. Development of r/nil/nxiii niniininllatii (alter Kowalevsky from Glaus), a Com- mencemeut of invayination ; / h cleavage-cavity. b Gastrula with blastopore ; ch rudi- ment of notochonl ; hi/ endodenn. (. Later stage ; Ek ectoderm ; ^V rudiment of neural- canal ; Eil' eiulodenn of future tail region, d Stage with body and tail ; Ed' endoderm of tail ; M muscular cells in tail, <> Just-hatched larva ; A eye ; Bl blood corpuscles ; _D k com- mencing intestine ; !;,! endo^uie. ; F opening of cerebral vesicle into mouth ; Ob cerebral vesicle with otolith projecting from its floor ; Hp papilla for attachment ; Kl one of the atrial invaginations ; O mouth ; I'll pharynx ; R g anterior swelling of post-cerebral region of nerve tube ; lim posterior part of nerve tube. / Two days larva, only the anterior part of the tail is represented ; 1 ks, 2 ks branchial stigmata Bb blood shins between them ; D in- testine. elongates in the future antero-posterior axis and the blastopore comes to lie on the posterior end of the dorsal surface (Fig. 15 b). A flat median groove of ectoderm appears along the DEVELOPMENT. 21 dorsal side of the already bilaterally symmetrical embryo extending from the blastopore forwards. This groove, into the hind end of which the blastopore opens, is the first rudiment of the central nervous system. It is known as the medullary groove. Its edges project and form the medullary folds which grow round and close the narrow blastopore, and gradually unite with one another from this point forward in such a manner as to convert the groove into a canal, the walls of which separate from the external ectoderm and give rise to the central nervous system. This canal is the medullary canal ; behind it is shut off from the exterior, ,>ut communicates with the cavity of the gastrula by way of the blastopore which is now called the neurenteric canal ; while in front it remains open for some time, but eventually clos3S. Before these changes are completed, the medio-dorsal endoderm cells of that part of the gastric wall which immediately underlies the posterior part of the neural canal (Fig. 15, c and rf, Ch) become different from the remaining endoderm cells and constitute the first rudiment of the notochord. Meanwhile the latero-dorsal endoderm, on each side of the notochord. has separated off the mesoderm as a solid * plate of cells (Fig. 16, ms). Tnes3, in the trunk, later become converted into a mesenchyme, occupying the space which now makes its appearance be- tween the ectoderm and endoderm, and give rise tD the blood-corpuscles, musculature, genital, and excretory organs of the body, while the caudal part becomes the musculature of the tail. The distinction between the caudal and trunk region of the em- bryo is now very apparent. The notochord is confined to the tail. The caudal part of the enteron be- FlG 16 ._ Tran ^ e rse section through cnlirl /TTirr 1 ^ rl J7rJ ' \ flir irrli an mbryo of Clavetina (after V. SOlld (JUg. Ida, MO, ), tllOUgll Beneden and Julin. from Korschelt rprnainincr rnntinnnnn with flip and Heider). eft rudiment of noto- remaining COnilllUOUS Wlttl me chord ; ec ectoderm :? endoderm ; inprlnllirv tnhp rnnnrl flip Viirirl pnrl ms rudiment of mesoderm; mia lllar y medullary folds ; n medullary plate. of the notochord. Eventually it disappears, apparently giving rise to blood corpuscles. * According to van Benedeii and Julin there is at first an enteric pro- longation in the front part of this. 22 PHYLUM TUNIC AT A (UROCHORDA). The anterior part of the enteron dilates and constitutes the rudiment of the pharynx (Fig. 15, Ph], from the hind end of which the intestine is developed as an outgrowth (D). In the further course of development the tail becomes greatly elongated and curved ventrally on the trunk (Fig. 15 e), and some ectodermai papillae are formed at the anterior end of the body for the future attachment of the larva (Hp). The anterior end of the medullary tube becomes dilated into the cerebral vesicle, in the wall of which two sensory structures -the eye and auditory organ are developed (Fig. 15 /, Gb). The part of the medullary tube immediately behind this acquires thickened walls and is called the trunk ganglion (Ry). This is followed by the narrow caudal extension of the tube which ultimately disappears. The mouth is formed as a perforation on the dorsal surface of the front end and the cerebral vesicle acquires an opening into the anterior part of the alimentary canal (Fig. 15 /, o). The atrium arises as two dorso-lateral invaginations of ectoderm (kl) into the left of which the anus opens. The atrial invaginations spread laterally round the pharynx but remain separate ventrally ; dorsally they coalesce, so that the single atrial aperture arises. The gill-slits or branchial stigmata arise as a pair of perforations of the wall separating the atrial cavity from the pharynx. They subsequently become more numerous, partly by formation of new I ic :f orations and partly by division of those already existing. The endostyle arises as a groove on the anterior (Willey) or antero-ventral wall of the pharynx, but subsequently, as a con- sequence of the rotation which the body undergoes at the meta- morphosis (Fig. 18). becomes entirely ventral. The development of the t /limr/Httm. which as we have seen above is probably to be regarded as the coelom of the animal, seems to take place in different ways in different forms. In CioiKi it arises from the hind end of the pharynx as two diverticula which remain separate throughout life and invest the digestive viscera like a perivisceral cavity. In ('In r, linn (Fig. 17, ep) it appeals to arise as a single diverticulum of the pharynx between the end of the endostyle and the oesophagus, the front end of which becomes double. In some cases however it is apparently delainiiiated from the phaiynx and is at first solid. The peri- cardium in a great number of eases, if not universally, is nipped of) from 1 he epieardium either from its posterior unpaired portion DEVELOPMENT OF GANGLION AND NEURAL GLAND. 23 or from one of its anterior limbs. The heart arises as an in- vagination of the dorsal wall of the pericardium (Fig. 17, pc). The tunic is laid down as a cuticular excretion of the ectoderm, and the larva is hatched. After a few hours of free life, during which it swims by the movement of its tail, it attaches itself by the three anterior ectodermal papillae and the tail shrinks up to a stump and atrophies (Fig. 18). The cerebral vesicle largely breaks down. The dorsal part of it which communicates with the pharynx by the new neuropore persists and gives rise to the neural gland. The dorsal wall of this thickens and forms the -_j .___ ur . -,^* " e-y pp ft, FIG. 17. Left side view of a Clarelina embryo (after Seeliger from Korschelt and Haider* uu eye ; ch notochord ; e atrial aperture ; ed rectum ; ep epicardial outgrowth of pharynx es endostyle ; / folding of the body surface in anticipation of the rotation that take place after fixation ; fg duct of neural gland ; h adhering papillae ; i mouth ; ks gill slits ; m stomach ; mz muscle cells of the tail ; oe oesophagus ; ot auditory organ ; p atrial cavity ; pc pericardium ; s larval tail ; sb cerebral vesicle. ganglion of the adult. If this account is correct it follows that the neural gland is a part of the original cerebral vesicle. It has been compared to the hypophysis (pituitary body) of the Vertebrata ; but it differs from this in the fact that it is, in its origin, actually a part of the embryonic brain, which the pit- uitary body never is. The trunk ganglion becomes solid and persists as the visceral nerve. At the same time the fixed larva undergoes a peculiar change which is illustrated in the annexed diagram (Fig. 18). The mouth which was close to the point of attachment becomes displaced to the opposite end, and the original hind end becomes placed close to the point of attach- ment. In a few monascidians (Culeolus, etc.) this change does 24 PHYLUM TUMCATA (UROOHORDA). FIG. 18. Diagram illustrating the metamorphosis ot the larva of Clavelitia during and after fixation (f ri 1 1 Eorschell ami Eeider, after Seeliger). A Free-swi inji larva. B Larva just attached. C Older metamorphosed stage, ch notochord ; e atrial aperture ; ed in- testine ; c/( r|ijcardium ; cs cndnsU lr ; / i-ctudcrmal fold ; ft duct of neural gland ; g ganglion ; /< heart ; hp adhesive papillae ; i mouth ; A-.S gill-slits ; // atrial cavity ; r trunk-ganglion ; v part j tii. 11 wall of stolon ; sb cerebral vesicle ; . cast cell 11 1< e ti>>m' of tail ; st stolo prolifer ; r larv.-il tail degenerating. ASCIDIACEA. 25 not take place, the animal remaining throughout life attached by a stalk which arises close to the mouth (Fig. 19, 2). The following classification has been adopted : Order 1. Ascidiacea. Tribe 1. Ascidiae simplices, Monaseidia. ,, 2. Ascidiae compositae, Synascidia. ,, 3. Ascidiae salpaeformes, Ascidiae Luciae. Order 2. Thaliacea. Sub-order 1. Hemimyaria, Salpida. 2. Cyclomyaria, Doliolida. Order 3. Appendiculariae (Perennichordata, Larvacea, Cope- lata). Order 1. ASCIDIACEA (TETHYODEA). Fixed or free-sivimming , solitary or colonial Tunicnta, which in the adult are never provided with a tail and have no trace of a notochord. The free- swimming forms are colonies and the solitary forms are fixed. The test is permanent and well developed ; as a rule it increases with the age of the animal. The musculature of the mantle is in the form of an irregular network, there being no regular circu- lar bands. The pharynx is large and well developed. Its walls are perforated by numerous apertures opening into a single atrial (peribranchial) cavity, into which the anus opens and which communicates with the exterior by an atrial aperture. The colonial forms reproduce by gemmation, and in most the sexually produced embryo develops into a tailed larva. The order is divided into three groups, the Ascidiae Simplices, the Ascidiae Compositae and the Ascidiae Salpaeformes. These three groups can only be regarded as tribes, for they are closely interrelated, and are distinguished, the two first oy the presence or absence of the power of budding and the last by being free-swimming. If other and more general anatomical characters w r ere taken, quite a different grouping of the families would be obtained, and there can be no question that some of the families of synascidians are more closely related to cer- tain families of the monascidians than to each other. But if in the grouping of the families account were taken of these facts it is difficult to say that a more natural system of classification would be obtained ; cross-relation- ships between the groups constituted would still exist, and would by many be considered to be of sufficient importance to justify a different 26 PHYLUM TUNIC AT A (UROCHORDA) grouping, so that it seems best to adopt a plan which has at least the merit of being simple and easy of application. Tribe 1. ASCIDIAE SIMPLICES (MONASCIDIA). Solitary usually fixed forms, incapable of budding in the adult state ; with large 'pharynx, and numerous branchial stigmata. These are the typical sea-squirts. They are solitary forms usually of considerable size and attached by their tunics to rocks and sea-weeds. A few however, e.g. Molgula, are not attached save to grains of sand by processes of the test, and in a few stalked forms the stalk arises from near the mouth (Boltenia. Culeolus, Fig. 19). The tunic is usually somewhat thick and opaque and may have a cartilaginous consistency. When touched they frequently discharge two streams of water which proceed from the two openings, the one. the mouth or inhalent opening which is terminal and at the free end, the other the cloacal or atrial aperture, which is placed on the dorsal surface a little distance from the free end. The protective covering of mud or sand is generally associated with fibrous processes of the test (Molgulidae, Polycarpa molguloides. Ascidia conchilega), but in some cases the sand adheres directly to the test. In Culeolus, Fungal us and Bath your us. abyssal forms belonging to different sub-families of the Cynthiidae, the pharynx presents the remarkable modification of being without the fine longitu- dinal bars (Fig. 20). The dorsal lamina varies from the con- dition of a membrane to that of a series of languets ; an inter- mediate condition of a toothed membrane being found. The viscera are placed in the body- wall at the level of the hinder part of the pharynx (except in Ciona), usually on the left side. They project into the atrial cavity, sometimes being attached to its wall by a mesentery. The atrial cavity is always traversed by vascular strands passing from the pharynx wall to the mantle \vall. The vascular system is well developed, the sinuses sometimes having the aspect of vessels. There is usually a tailed larva. Fain. I. Ascidiidae. Usually sessile, rarely pedunculate d : month usually with S lohes, atrial aperture \\ it li <> : pharyngcal wall without folds, \\itli internal Inngit u< linal hars hearing papillae; stigmata straight 01 enr\ ec I ; tentacles unl>raiiclu > d : gonads pi act d in t he intcst inal loop. ( 'iona approximate^ t<> the ( 'la \ elinidae l>\ the position c t' its \ is< era ;iiid the presence of an epieanlillin (pp. !.">. '.{'). MONASCIDIA. 27 Pharynx with internal longitudinal bars and straight stn/ma/a, viscera on the left side of pharynx, dorsal lamina as a membrane. Ascidia L. (Figs. 2, 12), test soft, A. mentula O.F.M. : Phallusia Sav. , P. mammillata Cuv., pharynx recurved on itself posteriorly on the left side, Seas of E u r . ; Pachylaena Herd. , test hard. P h a r ij n x a s above, dorsal lamina as languets. dona F 1 e m . (Figs. 1, 3, 1, GA, 11, 13, 14), viscera posterior to phar- ynx, tunic gelatin- ous, renal cells grouped near tl it- generative orifice, C. intestinalis L. most seas. Rhodo- oma E h r e n b erg (Chev reulius L.- Duth.) (Fig. ID), test folded at the anterior end jso as to form a kind of bivalve operculum covering the mouth and atrial opening, viscera on right side, Med. Abyssas- cidia Herd., at- tached by ventral surface, apertures far apart, deep-sea, 2,000 to 2,(KlOfms., viscera on either right or left side of FIG. 19. 1. Holgula (Anurella) so.'enota. from the right side (nat. nharvnv size). ^1 atrial pore, B mouth, at the end of papilliform prominences Ti, Ti ; ai, r alimentary canal: /s, f* liver ; r reproductive gland?. Pharynx with 2. Boltenia wiformis x j. 3. Rliutlnxnitin (Chevreulius) callense x 3 ; internallongiludinal a moutll: n 8 an 8 lion : o atrial aperture ; y operculum (from Perrier). bars, curved stigmata and viscera on right side of pharynx, bars. inata straight ; tentacles simple or branched ; intestine on the left side, only slightly or not at all attached to the mantle ; goiiads on the inner surface of the mantle, either on. both sides or on one only. In the genera. Culeolus, Ci/xtint/i. 1-1. J''l<;. % J1.- Filanii nt> nt tin' tuiiii- i grains nf >aml arr attarln'il (from M<:/:/u/ii riixi-nritn. shnvvin 1 -' tin 1 - an IOHL; stalk, fioltona Sav. (Fig. 19), pharynx wall \\ith the line longitudinal bars, stalk arises on ventral surface near mouth. X. Atlantic, .\nstial.i-ia. Arctic. ('t ra Herd., stigmata not in spirals, lobes of the apertures plain, abyssal, test without processes, stalked. Ker- iHielen. Ctenicella L.-Duthiers, apertures laciniated. Aled. Eiigyra Alder and Hancock, pharynx not folded, but with saccular divert i- cula in longitudinal rows, gonads impaired, most seas. Paramolr/ula Traustedt, pharynx without folds, stigmata spirally coiled and in in- fundibula. gonads paired. Bostrichobrunrlnix Traustedt: (l(mni*ll Coelocormus) colonial forms, tin individuals of n-li i<-li n /it-oil nci by gemmation and are embeddnl in common test (except Claveli>iir in a special incubatory pouch which they do not leave until they have developed into the tailed larva. Vain. 1. Botryllidae. <'olon\ usuallv thin and encrusting, sometime- in the form of thick flesh \ masses; zooids arranged in systems (Fig. 23) circular or elliptical or in branching lines, the x.ooids of a >\stem openinu int 1 1 one CO! on cloaca ; C moil cloaca I openings (list ii let . usually lobed : /.odi the sui-face nf the colons- : intestine on the left side of the posterior part of 'I' In folds arc really longitudinal rows of saccular project ions ( inhnnli - i> nli i) of t lie phar\ ns_eul \\ all into t lie peril ira nchial ca\ il \ . i' I'i/.oii, Histoire dc la Uastonenese chc/. les Hotryllidcs. Ann. Sd. \n/. (7). 14, 1SII3. /S4. SYNASCIDIA. 31 the pharynx ; test usually soft, with numerous vessels ending in terminal knobs and joined to the body of each zooid at two points ; pharynx large with 3 internal longitudinal bars on each side and numerous stigmata ; dorsal lamina as a membrane ; tentacles simple, not more than 1 6 ; gonads on both sides in the mantle (except Symplegma) ; gemmation lateral, from the bodies of the zooids* ; the neural gland is dorsal to the ganglion in Botryllus ; the stomach has an hepatic caecum. The budding of Botryllus differs from that of other synascidians in that the endoderm does not participate ; the bud being formed as an outgrowth of the atrial cavity and consisting only of outer ectoderm of the body, ectodermal lining of the atrial cavity and interposed mesoderm (Fig. 24). The process begins in the larva before hatching, as a pair of ventral out- growths of the atrial cavity. After fixation of Hie larva the left of these atrophies and the right alone de- velops. The zooid produced from it gives rise to two buds by a process which is described below and is re- peated in all the subsequent budding. The zooids of the colony thus in- crease in geometrical ratio, but in all cases when the buds are developed, the form which has produced them dies ; thus the fixed larva dies when it has produced its bud, and the latter dies when it has developed and produced its two buds. The zooids produced arrange themselves so that their atrial cavities are tiirned to- wards one another and open into the common cloaca which is a depression of the surface of the colony as in Pyrosoma. A system of zooids thus arises. The number of zooids in a system is limited ; when the limit is reached, of the two buds which each zooid produces one atrophies, and the other, instead of taking up its Fl - 23. Botryllus violaceus (liter M. Ed- wards from Clans). mouth ; A opening position in the system, moves away O f common cloaca of a system. from it and becomes the centre of a new system in the same colony. The budded zooids do not separate from the parent as in most synascidian colonies. It is true that the pharynx of the bud loses its connection with the atrium of the parent, but the outer c todermal connection persists and becomes an elongated and slender tube by which the vascular systems of the parent and bud remain in continuity. On the atrophy of the parent zooid the two tubes which connected her to her offspring become directly continuous owing to the fact that though the internal organs of the parent break down, its ectoderm persists. As stated above the buds are formed as diverticula of the atrial cavity on its ventral side (Fig. 24). They form hollow vesicles, the cavity of each of which divides into two ; one of these becomes the pharynx and th<- * In Sarcobotrylloides Herdman describes stolonial budding from the vessels of the test. PHYLUM TTJNICATA (UROCHORDA). aie other, saddle-shaped, becomes the atrium of the bud. The organs formed in the usual way, the pericardium and intestine as outgrowths oi the pharynx and the from The Eel -1" FIG. 24. Diagram <>i a ilorsal view of a young oozoid or of a bud of Botryllus violaeeus (after Pizon, from Perrier). B fu- ture mouth ; bl rudiment of new blasto-zooid of one side ; c.vib vibratile tube ; ect ectoderm ; gh gonad ; i intestine ; perb peribranchial sac (atrial cavity) ; pv epicardical di- verticulum ; vb pharynx ; Vcl cloacal part of future atrial cavity. nervous system the endoderm. original connection re- mains throughout life as the vascular ecto- dermal tube referred to above. Botryllus Gartner and Pallas (Fig. 23), colony thin, encrust- ing, systems circular, gonads paired, placed laterally, littoral, Eur., Med., N. Amer. Poly- cyclus Lamarck, as last, but colony thick and fleshy, Eur. and Med. Botrylloides H. Milne - Edw., colony thin encrusting, sys- tems elongated or branched irregularly, gonads paired, lateral. Sarcobotrylloides v. Symplegma Herdman, Drasche, like *last, but colony thick and fleshy. colony stalked, gonads unpaired, in intestinal loop, Bermuda. Fain. 2. Distomidae. Colony rounded or massive, rarely encrusting, either sessile or with long peduncle ; systems irregular, inconspicuous or absent, both mouth and atrium usually opening on the surface of the colony. Zooids divided into thorax and abdomen, and sometimes provided with long vas- cular ectodermal appendages ; test gelatinous or cartila- ginous, sometimes with non-stellate calcareous spicules ; pharynx without internal longitudinal bars ; dorsal lamina as langxiets ; gonads and heart in or alongside intestinal loop, spermatic vesicles numerous, vas de- ferens straight. In the Distomidae the budding is epicardial. The epieardiuni arises as a pair of diverticula of the pharynx, one v tissioii and eventually develop into new zooids.* In < 'olclla some of the buds, those placed in the deeper p irts of the stalk, have a store of reserve food material in their ectoderm. FIG. 25. Distaplia (after Herdman ironi Delage and Herouard). , op. c'lt. J uliii, 1 nt (Jong. Zoology, Leyden, JStKi. SYNASCIDIA. 33 Distaplia Delia Valle (Fig. 25), with a common cloaca and atrial languets, incubatory pouch as diverticulum of cloaca, colony sessile or only shortly peduncnlated ; Med., AtL, Ind. Ocean. Julinia Caiman, Antarctic. Dis- toma Gartner, with atrial siphon, without, pinnies in th t.p=t-, and incubatory pouch, colony sessile or only shortly pedunculated ; Med., Eur., Torres Straits. Heterotrema Fiedler, Cystodites v Drasche, Colella Herd., with an incubatory pouch and a well- marked peduncle, S. Ocean. Oxycornia v. Drasche. Chondro- stachys Macdonald, the zooids project beyond the test, Aust. Archi- distoma Garstang, the zooids arise from an incrusting basal layer, either singly or united in a common test, with- out common cloaca ; covered with sand ; may be compared to a Glavelina in which the tests of the zooids have fused in groups ; Plymouth. Finn. 3. Polyclini- dae. Colony usually massive, sometimes en- crusting, sometimes lobed or even stalked; systems of various shapes, sometimes ir- regular or absent ; common cloaca! aper- tures usually indistinct or absent ; zooids placed perpendicular- ly to surface, usually divided into thorax, abdomen, and post- abdomen (Fig. 26) ; mouth 6 or 8-lobed, atrial aperture often with atrial languet ; test sometimes stiffened by embedded sand grains; pharynx small with small stigmata, without internal longitudinal bars, with horizontal membranes ; tentacles small, few ; dorsal lamina as languets; gonads in post-abdomen, testis as spermatic sacs attached to FIG. 26. Diagrammatic section through a portion of a Poly- clinid colony (after Delage and Herouard). 1 openum of common cloaca ; 2 atrial openings ; 3 ganglion ; 4 mouths ; 5 tentacles; 6 peripharyngeal band ; 7 endostyle ; 8 languets; y anus; 10 male; 11 female generative opening; 12 epic;inli.i! tubes; 13 stomach; 14 pyloric gland ; 15 ovary ; Ifi tr.-ti- : 17 epicardium ; IS dorsal; 19 ventral limb of pericardium ; 20 posterior bifurcation of epicardium ; 21 heart. Z III D 34 PHYLUM TUNICATA (UROCHOBDA). large vas deferens ; gemmation from post-abdomen which lias the heart at its extremity. The post-abdomen contains the epicardium, which divides it into a right and left portion, the heart and pericardium, and in its upper part the gonads (Fig. 26). The pericardium (19) is U-shaped and contains the heart (also U-shaped) in its posterior part. The epicardium (17) bifurcates in front into two tubes which abut against the pharynx but do not open into it in the adult ; posteriorly it bifurcates so as to grasp the bend of the pericardium (20). Budding is effected by the separation of the post-abdomen from the abdomen and its fission into pieces (Fig. 27). The portion of the heart and pericardium in eacli piece atrophies and the epicardium develops in the usual way into pharynx, atrium and intes- tine of the bud. In Amaroucium, and possibly others, budding takes place in the summer ; towards winter this ceases and sexual organs are formed. The budding zooid regenerates a new post-abdomen. This method of bud- ding is clearly of the stolonic kind, such as we find in Clavelina and Distomidae, differing only by the presence of the heart in the stolon. With intestinal loop twisted and generally smooth-walled stomach. Polyclinum Sav. (Fig. 26). Systems simple, stomach smooth-walled, most seas. (Jlossophorum Lahille, hori- zontal membranes of pharynx denticulated, each colony of a single system ; Med. Aptogaster Herd., Polyclinoides v. Drasche, Aplidiopsis Lahille, with non-twisted intes- tine and smooth-walled stomach ; Med., Atl. Pharyngo- dictyon Herd., pharynx with simple meshwork of rectangular meshes, fine longitudinal bars supposed to be ab- sent (see p. 11); Antarctic, 1,600 fms. ; Tylobranchion Herd. Intestinal loop not twisted, wall of stomach grooved or areolated. Ajiiliilinni Sav., without languets, most seas. Psammapilidium Herd. Amaroucium M.-Edw. (Amaroecium) (Fig. 27), atrial aperture with a laii- guet, post-abdomen not separated off by constriction, most seas. Sigillina Sa\ ., Fragarium Giard and Fragaroides Maur.. with 8 buccal lobes, Med. Sid- nyum Sav.. <> buccal lobes, areolated stomach ; N\ arid S. Atl. Morchellium Giard. Synoiciim Phillips, each colony forms a separate club-.shapod mass. Parascidia M.-Edw. Cir- cinalium Giard. Morchellioides Herd. Polyclinopsis Gottschaldt. Fain. 4. Didemnidae. Colony usually thin and encrusting, rarely thick, never stalked : systems irregular, inconspicuous or absent, common cloacal apertures usually conspicuous : zooids placed perpendicularly or "Uiquely to surface, divided into tlmr.ix ami abdomen; mouth 6-lobed, atrial opening plain or with atrial languet ; test usually with stellate cal- careous spicules ; ectodermal processes well developed, with muscles which FIG. 27. Amaroucium -hi '\\ing seg- mented post-abdomen, JT, >j lunU. k anterior swollen end of epicanlium in the bud (after Kowalevsky >. SYNASCIDIA. 35 act as retractors ; pharynx small, with 3 or 4 (rarely (i) rows of stigmata ; gonads alongside intestinal loop, ova large, single testis round which vas deferens is coiled spirally ; gemmation from pyloric region, thorax and abdomen formed from separate buds ; larval gemmation feeble. The budding* (Fig. 28) in the Didemnidae is very remarkable. The buds appear to arise in two separate portions from distinct parts of the body ; the one of these, known as the thoracic bud (b), arises as a diverti- culum of the outer wall of the lower end of the atrial cavity, on the right side opposite the stomach ; the other, called the abdominal bud (c), is a diverticulum of the oesophagus which also projects towards the right side. The thoracic bud is formed of outer ectoderm, ectodermal lining of the atrial cavity and interposed mesoderm (like the bud of the Botryllidae) ; it gives rise to the thoracic portion of a new zooid, i.e. to the pharynx, atrial cavity, rectum and a portion of the oesophagus (Fig. 28, -B). It separates from its place of origin and acquires new relations : these are as follows : the mouth opens on the surface of the colony, the atrial aper- 28. Three successive stages in the budding of Didemnuin (from Delage and Herouard, 4 . after D. Valle). a parent ; b pharyngeal bud ; c abdominal bud ; cl atrial aperture ; O oeso- phagus ; r rectum ; s mouth of bud. ture into the common cloaca of the colony, the portion of oesophagus joins the oesophagus and the rectum j oins the rectum of the parent. The abdominal bud meanwhile has formed itself into a loop connected at both ends with the oesophagus of the parent ; one end however separates itself from the oesophagus, joins the rectum close to the point of union of the latter with the rectum of the thoracic bud ; and the whole loop forms itself into a new oesophagus, stomach and intestine. A new heart and pericardium are formed in the abdominal bud. Later when the new pharynx and new intestine have developed so as to be equal in size to those of the parent, they become detached from its oesophagus and rectum and join up in such a way that the oesophagus and rectum of the abdominal bud become continuous with the oesophagus and rectum of the thoracic bud. Such is this extraordinary phenomenon. The details are not yet fully worked out, and we await future observations for a more complete understanding of it. It sometimes happens that the two parts of the bud are not equally * Delia Valle, Mem. R. Accad. Lincei (3), 10, 1881, and Arch. Ital. Bio- logie, 3. 1883. Caulery, Comp. Rend., 1895-97. Salensky, Naples Mit., U, 1895. 36 PHYLUM TUXICATA (UROCHORDA). 13 developed, or one part alone may develop, so that zooids may be found with two sets of intestines. In such cases the old intestines disappear and the phenomenon may be regarded as one of regeneration. Didemnum Sav., colony thick and fleshy, pharynx with three rows of stigmata, most seas. Diilrioides\'. Drasche. Sarcodidemnoides Oka. Leptoclnnnn M.-Edw., colony thin and encrusting, 4 rows of stig- mata. Tetradidenmum Delia Valle. Poli/- fiijiirrutoii Xott. Kuiwlitim Sav.. rows of stigmata, Med. and Red Sea. Hy purr/on Sollas, with faecal pellets included in the test, Malay Peninsula. Fain. 5. Diplosomidae. Colony usually thin, transparent and encrusting, usually without spicules in the test, pharynx with 4 rows of stigmata ; body divided into thorax and abdomen ; vas deferens not spirally coiled ; gemmation* as in the Didemnidae : the larva produces a well-developed bud be- fore fixation. Diplosoma Macdonald. Med., Atl., Pac., Austr. Diplosomoides Herd., with spicules, Med. Brvrixtt Ilium Jourdain. As- tellium Giard. Pseudodidemnum Giard. Fain. f>. Coelocormidae. Colony massive, deeply concave on upper surface, not at- tached ; zooids large, scattered all over the surface, mouth .Vlobed : test soft with cal- careous spicules near surface of colony ; pharynx large, dorsal lamina as languets : intestine extending behind pharynx but not forming distinct abdomen ; testis as pyriform vesicles which join a spirally coiled vas de- ferens. Coeloconnus Herdman, 8. Atlantic, GOO fms. : . $$ UIJSJ IS' - G FIG. 29. A blastozooite of Clavelina, side view, diagrammatic (after Delage and Herouard), 7 atrial aperture ; 2 subneural gland ; 3 inoutli; 4, 5, epicardial tube ; G heart ; 7 stolon ; S sto- lonic septum ; 9 stomach ; in ovary ; // test is ; 12 oesophagus ; 13 anus. the recent account of Pizon (('<>n> i>t<- Ri'iidus, 137, 1903, p. 750) which is somewhat condensed, the budding appears to be as follows : The larva buds before fixation. Of the two individuals so formed, O 1 and B l , the oozoid O 1 or zooid into which the larva itself develops buds a new thorax (pharynx, oesophagus and rectum) and so becomes bithorurir. After twenty tour hours the old thorax degenerates, while its abdomen V persists and retains its connection with the budded thorax. There is thus formed a new individual O-. which in its turn produces two buds; from one of 1 hese proceeds a ne\\ t borax O 3 from the other a new abdomen V-. After about twenty-four hours these separate in such a way that the new thorax O :! takes I he old abdomen V, \\ hile the thorax of O 2 keeps the new abdomen V-. ()-' having thus acquired a new abdomen buds a new thorax ()' and then after twenty-four hours loses its old thorax. A ne\\ y.ooid \\ ith t he old a I id omen V- and t he new thorax O 1 is thus formed. ( )' L">es through the same scries of changes as those described for O 2 ; i.e. it first doubles itself and nixes rise to t \\ o new x.ooids, of which that with the old thorax O 3 becomes hithoracic ;md then loses its old thor-ix. The yooid B 1 produced by the free |ar\ a behaves like O 2 . SYNASCIDIA. 37 Fam. 7. Polystyelidae. Colonies massive or encrusting, rarely stalked or formed of small masses connected by stolons, without common cloacal cavities ; both apertures 4-lobed, opening directly to the exterior ; pharynx large, with strong internal longitudinal bars ; tentacles small, numerous ; dorsal lamina as membrane ; intestine alongside pharynx, rarely extend- ing behind it ; gonads as polycarps in mantle projecting into atrium ; gem- mation doubtful, probably from vascular stolons ; in Goodsiria it is stated to be pallial (Ritter). It is not certain that all the genera here included have the power of budding, and it is possible that some of them are close aggregations of simple Ascidians. By their structure they approach Poly- carpn (Styelinae) of that group. Goodsiria Cunningham, Chorizocormu? Herd., Oculinaria Gray, Thylacium Cams, Poly sty ela Giard, Synstyela Giard. Fam. 8. Clavelinidae. The zooids are not embedded in a common test but are attached to creeping stolons or to a stolonial mass from which new ascidiozooids are formed by gemmation ; test xisually gelatinous, thin and transparent ; pharynx often without internal longitudinal bars, which are without papillae ; tentacles simple, dorsal lamina as languets ; intestine usually behind pharynx as abdomen ; gonads in intestinal loop. This family comes closest to the genus Ciona of the simple Ascidians, with which groxip it is often united. The stolon of* the oozoiteof Clavelina .(form produced from the egg) is divided into two parts by a septum, the stolonic septum. This is a collapsed continuation of the epicardium and therefore contains endoderm. At the free end of the stolon the septum ceases so that the cavi- ties (blood spaces) on either side of it are in communication. At the other end the septum is continuous with the hind end of the single epicardial tube (in the blastozooite with the hind end of the pericardium, Fig. 29), the pos- terior part of which is applied to the dorsal side of the pericardium, while the front end forks to open by two openings into the hind end of the pharynx just ventral to the oesophagus. In the blastozooites or forms which ha,ve been produced by budding (Fig. 29), the arrangement is the same except that the stolonic septum is connected with the pericardium and not with the epicardium ; but this is not a matter of any great importance, inasmuch as the pericardium develops from the hind end of the epicardium, with which it remains in close contact. The point is that the stolonic septum is an endodermal structure continuous with or developed from the pharyngeal wall. The form produced from the egg remains asexual, the zooids (blastozooites) which are budded from the stolon of this become sexual (Fig. 29). The budding takes place in this way. The stolon produces on its upper side a small diverticulum (Fig. 31), into which the septum sends a hollow endodermal prolongation. This FIG. 30. Diagrammatic transverse section through the posterior part of a bud of Clavelina (from Kor- schelt and Heider). ep epicardium ; h heart ; i intestine ; m stomach ; n viscera] nerve cord ; pc pericardial vesicle. ' Kowalevsky, Sur le bourgeonnement du Perophorn listcri, Rev. Sci. Nat. Monlpellier, 1874, and Ueb. d. Knospung d. Ascidii-!). Arch./. JI//A-. Anat., 10, 1874. 38 PHYLUM TUNIC ATA (UKOCHORDA). endoderrnal vesicle becomes constricted into two parts which remain connected by a narrow neck : the upper of these gives rise to the pharynx and atrial cavity, and the lower to the epicardium from which the pericardium subsequently proceeds. The digestive tube arises from the pharynx in the ordinary way, while the central nervous system appears to develop from the pharyngealendodenn,* as it does also in the buds of Distaplia t and Botryllus i and probably in those of all Ascidiae Compositae. It would thus appear that in the Ascidiae Compositae the atrial cavity and the central nervous system,which in the embryo develop from the ectoderm, in the bud de- velop from the endoderm. This contrast is highly remarkable, and due weight will have to be given to it, when we are considering in the General Part, the theory of the embryonic layers. Clavelina Sav., stolons delicate and branched, body divided into thorax and abdomen, without pe- duncle, pharynx without inner longitudinal bars ; 2-3 cm. in length ; Eur. and Med. Podocla- i-illn Herd., with peduncle, Arctic. Australia. Stereoclavetta Herd., stolons and hind end of body in a common mass of test, Atlantic, Australia. Pycnw:lurdla Garstang, (.I.M.B.A. (-2), 2, 1891), similar to preceding, Plymouth. Perophora Lister, without abdomen, short peduncle, pharynx with 4 rows of stigmata, transverse bars of pharynx with papillae which branch and may form imperfect internal longitudi- nal liars (Fig. 8), zooids 3-5 mm. ; All.. ,Med.. X. Amer., Austr. Pero- l>lif>r<>/iN/ft Lahille, with 15 or 16 rows of stigmata, Med. Ecteina- ni-iilin Herd., without abdomen. test without blood-vessels, pharynx with internal longitudinal bars, most seas. Slttftcn'a v. Ben. Dia- zoiiii Sav.. many zooids united by tests posteriorly to form a large colony, abdomen present and embedded in con i test. phar\n\ with internal longitudinal bars, zooids 3-6 cm. FIG. 31. Portion of a proliferating .-toloii > of i'l'm/ilinrii latti-r l\n\\:ili-\-k\ troin K..r- l"-!l .-MM! llchli'l |. 61 toili-nil ; i')l cll- doderiu ; kn lm KM. Ritter, Jowrw. 1896. t Kowalcvsky. Arch. Mil:. Ann/.. Id, 1x71. i Hjort. Znnl. Anz., 10, 18!)2, p. 32s. PYROSOMA. 39 Eur. and Med. Rhopalaea Philippi, zooids few, not united by tests, stolons as foliaceous expansions ; 5-12 cm. ; Med. Rhopalopsis Herd. Tribe 3. ASCIDIAE SALPAEFORMES (ASCIDIAE LUCIAE). Free-swimming pelagic colonial Ascidians. The colonies have the form of hollow cylinders, closed at one end, open at the other, and slightly tapering towards the closed end. The closed end is rounded ; the open end is flat and its edges project inwards to form a diaphragm (Fig. 32). The cavity of the cylinder is the common cloaca of the colony, and its opening, which can vary in size, is the common cloacal opening. The test is trans- parent and gelatinous ; it bears on its outer surface a number of projections (Fig. 31 bis), while the inner surface is perfectly smooth. The zooids are elon- gated antero-posteriorly and placed in the thickness of the wall of the cylinder in a single layer at right angles to the surface. Their mouths open on the outer surface, each at the base of one of the test processes ; and their atrial apertures are at the opposite end and open into the common cloaca. The colonies vary in length from a few inches to four feet (Pyrosoma spinosum). They float horizontally in the sea, and have a slight power of movement with the closed end forward. The movement appears to be caused by feeble longi- tudinal Contractions of the Wall of the FW. 31 tns. Pyrosoma elegans, x i (from Perrier). a cloacal hind end of the cylinder. They are opening ; a buccai appendages J of the zooids. phosphorescent. When they are at rest the light is very feeble, but on stimulation by contact or other means, the light, passing through a red and green stage, be- comes white. Moseley * states that he traced his name on the * Notes by a Naturalist on the " Challenger," London, 1879, p. 574. 40 PHYLUM TUNICATA (UROCHORDA). surface of a large specimen with his finger, and describes how "in a few seconds his name came out in letters of fire." The light is emitted by two groups of cells in the region of the mouth. There is a single genus Pyrosoma Peron. The zooids usually lie with their ventral surfaces towards the closed end of the colony, but in P. elegans they have the opposite position. There is a buccal and atrial sphincter muscle and a few delicate bands of muscular fibre in the body wall. The test contains cellular elements and some fibrous bands, and the zooids near the common cloacal opening give off two tubular prolongations of the dorsal body-wall containing muscular fibres, which appear to have the function of acting on the dia- phragm. The structure of the zooids is well shown in Fig. 33. The viscera are behind the pharynx, and the anus opens into the posterior part of the atrial cavity. The atrial cavity sends forwards two diverticula, one on each side of the pharynx, but t hese are not continuous with one another on the dorsal side of the pharynx, as in other Ascidians. The lateral walls of the pharynx consist of from 20 to 50 transverse bars, crossed by from 15 to 30 longitudinal bars. The stigmata are quadrangular, and the longitudinal bars are supposed to be inter nal, fine longitudinal bars being absent, as in Culeolus, etc., among simple Ascidians and in Pharyngodictyon among the Synascicliaiis. however if this interpretation is correct (see p. 12). There is the N FIG. 32. Diagram of a section through a Py- rosoma colony (after Delage and Herouard). 1 buccal appendage ; 2 common cloaca ; 3 zooids ; 4 diaphragm ; 5 cloacal opening. It is doubtful of of Wh A usual row tentacles which the ven- tral median is the longest (Fig. M). The dorsal 1 am i 11 a has the form of I _' I II II U' II e t s which slio\\ no l-i'lat ion to the transverse liars. Tin- intest ine is curved, there is a pylorie gland ra 1 1 1 i f y i n g on the intest ine. The heart is placed a little behind the end of the endostyle (C). The gonada are placed in the body wall on the ventral side behind the intes- The ovary contains one ovum which ripens before the test is. The Br C T End FIG 33 \ zooi'l df I'l/nixiiimi (from Pt-rrior). A e\h;ileut aperture : 4/anus Br pharynx . Cheart ; Endendostyle : N nerve ganglion ; ,0 mouth Or civiiry ; >' *t.i].m ; 7' test is ; \\'l, p.-ripharyngeal ring. ASCIDIAE SALPAEFORMES. 41 ft budding stolon (st) is a projection of the ventral body wall and contains a diverticulum of the pharynx which comes off just behind the endostyle, a genital cord, and a prolongation of the pericardium which latter structure is said not to give rise to any organ in the bud. The ganglion carries on its ventral side a pigmented organ which is interpreted as an eye, and the phosphorescent organs (Fig. 33) are paired (right and left) masses of cells containing a fatty substance at the level of the peripharyngeal band in the blood space there found. The development* of Pyrosoma is very remarkable. The ovum is large and full of yolk. The cleavage is partial and gives rise to a disc of cells lying upon the yolk. There is no tailed stage and the development is very different from that of other Ascidians. It takes place within the egg- follicle and leads to the formation of an imperfect individual called by Huxley the cyathozooid. This while still in the maternal tissues produces a sto- lon (Fig. 34), which immediately undergoes imperfect transverse fission into four parts. Each of these develops into a typical Ascidian zooid, called by Hux- ley the ascidiozooid (Fig. 35). The four ascidiozooids arrange p la . 34. Diagram of a Pyrosoma embryo showing the devel- themselves equatori- P m g cyathozooid lying on the yolk with some of its organs developed, and the stolon st with commencing fission. ally round the parent ,/ enteron ; do yolk ; e atrial aperture ; g ganglion ; h heart ; cvathozooid (Fig 35) fi ciliated pit, of cyathozooid ; r edge of germinalfdisc grow- ing over the yolk of cyathozooid ; stfstolon. acquire a common cloaca and form the first individuals of a new colony. The cyathozooid now disappears, the ascidiozooids lose their primary connection, and the little colony passes into the atrium of the parent and thence to the ex- terior. Each of the zooids composing it possesses a ventral stolon which immediately begins to bud. As all the successively produced zooids have ventral stolons, the colony constantly increases in size and number of individuals until the limit of growth is reached. Budding by which the adult colony is formed, f Each of the f our primary ascidiozooids possesses a stolon at the hind end of its endostyle. This stolon is the pedicle which in an earlier stage connected the zooid to that next it in the chain, i.e. it is a part of the original stolon which does not develop into an ascidiozooid. It contains (Fig. 36) a diverticulum of the pharynx (d), two lateral bands of mesoderm, an unpaired cord of mesoderm called the genital cord (g) and the elaeoblast. Of these structures the mesodermal bands break up and are said not to give rise to organs, while the cord of genital mesoderm which is at first * Kowalevsky, Arch. f. Mik. Anat., 11, 1875, p. 597. t Seeliger, Jen. Zeitsch., 23, 1889, p. 595. 42 PHYLUM TUNIC ATA (UROCHORDA). en. FIG. 35. Two stages in the development of Pyrosoma (after Kowalevsky from Knrsi licit and Heider). In A the yolk of the ovum is partly surrounded by the cyathozooid c which has developed on it (see Fig. 34) ; in B the yolk is entirely enclosed by the cyathozooid. c cyatho- zooid, d its atrial pore, d its alimentary canal, do its yolk ; ec ectoderm ; el elaeoblast (a raesoblastic mass in the buds) ; en endostyle of ascidiozooid ; ft ciliated pit and g ganglion of cyathozooid ; i mouth, ks gill-slits of ascidiozooid ; I vascular space of cyathozooid; TO test ; ganglion of ascidiozooid ; p peribranchial (atrial) cavity of ascidiozooid : xn lateral nerve ; v endodermal canal connecting aseidiozooids ; z mesoblast cells. *?*- y FIG. 36. Three stages in the development of the proliferating stolon of Pyrosmim (frmu sclielt mid Ileider, after Seeliger). In C the division of the stolon into two buds, I and II. is indicated, d pharyngeal process (epicardium) ; ec ectoderm; es endostyle of parent; i genital (mesoderm) band ; kx first gill-slits ; m rudiment of digestive canal ; . rudiment eural tube. ASCIDIAE SALPAEFORMES. 43 continuous with the rudiment of the genital organs of the parent, gives rise to the peribranchial (atrial) cavity, to the central nervous system (n), to the gonads, and to the mesodermal structures of the budding zooid. The pharyngeal diverticulum becomes the pharynx and develops the intestine as an outgrowth (Fig. 37). It is clearly homo- logous with the epicardium found in many synascidians. The pericardium is developed as a vesicle formed of mesodermal tissue on the right side. It comes to lie on the dorsal side of the pharyngeal process which projects into the stolon of the fully formed zooid (Fig. 37 pc), and thus differs in FIG 37 A chain of three zooids in the budding of Pyrosoma (from Korschelt and Heider after Seeliger), I youngest, II middle and III oldest bud (nearly developed). D point at which the pharyngeal process of the parent enters ; d pharyngeal process (epicardium) ; dc rudiment of digestive canal ; dm elongated cell mass (? phosphorescent and haema poietic organs) ; e rudiment of atrial aperture ; eb elaeoblast ; EC ectoderm of parent ; oe e< derm of new stolon ; Es endostyle of parent ; es endostyle of bud ; fg ciliated pit ; g gangli h testis ; hd intestine ; hz heart ; i rudiment of mouth ; U atrium ; ks gill-clefts ; If internal longitudinal gill bars : Im phosphorescent organ ; m stomach ; ms genital strand ; nient of nervous system ; o ovary (egg follicle with egg) ; oe oesophagus ; p dotted dicating the boundary of the atrial cavity (peribranchial sac) ; pc pericardium ; rz langue tentacle rudiment ; v tube connecting the enteron of the first and second I position from the pericardium of other forms, which is placed ventral to the epicardium (see p. 15). The nervous system develops from the distal end of the genital cord (Fig. 36) which passes round the front end of the pharyngeal process on to the dorsal surface of the bud and becomes hol- lowed out to form a vesicle. The ganglion arises from the thickened dorsal wall of this vesicle, the cavity of which acquires a tubular communication with the pharynx and forms the subneural gland and duct (Fig. 37, fg} The elaeoblast also develops from the mesoderm, as paired masses which 44 PHYLUM TUN 1C ATA (UROCHORDA). join round the pharyngeal process (Fig. 37, eb). The test of the primary tetrazooid colony is said to be formed entirely by the cyathozooid, but in the subsequent budding each new zooid plays its part in adding to it. The stolon undergoes fission by transverse constrictions into buds. First one bud is marked off (Fig. 36 A), thenanother between the first and the parent (Fig. 36 C), and so on until five have been formed, the youngest bud being always next the parent. When five zooids have thus been marked off, the distal one has acquired full development and becomes de- tached and a new constriction is formed at the base of the stolon. The new zooids when detached pass round and take up their position near the common cloacal opening of the colony, where the youngest zooids lie. As they increase in age they become further and further removed from this opening by the interposition of the continually forming new zooids. The points in the above account which must be received with caution are those relating to the origin of the nervous system and peribranchial tubes. In the stolon of the cyathozooid these organs arise as ectoderm invaginations : in the stolons of the later zooids which are directly derived from that of the cyathozooid, they are said to arise from mesoderm. It is difficult to believe that this difference really exists. Order 2. THALIACEA.* Free-swimming solit'iry pelagic forms, which in the, adult are never provided with a tail or a notochord. Metagenesis always occurs, and the sexual forms ty pica/I ;/ r< main for some time con- nected to a process of the body of the asexual form by which they are budded. The mouth and atrial apertures are at opposite ends of the body. Tin- test is permanent and transparent, and closely adherent to the body. The musculature of the mantle is in the form of * Traustedt, Spolia atlantica, Bidray. til Kundskah om Salperne, Danskc \'iil. >'. /,s7.-. Sl.Tiit.. INS."). p. :53'.. Kowalevsky and Barrois, .Matcriaux pour s -i-vir a I'liistoire dc /' . 1 nrltiitin. .lonni. de V Anal. Phys., 19, lss:i. rijanin. l>i< Arten der Hutimiii Doliolion ///< <;<>!/< v. Xm/tfl, Leipzig, 1884. I'.arrois, Sui- le eycli<|\ic n<'- net i< pic ct la bourgeonnement de 1'Anchinie, Journ. de I'Anat. /'/;.:!. Seeli^ -r. Die Knospung der Salpcn, .l'ii. Zeitsch., 19, 188(i. Korotnet'f, Die Knospung der Anrliii/i'* Mit., 10, I s'.il . p. IST. Id.. Kml>r\ nlotiie der Solpa >'< mocratica, Z./.ir.Z.. .">!). p. _>!), L895. Id.. Tunicatenstudien, etc., JVapZes JWii., 11 and 12, 1895-6. Id., Zur Einb. v. Salpa runcinala Eusiformis, /..i.n-.K., (i'2. ISiKi. Brooks, The Genus Salpa. M> m. liinl. Lnli. .latin* ff'>/>kni.-< I r niveraity,2, I8!C?. Saleiisky, A scries of pa | HTS on the de\ eli .pi MCI it and I >u< l< IIML' "!' SH//IU, I'iih', Z.f.ir.T,., 17, l*7i, :'.(i. ls7s. Morph. Jahrb., :t. ls77, -'(), is'.i:?. \!'>_ ic von Salpa t'lisifonnis, \>>h. Senck. Oes. Fr. THALIACEA. 45 more or less complete circular bands, by the contraction of which locomotion is effected. The pharynx has either two large, or many small apertures leading into a single atrial cavity, which opens to the exterior by the atrial aperture. The anus opens into the atrial cavity. Alternation of sexual and asexual genera- tions (metagenesis) occurs in the life-history and may be com- plicated by polymorphism (Doliolidae). They are only occa- sional visitors to British shores. A few species of Salpa and of Doliolum have been taken in the seas around the Hebrides. They are divided into two groups, the Cydomyaria and the Hemimyaria, which seem to be sufficiently distinct to merit the rank of sub-orders. Sub-order 1. Hemimyaria (Salpida). Thaliacea in which the pharynx is reduced to its median dorsal and ventral (endostylar) walls, the lateral icalls being absent. The asexual form produces chains of sexual individuals, which give rise again to the asexual form. The muscular rings are usually incom- plete ventrally and a tailed larva is not developed. The salps are transparent pelagic organisms, coloured and opaque at one spot the nucleus, where the digestive organs and heart are placed. The test is soft and gelatinous and the mantle or body- wall closely adheres to its inner surface. The body is some- what elongated, with the mouth in front and the atrial aperture behind and slightly dorsal. The muscular tissue (Fig. 38) of the body-wall is arranged in hoop-like bands usually six to nine (there may be as many as twenty and as few as four) which are continuous dorsally, but usually not ventrally, except behind where they constitute the atrial sphincter. At the mouth they are modified in a different way, being prolonged forward into the lips. All these muscles are transversely striated. Dorsally some of them frequently join or approach one another. It is by the contractions of these bands with the mouth closed that water is expelled from the atrial cavity, causing the movement of the animal in a forward direction. The mouth is not lobed, but is bounded by mobile upper and lower lips. It leads into the pre- branchial part of the pharynx, which contains but a single ten- tacle the so-called languet on its dorsal side (Fig. 40, 11} 46 PHYLUM TUNIC ATA (UROCHORDA). Emb J FIG. 38. a sexual, & asexual form of Saipa (Thalia) democratica-miicronata (from Clans). mouth ; A atrial aperture ; N ganglion ; Br gill ; End endostyle ; Wg ciliated pit ; Ma mantle ; .V nucleus ; C heart ; Emb embryo ; st.p stolon. w I- ni. 39. a Himl c-nd nt Sul/m democratica, ventral view. St.p stolon; iVw nucleus, b Ter- minal part (it stolon lyuiiiin chain) nia^niliril. o mouth ; A atrial aperture ; A T ganglion,'; ll'i/ ciliated pit ; II / peripharyngeal band ; End endostyle ; Af anus ; Br gill ; ^Vw nucleus ilr nvar\ ; (' heart (Irnin Clans). SALPIDA. 47 Behind this is the peripharyngeal band (16), which is grooved in the usual manner. The nervous system and the organs of sense, in correspondence with the power of free locomotion, present a higher grade of development than in the Ascidiacea. The ganglion (13) with its numerous nerves lies dorsal to the anterior attachment of the : ' gill " and is of considerable size. It has on its dorsal side a horseshoe-shaped brownish-red pigment band, in which are con- tained besides the pigment numerous rod-shaped structures, the whole directly resting on and being part of the ganglion. There can be but little doubt that this structure is an eye. The dorsal tubercle (ciliated pit) is a short diverticulum of the pharynx in front of the peripharyngeal band ; it is without a FIG. 40. Sal'pa in longitudinal vertical section, seen from the left side, diagrammatic. 1 mouth ; 2 atrial aperture ; 3 anus ; 4 pharynx (branchial sac) ; 5 " gill " (dorsal lamina) ; 6 duct of subueural gland ; 7 endostyle ; 8 heart ; 9 oesophagus ; 11 languet ; 12 muscle - bands ; 13 nerve ganglion ; 14 embryo in ovisac ; 15 atrial cavity ; 16 peripharyngeal ring ; 17 stomach ; 18 testis ; 19 test ; 20 subneural gland ; 21 dorsal tubercle. glandular part and does not extend as far back as the ganglion (21). On the ventral side of the ganglion there are two glands (subneural glands) which open by separate orifices into the pharynx (or atrial cavity ?) just in front of the peripharyngeal band (6) and behind the representative of the dorsal tubercle. There are no otocysts in salps. The pharynx is without any side walls or stigmata and freely communicates with the atrial cavity on each side. The endo- style (7) is present in the usual form, and the dorso-median wall of the pharynx is represented by the so-called gill (5) which takes an oblique course across the body from near the ganglion to the opening of the oesophagus where it terminates. The oesophagus (.9) leads into a gut which possesses a stomach and pyloric gland. 48 PHYLUM TUNICATA (UKOCHORDA). is twisted on itself and constitutes the nucleus. The anus (3) opens into the atrial cavity in the posterior region. In Cyclo- salpa there is no distinct nucleus ; for the intestine is not twisted but passes forward, along the endostyle in the sexual form, and along the gill in the asexual form, to open anteriorly into the atrial cavity. The heart is placed in the nucleus in front of the stomach. In the asexual or solitary form (see below) the pharynx gives off a diverticulum at the hind end of the endostyle which enters the stolon. The stolon is a process of the ventral body- wall anterior to the nucleus. It lies in an excavation of the test and contains in addition to this pharyngeal diverticulum a number of other structures which will be described below. Its function is to segment into a number of zooids, which develop sexual organs and differ, when fully formed, in certain features from the asexual animals which arise from the egg. The sexual forms which originate from the stolon remain adherent by pro- cesses of the test, and form chains, from which they break off in sections. They are therefore generally found joined to others, but in some cases they become eventually entirely separate from one another. We thus get in the salps the regular alter- nation of a sexual and an asexual generation. The asexual form is solitary, the sexual usually joined to others in chains. As these two forms differ slightly in anatomical structure and have often been described before their genetic connection was known, they have generally received different specific names. To indicate this connection when it has been discovered and to render the matter perfectly clear, both these specific names are used in the name of the animal ; thus Thalia democratica mucronata is a species of which the asexual form was originally described as Salpa democratica and the sexual or chain-form as 8. mucronata, their genetic connection being subsequently dis- covered. The alternation of generation in salps was discovered at the beginning of last century by the poet Chamisso,* and sub- sequently rediscovered by Steenstrup. The sexual form, often called chain-form, proles gregaria, or blastozoite,! dillers from the asexual form, often called solitary / ' I if uiii/intliliiix i/ii\ different authors, >.jr. the blastozoite is applied to the sexual form 1 ause it is produced liy liiiddinu niid t> (lie iisexiiid form heeause it '"ids. It is host therefore to discard these terms. SA.LPIDA. e , proles solitaria or oozoite, in a number of characters, among the most important of which must be mentioned the presence of generative organs and the absence of a stolon. The salps, as a rule, develop only one ovum (lasis produces several ova and embryos). The ovary is in the young form placed in the nucleus by the right side of the intestine and is connected with the epithelium of the atrial cavity by a long stalk, consisting for the most part of a single cord of cells (Fig. 41). The . J X \^ .,/ opening of the oviduct will be formed later at the point where the stalk joins the atrial epithe- lium on the dor- sal side of the atrial cavity, somewhat to the right behind the penultimate muscular ring. Before fertilisation the cord con- necting the ovary to the atrial wall shortens and draws the ovary from the nucleus to the dorsal side. At the same time it acquires a lumen which opens into the atrial cavity. The ovary now has the form of a simple epithelial sac opening into the atrium and containing the ovum. Soon after fertilisation the oviduct loses its opening and the whole epithelial wall both of oviduct and sac constitutes the follicle of the ovum (Fig. 42 C, b). The testes which develop and ripen after the ovary are a pair of branched tubular glands lying in the nucleus on each side of the digestive organs. They open separately into the atrium. Amongst other points of difference between the sexual and asexual form attention may be called to the adhesive papillae on its ventro-lateral surfaces by which the sexual form is attached to its fellows in the chain. There are eight pairs of these. In isolated individuals they disappear. The muscles are less numerous and less developed, and finally the visual apparatus is FIG. 41. Side view of Salpa democratica-mucronata, sexual form (combined after Claus and Salensky from Korschelt and Heider). d atrial cavity ; e atrial aperture ; end endo- style ; / peripharyngeal baud ; i mouth ; k gill ; n ganglion ; nu nucleus ; od oviduct ; ov ovary ; ph pharynx-cavity ; x aperture of oviduct. z in E 50 PHYLUM TUNICATA (UKOCHORDA). much more elaborate, consisting of eight pigmented structures on the dorsal side of the ganglion. The development takes place within the follicle described above and is remarkable for the fact that the follicle cells play an important part in transferring nourishment from the maternal organism to the embryo. They do this in two ways : (1) the follicle cells, on the side of the embryo which future development shows to be ventral, proliferate and form a thick mass which soon assumes a cavernous structure. This is the pla- centa (Fig. 43). Maternal blood passes through its spaces and it is con- tinually detaching small masses which, charged with nutritive matter, pass into the vascular system of the embryo and afford nourishment for the growing tissues. ' (2) The follicle-cells in other parts of the follicle pro- liferate and migrate in amongst the blastomeres, with which they become inextricably intermingled, so that it is difficult if not impossible to dis- tinguish between them. These cells are called calymnocytes (gonoblasts). Their fate is much disputed, but there can be but little doubt that they, like the plac.-ntal tissues, afford nourishment to the growing embryo. FIG. 42. Stages in the cleavage of Salpa democratica-miicronata (after Salensky from Kor- schelt and Heider). In A and B the oviduct is open, in C it is closed, and the oviducal and follicular epithelia have combined to form the follicle, a atrial epithelium ; b shortened oviduct ; c follicle-epithelium ; / blastomeres ; z calymnocytes. In C, c points to the rudi- nic-iit ni the placenta. Many views have been held as to the nature of the calymnocytes. Sal- ensky is of opinion that they actually form the embryo, and that the em- bryo of Salpa is a follicular bud and not a true embryo. Brooks thinks that for a time they actually form the embryonic organs but are eventually replaced and probably consumed by the blastomeres. Meanwhile the embryo has been growing and projecting more and more into the atrial cavity. The follicle cells and the atrial epithelium over it become thin and eventually rupture so that the embryo lies free in the atrial cavity attached only to the placenta (Fig. 43). The formation of the layers is impossible to follow, but the organs gradu- ally make their appearance in the confused mass of embryonic and follicular cells. With regard to them it is only possible to note a few points here. The pericardium is. according to Salensky, developed from the mesoderm ; but Korotneff asserts that it arises as a diverticulum of the pharynx, which is perhaps the more usual way in Tunicata. The ganglion is by most observers described as arising from a mesodermal mass which becomes secondarily penetrated by an endodermal diverticulum of the pharynx. It is at first hollow and its cavity communicates with the pharynx. SALPIDA. 51 Later it becomes solid and detached from the pharyngeal diver ticuluin, which persists as the ciliated pit. The two subneural glands arise as two invaginations of endoderm further back where the ganglion is in contact with the pharyngeal epithelium. The elaeoblast (Fig. 43, eb) is a mass of apparently large mesodermal cells in the nucleus in later embryonic life. The large cells of which it consists are filled with nutritive matter, and it is probably to be regarded as reserve of food material. A similar tissue is found in the stolon of Pyrosoma. By Salensky it is regarded as a vestige of the tail and notochord of a larval stage. The stolon makes its appearance in the embryo as a diverticulum of the ventral wall of the pharynx between the end of the endostyle and c FIG. 43. Late stage of the embryo of Sal-pa democratica-mucronata (after Salensky from Kor- schelt and Heider). e atrial aperture ; eb elaeoblast ; ed intestine ; es endostyle ; ft ciliated pit ; i mouth ; k gill ; m stomach caecum ; n ganglion ; oe oesophagus ; p pericardium ; pi placenta ; st stolon ; t basal plate of placenta. the opening of the oesophagus (Fig. 43, st). This soon raises the ectoderm and a small process of the ventral body-wall is formed projecting into the tunic. As this process lengthens a cavity appears between it and the tunic, and acquires an opening to the exterior so that the young stolon lies in a tube formed by the tunic. In some species this tube is directed straight forwards along the ventral side of the body (Cyclosalpa pinnaia. Salpa affinis, etc.). In others it is at first directed forwards, then turns to the left and runs backwards on the left side of the nucleus, opening behind this organ (lasis tilesii, etc.). In yet others (Fig. 39) the tube winds spirally round the nucleus (Thalia democratica-mucronata, etc.). With later growth the stolon extends beyond this tube and projects freely uncovered by tvinic. Later on the zooids produced on the stolon acquire tunics as a result of their own activity. The projecting part of the chain of zooids is usually 52 PHYLUM TUNIC ATA (UROCHORDA). in the form of an elongated band, but in Cyclosalpa it has the form of a ring of from seven to twelve zooids attached together. The first rudiment of the stolon contains, in addition to the endodermal tube, a mass of meso- derm derived from the mesoderm of the parent in the neighbourhood of the elaeoblast and possibly from the elaeoblast itself. Stolons a little older contain on the ventral side of the pharyngeal diverticulum a mass of cells called the genital cord, a pair of peribranchial tubes, one on each side of the pharyngeal diverticulum, a pair of mesoderm bands placed immediately outside the peribranchial tubes, a nerve tube on the dorsal side, and a pericardial tube ventral to the pharyngeal diverticulum on the right side. In addition there are two blood-sinuses, one on the dorsal and the other on the ventral side of the pharyngeal diverticulum ; these communicate with each other at the free end of the stolon and proximally with the blood sinuses in the body of the parent. The neural and peribranchial tubes and mesodermal and genital bands are confined to the stolon and are not continuous with the tissues of the parent. Their origin is doubtful : according to Seeliger they are all differ- entiations of the mesoderm of the stolonic rudiment ; accord- ing to Brooks the peribranchial and neural tubes are derived from the ectoderm of the base of the stolon, while Korotneff maintains that the peribranchial tubes are derived from the en- doderm tube. The fate of these structures is as follows : the en- dodermal tube becomes the ^Transverse section through a young p h arvnx au d give rises to the mi nt Halpa (after Brooks, troin Korschelt J and Heider). c perihranchial tubes ; ec ecto- digestive canal, the neural tube derm; e endoderm tube ; genital cord ; , yields the ganglion and ciliated mesoderm hand; n neural tube ; <> upper, and // lower i.l 1-sjniis. pit, the peribranchial tubes the atrial cavity, while the genital cord gives rise to the genital organs., and the mesoblastic bands to the muscles, pericardium, heart, elaeoblast and mesoderm generally of the future zooids. In t he st met nre of the stolon ami it s relation to the parent, and in the for- mation of t he organs, the budding of salps is not unlike that of Pyrosoma. Tli" ph.ii vn-eal diverticulum clearly corresponds to the epicardial tube. The two processes differ mainly in the much larger number of buds pro- dueed liy salps before separation from the stolon. The formation of the buds takes place in the following way. The stolon becomes marked into segments by transverse constrictions due to unfold- in _' of the ectoderm. These .ire numerous (50 to 100), very close together, am I all appear at the same ti (Kiii. ">) They nre t he buds which gradu- ally enlarge and develop into the sexual zooids. All the organs of the Flo. SALPIDA. 53 stolon are divided by the segmentation except the outer ectoderm, the blood-sinuses, and the endodermal tube. These are continuous from end to end of the stolon, being found in the narrow stalks which connect the buds as well as in the buds themselves (Fig. 45). The stolon has therefore now the form of a chain of developing zooids connected together, the dorsal surface of one to the ventral surface of the next, by short narrow stalks. Meanwhile, the proximal part of the stolon, between the zooids and the parent, has been growing, and undergoing a similar process of segmenta- tion into a group of buds. This process is repeated, so that eventually a chain is formed consisting of groups of from 50 to 100 zooids. The indi- viduals of each group are of the same age, but the groups next the parent consist of younger individuals than the distal and earlier formed groups. As development proceeds a change in the attachment of the zooids is effected. At first they are arranged in a single row, the intermediate con- stricted part of the stolon passing from the ventral surface of one to the dorsal surface of the next (Fig. 45). Soon, as a result of growth, by which the relations of the parts are changed, and of rotation, they become arranged in a double row, the zooids of one row alternating with those of FIG. 45. Diagram representing a stolon of Salpa as it would appear if no secondary shifting of the individuals were to take place (after Brooks, from Korschelt and Heider). P solitary form (parent) ; /, //, ///, first, second and third group of individuals (the number shown in each group is much smaller than actually occurs) ; b, b", b'" pharynx ; c", c'" atrium ; d digestive canal ; ee ectoderm ; el elaeoblast ; en endoderm of the connecting stalks ; es endostyle ; g gill ; h heart ; ganglion ; o", o'" ovary. the other. Later the intermediate constricted part of the stolon, which joins them, atrophies and the zooids remain adherent by processes of the body- wall and test, which have become developed for the purpose. When this happens the zooids of the group affected become detached from the proxi- mal part of the stolon and form a free-swimming chain of sexual individuals all of the same age and connected by processes of the body wall and test. The sexual zooids when full grown are of about the same size as the asexual. The ovaries are formed early and the ova are fertilised before the testes are developed. Fertilisation therefore must be effected by spermatozoa produced by another chain. Cyclosalpa Blainville, digestive tube running antero-posteriorly and not coiled up to form a nucleus ; the chain-forms are attached together in a circle. G. pinnata Forsk., with linear bands, C. affinis Cham., without linear bands. Salpa Forskal, digestive tube coiled up in the nucleus, one embryo, chain as elongated band, S. africana-maxima Forsk. Thalia 54 PHYLUM TUNIC AT A (UROCHORDA). Blumenbach, very like Salpa, Th. detnocratica-mucronata Forsk. Pegea Sav., very like Salpa but with supposed traces of stigmata (hemitremata) on its dorsal wall. lasis Sav., like Salpa, but with several embryos at different stages of development in the same sexual animal. The abyssal form Octacnerntts Moseley, discovered by the Challenger Expedition and described by Moseley* is placed here ; why, it is difficult to for it does not appear to possess any important salp-like feature. Octacnemus bythiu* .Moseley, body octoradiate, medusa-like, flattened antero-posteriorly, probably attached, test gelatinous, thin, transparent ; pi larynx with no stigmata or openings into the atrial cavity ; digestive 'ir-ans coiled up to form with the reproductive organs a nucleus which is pl.ir-ed posteriorly ; S. Pacific. Our knowledge of its structure is limited, .MM! nothing is known about its budding and reproduction. A second imperfectly known colonial speciesf O. patagoniensis Metcalfe from 1,000 fms. off the coast of Patagonia has been assigned to this genus. Sub-order 2. Cyclomyaria (Doliolida). Barrel-shaped Thaliacea with thin test, and pharynx with two rows of stigmata on its posterior watt. The asexual form has a ventral stolon which buds and produces three kinds of zooids, one of which develops sexual organs. The muscular rings are complete ventralli/ and a tailed larva is ahvays developed. Doliolum is a transparent pelagic organism with a thin test in \vhich there are no protoplasmic elements and no cellulose. The pharynx (Fig. 46) occupies the anterior part of the body and the atrial cavity the posterior, the digestive canal and the heart (5) together with, in the sexual form, the gonads being aggregated together in an inconspicuous nucleus behind the pharynx on the ventral side of the atrial cavity into which they project. The mouth (1) and atrial (17} openings are at opposite ends of the body and terminal, and the edges of both are lobed. There are no tentacles The dorsal tubercle (22) is surrounded by the spirally coiled, dorsal ends of the peripharyngeal bands, and the pharynx, which has an endostyle (3) but no dorsal lamina, possesses only two rows of stigmata placed in its posterior wall (20). There is a ganglion and closely adjacent subneural gland (21), which opens in front of the peripharyngeal band by the dorsal tubercle (22). The muscular bands of the body-wall are in the form of complete hoops, of which there are nine in the Trim*. Linn. Soc. {-). I. p. -_'s7, IX7t'>. See also Herdman Challenger I, 1 ' ports, Timi.-iitM. Pt. Ill, 1888. t Metealfe, Jo////> Hopkins Univ. ('ire., 12. IS't.X Fie. 8. DOL1OLIDA. 55 20- asexual and eight in the sexual forms, the anterior and posterior of them acting as sphincters. A tailed larva is formed, and develops into an asexual form which produces a budding stolon. The buds are polymorphic, there being three kinds, one of which alone becomes sexual. The asexual form (Fig. 46) presents the following features : It has nine muscular rings, and the ganglion is placed behind the fourth. The mouth is surrounded by ten lobes, the atrium by twelve. There is an otocyst (4) in the body-wall on the left side which contains an oto- lith and in some species opens to the exterior. The body carries two median appendages, a dorsal one (cadophore), arising far back near the atrial opening (19). and a ventral one, the stolon (8), arising behind the fifth muscular band. The stolon will be described below. The dorsal process contains a blood space divided into two by a septum, and its dorsal ectoderm is columnar. In the sexual form (Fig. 47) the mouth is surrounded by twelve lobes, the atrial aperture by ten ; there are eight muscular rings and the ganglion is placed behind the third. There is no otocyst, both the stolon and dorsal appen- dage are absent, and there is a greater number of stigmata. There is an ovary and a testis, both unpaired ; they open a little to the left of the middle line behind the anus. The ovary never contains more than one ripe egg at the same time, but as soon as this is expelled a second is formed, and after that a third. After the third egg is laid the ovary atrophies 15 J9 16 FIG. 46. Diagram of a longitudinal-vertical sec- tion through the asexual (budding) form of Doliolitm (after Delage and Herouard). ] mouth ; 2 peripharyngeal band ; 3 endostyle ; 4 otocyst ; 5 heart ; 6 left pharyngeal diverti- culum ; 7 mesoderm ; S stolon ; 9 diverticulum from left side of cloaca into stolon ; 10 aper- ture of oesophagus ; 11 pyloric glands ; 12 sto- mach ; 13 intestine ; 14 anus ; 15 nerve ; 76 tac- tile process ; 17 exhalent aperture ; 19 dorsal appendage ; 20 hinder wall of pharynx with stigmata ; 21 subneural gland with the closely adjacent ganglion ; 22 opening of duct of sub- neural gland (dorsal tubercle). 56 PHYLUM TUXICATA (UROCHORDA). and the testis ripens. The egg surrounded by its layer of follicle cells passes into the cloaca and thence into the sea. After ' at I SO si ov FIG. 47. Doliolum denticulatiim, sexual form, from the left side (after Herdmau). at at rial aperture ; at.l atrial lobes ; br mouth ; br.l buccal lobes ; brs pharynx ; dt dorsal tubercle ; end endostyle ; h heart : i intestine ; >;' m* muscular rings ; n nerve ; ng ganglion ; ov ovary ; pbr atrial cavity; pp peripharyng"al liand ; g stigmata ; sgl subneural gland ; so sense organs ; st stomach ; tes testis. fertilisation it secretes a vitelline membrane and falls to the bottom. Here it develops into a tailed larva by a process, of which all the stages have not been followed. It begins to swim by means of its tail while still within the vitelline membrane, Flo. 48. Old larva of Duliolinn i-/ifi-ni>r,;/i, ;itt.-r I'ljanin, from Korschelt and Heider). ch notochord ; d dorsal appendage ; ' endnstyle ; h heart and pericardium ; m vitelline mem- brane ; n ganglion ; r rudiment "I -tnlmi in^cttc-shaped organ). DOLIOLIDA. 57 which eventually ruptures and allows the larva to go free. The anterior end of the body assumes the barrel-shaped form of the adult, with its dorsal process, but still preserves its larval tail and notochord (Fig. 48). These eventually atrophy, persisting for a time as a stump, which resembles in structure and position the organ called elaeoblast in the salps and Pyrosoma. The principal features in the development are as follows. Cleavage is total and is followed by an invaginate gastrula. In the next stage observed the embryo appears to contain three cell-groups which constitute the rudi- ments of the nervous system, the mesoderm and the notochord. The alimentary canal and endoderm are formed later as a se- condary invagination of ectoderm. The atrium is formed by another invagination of ectoderm. The nervous r u d i ment elongates. Its median part remains bulky and gives rise to the ganglion of the adult and the subneural gland (ventral lobe) ; its anterior end narrows, acquires a lumen which opens into the pharynx, and forms the duct of the subneural gland and the dorsal tubercle ; the posterior portion also becomes narrow and persists as the posterior unpaired nerve (nervus branchialis). The pericardium is developed as an excavation of a portion of the mesoderm, and the stolon is formed as a small ventral process. The stolon in its earliest stage in the embryo consists of a small mass of mesoderm applied against the ectoderm in the neighbourhood of the heart. This is very soon reinforced by two pairs of out- Arvf f'"' ''^^ growths, one from the atrium and the other from P "$?. ".l|a the pharynx. These, carrying with them the mesoderm, cause a projection of the ventral body wall, which extends into, but does not pierce, the tunic. Later the stolon pierces the tunic and pro- jects freely, and the five strands contained in it be- come increased, by development of the cloacal and now FIG. 49. Dorsal region of a larva of Doliolum Mullen (after Uljanin, from Korschelt and Heider). cl atrium ; fl, dorsal tubercle and duct of subneural gland ; m muscle-hoops ; ganglion ; nb branchial nerve. R I FIG. 50.- CCL -Probudof Dolio- ; showing its trans- pharyngeal tubes, to seven. The free end " wn 1 DeSe 1 and segments into small bodies (Fig. 50), which become Herouard). < trans- detached and are known as probuds. The probuds f contain a portion of the seven cell masses sur- rounded by ectoderm. They wander by means of the pseudopodial activity of certain of their ectoderm cells (Fig. 50, ca) along the right side of the animal to the dorsal side of the base of the dorsal process (Fig. 51). Here they divide into from fourteen to twenty buds, which attach themselves on each side of the middle dorsal line. This 58 PHYLUM TUNICATA (UROCHORDA). the stolon (Fig. 52). attachment is an epithelial one ; they perforate the tunic and their ecto- derm becomes adherent to the columnar ectoderm of the dorsal side of By the growth of the stolon in girth they become displaced outwards and give rise to the lateral buds (Fig. 52, G) of the dorsal stolon. As the stolon is continually growing in length the buds which first affix themselves are carried away from the body and new buds are at- tached between them and the base of the stolon. Thus it happens that the buds increase in age as we travel from the body along the stolon. The probuds which arrive first at the dorsal process give rise to lateral btids (G). Those which arrive later give rise to median buds, which are attached nearer the middle line and are not displaced so far outwards (Fig. 52, P). The fate of the lateral and median buds is as follows. The lateral buds develop into small doliolum - like zooids with the rudiments of gonads, which however soon atro- phy, and with a well-de- veloped alimentary canal, but without a closed atrial cavity (Fig. 52, G). They actively take in nourish- ment, and serve for the nutri- tion of the growing stolon. They are therefore called qastrozooids. They are in osmotic relation, through their epithelial attachment to the dorsal stolon, with the blood of the parent. This relation, is of some im- portance, because the a- sexual parent (oozoite) has lost its pharynx and diges- tive canal by atrophy and. retaining only its ner\ mis system, heart, and muscles, has become converted into u. ML. .">!.- -Dorsal view of the posterior part "t an asexual form (nurse) of Doliolum showing the migra- tion of the probuds (after Barrels, from Korschelt and Heider). I lateral liuds ; m median buds ; ;/ ..f.M.S., 4, lSf>(>. Fol, Etudes sur les Appendiculaires tin detroit de .Messiue. Mnn. de la So<-ihioxus. The blood-channels are without epithelial walls. APPENDICULARIAE. 63 and the blood appears to be entirely devoid of cellular elements. The mouth opens into a spacious pharynx (Fig. 55), and leads behind by a short oesophagus into a dilated stomach, from which passes forward an intestine to open on the ventral surface by the anus. The hind end of the intestine may be marked off as a rectum. There is no pyloric gland or liver except in Stegosoma and Megalocercus in which a gland, identified as liver, opens by a narrow duct into the stomach. The pharynx has, on the anterior part of its ventral wall, a short endostyle (o), consisting of a groove bounded by large glandular cells without cilia except at its front end. From the front end of the endostyle there arises on each side a ciliated band which passes backwards and dorsal - wards to join its fellow at the oesophageal opening. There is no dorsal lamina and no row of tentacles. From the hind end of the endostyle a ventral band of cilia, corresponding to the ventral groove of other Tunicata, passes backwards to the oeso- phageal opening. In the ventral wall of the pharynx behind the endostyle are the internal openings of the two gill-apertures ; these lead into two short ciliated tubes which open on the ventral side of the body by the spiracles (n). There is a small ciliated diverticulum of the dorsal wall of the pharynx, placed on the right side of the ganglion and corresponding to the dorsal tubercle of other Tunicata. The nervous system consists of a ganglion (Fig. 55) placed on the dorsal side of the anterior wall of the pharynx, and of a dorsal nerve cord (d) which passes back from the ganglion on the right side of the stomach to the tail. In the tail, to the hind end of which it extends, the dorsal cord lies on the left side of the notochord and presents a variable number of swelhngs due to the presence of nerve cells and called caudal ganglia. The number of these caudal ganglia varies from eight to forty. The first of them is usually larger than the others. Nerves are given off by the cerebral ganglion, and by the caudal ganglia. It has been stated that the cerebral ganglion and dorsal cord contain a minute canal, but this is doubtful The position of the caudal nerve cord on the left side of the notochord has suggested the view that the tail has undergone rotation through 90 so that its true dorsal surface has come to lie to the left. An otolithic vesicle (a) lies on the left side of the ganglion, so closely in contact with it that at one place the inner wall of the 64 PHYLUM TUNICATA (tJROCHORDA). vesicle appears to be formed by the ganglion. There are no visual organs. The notochord (h) is confined to the tail. It consists of an axial hyaline substance of cartilaginous consistency, surrounded by a protoplasmic nucleated membrane. The nuclei project on the inner side into the hyaline axial substance. The musculature of the tail is well developed. It consists of two broad bands of muscular substance, one on each side of the noto- chord. Each band consists of an outer protoplasmic layer con- taining the nuclei and an inner layer of striated contractile substance. The nuclei are ten in number in each band, arranged in a row at regular intervals, thus indicating that the muscular bands are composed of ten cells. The limits of these cells can- not be made out in the adult ; though it is said that after certain treatment indications of them may be seen as transverse lines through the muscle substance between the nuclei. (Fig. 55, k). It has been suggested that these lines represent the limit of segments and that the tail is segmented. This however is very doubtful. There is no correspondence between the caudal ganglia and the supposed muscle segments. The pericardium is a simple epithelial sac placed on the ventral side of the stomach. Its dorsal wall is contractile and slightly invaginated into its cavity. A concave contractile lamella (Fig. 55) bounding a blood sinus, the dorsal wall of which is formed by the stomach, is thus established. This concave contractile membrane constitutes the heart. When it contracts it propels the blood from the super jacent blood sinus into the blood channels generally and so acts as the central organ of the circulation. It is said to reverse the direction of its action as in other Tunicates, but this is doubtful. The wall of the pericardium is a simple protoplasmic membrane con- taininn nuclei. Dorsally it contains in its outer layer striated contractile fibres. It differs from the heart of other Tunicates merely in the small <-\tent to which the dorsal wall of the pericardium is invaginated. The gonads are contained at the posterior end of the body (Fig. 55, /, g). Excepting Oikopleura dioica all members of the group are hermaphrodite, the male organs maturing first. There is a single or double ovary contained between two testes. There is no oviduct. The eggs when ripe dehisce into the vascular cavity from which they escape by dehiscence of the body-wall, APPENDICULARIAE. 65 causing the death of the animal. The testes acquire an opening to the exterior at the time of sexual maturity. Budding does not take place. What little is known of the development * appears to indicate that it does not differ essentially from that of other Ascidians. The first Appendicu- larian was discovered by Chamisso, who gave it the name Appendicularia. For a long time the systematic position of these forms remained obscure. By many zoologists they were regarded as larval Tunicates, but it was Huxley who dis- covered their spermatozoa and to whom belongs the merit of having first recognized them as members of the tunicate phylum. Appendicularia Cham., capsule ovoid, stomach unilobed, rectum enor- mous, testis single. Oikopleura Mertens, capsule large, stomach bilobed, mouth with ventral lip (Fig. 54) ; tail very long. Vexillaria J. Muller, special muscles traverse the body and are inserted on to the viscera. Stego- soma Chun, capsule unknown, stomach with liver, surface-waters and deep sea to 1,000 fms. Megalocercus Chun, the largest of all known genera, length of body 8 mm., total length 30 mm., capsule unknown, with a body- wall musculature similar to that of the Salps ; Med., 600-900 fms. Folia Lohmann, capsule unknown, tail long, genital mass single ; Atl. Althoffia Lohmann, warm parts of Atl. Fritillaria Q. and G., body elongated, con- stricted in the middle at the attachment of the tail, a fold of the dorsal integument forming a hood over the head (Fig. 55), testis usually single. Kowalevskia Lahille, capsule large with single orifice, and interior marked with projecting ribs ; endostyle, peripharyngeal band, pericardium and heart absent ; pharynx with 4 longitudinal rows (a dorsal and ventral on each side} of solid, ciliated comb-like processes which act as strainers ; Med. and E. Atl. Goldschmidt, Biol. Centralbl., 23, 1903, p. 72. Z III CHAPTER II. PHYLUM ENTEROPNEUSTA. Unsegmented Chordata with a tripartite division of the body rind coelom, a dorsal preoral lobe, and a notochord-like structure which is confined to the anterior (proboscis) region of the body. Pharyngeal branchial apertures are present in all except Rhab- popleura. The Enteropneusta are Chordates * which present the fol- lowing important features. The body is divided into three regions, the proboscis which is a dorsal preoral lobe, the collar, and the trunk. In the anterior part of the trunk paired lateral apertures the gill-slits are present putting the alimentaiy canal in communication with the exterior (except in Rhabdo- pleura). The coelom, in the cases in which its development is known, arises as five diverticula of the embryonic enteron, viz., one unpaired pouch, which extends into the proboscis and is called the proboscis cavity ; one pair of pouches which occupy the collar-region and are known as the collar- cavities ; and finally a posterior pair which occupies the whole of the trunk region and constitutes the trunk-cavities. The proboscis cavity opens to the exterior by a single or double pore ; the collar Cavities each by a pore ; while the trunk-cavities are devoid of an external opening. The anterior part of the alimentary canal sends into the base of the proboscis an unpaired diverticulum which has a charac- teristic structure and has been identified as a notochord. The central nervous system lies in the ectoderm, and there is a special concentration of it in the dorso-median line of the collar. This concentration comes, in the majority of forms, to lie in the wall of a canal which is open at both ends. There is always a * See vol. ~2. oh. i. 66 AFFINITIES. NOMENCLATURE. 67 system of tubes which are identified as vascular, and in all there is a curious organ in the base of the proboscis which consists partly of the before-mentioned gut diverticulum and partly of vascular and glandular tissue. This organ is called the central- or proboscis-complex. A remarkable feature of the phylum and one which it has in common with the Cephalochorcla is the absence or small amount of stellate or other connective tissue cells between the coelomic epithelium and the epithelium of the ectoderm and endod'erm (p. 91). The group Enteropneusta was established by Gegenbaur in 1870 and was placed among the Vermes. They have been re- garded by some authors as especially related to the Nemertines and by others as possessing leanings towards the Annelids. Aletschnikof}' * in 1869 pointed out their resemblances to Echinoderms, while Gegenbaur (1874) and Huxley f (1877) were the first to call attention to the Tunicate affinities, Huxley going so far as to include the two in a new group, the Pharyngo- pneusta. The connexion with the Vertebrata was suggested by Sedgwick 1: in 1884, and their inclusion in the group Chordata was accomplished by Bateson in 1885. The Echhioderm affinities first pointed out by Metschnikoff undoubtedly exist and are dealt with below (p. 99) arid in the section dealing with affinities in the general account of Echinoderms. The suggested affinities with Phoronis are discussed at the end of this chapter. The phylum is divided into two orders, the Balanoglossida and the Cephalodiscida. The establishment of these orders has been necessitated by the discovery of the genera,' RJiabdopleura and Cephalodiscus, which present all the important Enteropneust features (except the gill apertures in Bhdbdopleura) . A great variety of names has been proposed for the phylum here called Enteropneusta. We have selected Gegenbaur' s term on the grounds of priority and usage. The discovery of Cephalodiscus and Bhabdopleura, and consequent necessity for the establishment of two orders within the phylum, have induced some authors to restrict the term Enteropneusta to the forms which we have here included under Balanoglossida, and to employ another name, Hemichordata, for the phylum. We object to this terminology for two reasons : in the first place Cephalodiscus is as much ' Ueber Tornaria, Nachr. K. Ges. Wiss. u. Univ. Gottingen, 1809, p. 287- "292, and Z. f. w . Z., 20, 1870, pp. 131-144. t Journal Linn. Soc., Zool., 12, 1877, pp. 199-226. J Q.J.M.S., 24, p. 70. Q.J.M.S., 25 sup., p. 111. 68 PHYLUM ENTEROPNEUSTA. enten.pneiist as Itiiluiiniiluxxii.i or Pttjchodera and cannot logically be divorced from the group to which that name is applied. In the second place, Hemichordata is a bad term, because it is not certain that the struc- ture identified as notochord is really such, and because, even if it is a noto- , it cannot fairly be said to be " hemi." It might perhaps be termed ri." but no one. s<> far as we know, has suggested this or applied tin- nanif Proboscichordata to the phylum. Order 1. BALAXOGLOSSIDA.* Enteropneusta irilh vermiform, cloifjahd body and tim/i// /' ilu . I mi rican Acad. of Arts and Sciences, 9, 1873. W. Bateson, A Series of Studies in the Anatomy and Development of Balanoglossus in the Q.J.M.S., 24, 25, and 26, 1884-6. J. W. Spengel, Die Enteropneusten, Fauna und Flora des Golfes von Neapel, 18th Monograph, 1893. Id., Die r-'-iiennung der Enteropneusten Gattungen, Zool. Jahrb., Syst., 15, 1901, p. 209. Id., Neue Beitrage, etc.. ii.. iii., iv., Zool. Jahbr., Anat., 20, p. 1, 1 1. .'!!."), p. 413. J. P. Hill, The Enteropnexista of Funafuti, Memoir* Aux- ti-nlnni Miiximn, 3, 1898, pp. '2(1.") and 335. A. Willey, Enteropneusta tnuu the S. Pacific, Willci/'* Zoological Result.*, |>t. 3, 1899, p. 32. R, C. I'mmcft, The Entcropneusta, < luril iii<'/(it/. S. F. Harnicr. Hemichordata, in Cambridge Natural History vol. 7. l!(4. p. 3. REGIONS OF THE BODY. SKIN. 09 , slightly over the anterior end of the trunk, so that there is a circular groove at the junction of the collar and trunk. In some species this hind end of the collar may project so far back as to cover three or four gill- slits (Dolichoglossus koivalevskii), and in this case the projection has been called by Bateson the atrial fold, and the space which it encloses the atrial cavity (cf. Ampliioxus). The trunk itself is also indistinctly divided into re- gions : these are the bran- chiogenital immediately succeeding the collar (Fig. 56, k, g) ; next, the hepatic (vb), in which the intestinal walls contain a green or brown pigment ; and lastly the abdominal region which forms the hind end of the body. The branchiogenital re- gion possesses in its anterior part on each side of the dorsal middle line a double row of pores ; these are the branchial pores which place the gill-apertures of the alimentary canal in com- munication with the ex- terior. The branchial pores are placed in a slight groove, the branchial groove. Posteriorly the branchiogenital region is without branchial apertures. The genital organs are also found in. the lateral walls of the body in this region. The skin consists of a layer of ciliated epidermis or ectoderm beneath which is a structureless basement membrane. The epidermis is in most parts a thick layer and appears to consist of very long and narrow cells, extending indeed through the whole thickness (Fig. 57) with interspersed unicellular glands. Moreover the nuclei occur at different levels, thus suggesting at FIG. 56. Dolichoglossus kowalevskii (after A. Agas- siz, from Korschelt and Heider). e proboscis ; kr collar ; k anterior part of branchial region ; g genital region ; db dorsal, vb ventral middle line. 70 BALANOGLOSSID A . first sight the view that the epidermis is many-layered. These long ceUs, into which the epidermis breaks up when teased or otherwise ill-treated, are connected together by numerous lateral processes (Fig. 57), and are prolonged internally into fine fibres which enter the nervous felt-work found in the deeper part of the epidermis over the greater part of the body. The thickness of the epidermis and the number and disposition of the gland cells varies in different places. The muscles and connective tissue are entirely derived from the walls of the coelomic sacs and will be considered in connexion with those struc- tures. The nervous system. In the lower stratum of the ectoderm just external to the basement mem- brane there is the feltwork of fine fibres almost en- tirely without nuclei, which has just been men- FlO. 57. Preparations of the epidermis of the collar of BuJano- tidied. This tlS- ylosstis clavigerus (after Bateson). A, a similar preparation more highly magnified, showing the long spindle-shaped cells g\jg jg Supposed to passing internally into fine fibres ; -B, portion of epidermis teased and pressed out. showing elongated cells united at their fog nei'VOUS in outer ends l>y numerous connection-. function and strongly recalls the corresponding layer found in many Xcmertini. It is thickened along the dorsal middle line of the collar and trunk and along the ventral middle line of the trunk. These constitute the dorsal and ventral nerve cords (the collar portion of the dorsal cord is invaginated, see below). They are connected at the junction of the collar and trunk by lateral commissures which like them are special concentrations of the nerve plexus and lie in the ectoderm at the base of the collar or atrial groove. The dorsal nerve cord extends forwards on to the base of the proboscis to become continuous with the general NERVOUS SYSTEM. 71 epidermal nerve-plexus of that organ. This is specially de- veloped round the stalk of the proboscis to form the somewhat FIG. 58. Transverse section through the collar nerve cord of Glossobalanus sarniensis. showing the central canal c with its cuticular lining ; ga giant ganglion cell, dw dorsal rout. ill-defined structure which is frequently spoken of as the proboscis nerve-ring. The part of the dorsal cord contained in the collar has different relations from the rest of the nerve trunks. It nc FIG. 59. Transverse section of the collar nerve cord of Glossobalanus minutus (after Spengel). showing two of the isolated cavities with their cuticular lining, and the radiating lines sur- rounding them ; nc canals in the fibrous substances which contained processes of the giant cells. lies in the ventral wall of a tube or rod (Figs. 58, 59, 60) of ectoderm which is derived in the embryo from the delamination 72 BALANOGLOSSIDA. or imagination (according to the species) of a dorsal median tract of collar ectoderm. This collar nerve-cord, which is called the medullary cord, is always continuous with the outer ecto- derm at the front and hind end of the collar (and in some species by the dorsal roots, see below). At these points (Fig. 60) there is almost always a canal the medullary canal which in some species is continued through the whole length of the cord. The central canal, whether it is confined to the anterior and pos- terior ends or extends throughout the whole length of the cord, opens to the exterior by the anterior and posterior nerve pores, the former being placed at the base of the proboscis stalk within the anterior projecting rim of the collar, and the latter at the Imttom of the groove (atrial) which is found at the junction of the collar and trunk (Fig. 60). The fibrous matter of the trunk dorsal nerve-cord is continued along the ventral side of the medullary cord. In the genus Glossobalanus, however, the fibrous matter is continued right round the medullary cord, being found on its dorsal surface also (Figs. 58, 59). In those cases in which the central canal is not present as a continuous tube, it appears to be represented by a number of isolated cavities which are frequently surrounded by elongated radiating cells.* In Schizocardium peruvianum and Dolichoglossus merechkovskyi the medullary cord is quite solid, there being neither neuropores nor isolated cavities. There is frequently present at the anterior and posterior ends of the medullary cord an ectodermal pit, with the ectodermal lining of which the nerve cord is continuous. The anterior and posterior neuropores when present open into these pits, or if the cord is entirely solid, as in the two >l>ecies just named, the medullary cord is continuous with the ectoderm lining them. It is difficult to say whether the pits should be regarded as terminal parts of the central canal or as independent structures. Some importance has been attached to them by Spengel, but as in many cases it is exceedingly difficult to distinguish them from the adjacent parts of the medullary canal, it is perhaps better to regard them as the terminal and, if the central canal is elsewhere obliterated, as persistent portions of this 'anal. The dorsal nerve cord lias the same relation to these pits as it has to the cord, i.e. it lies in their \entral \\alls. They are sometimes dilated and the anterior pit may project backwards so as to overlap the dorsal side of the adjacent p : irt of the medullary cord (Spengelia alba, Schizo- i-nriii mil ). thus recalliii!_ r in some decree a cerebral vesicle of the vertebrate brain. The condition of the central canal varies much even in different species of the same genus. It is present as a complete canal, traversing the whole length of the medullary cord and opening at each end by a neuropore, only in the genus Ptyclwdera(Pt. flam, erythrea, Imhiinx //*/*). in Glossobalanufi In some, cases these isolated ca\ itics are said to coexist with a central canal without however communicating with this latter (e.g. S'c/i. brasil- " n-' in some individuals). NERVOUS SYSTEM. 73 tiarniensis, hedleyi, ruficollis, and in Balanoglossus apertus. In other species it is as a rule only present at the anterior and posterior ends, being repre- sented in the middle by a number of cavities (sometimes arranged in a double row [Fig. 59] ). In two cases, already mentioned, all traces of the central canal are absent (unless the anterior and posterior pockets are regarded as parts of it), the cord being solid. But even within the limits of the same species it frequently happens that the central canal or its repre- sentative exhibits considerable variation. In the Ptvchoderida. 1 the dorsal side of the medullary cord is / *. connected with the dorsal ectoderm by a variable number of cords which constitute the dorsal nerve roots (Figs. 60, 8). These roots are frequently hollow (Fig. 58), particularly those which arise from the anterior end of the cord, and their axial canal communicates with the central canal of the cord or with one of the isolated cavities which represent that structure. The axial canals of the roots do not however open to the exterior, but in some species they may be continued for a short distance in the epidermis as intraepidermal canals (Balanoglossus canwsus). The dorsal nerve roots vary much in number and in the course which they take in reaching the ectoderm. In Glossobalanus rnfiro//fs there may be from 12 to 18 of them, in Ptychodera flava 1. to 6. As to course, they may be straight or directed obliquely backwards or even curved. In the other families dorsal roots are rarely present. There seems to be a trace of one such structure (not, however, reaching the skin) in SpengeMa porosa. In Dolichoglossus kowalevskii the posterior part of the medullary cord possesses a dorsal keel-like projection, which is solid and does not reach the epidermis. The roots whether solid or hollow consist of cellular prolongations of the cord and are coated with fibrous matter which is a prolongation of that of the cord and is continuous with the nervous stratum of the epidermis. The ectoderm over the dorsal and ventral nerve cords of the trunk is thickener) and sometimes projects as a ridge : in some cases it is partially invaginated and so placed at the bottom of a groove. Moreover it differs from the rest of the epidermis in containing fewer gland cells. As an indi- cation of the fact that the nerve cord of the collar is derived from the com- pletion of such an invagination by the union of the sides of the groove over the gland-free nerve-cord ectoderm, may be mentioned the fact that the ectoderm of the floor of the central canal of the collar cord is more or less free from gland cells, while they abound in the roof. Further, as already mentioned, fibrous nerve matter is usually found only in ^he ventral wall of the ventral canal. The ectodermal lining of the central canal and iso- lated medullary cavities possesses a cuticle-like border. It is impossible to arrive at any satisfactory conclusions as to the structure of the nerve tissue. The most that can be said is that the external ends of some of the superjacent ectoderm cells extend into and branch in it. According to Bateson, processes from the fibrous matter of the nerve cords may be traced through the basement membrane into the adjacent muscles. There do not appear to be anv organs of special sense. 74 BALANOGLOSSIDA. The alimentary canal extends as a straight tube from mouth to anus. The mouth (p. 68) leads into a wide tube which traverses the collar region and is called the buccal cavity. The buccal cavity passes at the junction of collar and trunk into the pharynx in the side walls of which the branchial slits are placed ; this leads in the hinder part of the branchiogenital region into the oesophagus which passes behind into the intestine. The anterior part of the intestine is distinguished as the hepatic intestine from the posterior part which opens by the anus. The part of the intestine next the anus is distinguished as the rectum. The wall of the hepatic part of the intestine contains a greenish or brown pigment, and in the Ptychoderidse and in Schizocardium fit*. 60. Diagrammatic Longitudinal-vertical section of Ptychodera (from MacBride). C col- lar ; Pr proboscis ; Tr trunk. 1 proboscis coelom ; 2 glomerulus ; 3 heart ; 4 pericardium ; i proboscis pore ; 6 collar coelom ; 7 collar nerve-cord ; 8 dorsal nerve roots of 1 ; D dorsal iiluncl-vfssrl ; 10 dorsal mesentery ; 11 external opening of the branchial sac (not clearly shown), 12 U-shaped internal opening of the same ; 13 ventral blood-vessel; 14 ventral part (hypobrauchial groove) of the branchial region (pharynx) of the alimentary canal: 15 dorsal part of the branchial region of the same ; 16 buccal cavity ; 17 mouth ; 75 noto- chord ; 19 euticular sheath of the same ; 20 ventral pocket of proboscis coelom ; 21 ventral n nt 1 proboscis coelom. possesses on each side a row of hepatic diverticula which cause the hepatic sacculations of the body wall visible in these genera . The alimentary canal is lined by a more or less columnar epithe- lium and is ciliated throughout. It is attached to the body wall in the dorsal and ventral middle lines by longitudinal mesen- teries ; which however in the adult are deficient in certain parts of the body (see below, pp. 87, 89). The buccal cavity gives off from the anterior part of its roof, in the dorsal middle line, a forwardly directed caecal diverticulum (Figs, 60, 64), which, following Bateson, we shall call the noto- chord. The notochord is a tubular structure which extends through the neck of the proboscis and projects into its base. It is divisible into two regions : the narrower neck which lies NOTOCHORD. PHARYNX. 75 in the stalk and opens into the buccal cavity and the wider head which projects into the proboscis-cavity and in which the lumen may be much reduced and not easy to see. At the junction of the two there is a ventrally directed csecaldiverticulum (Figs. 60, 64). The walls of this structure are formed of endoderm which is continuous with the epithelium lining the buccal cavity and consists of cells which have undergone a peculiar modifica- tion, very similar to that of the notochord cells in Amphioxus and in the embryos of the Vertebrata. That is to say, they are much vacuolated and the protoplasm and nucleus occupy but a small portion of them. There is a basement membrane round the whole structure. This, which may be compared to the sheath of the vertebrate notochord. is much developed as a skeletal tissue ventrally and posteriorly (Fig. 60, 19, Fig. 64, n). We shall describe it later on. There is considerable variability in the structure of the notochord. In transverse section the neck is often semilunar, but its cavity may be broken up into several separate spaces. This is particularly liable to happen in the neighbourhood of the ventral appendage which in some species is very massively developed. In some species portions of the proboscis skeleton may penetrate the substance of the notochord (Harrimania kupfferi, Balanoglossus biminiensis). In Ralanoglossus carnosus some of the noto- chordal cells lining parts of its lumen, which are quite cut off from the rest, are densely ciliated. In the same species the hind end of the notochord splits up into three minute tubes which unite again before joining the mouth. In Stereobalanus canadensis the neck is absent, and the head is therefore quite disconnected from the buccal wall. Speaking generally, it may be said that the notochord behind its ventral diverticulum or caecum, quickly loses its chorda-like character. It is interesting to notice that it is in this posterior portion the skeletal development of the sheath is found. May we draw the inference that anteriorly it has a supporting function, while posteriorly this function is taken over by its sheath ? In Schizo- cardium, Glandiceps, and Spengelia the head is continued forwards as a narrow process which traverses the axis of the proboscis and is either solid or contains a small indistinct lumen. This anterior extension is called the vermiform process of the notochord. The ventral caecum of the noto- L'hord frequently has a considerable lateral extension, whence arise its lateral pouches (e.g. Glossobalanus ruficollis). The lumen of these struc- tures is often cut off from the main lumen. The pharynx. The anterior part of the branchiogenital region is occupied by the pharynx, which, as in other Chordata, con- stitutes the branchial * portion of the alimentary canal. The gill- : Punnett suggests, as it seems to us with some truth, that the 'unction of the branchial wall with its perforations is more that of a sieve than of a branchial organ, allowing the water which enters by the alimentary 76 BALANOOLOSSIDA. apertures are, except in Ptychodera and Stereobalanus canadensis, tubes, which lead outward to open externally by the branchial pores placed in the branchial groove on each side of the middle dorsal line. The internal openings (gill-slits) are U-shaped, the 1011 1 .0. 111. aliifsoliiiltniiis iiinnitiix. transverse section through the lirauchuil region, .-oinewliat diagrammatic (alter Spem-el in mi Lang). / dorsal nerve eord ; 2 dorsal blood-vessel: 3 bran- chial (submedian) groove ; /ectoderm; .j^ dorsal P ar t "I <-avil\ oi pharynx; Hi h\ pobraiu-liial groove of pharynx, in this region very largely developed. dors il margin having a tongue-like projection, forming a tonguc- 1 r (Figs. 60, 62) as in Amphioxus. This loads into a pouch canal \vith the food to escape. On this view, the skin, which contains a rich plexus lit' liliiud \-(.-sscls. vvmild art as tlu> main nroan of respiration. PHARYNX. 77 1 having the same vertical extent as the gill-slit and placed close to the pharyngeal wall (Fig. 61). This branchial pouch opens to the exterior by a small aperture the branchial pore. The whole branchial passage is lined by endoderm and is developed as a pharyngeal diverticulum. In the forms just mentioned, the U-shaped gill-slits open directly to the exterior and there are no branchial pouches or pores distinct from them. This is due to the fact that the original pharyngeal outgrowth remains short and its external opening becomes as wide as the internal. The gill- apertures are added to throughout life and arise as small round perforations at the hind end of the pharynx. The tongue is de- veloped as a down-growth of the dorsal wall as in Amphioxus, but hangs freely, not as in that animal becoming con- nected with the ventral margin (Figs. 60, 62). It receives a prolongation of the trunk coelom (p. 92). Synapticula (Fig. 62) are present in some species, traversing the two limbs of the U-shaped aperture. They are connected to the tongue bar, and vary considerably in number in the different species ; in Bal. carnosus there may be as many as 30 on each side of a tongue bar ; 10 to 1 1 seems however to be the usual num- ber. They are absent in the Harri- maniidae and in Glandiceps. In most forms the gill - slits are . not quite straight, being slightly bowed with the concavity directed forward. The successive branchial pouches are placed close together, and their opposed walls constitute the branchial septa. The first of them always receives the opening of the collar pore (see below), and in Bal. carnosus the first branchial pore opens into the hind end of the medullary canal. The gill-apertures are very numerous. The number varies of course with the growth of the animal, and appears to be highly variable in the different species. It varies from 10 or 11 pair;;, the number found in Dolickoglossus sulcatus, to 700. the number recorded for Bal. aurantiacus. FIG. 62. Diagram of two gill- slits of Ptychodem viewed from the inner sick- (from MacBride). 1 gill-slit; 2 tongue-liar ; 3 synapticulum ; 4 septal or primary liar : -J skeletal rod of primary bar forking ventrally; H skeletal rod of tongue-bar, double. 78 BALAJSTOGLOSSIDA. The walls of the pharynx are supported by a definite skeleton recalling the branchial skeleton of Amphioxus (see p. 91). The gill-slits of the two sides are separated from one another dorsally by a narrow streak of endoderm called the epibranchial streak (Fig. 63, 1). Their ventral extension is more variable. In Schizocardium they almost meet ventrally, being separated only by a narrow band of endoderm which is called the hypobranchial streak (Fig. 63, C). In Dolichoglossus and Glandiceps they extend half way down the pharyngeal wall, being separated by a wide tract of continuous endoderm, which may also be called the hypobranchial streak (Fig. 63, A and B). Lastly in the Ptychoderidie and in Spengelia. they only extend half way down the pharyngeal wall, and the hypobranchial portion is wide as it is in Dolichoglossus ; but at the junction of hypobranchial streak (from MacBride). the two parts the lining of the pharynx is much thickened, form- ing the so-called parabranchial ridges (Fig. 61, 13} and its lumen is constricted in such a way that the whole pharynx has in section a figure-of-8 shape. The result of this arrangement is that the pharynx is almost divided into two tubes a dorsal tube in connexion with the gill-slits and a ventral tube com- parable to the hypobranchial streak of other forms. In ffldiidiceps two conditions are found ; in some species the arrange- ment resembles that found in Dolichoglossus, in others the condition described for Schizocardium is repeated. It seems obvious to compare the epibranchial streak with the hyper- l>liupyngeal groove of Amphioxus and the hypobranchial which is so much developed in Ptychodera and Spengelia with the hypopharyngeal irroove or endostyle* of that form. The pharynx sometimes retains its * The term oesophagus is sometimes applied to the ventral portion of the pharynx, but this is obviously a misnomer and must be rejected. INTESTINE. 79 distinctness for a short distance behind the gill-slits as a groove on the dorsal wall of the oesophagus ; this has been called the postbnnn-hinl canal. The pharynx is followed by the oesophagus (or afferent in- testine as it has been called), which extends throughout the hinder part of the branchiogenital region and leads into the hepatic intestine. This structure occupies the anterior part of the abdominal region of the body and is distinguished by the fact that its walls contain green or brown pigment globules. In the Ptychoderidfe and in Schizocardium, it gives off the hepatic diverticula already referred to. These are arranged in two dorso-lateral rows, one on each side, and open into the intestine by transverse slits. The boundary between the oesophagus and hepatic intestine is not very sharp, especially when the hepatic diverticula are not present. In Glandiceps hacksi there is a kind of siphonal tube or accessory in- testine in the hepatic region on the dorsal side. It leaves the hepatic in- testine at about the middle of its length and joins it again at its hind end. Spengelia alba is distinguished by possessing hepatic diverticula which do not cause external sacculations of the body wall. In some species the oesophagus and anterior part of the hepatic intestine - sends off dorso-laterally short canals which open to the exterior by pores which are placed on each side of the dorsal surface in a line with the branch- ial pores. These pores are in two sets. The anterior set comprises the so-called unpaired intestinal pores. These are placed in the part of the branchiogenital region immediately following the pharynx, i.e. on the anterior part of the oesophagus, and as their name implies are usually unpaired. They are found in Schizocardium brasiliense, \ 3 on one side and 16 on the other ; in Glandiceps hacksi 9 altogether, one on the right side and the rest on the left. The posterior set is known as the paired intestinal pores. These are placed at the hind end of the. branchiogenital region, or front end of the hepatic, and communicate either with the hind end of the oesophagus or front end of the hepatic intestine. They are found in Schizocardium brasiliense (one pair), Glandiceps hacksi (three pairs), Dolichoglossus kowalevskii (four to six pairs) ; also in D. mere- schkowskii and in Spengelia alba. The hepatic intestine passes without any strong line of de- marcation into the intestine proper or efferent intestine which occupies the hinder part of the abdominal region. This section becomes somewhat narrower behind where it forms the rectum which opens outward by the wide terminal anus. In most Ptychoderidse a median rod of endoderm is marked <> ft' from the ventral wall of the hinder part of the intestine. It occurs as a thickening of the gut wall, or as a tube with an interrupted lumen opening at both ends into the gut, and has been called the pygoehord. It lies in what may be called the caudal part of the body and extends back to the anus. 80 BALANOGLOSSIDA. The coelom, though largely filled up by muscular and con- nective tissues in the adult, presents the arrangement which, as we have seen in the first chapter of the second volume (p. 7), is on the whole characteristic of the Chordata (see Fig. 72). It is present in three distinct parts which remain separate from one another : these are the unpaired preoral or proboscideal sac, the I cured collar sacs, and the paired trunk sacs. In the embryo these sacs all contain distinct cavities and have epithelial walls (Figs. 72, 73). In the adult the walls have increased in thickness and have become differentiated into the muscular and connective tissues of the body ; while the cavitiss are much reduced, being encroached upon and partly obliterated by these tissues. The coelom of the proboscis. The wall of the coelomic sac of the proboscis consists of a thin layer of circular muscles lying next the basement membrane of the ectoderm and inside this of a thick layer of longitudinal muscles, which are often arranged in bands and which surround the central cavity (Fig. 65). Within the longitudinal muscles and extending amongst them is a layer of loose stellate connective tissue which partly fills up the proboscis anteriorly. Into the posterior region of the proboscis cavity there projects a complicated structure known as the central complex or the basal organ of the prDboscis (Figs. 60, 64). The central complex. \\ liich consists of the notochord and certain vascular organs, is almost entirely covered towards the proboscis cavity by a layer of coelomic epithelium, and is attached to the body wall in the dorsal and ventral middle lines by the dorsal and ventral septa of the proboscis (Fig. 65, i, 6). The posterior part of the pro- boscis cavity is therefore double and lined, at any rate posteriorly in the region of the dorsal and ventral canals (see below), with coelomic epithelium ; while the anterior part (solid in Harrimania kupfferi) is single (except in Willeyia in which it is double) and without, so far as can be seen, a distinct layer of coelomic epithe- lium. Each of the posterior halves of the proboscis cavity is further subdivided, in consequence of lateral expansions of the notochord which meet the body- wall, into a dorsal and a ventral portion (Fig. 66).* In this way are formed the two nl and the two ventral canals of the proboscis coelom. In Fig. 06 the t\vo ventral canals are joined, see below. PROBOSCIS COELOM. 81 The two dorsal canals are lined by coelomic epithelium and extend back into the proboscis stalk (eh), where the left of them, and sometimes the right also, opens to the exterior on the dorsal side close to the junction of the stalk with the body, by the proboscis pore or pores. The ventral canals also extend back to the hind end of the proboscis stalk as far as the mouth, where they end blindly (racemose organ) or, at most, join one another owing to a deficiency in the ventral septum posteriorly (Fig. 66, eh v). The anterior end of the central complex projects into the proboscis coelom to a variable extent in the different species. anp. pup. vrv TIG. 64. Diagrammatic reconstruction of the anterior end of Spengelia porosa, x 10 (after Ptmm-tt). The dotted line in front of the notochord indicates the forward extension of the pericardia! auricles. The heart is omitted, anp anterior neuropore ; en posterior cornu of nuchal skeleton ; ds dorsal septum ; mdv dorso-ventral muscles (muscle-plate) ; ml longitudinal muscles ; n nuchal or proboscis skeleton ; p proboscis coelom ; -per peri- cardium : ///,') posterior neuropore : r dorsal root of medullary nerve cord ; vrv ventral recurrent vessel ; vs ventral septum. Ill some species the notochord is continued forwards in front of the other organs of the central complex for a considerable dis- tance as the vermiform process (p. 75). However far it may extend this anterior extension of the notochord in front of the pericardium (see below) is connected to the dorsal and ventral proboscis wall by a median septum consisting of dorso-ventral muscle fibres which traverse the proboscis cavity and are in- serted on to the notochord. These muscle fibres constitute the dorso-ventral muscle plate of Spengel (Fig. 64, mdv).* An- * The transverse sections figured pass behind this muscular septum, excepting possibly dorsally in Fig. 65 (see footnote, p. 83). /. Ill O 82 BALANOGLOSSIDA. 11 teriorly this septum ends with a free edge, the anterior part of the proboscis cavity being single, but posteriorly it overlaps, though it does not join, the dorsal and ventral septa above referred to. In this overlapping part the muscle fibres of the dorso-ventral muscle plate diverge on reaching the central complex and em- brace that structure. The ventral septum consists of a layer of basement membrane which is continuous with the basement membrane of the body wall and with that surrounding the notochord (Fig. 64, vs ; Fig. 65. 6 ; Fig. 67. .r). It ends with a free edge in front, rarely extending be- yond the front end of the central complex, some- times not so far. Posteriorly it either extends to the very hind end of the ventral canals, c o m - pletely separat- ing them, or it stops short of this, ending with a free edge, so that the ventral canals join pos- teriorly (Fig. 66). The basement membrane of the ventral sep- tum at the point where it joins the basement membrane of the ventral proboscis wall is pierced by the fibres of the circular muscular layer. The dorsal septum, which separates the dorsal canals, is formed by the walls of the pericardium. 1 lie basement membrane of which is continuous with the basement membrane of the body wall in the dorsal middle line (Fig. 65). It is usually continued forwards as an effective septum by the fibres of the dorso-ventral muscle plate (Fig. 64), which may overlap the anterior portion Kic. ita. Diagram of a transverse section through the proboscis "i ;i Itulunoglossid (from L;mg after Spengel). 1 dorsal septum ; 3 ectoderm; '', i lood vessels of the integument ; 4 circular, .; longitudinal muscular layer; 6 vcnti.-il -I'ptuin : 7 noto- chord ; 8 coelom ; 9 heart ; 10 alomerulus ; 11 pericardium. PROBOSCIS COELOM. 83 of the pericardium,* in the region anterior to the dilated part of the notochord, where the dorsal canals acquire their coelomic epithelial lining. f The extension of the ventral septum and of the median muscle septum of the proboscis exhibits considerable diversity in the group. It is now necessary to describe more in detail the posterior more definite portions of the proboscis coelom, where an epithelial lining is pre- sent. First of all the ventral canals ; they nearly always join posteriorly t o f o r m the ventral ccectun which extends back to the root of the proboscis (Fig. 66, ehv), and the hind end of which in --eh Sk 3 ehv FIG. 66. Transverse section through the neck of the proboscis of Balanoglossus gigas behind the ventral septum (after Spengel). eh dorsal canal of coelom, leading on the left side to the pore ; h tissue in the pericardium ; div notochord (buccal diverticuluni) ; sk pro- boscis skeleton ; ehv hind end of ventral canal (ventral caecum). The efferent vessels of the proboscis are shown on each side dorso- laterallv of the notochord. some species forms a small projection (the racemose or- gan) from the roof of the buc- cal cavity. The dorsal canals are more complicated. They pass backwards in the neck of the proboscis as narrow canals lined with a ciliated epithe- lium (Fig. 66 eh), and they either end blindly or open, at the hind end of the neck, into terminal bladder-like dilatations. These two end-vesicles open to the exterior by the proboscis pores. As a rule there is only one proboscis pore and end- vesicle, viz. on the left side, but when there are two, those of the * It appears to have done this in the section from which Fig. 65 was taken. f It will be remembered that anteriorly the proboscis coelom appears to be without epithelium in the adult (see p. 80). 84 BALANOGLOSSIDA. left side are nearly always larger than those of the right. The lining of the hind end of the dorsal canals, whether they end blindly or not. appears to give off nests of cells into the chondroid skeletal tissue (see below), which is found in the neck of the proboscis. There is the greatest variation in the proboscis pores and in the dorsal IM mils ;ui) The right vesicle is larger than the left, and neither of them are in communication with their dorsal canals. Finally, in Bal. carnosus, in which only the left pore and vesicle are present, the end vesicle is continued behind the an- terior neuropore as two csecal pockets placed ventral to the neural tube, and in some specimens the pore opens into the front end of the medullary t nl ie (for the behaviour of the first branchial pore in this species, see p. 77). \Vhen there are two pores the right is said to arise later in development than the left. The lining of the end vesicles is said to be due to an ingrowth of ectoderm : if this is true the actual coelomic pore would be th" aperture connecting; the vesiele \\ith the dorsal eoelomic canal. The central complex of the proboscis (Figs, 60, 64, 65, 67). Before proceeding to our account of the coelom in the collar, it will be convenient to describe this remarkable structure. It occupies almost the whole of the neck of the proboscis and in the dorsal middle line is in contact with the basement membrane of the ectoderm (Fig. <>7). On the ventral side of it is the ventral extension of the proboscis coelom. In some species (Ptychodera flava and cn/!//rt, etc.) the ventral canal of the coelom extends back as far as the union of the notochord with the buccal epithelium, but in most it stops short of this and the posterior pirt of the central complex is in contact in the ventral middle CENTRAL COMPLEX. 85 lm line with the basement membrane of the ventral ectoderm of the proboscis stalk. The central complex consists of the following parts : (1) ventrally the notochord and, in the stalk, the skeletal develop- ments of the basement membrane of the notochord (Fig. 64, n ; Fig. 66, sk ; Fig. 60, 19), (2) the heart (Fig. 67, b, see also Figs. 60, 65), which is a blood vessel lying on the dorsal side of the notochord, between it and (3) the pericardium (Fig. 64, per, Fig. 67, A. see also Figs. 60, 65), a closed sac of con- siderable size and of doubtful meaning, and (4) the glomerulus or proboscis gland which is the well - developed and folded coelomic epithelium lying on the front wall and anterior part of the sides of the pericar- dium (Fig. 65, 10, Fig. 60, 2). The notochord has already been described. [t extends along the whole length of the cen- tral complex and varies very considerably throughout the group. The basement membrane which lies immediately outside its epithelium forms its sheath and is especially thickened on the ventral side of the posterior part of it (Fig. 64, n, Fig. 66, sk). This specially thickened part of the notochordal sheath is continuous with a specially thickened part of the basement membrane of the dorso-lateral buccal wall (Fig. 64 en), and the whole constitutes the proboscis or nuchal skeleton (Fig. 68). The proboscis skeleton consists of a median body and of two limbs diverging backwards. The body is placed ventral to the posterior part of the notochord in the neck of the epi 7' sv FIG. 67. Transverse section through the hind end of the proboscis, near the.neck, of Glossobalanus minutus (after Spengel). The section goes behind the glomerulus. The epidermis is dotted and the nervous layer of the epidermis is diagonally shaded, the basement mem- brane of the epidermis "is unshaded, adv blood vessels of skin ; 6 heart ; div notochord ; eh proboscis coelom divided here into a right and left half by the anterior end of ventral septum sv ; epv coelomic epithelium ; h pericardium ; lm longitudinal muscles of proboscis wall ; rev efferent vessels of proboscis ; rm circular muscles ; sv ventral septum. 86 BALANOGLOSSIDA. i proboscis. It extends from the hinder side of the notochordal caucum in front to the junction of the notochord with the mouth behind (Fig. 64, n). At this point it divides into the two limbs which diverge backwards into the collar region lying immediately within the buccal epithelium. The limbs may be long, extending al- most through the whole length of the collar, or they may be short reaching but a little way. The proboscis skeleton is strengthened laterally in the hinder part of the neck of the proboscis by the chon- droid tissue. This tissue is the enlarged basement membrane be- longing to the anterior wall of the collar coelom and the posterior part of the dorsal canals of the FIG. 68. Proboscis skeleton of oios- proboscis coelorn. It differs from sobalanus sarniensis (after Spengel). the rest of the skeletal tissue of the body in containing nests of cells which have migrated into it from the adjacent coelomic epithelial walls (p. 84). The chondroid tissue, which has a certain resemblance to hyaline cartilage, is specially developed in Schizocardium and ''/(indiceps. The heart is a blood vessel lying in the basement membrane between the notochord and the pericardium (Figs. 65, 67). Its connexions will be described later on when we are dealing with the vascular system. The pericardium or proboscis sac is a closed vesicle lying over the anterior end of the notochord and extending only for a short distance backwards into the stalk (Figs. 64, 65, 67). This posterior part of it is traversed by transversely directed fibres ; its anterior part is filled up with a peculiar loose tissue of unknown function while its main body contains a clear fluid. It possesses an epithelial lining which in some cases may proliferate so as almost entirely to fill up its cavity. It has no connexion with the vascular system, and the latest observers have asserted that it is quite closed off from the coelom, but Bateson maintained that its cavity com- municates with the coelom through the loose tissue filling up its anterior part. This loose tissue sometimes occupies a large GLOMERULUS. COLLAR COELOM. 87 portion of its cavity (e.g. Glossobalanus ruficollis, etc.). Its ventral wall is concave, slightly enveloping the dorsal sides of the notochord, and contains a number of transverse muscular fibres which being in contact with the heart tube very likely cause the pulsations by means of which the blood fluid is moved. In Schizocardium (hence the name) and in Glandiceps the anterior lateral portions of the pericardium are produced into two tubes which extend along the vermiform process of the notochord and constitute the auricles of the pericardium. In Bal. carnosus the pericardium is bifid anteriorly, being produced into two pouches, accompanied by glomerulus tissue, beyond the anterior limit of the notochord. As already stated the dorsal wall of the posterior part of the pericardium is in contact with the basement membrane of the ectoderm, thus giving rise to the dorsal septum ; but its anterior wall and the anterior part of its dorsal wall, and its side walls are covered by coelomic epithelium which is much folded and composed of large cells containing pigment grains. This is the glomerulus. Within the folds, i.e. between their epithelium and the pericardial wall are some blood vessels which are in direct communication with the heart. If the glomerulus is an excre- tory organ, as is supposed by some, its secretion must pass into the proboscis coelom, reaching the exterior by the proboscis pore. The collar coelom is in many species almost entirely filled up by the muscular and connective-tissue development of its walls, and, except in the parts of it known as the collar canals, is devoid of a lining epithelium. The two lateral sacs of which it originally consists meet dorsally above the collar nerve-cord and ventrally below the gut (Fig. 69). The longitudinal mesenteries formed by the opposed walls of these sacs persist only in part : the ventral mesentery is present in the posterior region for a short distance, but is deficient over the greater part of the collar ; the dorsal mesentery which extends between the nerve cord and the skin persists over the greater part of its length, but is deficient anteriorly. In Harrimania kupfferi both mesenteries are absent in the adult, and it is quite possible that there may be consider- able variation in them in other species. In some genera (Ptychodera, Schizocardium, Spengelia and Dolichoglossus), the coelomic sacs of the collar are in parts separated from the gut wall by the peripharyngeal cavities (Fig. 69) 88 BALANOGLOSSIDA. \vhich are forward extensions of the trunk coelom and will be described later. The collar canals are two short tubes lined by a ciliated epithe- lium, and placing the collar-coelom in communication with the .-xterior. They occur dorso-laterally at the hind end ot the collar, and they open externally by the collar pores into the first branchial pouch. Internally they open by a funnel-shaped aperture into the collar coelom, which, in the neighbourhood of the funnel, is free ..-- 1 ---8 from the tissue ele- ments which pervade it elsewhere. The col- lar canals are covered towards the coelom by a layer of flat epithelium, and their dorsal walls are often infolded so as to pro- ject into the cavity of the canal.* The visceral wall of the collar coelom pre- sents along a certain line on each side a fold containing blood vessels and called the vascular fold. In the Ptychoder ida? the vascular fold begins at the hind end of the collar region and extends along the ventral surface of the alimentary canal in the middle line to a point not far removed from the anterior end of the collar. Here it divides into two folds which pass transversely one on each side to the dorsal surface. In the other genera the fold is double throughout its course. It begins at the hind end of the collar ventrally and ascends obliquely dorsalwards and forwards on each side to the point where the neck of the proboscis joins the body. The vascular folds contain blood vessels which unite the median FIG. (19. Diagram of a transverse section through collar region of a Balanoglossid to show the relations of the body-cavities in the collar (from MacBride). The coelo- mic Spaces are represented as clear of tissue and the ventral mesentery as persistent. 1 medullary ner\e cord ; 2 dorsal blood vessel ; 3 perihaemal cavity ; 4 notochord at its point of attachment to the huccal wall ; 5 posterior diverging crura of the proboscis skeleton; 6 collar coduin ; r Imccal cavity; g peripharyngeal cavity; 9 ventral vessel. This fold ivi-.ill-. .1 similiir -inn-lure projecting into the sand canal of TRUNK COELOM. 89 ventral vessel of the trunk with the efferent vessels of the heart. They are obviously to be compared with the glomerulus of the proboscis region. The trunk coelom extends throughout the whole of the body behind the collar. It is a more continuous cavity than that of the collar and is lined throughout by an epithelium and does not open * to the exterior. The dorsal and ventral mesenteries are persistent over the greater part of their course. The coelom of the trunk, though completely separated from that of the collar, sends forwards diverticula into the collar region. There are two pairs of these extensions. Dorsally on each side of the dorsal blood vessel of the collar, between the collar nerve-cord and the gut, there is a tubular extension of the trunk coelom which extends throughout the whole collar region as far as the insertion of the neck of the proboscis ; these are the parihaemal cavities (Fig. 69, j). They are largely occupied by longitudinal muscular tissue which is almost entirely developed from their dorsal walls. The other pair of collar extensions of the trunk coelom constitute the peripharyngeal cavities (Fig. 69, 8). These are not present in all genera (see above, p. 87). They lie between the ventral wall of the collar coelom and the gut. In the Ptychoderidse they entirely surround the gut except in the dorsal and middle ventral lines as far forwards as the transversely directed vascular folds described on p. 88. In the other genera in which they are present, they also extend as far forwards as the vascular folds, but these being obliquely directed, they occupy a triangular area on each side of the pharynx. In the Ptychoderidce the anterior part of the trunk coelom is divided on each side into a dorsal and ventral portion by a septum called the lateral septum or accessory mesentery (Fig. 70). The lateral septum passes in the hinder part of its extent from the body-wall, where it is inserted along the submedian line, to the gut wall ; while anteriorly it shifts its gut- wall attachment to the body wall near the dorsal mesentery (Fig. 70, B). Moreover, the dorsal section of the coelom gradually dwindles anteriorly and ceases altogether in the branchial region. Posteriorly the lateral septum ends freely and the dorsal cavity opens into the posterior part of the coelom. * Willey has described structures in Spengelia, in relation with the first gill-pouch and the perihaemal cavities, which he regards as vestiges of a pair of coelomic pores for the trunk coelom. 90 BALANOGLOSSIDA. FIG. 70. Diagram- <>f transverse s.-rti<>ns through the trunk of Ptychoderu in show the insertions dy wall of the collar are developments of the parie- tal wall uf the collar coelom. They consist of an external layer of longi- tudinal fibres and in the anterior region of an inner circular layer as well. In the trunk the body wall muscles diminish in strength posteriorly. They consist of a layer of longitudinal fibres only, except in the Ptychoderidce in which there is an outer layer of circular fibres as well. The layer of longitudinal fibres is interrupted in the dorsal and ventral middle lines and in the submedian lines (where the branchial pores and gonads open). The visceral muscles of the collar are rather complicated, part of them being derived from the extensions of the trunk coelomic sacs. They are partly longitudinal and partly transverse. Briefly, it may be said that the visceral wall of the collar coelom where it is in contact with the outer wall of the peripharyngeal extensions consists of longitudinal muscular fibres, whereas, where it is in contact with the gut wall, it gives rise to trans- versely directed fibres. For instance, in Ptychodera all the buccal wall behind the vascular fold possesses longitudinal fibres, whereas in front of the vascular fold, transverse fibres only are found. In the other genera the longitudinal fibres are found over the triangular area referred to above (p. 89), while only the transverse fibres are found on the part in front of the vascular folds. The inner wall of the peripharyngeal cavities con- sisN dt' circular fibres which, together with the partial layer of circular fibres belonging to the visceral wall of the collar coelom, complete the trans- verse layer of fibres round the gut in the collar region. In addition to the above muscles, radial fibres are found passing across the collar coelom from the visceral to the parietal wall. In the trunk 1'egioii the coelomic space i> traversed by radial muscular band* and the visceral muscles. The dorsal walls of the periliaemal cavities give rise to longitudinal muscles which almost entirely iill them. Muscles art? also in some genera i le\ eloped in the ventral walls oi these cavities. There are also fibres. \\hidi traverse the periliaemal cavitio. CONNECTIVE AND SKELETAL TISSUES. 91 The muscular fibres seem to be unstriated, but Bateson de- tected signs of a faint transverse striation in Dolichoglossus koivalevskii. In addition to the muscles above described, muscular fibres are present in the walls of the branchial pouches and in the anterior wall of the collar coelom. The connective and skeletal tissues and basement membrane. The Balanoglossida present the remarkable feature of having no connective tissues in the ordinary sense of the word, excepting the delicate reticulate tissue found in the coelomic cavities and extending amongst the muscles. This connective tissue is, as already mentioned, a product of the walls of the embryonic coelomic sacs and appears to be of the nature of mesenchyme. On the other hand all the epithelia of the body possess a kind of internal cuticle, that is to say they secrete on their inner- surfaces a continuous structureless membrane devoid of nuclei and fibrous structures. This structureless layer is called the basement membrane. It is formed not only on the internal side of the ectoderm cells but also on the inner side of the endo- derm and of the walls of the coelomic sacs, even in cases in which these walls have become entirely converted into connective tissue and muscles. It follows from this that not only must there be a layer of basement membrane interposed between all the organs of the body, but also that this interposed membrane must, theoretically at any rate, be in all cases a double membrane, one of its laminae being derived from one and the other from the other of the two organs which are in contact. As a matter of fact this doubleness of the basement membrane is not as a rule discernible, the two laminae of which it is theoretically composed having completely fused. But a separation persists in places and the cavities thus formed constitute the blood vessels (see below). The skeletal tissues are entirely derived from this membrane. The proboscis skeleton has already been described. Here it is only necessary to call attention to the fact that one part of this skeleton the chondroid tissue differs from the rest of the basement membrane of the body in containing strings and nests of nuclei which have migrated into it from the adjacent coelomic walls. The branchial or pharyngeal skeleton consists of a special 92 BALANOGLOSSIDA. thickening of the basement membrane of the walls of the pharynx on each side the gill-slits. It is made up on each side of a series of structures placed vertically in the walls of the pharynx and presenting a resemblance to a three-pronged fork (Fig. 62). The three prongs of each piece are joined together dorsally and end freely ventrally. The central prong is single (though presenting the appearance when closely examined of being composed of two fused rods) and forks ventrally. It is contained in the septum separating two gill-slits. The anterior and posterior prongs lie in the adjacent tongue bars, which also contain a corresponding prong of an adjacent skeletal piece. The anterior and posterior prongs do not fork ventrally. It would thus appear that each tongue bar contains two skeletal rods belonging to adjacent skeletal pieces. The whole skeleton lies close to the wall of the pharynx. Coelom is present in the tongue bars and its epithelium contributes with that of the adjacent endoderm to the formation of this skeletal basement membrane (Fig. 61). In the primary or septal bars there is no coelom in the adult, so that the skeletal rods in them are exclusively formed by the pharyngeal epithelium. The vascular system consists, as already explained, of channels hollowed out in the basement membranes of the body. It might be said to represent a persistent portion of the spare which theoretically occurs between the two laminse of which all or nearly all the basement membranes consist. But as the basement membranes do not show a composition of two lamella 1 it will be perhaps safer at this point to take the first statement as representing all that we actually k>itr on this subject. The blood is usually described as colourless, but it appears to be coloured red in some forms at least, if we may judge from the fact that a red line can be seen through the body wall along the lines of the chief blood vessels ; and it contains a few floating amoeboid cells of the ordinary type. An epithelial lining is present in the blood vessels of a few forms (Ptychodcra, etc.), but frequently nothing of the kind can be detected. Some of the larger vessels are provided with muscles (usually circular), which arc furnished by the walls of adjacent portions of the coelom. The principal blood vessels are as follows : (1) A longitudinal dorsal vessel running through the body and collar, and passing into the proboscis to become VASCULAR SYSTEM. GONADS. continuous with the heart. The proboscis portion of this vessel is called the afferent vessel of the proboscis. It runs in the dorsal mesentery, or in what remains of it, in the trunk, while in the collar it is placed between the two perihaemal cavities (Fig. 69). In the proboscis it runs between the notochord and the hind end of the pericardium where it is enlarged to form the heart (Figs. 60, 65). (2) A ventral longitudinal vessel extending through the whole trunk and into the hinder part of the collar (Fig. 69). It runs in the ventral mesentery or in what remains of it. In the collar the ventral vessel divides and becomes con- tinuous with the two plexuses of vessels which are contained in the vascular folds already described. These pass in front into the efferent vessels of the proboscis (Fig. 67 rev) which run, one on each side, through the chondroid tissue in the neck of the proboscis to the hind end of the glomerulus, with the blood spaces of which they are continuous. There is in all the basement membranes a capillary network by means of which these main vessels are connected. This network is principally developed in the body wall, in the gut wall, particularly in the hepatic region, and in the gonads. That part of this network which lies in the pharynx wall and is presumably of importance in respiration is of course directly connected with the dorsal vessel. The dorsal vessel is supposed to be contractile and its contractions prob- ably travel forwards, so that the blood passes from it to the heart. From the heart it passes, driven by the contraction of the transverse muscle in the ventral wall of the pericardium, to the spaces within the folds of the glomerulus. Thence it passes into the efferent vessels of the proboscis which are continued from the hind end of the glomerulus through the chondroid tissue into the vessels of the vascular folds and so into the ventral trunk vessel. There are other main vessels in addition to the two mentioned : e.g., in Ptychodera there is on each side in the lateral septa a vessel which arises in the skin in front and enters the intestinal plexus behind ; it supplies the i:i>ii,ids. There are also definite main vessels both in the tongue and prim- ary bars of the pharynx, supplying the capillary network. The reproductive organs. The sexes are separate and the reproductive organs are similar in form and arrangement in the two sexes. They are simple or branched sacs which project into the coelom and are placed in the lateral part of the trunk in the 94 BALANOGLOSSIDA. branchiogenital region (p. 69). In the simplest cases they form on each side one series, the series of primary gonads, which open to the exterior on the dorsal surface along a line which may be called the gonaducal or sub-median line. The submedian line lies along the insertion of the lateral septum, where the longi- tudinal muscular layer is broken, and in the branchial region frequently coincides with the branchial groove, but in Ptycho- deridae it is placed at a greater or less interval from the branchial groove and the branchial pores perforate the dorsal longitudinal muscles. In the branchial region the primary genital sacs open externally to the branchial pores (Fig. 61), with which they correspond roughly in number. The wall of the genital sacs consists internally of a layer of germinal epithelium which is continuous with the ectoderm through the external opening. Outside this is a layer of base- ment membrane containing an abundant capillary plexus or even a continuous sinus. Then come some muscular fibres and a layer of coslomic epithelium towards the body cavity. It would appear that new gonadial sacs are continually being formed at the hind end of the genital region. In Ptychodera asymmetrica gonads are present only on the left side * (Punnett). The arrangement just described, in which the so-called primary series of gonads alone is present, is the simplest found and is possibly charac- teristic of all in the young state. It usually becomes more complicated with growth in the following way. The genital sacs become lobed and the lobes acquire independent openings to the exterior (Schizocardium, brasil- iense, Glandiceps talaboti). The secondary pores, as these are called, are not placed in the line of the primary pores, but are internal or external to them, except in the branchial region, where there are no secondary pores internal to the primary pores. The secondary pores usually open on the submedian line, but some of them may perforate the dorsal longitudinal muscles. A further complication is reached by the complete separation off of some of the lobes, to form a number of accessory glands in addition to the original gland which is called the primary gland. These accessory glands are not pl;irrd in the series of primary glands but in longitudinal rows there may bo several either external or internal to the primary row. They acquire their own external openings, which, inasmuch as they belong to the accessory grate the longitudinal muscles. * Agreeing in this rcsjr-ct with some species of Amphioxus (see Gold- -'hmidt on Amphioxid.es, Wiss. Ergebnisse der " Valdivia " Expedition, Bd. \->, Lf. 1, 1905, and Zo,,l. !/,;, iger, 30, 1906, p. 443). GONADS. OVA. TORN ARIA. 95 In Stereobalanus canadensis there are several rmvs of gonads, both inside and outside the branchial pores. Their openings are all in the submedian line which is much widened out. In Ptychodera there are usually several rows of accessory gonads which are for the most part external to the primary gonads. The extension of the genital organs varies in the different species. In some species (e.g. Ptychodera flava) they are found throughout the branchio- genital region, beginning just behind the collar ; in other species (e.g. Glandiceps abyssicola) they are confined to that part of the branchio-genital region which lies behind the branchial pores ; while in Stereobalanus cana- densis they are coextensive with the branchial region. In the majority of species, however, they begin in the branchial region at a variable distance behind the collar. In almost all cases the branchio-genital region over- laps the hepatic region, so that genital sacs are found at the anterior part of the latter. In Bal. carnosns, however, the transition is abrupt and no genital sacs are found in the hepatic region. In most of the Ptychoderidse the side walls of the body in the branchiogenital region are prolonged into winglike folds the genital pleurae. The genital sacs are for the most part con- tained in these wing-like outgrowths, but not entirely, for in some species they are found in the body as well, mediae! of the insertion of the lateral septum. The genital pleurse arc extremely mobile structures and may be bent up so as to cover the back of the animal. In Bal. carnosus their free opposed edges may become united by mucus, so that a cavity, widely open behind and receiving the branchial pores, is formed on the dorsal side of the branchiogenital region. The gametes are produced from the walls of the gonads and when ripe are passed out to the exterior through the genital pores. The spermatozoa have spherical or oval heads and active flagelliform tails. The ova are provided with a close- fitting egg-membrane, and are of two sizes according to the amount of yolk and manner of development. In most genera the egg is small, varying from '06 mm. in diame- ter in Ptychodera flava to 15 mm. in Bal. carnosus. The develop- ment of such eggs is probably indirect, passing through a free larval stage, called the Tornaria (Fig. 74). In some cases, however, the eggs are much larger, varying in their longest diameter from - 4 mm. in Dolichoglossus kowalevskii to 1*5 in Harrimania kupfferi. The early stages of the small eggs which probably develop into tornaria are not known, but thanks to the researches of Bateson all the stages of the development of Dolichoglossus 96 BALANOGLOSSIDA. kowalevskii are known. The eggs of this species are laid singly and deposited in the muddy sand which the animal inhabits. Segmentation is regular and complete and leads to the forma- tion of a spherical blastosphere which becomes a gastrula by i imagination (Fig. 71, -4). The blastopore narrows and a circle of cilia is formed round it ; it eventually closes up at that pole of the egg which will become, as shown by the persistent ciliary circle, the future hind end ; i.e. it closes in the posi- tion of the future anus. The embryo becomes covered with fine cilia and a tuft of longer cilia is formed at the anterior end. The embryo elongates and 1 B - e Ar ~ two transversely directed grooves appear, marking out the three regions ; proboscis, collar, and trunk (Fig. 71, B). The embryo hatches at about this stage and lives as a free larva which at first swims in the mud by means of its cilia, and later burrows in the mud by means of its proboscis (Fig. 71, C). A groove appears in the collar re- gion in the dorsal middle line marking the position of the medullary nerve- cord which is now being developed. This groove is a temporary structure and does not appear to participate in the formation of the collar nerve-cord which arises by delamination from the ectoderm. The first gill-slits now appear just behind the collar as a pair of perforations, and the mouth is formed as a minute perforation in the groove between the collar and proboscis. The anus is formed rather later at the hind end. Meanwhile the five coelomic pouches have made their appear- ance ; one unpaired pouch in front and two pairs behind (Fig. 72). These become closed from the gut and persist as the coelomic sacs. All tin- mesodermal tissues are derived from the walls of these sacs. n;. 71. Free-swimming larva; of Dolichoglossus kii'i-ntevskii in different stages of development (after Bateson, from Korschelt and Holder), e proboscis; & branchial pore; kr collar. DEVELOPMENT. 97 The collar nerve cord is delaminated from the ectoderm beneath t-lio transient groove already mentioned. It remains however connected with the ectoderm throughout life at its front and hind ends. It appears to increase in length at the front and hind ends of the collar by invagination of the median ectoderm, by which pro- cess the continuous central canal pre- sent at these points is formed. In the subsequent changes the gill- slits increase in number from before backwards and the adult form is gradu- ally assumed (Fig. 71, C). A larval organ, constituting a suctorial tail for attachment, is developed and lasts for a short time ; it is placed at the hind end ventral to the anus. The noto- chord is formed as a dorso-median groove of the anterior part of the gut which becomes partially constricted off from before backwards to form a tube ; it later extends into the base of the proboscis. The pericardium is devel- oped from a solid mass of cells derived FIG. 72. Diagram of a longitudinal sec- from the wall of the posterior part of the proboscis coelom. The first trace of the generative organs is seen when there are 10 gill-slits. Their exact ori- gin was not made out in this species, but from observations of other species they are probably mesodermal, though by some observers they are believed to be derived from the ectoderm, with which in any case they soon become connected. So far they have not been traced into connexion with the coelom. The proboscis pore is first indicated by a thickening of ectoderm on the base of the ":,7 proboscis. This thickening acquires a cavity which later opens to the exterior and to the proboscis coelom. The collar pores arise by the per- foration of thickened patches of ectoderm in close connex- FIG. 73. Transverse section through the middle ion with the opening of the part of the collar of a larva of Dolichogtosstts kowa- levskii which is at about the stage of Fig. 71 B tion through a larva of Dolichoglossus kowalevskii (after Bateson. from K. and H.). ei anterior (proboscis), en middle (collar), em posterior (trunk) coelomic diverticula, d enteron. n. fir ,, _;il di't At tho nnst prior farst g ul - sllt - At tne P . (after Bateson, from K. and H.). Above is seen the end of the proboscis coelom coelom r n ; f the point where the coelom was connected with the gut in an earlier stage (Fig. 72), a proliferation of the coelomic wall occurs in the dorsal middle line ; this causes a projection into the cavity which Z III. H 98 BALANOGLOSSIDA. Fm. 74. Tornaria larva (after Metschnikoff), a from the side ; 6 from the dorsal surface ; A anus ; C pericardium ; P, P' coelomic sacs ; O mouth ; S apical plate ; W rudiment of the proboscis coeloni (so-called water-vascular sac). eventually reaches the ventral wall. The hinder part of the proboscis coelom thus becomes divided into two parts by a vertical septum. of the cells of this septum give rise to the pericardium, which wh< formed is solid, and later on the notochord grows into it. The tornaria larva. In most species which have eggs of the smaller size, the development is indirect and a transparent pela- gic larva known as the tornaria is developed. This larva was first described by J. Miiller, who took it for an Echinoderm larva. Its real nature was determined by Metschnikoff. The early stages of this development are not known, but the later stages by which the larva passes into the adult have been more or less worked out. In the youngest stage known, the tornaria has a somewhat ovoid form with a ventral mouth, a terminal anus, and an alimentary canal divided into three re- gions, viz., oesophagus, stomach and intestine (Figs. 74, 75). There are two ciliated bands, one of which is preoral and encircles the preoral lobe, and the other postoral but longitudinal in direc- tion. These two bands touch one another dorsally and anteriorly, the ectoderm at the point of junction being thickened to form the apical plate (S). The apical plate possesses nerve fibres and ganglion cells and soon acquires a tuft of immobile cilia. It also contains a pair of pigmented eye spots. Internally there is a spacious blastocoel, in the front part of which is a vesicle called the water-sac (Fig. 74). This is the FIG. 75. Early stage of a tornaria larva in longitudinal section (from Balfour after Goette). m mouth ; an anus ; tv rudiment of pro- boscis coelom (so-called water-vascular vesicle). TORNARIA. ECHINODERM AFFINITIES. 99 rudiment of the proboscis coelom. It lies on the hinder part of the oesophagus and is connected to the apical plate by a muscu- lar band, and opens on the dorsal surface by a short ciliated tube, the rudiment of the proboscis canal and pore of the adult. The pericardium (heart- vesicle), the origin of which is uncertain, lies at first near the skin to the right of the proboscis pore. It soon leaves the surface and becomes surrounded by the proboscis coelom. By some authors it is regarded as the right proboscis cavity. Later a transverse circumanal ciliated ring is formed, and sometimes a second less dis- tinct ring behind this. Moreover the course of the two first described ciliated bands fre- quently becomes complicated and folded, and in Tornaria grena- cheri they acquire ciliated lobe-like projections. In the later development the form of the adult is gradually at- tained ; the cili- ated rings disap- pear and the two posterior pairs of coelomic sacs are developed in a manner which has not been thoroughly ascertained. The collar nerve cord, gill-slits etc., are formed as they are in the direct de- velopment. The tornaria has by its ciliated bands and general form a strong external resemblance to the bipinnaria larva of an echino- derm, a resemblance which is increased by the presence of the so-called water- vesicle with its dorsal pore ; and when first dis- covered it was described as belonging to that group. A close examination of its structure, though revealing points of difference FIG. 76. Two later stages in the development of Tornaria larva, side view. A larva with one pair of gill apertures (after Metsch- nikoff ) ; B larva, with four pairs of gill apertures (after A. Agas- siz) ; Bo gill apertures ; C heart ; O mouth ; P coelomic sac ; W rudiment of proboscis-coelom. 100 BALANOGLOSSIDA. such as the presence of eye-spots and an apical plate, does not nullify this resemblance and renders it extremely probable that the Enteropneusta have affinities with the Echinodermata as well as with the Vertebrata a view which is expressed in this work by the juxtaposition of the chapters dealing with them. The Balanoglossida live in sand or mud* and are found in most seas. Their burrows, the walls of which appear to be cemented into some consistency by the mucus which the animal secretes, appear frequently to be U-shaped, opening on the surface at two points. At one of the openings the casts formed by their faeces are discharged, while the anterior end is in relation with the other opening. In Dolichoglossus kowalevskii, Bateson describes them as being highly coloured (collar a bright orange) and as living at about a depth of eight inches with the anterior end of the body (to the branchial region) and the hind end both vertical, the middle portion being coiled in a corkscrew spiral. They usually possess a disagreeable smell, described as re- sembling that of iodoform or sometimes that of chloride of lime with a faecal admixture. It frequently happens that two species live in association : thus Willey records that Glossobalanus ruficollis inhabits the same burrow as Bed. carnosus in New Britain, while the latter is taken with Spengelia porosa at Lifu. They actively burrow in the ;sand by means of their proboscis and collar, in which the chief muscular development occurs, drawing the hinder part of the body passively after them. The proboscis and collar can be readily rendered turgid or the reverse, and they clear out the sand and mud by swallowing them. Thus their locomotion and alimentation are effected by the same means. It is probable that the inflation of the proboscis and collar which appears to be a necessary condition of their use in locomotion is effected by the taking in of water through the proboscis and collar pores. The body very readily breaks into pieces particularly the hinder part, and some species appear to be able to practise autotomy. In presence of this fragility of body, we should expect the power of regeneration to be considerable, and this appears to be the case. Animals are frequently taken showing * See, however, note on p. 103. REGENERATION. PARASITES. 101 features which are only explicable on the view that injuries are being repaired. Thus individuals are found with a very small proboscis and collar, or with an imperfectly formed collar indicating that these parts are being regenerated. It is asserted that in the regeneration of the collar of Ptychodera flava the collar nerve cord is formed as an open groove which is gradually con- stricted off, and that in the regeneration of the proboscis of the same species the proboscis pores may be equal in size. The first fact has its- counterpart in some tornaria larvae in which the medullary cord arises by the constriction off of an open groove, but the latter has no counterpart so far as is known in the early developmental history, no tornaria being known with two water-pores. Various parasites (Gregarines, flagellate Protozoa, Trema- todes, Nematodes, etc.) are found in the Balanoglossida, the most remarkable perhaps being a parasitic Copepod Ive balano- glossi sometimes found in the genital pleurae. The Balanoglossida are remarkable for the complexity of their structure in contrast with the simplicity of their mode of life and the absence of organs of special sense. There are no known organ- isms whether animal or vegetable, at present living, which pre- sent such complexity of organization combined with such simplic- ity of habits. They possess all the three features of the Chordata (see vol. 2, ch. i.) : namely a notochord developed from the dorsal wall of the enteron ; a tubular central nervous system ; and paired branchial apertures leading outwards from the anterior part of the alimentary canal. In addition they possess complex vascular and muscular systems, and a coelom which, largely filled up by muscular and connective tissue in the adult, consists of three main divisions, viz. an unpaired chamber in the pro- boscis, two paired chambers in the collar, and a similar number of chambers in the trunk. All the mesodermal structures including the gonads develop from the walls of the coelomic sacs, but no relation has as yet been traced between the cavities of the gonads and those of the coelom. There is no relation between the repetition of any of the organs and that of the coelomic sacs. Indeed all the repeated structures of the adult, viz. the gill-slits and the gonads occur in the region of the posterior or trunk sacs. Although there is no regularity in this repetition it is interesting that it should occur here, for it is these posterior sacs which in AmpJiioxus 102 BALANOGLOSSIDA and possibly in the Vertebrata become segmented into the mesoblastic somites. Finally, we must not forget to call attention to the great variability presented by different members of the group. This variability not only occurs among different species, but is also shown by different individuals of the same species. Moreover, and this is the remarkable point, it affects features of structure which to judge by the standard of higher Chordata are of great importance, and are usually perfectly constant through large groups. In illustration of this we may call attention to the condition of the central canal and dorsal roots in the nerve cord of the collar (many species of Ptychoderidse), to the canals and pores which connect the proboscis coelom with the exterior (species of Ptychodera), and to the length of the branchial region as compared with the rest of the body (macro- and brachy- branchiate varieties of Ptychodera) : these features are variable in some cases within the limits of the same species. As examples of characters which vary in allied species and which we should otherwise judge to be important, we may refer to the presence or absence of liver diverticula, to the condition of the notochord, and to the presence or absence of external protective covering of the gill-slits, and finally to the presence or absence of genital pleurae. It is in consequence of this remarkable variability that in our treatment of the group we have entered into much greater detail than has been our custom in this work. Farn. 1. Ptychoderidae. Proboscis usually shorter than the collar: cornua of proboscis skeleton do not extend backwards beyond the middle of the collar. Dorsal unpaired roots unite the medullary cord of the collar with the epidermis. Efferent vessels of proboscis united in one transverse plane by a circular vessel with the ventral blood vessel of the collar. Peri- pharyngeal spaces contain circular muscles and completely surround the buccal cavity continuously up to the level of the mouth opening. Peri- haemal cavities without transverse muscles. Circular muscles outside the longitudinal are usually present in the body wall of the trunk. Hypo- pharyngeal streak as a well-marked groove on the ventral side of the pharynx (Fig. (il). Genital pleurae well developed or small. Lateral mesenteries present in the trunk coelom. External liver-saccules present (except in (llnssobalanus ruficollis, Willey). Ptychodera Esch., the gill- slits open directly to exterior and the genital pleurae have a ventral origin, generally with continuous axial canal in the medullary cord of the collar, Pt. fiava Esch., erythraea Speng., bahamensis Speng. Balanoglossus D. Chi., gill-slits open into pouches which discharge to exterior by dorsal i_nll pm-es, genital pleura* with dorsal origin, B. apertus Speng., clavigerus D. Chi., gifja* F. v. Mull., aurantiacus Girard, australiensis Hill, carnosus BALANOGLOSSIDA. 103 Willey, bimimensis Willey, jamaicensis Willey. Glossobalanus Speng., gill apertures as in Balanoglossus, genital pleurae reduced to ridges, Gl. minutus Kow., sarniensis Koehler, hedleyi Hill, ruftcollis Willey. Fam. 2. Glandicipitidae. Proboscis longer than collar. Notochord fre- quenthy produced anteriorly into a vermiform process. Cornua of proboscis skeleton extend to posterior region of collar. Nerve -roots absent or vestigial. Efferent vessels of proboscis pass obliquely downwards to posterior end of collar. Peripharyngeal spaces separate, vestigial or absent. Perihaemal cavities contain transverse muscles. Circular muscles of body-wall lie inside the longitudinal muscles. Genital pleurae and lateral septa of trunk coelom absent. External liver saccules present or absent. Schizocardium Speng., right and left peripharyngeal cavities and synap- ticula present, ventral septum of proboscis extends to end of vermiform process, external liver saccules present, medial gonads absent, pericardial auricles highly developed, hypopharyngeal streak of pharynx reduced to narrow band (Fig. 63) ; Sch. brasiliense and peruvianum Speng. Spen- gelia Willey, peripharyngeal cavities and synapticula as in Schizocardium, ventral septum of proboscis does not extend to the vermiform process, external liver saccules absent, medial gonads present or absent, pericardial auricles reduced (Fig. 64), dermal pits in the genital region, hypopharyngeal groove deep and well marked as in Ptychodera ; Sp. porosa and alba Willey. Glandiceps Speng., peripharyngeal cavities and synapticula absent, ventral septum of proboscis and external liver saccules and medial gonads as in Sp., pericardial auricles reduced, hypopharyngeal streak of pharynx re- duced to broad tract (Fig. 63) ; G. talaboti and hacksi Marion, G. abyssicola- Speng. Willey ia Punnett, branchial part of pharynx small compared with ventral portion, without dermal pits, synapticula, medial gonads. Fam. 3. Harrimaniidae. Boreal forms with large eggs and direct development, vermiform process of notochord and dorsal roots of medul- lary cord absent ; cornua of proboscis skeleton, efferent vessels of proboscis and perihaemal cavities as in Glandicipitidae ; perihaemal spaces present or absent ; no circular muscles in body-wall of trunk ; synapticula and external liver saccules absent. Harrimania Hitter, proboscis short, proboscis pores paired, peripharyngeal spaces absent, medial gonads present ; H. kupfferi v. Will.-Suhm ; H. maculosa Hitter. Dolicho- glossus j Speng., proboscis long, proboscis pore unpaired, peripharyngeal spaces present, medial gonads absent ; D. kowalevskii A. Ag., D. mereschkoivskii N. Wag., D. sidcatus * Speiigel, with a dorsally grooved proboscis and 10 or 11 pairs of gill-slits; D. ruber Tattersall, from west coast of Ireland. Stereobalanus Speng., proboscis short, two proboscis pores, two pairs of genital pleurae. Fam. 4. Protobalanidae.t The coelom preserves its primitive arrange- ment, is free from mesenchyme and its mesenteries persist. Lateral septa are absent from the trunk, and perihaemal and peripharyngeal cavities from the collar. Gonads in a single row. The other characters as in the Harrimaniidae. Protobalanus Caullery and Mesnil, P. koehleri, 4-6 cm. in length, St. Martin's Bay, Cap de la Hague, north coast of France. * Dolichoglosstis otagoensis Benham (Q.J.M.S., 42, 1899, p. 497), a form recently described from New Zealand, creeps on seaweed by means of a very contractile proboscis, which is grooved dorsally as in D. sulcatus ; it possesses only 12 pairs of gill-slits. t Caullery et Mesnil, Zool. Jahrb. Anat., 20, 1904, p. 227. 104 PHYLUM ENTEROPNETJSTA. Order 2. CEPHALODISCIDA. Enteropneusta in which the collar is prolonged into paired tenta- culiferous arms and the trunk is much shortened in its antero- posterior axis, with at most one pair of branchial apertures. The animals live in colonies and inhabit tubes which are formed by the proboscis. The Cephalodiscida comprise two genera, Cephalodiscus Mc- Intosh and Ehabdopleura Allman. They are colonial animals which possess the power of budding and inhabit tubes secreted by the ectoderm of the flattened proboscis. The two genera while agreeing in most points of structure differ in the presence or absence of branchial apertures. In Cephalodiscus there is one pair of lateral openings leading outwards from the pharynx ; in Rhahdopleura branchial apertures are absent. In the following account the two genera are dealt with separately. Cephalodiscus* resembles the Balanoglossids in the main plan of its organisation, but differs from them in the fact that the trunk region, though exceedingly shortened antero-posteriorly, is much elongated in the dorso-ventral direction, and in the restriction of certain trunk organs, which are repeated in the Balanoglossida, to a single pair, e.g. gill-slits and gonads. More- over, it has the power of reproducing itself by budding ; and several individuals live in association in a single tube system. Cephalodiscus was discovered by the Challenger in 1876, at a depth of 245 fathoms, in the Strait of Magellan. It has since been found in other localities (Japan, and from comparatively shallow water in the Malay Archipelago and in the Antarctic). It was at first thought to be a compound Ascidian. Then it was referred to the Polyzoa, and it was not until 1887 that its structure was satisfactorily elucidated by Harmer and its real nature as an ally of the Enteropneusta demonstrated. A considerable number of individuals, probably all produced by budding from a single original individual, live together in a system of ramifying, sometimes anastomosing tubes which ; \V. ('. .Mdiitosh, Cephalodiscus, ('Imll, ifjer Reports, vol. 20, 1887. S. F. Harmer, Appendix to the preceding. Id., On the Notochord of Cophalodiscus, Zoolog. Anzeiger. 1897. Id., The Pterobraiichia of the Siboga-Expodition, Sihfnja-E.>-/>crlitie, vol. 2(1 bis, 1905. A. T. Masterman, On the Structure of Cephalodiscus, Q.J.M.S., 40, p. 340, and Transactions of Royal Soc. of Edinburgh, 39, ISDN. Id., Q.J.M.S., 46, 1903, p. 715. CEPHALODISCTJS. 105 consist of a flexible, brownish semitransparent material. The tubes are composed of superposed lamellae and are probably secreted by the proboscis of the animal (see below). The organ- isms are not attached to each other or to the wall of the tube in anyway, but appear to have the power of moving freely about inside it. Scattered about here and there on the tubes are large rounded apertures, near which the individuals are often found, and through which they can protrude their tentacular tufts. The tubes are covered with tapering spinous processes of their walls. The cavity of the tube may be continuous or it may be divided up into cham- bers, one for each in- dividual or zooid. With regard to the size of a colony it may be mentioned that in Ceph- alodiscus dodecalophus the network of tubes covered an area of 9 inches by 6 inches. The stems have a dia- meter of from 4 to 10 mm., and the whole colony appears to have been attached to mar- ine objects such as stones, sponges, etc., by vertical stems which descend from the underside of the net- work to the substratum. The natural position of the colonies would thus appear to be horizontal. The full-grown individuals of C. dodecalophus measure about 2 mm. in their longest diameter. Cephalodiscus may be described as an animal in which the or- anal axis of the body is very short and the ventral surface behind the mouth is produced into a large hump in which the alimentary canal is continued (Figs. 77, 78). This ventral hump terminates at its ventral and anterior end in a pedicle on which buds are continually being formed. The mouth is ventral and anterior and is overhung by a large FIG. 77. Cephalodiscus dodecalophus, anterior view (after Mclntosh). 1 tentacles ; 2 proboscis (buccal shield) ; 3 pigment baud on proboscis ; 4 buds ; 5 pedicle ; 6 trunk. 106 PHYLUM ENTEROPNEUSTA. preoral lobe (Figs. 77, 2, 78, 18) called the Duccal disc or proboscis. Behind this there is a region of the body which, except in the young bud, is not very distinctly marked off either from the proboscis in front or from the trunk behind. This is the collar region. 16 17 8 F IG yg. Median longitudinal vertical section through Cephalodiscus dodecalophus (after Harmer, from Lang). The figure is diagrammatic, especially in the fact that certain struc- tures not in the middle line are shown. 2 nervous system ; 3 collar coelom ; 5 postoral lamella (operculum) ; A' trunk coelom ; 9 pharynx ; 10 oesophagus ; 11 stomach ; 12 in- testine 13 buccal cavity ; 14 pedicle ; 15 ovary ; IP, anus ; 17 oviduct ; IS proboscis (buccal shield) ; 19 proboscis coelom ; 20 one of the proboscis pores ; 21 notochordal diver- ticulum of the pharynx. The collar forms no projection round the base of the proboscis in the mid-dorsal region, but its ventral and lateral portions pro- ject forward, forming a kind of lip (Figs. 78, 79). This lip is called the operculum* : it forms the ventral and lateral edge of the * The operculum can be folded backwards, a peculiarity which com- bined with the inronspicuousness of the ventral part of the collar has led some authors to describe it as a projection of the posterior edge of the collar. CEPHALODISCUS. 107 mouth- opening which lies between it and the base of the proboscis stalk. The arms are prolongations of the dorso-lateral parts of the collar commencing at the point where this lip-like projection passes into the dorsal part of the collar. They are tentaculi- ferous and vary in number from four to six pairs (one pair in the males of C. sibogae). In C. dodecalophus they are slightly swollen at their ex- tremities. Each arm contains a pro- longation of the collar body- cavity (Fig. 79) and is grooved on its ventral surface, the grooves uniting in pairs and converging to the corners of the mouth, near which they terminate. The ectoderm of the swellings at the end of the arms of C. dodecalo- phus contains a number of ovoid globules of a clear substance which have been variously interpreted as organs akin to the rhabdites of Turbellaria, and as the refractive structures of a rudimentary visual organ. Similar refringent vesicles are found along the whole course of the arms of the males of C. sibogae. FIG. 79. Transverse section through C'ephalo- discus dodecalophus (after Harmer). 1 central nervous system of the collar region ; 2 oper- culum turned backwards ; 3 pharynx ; 4 oesophagus ; 5 stomach ; 6 intestine ; 7 trunk-coelom ; S gill-slit ; 9 collar-pore ; 10 collar-coelom ; 11 arm. Behind the collar region is the trunk, the ventral portion of which is prolonged into the huge hump before referred to. The single pair of gonads (Fig. 78, J5) are placed in the anterior part of the trunk, just behind the collar, and open dorso-laterally one on each side. The anus is at the hind end of the dorsal surface (Fig. 78, 16). The proboscis presents a broad disc-like surface towards the front (Fig. 77) and is marked by a crescentic band of pigment in its ventral portion (j). Its anterior wall is covered with a glandular ectoderm, which appears to play some part in the formation of the tube, as in Rhabdo pleura. The mouth is on the ventral side at the junction of the proboscis and collar (Fig. 78) 108 PHYLUM ENTEROPNEUSTA. and is usually entirely hidden by the ventral portion of the over- hanging proboscis. The alimentary canal (Fig. 78) consists of mouth-cavity, pharynx, oesophagus, stomach, and intestine which opens by the posterior and dorsal anus. The bulk of it is contained in the great ventral hump so characteristic of this animal. The anterior end of the pharynx gives off a forward diverti- culum which extends through the collar region into the proboscis : this is the notoehord (Fig. 78, 21). There are two gill- slits, one on each side leading outward from the pharynx and opening ventro-laterally on the anterior part of the trunk. The body- cavity is arranged as in other Enter opneusta, i.e. it consists of an unpaired chamber in the proboscis which opens to the exterior by a pair of pores at the junction of the proboscis and collar region of the body (Fig. 78, 20) ; of a pair of chambers in the collar region, which also open outwards by a pair of collar pores (Fig. 79, 9), and of another pair in the trunk which are in rela- tion with the greater part of the alimentary canal and with the ovaries, and which do not open to the exterior. The two halves of the body-cavity of the collar overlap dorsally the hinder part of the proboscis region (Fig. 78, j). The dorsal and ventral mesenteries appear to persist, completely or incompletely, in the collar and trunk regions. The sexes are separate and all the individuals of one colony are of the same sex (except in C. nig- rescens, p. 109). The ovaries are paired sacs placed dorsally in the anterior part of the trunk region. They open to the exterior by two short oviducts, the walls of which are richly pigmented (Fig. 78, ij}. The male colonies appear, in some species (e.g. sibogae), to be dimorphic, containing male individuals which are characterized by possessing two arms without tentacles, a long pedicle and an abortive alimentary canal, and neuter individuals which are like the females but without ovaries. The coelom of the proboscis and collar are to some extent obliterated by muscle and connective tissue, as is the prolongation of the trunk body- cavity in the pedicle. The organs appear to be separated by a basement membrane as in the Balanoglossida and there is no connective tissue of the usual kind except in. the coelomic sacs in the dorsal region of the collar. The central nervous system is contained in the ectoderm. It extends forwards on to the hinder part of the proboscis and the proboscis pores perforate its CEPHALODISCUS. 109 anterior portion (Fig. 78). It is moreover continued as a nervous tract along the dorsal side of each arm. The pedicle is a process from the ventral surface. It is said to be largely filled with longitudinal muscular fibres, but it appears to contain a prolongation of the posterior body-cavity. Buds are formed upon its terminal portion. From one to three buds are found upon almost all full-grown individuals. The endo- dermal tissues of the parent do not extend into the pedicle and play no part in the formation of the bud, the alimentary canal of the latter being entirely formed by an invagination of ecto- derm. Cephalodiscus has, in its proboscis and in close contact with the noto- chord, structures corresponding to the pericardial sac, glomerulus, and central blood-sinus (heart) of the Balanoglossida. The pericardial sac is at the anterior end of the notochord and contains the heart in its interior. Blood vessels have been made out in other parts of the body as spaces the walls of which are probably formed by the walls of the mesodermic cavities, or they may be spaces between the same structures and the ectoderm or.endoderm. Free eggs and embryos are found in the tubes of some species. The ova are of a fair size and contain a considerable quantity of yolk. The cleavage is complete and the embryos leave the colony at an early stage as ciliated planulas. The following species are known. C. dodecalophus M'Intosh, with 12 arms, with end bulbs and vesicles ; cavity of tube continuous ; Straits of Magellan, 448 metres. C. levinseni Harmer, with 12 arms, without end- bulbs or vesicles ; cavities of the tubes divided up into chambers, one for each zooid ; sea between Japan and Corea, 183 metres. C. gracilis Harmer, very small ; with 10 arms, apparently without end-bulbs and vesicles in adult ; E. coast of Borneo, reef. C. sibogae Harmer, male-colony only known, with dimorphic individuals (see p. 108) ; S.E. of Celebes, 75-94 metres. C. nigrescens Lankester,* colony large and massive, nearly transparent, with tubes projecting from its surface, each tube being cut off from the rest and containing one full-grown zooid ; the zooids are deeply pigmented and large (4' 5 mm. X 1 mm.) and have 6 to 8 pairs of arms which are without terminal swellings ; each colony contains male, female, and hermaphrodite individuals, the latter having one ovary and one testis ; Antarctic Ocean, in 100 fms. The genus Ehabdo pleura^ Allman must also, in view of Fowler's * Proc. Roy. Soc., 1905, 76 B, p. 400. t Allman, Q.J.M.S., 9, 1869, p. 57. Sars, Q.J.M.S., 14, 1874, p. 23. Lankester, Q.J.M.S., 24, 1884, p. 622. Fowler, Proc. Roy. Soc., 52, 1893, p. 132, also in Leuckart's Festschrift, Leipzig, 1892, p. 293, and Q.J.M.S., 48, 1904, p. 23. Schepotieff, A. Zur Organisation von Rhabdopleura, Bergens Museums Aarbog., 1904, No. 2, and Zool. Anzeiger, 28, 1905, p. 795. L10 PHYLUM ENTEROPNEUSTA. work, be placed in alliance with the Enteropneusta. It differs, however, from the other members of the phylum by the absence of pharyngeal apertures. It exists in the form of colonies, the zooids of which are con- nected with one another by living substance and are contained in trans- parent tubes of a chitin- like material. The colo- nies have the form of a branching axis which lies upon the substratum and gives off at irregular intervals the terminal branches. These, for the most part, after adhering for a short distance to the substratum, rise up and project freely into the water. At their free ends are found the openings of the tubes by which the zooids come into relation with the exte-rnal world. The zooids consist of a body and a stalk (Fig. 81) The body is small ("12 mm. in diameter) and possesses at its anterior end a preoral lobe the buccal disc or proboscis, a pair of tentaculiferous arms arising laterally at the level of the mouth from the collar region (see below), a mouth on the ventral side of the buccal disc (Fig. 82, 8) and an anus (j) placed at about the same level as the mouth on the dorsal side. The stalk is the narrow ventral end of the body which passes down to join the common axis of the colony in the creeping stolon. Its connexion with the body is somewhat an- Fiu. 80. Portion of a living colony of Rhabdopleura normani (Allman), Lofoiru Islands, x 16. a the terminal branches with the zooids in different states of protrusion : c the proboscis (cephalic disc) ; d the two tentacular arms ; / the stomach ; g the intestine; h the stalks of the zooids ; i the axial rod in the creeping stolon. EHABDOPLEURA. Ill teriorly placed. The tube is secreted by the buccal discs of the zooid. It consists of a series of rings, each of which is secreted separately by the buccal disc and added to its predecessor. It is therefore quite independent of and separate from the bodies and FIG. 81. Rhabdopleura nor- mani from the right side (after Lankester, from Lang). 7 buccal disc (pro- boscis) ; 2 arms with ten- tacles ; 3 region of the collar pore ; 4 anus ; 5 trunk ; 6 stalk or pedicle. stalks of the zooids, and the body of the animal can move freely within its tube. It can be retracted into the tube by the con- traction (into a spiral) of a muscle in the stalk, and it can crawl up the inside of its tube by help of its buccal disc and can protrude its arms from the terminal opening. 112 PHYLUM ENTEROPNEUSTA. The common axis consists of the stalks of the zooids in earlier stages of the growth of the colony. It lies within the tube and the older parts of it possess a cuticle on its ectoderm. This cuticle becomes hardened and of a dark colour. It fuses second- arily with one side of the wall of the tube and forms a con- spicuous object as a dark brown, thin rod easily visible to the naked eye in the creeping stolon of the colonies. Some of the terminal branches of the colony do not rise up, but the tubes with their contained axis continue along the substratum and end in open mouths. These are actively growing proliferous branches. The axis in them terminates in an imperfect zooid, the proboscis of which adds to the tube, while its stalk has lateral wart-like buds. Transverse septa are formed across the tube between the buds, and the buds increasing in size burst through the wall of the tube and grow outwards from the creeping axis as imperfect zooids. Most of the new zooids so formed rise up and develop into the perfect form ; a few no doubt adhere to the substratum and form new proliferous branches. It follows from what has been said that the creeping (stolonic) part of the tube is divided up into chambers by septa which do not however interrupt the con- tinuity of the living substance of the stalk, and from each of which one zooid-bearing branch arises. There are no branchial apertures, but there is on each side a groove, between the inser- tion of the proboscis and that of the arms, which leads through the mouth into the oesophagus, on the side-wall of which it can be traced. This groove is called the branchial groove. The anatomy of the animal may be understood at a glance by inspection of Fig. 82. The alimentary canal is bent on itself. The buccal cavity is provided with an anteriorly directed diverti- culum which is continuous with a rod-like structure, apparently half-cellular and half-gelatinoid (Fig. 82, 9). This, which is clearly comparable to the notochordal diverticulum of other Enteropneusta, projects into the base of the preoral disc (pro- boscis). The body cavity, which appears to have a cellular lining, is divided into five chambers arranged as in other Entero- pneusta ; viz. an unpaired chamber (10) in the preoral disc (pro- boscis), a pair of chambers separated from one another by dorsal and ventral median septa and placed in what may be called the collar region of the body (Fig. 82, 2), and finally a pair of chambers in the trunk in relation with the greater part of the alimentary RHABDOPLEURA. 113 canal (./). The collar-cavities are continued into the arms, which arise from the collar region, and the trunk cavities appear to be continued into the stalk. At any rate the stalk contains in addi- tion to its muscular tissue two cavities separated by a septum.* The proboscis-cavity possesses a pair of pores, and the collar- cavities communicate with the exterior, each by a collar canal. On the dorsal side of the collar region behind the collar coelom (Fig. 82), there is an ectodermal thickening which seems to corre- spond with the nerve plate of Cephalodiscus and the medullary nerve cord of the Balanoglossida. Mesoblastic skeletal tissue, ap- parently similar in character and relations to that of the Balano- glossida, occurs in the arms and tentacles, and in the axis of the stalk. Most of the specimens hitherto found have been devoid of generative organs, but in a few there was a testis in the form of an elongated sac lying parallel to the intestine on the right side, and forming a projection on the surface of the body ; it opened close behind the anus. A vascular system consisting of a dorsal and ventral vessel has been described, and there appears to be a pericardial vesicle and heart at the anterior end of the notochord as in other Enteropneusta. Rhabdo pleura^ is a marine animal and is found in compara- c Fowler thinks that the stalk contains a cord of endoderm which is continuous with that of the alimentary canal. If this is so, the endoderm may participate in the budding. t It has been suggested that the section of the Graptolites known as Monograptidae are nearly related to Rhabdopleura (see Alhnan, On the Morphology and Affinities of Graptolites, Ann. and Mag. Nat. Hist., 1872. and Schepotieff, Neues Jahrbuch /. Mineralogie, 2, 1905, pp. 79-98). Z III I FIG. 82. Rhabdopleura normani in longi- tudinal vertical section (after Fowler from Lang). 1 arm of one side, indi- cated by dotted lines ; 2 anterior paired (collar) coelom ; 3 anus ; 4 posterior paired (trunk) coelom ; 5, 6 alimentary canal ; 7 buccal cavity ; 8 mouth ; 9 anterior diverticulum of buccal cavity(notochord); 10 anterior impaired (proboscis) coelom ; 11 buccal shield. 114 PHYLUM ENTEROPNEUSTA. tively deep water (40 to 120 fms.). It is widely spread, having been recorded from the Norwegian Fiords, off the Lofoten and Shetland Islands, off Tristan d'Acunha, South Australia the Malay Archipelago, the Azores, Ireland and Brittany. The question of the relationship of Phoronis to the Enteropneusta was dealt with on p. 547 of vol. i. of this work. Since that passage was written several memoirs * dealing with the development and structure of the Actinotrocha larva have been published. The upshot of these is to justify the criticism there set forth. To take the points seriatim, the exist- ence of a " neuropore " and " subneural gland " has not been confirmed : on the contrary it has been denied by Ikeda and Goodrich. The relations of the stomach caecum of Actinotrocha have been shown to be not those of a notochord. No evidence is forthcoming to contradict Caldwell's original statement that the cavity in the preoral lobe of Actinotrocha is haemocoelic (so-called primary body cavity of the trochosphere), and not a coelomic sac. It seems fairly clear that nothing comparable to proboscis pores or to collar pores are present. The only new point that has come out in this connexion is that the preseptal body-cavity of the adult is present in the old larva as a horseshoe -shaped cavity which underlies the ring of tentacles and send extensions into them, but the development of this cavity has not been ascertained. Finally, it has been shown that Cald- well was right in his account of the larval excretory organ as a ciliated canal not opening into the body-cavity, but terminating internally in some peculiar cells which recall the so-called flame cells of Platyhelminthes and Annelid larvae, etc. (solenocytes of Goodrich, see vol. 2., pp. 27, 28). * Masterman, Q.J.M.S., 43, 1900, p. 375. Ikeda, Journ. Coll. Sci. Imp. Univ. Japan, 13, 1900-1, p. 507. Goodrich, Q.J.M.S., 47, 1903, ]>. 103. Schultz, Z.f. w. Z., 75, 1903, pp. 391, 473. M. de Selys Long- champs, Mem. Classe Sci. Acad. Belgique, i. 1904. CHAPTER III. PHYLUM ECHINODERMATA* With a radial, usually pentamerous arrangement. The body- uraU contains calcareous plates and generally bears spines. The coelom is divided into two well-marked portions the perivisceral cavity and the watervascular system, and the gonads are noi con- nected with it in the adult. The Echinodermata, like the Coelenterata, present a radial arrangement of their principal organs and were for that reason united by Cuvier with the polyps and medusae in the group Radiata. With the progress of anatomical knowledge, it soon became apparent that the organization of the Echinoderms differs totally from that of the Coelenterates and belongs to a much more complex grade of development. This fact, which was first recognized by R. Leuckart, led to the separation of the two groups and to the establishment of two independent phyla, the Coelenterata and Echinodermata. A trace of the old view * Fr. Tiedemann, Anatomic der Rohrenholothurie, des pommeranzfar- benen Seesternes u. des Steins eeig els, Heidelberg, 1820. De Blainville, Manuel tf Actinologie, Paris, 1834. L. Agassiz, Monogr. d' Echinoderm.es vivants et fossiles, Neuchatel, 1 838-42. L. Agassiz et E. Desor, " Catalogue raisonne des families, des genres et des especes de la classe des Echinodermes" Ann. Sci. nat. (3), 6, 7, a,nd 8, 1846-7. E. Forbes, A History of British Starfishes, London, 1841. R. Leuckart, Ueb. d. Morphologic u. Verwandt- jchaftverhaltnisse d. wirbellosenthiere, Braunschweig, 1848. J. Miiller, "Ueb. d. Bau d. Echinodermen," Abh. d. Berlin Akad., 1853. Id., " Sieben Abh. iib. die Larven u.d. Entw. d. Echinodermen," Abh. d. Berlin Akad.. 1846, 1848, 1849, 1850, 1851, 1852. A. Agassiz, " On the Embryology 01 Echinoderms." Memoirs of the American Academy, 1864. E. Metschnikoff, *' Studien ub. die Entwickelungsgesch. der Echinodermen u. Nemertinen," St. Petersburg, 1869. H. Ludwig, " Morphologische Studien an Echino- dermen," Z. /. w. Z., 1876-82. Id., " Echinodermen," in " Bronns Klassen u. Ordnungen," in progress. O. Hamann, " Beitrage z. Histologie d. Echinodermen," Jena. Zeitschrift., 1884-89. L. Cuenot, "Etudes morphologiques sur les Echinodermes," Arch, de Biologic, 11. 1891. ^ H. E. Durham, " On Wandering Cells in Echinoderms, etc.," Q.J.M.S., 33, 1892. P. H. Carpenter, several memoirs in the Q.J.M.S. and other Journals, 1875-1890. E. Ray Lankester, A Treatise on Zoology, Pt. 3, The Echino- derma, London, 1900. E. W. MacBride, Echinodermata. in vol. I of the Cambridge Natural History, 1906. 115 116 PHYLUM ECHINODERMATA. has until quite recently survived in the juxtaposition of the two phyla which used frequently to be found in works on Zoology and Comparative Anatomy. But of late years it has been recognized more and more clearly that this juxtaposition is not warranted by the facts and that the affinities of the Echino- dermata, in so far as any can be traced, are rather with the higher phyla of the Metazoa, than with the lower. Expression was given to this view in 1877 by Huxley in his Anatomy of Inrcrtcbrata and in 1880 by F. M. Balfour, who in his Comparative Eni'ift/ology placed the Echinodermata at the end of the volume dealing with the invertebrate groups, in the neighbourhood of the Enteropneusta and Chordata. This example was followed in 1890 by A. Lang in his textbook of Comparative Anatomy, and now we, in the light of the most recent work on the subject, have thought it right to take the same course. The Echinodermata are radiately symmetrical animals (see p. 117) in which the number of radii is nearly always five or some multiple of five. This symmetry is, however, characteristic of the adult only, for in the youngest state all members of the group are bilaterally symmetrical and in nearly all there is a free-swimming bilateral larva. The view usually taken and adopted by us when referring to the matter in the chapter on Mollusca in the first volume of this work (p. 317) is that the bilateral symmetry is the primitive symmetry possessed by some adult ancestor and that the radial arrangement is to be regarded as a distortion from the original condition. Now, however, after a more com- plete study of the group, we see reason to suspend our judgment on this matter, and though we should hesitate to adopt the view that the Echinodermata have been derived from asym- metrical or from radiately symmetrical forms, and that the larva has been especially produced and modified for a free- s\\ i mining life, we are of opinion that there is at least as much to be said for it as for the older and more usually adopted view that the ancestral form was a form in which bilateral symmetry had been completely evolved.* The radial symmetry is expressed not only in the external appearance, but also in the arrangement of most of the internal organs. It is, however, never completely carried out, and in ' For a discussion of this question, see the section on " Affinities,"' p. 160. RADIAL STRUCTURE. 117 some forms (certain Echinoids) there is a tendency to bilateral symmetry. Speaking generally the body may be described as spherical or discoidal in form, with the mouth in the centre of the lower * surface and the anus at or near the centre of the upper surface. An oral and aboral surface or pole may thus be distinguished. The anus varies more in position than does the mouth. The mouth is nearly always in the centre of the oral surface or at the oral pole ; in a few Holothurians, some Echinoids and in Actinometra among Crinoids it is slightly shifted from this position. The anus on the other hand is central only in Holothurians, and even in some of these it is slightly displaced ; in Asteroids and regular Echinoids it is very near the centre of the aboral surface, but always slightly excentric ; in irregular Echinoids it is at some distance from the central point and sometimes on the oral surface ; and in Crinoids it is always on the oral surface. In a few Asteroids (Astropectinidae, etc.) and all Ophiuroids the anus is absent in the adult. The radial structure is indicated externally by the rows of tube- feet (p. 129) which extend outwards from the mouth towards the aboral pole. The surface of the body is thus marked into radii along which the tube-feet are arranged, and into interradii the portions between the tube-feet rows. In Asteroids, Ophiu- roids, and Crinoids the radial portions of the disc are prolonged into processes, which constitute the arms ; in Holothurians and Echinoids the radii are not so prolonged and there are no arms. The rows of tube-feet never extend quite to the aboral pole ; it is therefore possible to distinguish that portion of the surface of the body from which tube-feet project, as ambulacra! or actinal, from the antambulacral or abambulacral or abactinal surface which is without tube-feet. In the brachiate Echiiioderms, that is in Asteroidea, Ophiuroidea and Crinoidea, these two regions or surfaces are about equal in extent, the whole of the oral surface being ambulacral, and the whole of the aboral surface antambulacral. In such cases it is customary to call the lower or ambulacral surface ventral, and the upper or antam- bulacral surface dorsal. But it is better not to use the terms dorsal or ventral in adult Echinoderm morphology without prefixing the word adult, because dorsal and ventral are used * In Crinoids and their allies the oral surface is turned upwards in the natural position of the animal. 118 PHYLUM ECHINODERMATA. in an entirely different sense in describing the larvae and in adult Holothuriaiis, and cannot be used satisfactorily in Echin- oids at all. In Holotlmrians and Echinoids the tube-feet extend almost to the aboral pole, and the antambulacral surface of the body is restricted to the very small region round the anus. Moreover in Holotlmrians, in which the body is cylindrical and lies with its whole length applied to the substratum, it is usual to call one surface of the cylinder, viz. that on which the single genital opening and madreporite are placed, the dorsal, and the other side the ventral. In many Holothurians the animal always- lies with the three radii (trivium) of the so-called ventral surface directed towards the substratum and uses the tube-feet of these radii for adhesion, the tube-feet of the two dorsal radii (bivium} being without suckers and probably used for respiration and sensation only. In pentameral Echinoderms the words bivium and trivium are frequently used to designate the two groups into which the five radii may be divided. These words are used however in a somewhat vague sense and it must not be supposed that the arms of the bivium and trivium of one class are necessarily the same as the bivial and trivial arms of another class. The water-pore or madreporite is generally single and always interradial in position. It is generally abactinal, but in Crinoids and Ophiuroids it is on the ambulacral surface. In Holothurians the madreporitic interradius occupies the middle of the so-called dorsal surface. In Ophiuroids as in Asteroids the water-pore is generally single. but in Crinoids it is always multiple, there being one (Rhizo- crunt*} or more than one in each interradius. The number of water-pores (madreporites *) is however subject to variation in all classes of Echinoderms except Echinoids, in which normally there is never more than one. In Echinoids, Holothuroids. Neocrinoids and Blastoids the number of radii is constantly live except in abnormal individuals. t In Asteroids. Ophiuroids, Crinoids and Cystids this number may romplete cxpla.nat inn df the terms water-pore and madreporite sen i>fl(>\v. ]>. 1:>7. It may be mentioned here that the numerous perfora- tions in the madreporitic plate of Echinoids and Asteroids represent one water-pore only. '<- See Bateson, Mutcn'rifn far thr Study of Variation, London, 1894. p. 43.J et seq. In Echinoids 4- and 6-rayed abnormalities are not uncommon, and in Holothurians Ludwijj; found half a dozen 6-rayed individuals in ].">(> specimens of Cin-mnaria plane/. NUMBER AND ENUMERATION OF RADII. 119 be departed from.* In most cases the number of radii is deter- mined early in development, but in a few forms (e.g. Labidiaster) the number increases with the growth of the animal, at any rate in the early stages of the adult (Perrier). In Cystids the number of radii varies more than in any class, the number two or three sometimes occurring. Enumeration of Radii. After some hesitation we have decided to adopt a definite enumeration of the radii for all classes of Echinoderms. Our enumeration, which is shown in Fig. 83, is based on the assumption that the position of the stone-canal and primary water-pore is the same in all classes. If that assumption is incorrect, which it may very well be, j" the homolo- gies which might be deduced from it fall to the ground. In any case we desire to warn the reader against attaching too much importance to the determination of homologies based solely on this assumption. We have adopted the enumeration because it conduces to clearness and enables us to impart a greater precision to our descriptions and not because we think that it is of any importance from the point of view of determining homologies. The interradius in which the hydrocoel closes in the larva presents remarkable variations, in relation to the position of the madreporite, as the diagram indicates, and it is probable that later researches will show an even greater variability in this respect, the number of species hitherto examined from this point of view being small. Our enumeration is identical with that of Loven excepting in its applica- tion to Crinoids (see p. 272). The first-formed water-pore of Criiioids and the anus are in the same interradius. As our enumeration is based on the position of the water-pore, we assign the anus of Crinoids to inter- radius II. Ill, and not to interradius I.V, as is done by Loven. One of the most important characteristics of the Echino- dermata is the presence of calcareous plates in the dermis. These may have various forms, from the isolated plates and * In Qphiuroids and Crinoids, though the number of radii is usually five, the arms in some cases branch. f It is quite likely, as suggested by MacBride, that the interradius of closure is the fixed point, the position of the madreporite being variable. He points out to me that the stone-canal opens on the inner side of the hydrocoel ring, and that the position of the opening might easily shift. The madreporite being on the dorsal surface might also easily change its interradius, as the anus undoubtedly does. 120 PHYLUM ECHINODERMATA. spicules of Holothurians, which are remarkable for their definite and symmetrical form, to the compact and complete dermal skeleton of Echinoids. The plates arise as deposits in the form Trivium r in CLOSURE IN OPHIURIDS AND HOLOTHURIANS ? CLOSURE IN ANTEDON IV CLOSURE IN ECHINIDS AND HOLOTHURIANS ? POSTERIOR INTERRADIUS CLOSURE IN ASTERINA ANUS OF EXOCYCLICA Bivium FlO. 83. Diagram uf an Kchinoderm viewed from the aboral pole, to show the enumeration of the radii adopted in this work. The diagram also shous the interradius in which the liydrocoel closes in certain observed eases (vide Bury. (J..I.M.S.. '2'.1. p. 431). I-V the radii ; M the inadrejiorite. In MacBride's enumeration of the liydrocoel lolies adopted hen 1 in the succeeding account of the development of Asterhui, lulie no. 1 is radius II. lolie no. ~ is radius III, ami lolic no. f> is radius I. In Cue-not 's enumeration the equivalence is A = V. B = I, C = II, D = III. and E = IV. Radius I is the right anterior, and radius IV is the left anterior radius of Crinoids. The plane of symmetry of Holothurians and Crinoids passes through radius V and interradius II. Ill, that of Asteroids through IV and I. II. and that of exocyclic Echinoids through 111 and V.I (Loven's plane of symmetry). SKELETAL SYSTEM. 121 of networks or spongeworks of calcareous matter in the con- nective tissue. In many of the classes (Asteroids, Ophiuroids and Echinoids) these plates carry spines and processes which project on the surface, and in the young state at least are covered by the epidermis. The majority of the spines are movably articulated with the plates, and in Asteroids and Echinoids some of them are specially modified as snapping organs the pedicellariae. The function of the spines is probably mainly protective, but in the case of the long spines of Echinoids it is locomotory. The pedicellariae are protective, seizing foreign organisms (p. 224). They are said not to bite animals of the same species (autodermophily, Uexkiill). The epidermis is generally ciliated, but in Ophiuroids, and on the aboral surface of Crinoids it is difficult to distinguish it as a layer distinct from the cutis. The description of the skeletal system will be best dealt with in connexion with the different classes of the phylum. It will be well however to call attention here to certain plates which are supposed to be homologous throughout the group : these are the plates of the oral and apical systems. The oral plates are five in number and are placed interradially round the mouth on the oral surface. They can be distinguished in many Crinoids and in Ophiuroids, but are not clearly distinguishable in any other class. The plates of the apical system are placed at the aboral pole, and when present are usually discernible at a comparatively early stage of development. In typical cases (Ophiuroids, Fig. 84, many Crinoids, Fig. 85) they consist of a central plate surrounded by ten plates, five of which are radial in position and five interradial : the former are called infrabasals, the latter basals. Beyond this circle there are five radially placed plates called the radials. Of these plates the infrabasals, or under- basals as they are sometimes called, are frequently absent. In Echinoids the central plate is not pierced by the anus which is placed on one side of it ; moreover it is frequently very difficult to distinguish it, owing to the presence of a number of small plates, called the periproct plates, at the apical pole. In Crinoids the position of the central plate is not certainly known, but it is supposed to be represented by the so-called dorso-central which is found at the peripheral end of the stalk. In Asteroids a complete 122 PHYLUM ECHINODERMATA. apical system can sometimes be made out, but as a rule the plates are indistinguishable from the other plates of the antambulacral surface. In Holothurians the plates of the oral and apical systems are absent both in the larva and in the adult. In Echinoids, in which a central plate can often be made out, the five interradial basals (genital plates) are always present, but infrabasals and radials are not found. The oculars of Echinoids. which are often called radials, are rather comparable to the terminals, which must now be referred to. Whereas the plates of the apical system, viz., the central, the infrabasals, the \n/-Mifi3Ite( ce ism^Fl Flo. 84. Apical system of a young Ophiurid Ainphiura squamata (after P. H. Carpenter), ba basal : ce cen- tral : ib int'rabasal ; r radial ; t ter- minal. FIG. 85. Apical system of a Crinoid (Cyatho- crini'x). IKI basal ; c2 second costal (primi- brach) : ib infrabasal ; ian anal interradial ; r radial (from Lang). basals and the radials are developed round the right coelomic vesicle of the larva, the terminals are five radially placed plates on the oral wall of the left coelomic sac. They become the oculars of Echinoids and those plates of Asteroids and Ophiu- roids which are placed at the ends of the arms on the aboral side of the projecting end of the radial water- vascular canal. The question as to whether these plates are in all eases homologous is a very difficult one to answer. It was originally suggested by Loven and has been maintained by many of the later students of the group- notably by P. H. Carpenter that there is a general homology between these plates in the different classes ; but recently some doubt has \>i-> -ii thrown upon, this view of them (see p. 292). It is possible NERVOUS SYSTEM. 123 that the very remarkable similarity in their arrangement may be clue to the similarity in the general structural conditions, viz. the presence of a central point round which the plates are regularly arranged. This view derives some support from a consideration of the fact that when the matter is minutely examined there is a considerable amount of variation in the arrangement of the apical plates. For in- stance the apical pole may be occupied by a central plate (many Echin- oids, larva of Antedon, many Asteroids and Ophiuroids), or by five interradially placed plates (basals in many irregular Echinoids), or by five radially placed plates (infrabasals of some Crinoids). Again the number of circles of these plates varies considerably. In many Crinoids there are three, viz. infrabasals, basals, radials. The same arrangement occurs in many Asteroids and Ophiuroids, but in these classes there is the greatest variation in the arrangement of the plates near the apical pole, and in many of them the plates in this region are small and numerous and the typical apical plates cannot be recognized. In Echinoids, on the other hand, there is never more than one circle, the interradially placed basals. The alimentary canal. The variations in the position of the mouth and anus have already been indicated (p. 117). The alimentary canal passes between the two, and is chiefly remarkable for the very general absence of separate glandular appendages. The anus is absent in the adult of Ophiuroids and of a few Asteroids. For details the reader is referred to the account of the different classes. The central nervous system consists of three parts which are variously developed in the different classes. (1) The ventral, (2) the deep oral, (3) the apical. The ventral system consists of a concentration of a diffuse subepithelial plexus, found in most parts of the body in both ectoderm and endoderm. It is mainly a sensory system and supplies the skin, the feet and the gut. The ectodermal part of this plexus, which is continuous with the endodermal and may be called the ectoneural, is especially concentrated in an annular tract round the mouth (circumoral nerve ring) and in prolongations of this along the whole length of the radii (radial or ambulacral nerves). These concentrations constitute what we have called the ambulacral central nervous system. The general disposition of the ectoneural plexus is well shown in Fig. 86 : it extends into the tube-feet, in some of which, especially in the pointed variety, it is well developed, as well as into the spines, papulae and pedicellariae. The endodermal part of this plexus, which is called the endoneural and presents in Asteroids special concentrations round the edge of the mouth 124 PHYLUM ECHINODERMATA. opening (perioesophageal nerve ring), is found throughout the endoderm and is apparently continuous at the mouth and possibly at the anus with the ectoneural plexus. In Echinoids, Holothurians and Ophiuroids, the circumoral ring and radial nerve trunks are separate from the superficial ectoderm and placed in the wall of a canal, the epineural canal (Figs. 140, 169, 180), which is actually formed in development by the invagination of the larval ectoderm along the centre of each radius* (p. 150). So that in these forms the ectodermal part of the central nervous system attains its internal position by invagination as it does in Enteropneusta and Verte- Flo. 86. Scheme of the nervous system of the arm of a starfish (after Cuenotl. a wall; b body-cavity of arm ; c ampulla of tube foot ; rftube foot ; e radial canal of water- vascular system ; 1 radial portion of ectoneural central nervous system ; 2 ectoneural plexus of tube-foot ; 3 ectoneural plexus of skin ; J Lange's nerve cords (deep oral) ; 5 mesoneural plexus just beneath the longitudinal muscle. brata, and the epineural canal may be compared to the central canal of the nervous system of those animals. Special nerve trunks pass from the central parts of this system to the skin, tube-feet, etc. At the end of the radii these radial trunks pass to the surface and lie in the ectoderm covering the terminal tentacle, if such is present. In the same three classes the apical nervous system (see below) is not present, or at any rate not developed in the same marked manner that it is in Crinoids and Asteroids. The deep oral nervous system consists of a double cord in each * This has been shown for Echinoids by Mac-Bride, for Ophiuroids by Grave, and for Holothurians by Clark. SENSE ORGANS. 125 radius just within the radial nerve thickening of the ectoneural system, being only separated from the latter by a thin layer of connective tissue (Fig. 86, 4). These are called Lange's cords after their discoverer. They lie in the outer wall of the peri- haemal canal and are mesodermal in origin. They are said to be exclusively motor in function. Round the mouth there ap- pears to be a more or less complete ring belonging to this system in Asteroids and Ophiuroids, but this is absent or much less developed in Echinoids and Holothurians. The deep oral system is not present in Crinoids, or, if it is present, it lies deep on each side of the water-vascular canal (Fig. 197). The apical nervous system is also mesodermal and motor. It has the form of a cord in the dorsal middle line and is developed from the dorsal peritoneum with which it sometimes remains continuous (Asteroids, Fig. 86). It is best developed in Crinoids (p. 283), where it is separate from the peritoneum ; it is not found in Holothurians. The sense organs, so far as they are understood, are mainly of a tactile nature. There are the tube-feet, which are all highly sensitive, and in some cases (Crinoids, Ophiuroids, certain tube- feet of Echinoids and Holothurians) exclusively concerned with sensation (and respiration). There is the unpaired tentacle, formed by the projecting end of the radial water-vascular trunk, at the end of the radii ; in the early larvae this is the only tentacular or tube-foot structure present. Lastly there are the circumoral tentacles of Holothurians, which contain prolonga- tions of the water-vascular system and are used for grasping, and the buccal tube-feet of regular Echinoids. In some Crinoids there are special tube-feet near the mouth which may be regarded as circumoral tentacles. The circumoral tufts of Echinoids are respiratory in function and not connected with the water- vascular system. Of other tactile structures we must mention the spines which are richly provided with nervous tissue. Special organs of sense are not numerous. The pigment spots at the end of the arms of Asteroids are probably visual in function (p. 175), as are also the shining spots on the skin of Diadema (anEchinoid, p. 230). The ocular plates of Echini have no visual structures on them and are merely perforated by the terminal tentacle of the water- vascular canal. 126 PHYLUM ECHINODERMATA. Otocysts are found in Holothurians (Synaptidae and Elasi- poda), in connexion with the radial nerves (p. 256). Finally the sphaeridia of Echinoids (p. 227) are probably organs of some special sense. They are modified spines, and are richly provided with nervous tissue. It has been suggested that they are organs of orientation, enabling their possessor to perceive its position in space. The muscular system is very variously developed in the different classes. In Holothurians. only, is there a well-marked derrno- muscular body-wall. In all other Echinoderms the skin is not contractile or but slightly contractile, and the muscles are restricted to special bands acting on particular skeletal plates or other part of the body, and to the muscular elements in the walls of the tube-feet and water-vascular system generally and of the viscera. In the spiniferous forms there are special muscles attached to the base of the spines and pedicel] ariae. The muscular tissue consists of smooth contractile fibres with- out transverse striation, except in the case of the muscles of some of the pedicellariae and of a few spines. The Coelom. Our knowledge of the coelom of Echinodermata has been of slow growth. It began witli the discovery of A. Agassiz * that the enteron of the larva of Asterias berylinns gave rise not only to the alimentary canal of the adult, but also to the body-cavity and water-vascular system. This discovery was established and extended by the work of Metsclmikoff on Asteroids. Echinoids, and Ophiuroids, of Kowalevsky and Selenka on Holothurians. of Bury on Crinoids. and of Ludwig on Asteroids. Finally the recent work of MacBride has not only elucidated the relations of the genital organs and of the so-called perihaemal spaces, but has thrown light upon the nature of that organ which was formerly called the '"heart" but is now more commonly referred to as the axial organ or ovoid gland. The coelom arises from the enteron of the embryo as a single pouch. This pouch soon separates from the enteron and divides into a number of sacs, one or one pair of which constitutes the hydroeoel, the others the splanchnocoel. The left hydrocoelf gives rise to the water- vascular system, while the splanchnocoel * " Embryology of the Starfish," Contributions to the Natural History of the United States, 5, 1864. f The right hydroeoel, when present, is always very small, and takes no part in forming the water-vascular system. COELOM. 127 becomes transformed into the perivisceral cavity and its associated spaces (axial sinus, aboral sinus, perihaemal canals) of the adult. All the coelomic spaces are lined by an epithelium, which in the case of the water-vascular system and perivisceral cavity is ciliated, and they all contain an albuminous fluid in which amoeboid cells float. Nothing of the nature of nephridia is known in the group, but one of the divisions of the splanchnocoelic part of the coelom, viz. that known in the larva as the anterior coelom, has an open- ing to the exterior. This opening is called the primary water pore or the madreporitic pore. The primary water-pore, which is frequently in special relation with a dermal plate called the madreporite, and is always interradial in position, may be a small simple opening or it may be subdivided (Asteroids, Echinoids, many Ophiuroids) into a large number of minute secondary pores, which are more or less closely aggregated together on the madre- porite. Though the generative organs are separate from the coelom in the adult, the primitive germ cells which give rise to them have a coelomic origin. We may now proceed to consider the different parts of the coelom in greater detail, and first of all we will treat of the spaces derived from its splanchnocoelic division. These are three in number : the perivisceral cavity, the axial sinus and the perihaemal spaces. 1. The perivisceral cavity or body cavity proper is always well developed and in relation with the alimentary canal and principal viscera. It is developed from the right and left posterior coeloms of the larva and it never communicates with the exterior except in Crinoids in which the water-pores open into it (see p. 287). It is often traversed by complete or incomplete mesenteries (Holothurians) or by strands of connective tissue, which pass from the body wall to the wall of the alimentary canal. In the s / brachiate forms the arms always contain prolongations of the perivisceral cavity. As stated above, it usually has a ciliated lining and contains an albuminous corpusculated fluid. The Amoebocytes of the coelomic fluids and possibly of other organs play an important part, as was first shown by Durham (op. cit.), in remov- ing foreign bodies from the organism. They act as phagocytes and pass to the exterior by diapedesis through the walls of the papulae on the outside of which they disintegrate. PHYLUM ECHIXODERMATA. Several kinds of amoebocytes,* differing in size, character of granula- tions and shape, are found in the coelomic fluids and tissues of the body. In some cases spherical cells provided with a long cilium have been ob- served (e.g. coelomic fluid of Dorocidaris). In Echinus the amoeboid cells have been observed to unite and give rise to plasmodia and networks, f In the larva they play an active part in the absorption of the calcareous skeleton, and in both larva and adult isolated amoeboid cells take part in tissue formation. 2. The axial sinus is distinct from the perivisceral cavity in all classes except in the adults of Crinoids and some Holothur- ians (see pp. 129, 152). It is developed from the anterior coelom of the larva and communicates with the exterior by the water- pore or pores, and the stone-canal opens into it. It will be convenient to reserve what we have to say about the two latter structures until we have dealt with the water-vascular system. In Asteroids the axial sinus (i.e. the space into which the stone- canal and axial organ project) is continued ventrally round the mouth and forms the so-called inner circumoral perihaemal space (Fig. 131). This space is not connected with and must not be confused with the perihaemal space about to be described. 3. The perihaemal spaces (sinus system) usually consist of an annular circumoral space which is called the outer in contra- distinction to that just mentioned, and of five tubes placed along the radii between the radial nerve cord and the radial water-vascular trunk (Fig. 131). This system of spaces appears to be unrepresented in Crinoids. It is lined by an epithelium and has been definitely traced in Asteroids to outgrowths of the posterior and anterior body-cavities of the larva (p. 145). The perihaemal system must not be confused with the epineural canal of Ophiuroids, Echinoids, and Holothurians (p. 124), which lies superficial to the radial nerve cord and has an epithelial lining which must be regarded as ectodermal. The aboral circular sinus (Fig. 131) might be included amongst the perihaemal spaces ; it is a development of the left posterior coelom and is in relation with the generative rachis (see p. 146). The water-vascular system is derived from the left hydrocoel (p. 144). It is very similarly developed in all classes of Echino- clerms. It is the organ which pre-eminently displays the radial structure of the body, and is the first to show it in the larva. It consists of a circular vessel round the mouth with as many * Cuenot, Arch. Zool. K.rp. (2), 1), 1891, p. (i!3. f Theel, Festskrift Lilljeborg Upsala, 1896, 3, p. 47 WATER- VASCULAR SYSTEM. 129 tubular prolongations as there are radii. These are the radial water- vascular vessels ; they give off lateral branches to the tube-feet all along their course (Fig. 133). The tube-feet are hollow, cylindrical or conical processes of the body wall, and the space within them is continuous with the water- vascular system by the just mentioned lateral branches of the radial vessels. At their inner ends they are connected with small vesicles the ampullae (Fig. 129, 26). Ampullae are absent in Crinoids and Ophiuroids ; they have muscular walls and their function is to drive the fluid into the tube-foot and so to cause its extension. The retraction of the foot is brought about by the contraction of the muscles in its wall, the fluid in the tube- foot passing into the ampulla.* In the forms with ampullae- the tube-feet are locomotive and adhesive organs. Their free ends terminate in sucker-like discs which adhere to foreign bodies. When ampullae are absent (Crinoids and Ophiuroids} the tube-feet are purely sensory and respiratory in function and cannot be used for adhesion and locomotion. Calcareous; bodies are often present in the connective tissue layer of the tube-feet, particularly at the sucker-like termination. The circumoral vessel frequently possesses accessory structures opening into it in its interradial portions ; these are the bladder - like polian vesicles (p. 184) and the gland-like Tiedemann's bodies (p. 185). The water-vascular system is indirectly connected with the exterior by a canal called the stone- or sand- canal (Figs. 133, 136). The stone-canal owes its name to the fact that it frequently contains in its walls a large amount of calcareous matter, which readily breaks through the lining and falls into the canal, where it is found as gritty matter. The connexion with the exterior is effected in the following way (Fig. 132) : the stone-canal passes off from the circumoral vessel in one of its interradii to open, in the larva, into the anterior coelom which is a portion of the splanchnocoel. The anterior coelom, which, as already explained (p. 127), opens to the exterior by the water-pore, persists into the adult in Asteroids, Ophiuroids, and Echinoids as a distinct space, the axial sinus, f In Crinoids it is a distinct space in the * For the valve assisting in this process, see p. 182. f In the adult many of the pores of the madreporite come to open directlr into the dorsal end of the stone-canal, but the opening of the stone-canal into the axial sinus is always maintained. Z III K 130 PHYLUM ECHINODERMATA. young larva, but later becomes merged in the general perivisceral cavity, so that in the adult both the stone-canals and the water pores open into the perivisceral cavity. In Holothurians the anterior coelom is very feebly developed, but as Bury has shown it can be traced through the whole development (p. 152) and has the appearance of a small appendage of the stone-canal (Fig. 108). In a few Holothurians it retains its communication (by the water-pore) with the exterior, but in the majority the water-pore closes in the adult, and the walls of the anterior body-cavity largely break down and give rise to the so-called internal madreporite. It thus comes about that in most Holo- thurians the stone-canal appears to open directly into the general body cavity. It sometimes happens (some Asteroids and Ophiuroids, Crinoids) that there is more than one primary water-pore ; in such cases there is a cor- responding increase in the number of madreporites and stone-canals, and they are generally placed in different interradii. In Crinoids there are never less than five primary water-pores, and generally they are much more numerous. In the former case there is one water-pore and one stone-canal in each interradixis ; in the latter case there are many of both structures in all the interradii, and an exact correspondence between the stone-canals and water-pores, though it may exist, cannot be shown. The water- vascular system is lined by a flat ciliated epithelium and contains an albuminous fluid with leucocytes very similar to that found in the perivisceral cavity and sinus system. Its function is mainly locomotory, but to this must be added, especially in forms in which the tube-feet are without ampullae, a tactile and respiratory function. By some authors it has been regarded as in part excretory, but there seems to be little evidence in support of this. There might be something to be said for this view, if the current through the madreporite set outward, but the reverse appears to be the case (Ludwig). If there is any current through the water-pore, it appears to be an inward one as a result of which sea-water is drawn into the system. But it may be that on occasion the direction of this current is reversed. Particles of carmine injected into the water-vascular system are taken up by the cells of Tiedemann's bodies (Kowalevsky). The axial organ, which goes by various names, e.g. dorsal organ, ovoid gland, heart, etc., is found in Echinoids, Asteroids (Fig. 130), Ophiuroids and Crinoids. It is developed as a fold EXCRETION. VASCULAR SYSTEM. 131 of the wall of the axial sinus (except in Crinoids, p. 159), and consists of connective tissue and of cells which have grown into it from the genital rudiment of the larva, It is covered towards the axial sinus by the coelomic epithelium. Its walls are folded so that it appears to be penetrated by tubular pro- longations of the epithelium of the axial sinus and of the peri- visceral cavity. Some observers have attributed to the axial organ a lymphatic gland function, and have supposed that it buds out amoeboid cells into the axial sinus, but this is ex- tremely doubtful. In Echinoids it takes up carmine injected into the body cavity (Kowalevsky). As stated above the primitive germ cells grow into it, so that it is connected in the adult with the generative rachis. The axial organ is absent in Holothurians. Excretion. Very little is known about excretion in Echino- derms. There do not appear to be any special organs devoted to it. It is possible that there may be some organ in connexion with the water-vascular system or with the axial sinus which is concerned with the elimination of the nitrogenous waste, for these organs open to the exterior by the water-pore but no such organs have been certainly identified. The so-called vascular system, which is found in all classes and appears to be specially well developed in the Holothurians and Echinoids, is formed of a peculiarly modified connective tissue in which the fibres are sparse, and which contains intercom- municating spaces without an epithelial lining. The fluid in these spaces does not appear to undergo any definite movement. The real nature of this tissue is doubtful. By some observers it has been regarded as a lymphatic gland, a view which is sug- gested by the appearance, sometimes found, of amoeboid cells being budded off from it. Typically there is a circumoral tract of it with radial prolongations which lie between the radial water- vessel and the radial nerve cord ; an annular aboral tract of it, in which the generative rachis is embedded and which sends extensions to the genital organs ; and in Holothurians and Echinoids a considerable development of it in the mesentery and on the gut wall. An account of its occurrence will be found in the description of the different classes. The generative glands * of Echinoderms are peculiar in the * G. W. Field, on the Morphology and Physiology of the Echinoderm spermatozoa, Journ. Morph., Boston, 2, 1895, p. 235. 132 PHYLUM ECHINODEEMATA. fact that they open directly to the exterior and apparently have no relation to the coelom. The researches of MacBride have however shown us that the cells composing them are, in Asteroids, Echinoids and Ophiuroids, coelomic in origin, being derived from lining cells of the body-cavity, and this will probably be found to hold throughout the phylum. Another peculi- arity of them consists in the fact that in all classes, with the possible exceptions of Holothurians (see below), they are connected with a cellular cord, disposed in different ways in the different classes and called the generative rachis. Indeed they may be regarded as swollen, fertile portions of this rachis, the rest of which is sterile and does not produce generative cells. As stated above the axial organ contains a prolongation of the generative rachis. There are no accessory glands of any kind. The generative products when ripe are in nearly all cases discharged directly into the sea, where the ova are fertilized and undergo their development. In rare cases the eggs are retained in special brood pouches or on the skin of the mother. Secondary sexual characters are as a rule not developed. The generative organs usually consist of tufts of branched tubes which open directly on the surface. In Crinoids they are contained in the pinnules as hollow structures without any opening to the exterior, and the manner of escape of the genera- tive cells is not certainly known, but it probably takes place by dehiscence. In Asteroids, Ophiuroids, and Echinoids they are pentamerously arranged in the interradii and parti- cipate in the radial symmetry of the body ; they open inter- radially and sometimes dorsally. In Holothurians, which are sometimes said to be without the generative rachis,* they are present as a single tuft of tubes which open in the adult-dorsal middle line. The Echinoderms are exclusively marine animals. f They are found on the floor of the ocean from the littoral to the abyss. Very few pelagic forms are known. They are almost all slowly * Holothurians possess a cord containing germ cells and passing from the point of union of the genital tubes along the generative duct towards the body wall (Theel, Bih. Svenska Akad. Handl., 27, Afd. 4, No. 6, 1902). This is probably the generative rachis. | Synapta similis inhabits brackish water of the mangrove swamps of Bohol (Ludwig). CLASSIFICATION. 133 creeping creatures, moving by means of their tube-feet or by their arms (Comatulids, Ophiuroids). Most Crinoids how- ever are attached by a stalk which is a prolongation of the aboral surface of the body, but they may become detached and acquire a new attachment (Pentacrinus). From a psychical or nervous point of view, the Echinoderms are extremely low in the scale of life, but judging them from the complexity of their organization apart from the nervous system, they approximate to the so-called higher animals. Geologically they are of great interest, the structure of the body wall lending itself very readily to their preservation as fossils. They make their appearance in the Cambrian and in the Silurian with a range of structure not unlike that which characterizes living forms. The power of regenerating lost parts is considerable in almost all Echinoderms, and many of them, especially those with long arms, possess the power of autotomy. It is least developed in Echinoids, but in Asteroids and Ophiuroids it is in some cases so extensive that not only may arms and portions of the disc be regenerated, but the whole body may be reformed from a single arm. In Crinoids and Holothurians the viscera even are capable of regeneration. Asexual reproduction is found in many Asteroids, Ophiuroids and Holothurians. It takes the form of fission into two equal halves, in the two former classes the plane of fission passing through the disc, in the latter transversely to the long axis of the body. In this work the Echinodermata are divided into seven classes : Asteroidea, Ophiuroidea, Echinoidea, Holothuroidea, Crinoidea, Cystidea, Blastoidea. The mutual relationships of these are discussed under the accounts of each class. The group Pelmatozoa is considered and the reasons for not adopting it are stated on p. 303 et seq. Development.* In the great majority of Echinoderms the * For a more complete account of the developmant of the Echinoder- mata, the readsr is referred to the following works, where the older litera- ture will be found : F. M. Balfour, A Treatise on Comparative Embryology, vol. 1, London, 1885; E. Korschelt and K. Heider, Textbook of the Em- bryology of Invertebrates, English Translation, London, 1895. A. Lang, Text-Book of Comparative Anatomy, English Translation, vol. 2, London, 1896. Of recent workers the following may be cited : E. W. MacBride, The development of Asterina gibbosa, Q.J.M.S., 38, 1896, p. 134 PHYLUM ECHINODERMATA. egg is small and the entire development takes place in the sea. and in most of them there is a free-swimming, externally bisym- metrical larval stage. In about fifty species however the young remain in connexion with the mother during their development and care of the brood occurs, and in some cases (e.g. Asterina gibbosa) a free-swimming larva is not formed. In such cases the egg is usually larger than when a free-swimming larva is formed, owing to the presence of a greater quantity of food-yolk.* When care of the brood occurs, f the young may either creep about freely on the surface of the mother, or they may be attached to special parts of the body wall, e.g. the neighbourhood of the mouth, among the spines of the back, in special depres- FlG. 87. Two stages in the development of Holuthitriu tiilndnxa viewed in optical section (from Balfour, after Seleuka). A, Blastosphere at the close of segmentation. B, com- mencing gastrula stage, ae archenteron ; bl wall of blastosphere ; ep ectoderm ; ft vitelline membrane ; hy endoderm ; mr micropyle ; ms protoplasmic immigration into the seg- mentation cavity from the invaginating endoderm ; sc segmentation cavity. sions of the skin. In some Holothurians the development takes place in the body cavity, and in the dioecious Chiridota contorta in the genital tubes. In the starfish Stichaster nutrix the young 339. Id., The development of Echinus esculentus, Phil. Trans., 1903, p. 285. H. Bury, The Metamorphosis of Echiiioderms, Q.J.M.S.* 38, 1895. H. Theel, The development of Echinocyamus pusillus, Nova Acta R. Soc. Sci. Upsala, 1892. Id. Prelim. Account of the development of Echinus miliaris, Proc. Roy. Swedish Acad., 28, 1902. G. W. Field, The larva of Asterias vulgaris, Q.J.M.S., 34, 1893, p. 105. S. Goto, The metamorphosis of Asterias pallida, etc., Journal Coll. Sci. Imp. Univ., Tokyo, 10, 1898, p. 239. Id., Metamorphosis of Asterina gibbosa, Ibid.. 12, 1898, p. 227. Th. Mortensen, Echinodermlarven, Ergeb. d. Plankton- Exped. d. Humboldt-Stiftung, 2, 1898. * In Benthodytes sanguinolenta the eggs measure 2 mm. in diameter. f Ludwig, Zoolog. Jahrb., Suppl. Bd., 7, 1904, p. 683. DEVELOPMENT. 135 undergo their early development in the stomach of the mother. Care of the brood is found most frequently in forms inhabiting the colder seas. With a few exceptions the early development, so far as it is at present known, may be summarized as follows. The egg is fertilized in the sea, undergoes a total cleavage, and becomes transformed into a hollow one-layered blastosphere. At one pole of the blastosphere an in vagination makes its appearance and a typical gastrula is formed (Fig. 87). The invaginated cells constitute the endoderm, the cavity bounded by them the archenteron, and the opening of the archenteron the blastopore. The archenteron rarely if ever fills the segmentation cavity, but the lat- ter soon becomes traversed by a nucleated proto- plasmic network (Fig. 88), which is continuous both with the endoderm and with the ecto- derm (outer cells of the gastrula), and constitutes the first trace of the meso- derm. This meso- dermal network, which is reinforced from the later appearing enterocoelic vesicles is commonly called mesenchyme and gives rise to the connective tissues, to some of the muscles, and to the calcareous structures of the body. Different views may be held as to the origin and- structure of this early appearing mesodermal network. The usual view is that it consists of amoeboid cells which arise by the proliferation of the epithelial cells of the embryo in the blastosphere and later stages, and wander into the hlastocoel. Another view for which there is much to be said is that the blastocoel at its very first appearance is traversed by protoplasmic pro- cesses of its walls,* and that the apparent proliferation of cells of the \ FIG. 88. Gastrula stage of Tox&pneustes bremgpinosus (from Korschelt and Heider, after Selenka). The mesodermal net- work (so-called mesenchyme) is shown traversing the blasto- coel and continuous with both ectoderm and endoderm. * See C. Shearer on the connexions between the blastomeres of embryos in the Proc. Roy. Soc., 1906. 136 PHYLUM ECHINODERMATA. endodermal pole of the blastosphere or of the invaginating endoderm, etc., is really only a proliferation of nuclei, and that the migration of cells into the blastocoel is nothing more than the shifting of these nuclei along the strands of the protoplasmic reticulum which has traversed the blastocoel from its first appearance. On this view the embryo must be regarded not as an aggregation of separate cells, but as a continuous structure with different densities of protoplasm in different parts. Thus in the blasto- sphere stage the bulk of the protoplasm is concentrated in a peripheral layer to which the nuclei are confined, the central portion being occupied by a sparse non-nucleated reticulum. As to the meaning of this peculiar peripheral aggregation of the protoplasm and nuclei, which is characteris- tic of the blastosphere of so many pelagically developing animals, we are quite in ignorance. It has been suggested that it represents some ances- tral stage of structure common to all the organisms in which it occurs. On the other hand it may be argued, perhaps with greater force, that the condition is due to some physiological necessity, possibly of a nutri- tive nature, which is felt by all organisms developing in the open sea. The undoubted fact of the presence of free leucocytes in the fluids of the body must not be held to be at variance with the view of the nature of the mesenchyme mentioned above. The view of the essential 'continuity of the mesenchyme is based upon Comparative Embryology. In all cases in which its structure has been thoroughly made out, it has the form of a network continuous both with ectoderm and endoderm. The free leucocytes which are found in the body fluids are products of this network, formed no doubt for a definite purpose. What that purpose is has been largely explained by the remarkable and penetrating observations of Metschiiikoff,* though the causes which govern the budding off of the leucocytes have not been ascertained. The eggs when laid are enclosed in the vitelline membrane (Fig. 87), but they soon acquire a coat of cilia and become free. The gastrula has a somewhat oval form with the blastopore at one end and one surface slightly flattened. The end carrying the blastopore is posterior and the flattened surface is ventral. The archenteron now bends towards the ventral surface anteriorly, where, its wall having fused with the ectoderm, it acquires an opening to the exterior ; this is the larval mouth, the posterior opening or blastopore persisting as the larval anus (except in Crinoids). The fate of the larval mouth and anus varies in the different classes. In Holothuroidea they usually persist into the adult. In Ophiuroids the larval mouth persists but the anus closes, the adult being without one. In Asteroids and Echinoids the larval mouth and anus close and the oesophagus atrophies, the adult mouth, oesophagus, and anus being new formations. It is possible, however, that in some Asteroids they may persist. In Crinoids there is 110 larval mouth or anus, and the blastopore closes at an early stage. * See the numerous papers by this author dealing with the importance of the amoeboid cells in the organism and especially his great work on Immunity, English Translation, Cambridge, 1905. DEVELOPMENT. 137 pr.c Later the anus shifts on to the ventral surface, and the cilia become confined to definite tracts or bands. According to the arrangement of these ciliated bands, two main types of larvae can be distinguished. In one (A, Fig. 89) there is a single longitudinal band of cilia which passes across the ven- A 8 tral surface immediately in front of the mouth and in front of the anus, and bounds a depression on the ventral surface into which the mouth opens. This is the Auricularia type of larva which is found in Holothurians ; a modifica- tion of it is found in Echinoids and Ophiuroids, where it is known as the Pluteus larva (Fig. 94). In the other (B, Fig. 89) there are two bands of cilia, the one preoral and encircling the preoral lobe, the other longitudinal in appearance, but really postoral, forming a complete circuit of the body between the mouth and the anus ; this is the Bipinnaria type of larva and is characteristic of Asteroids. The larvae of both these types are externally bilaterally symmetrical and have been so since the ventral sur- face became distinct. They remain symmetrical externally for some time longer, but at about this stage, or in some cases even sf- FlO. 89. A, the larva of a Holothurian (Auricularia type) ; B of an Asteroid (Bipin- naria type), both seen from the left side (from Balfour). a anus ; m mouth ; st stomach ; l.c longitudinal ciliated band ; pr.c preoral ciliated band. tn. ten FIG. 90. A series of diagrams representing the evolution of an auricularia larva from the simplest Echinoderm larval form, ventral view (from Balfour). The black line represents the ciliated band. The shaded part is the oral side of the ring (oral depression, see Fig..89), the clear part the aboral side of the larva, m mouth ; are anus. before this stage is reached, their internal structure begins to show traces of that asymmetry which is so characteristic of the later larvae and eventually leads to the establishment of the radial symmetry of the adult. Before, however, considering 138 PHYLUM ECHINODERMATA. these internal changes it will be well to consider briefly the various forms which the bilateral larva may assume. The auricularia larva. In the Holothurians the ciliated band becomes elongated and sinuous, as is shown in the series of diagrams in Fig. 90. A completely developed auricularia larva of a Holothurian is shown in Fig. 91. The sides of the body are prolonged into processes which correspond to the arms of the bipinnaria and pluteus, and there is a well-marked preoral lobe. Calcareous structures in the form of spheres, wheels and star-shaped bodies are formed, but there are no calcareous rods. In some species (Auricularia stelligera and sphaerigera) peculiar elastic spheres, the nature of which is not known, are present. -ffl FIG. 91. Auricularia stelligera, ventral view (after J. Mullen. 1 frontal area ; 2 preoral pro- cess ; 3 anterior, 4 posterior portion of the ciliated band : 5 postoral process ; 6 anal area ; 7 postero-lateral pro- cess ; 8 postero-dorsal pro- cess ; 9 oral depression ; 10 dorso-median process ; 11 antero-dorsal process. FlO. 92. Pupal stage of the larva of Synapta digitata (from Kor- schelt and Heider, after Se- inon). ed hind-gut ; ent right body-cavity ; m oral funnel ; w water-vascular ring, with out-growths into tentacles t and radial vessels p. The larva now enters the so-called pupa-stage, in which it has the form of a barrel with five ciliated hoop-like bands (Fig. 92). The pupa is formed from the auricularia in the following way : The ciliated ring breaks up (Fig. 93) into pieces which rearrange themselves into the five rings of the pupa and the oral ring ; the mouth and surrounding parts, including the oral ring, retreat into the interior, giving rise to the oral vestibule (atrial cavity), the opening of which narrows and passes to the left side. Event- ually the atrial opening becomes terminal by the atrophy of he small preoral lobe, and the epithelium of the oral ring DEVELOPMENT. 139 bscomes the epithelial covering of the tentacles which at first project into the atrial cavity. When the adult condition is FIG. 93. Synapta larva, showing the break up of the ciliary band of the auricularia. a anus ; ed prootodaeum ; ent euterocoel ; kr calcareous wheels ; m oral funnel or atrium ; mg stomach ; n nerve bands ; 10 water-vascular ring with prolongations (from Korschelt and Heider). attained the ciliated bands disappear, the oral vestibule opens out and the tentacles project. The pluteus larva. In the Ophiuroidea and Echinoidea the larva resembles the auricularia in possessing a single ciliated band, but differs from it by the long arm-like processes of the FIG. 94. Diagrammatic figures showing the evolution of an ophiuroid pluteus from a simple Echinoderm larva (from Balfour, after J. Miiller). The calcareous skeleton is not represented, m mouth ; an anus ; d antero-lateral arms ; d' the long postero-lateral arms ; ' postoral arms ; g' postero-dorsal arms. margin of the body on to which the ciliated band is continued, by the small size of the preoral lobe (frontal area) and by the 140 PHYLUM ECHTNODERMATA. 8 FIG. 95. Echinopluteus of a Spatangid (after J. Miiller). 1 frontal area ; 2 preoral arm ; 3postoral arm ; 4 anterior, 5 posterior transverse portion of ciliated band ; 6 unpaired posterior ami ; 7 anal area ; 8 postero-lateral arm ; 9 oral area ; 10 postero-dorsal arm ; 11 antero-dorsal arm ; 12 antero-lateral arm. great development of the postanal part of the body (Fig. 94). The arms are supported by calcareous rods. The plutei of Ophiu- roids and Echinoids are distinguished as ophio- plutei and echinoplutei respectively. In ophio- plutei the postero-lateral arms are always the largest and directed for- wards (Fig. 94), and pre- oral and antero-dorsal arms are absent. In echinoplutei (Fig. 95) the postero-lateral arms when present are directed back- wards or outwards, preoral arms are present, and in Spatangid larvae (Fig. 95) there is an unpaired posterior arm on to which the ciliated band is not continued. Moreover in some genera (Echinus, Strongylocentrotus, Sphaer echinus) there are in old larvae four ciliated projec- tions, called ciliated epau- lettes (Fig. 96), at the base of the postoral and postero- dorsal arms, and sometimes an additional pair at the hind end of the body. Ciliated lobes (auricular appendages, auricles) are cutaneous expansions be- tween some of the arms in certain echinoplutei, and pedicellariae may appear before the larval charac- ters are lost. In ophio- plutei the calcareous skele- FlG . 96 ._p lu t eus larva of Echinus KM with ton i=? in two Vmlvps parli four ciliated epaulettes TF (after Metschnikoff) -On IS m tWO naiVCS, eaCIl from the ventr al side. mouth ; A anus. u DEVELOPMENT. 141 half proceeding from one calcification centre ; in echinoplutei the skeleton arises from five or six centres. The bipinnaria larva, found only in the Asteroidea, has two ciliated rings, one preoral and the other longitudinal and postoral (p. 137 and Fig. 89 B). In possessing a well-marked preoral lobe it resembles the auricularia, and its two ciliated rings must be regarded as having been derived by the division of the single band of the auricularia, a view which is supported by the fact that in some species (e.g. Asterias rubens, A, glacialis, Astropecten) they are at first connected dorsally. Further, Driesch* has shown that in some artificially reared bipinnariae there is only one band. As in the pluteus a series of arms is formed along the lines of the two ciliated bands, and sometimes three arms, not connected with the bands and covered with warts, are formed in front of the preoral arms at the anterior end of the frontal area ; these are the brachiolar arms and the larva bearing them the Brachiolaria larva, f The arms are with- out calcareous rods. 12 The median brachiolar arm replaces the ven- tral median arm and bears at its base a sucker J by wliich the larva attaches itself during the metamorphosis. The ciliated band appears, as a rule, not to pass on to these arms (except in Bipinnaria papillata). FlG. 97. Bipinnaria elegans. frontal area ; 2 preoral arm ; 3 anterior, 4 posterior trans- verse portion of the ciliated band ; 5 postoral, 6 postero- lateral, 7 postero-dorsal arm ; 8 anal area ; 9 oral depression ; 10 antero-dorsal, 11 ventro- median, 12 dorso-median arm (after Mortensen). In the bipinnaria (and brachiolaria) the frontal area is well developed and surrounded by the preoral ring of cilia. There is a median ventral anterior arm and a median dorsal anterior (Fig. 97), neither of which is present in the auricularia or in the * Arch. f. Entwickelungsmechanik, 20, 1906, p. 13. f It is possible that the brachiolar arms, which serve for temporary attachment, are present in a later stage of all bipinnariae, for in the only two cases in which the life-history is fully known (Asterias glacialis, A. vulgaris) they are formed. J It is probable that in all bipinnariae this disc, which serves for fixation during the metamorphosis, is present at a later stage in all cases, for not only is it present in the two cases mentioned in the last note, but it occurs in all cases thus far examined where a shortened development due to food yolk has brought these stages within range of easy observation (Ludwig, op. cit. p. 134). (Cf. Anasterias, Asterias antarctica, Asterina gibbosa, Cribrella sanguinolenta.) U2 PHYLUM ECHINODEEMATA. pluteus. Antero-lateral arms are never present. The arms have muscles and are contractile, a fact which renders it diffi- cult to preserve these larvae without distortion. These are the principal types of larvae of the Echinodermata, but we must not omit to mention the vermiform Ophiurid larvae which are pelagic and the vermiform larvae of those Asteroids in which care of the brood occurs. The larva of Asterina gibbosa may be classed with the latter. Echinoderm larvae are found principally near the coast and in small waters : they are not characteristic of the plankton of the high seas. Only a few of the larvae are taken in more than one locality. In the nomenclature of Echinoderm larvae when the adult is not known, the name of the kind of larva is used as that of the genus and a specific name is added. In this sense there are four genera of larvae, viz. Auricu- laria (confining this term to Holothurian larvae), Ophiopluteiis, Echino- pluteus, and Bipinnaria (including brachiolaria). When the adult is known specific names are not required. j-j G . 98. Longitudinal-horizontal sections through three successive stages of the larva of Asterina gibbosa (from Korschelt and Heider, after Ludwig). A, showing the first origin of the enterocoel Vp from the enteron ; B, a later stage, the enterocoel is still in communi- cation with the enteron D ; C enterocoel cut off from the enteron and consisting of a median unpaired portion in front the anterior body cavity, and two paired portions behind, the right and left posterior body cavities. Bl blastopore; D intestine; Vp entero- coel (vasoperitoneal vesicle) ; r and I right and left side of the larva. The hydrocoel has not yet made its appearance. Development of internal organs and the metamorphosis. The coelom arises soon after the establishment of the gastrula as a single evagination of the enteron* (Figs. 98, 99). This soon separates from the enteron, and becomes divided into a number * The term vaso-peritoneal is sometimes applied to the enteric pouch DEVELOPMENT. 143 of sacs, one* of which is called the hydrocoel because it gives rise to the water-vascular system and its lining ; the others constituting the splanehnocoel, because they give rise to the body-cavity and its associated spaces of the adult. The mode of division is very similar throughout the group except in Crinoids, and is described below. In Asteroids we may take as type Asterina gibbosa in which the development has been so fully worked out by Ludwig and later by FIG. 99. Longitudinal section of three stages in -. .- T- i mi i i the formation of the enterocoel of Echinus mili- MaC-Bride. 1 he eggS Which am (from Korschelt and Heider). are 0'5 mm. in diameter are attached to stones to which they adhere by means of the vitelline membrane. On about the fourth day the embryo ruptures the vitelline membrane and escapes. It is then found to possess a large preoral lobe (Fig. 100) the edge of which is thickened, constituting the larval organ. The larval organ surrounds a central depression and is covered with specially long cilia, by means of which the larva can swim. Later, in the centre of the concavity of the larval organ, a small eleva- tion is formed. The ectoderm of this contains gland-cells which secrete an adhesive substance by means of which the larva fixes itself during the m e t a m o rphosis. The temporary fixation which is possible during larval life appears to be due to a kind of cupping action brought about by the application of the preoral lobe to the sub- stratum and the retract- ion of its central portion. FIG. 100. 'Larvae of Asterina gibbosa (after Ludwig). A, a younger stage, oblique ventral view. B, older stage from the side. Lu larval organ ; m larval mouth. and to the sacs which arise from it, the word enterocoel being retained for the peritoneal or body-cavity portions of the latter. We, however, prefer to call the totality of sacs enterocoelic, distinguishing the water- vascular portion as hydrocoel and the perivisceral portion as splanehnocoel. * Or one pair (see footnote on p. 126). 144 PHYLUM ECHINODERMATA. The early development and formation of the larval mouth and anus (p. 135) have already been dealt with. The enterocoel or coelom has the form of an anterior unpaired diverticulum of the enteron extending into the preoral lobe and sending back two lateral prolongations, one on each side of the larval gut (Fig. 98) . This soon separates from the gut, and the posterior prolonga- tions of it eventually meet one another ventrally, their opposed walls forming a ventral mesentery. The primary water- pore is now formed as a pocket of the unpaired portion of the coelom. It meets the skin of the dorsal surface just to the left of the middle line, and a perforation is formed at the point of contact. At the same time there is formed, first on the left side of the enteron and then on the right, a septum, by which the hinder parts of the two coelomic prolongations referred to above become separated off from the anterior part and form closed sacs called the right and left posterior coeloms respectively. The unpaired portion of the coelom, together with the anterior portions of the two prolongations, constitute the anterior coelom of the larva into which opens the primary water-pore. The hydrocoel is developed as an outgrowth from the hinder end of the anterior coelom on the left side, and while it is yet but faintly marked indications of its five prirnarj 7 lobes appear. These, which are numbered* 1, 2, 3, 4, and 5, Xo. 1 being the most dorsal, are arranged in a curve open anteriorly, the hydrocoel sac assuming a horseshoe shape (Figs. 101, 102). The two limbs of the horseshoe eventually come together, enclosing the stalk of the preoral lobe and the oesophagus of the adult, when that is formed on the left side of the larva. In this way the circumoral water-vascular vessel is developed from the horseshoe-shaped hydrocoel. The radial canals are developed from the lobes of the hydrocoel. Shortly after the formation of the left hydrocoel or hydrocoel proper, a rudimentary right hydrocoel is developed as an outgrowth from the hinder wall of the anterior coelom to the right of the middle line (Fig. 101, rJiy). It loses its connexion with the anterior coelom and persists into the adult as a small thin-walled sac beneath the madreporite. * The relations of these numbers to the enumeration adopted for the adult is explained on p. 120. The numbering here adopted has relation to the fact that the hydrocoel before it closes to form a ring has two ends, one of which may be called anterior. DEVELOPMENT. 145 The stone-canal arises as a groove along the anterior face of the posterior wall of the anterior coelom. The central portion of this groove be- comes closed to form a canal which opens at its posterior end into the hydrocoel between lobes 1 and 2 and at the other into the anterior coelom, which as we have seen opens to the exterior by the primary water-pore. Save for this communication the anterior coelom and the hydrocoel become completely separate. The anterior coelom persists into the adult as the axial sinus. When the primary water-pore becomes con- verted by subdivision into the numerous pores of the madreporite of the adult, most of these are found to lead directly into the stone-canal, but some open into the axial sinus. The stone-canal retains its opening into the axial sinus (anterior coelom). The left posterior coelom undergoes a very complicated development. It gives off dorsal and ventral outgrowths which grow on to the right side of the larval body, and a diverticulum which extends round the oesophagus and is called the oral coelom. These all become indistinguish- able in the adult and merely persist as part of the perivisceral cavity. In addition to these the left posterior coelom gives off four interradial prolongations, each of which bifurcates to proceed to the adjacent arm rudiments. These are the first traces of the outer perihaemal ring and of the radial perihaemal canals (Fig. 131). There is a fifth interradial coelomic prolongation, viz. that which lies between lobe 1 and 2 of the hydrocoel and is distributed to arms 1 and 2 ; this lies in the interradixis of the axial sinus and stone-canal and is an evagination of the anterior coelom. Lastly the aboral sinus is an outgrowth of the left posterior coelom, from which it becomes completely cut off in the adult (p. 146). The right posterior coelom remains much smaller than the left. It persists in the adult, as the part of the body-cavity (sometimes called epigastric) which lies between the gut and the aboral body wall of the starfish. It is prolonged into each arm as the space between the two mesenteries of each pyloric caecum, Z III L FIG. 101. Longitudinal horizontal section of a larva of Asterina gibbosa showing the origin of the hydrocoel and the relation of the coelomic sacs (after MacBride). a anterior body-cavity in the preoral lobe ; Ipc left pos- terior coelom ; Ihy 1. 2. 3 lobes of the hydrocoel, no. 1 being the most dorsal, no. 3 the posterior, no. 5 (not shown) is at the other end of the incipient ring and is the most ventral ; or.coral coelom ; rhy right hydrocoel ; rpc right pos- terior coelom. 146 PHYLUM ECHINODERMATA. so that these mesenteries are really remains of the septum between the right and left posterior coeloms. In the central region of the disc this septum seems to have disappeared in the adult. The anterior coslom, with the atrophy of the preoral lobe, becomes much reduced in size. As stated above, it persists as the axial sinus. The inner perihaemal ring is an outgrowth of the axial sinus, with the ventral end of which it remains continuous (Fig. 105, int.p.r.). The axial organ makes its appearance as a ridge projecting into the axial sinus and containing jelly, fibres and leucocytes. Later there is formed a thickening of the epithelium of the left posterior coelom near the aboral end of the incipient axial organ. The cells of this thickening are the primitive germ-cells. They become invaginated into the septum separating the axial sinus from the left posterior coelom, and thence grow out in two directions, viz. (1) into the ridge of the axial organ forming its core, and (2) as a cord extending in a direction parallel to the surface of the disc and constituting the generative rachis. The generative rachis becomes enclosed by a flap of peritoneum in a space cut off from the left posterior coelom and known as the aboral sinus. It forms an aboral ring on the dorsal side of the stomach and sends off, as it passes each interradius, two branches enclosed in corresponding branches of the aboral sinus. These branches end in swellings which become hollowed out, acquire a communication with the exterior and form the generative glands. The portion of the aboral sinus round the gonads be- comes cut off by a septum from the rest. The metamorphosis of Asterina gibbosa begins on the eighth day of its development. The larva fixes itself by its preoral lobe to the substratum by means of a thin mucilage which appears to be secreted by the adhesive disc, and it remains attached during the whole of the metamorphosis (Figs. 102 to 104). The following changes occur : (1) The constriction of the body into disc and stalk, the latter being formed from the preoral lobe. (2) The sharp flexure of the disc on the stalk, the disc being bent obliquely and to the left, so that the left side of the body is turned towards the substratum. (3) The prepon- derating growth of the organs of the left side, the left posterior coelom and the left hydrocoel having both sent out dorsal and ventral horns, which meet so as to form complete circles, while the right hydrocoel and the right posterior coelom remain small. (4) The gradual atrophy of the stalk. (5) The out- METAMORPHOSIS OF ASTERINA. 147 growth of the adult oeso- a phagus and the formation of the new mouth on the left side. Even before the meta- morphosis the lobes of the hydrocoel are visible on the left-hand side of the larva in a curve open anteriorly (Fig. 102), the dorsal lobe being no. 1, the posterior no. 3 and the ventral no. 5. On the right side are visible the first rudiments of the arms, lettered A-E, the dorsal being A and the ven- tral E ; these arm rudiments are developed over the left posterior coelom. The open curve of the hydrocoel eventually closes into a circle, the two ends coming together round the base of the preoral lobe (stalk), so that the preoral lobe arises from the oral surface of the developing disc (Fig. 105). In this closure the rudiment of arm E shifts so that it comes to lie directly on lobe no. 1 of the hydrocoel. When this fitting together of the two surfaces of the future disc is complete, the oral and aboral sur- faces of the starfish are fashioned, and it is seen that the oral surface is derived from the left and slightly ventral side of the larva, the aboral from the right and slightly dorsal side. Meanwhile the first cal- FlO. 103. Larva of Asterina gibbosa of T T i n nine days from the right side. The CareOUS plates are laid down Oil letters denote the rudiments of the j_\ i i arms, the arabic numerals the lobes of tne aboral Surface as SllOWn in the hydrocoel. The first formed tube- -pi- 1A/1 75 .C snip. Lees. ..../' snt.C FIG. 105. Longitudinal vertical section through a metamorphosing larva of Asterina gibbosa (after an original drawing by E. W. MacBride). The section passes through the interradius of the preoral lobe and through the opposite radius, and shows that the preoral lobe of the larva is attached to the oral surface of the adult, ad. in. adult mouth ; amp. ampulla of a tube-foot; ant.c. anterior coelom in the preoral lobe; ii/t./i.r. internal perihaemal ring- canal ; l.oes. vestige of larval oesophagus; l.p.c. left posterior coelom; p.or.c. perioral coelom; pyl. pyloric sac; r. />./. right posterior coelom; slum, stomach; t.f. tube-foot; w.v.ring water-vascular ring canal. completely double (Fig. 106), whereas in Asterina the anterior coelom is single from the first, the posterior alone being double. Later, after the division of the coelom into anterior and posterior in bipinnaria, the two COMPARISON WITH ENTEROPNEUSTA. 149 anterior coeloms unite in the front part of the preoral lobe to form a single cavity.* In bipinnariae with brachiolar arms, which are it will be remembered processes of the preoral lobe, temporary fixation f during development takes place by these arms. During the metamorphosis of bipinnariae fixation occurs by means of a median oval sucker-like disc on the preoral lobe (p. 141). Attachment by the preoral lobe has also been noticed by Perrier in Asterias spirabilis, in which the larvae adhere to the buccal membrane of the mother. From the above account f it is clear that in the larvae of Aste- roids which become attached dur- ing development, the fixation is effected by the preoral lobe, and that the stalk so formed is sur- rounded by the hydrocoel and springs from what will become the oral surface of the starfish. The knowledge of this fact, so impor- tant for a proper comprehension of the morphology of the class, we owe to MacBride. The further statement may be made that the coelom, which arises by a single diverticulum from the enteron, becomes segmented into three pairs of chambers, viz;, the anterior coelom, the hydrocoel, and the posterior coelom. Of these the anterior coelom is at first single or soon becomes so, and, as in the Enteropneusta, acquires an opening or two openings to the exterior (water pores). Of the second pair the chamber of the left side becomes much larger than that of the right, retains its connexion with the anterior chamber by the stone-canal, and becomes the hydrocoel, while the right remains as a small apparently functionless sac ; this pair of chambers may be compared to the collar-cavities of the * There are according to Field (op. cit. ) two water-pores in many of the bipinnariae of Asterias vulgaris, but it is doubtful if this can be regarded as a normal occurrence (see note, p. 166). t Bury, op. cit., 1895 ; Delage, Arch. Zool. Exp. (4), (2), 1905, p. 27. f The preceding account of the development of Asterina is taken largely (often directly quoted) from the important work of MacBride (loc. cit.). 4-f \- FIG. 106. Optical section of an Asteroid larva (Bipinnaria) seen from the dorsal surface (after Bury). There are two anterior coelomic sacs, the water pore is shown opening into the left of them. The hydrocoel is indicated but not yet completely separated from the left anterior coelom. 1 left anterior coelom ; 2 water pore ; 3 hydrocoel : 4 second terminal plate (arm B) ; 5 left posterior coelom ; 6 mesentery separating the two posterior coeloms ; 7 right posterior coelom ; 8 right anterior coelom ; 9 euteron. PHYLUM ECHINODERMATA. Enteropneusta, which in the Cephalodiscida furnish the tentacles. The chambers of the third pair both persist and furnish the perivisceral cavities of the adult, but the left be- comes, in accordance with the predominance of the organs of the left side of the larva, much larger than the right, and alone furnishes the germ cells from its lining. The knowledge of these facts we owe mainly to Bury and MacBride. Bury was the first to grasp the importance of the anterior coelom and its trans- ference into the axial sinus, while MacBride went a step further in showing that the coelom was segmented into three chambers on each side, though Metschnikoff * preceded him in having demonstrated the presence of the right hydrocoel in Amphiura squamata. In Echinoids (Fig. 99) the single enterocoel divides into two, one on each side. Each of these again di vides into two, the hindermost of which lie at the sides of the stomach and constitute the posterior coeloms. The left anterior division develops a water-pore, which is placed on the left side of the dorsal surface, and then becomes constricted into two parts, of which the anterior retains the water-pore and persists into the adult as the madreporitic ampulla and axial sinus, corresponding to the left anterior coelom of Asterids, while the posterior becomes the left hydrocoel. The separation between these two vesicles appears not to become complete, the connecting tube persisting as the stone-canal. The left hydrocoel has at first the form of a disc which is soon transformed into a ring (ap- parently by bacoming notched on one side) through which the adult oesophagus later grow s. The right anterior division likewise constricts into two parts, but does not develop a water-pore. The anterior of these is the right anterior coelom ; its fate is unknown. The posterior portion is the right hydrocoel, which remains small, never develops lobes, and persists into the adult as the " dorsal sac " which lies beneath the madre- porite. The segmentation of the coelom therefore proceeds in a very similar manner to that of Asterids, and as in them the segmentation of the left side precedes that of the right. The two posterior coeloms give rise to the general perivisceral cavity. The lantern coelom, which is homologous with the outer perihaemal ring of Asterids, develops as five evaginations of the left posterior coelom. The teeth and jaws are developed from the walls of these pockets and the radial perihaemal canals are outgrowths of them. The genital rachis, genital organs and aboral sinus are developments of the left posterior coelom, exactly as in Asterids. An invagination of ectoderm which becomes closed is formed on the left-hand side of the larva. Its cavity is known as the amniotic cavity, its outer wall becomes thin and is called the amnion, while its inner wall or floor applies itself to the hydrocoel and forms the ambulacra! surface of the adult. The epineural canals are developed as imaginations of the floor of this cavity. There is no fixation of the larva during the metamorphosis. Lastly it must be men- * Studien iib. d. Eat. d. Echinodermen u. Nemertinen, Mem. de VAcad. de. St. Petersbourg, 14, 1869. DEVELOPMENT OF OPHIUBIDS AND HOLOTHURIANS. 151 tioned that there is a well-developed nervous system in the larva of Echinus. It has the form of an apical plate of neuro-epithelium, placed on the preoral portion of the body dorsal to the ciliated band and between the preoral arms. It is not recognizable till the larva is three weeks old. In the metamorphosis the animal falls to the bottom, the amnion rup- tures and shrivels up, the larval arms are absorbed by phagocytic amoe- bocytes, the larval mouth and anus close and the somewhat spherical form of the adult is assumed. At first the metamorphosed animal is with- out a mouth and anus ; these soon appear, the mouth first. The anus is formed in the centre of the antambulacral surface which is at first equal in area to the ambulacral. In Ophiurids a single coelomic sac is budded off from the anterior end of the archenteron. This soon divides into a right and left sac. Each of these divides into anterior and posterior coeloms. The hydrocoels FIG. 107. Longitudinal vertical sections of the two stages of the larva of Synapta digitate showing the formation of the enterocoel and of the primary water-pore P. Bl blastopore (after Selenka). then arise from the hind end of the anterior coelom in the usual way (MacBride) and the left anterior coelom develops a water-pore. The left hydrocoel alone acquires lobes and develops into the water-vascular system. As in Asterina the rudimentary right hydrocoel in abnormal specimens occasionally acquires a form similar to the left. In Holothurians the single pouch acquires an opening to the exterior by a pore placed on the dorsal surface and just to the left of the middle line. This is the primary water-pore. In Synapta it may even be formed before the enterocoel has separated from the enteron (Fig. 107). Soon after the formation of the pore, the enterocoel divides into two portions, an anterior and a posterior ; the anterior remains in connexion with the water-pore and constitutes the combined anterior coelom and hydrocoel, while the 152 PHYLUM ECHESTODERMATA. posterior is the splanchnocoel. The splanchnocoel then divides into two sacs which apply themselves to the gut, one on the right and the other on the left, and give rise to the perivisceral cavity and its lining. The tube connecting the combined anterior coelom and hydrocoel with the water- pore elongates and a small swelling appears on its 5 anterior wall (Fig. 108). This is supposed to be the reduced anterior coelom. The vesicle itself becomes lobed and forms the hydrocoel. It eventually surrounds the oesophagus to form the circumoral water-vascular vessel and gives off five outgrowths which become the radial canals. The canal connecting the hydrocoel with the small anterior coelom must be regarded as the stone-canal. In the forms with a so-called inter- Fio. 108. Diagram of the na [ madreporite, it must be supposed that the hvdrocoel and anterior , . , , /. ., n coelom of an old Holo- canal (water-pore) distal 01 the small anterior thurian auricularia (after coelom breaks down and that the anterior coelom Bury). 1 water-pore ; 2 .^.-^1^1 i pore canal ; 3 sand-canal ; acquires a tree communication with the general 4 polian vesicle ; 5 an- perivisceral cavity by the rupture of its walls, as it terior coelom. j n -j / i KO\ does in Urmoids (p. 158). Development of Crinoids. The development of Crinoids differs considerably from that of the other classes and requires separate treatment. The principal points of difference concern (1) the form of the larva ; (2) the relation of the larval preoral lobe to the adult surfaces ; (3) the fact that, though the hydro- coel at first undergoes a displacement to the left side, the oral surface of the adult is the posterior surface of the larva and the left posterior coelom does not exceed the right in size ; (4) the posterior coeloms arise from the enteron independently of the common rudiment of the anterior coelom and hydrocoel ; and (5) the absence of any trace of a right hydrocoel. In Antedon, the only Crinoid the development of which is known, the egg, as in most other Echinoderms, is fertilized and undergoes its whole development in the sea-water, but it remains for some time within the vitelline membrane attached to the pinnules of the parent. The total cleavage leads to the forma- tion of a hollow blast osphere from which a gastrula arises by invagination. The blastopore closes completely at or near the hind end of the embryo, and the uniform ciliation gives place to five ciliated bands which encircle the body transversely and to a ciliated tuft at the anterior end (Fig. 109). The ciliated tuft springs from a thickened patch of ectoderm which con- stitutes a neural apical plate. In the deeper layers of this neural plate nerve-cells and fibres are formed and constitute the larval DEVELOPMENT OF ANTEDON. 153 C.T nervous system. The embryo usually leaves the egg-membrane on the seventh day, and becomes the free-swimming larva, which shows bilateral symmetry, the ventral surface being slightly flattened. The anterior ciliated ring is incomplete ventrally, and between the second and third, which are separated by a wider interval than the others, there is a ciliated depression (Lm), called the vestibular depression (so-called larval mouth) and supposed to correspond with the now closed blastopore. On the ventral surface between the first and second rings, there is a small pit, the adhesive pit (6V), by the secretion of which the larva, after a free-swimming life of from 12 to 48 hours, attaches itself. After attachment the larva at first lies with its entire ventral surface turned towards the surface to which it is attached ; soon however it erects itself and projects at right angles to the substratum. The attached, i.e. anterior, end of the larva now be- comes narrow and elongated into the stalk, while its free, i.e. posterior, end becomes broader and constitutes the rudiment of the calyx (Fig. 115). The vestibular depression has during these changes become cut off from the ectoderm and forms a closed ecto- dermal vesicle (Fig. 114, 4), which constitutes the larval vestibule. At first this vesicle is placed on the ventral surface, but soon it comes to occupy the free end (Fig. 115). This change in posi- tion is shared by certain internal organs, and is doubtless due to the relative growth of parts by which the ventral surface of the larva comes to occupy the free end. The floor of the vestibule applies itself to the adjacent internal organs and eventually becomes perforated by the mouth opening. Mean- while processes from the water-vascular ring, the tentacular canals, push before them the floor and project as tentacles into the cavity of the vestibule. Eventually the roof of the vesti- Flo. 109. Larva of Antedon rosacea with ciliated bands, anterior tuft of cilia and rudiments of the skeletal plates (from Korschelt and Heider). Gr. adhesive pit by which the larva attaches itself ; Lm. the vestibular depression (larval mouth). 154 PHYLUM ECHINODERMATA. bule disappears and the tentacles and mouth become freely exposed. We thus reach the so-called Cystid stage of develop- ment (Fig. 110). The larvae of Cri- noids, then, become attached by the ventral side of the anterior end, and, as was shown by Bury, the stalk of the adult is a de- velopment of the preoral lobe. They therefore resemble the larvae of Aste- roids in the fact that attachment takes place by the preoral lobe, but differ entirely from them in the relation which the preoral lobes bear to the arrangement of the organs in the adult ; for whereas in Aste- roids the preoral lobe is encircled by the water- vascular ring, and its with- ered vestige springs from the oral sur- face of the adult disc (Fig. 105), in Crinoids it is quite FIG. 110. Cystid larva of Antedon (after Thomson). free of the circum- oral vessel and arises from the apical or aboral surface of the adult. Our knowledge of the development of the coelom is mainly due to Bury, who in his memoir on the development of DEVELOPMENT OF ANTEDON. 155 Antedon* first made us familiar with the conception of the ante- rior coelom and so paved the way for the modern views, largely due to him and MacBride, on the segmentation of the coelom into three chambers on each side of the body (p. 149). B _-*r"v VT mes-. A> <^.: \ ^S?BB3j3 $T $#. * M<^&iM -mot _sr t.1 Fia. 111. 4 longitudinal vertical section of gastrula of Antedon at the end of the second day showing the formation of mesenchyme. bl blastopore (after Bury). B longitudinal section of a later stage showing the division of the archenteron. mes mesenchyme ; ent first-formed enterocoel which gives rise to the right and left posterior coeloms (after Barrois). After the closure of the blastopore the archenteron which is placed at the hind end of the embryo divides by a constriction into an anterior and a posterior portion (Fig. Ill B). The posterior portion is the first enterocoel vesicle ; it lies close to the hind end of the embryo and soon divides into a right and left part, which constitute the right and left posterior coeloms respec- tively. The anterior portion or vesicle into which the archenteron has divided de- velops two outgrowths, a dorsal and a ventral (Fig. 112, 5, 7), which give origin to the intestine ; and itself becomes constricted into two portions, a ventral one which is the rudiment of the hydrocoel, and an anterior one which is the anterior coelom (Fig. 112). These become separated, in * Phil. Trans., 179, 1888. ia. 112. Posterior end of an embryo of Ante- don of sixty hours, from the right side. 1 the outline of the right posterior coelom ; 2 rudi- ment of anterior coelom ; 3 rudiment of hydrocoel ; 4 mesenchyme ; 5 ventral, 7 dorsal part of enteron, which clasps 6 the still persistent connecting portion between the incipient right and left posterior coeloms (from Lang, after Seeliger). 156 PHYLUM ECHINODERMATA. continuity with one another, from the rudiment of the intestine. Soon after this has happened the anterior coelom separates from the hydrocoel, which, placed between the enteron and the ectoderm on the left-hand side of the body (Fig. 113), at once acquires the characteristic horseshoe form. The anterior coelom then acquires its external opening, the water-pore, which is placed on the left side of the body just in front of the fourth ring of cilia. Later the hydrocoel develops an anteriorly directed outgrowth which acquires an opening into the anterior coelom and forms the stone-canal. The posterior coeloms now shift ; the left-hand one moves posteriorly and comes to lie like a cup over the hind end of the enteron (Fig. 113, <5), while the right sac extends ante- riorly and following the enteric wall reaches on to the left side (Fig. 114). The right posterior coelom gives off from its anterior end five forwardly directed diverticula (Fig. 113, 2). FIG. 113. Longitudinal-vertical section of a free- Tlif>f> hppnmp nut nfif from swimming larva of Antedon.. 1 stem-joints; 2 anterior prolongation of right posterior coelom if a f r, lofor eta o-p anrl to form the chambered organ ; 3 right posterior ] coelom ; 4 enteron ; 5 left posterior coelom 0iyp rise to the Chamb8r6d 6 hydrocoel ; 7 anterior coelom. organ. The calcareous plates make their appearance in the embryo on the sixth day. They are shown in Fig. 114, which however is taken from a larva after attachment. There are five orals (Fig. 114, or) arranged in a horseshoe curve near the posterior end. The horseshoe is set obliquely to the long axis, its dorsal end being posterior to its anterior end, and the open end of it is directed ventrally. Parallel to this row, but anterior, are the five basals (ba) set in a similarly disposed horse- shoe curve. Both the orals and basals are, as shown by the later development, interradially placed. In front of the basals ---.-6 DEVELOPMENT OF ANTEDON. 157 but deeper and in the axis of the body is a row of stem-joints. The anterior of these is the dorsocentral (Fig. 114, 7), which becomes the terminal joint of the stem. The stem-joints rapidly increase in number, the new pieces being added at the posterior (proximal or calycine) end of the row. A little later, on the seventh day, the underbasals (ib), three in number (rarely four or five), are formed. They lie in front of the basals and a good deal deeper and eventu- ally fuse with the top (posterior) stem joint to form the centro- dorsal plate. At about this stage the larva hatches and undergoes its brief free-swimming life. It then attaches itself (p. 153), the ciliary rings, preoral tuft and apical plate atrophy, and the anterior end begins to become narrower and longer, and to mark itself out as the stem from the pos- terior end, which becomes the calyx. In fact the larva be- comes club-shaped, the swollen free end forming the rudiment of the calyx (Fig. 114). The vestibular depression (p. 153) on the ventral surface becomes deeper and converted into an ectodermal invagination which occupies the greater part of the ventral surface, remaining open for some time anteriorly. It eventually closes, and soon after attachment shifts on to the posterior end of the larva (Fig. 115, 5). The hydrocoel follows this shift of the vestibule and lies at the posterior end immediately beneath the floor of the vestibule (Fig. 114, 3). It is still an open horseshoe, the opening being towards the water-pore ; but its ends have approximated and its five lobes, each of which soon becomes trilobate, pushing up the ectoderm of the floor of the vestibule, project into the vestibular cavity as the primary tentacles. Five additional pairs of tentacles are formed later at the base of these (Fig. 115, 7). FIG. 114. Young attached larva o Antedon from the left side (from Lang, after Seeliger). The vestibule is closed. ba 1-3 basals ; or 13 orals ; ib under- basals ; 1 dorsocentral plate ; 2 anterior coelom ; 3 lobes of hydrocoel ; 4 vestibule ; 5 enteron ; 6 left, 7 right posterior coelom ; 8 stem-joints ; v. ventral ; d dorsal surface of the larva. 158 PHYLUM ECHINODERMATA. The mouth is formed as a funnel-shaped depression of the vestibule, which passes through the hydrocoel ring and opens into the enteron (Fig. 115, 8). On the shifting of the vestibule to the hind end, the oral plates which, like the basals, have become arranged in a circle, also shift backwards. They come to lie in the thin roof of the vestibule, and when the latter ruptures and splits into five lobes, each lobe contains one oral plate. The stage we have now reached is sometimes called the cystid stage (Fig. 116, b). It is characterized by having a mouth overhung by five oral plates, an absence of arms, and an anus which has been formed as a lateral perforation through the body wall outside the circle of the orals. There are twenty-five ten- tacles, which at first arose in five groups, but now all spring sepa- rately from the water - vascular ring. As to internal changes, we may mention that the mesenteries be- tween the right and left posterior coeloms, which, by shifting, have some time before become aboral and oral, break down so that the perivisceral cavity is a continuous cavity. Further, the anterior coelom loses its walls and becomes merged in the general body-cavity. The result of this is that the primary water-pore and the sand- canal, which at the previous stage both opened into the anterior coelom, now open directly into the body-cavity. The primary water-pore and the anus lies in the same interradius. Later each of the other interradii Flo. 115. Attached larva of Antedcm after the separation of the stalk and calyx (from Lang, after Seeliger). The calcareous plates are not shown. d dorsal, v ventral side. 1 right pos- terior (aboral) coelom ; 2 stomach ; 3 left posterior (oral) coelom ; 4 sac- culi ; 5 vestibule ; (i primary ten- tacles, 15 in number derived from the 5 lobes of the hydrocoel each of which becomes 3-lobed ; 7 the secondary interradial tentacles (in 5 pairs) ; 8 oesophagus ; 9 hind gut ; 10 axial organ ; 11 fibrous strands in the stalk. DEVELOPMENT OF ANTEDON. 159 acquires a water-pore and sand-canal which lead from their first appearance into the general body cavity. The genital stolon (axial organ) develops as a thickening in FlO. 116. Larvae of Antedon (after Thomson), a free-swimrning larva with rings of cilia; b attached cystid stage ; c older stage described as Pentacrinus europaeus with arms and cirri. the epithelial wall of the aboral coelom. The genital rachis is probably developed as an outgrowth of this. The larva becomes converted into the pentacrinoid larva 160 PHYLUM ECIIINODERMATA. (Fig. 116, c) by the formation of the arms, which grow out from the sides of the body aborally to the water-pore and anus and between the oral and basal plates. The result of their out- growth is that the oral surface of the body is much enlarged by the formation of the tegnien calycis beyond the circle of the oral plates which become reduced in size and eventually disappear. On the aboral side of the calyx, calcareous plates, the radials, are formed to support the growing arms. Finally the calyx be- comes detached from the stalk, and the free adult state is reached. The Crinoids, in so far as relates to the development of the coelom, differ from all other Echinoderms in the fact that the rudiment of the posterior coeloms is budded off from the gut independently of the anterior coelom and hyclrocoel. They resemble certain of the Holothurians in the fact that the anterior coelom becomes merged in the general body cavity, but differ from these in the retention of the water-pore. In becoming attached by the preoral lobe they resemble the larvae of certain Asterids, but they differ, as already explained (p. 154), from these in the relation which the attaching surface bears to the adult structure. Affinities. The fundamental fact in the moi-phology of the Echinoderms is the enterocoelic origin of the coelom. In this, as has been already pointed out in the second volume of this work (chap, i., p. 7) they are associated with the Brachiopoda Chaetognatha, Chordata and probably the Phoronidea. Whether we are to regard this fact as indicating affinity it is difficult to say. In the absence of evidence tending to unite any of these groups more closely or as closely with any other group of the animal kingdom, we may perhaps consider this common feature as sufficient justification for treating them in immediate succession to one another, but we must not attribute too much importance to it, for it is absent from both vertebrate and tunicate development, nor is it found in Annelids, Arthropods and Molluscs, the coelom of which is clearly the homologue of the coelom of enterocoelic forms. The question now presents itself, do the Echinodermata possess any features which enable us to associate them more closely with any particular phylum of the Enterocoela than with the others ? It has been pointed out by some zoologists, amongst whom I may specially mention MaoBride, that in the primitive disposition of their coelomic AFFINITIES. 101 sacs they present a certain resemblance to the Chordata. In all the members of that great group with the exception of the Tunicata, the coelom in its first state in the embryo presents traces more or less marked of three divisions : these are (1) the anterior or proboscis coelom, which in Vertebrata and Entero- pneustais single, in Amphioxus double, (2) the collar or middle coelom which is always double, and (3) the trunk coelom which is double and which in Vertebrata and Amphioxus becomes metamerically segmented. In Echinodermata we seem to be able to make out indications at least of a similar tripartite division. We have (1) the anterior coelom which is sometimes single (Asterina), sometimes double (Echinus], (2) the hydrocoel which is probably fundamentally double though in some cases only one hydrocoel sac is formed (Holothurians, Crinoids), and (3) the posterior coelom which is always paired. Of these the hydrocoel presents the peculiarity of growing out into ten- tacles a feature which is also presented by the middle division of the coelom in the enteropneust genera, Cephalodiscus and Rhabdopleura. But with these similarities we have to note certain differences. In the first place in the Chordata, in which the enterocoelic origin of the coelom is clearly presented, these three divisions of it always come off from the enteron separately, whereas in Echinoderms the enteron at most gives off only one pair of coelomic sacs. Further, whereas in the Chordata the middle coelom (collar) is never associated more closely with the anterior than with the posterior, in the Echinoderms it is always closely associated with the anterior coelom, being developed from it and remaining connected with it by the stone-canal throughout life. With regard to these differences we have only to say this : that they are differences such as we might expect from the greater remoteness of the Echinoderms from the Chor- data than of any of the Chordata from each other, but that they are not sufficiently great to put out of court the homologies suggested by the comparison. To turn to other points of resemblance : we have the resem- blance (1) in the central .nervous system, (2) in the skeletal system, (3) in the shifting of the mouth and in the asymmetry of the body, and (4) in the larval form. To take these in order : (1) In the Chordata, as is well known, the central nervous system z in M 162 PHYLUM ECHINODERMATA. never becomes separated by mesodermal tissues from the tract of ectoderm which gave it origin in the embryo. This is a feature of all Echinoderms in so far as the ventral nervous system, which is the predominant central nervous system, is concerned. When this nervous system is removed from the sur- face, the removal is effected by invagination (p. 124). (2) The presence of calcined skeletal tissue in the meso- derm of the body wall is a character found in Echinodermata and Vertebrata alone among Coelomata. This has already been pointed out by MacBride, and though not perhaps a very important indication of affinity is one which from its rarity deserves mention here. (3) In all Echinoderms the mouth shifts from the ventral surface of the larva on to the left side of the body. This can be demonstrated in all classes except Crinoids, and in Crinoids it may fairly be inferred. In Chordates a similar though not identical phenomenon is presented by > Amphioxus. In this animal the mouth actually makes its appearance on the left side in an animal otherwise bilaterally symmetrical, but the phenomenon differs from that of Echinoderms in the fact that the left-sided position of the mouth is not preceded by a con- dition in which it is in the middle ventral line. The feature then which Echinoderms have in common with Amphioxus is the sinistral position of the mouth. Here again we have a character which strikes us from its very rarity, for it is found in no other Coelomate nor so far as we know in any other mem- ber of the animal kingdom. It also strikes us by its strangeness and inexplicableness. In Amphioxus no serious attempt has been made to explain it. In attempting to explain peculiarities of this kind we are accustomed to take into consideration two factors which must be kept distinct. Firstly we have peculiarities in habit, secondly associated peculiarities in other organs. Now in Amphioxus the asymmetry of the mouth is accompanied by no peculiarity in habit, for the animal while it has this monstrous mouth behaves more after the fashion of a bilaterally symmetrical animal than it does in later life, when the mouth has acquired a more median position and it has taken to burrowing in sand. Nor do the peculiarities in some of the other organs lend us any assistance, for no one, so far as we know, has ever attempted to bring the RADIAL STRUCTURE : ASYMMETRY. 163 extraordinary features in the development of the gill clefts, of the endostyle, of the head-cavities, the asymmetric positon of the anus and olfactory pit, into relation with the asymmetry of the mouth. The thing cannot be done. There is no sort of con- nexion between these various asymmetries. They seem to> occur without rhyme or reason. The mouth which should be- a median structure is from the first on the left side ; the gill- clefts, which are on the left in the adult, appear in the median line and at once pass on to the right side ; the endostyle which is a median structure in the adult appears as an entirely dextral organ. In Echinoderms on the other hand the asymmetry of the mouth is accompanied by changes in habit and by change of other organs which seem to be connected with the change in the mouth. The animal here becomes sessile or semi-sessile and acquires an entirely different symmetry in which other organs of the body participate in an intelligible manner. But though we can understand to a certain extent that the shift of the mouth might indent the left hydrocoel and bring about an inequality in the posterior coeloms, no adequate attempt has ever been made to show how the sessile habit and the radial structure is connected with the shifting of the mouth on to the left side. We have here three factors : the sinistral mouth, the radial structure and the sessile habit. Can these factors be brought into the relation of cause and effect ? (1) Can the sessile habit be regarded as the cause of the other two, even if we accept the view that all Echinoderm classes have passed through a fixed stage in their phylogeny. We can only point out in reply that no such results have followed fixa- tion in any other group of the Coelomata : they have not followed in Cirripedes, Brachiopods or in Tunicates. (2) Can the left-handed mouth be regarded as the cause ? In Amphioxus, the only other animal in which the mouth is sinistral, it is accompanied neither by the sessile habit nor by the radial symmetry. (3) Lastly, can the acquisition of radial symmetry, to whatever cause due, have brought in its train the shifting of the mouth and the sessile habit ? In the only other animals which can lay claim to a radial symmetry, the Coelenterata, no such result has followed. 161 PHYLUM ECHINODEEMATA. It is therefore no more possible to explain the sinistral position of the mouth in Echinoderms than it is to account for the same phenomenon in Amphioxus. But that conclusion does not in any degree diminish the importance of the character as an indi- cation of affinity. On the contrary it increases it. For, if it cannot be shown to be connected with habit of life or with other peculiarities of structure in the animals presenting it, the pre- sumption that it is a property which was possessed by the ancestral matrix from which Echinoderms and Amphioxus have emerged is increased. We are thus brought back to the ques- tion which we touched upon on p. 116, are the Echinoderms descended from asymmetrical or from bilaterally symmetrical forms ? This discussion of the asymmetry of Amphioxus and Echinoderms has elicited facts which are not without a bearing upon this question. We have seen that in Amphioxus there is hardly a single organ of the body which displays complete bilateral symmetry at all stages of existence, and in the adult traces of this asymmetry slight traces, it is true, but all the more striking on that account are present (position of olfactory pit and anus just to the left of the middle line, preoral hood and other small distortions). We have also seen that the asymmetry of one organ is entirely independent of the asymmetry of the others. Very similar statements may be made about Echino- derms : in these also development begins with a transitory bilateral symmetry which is almost at once followed by asym- metry, at least of the internal organs, and the asymmetry then initiated is never completely got rid of, for the radial symmetry of the adult is in all classes imperfect (least so in Holothurians), and some of the adult distortions, such as the position of the anus, recall the similarly slight distortions found in the adult Amphioxus. Now the upshot of these considerations is to make us pause in accepting as final the conclusion that the ancestral Echinoderm was a bilaterally symmetrical animal. (4) The striking resemblance of the bipinnaria larva to the tornaria of Enteropneusta has already been referred to (p. 99). It is impossible to estimate its value, but it clearly cannot be passed over in a discussion of this kind, and taken in conjunc- tion with the other facts mentioned must be admitted to have considerable weight. We have now passed in review all the points of resemblance AFFINITIES : PHYLOGENY. 165 between the Eehinodermata and the Chordata and we have seen that they all, except possibly the last, are concerned with fundamental, not superficial, traits. Some of them link the Echinoderms to all the Chordates or to all except the Tunicata, e.g. the relations of the central nervous system and the general relations of the coelomic sacs ; another they hold in common only with Vertebrata, viz. the presence of calcified plates in the mesoderm of the body wall ; a third the sinistral position of the mouth is found again only in Amphioxus ; and lastly there is the striking resemblance to Enteropneusta by the tornaria larva. The accumulated weight of these facts is over- whelming and leaves us no choice but to consider not only that the Chordata are the nearest allies of the Eehinodermata, but that the Eehinodermata are of all Coelomata the nearest to the Chordata. 8 '-7 '4 '5 Fia. llGbii. Diagrammatic representation of the supposed Dipleurula ancestor of Echino- derms, seen from the left side, with the ventral surface towards the substratum (after Bather). 7 right and left water-pore; 2 preoral lobe ; 3 nerve plate of preoral lobe; 4 anterior coelom of the left side ; 5 month ; 6 left posterior coelom ; 7 anus ; 8 right posterior coelom ; 9 right hydrocoel ; 10 left hydrocoel. We now come to a consideration of the so-called Dipleurula, a hypothe- tical form which has been imagined by some zoologists as the bilateral ancestor of all Echinoderms. The structure of this hypothetical animal will be understood at a glance from an inspection of Fig. \\Qbis, which represents it as seen from the left side with its ventral surface turned towards the substratum. It is a bilaterally symmetrical animal with a preoral lobe (2) carrying a nervous plate (.3), a ventral mouth (5), a terminal or ventral anus (7), three coelomic vesicles on each side (4, 10, 6), and two water-pores (1). How has this form been arrived at ? It is not arrived at by selecting features common to all the free-swimming bilateral larvae (sometimes called Dipleurula larvae) of Echinoderms ; but by picking and choosing from among the characters of different larvae those which, according to the preconceived ideas of its authors, the common ancestor might be supposed to have possessed, and adding one or two characters which none of them possess. For instance, the preoral lobe is. very, small in pluteus larvae, though well developed in bipinnaria and- in the larva of Crinoids ; a preoral nervous system has been detected 166 PHYLUM ECHINODEEMATA. only in Crinoid and Echino-pluteus larvae, no Echinoderm larva has a right hydrocoel equal in size to the left, and in Crinoids and Holo- thurians there is no trace of a right hydrocoel at any stage of existence. Lastly, in no normal* Echinoderm does the right hydrocoel ever possess a water-pore. We do not wish to be unduly critical, but we think it not unreasonable to point out that, in the absence of any test which enables us to decide which characters are ancestral and which secondarily acquired, an ancestor constructed by this somewhat one-sided application of the recapitulation theory can have very little value in advancing zoological knowledge. We do not say that this kind of speculation has no value, for it is a source of delight and stimulus to many minds ; but we think that it is most important that its value should not be overrated and that it should not be allowed to divert attention from more important and more practicable problems. To continue the imaginary history of the dipleurula ancestor. The next change is due to its fixation, which is supposed by Bather to have taken place by the right side of its preoral lobe, though the fixation actual occurs in the middle line. This led to the passage of the mouth to the left side and to the establishment of the radiate structure of most organs except the genital and to the shifting and asymmetry of the coelomic sacs. This brings us to the so-called Pentactaea.t Now came the diver- gence into types. The Holothurian type in which the generative organs never acquire a radial arrangement, was the first to separate. In this type the attachment was entirely lost from the whole life history. Next, after the acquisition of radial structure by the gonads, the Asteroids and Echinoids separated off ; of these the Echinoids entirely lost their attach- ment, while the Asteroids appear to have retained it in some if not in all casss (larval attachment, p. 149). Lastly, or perhaps as a continuation of the main stem but little modified, came the Crinoids, in which the attach- ment is retained. This type further presents the following remarkable feature which may or may not have been primitive ; the mouth shifts from the left side to the hind end, where it lies alongside the anus. Class ASTEROIDEA J Star-shaped or pentagonal forms with the body flattened in the or-anal axis. The arms are not sharply marked off from the disc and have an ambulacral groove from which the tube feet project. The madreporite is on the abactinal surface. * Brooks and Field have asserted that in bipinnaria a second madre- poric pore normally occurs, but this statement has not been confirmed. f See Lang's Comparative Anatomy, Pt. 2, p. 548. j E. W. MacBride, " The development of Asterina gibbosa," Q.J.M.S., 38, 1896, p. 339. H. Ludwig, " Asteroidea " in Bronris Thierreich, Leipzig, 18948. Id., Die Seesterne des Mittelmeeres, Neapel, 1897. S. Goto, Metamorphosis of Asterias pallida, etc., Journ. Coll. Sc. Japan, 10, 1898, p. 239. Id., Metamorphosis of Asterina gibbosa, ibid., 12, 1898, p. 227. W. P. Sladeri, " Report on the Asteroidea," Challenger Reports, vol. 30, 1889. E. Perrier, " Echinodermes," in Exped. Sci. du Travailleur et du Talisman, Paris, 1894, and in Mission Scient. du Cap Horn, vol. vi., Paris, 1891. See also works of Lxidwig, Cuenot, Hamann, Delage et Herouard, loc. cit. ASTBROIDEA. 167 The form of body varies from that of a pentagonal disc in which the rays are only marked by the angles (Pentagon- asteridae, Fig. 117, Pterasteridae, species of Culcita), to that of a star (Fig. 139), in which the disc is small and the arms as sharply marked off from it as in the Ophiurids (Brisingidae). The number of rays varies in living forms from five to forty-five (Labi- diaster). Five is the most usual number and is especially constant in the discoidal forms and in those with well-developed marginal plates (Palmipes rosaceus is exceptional in having eleven and Culcita tetragona in having four). Four is sometimes found as an individual variation. The number of rays shows a distinct tendency to increase in families, in which the arms are long, the disc- small, and the marginal plates feebly developed (e.g. Brisingidae, Heliasteridae, Asterii- dae, Echinasteridae). When there are more than six rays individual variations in their number are fairly common. In Labidiaster in which the arms are very numerous, the number of them increases with the growth of the animal ; but in most if not all other cases the full num- ber is laid down in the embryo. The body is usually compressed dorso-ventrally. On the actinal surface, reaching the whole length of the radii and terminating in the centre of the disc in the oral area or depression, are the ambulacral grooves, and from them project the two, more rarely four, rows of J Fia. 117. Pent'tgonaster Parkinson. tube-feet. The mouth is placed Forbes (after Perrier). Seen from be- low and trom the side. on the actinal surface in the centre of the disc in the oral depression. There are no circumoral tentacles or tentacular prolongations of any kind round the mouth, but at the distal end of each ambulacral groove there is a red pigment spot which is called the eye or ocellus and over which projects an unpaired tentacle-like structure ; this is the ocular tentacle and contains the end of the radial water- vascular trunk. The dermal skeleton is well developed, and carries, especially on the abactinal surface, numerous spines and usually pedicellariae. The anus, which is absent in the Astropectinidae and probably in the Porcellanasteridae, is on 168 PHYLUM ECHINODEKMATA. the abactinal surface of the disc in interradius I. II, very nearly, but not quite, at the central point (Fig. 83). The flattening of the body is carried furthest in Palmipes membranaceus, which has the form of a pentagonal sheet of card- board. While the oral surface is generally flat, the aboral sur- face is often arched, sometimes considerably so (Pteraster, Hymenaster, Marginaster, Pentaceros, Culcita, many Solasteridae FIQ. 118.Echinastersentus from the oral surface (after A. Agassiz). mouth. A f ambulacra! feet. and Asteriidae). The rays are usually approximately equal in size ; inequality generally implies recent mutilation and re- generation. Starfishes vary in size from those with an arm-radius (centre of disc to extremity of ray), of 1 cm. or less (Marginaster pentagonus, '3 cm.) to those in which the same dimension measures 45 centimetres (Luidia savignyi to 37 cm., Freyella remex to 45 cm.). ASTEROIDEA. 169 The body- wall consists (1) of a single layer of ciliated columnar ectoderm, with a cuticle on the outer side and a basement mem- brane on the inner ; (2) of a dermis formed of a gelatinous matrix containing fibrillar connective tissue and calcareous plates and muscular elements ; (3) of a layer of ciliated peritoneal epithelium which lines the body cavity. In the ectoderm are sense-cells and gland-cells, and in some parts of the body nerve- fibres and nerve-cells are present in the deeper parts of the same layer. The integument is more or less hard and stiff owing to the presence of the calcareous plates. These structures may be regularly arranged and in contact with each other, or they may be irregularly disposed rods forming a kind of loose network, through the variously shaped meshes of which such delicate processes of the body-spaces as the tube-feet and dermal branch- iae project, or they may be isolated from one another. The calcareous plates may be deeply imbedded in the dermis and not visible from the exterior, or they may lie just beneath the epidermis, so that their shape is more or less completely dis- cernible in surface view. As a general rule some or all of the dermal plates bear granules, or processes and spines of various shapes, or pedicellariae. When these structures project from the surface, as they generally do, the skin has a rough or even spiny appearance ; but sometimes, especially when the plates lie deep, they project but little in the fresh state, and the skin is nearly smooth ; though even here, in dried specimens, the skin is rough and the spines are discernible from the surface (e.g. Tylaster willei, Porania, Culcita). In addition to these skeletal structures of the general in- tegument, small calcareous pieces are found in the walls of the tube-feet. The skeleton of the plates or ossicles falls under two heads, (a) the auibulacral, (b) the ambital. (a) The ambulacra! skeleton consists of the ambulacral and adambulacral ossicles, together constituting four rows of plates in each ray (Fig. 110). They form the roof and sides of the ambulacral groove (Fig. 121). The ambulacral ossicles (Fig. 119, A) are two rows of rod-shaped structures which meet and are articulated together in the middle line above (Fig. 121), and diverge from one another on each side, abutting at their outer ends upon the adambulacral ossicles (A'). These, which correspond in number with the ambulacral ossicles, though they usually alternate with them in position (Fig. 119), form the edges of the ambulacral grooves and carry 170 PHYLUM ECHINODERMATA. the adambulacral spines (Fig. 121, 8). The ambulacral ossicles are always without any form of spinous appendage, and the tubes connecting the ampullae and tube-feet pass be- tween them in such a way that never more than one tube passes between two successive ossicles. When these apertures are ar- ranged in a straight line on each side, as usually happens, there are two rows of tube-feet in each ambulacrum (Fig. 119) (biseriate arrangement). When on the other hand the apertures are alternately near to and remote from the middle line of the ray (Fig. 120), there appear to be four rows of tube feet (quadrise- riate arrangement, Asteriidae, Heliasteridae). In certain fossil starfishes (Palaeasteroidea) the ambulacral ossicles of the two sides are not opposite one an- other, but alternate. In living forms (Euasteroidea) they are opposite. The ossicles of both series increase in number with the growth of the starfish in such a way that the oldest, i.e. first formed, ossicles are next the mouth, while the youngest are at the end of the arm. The young- MV IG. 119. Diagram showing the skeletal pieces of the arm of a starfish with adambulacral peristom and biserial tube-feet, when viewed from the actinal surface (from Ludwig). .41, A2, etc., first, second, etc., ambulacral ossicle ; A'l, A' 2, A' 3, first, second, third adambula- cral ossicle ; inner intermediate piece (oral plate); MV inframarginal plates; T terminal plate; VI, V2, VZ ventrolateral plates. est ambulacral ossicle lies on the ventral side of the terminal plate. The superambulacral pieces, which may be mentioned here, are internally placed and connect the ambulacral ossicles with the marginal plates (Fig. 121, 5) ; they are found in the Astropectinidae, many species of Linckiidae, and in some species of the Porcellanas- teridae and Archasteridae. The first two ambulacra] ossi- cles of each side are more or less fused with one another (they are represented separate in the diagram) and form with the first of the adambulacral series the dental apparatus of the peris- tom. When the adambulacral pieces of this system are more prominent than the ambulacral, and form the buccal angles, the peristom is called adambulacral (Fig. 119) ; when, on the other hand, the .ambulacral pieces are the more prominent and the buccal angles are less FIG. 120. Diagram showing skeletal pieces of the proximal part of the arm of a starfish with ambulacral peristom and quadriserial tube- feet (from Ludwig). Al, A2, etc., first, second, etc., ambulacral ossicle ; A 1, A'2, A'3, first, second, third adambulacral ossicle ; o inner intermediate piece (oral plate). ASTEROIDEA. 171 FIG. 121. Diagram of a transverse section through the arm of Astropecten, the hepatic diverticula being omitted. 7 ambulacral, 2 adambulacral ossicle ; 3 lower, 4 upper marginal plate ; 5 superambulaeral ossicle ; 6 integument of abactinal surface ; 7 paxil- li ; S adambulacral spines ; 9 spines of the lower side of the ini'ramarginal plate ; 10 lower, 11 upper marginal spine ; a radial water-vascular vessel ; 6 so-called radial blood-vessel ; c radial nerve ; d ampulla of tube- foot ; e tube-foot ; / perivisceral cavity of arm. developed, we get the ambulacral peristom (Fig. 120). Whatever may be the condition in the adult, all young starfishes begin by possessing an ambulacral peristom. The adambulacral pieces of the peristom carry opines. The terminal plates may be mentioned here. They belong to the actinal surface, and one of them is found at the end of each arm on the abactinal side of the last ambulacral ossicles (Fig. 119 T). They are especially conspicuous and important when the skeletal plates are first making their appearance ; and it can then be seen that they are laid down round the left coelomic ;sac, i.e. on the actinal surface of the future star- fish (p. 147). (b) The ambital skeletal. The rest of the skeletal plates are classified as ambital. They consist of the interambulacral plates and the antambulacral. The interambulacral plates are of three kinds : ( 1 ) the inner intermediate pieces (o, Fig. 119), of which there is one in each interradius ; they lie on the abactinal side of the two adambulacral pieces which form the buccal angles ; (2) the ventrolateral plates, which lie between the adambulacral and the inferior marginals (V, Fig. 119); the ventro-lateral plates are often only found on the interradial portions of the actinal surface of the disc ; (3) the inferior marginals, which constitute a row of plates placed at the edge of the arm, just external to the ventro-laterals (Fig. 119, MV). The antambulacral plates (Fig. 122) constitute the bulk of the ambital skeleton. They consist of (1) the superior marginals '(M), which may either be in contact with the inferior marginals or separated by intercalated plates; (2) the eleven primary plates of the apical system, viz. five radial (R), five interradial (JR), and one central (C) ; (3) the secondary radial plates of the arms and disc (Rb and Rd) (carinalia of Perrier, medioradials of Sladen) ; these are often not distinguishable in adults from the other antambulacral plates, e.g. many Archasteridae, Porcellanasteridae, Asteriidae, Solasteridae, Echinasteridae ; (4) the dorsolateral plates (dl), corresponding to the ventro-laterals of the interambulacral skeleton ; (5) the supplementary plates of the arms and disc (s) which may effect the connexion of the dorsolaterals into a meshwork (hence reticularia of Perrier) ; (6) the madreporic plate (Md) placed in the left anterior in- terradius (the mouth being downward, and the anal interradius forward) ; the madreporic plate may be either the interradial plate of the apical system (JR, MD), or it may be outside this plate (Md), or it may be fused with it (Md 1 ). The primary plates of the apical system are quite distinct in the young starfish, but in later growth are only rarely distinguishable by size or form .(some Asterinidae and Pentacerotidae, many Pentagonasteridae) from 172 PHYLUM ECHINODERMATA. V ra the other plates of the antambulacral surface. Sometimes they remain throughout life as the sole plates of the antambulacral system (Cnemidaster wyvillei], and sometimes only a few other plates are developed between them (Neomorphaster talismani, Korethraster setosiis, different species of Marginaster}. The dorsolaterals may be connected into an irregular network, and fit closely into one another, and when the marginals are not dis- tinctly developed they pass gradually into the ventrolaterals at the sides of the arms. The external skeletal structures are spines, spinelets, scales, gran- ules and pedicellariae. Like the plates they are dermal structures and are covered by a layer of ectoderm. They are processes of or movably - attached to the subjacent skele- tal plates. The spines are elon- gated pointed rods, which project, usually singly, from the sub- jacent plates, to which they are attached by muscles and often by a ligament. They are found mainly on the larger plates of the abactinal surface and on the marginal plates. The spinelets are simply small spines. They often occur in tuft-like FIG. 122. Diagram of a starfish viewed from the abactinal side to show the antambulacral skeleton (after Ludwig). The arms are numbered i-v (see Fig. 83). A anus ; C central plate ; dl dorso-lateral plates of arms ; ill, I ditto of disc ; JR interradial plate of apical system ; JRA inter- radius in which the anus is placed ; JRM iuterradius con- taining the madreporite ; JR, Md madreporite and inter- radial as one and the same plate ; M marginal ; Mil madreporite separate from the interradial : Md' madre- porite as fused with the interradial ; R radial ; Rbl. rtr. secondary radials of arms ; Rdl, etc. secondary radials of disc ; s supplementary plates ; T terminal plate. aggregations. Some- times the plates carry- ing such tufts project in a column-like manner : such projecting columns with their tufts of spines are called paxilli * (Fig. 121). The spinelets are found principally upon the adambulacral plates and upon the actinal and abactinal intermediate plates ; they are attached to the subjacent * For a somewhat similar arrangement of spines in the Solasteridae and Pterasteridae, see p. 191. ASTEROIDEA. 173 plates by muscles and are covered with a glandular and nervous epider- mis. The scales are flattened spinelets. The granules are short rounded spinelets ; they occur set closely together on the skeletal plates. The pedicellariae are pincer-like calcareous structures consisting of two, rarely three, blades articulated to a plate in the clermis, and capable of executing snapping movements by means of a special set of muscles at- tached to their base. They are contained in the dermis and are covered FIG. 123. Pedicellariae of Asteroids (from Lang). A-F grouped spines resembling pedice lariae. G sessile pedicellaria of Gymnasteria carinifera (after Cuenot). H stalked straight pedicellaria, diagrammatic (after Cuenot). J basal piece of a stalked crossed pedicellaria of Asterias rubens (after Perrier). K stalked crossed pedicellaria of Asterits glacialis (after Cuenot). 1 calcareous blade of forceps ; 2 basal piec3 ; 3 occlusor muscle ; 3 axial muscle of the blades attached to the basilar piece and acting as an occlusor muscle ; 4 opening muscle ; 5 axial band ; in K this band divides, each branch being inserted into the base of one of the blades in such a way as to strengthen the grip when the pedicellaria is pulled ; 6 ectoderm ; 7 body wall ; S stem. by ectoderm. They may attain a size of 4 mm., but are usually much smaller. They are modified spines, and sometimes small groups of spines ar,e so associated that they can be moved towards one another and act like pedicellariae (Fig. 123, A-F). Pedicellariae are entirely absent in the genera Solaster, Echinaster, Cribrella, Mithrodia, Ophidiaster, Scytaster, Astropecten. Pedicellariae are of two kinds, sessile and stalked. Sessile pedicellariae (Fig. 123, (?) arise direct from the integument, to one of the plates of which 174 PHYLUM ECHENODEKMATA. they are attached. Stalked pedieellariae are at the end of a short soft stalk, the blades articulating with a calcareous piece at the end of the stalk (Fig. 123, H and K, 2). Stalked pedieellariae are either straight (forficiform) or crossed (forcipiform) (Fig. 123, H and K). In the forcipi- form pedieellariae the blades cross at their lower ends. Three-bladed pedieellariae constructed after the fashion of the straight variety are occasionally found. The blades of pedicellariae may be longer than they are broad, in which case they ar& said to be forcipate ; or they may be broader than they are long, in which case they are valvulate (Fig. 124). Valvulate pedicellariae are always sessile.. ^^^ Alveolate pedicellariae are sessile pedicellariae in which the 'blades are inserted into a depression or Fm. 124. Valvulate ,. , pedicellariae of alveolus in the calcareous plate. otalked pedicel- Culcita grex (after l ar i ae are confined to the Brisingidae, Pedicellas- IVrner). a one of .. . ' ... the five pedicel- teridae, Heliasteridae, Astemdae, Zoroastericlae, lariae shown in Stichasteridae, which are united together by Perrier the figure. as Forcipulata. Spines are movable : those at the edges of the ambulacral grooves, can be bent over the groove. Pedicellariae are no doubt related to such movable spines. However this may be, it is highly probable that they are defensive weapons, seizing and damaging other organisms which come in contact with them, and that they also serve for keeping the skin clear of foreign growths (see the account of pedicellariae under Echinoidea). The starfishes have some power of bending their arms and the tips are generally turned upwards so as to expose the eye to the light. Muscles of the body wall. Muscular fibres are found in the inner parts of the dermis near the peritoneum on the dorsal and lateral parts of the arm and disc. They are arranged in an inner longitudinal and outer circular layer. Some of the fibres of the circular layer pass outwards to be inserted into the cal- careous plates. The longitudinal layer is specially thickened on the abactinal side of the arms and disc to form a kind of longitudinal muscular band. These bands unite at a central point in the disc. Skeletal muscles. Special muscles passing between various, parts of the skeletal system are present. The most important are : (1) Muscles of the spines and pedicellariae, (2) muscles which pass between the two ambulacral ossicles of a pair, (3) muscles which connect the successive ambulacral ossicles. All three parts of the central nervous system are present, viz. the superficial, the apical and the deep oral (Fig. 125 and p. 123). The ectoneural tracts consist of a circumoral ring and of radial ASTEROIDEA. 175 nerve trunks, the whole lying in the deeper parts of the ecto- derm. It is in connexion with a diffuse ectoneural plexus found throughout the ectoderm, and at the mouth with an endoneural plexus. The apical or mesoneural nervous system and the deep oral system have already been sufficiently described (p. 123). a FIG. 125. Scheme of the nervous system of the arm of a starfish (after Cuenot). a wall, b body-cavity of arm ; c ampulla of tube-foot ; d tube-foot ; e radial canal of water- vascular system ; 1 radial portion of ectoneural central nervous system ; 2 ectoneural plexus of tube-foot ; 3 ectoneural plexus of skin; 4 Lange's nerve cords (deep oral) ; 5 mesoneural plexus just beneath the longitudinal muscle. The structure of the ectoneural plexus and its relation to the ectoderm are best seen by examining the annular (circumoral) nerve ring or the radial prolongations of it. The ectoderm here is much thickened and consists of elongated columnar cells with their nuclei near their outer ends ; the inner ends of the majority of the cells, which may be called supporting cells, taper and form a supporting tissue for a plexus of fine nerve fibres and scattered ganglion cells, which are here especially conspicuous and cause the thickening of the ectoderm. Some of the ectoderm cells are sense cells, and their inner ends do not form supporting fibres but branch out and join the nervous plexus. There are special aggregations of these round the ends of the tube-feet. The ectoderm along these central concentrations of the ectoneural sys- tem is much thickened, and its cells contain yellow pigment grains which give the whole tract in both its annular and radial por- FlG i26.Astroi>ecten aurantia- tions the appearance of a yellow streak. At the ends of the arms, this thickened tract of nervous ectoderm is continued over the terminal tentacle-like process (ocular tentacle), which forms the pro jecting end of the radial water-vascular trunk. The eye is on the oral side of the base of this tentacle ; it is coloured bright red and formed by a special thickening of the radial nerve tract Oc ctis, end of ray with the eye Oc surrounded by spicules (from Claus, after Haeckel). J76 PHYLUM ECHLNODERMATA. and subjacent connective tissue. In a typical case* each eye consists of a number (50 to more than 100) of funnel-shaped ectodermal pits, the cells of which contain the red pigment and end in a clear highly-refractile rod which projects into the cavity of the pit. The pit is closed towards the exterior by the cuticle, which may have on its inner side a lens-like thickening. The pigment cells are continued internally as fine fibres which join the nervous tissue of the radial nerve. The pits increase in number with the growth of the animal and they appear to contain a transparent gelatinous tissue. In some cases there are no pits and the pigment cells are distributed uniformly over the surface of the ocular cushion (Astro- pecten pentacanthus, miilleri}. The alimentary canal begins with a mouth which is placed in the centre of the actinal surface of the disc in the buccal membrane, and leads into a short oesophagus or directly into the spacious stomach (Figs. 127, 132). The oesophagus passes quite gradually into the stomach from the abactinal part of Nd FIG. 127. Longitudinal section through the disc and an arm of Solaster endeca (from Clans, modified after G. 0. Sars). mouth leading directly into the wide stomach ; A anus ; L hepatic diverticulum of the stomach ; G gonad ; Aid madreporite ; Js rectal caecum ; Aj tube-foot. which two caecal diverticula (hepatic) are given off into each arm. From the stomach a short rectum, which receives the rectal caeca, leads to the anus, which opens on the abactinal surface in interradius /. // (Figs. 83, 122). The buccal membrane is the part of the body-wall round the mouth in the oral depression and is devoid of calcareous structures ; the circular muscular fibres in it act as a constrictor and the longitudinal as dilators of the mouth opening. The oesophagus passes without any line of de- marcation into the stomach ; it is beset with ten glandular diverticula in Echinaster and Cribrella. The stomach is a spacious, thin-walled sac, divided by a horizontal fold (absent in the Astro jii'i-tiindae without an anus) into an oral (cardiac) and aboral portion (pyluric- s K ). Its walls are often considerably folded, so that it appears lobed. From its aboral portion the stomach gives off the hepatic or pyloric caeca (Fig. 128), one pair into each arm. These are tubular structures beset with numerous * Pfeffer, Zool. Jahrb. Anat., 14, 1901, p. 523. ASTEROIDEA. 177 secondary and tertiary glandular diverticula, and suspended to the abactinal body-wall of the arm by two mesenteries (Fig. 129, 24). The two caeca of a pair usually arise separately from the pyloric sac, but in some cases (Asterias, etc.) they are united near the stomach and arise from it by a common tube. The proximal end of each of these caeca gives off on its actinal side a pouch, called Tiedemanris pouch (specially developed in the Echinasteridae and Asterinidae). The abactinal side of the stomach is closely applied to the body wall, but between the two is a variable number of inter- radially placed glandular diverticula : these are the rectal caeca (Fig. 127, Js). They vary considerably in form, size and number (from two to five, in Luidia they are absent), and they open into the rectum except in the Astro pectinidae, in which the rectum and anus are absent and the caeca open directly into the stomach. The rectum is a short tube which leads from the stomach to the anus. The anus (p. 167) is absent in the Astropectinidae and Porcel- lanasteridae. The alimen- tary canal is lined throughout by a glandular ciliated epithelium, and, in some forms at least, a few calcareous bodies are found (Culcita, Ophidi- aster chinensis) both in its walls and in the mesenteries which attach it to the body-wall. The mesenterial attachments of the alimentary canal are as follows : ( 1 ) a pair of muscular strands run from the oral wall of the stomach in each radius to be inserted along the ambulacral ossicles of the ray (Fig. 132, ret) ; they serve to retract the stomach when, it has been evaginated for the prehension of food ; (2) a number of fine cords passing from the aboral wall of the stomach to the body-wall (Fig. 132, mes) ; these are remains of the septum between the right and left posterior coeloms of the larva ; (3) the mesenteries of the hepatic caeca (p. 177 and Fig. 129, 24) ; these fuse with one another distally, but proximally end freely so that the space between them is in open communication with the general body-cavity. They are probably remains of the mesentery which separated the right and left posterior body-cavities of the larva. On this view the spaces between the two mesenteries of a pair are parts of the right posterior body-cavity, and, if the division between the two body-cavities were retained in the adult, the two mesenteries of one caecum should join respectively the mesenteries of the caeca on each side of it in the disc, so as to cut off just above the stomach a circular patch of body-cavity with two prolongations into each arm. (4) Two bands which pass from the aboral end of the wall of the axial sinus to the stomach. These were formerly supposed to be processes of the aboral sinus. Z III N FlO. 128. Asterina gibbosa with the ahactinal body- wall removed (from Claus). Ld hedatic caecum G gonad. 178 PHYLUM ECHINODERMATA. The coelom. The relations and development of the coelom have been more fully elucidated in Asteroids than in any other class of Echinodermata. It represents the following parts : * (1) the general or perivisceral body-cavity, (2) the axial sinus and inner circumoral perihaemal ring, (3) the so-called outer circumoral perihaemal ring with its radial prolongations between IS 27 FIG. 129. Diagrammatic view of a transverse section through the arm of an Asteroid (fro 11 Lang). 1 trunks of the deep oral nervous system (Lange's nerves) ; 2 radial water-vascular trunk ; 3 tissue of so-called radial blood-vessel ; 4 radial nerve of the superficial oral system ; 5 radial perihaemal canal ; G and 7 branches of the same to the walls of the tube- feet ; 8 stalked pedicellaria ; 9 spine ; 10 genital aperture ; 11 papula with contained body cavity 13 ; 12 sessile pedicellaria ; 14 hepatic caecum; 15 peribranchial cavity ; 16 supra- marginal; 17 inframarginal ; IS adamlmlacral plate; 19 marginal canal communicating at 20 with the perivisceral coelom ; 21 peritoneal epithelium ; 22 genital sinus ; 23 gonad ; 24 mesenteries of hepatic diverticula ; 25 tube connecting tube-foot with 26 ampulla ; 27 cavity of tube-foot ; 28 upper and lower transverse muscles connecting the ambulacral plates; 29 branches of the deep oral nerve trunks ; 30 ambulacral plates; 31 perivisceral cavity ; 32 longitudinal muscle ; 33 apical nervous system. the ambulacral nerves and the radial water-vascular trunks, (4) the aboral sinus surrounding the generative rachis and the sinuses in the walls of the gonads, (5) the water- vascular system, and (6) the gonads. The general body-cavity which is derived from the right and * For the general relations and development of these structures the reader is referred to p. 126 et seq. and p. 142 et seq. ASTEROIDEA. 179 - c 4 left posterior body-cavities of the larva is a perivisceral space in relation with the stomach, rectum and pyloric caeca : it is a continuous space, is found in the disc and in the arms on the abactinal side of the ambulacral ossicles and is traversed by certain mesenteries which have already (p. 177) been described It is lined by a ciliated epithelium, and contains an albuminous fluid with floating amoeboid cells. The papulae or dermal branchiae (Fig. 129, 13) are thin projections of the body- wall found principally on the abactinal surface of the disc and arms and con- taining prolongations of the perivisceral cavity. Round the bases of the papulae there is generally present in the body- wall a space lined by an epithelium and supposed to be developed as a diverticulum of the perivisceral cavity. These St.... a FIG. 130. Diagrammatic transverse section through the axial sinus of a starfish (from Lud wig), a, b the lamellae of the inter- brachial septum ; C axial organ ; St stone-canal ; H axial sinus. The upper side of the section is the body-wall side. Flo. 131. Diagram showing the arrangement of the pcrihaemal spaces, etc., of a starfish, seen from the aboral side (after MacBride). a' axial sinus and inner perihaemal ring ; ab aboral circular sinus; yen.r generative rachis ; ov.g axial organ; ph. I. II . . . ph.IV.V the five parts of the so-called outer perihaemal ring; rhy vestige of right hydrocoel; lY the arms numbered. 180 PHYLUM ECHINODERMATA. spaces are called the peribranchial cavities (Fig. 129, 15, Fig. 132, pbr). The axial sinus is a tube passing from the abactinal to the actinal surface in the interradius of the madreporite (Fig. 132, ax.s). It is contained in an irrterbrachial septum (Fig. 130) and is lined by a ciliated epithelium. The stone-canal and axial organ project into it, being attached to its inner wall, i.e. to its wall next the oro-anal axis, by membranes (st.c). Orally the me* Mils st.c my tf&i.r y f "' h fl ft mv < -;'" ' ,jf> br ip.c. FIG. 132. Longitudinal vertical section through the disc and one arm of a young starfish (after an original drawing by Prof. E. W. MacBride, F.R.S.). The section passes through the interradius of the stone-canal and madreporite. ab aboral sinus ; ax.s axial sinus ; az.t terminal unpaired tentacle ; br papula (dermal gill) ; ext.p.r external perihaemal ring- canal ; gen.r genital rachis ; gen.st. genital stolon (prolongation of rachis into axial organ) ; int.p.r internal perihaemal ring-canal; l.p.c left posterior coelom of larva, hypogastric coelom of adult ; mes remains of the mesentery separating the left posterior (hypogastric) and right posterior (epigastric) coeloms ; mp madreporic pores ; mv madreporic vesicle (right hydrocoel) ; nr nerve ring ; oc eye ; pbr peribrauchial space ; per.rad.c radial peri- haemal canal ; p.or.c perioral coelom, an outgrowth of the hypogastric coelom, from its walls are formed the retractors of the stomach ; pyl pyloric sac ; pyl' the opening of a pyloric caecum ; rail. n radial nerve ; reel rectum ; ret retractor of the stomach ; r.p.c right posterior coelom (epigastric of the adult) ; st.c stone-canal ; stow, stomach ; tf tube-foot ; w.v.rad.c water-vascular radial canal ; w.v. ring water-vascular ring canal. axial sinus is continuous with the so-called inner perihaemal space which surrounds the mouth (Fig. 131, a l , Fig. 132, int.pr). At its aboral end it communicates with the stone-canal on the one hand and with the exterior through some of the pores of the madreporite on the other (Fig. 132, mp.). The axial sinus is derived from the anterior body-cavity of the larva. The outer circumoral perihaemal ring, which is separated by a slightly oblique septum from the inner perihaemal ring just described, is not really a ring at all. It is (Fig. 131, ph. I. II. etc.) ASTEROIDEA. 181 made up of as many interradial pieces as there are arms, each piece being prolonged into two adjacent arms. It thus comes about that each arm has a prolongation of two pieces of the sa-called outer perihaemal ring, the septum between these two radial prolongations being the septum (Fig. 129, 3) which divides the radial perihaemal canal into two parts and in which the so-called radial blood-vessel runs. This outer circumoral peri- haemal ring is derived for the most part from the left posterior body-cavity of the larva (see p. 145). It is lined by a flat epithelium, whether ciliated or not is unknown. There is said to be a canal in the body-wall at the edge of the ambulacra! groove which communicates at intervals with the radial perihaemal space and with the perivisceral cavity (Fig. 1'29, 19). The aboral sinus (Figs. 131and 132a&) is a circular or penta- gonal sinus placed on the aboral side of the stomach between it and the skin, and giving off in each interradius two prolonga- tions * to the generative organs (gen.r). In this sinus and its prolongations lies a peculiar cord of tissue, consisting partly of generative rachis and partly of vascular tissue. The space, formerly supposed to be a blood-vessel, in the wall of the gonads, is de- veloped as a part of this system, but in the adult there is a septum which shuts it off from the rest. The whole of the aboral sinus with its prolongation to the gonads and the sinus in the walls of the latter is developed as a part of the left pos- terior body cavity of the larva. The water-vascular system is lined throughout by a ciliated epithelium, has thin muscular walls, and contains a colourless albuminous fluid in which float amoeboid cells. It consists of a circumoral vessel, which is placed on the inner side of the buccal membrane close to the calcareous pieces of the peristom, sends prolongations the radial vessels into the radii, and communicates with the exterior through the stone-canal and axial sinus. The stone-canal passes aboralwards in interradius //. /// (the so-called left anterior). At its aboral end it opens into the axial sinus on the one hand and to the exterior through the pores of the madreporite on the other. The madreporite is * The two cords which pass from the point where the axial sinus abuts upon the aboral sinus to the stomach are not processes of the generative rachis and aboral sinus as was formerly supposed, but of the wall of the axial sinus (see p. 177). 182 PHYLUM ECHINODERMATA. a calcareous sieve-like plate perforated by many pores the secondary water-pores. Though in the adult most of the pores of the madreporite lead directly into the stone-canal, some of them open into the axial sinus. In the larva the primary water-pore opens into the axial sinus. With later growth this one pore becomes divided by folding into many, the epithelium of some of which becomes directly continuous with the epithelium of the stone -canal, but some of the pores retain their direct connexion with the axial sinus. The circumoral vessel bears two kinds of appendages the polian vesicles (p. 184) and Tiedemann's bodies (p. 185). The radial canals, which lie between the heads of the ambu- lacral ossicles in the angle formed by their apposition (Fig. 129), give off on each side as many lateral branches as there are tube- fcet. Each of these branches goes to a tube-foot and is con- nected by a canal which passes between two adjacent ossicles, with the internally placed am- pulla. In some forms there are two ampullae to each tube-foot (many Astropectinidae). The radial canals end blindly at the end of the arm in the ocular tentacle. The tube-feet are always pointed in the young, but in the adult they often terminate in suctorial disc-like expansions. Sometimes the distal feet of an arm are pointed and the proximal suctorial. The pointed feet are tactile, while the others are adhesive as well. The tube-feet have a well-developed ectoneural nerve layer, and longitudinal muscles only, except in the sucking disc where there are radial fibres, the contraction of which brings about the adhesion of the disc. At the junction of the lateral branches with the radial trunk there is a valve Flo. 133. Diagrammatic representation of the water-vascular system of a starfish (from Clans). Re circumoral vessel ; Ap polian vesicle ; Stc stone- canal ; M madreporite ; P tube-feet ; Ap' ampullae of the same. ASTEROIDEA. 183 which prevents the fluid passing back into the radial canal, so that when the ampulla contracts, it drives its contents into and so extends the tube-foot. The retraction of the foot is caused by the contraction of the longitudinal muscles of its walls, the fluid being driven back into the ampulla. If the tube-foot has a sucker, it is able to be attached to external objects and then by its contraction to draw the body of the starfish towards the object. Of course movement of the body in this way can only be effected when a number of feet are acting together in a co- ordinated manner. If a starfish be removed violently from its substratum, the attached sucker will be broken off and left on the stone and water will be ejected from the lacerated ends of the contracting tube-feet. FIG. 134. Diagrams of transverse sections through the stone-canals of various Asteroids (after Lang). 1 membrane by which the stone-canal is attached to the wall of the axial sinus ; 2 epithelium of the axial sinus ; 3 epithelium of stone-canal ; 4 connective tissue of wall of stone -canal. The stone-canal projects into the axial sinus (p. 180). On one side of it, viz. on the side opposite that by which it is attached to the wall of the axial sinus (Fig. 130), there is a longitudinally disposed fold of its lining membrane. In the simplest cases this fold projects into the canal as a ridge (Echinaster purpureus, Brisinga coronata (Fig. 134, .4). In other forms (Asterina gibbosa Penn, Cribrella, oculata Linck, etc.) the free edge of the fold splits into two lamellae (Fig. 134, B), which in yet other species (of Asterias, Pentaceros, Gymnasteria, etc. ) become coiled (C). A further complication is introduced by the fusion of the ridge with the opposite wall of the canal and the formation, from each surface of the septum so constituted, of a spirally coiled lamella (species of Astro- pecten, etc., D). Finally there are forms (Astropecten aurantiacus, species of Luidia and Culcita)in which these septa are present in great number and divide the whole lumen into many irregular chambers. All these ridges and lamellae vanish at the lower end of the canal, where it joins the circular vessel. The walls of the stone-canal and the lamellae, etc., which project into it, contain a strong deposit of calcareous matter. 184 PHYLUM ECHINODERMATA. The small sac beneath the madreporite (Fig. 132, mv), sometimes called the ampulla, is the small right hydrocoel (p. 144). The discovery of the communication between the upper end of the stone-canal and the axial sinus, we owe to the work of Perrier, Durham, and MacBride. The number of madreporites and stone-canals varies considerably. In the majority of cases there is only one of each, but there may be more, and this increase in their number is usually, though not always, associated either with an increase in the number of arms, whether such appears as an individual variation or as a constant specific character, or with the power of asexual reproduction by fission across the disc which some starfishes possess. Though the increase in the number of madreporites is generally associated with a greater number of arms than five, it is by no means always found in such forms. For instance, in the genera Heliaster and Labidiaster which normally have a large number of arms there is only a single madreporite. The increase in the number of madreporites is found most frequently in the families Asteriidae, Stichasteridae, Echinasteridae, and Linckiidae, and the following table* in which the abbreviation M is used for madre- porite shows some of the most conspicuous instances of it : Asterias calamaria Gray, with 6-12, usually 7 arms, has in 7-armed forms usually 1, rarely 2 M ; but in a 12-armed specimen 2 M and in an 11-armed one, 4 M have been observed. Asterias tenuispina Lam., has in 5- to 9-armed forms 1 to 3, rarely 4 M. Asterias capensis Perr., has 6 (rarely 5) arms and 3 M. Asterias rubens L., exceptionally 2 M in 5-armed forms. Stichaster polyplax M.Tr., has 7 arms and 1 to 5, usually 3 M. Stichaster albulus Stimps., has 5 to 7 arms and 1 to 2 M. Acanthaster echinites Ellis and Solander, has in 13- to 20-armed speci- mens 5 to 16 M. Acanthaster ellisii Gray, has with 11 to 19 arms 5 to 15 M. Echinaster eridanella M.Tr. and E. purpureus Gray, have in 5-armed individuals 1 M, in 6- or 7-armed 2 M. Ophidiaster germani Perr., has in 5-armed forms 2 M. Linckia multifora Lam., has often 2 M in 5-armed specimens. Linckia pacifica Gray, var. diplax M. Tr., and L. guildingii Gray, have as a rule 2 M in 4- to 7-armed specimens. In the above instances the madreporites are in different, either conti- guous or remote, interradii, but cases are known in which there are two madreporites and canals in the same interradius as an individual varia- tion (Linckia multifora Lam., Heliaster multiradiata Gray, and in a 6-armed example of Asterias glacialis O.F.M.), and Giard has described a specimen of Asterias rubens with one madreporite and two stone-canals passing off from it. Finally it must be mentioned that sometimes the madreporite is divided into several pieces, all however connected with the same stone-canal. Tiedemann's bodies are small yellowish glandular bodies attached to the inner wall of the circumoral vessel into which they open (Figs. 135, 136). They consist of branching tubes of * Taken from Ludwig's excellent account in Bronn's Thierreich to which the reader is referred for the facts and literature relating to the variation in the number of madreporites. ASTEROIDEA. 185 .T FIG. 135. Vertical section through an interradial region of the peri- stom of Asterias rubens L. (after Ludwig), showing the connexion of one of Tiedemann's bodies with the water-vascular ring. B tissue of the so-called vascular ring ; E outer perihaemal ring ; J inner perihaemal ring ; Mi interradial muscle of peristom ; M h huccal membrane ; N circ'umoral nerve- ring ; T Tiedemann's body ; 1!' water- vascular ring ; Zoral portion of deep oral nervous system (Lange's nerve). cubical ciliated yellow epithelium. There are usually two in each inter- radius, except that of the madre- porite, which has only one, and they are supposed to be of the nature of lymphatic organs and to bud off cells into the water-vascular sys- tem. Like so many other features of Asteroid anatomy the polian vesicles vary considerably both in number and arrangement throughout the class, even in closely allied forms. They are large vesicular structures with muscular walls and long stalks which open interradially into the circumoral vessel (Fig. 136). Some- times (Asterias rubens and glacialis, etc.) they are absent altogether ; sometimes there is one in each interradius, except that of the madreporite, in which there are none at all or two (species of Astropecten) ; finally there are cases of two, three or four or even more in each i n t e r r a dius ; in the latter case it is common to find several vesicles opening by one stalk. It is sup- posed by some that the polian vesicles are con- tractile structures acting as central organs to vary the pressure in the 10 FIG. 136. Circumoral water-vascular vessel with appendages of Asterina gibbosa (after Cuenot, from Lang), seen from the aboral side. 1 mouth in the centre of the buccal membrane ; 2 stone- canal ; 3 axial sinus ; 4 transverse muscles of the ambulacral ossicles ; 5 ambulacral ossicles ; 6 poliau vesicles ; 7 Tiede- A , r l ir .1 c n T TQ( C , 7a f m fe j bl [11 > mann's bodies ; 8 circumoral vessel ; 9 ring of supposed vascular fi, c fV, f fl,,, tissue; 10 ampullae of tube-feet. Otlieib tnat tliey 186 PHYLUM ECHINODERMATA. are of the nature of lymphatic glands, and that their lining membrane which has the form of a connective tissue network containing dividing cells in its meshes buds off amoeboid cells which enter the fluid of the system. The axial organ (ovoid gland, heart) is a fold of the wall of the axial sinus, into which it projects. Its walls are secondarily folded so that it appears in section to contain prolongations of the axial sinus. It consists mainly of connective tissue and of cells derived from a prolongation of the generative rudiment (pp. 131, 146). The so-called vascular system (lacunar system) of Asteroids consists of tracts of connective tissue in which the fibres are sparser, the ground substance stains more deeply and the leucocytes are more numerous than in the ordinary connective tissue. There is a cord of it in the vertical septum dividing the radial perihaemal canal (Fig. 129, 3) : this is the so-called radial blood-vessel. This is continuous with a circumoral cord of the same tissue (Fig. 135, B). There is a certain amount of it in the wall of the stomach, which presumably is in connexion with the circumoral tract, and it is possible that the latter may send prolongations on to the axial organ. There is also a tract of it in close association with the generative rachis in the wall of the aboral sinus. The vascular tissue is sometimes described as consisting of bundles of anastomosing canals without any epithelial lining and containing a coagul- able fluid. A definite circulation of this fluid has never been observed. The sexes are separate. Asterina gibbosa however has been said * to be a protandrous hermaphrodite, the same gland pro- ducing in young specimens spermatazoa and in old ones ova. The generative organs consist of bunches of tubes which are attached on each side of the interbrachial septa to the abactinal body-wall (Fig. 127). There are therefore twice as many gonads as there are arms. Each gonad consists of one or of several tufts of tubes. In the latter case, the gonads extend into the arms, to the dorso-lateral wall of which they are attached. Each gonad opens to the exterior by one simple opening (rarely subdivided into several) in the abactinal wall of the disc, or, when there are several gonadial tufts, by as many pores as * Cuenot, L. Arch. Zool. Exp. et gen. 5 bis, 1888. ASTEROIDEA. 187 there are tufts, along the sides of the abactinal surface of the proximal parts of the arms. The gonad tubes are lined by a simple epithelium which gives rise to the generative cells, and the external openings are always on the dorsal surface except in Asterina gibbosa and Asterina pancerii Gasco, in which they open on the actinal surface, no doubt in correspondence with the fact that these animals attach their ova to foreign objects. The composition of the gonads of a single tuft or of several tufts is often a generic character, but sometimes both conditions are found in the same genus (e.g. Echinaster). The gonads are really the peripheral parts of the generative rachis. This structure consists of a cellular cord placed in the wall of the aboral sinus on the dorsal side of the stomach. It gives off a prolongation into the axial organ, and in each inter- radius two cords proceed from it to the gonads. These cords are accompanied by a prolongation of the aboral sinus, which reaches as far as the gonads and surrounds them (p. 181). Except in those forms in which there are arrangements for the care of the brood, external sexual differences are only occa- sionally present, and are then usually confined to differences in colour. A brood pouch is developed on the dorsal surface of the Pterasteridae (see p. 192). In some of the Astropectinidae the eggs pass into the interstices between the stalks of the paxilli and there undergo their development. In Stichaster nutrix Studer describes the young as being at first in outgrowths of the stomach where they undergo their early development, and then as passing on to the edges of the mouth. In Asterias perrieri Smith and other species the young are described as being attached to the oral surface of the disc in the neighbourhood of the mouth, and there undergoing their development ; the arms being slightly bent over them for protection. In species of Di/i/nnfrrins similar phenomena appear to occur. They are all marine and crawl upon the bottom of the sea. They capture their food by means of their tube-feet and many of them have the power of partially everting their stomach, the inner surface of which is applied to their prey. Many of them have the power of autotomously severing their arms from the disc and of regenerating arms so lost. The power of regenerating lost parts is great in all members of the 188 PHYLUM ECHINODERMATA. group, and in some cases it appears that a single arm can re- produce the whole animal. This appears to be the explanation of the so-called comet-forms, which consist of a large arm carry- ing a small disc with four small arms. Fission of the body through the centre of the disc into two parts sometimes occurs. The power of regeneration possessed by a wounded surface sometimes leads to the production of curious forms, e.g. in Linckia multifora the wounded surface of an arm has been described as forming a new disc with four arms. The development (p. 133) is rarely direct, and the young usually pass through the free-swimming larval stage called bipinnaria. Including brachiolaria larvae about twenty bipin- naria larvae are known. Most of these have not been related to their adults. The larva of Asterina gibbosa may be regarded as a much modified bipinnaria ; it has the power of swimming feebly with the cilia of its larval organ. Bipinnaria asterigera is the larva of Luidia sarsii ; it is the largest bipinnaria known (1-licm.). B. metschnikoffi and mulleri probably belong to species of Astropecten. B. russoi and buryi have been assigned to Asterias glacialis. The Asteroidea are found fossil from the Cambrian onwards, but the known fossil forms are not nearly so numerous as in the case of Oinoids and Echinoids. This is doubtless due to the fact that their tissues do not lend themselves so readily to pre- servation as do those of the above-named classes. The early forms do not differ essentially from those now living. The class is divided into two orders the Encrinasteriae, in which the ambulacral plates alternate on the two sides of the arm and the madreporite is on the lower surface ; and the Euasteriae, in which the ambulacral plates are opposite one another and the madreporite is on the dorsal surface. The Encrinasteriae, are exclusively Palaeozoic, while the Euasteriae include all the living forms and make their first appearance in the Silurian. Order I. ENCRINASTERIAE. With characters as above. Aspidosoma Goldf. lower Devonian ; Palaeaster Hall (Archasterias J. Mull), Silurian, Devonian and Carboniferous; Urasterella M'Coy (Stenas- ter Billings), lower Silurian ; Palasterina McCoy ; Palaeodiscus Salter ; Palaeocoma Salter, upper Silurian ; Salteraster, etc. ASTEROIDEA. ISO Order II. ETJASTEKIAE. With characters as above. Sub-Order 1. PHANEROZONIA. With large marginal plates. The supramarginal and inframarginal plates are in contact. Papulae restricted to the abambulacral surface within the area bounded by the supramarginal plates. Ambulacral plates usually broad. Tube-feet in two rows in each arm. Oral adambulacrals prominent. Pedicellariae when present sessile. Fam. 1. Archasteridae. Marginal plates thick, with spines or spini- form papillae. Adambulacral plates large and not compressed. Ventro- laterals and marginals with spines or paxilli. Superambulacral plates absent. Pararchaster Slad. ; Pontaster Slad. ; Cheiraster Studer ; Pecti- naster Perr. ; Lonchotaster, Dytaster, Plutonaster Slad. ; Archaster Miiller and Troschel ; Grnathaster Slad. ; Asterodon Perr. ; Odontaster Verrill ; Mimaster Slad. ; Goniopecten Perr. ; Leptogonaster Slad. ; Pseudar -chaster, Aphroditaster Slad. British species : Pontaster tenuispiins, Scilly, Faeroe Channel, etc. 90-60 fms. Plutonaster bifrons, Faeroe Channel, etc., 200-1,300 fms. ; PI. bifrons, N. of Ireland, 1,360 fms. Fam. 2. Poreellanasteridae. Marginal plates well developed, but thin and porcellanous in appearance, and apparently naked or covered only with a thin epidermal layer. Abactinal area covered with membrane and carrying in its centre an epiproctal prominence (Fig. 137). Anus said to be absent. Actinal surface of disc is covered interradially with squamiform plates. Cribriform organs (1 to 14 in each inter- radius) present. Adambulacral plates large with simple mar- ginal armature uniserially dis- posed. Excepting Ctenodiscus all genera are exclusively from the deep sea. Cribriform organs are situated on the marginal plates in the inter-brachial region of the disc, and extend when numerous on to the base of the arms. They consist of a number of parallel vertically arranged calcareous lamellae equal in length to the height of the two series of marginal plates. Porcellanaster W. Thorns. (Fig. 137) ; Styracaster, Hyphalaster, and Thoracaster Slad. ; Pseudaster Perr. ; Ctenodiscus Miill and Trosch. British species : Ctenodiscus cristatus, Faeroe Channel, 312 fms. Fam. 3. Astro pectinidae. With large marginals bearing spines or spiniform papillae. Actinal interradial areas small. Abactinal skeleton with paxilli (Fig. 121). Tube-feet conical. Superambulacral plates present. Anus absent. Pedicellariae rarely present. In Leptoptychaster kerguelenensis the eggs pass into the spaces between the groups of paxilli and there develop (W. Thomson, J. Lin. Soc. London, 1876, 13). Craspidaster Slad. ; Leptoptychaster Smith ; Moiraster Slad. ; Blaki- aster Perr. ; Astropecten Linck ; Psilaster Slad. ; Phoxaster Slad. ; Bathy- biaster Dan. and Kor. ; Ilyaster Dan. and Kor. ; Luidia Forbes ; Platas- terias Gray. British species : Leptoptychaster arcticus, Faeroe Channel, 1,312 fms. Astropecten irregularis (aurantiacus) Atl. and Med., 10-1,000 FIG. 137. Porcellanaster gracilis Sladen, side view, showing three cribriform organs and the epiproctal prominence. 190 PHYLUM ECHINODERMATA. fms. Luidia ciliaris, E. N. Atl., to 87 fms., L. sarsii, E. N. Atl., to 374 fms. Fam. 4. Pentagonasteridae. Interbrachial region of disc well de- veloped, so the body is pentagonal with more or less concave sides. Mar- ginals well developed. All the plates, both dorsal and ventral, form a close mosaic, and are granular or naked. Anus present, but often hidden by paxilli. Pentagonaster Linck. (Fig. 117) ; Stephanaster Ayres ; Astrogonium M. and T. ; Calliaster Gray ; Chitonaster Slad. ; Calliderma Gray ; Iconaster Slad. ; Gnathaster Slad. ; Nymphaster Slad. ; Paragonaster Slad. ; Mediaster Slad. ; Nectria Gray ; Stellaster Gray ; Ogmaster v. Martens ; Leptogonaster Slad. ; Goniodiscus M. and T. ; Mimaster Slad. ; FIG. 138. Porania glaber Sladcn, abaetinal view (after Sladen). Anihenoides Perr. ; Ho plaster Perr. ; British species : Pentagonaster greeni, Faeroe Channel, 440 fms. Mimaster tizardi, Faeroe Channel, 550 fms. Fam. 5. Antheneidae. With well developed marginals, which may bear granules or tubercles. Actinal interradial areas large and covered with pavement-like plates, which bear large valvate pedicellariae. Anus distinct. Anthenea Gray ; Goniaster L. Ag. ; Hippasteria Gray. Fam. 6. Pentacerotidae. Dorsal marginals smaller than the ventral, often more or less hidden. Actinal interradial areas with large pavement- like plates which bear unequal sized granules. With small valvate pedi- cellariae. Abactinal skeleton reticulate. Anus distinct. Pentaceros Linck, Nidorcllia Gray ; Amphiaster Verrill ; Pentaceropsis ASTEROIDEA. 191 Slad. ; Culcita L. Ag. ; Asterodiscua Gray ; Choriaster Liitken ; Paulia Gray. Fam. 7. Gymnasteriidae. Marginal plates large. The whole body is covered with a thick membrane. Arms usually short. Actinal inter- radial areas with large regular plates. Abactinal skeleton tesselate. Anus distinct. Asteropsis M. and T. ; Dermasterias Perr. ; Gymnasterias Gray ; Tylaster D. and K. ; Porania Gray (Fig. 138) ; Marginaster Perr. ; Rhegaster Slad. ; Poraniomorpha D. and K. ; Lasiaster Slad. British species : Porania pulvillus. E. N. AtL, to 106 fms. Fam. 8. Asterinidae. With small, sometimes inconspicuous marginal plates. Abactinal skeleton composed of imbricating plates notched on one side and bearing spines on the free margin. Actinal interradial areas with imbricating plates bearing spines. No pedicellariae. Cycethra J. Bell ; Ganeria Gray ; Patiria Gray ; Nepanthia Gray ; Asterina Nardo ; Disasterina Perr. ; Palmipes Linck ; Stegnaster Slad ; Tremaster Verr. British species : Asterina gibbosa, E. N. AtL, to 35 fms. Palmipes placenta, shores of Britain, etc., to 30 fms. Sub-Order 2. CRYPTOZONIA. Marginal plates inconspicuous. The supra- and infra-marginal plates are often separated by intermediate plates. Papulae not confined to the area bounded by the supra-marginals, but found also between the mar- ginals and on the ambulacral surface. Ambulacral plates crowded and narrow. Tube-feet often in four rows. Ambulacrals or adambulacrals of the oral skeleton prominent. Pedicellariae stalked or sessile. Fam. 1. Linckiidae. Marginal plates comparatively well developed, and in contact. Disc small, arms long. Abactinal skeleton tesselate. Pedicellariae (rarely present) excavate or foraminate. Chaetaster M. and T. ; Fromia Gray ; Ferdina Gray ; Ophidiaster Ag. ; Pharia Gray ; Leiaster Peters ; Linckia Gray ; Phataria Gray ; Nardoa Gray ; Narcissia Gray ; Metrodira Gray. Fam. 2. Zoroasteridae. Marginal plates in contact. Disc small ; arms long, cylindrical, and tapering. Integumentary skeleton spiny. Abactinal skeleton tesselate, arranged in regular longitudinal and trans- verse series. Primary apical plates persistent and distinct in the adult. Tube-feet conical, terminated by a small sucker ; they are arranged in four series at the base of the arm, in two series distally. Pedicellariae (forcipulate) stalked. Zoroaster W. Thorn., for the most part from great depths ; Cnemidaster Slad. ; Pholidaster Slad. ; Mammaster Perr. ; Caly- caster Perr. Fam. 3. Stichasteridae. Marginal plates in contact. Disc small ; arms long, cylindrical and tapering. Integumentary skeleton for the most part granular. Abactinal skeleton tesselate arranged in longitu- dinal rows. Primary apical plates less distinct. Tube-feet usually cylindrical and with large terminal sucker, arranged in four rows all along the arms. Pedicellariae forcipiform and forficiform. Coelasterias Stimpson ; Stichaster M. and T. ; Tarsaster Slad. ; Neornor- pluister Slad. , Tonia Gray ; Nanaster Perr. ; Granaster Perr. British species : Stichaster roseus, Brit, coast, to 200 fms. Fam. 4. Solasteridae. Abactinal skeleton reticulated, with plates carrying on a projecting tubercle a bundle of divergent spines. Actinal intermediate plates more or less developed. Anus distinct. No pedi- 192 PHYLUM ECHINODERMATA. cellariae. Crossaster M. and T. ; Solaster Forbes ; Rhipidaster Slad. ; Ctenaster Perr. ; Lophaster Verrill ; Korethraster W. Thorns. ; Peribolaster Slad. British species: Solaster papposus, E. N. Atl. to 640 fms., S.endeca, ditto to 150 fms. Fam. 5. Pterasteridae. With a reticulated dorsal skeleton bearing paxilliform groups of spines. These spines are united together by a membrane (supradorsal membrane) which forms a continuous canopy over the dorsal surface. The chamber so enclosed is said to be a brood-chamber, and opens to the exterior centrally by a valvular aperture, and by a number of small contractile pores in the supradorsal membrane, and at the side of the arms by apertures regularly recurring over each adambulacral plate and called the segmental apertures. The canopy is present in all specimens hitherto examined ; it is uncertain whether these specimens are females or hermaphrodites or whether the canopy is present in both sexes. The supradorsal membrane may in some forms be wholly or partially aborted. Actinolateral spines when present united by mem- brane so as to form a web on the actinal surface. Pedicellariae absent. Pteraster M. and T. ; Retaster Perr. ; Marsipaster Slad., from the deep sea; Calyptraster Slad.; Hymenaster W. Thorns., almost entirely an abyssal form ; Benihaster Slad., from the deep sea ; Myxaster Perr. ; Cryptaster Perr. The above are disco-pentagonal in form and have a supradorsal membrane with segmental apertures. Pythonaster Slad., stellate forms without supradorsal membrane, actino-lateral spines and segmental apertures, from the deep sea. Fam. 6. Echinasteridae. Dorsal skeleton formed of plates disposed in longitudinal and transverse series, or in an irregular network bearing spines. Spines moderate, pointed, naked or covered by a thin membrane containing calcareous granulations. Arms long. Pedicellariae present only in Acanthaster and Valvaster. Acanthaster Gervais, with numerous arms (more than 10) and several (5-16) madreporic plates ; Mithrodia Gray ; Cribrella Ag. (Henricia Gray) ; Perknaster Slad. ; Echinaster M. and T. (Fig. 118) ; Plectaster Slad. ; Valvaster Perr., with wide-meshed arrangement of the calcareous plates and large groups of papulae. British species : Cribrella sanguinolenta, E. N. Atl., to 1.350 fms. Fam. 7. Heliasteridae. Arms very numerous (more than 25) and short, disc large. Abactinal skeleton reticulate. Tube-feet in four rows. Double interbrachial septa. Heliaster Gray. Fam. 8. Pedicellasteridae. Disc small, not sharply marked off from the arms. Abactinal skeleton of the arms reticulated. Tube-feet in two rows. With numerous, large forcipiform pedicellariae. Genital organs open on the disc. Pedicellaster Sars ; Coronaster Perr., with numerous arms ; Lytaster Perr. ; Gastraster Perr. Fam. 9. Asteriidae. With reticulate abactinal skeleton, bearing isolated or grouped spines. Tube-feet in four rows. Pedicellariae for- ficiform or forcipiform. Members of this family which have more than five arms are called heteractinides : in such species the number may be constantly six, or the number may be subject to individual variation. Pycnopodia Stimpson ; Coscinasterias Verrill ; Polyasterias Perr. ; Stolasterias Slad. ; Leptasterias Verrill ; Asterias L. ; Diplasterias, Perr. ;. Smilasterias Slad. ; Sporasterias Perr ; Anasterias Perr. ; Hydrasterias Slad. ; Cosmasterias Slad. ; Podasterias Perr. ; Uniophora Gray ; Cal- vasterias Perr. British species : Asterias glacialis, E. N. Atl., to 66 fms., ASTEROIDEA. 193 A. rubens, ditto to 110 fms., A. muelleri, ditto, 53 to 433 fins., A. murrayi, W. coast Scotland and Ireland. Fam. 10. Brisingidae. Arms long, marked off from the disc (Fig. 139). Marginals absent or vestigial. Abactinal skeleton absent or present only on the ovarian regions. Tube-feet biserial. Genital organs opening on the sides of the arms. Somewhat ophiurid-like in appearance. For the most part in deep water. FIG. 139. Odinia elegans E. Per., abactinal view (after E. Perrier). Hymenodiscus Perr. ; Gymnobrisinga Studer ; Brisinga Asbjornsen ; Odinia Perr. ; Freyelki Perr. ; Colpaster Slad. ; Brisingaster de Loriol ; Labidiaster Ltitken. British species : Brisinga endecacnemos and coronata, E. N. Atl., 100 to 1,300 fms. The starfishes of the abyssal regions of the ocean belong to the families Brisingidae. Pedicellastridae, Zoroaster idae, Stichasteridae, Pterasteridae, Pentagonasteridae, Ar chaster idae, Porcellanasteridae. But these are not exclusively abyssal, having littoral representatives in various parts of the world. Z HI O 194 PHYLUM ECHINODERMATA. Class OPHIUROIDEA * Brachiate Echinoderms with the body flattened in the or-anal axis. The arms are sharply marked off from the disc and are without an ambulacral groove. The madreporite is on the oral surface. The Ophiurids are brachiate Echinoderms, and the arms which are very rarely more than five in number (e.g. Ophioglypha hexactis with 6 arms, Ophiocantha vivipara 6 or 7 arms, 0. anomala and nodosa 6 arms etc.) are sharply marked off from the disc. The ambulacral grooves are absent or so slightly marked as not to be noticeable, and in most forms the ambula- cral nerve tracts are separated from the ectoderm by calcareous plates. The generative organs usually open into special pockets developed on the oral side of the disc at the base of the arms and called the genital bursae. The alimentary canal is without an anus and is not prolonged into the arms. The water-pore (or pores) is on the oral surface of the disc, and interradial in posi- tion. There is an axial sinus, axial organ, and generative rachis. The tissue of the so-called vascular system is feebly developed, and the tube-feet are without ampullae. As stated above the arms are rarely more than five in number, but in the Cladophiurae they may be much branched (Fig. 146). The integument is not ciliated, and except in the Cladophiurae and perhaps some Streptophiuridae the ectoderm is not present as a layer distinct from the dermis. The dermis is without muscles, but in all Ophiuroids, except the Cladophiurae, is richly provided with calcareous plates. As a general rule these plates form a complete dermal armour which may or may not be * A. Ljungman, Ophiuridea viventia hue usque cognita, Stockholm, 1867. H. Ludwig, Trichaster elegans, Z.f.w.Z., 31, 1878. Id., Das Mimdskelet der Asterien u. Ophiuren, Ibid., 32, 1879. Id., Zur Ent. d. Ophiurenskeletes, Ibid., 36, 1882. Id. Ophiopteron elegans etc., Ibid., 47, 1888. Th. Lyman, Report on the Ophiuridea, Challenger Reports, 5, 1882. Id., Ophiuridae and Astrophytidae Illustrated Catalogue of the Museum of Comp. Anat. of Harvard College, I. Cambridge, Mass., 1865. L. Cuenot, Etudes anatomiques sur les Ophiures, Arch. Zool. Exper. et gen (2), 6, 1888, p. 33. E. W. MacBride, Development of the genital organs, etc., in Amphiura squamata, Q.J.M.S., 34, 1893. F. Jeffrey Bell, A Contribution to the Classification of Ophiuroids, etc., Proc. Zool. Soc., 1892, p. 175. J. W. Gregory, Classification of the palaeozoic Echinoderms (Ophiuroidea), Proc. Zool. Soc., 1896, p. 1028. Liitken et Mortensen, The Ophiuridae, Mem. Mus. Harvard College, 23, 1899. O. Hamann, Ophiuroidea, in Bronn's Thierreich, 1900-1901. OPHIUROIDEA. 195 covered by a soft integument, but in the Cladophiurae and some Streptophiuridae the skeletal plates of both the disc and the arms are much reduced, and the integument is soft and thick. The skeleton of the arms consists of a double row of internally- placed ossicles which are generally fused with each other in pairs and are comparable to the ambulacral ossicles of Asteroids . 140. Transverse section through the arm of an Ophiurid, diagrammatic (from Lang). 1 tube-foot (ambulacral tentacle) ; 2 its cavity ; 3 epineural circular canal ; 4 circular ganglion, both at the base of the tentacle ; 5 under arm plate ; 6 radial epineural canal ; 7 radial nerve trunk of the superficial oral system (ambulacral nerve) ; S radial blood- vessel ; 9 radial trunk of the deep oral nervous system (Lange's nerve) ; 10 radial peri- haemal canal ; 11 peripheral branch of the radial nerve ; 12 spine ; 13 lower intervertebral muscle ; li lateral plate ; 15 ambulacral (vertebral) ossicle ; 16 upper intervertebral muscle ; 17 brachial coelom ; 18 specially ciliated strip of peritoneum ; 19 upper arm plate ; 20 radial water-vascular trunk ; 21 lateral portions of the brachial coelom which are repeated in each brachial segment ; 22 branch of water-vascular trunk to tube-foot ; 23 ganglion at base of spine ; 24 motor nerve from deep oral nervous system. (Fig. 140, 15). The water- vascular trunk, perihaemal canal and nerve cords lie on the lower side of these ossicles and a prolongation of the perivisceral space on the dorsal (77). The outer or ambital skeleton of the arm is segmented. Each segment consists of four plates ; a median under arm-plate, a median upper arm-plate, and two lateral arm-plates (Fig. 140). These plates join one another and the corresponding plates of 196 PHYLUM ECHINODERMATA. the adjacent segments, except in those forms in which the skeletal plates are deficient (Cladophiurae, some Strep tophiuridae) and the arms have a soft integument which contains only small skeletal pieces. The lateral arm-plates, which are generally compared to the adambulacral plates of Asteroids, carry spines, the others do not. The tube-feet emerge through openings between the under and lateral plates, one pair in each arm- segment. At the edges of these apertures are small scales. The ambulacra! or vertebral ossicles arise as separate pieces, which generally fuse together in pairs. In certain deep sea forms (Ophiohelus) each of these has the form of a curved rod joined to its fellow at each end. The successive ossicles movably articulate with one another and are attached by muscles (Fig. 140, 13 and 16). The articular surfaces vary in form and may develop processes and pits analogous to the zygosphenes and zygantra of an ophidian vertebra ; but sometimes the articulating surfaces are simple and the arms have a greater power of movement (Astrophyton, etc.). The radial water-vascular trunk lies in a groove on the lower side of the ambulacral ossicles, and its branches in passing to the tube-feet have a curved course through these structures (Fig. 140). Spines are present on the lateral plates of the arms, and occasionally on the upper surface of the disc and on the lower surface between the arms. In Ophiopteron elegans some of the arm-spines of each arm-joint are united together by a thin transparent membrane, thus forming a series of lateral fins. Pedicellariae of the ordinary type are not present, but in some forms (Cladophiurae) movable hooks, generally articulated to a pedicle, are found on certain parts of the arm. The hooks which occur in pairs are not opposed but move parallel to one another. Such modified pedicellariae are found in Astrophyton, Ophio- thrix fragilis, Trichaster elegans. The absence of true pedi- cellariae in Ophiurids would appear to militate against the view- that pedicellariae play an indispensable part in keeping the skin of Echinoderms clear of foreign growths and debris. The lower skeletal pieces of the disc constitute the oral skeleton and are very complicated. They comprise the proximal ambu- lacral and adambulacral (lateral) brachial plates, and the inter- radial buccal shields, on one or all of which are the water pores (see below), and a number of accessory pieces which belong to the ambital skeleton. The integument on the lower side of the disc between the insertions of the arms is either soft and contains small isolated skeletal pieces or granules, or is provided with a layer of imbricating plates. The bursal OPHIUROIDEA. 197 apertures which in some species are double (Ophioderma) are placed on the lower side of the disc, one on either side of the insertion of each arm. The genital plate (Fig. 141, gp) is a skeletal piece placed on the radial side of each of these slits. In a view from the inner side (Fig. 141) the angles are seen to consist of four plates; the two oral-angle plates (am , + ad Y ) and the peristomial plates (am 1 ). The oral -angle plates meet at the torus angularis (ta) and diverge outwards and towards the radii. Each of them has on the side towards the buccal fissure two depressions for the oral tube-feet (Fig. 141, pteb) and carries on its lower edge a number of small spines ; of these some project into the buccal fissures and are called oral papillae (Fig. 142, 10), ptel FIG. 141. Oral skeleton of Ophiopyren longispinus Lym., from within (from Lang after Lyman). am ambulacral ossicle ; am i peristomial plates, supposed to be the ambulacral ossicle of the first brachial segment ; am 2 + adi oral-angle plates ; D teeth ; fb bursal aper- tures ; gp genital plate ; ibr interbrachial region ; pteb depressions for oral tube-feet ; sge bursal plate ; to torus angularis. while others arise nearer the angle and project towards the centre of the mouth and are called dental papillae. The dental papillae are above the teeth of the torus angularis. The oral-angle plates are supposed to con- sist of the adambulacral (lateral) plate of the first brachial segment (next the angle) and of the ambulacral ossicle of the second brachial segment (outer part of oral-angle plate). The peristomial plates (am^) are on the inner side of the oral-angle plates and are supposed to represent the ambulacral ossicles of the first brachial segment. At the outer end of each buccal fissure is a plate (Fig. 142, 8) which is supposed to be the under plate of the second brachial segment) ; dorsally to this there may be sometimes made out another piece which is supposed 198 PHYLUM ECHINODERMATA. to represent the under plate of the first brachial segment. The mouth is a star-shaped aperture with the five interradial angles projecting into it (Fig. 142). These angles are formed externally of four plates : the oral plate or buccal shield (7), of two lateral buccal plates (5) lying on the sides of the buccal shield and meeting on the oral side of it, and of the torus angularis, which carries teeth. The lateral buccal shields are supposed to be the modified adambulacrals (laterals) of the second brachial seg- ment. The torus angularis consist of a vertical row of pieces which may fuse together (Fig. 141, ta). The slits between the angles are called the buccal fissures. Fm. 142. Lower surface of disc and base of arms of Ophiactis poa Lym. seen from the out- side (from Lang after Lyman). 1 under plates ; 2 spines of the lateral plates 4; 3 ten- tacle scales ; 5 lateral buccal shields ; 6 apertures of bursae ; 7 buccal shields ; 8 first under plate of arm (supposed to belong to tne 2nd brarhial segment) ; 9 torus angularis ; 10 oral papillae. The skeleton of the disc. The upper skeletal plates of the disc consist principally of the plates of the primary apical system, but there are generally other plates as well, and these may be so numerous as to completely obscure the primary plates. Moreover the completeness of the system of primary apical plates varies even in the same genus. In a typical case the ar- rangement of plates on the upper side of the disc is as follows (Fig. 143) : a central plate surrounded by five radials (r) and five basals (ba) ; the radials are separated from the central by the OPHiUEOIDEA 199 infrabasals, beyond this system are the ten radial shields (rs) and the five second interradials (ir), which do not belong to the primary apical system. The completeness of the apical system varies much in the group. In some forms, the embryonic condition is retained, but even here it may be reduced to the central and five radials, or central and five basals. Not infrequently it ( happens that the integument of the disc is soft and the plates small, scat- tered and incon- spicuous. The radial shields are perhaps the most constant and con- spicuous of the upper skeletal plates of the disc (including the apical system). They sometimes reach from the base of the arm to near the centre of the disc. FIG. 143. Plates of the upper surface of the disc of Ophiomusium validum (from Lang, after P. H. Carpenter), ce central plate ; r radial, ba basal plates ; the dotted plates betweeu the basals and the central are the infrabasals ; ir interradial plates ; rs radial shields ; br arms. At the end of each arm there is a median termi- nal plate which differs from the terminal plate of Asteroids in the fact that it not only lies over the terminal unpaired ten- tacle of the water-vascular system, but completely surrounds it. The alimentary canal, the mouth of which has already been described, leads by a short oesophagus into a large stomach (Fig. 144). The stomach is without any special glandular appendages and there is no anus. The nervous system is arranged very much as in Asteroids, except that the ectoneural plexus is absent in those forms in which the epithelium of the ectoderm is reduced, and that the ventral system (circumoral ring and ambulacral nerves) is re- moved from the surface and lies in the wall of an epineural canal (Fig. 140, 7). This canal, which is lined by ectoderm, being developed as an ectodermal groove, is covered ventrally by calcareous plates. It is found in connexion with the circumoral 200 PHYLUM ECH1NODERMATA. ring as well as with the radial cords. The latter are swollen into ganglia at the points where the nerves come off. The deeper oral system consists of two trunks in each arm placed close together and just above the radial trunks of the superficial system (Fig. 140, 9). They innervate the inter- vertebral muscles (Fig. 140) and are continued into a circum- oral band in the disc. The apical nervous system appears not to be represented, unless the genital nerve ring which runs in the wall of the aboral sinus belongs to it. The tube-feet are all sensory structures, and are supplied by a branch from the ambulacral nerve (Fig. 140, 4), which dilates into a ganglion at aB P# JCW PMS'PH JSi~ T Z Z T ' ! B 2 B* B 4 B FIG. 144. Diagrammatic vertical section through the disc and one radius of an Ophiurid (0 phinglypha), from Perrier, after .Ludwig. A^ peristomial plate (1st ambulacral) ; A' 2 - A 6 2nd to fith ambulacral plates ; aB generative rachis ; B l -B$ the 1st to the 6th ventral plates (under arm plates) ; Bi mesenterial filaments attaching the stomach to the body- wall ; Br radial blood-vessel ; D stomach ; ePH circumoral perihaemal sinus ; iPH perioesophageal part of perivisceral cavity ; K W body wall ; L body-cavity of disc ; L 1 of arm ; Li lip ; M, A/i 1 , .Mi 11 muscles of the oral skeleton ; ME oral angle plate ; MF and MF 1 first and second oral feet ; MS buccal shield ; j.V nerve ring ; A T r radial nerve ; mouth ; oB circular blood-vessel ; P polian vesicle ; PH aboral sinus ; rPH radial perihaemal canal ; S septum which separates the perioesophageal sinus from the rest of the perivisceral cavity ; T torus angularis ; IF water-vascular ring ; Wr radial water vessel ; Z teeth. their base. There is no eye, but the ambulacral nerve trunk becomes epithelial in position on the terminal tentacle. The general arrangement and relations of the coelom are the same as in Asteroids, the principal difference consisting in the ventral position of the madreporite. The general body-cavity or perivisceral cavity is in relation with the stomach in the disc and is prolonged into the arms on the dorsal side of the ambulacral ossicles (Fig. 140). The portion in the disc is traversed by connective tissue strands (Fig. 144), and is divided into two parts by a septum connecting the oesophagus with the oral skeleton (S). A small perioesophageal OPHIUROIDEA. 201 sinus (iPH) is thus cut off from the main portion. The brachial portion is dilated segmentally over each pair of ambulacral ossicles. The walls between these dilatations or chambers are imperfect and are traversed by calcareous plates which connect the ambulacral ossicles with the plates of the ambital skeleton. There is a streak of epithelium carrying specially strong cilia in the upper wall of the brachial continuation of the peri visceral coelom (Fig. 140, 18). The water-vascular system is almost exactly as in Asteroids. It consists of a circumoral vessel sending off a prolongation along each arm. This gives off lateral branches to the tube-feet which however are without ampullae and are purely sensory in function ; it ends in the terminal tentacle. The first two pairs of tube- feet are in relation with the mouth as oral tentacles (see p. 197), and are supplied by canals which arise from the circumoral vessel. There is a polian vesicle in each interradius except that of the stone-canal. In Ophiactis virens, which has several stone- canals, there are not only two or three polian vesicles in each interradius, but also a number of tubular prolongations (canals of Simroth) of the circumoral vessel which encircle the intestine and penetrate between the generative organs. These tubes are supposed to be respiratory in function, a view which is sug- gested by the fact that the genital bursae are absent in this species. The stone-canal (Fig. 145, 2), which however is without any calcareous matter in its walls, passes ventralwards to open into the ampulla (3), which corresponds to the whole of the axial sinus of Asteroids and opens to the exterior by the pore-canal (4) on the ventral surface of the disc, on one of the buccal shields (oral plates). The opening of the water-pore is placed asymmet- rically on the oral plate on an edge of it adjacent to a bursal slit. As a rule there is only one water-pore, but in some species of many genera of Ophiurae (Amphiura, Ophiolepis, Ophiopocus Ophionereis, Ophiocnida), and in all Astrophytidae there are several pores on the buccal shield concerned. In Trichaster elegans, there are five stone-canals and five water-pores, one in each interradius. In Ophiactis virens, which re- produces itself by division, the stone-canals are repeated in several interradii. In these cases of repetition of the stone- canals, the young forms are said to have only one. 202 PHYLUM ECHINODERMATA. Some or all of the corpuscles in the water-vascular fluid of Ophiactis virens are coloured red with haemoglobin. The axial organ has the same structure as in Asteroids. It is found on the side of the stone-canal turned away from the mouth (Fig. 145). It is in relation witli a section of the body cavity (Fig. 145, 7), which is quite unconnected * with the ampulla (axial sinus) and is probably the diverticulum of the left posterior body-cavity of the larva which is formed in the invagination of the primitive germ-cells (MacBride). The axial organ (Fig. 145, 8) is continuous with the generative rachis, which is contained in the aboral sinus (Fig. 144, PH) 16 (a portion of the left larval coelom as in Asterids) and takes a somewhat peculiar course round the disc. Radially it lies in the aboral part of the disc FIG. 145. Diagram of a vertical section through the j lp f wppn fV,p stomach cae- madreporitic interradius of Ampkium squamata, Between 1 showing the relations of the axial organ, stone- ,,, Qrir l f] 1P nrmpr in- canal and neighbouring sinuses (after MacBride). ' 1 circumoral water-vascular vessel; 2 stone- f pmlTrlf vnf wViprpn in pirli canal; 3 ampulla (axial sinus) ; 4 pore-canal ; tegument, Wiiei 5 closed sac which appears to represent the right infpwa/line if r1ir> rlnwn hydrocoel; 6 aboral sinus in its ventral position; mtCliaC 7 sinus derived from the left posterior coelom and U p 4- wppn f! 1p etnmaoli rappa sometimes called the axial sinus ; 8 axial organ ; DetWCen W16 9 genital rachis; 10 genital bursa ; 11 wall of f^.,, r ,1 Q f| 1p Inwpr <5iirfflf>p stomach ; 12 perioesophageal sinus ; 13 inter- OWarClS U16 radial muscle; 14 nerve ring; 15 teeth; 16 /f^: CT ~\ AA\ Tf liac Vippti mouth ; 17 oral surface of disc. suggested that this course is due to the fact that, in the interradii, structures which were originally on the upper side of the disc have moved on to the lower surface ; as an instance of this may be cited the water- pore which in the young form is dorsal, but in all adults has passed on to the ventral surface. The perihaemal system consists of a circumoral canal which is prolonged into a radial tube in each arm on the upper side of the nerve cords (Figs. 144, ePH ; 140, 10). There does * Cuenot, who thought that the two are in communication, appears to have been in error. His mistake is reproduced, together with an errone- ous figure, on p. 820 of the section on Echinoderms in Bronn's Thierreich (1900). OPHIUROIDEA. 203 not seem to be any representative of the inner oral perihaemal ring of Asteroids. The vascular tissue seems to be arranged very much as in Asteroids. In some forms the radial vessels are said not to be present. The generative organs. Most Ophiurids are dioecious, but Amphiura squama-ia is hermaphrodite. The generative glands are simple sacs which open, in all except Ophiopus and Ophiactis virens, into the genital bursae. The genital bursae are five pairs of sac-like invaginations of the interradial portions of the lower wall of the disc, and project into the body-cavity between the bulgings of the stomach. They open by slit-like apertures (double in Ophiura) placed one on either side of the insertions of the arms into the disc (Fig. 142). They have thin walls, lined by a ciliated epithelium and often containing calcareous matter. The gonads are small sacs placed on the walls of the bursae and opening into them. Each generative sac is connected with a branch from the generative rachis, the course of which has been described. The aboral sinus is continued with the rachis to each gonadial sac and invests it as in Asteroids. The generative cells pass into the bursae and outwards by their slit-like apertures. In many Ophiurids (Amphiura squamata and magellanica, Ophiacantha vivipara and marsupialis ; Ophiomyxa vivipara, etc.) the bursae act as brood pouches and the eggs develop in them ; but the principal function of the bursae seems to be respiratory, water being continually drawn in and ejected by the ciliary currents, and in some cases by the muscular elevations and depressions of the dorsal surface. In Ophiactis virens the bursae are absent * and the gonads open directly on the lower surface of the disc. They are replaced by the canals of Simroth, which have been already described (p. 201). The power of regenerating lost parts, e.g. arms, even a portion of the disc, is considerable, but apparently the disc cannot be regenerated from a single arm as in Asteroids. The brittle stars readily lose their arms, so that the power of regeneration is very important. Reproduction by fission through the centre of the disc has been observed in a few genera (e.g. Ophiactis). * Cuenot, Arch. Biologic, 11, 1891, p. 303. 204 PHYLUM ECHINODERMATA. Among preserved Ophiurids many specimens are found in which the dorsal surface of the disc is absent.* The significance of this fact is not understood. The larval form is the Ophiopluteusf (p. 140). In those forms in which care of the brood occurs there is no free larva. The ophiopluteus may present modifications in which the arms are reduced. Such are Ophiopluteus metschnikoffi and daparedei. The larvae known as Ophiopluteus annulatus,% 0. krohnii, O. oblongus are vermiform, without ciliated band and without or with only one skeletal rod. These larvae are pelagic. In- cluding the vermiform larvae about seventeen ophioplutei are known, most of which are as yet unrelated to any adult. The Ophiuroidea live upon the bottom of the sea and feed upon the minute organisms and organic matter contained in the surface mud, which they take up by means of the buccal tube-feet. They move fairly actively by means of the lateral flexion of their arms. In some forms the arms possess a power of vertical movement as well, especially towards the end. When the arms are very long they can be moved in a serpentine manner. The group first makes its appearance in the Upper Cambrian (Ordovician). Its affinities are with the Asteroids, with which it is sometimes united under the superclass Stelleroidea. It is indeed difficult to separate them, especially when the palaeo- zoic genera (Eophiura, Bohemura, etc.), recently described by Jaekel, || and such a form as Astrophiura are considered. The principal points of difference relate to the closing over of the ambulacral groove (a feature which is but slightly marked in * Jeffrey Bell, in Gardiner's Maldive and Laccadive Expedition, 1903, p. 223. Prof. MacBride states that he has seen an Amphiura squamata throw off the whole dorsal surface of the disc with the stomach. t Mortensen, op. cit. j This is J. Miiller's vermiform asteroid larva, with segmented body (Ueb. d. Larven u. d. Metamorphose d. Echinodermen, Abh. 3, p. 26, Abh. d. Pr. Akad. d. Wiss. zu Berlin, Abh. i.-vii., 1846, 1848-53), and appar- ently the larva described by Grave ( Mem. Johns Hopkins, 1900). The latter is at first uniformly covered with cilia, which later become restricted to four bands. These give it a resemblance to an Antedon larva, a resemblance which is heightened by the fact that the last trace of the larval organ (preoral lobe) is found at the edge of the aboral surface of the disc. The author does not state whether this organ is encircled by the water- vascular ring as it is in Asteroids. Preyer, Naples Mittheilungen, 7, 1886, p. 27. || Asteriden und Ophiuriden aus dem Silur Bohmens, Zeitschr. der deut- schen geol. Ges., 55, 1903, p. 106. See also Bather's description of a Devonian genus, Sympterura (Geol. Mag., 1905, p. 161.) OPHIUROIDEA. 205 the Lysophiurae and Ophioteresis), the sharp differentiation of the arms from the disc, the absence from the arms of any prolongation of the alimentary system, and the ventral position of the madreporite. Lastly the free larva has the pluteus form. The closure of the ambulacral groove and the presence of an epineural canal is a feature of some importance and one which the class has in common with Echinoids and Holo- thurians. The Ophiuroidea are divided into four orders : (1) the Lyso- phiurae, which are palaeozoic forms in which the ambulacral ossicles alternate ; (2) the Zygophiurae (brittle-stars) in which the system of dermal plates is well developed and in which the arms cannot be rolled up ; (3) the Streptophiurae, which are also brittle-stars, but which approach in some of their characters the Cladophiurae ; (4) the Cladophiurae (gorgon-heads) which have a thick integument with granular deposits without regular dermal plates. Order 1. LYSOPHIURAE. The ambulacral ossicles are alternate and are not united into pairs, but those of each segment are separate. There are no ventral arm plates. All extinct, Silurian and Devonian. Eophiura, Bohemura, Sympterura, Protester, Bundenbachia, Sturtzura, Eugaster, Ptilonaster, Taeniura, Palaeo- phiura. Order 2. ZYGOPHIURAE (OPHIURAE). Brittle-Stars. The surfaces by which the ambulacral ossicles of the arms articulate with one another are provided with processes and pits which fit into one another and limit the movement of the ossicles upon one another. Upper, under and lateral arm plates are present, and the arms are incapable of coiling round straight rods. The lateral arm-plates bear spines. 1. With arm-spines short, parallel to the long axis of the arms. Fam. 1. Ophiodermatidae. With numerous oral papillae, without dental papillae, with arm incisions on the disc. Ophioderma M. and T., Ophioncus Ives, Ophiogona Stud., Pectinura Forbes, Opliiopezella Ljg., Ophiopinax Bell, Ophiopeza Ptrs., Ophiopyren Lym. (Fig. 141), Ophioconis Ltk. Fam. 2. Ophiolepidae. With 3-6 oral papillae of which the innermost is rarely infradental, without dental papillae, with arm incisions on the disc. Ophiotrochus Lym. ; Ophiopaepale Ljg. ; Ophioceramis Lym. ; Ophiothyreus Ljg. ; Ophiolepis M. and T. ; Ophioplocus Lym. ; Ophio- zona Lym. ; Ophioplinthus Lym. ; Ophiolipus Lym. ; Ophiernus Lym. ; Ophiophyllum Lym. ; Ophiochaeta Ltk. ; Ophiopleura Dan. ; Ophiopyrgus Lym. ; Ophiomastus Lym. ; Ophiomusium Lym. (Fig. 143), no tentacle pores beyond the basal arm-joints ; Ophiotypa Khlr. ; Ophiura Lm. ; 206 PHYLUM ECHINODERMATA. Ophiocten Ltk. ; Gymnophiura Mrtsn. British species : Ophiura ciliaris, E. X. Atl., 7-100 fms. ; 0. albida, ditto, to 250 fms., O. affinis, ditto, 10-192 fms. 2. Spines at right angles to the arm axis. Fam. 3. Amphiuridae. With 1 to 5 oral papillae, the innermost often infradental. Arms inserted on the ventral side of the disc. Dental papillae absent. Ophiambyx Lym. ; Ophiopholis M. and T. ; Ophiostigma Ltk. ; Ophiochiton, Hemipholis, Amphilepis, Ophiocnida, Ophiophragmus, Ophiopla.r, Ophiochytra, Ophiomyces, all Lym. ; Ophiopits~Ljg. ; Ophiactis (Fig. 142) and Opliionereis Ltk.; Amphiura and Ophiopsila Forbes; Paramphiura Khlr. ; British species: Amphiura chiajii, to 120 fms ^ A. filijormis, to 120 fms. ; A. elegans, to 120 fms. Ophiactis abyssicola, 64 to 767 fms. ; 0. balli, to 203 fms. Ophiopholis aculeata to 300 fms. Ophiacantha bidentata, 20-2,335 fms. Fam. 4. Ophiohelidae. Disc with scales and sharp or blunt spines,, with teeth and oral papillae, without dental papillae. Ophiomitra, Ophiothamnus, Ophiocamax, Ophiotholia, all Lym. Fam. 5. Ophiacanthidae. Disc covered by soft skin, which more or less hides the subjacent scales ; dental papillae absent or few. Ophia- cantha il. and T. ; O. vivipara Ljn., 6 or 7 arms ; 0. anomala G. O. Sars, 6 arms. Ophiolebes, Ophiotoma, Ophiogeron, Ophiosciasma, Lym. : Ophio- blenna, Ophionema, Ophionephthys Ltk. ; O phiocentrus Ljg. ; Ophioscolex M. and T. ; Ophiotrema Khlr. Brit. sp. : Ophioscolex glacialis, 100-300 fms., 0. purpurei, 64 to 76 fms. Fam. 6. Ophioeomidae. With oral and dental papillae. Ophio- cymbium Lym. ; Ophiocoma L.Ag. ; Ophiarachna and Ophiomastix M. and T. ; Ophiopteris E. Sm. ; Ophiarthrum Ptrs. British species r Ophiocoma nirjra. to 87 fms. Fam. 7. Ophiotrichidae. With 8-10 dental papillae, without oral papillae. Ophiopteron and Opltiotrichoides Ludw. ; Ophiothrix and Ophiocnvmis M. and T. ; Ophiocampsis Dune. ; Ophiomaza and Ophio- psammium Lym. ; Ophiothela Verr. ; Ophiogymna Ljg. ; Luetkenia, Gymnolophus, Ophioaethiops, Ophiosphaera Brock. Brit. sp. : Ophiothrix frag His, to 52 fms. Order 3. STEEPTOPHIUEAE. Astrophytoii-like Ophiurae. The ambulacral ossicles articvilate with one another by means of a more or less simple ball-and-socket joint, and the arms can be moved in a vertical direction and be coiled towards the mouth. Upper, under, and side plates are more or less regularly developed : the side plates bearing spines. Fam. 1. Ophiomyxidae. With 3-7 oral papillae, without teeth. Arms covered with soft skin. Neoplax Bell ; Ophioteresis Bell, no under arm-plates ; Ophiobyrsa, Ophiochondrus, Sigsbeia and Ophiobrachion Lym. ; Ophiomyxa M. and T. ; Hemieuryale Martens ; Astrophis A. M. Edw. ; Ophiohclus. Several extinct families come here, viz., Ophiurinidae, Lapworthidae, Eoluididae, Onychasteridae, Eucladiidae. The Eucladiidae from th& middle Silurian of England have short arms, and a ventral madreporite. Each arm has 2 or 3 arm -like branches ; these are tube-feet with a flexible^ armour of minute spiny scales. Eudadia H. Woodward, Euthemon. Sollas. OPHIUROIDEA. 207 FIG. 146. Trickaster elegans (after Ludwig). Order 4. CLADOPHIUEAE (EURYALAE). Gorgon-heads. The ambulacral ossicles articulate with one another by means of hour glass-shaped surfaces and are covered by granular deposits in the thick integument. The arms may be simple or branched repeatedly. They can be moved in the vertical plane and 'coiled towards the mouth. There are no spines on the sides of the arms. Most of those with ion- branched arms have a mouth-shield at the inner angle of each lower interbrachial space, one of which serves as the madreporite. Those with branched arms have often no mouth-shields, and the madreporites, sometimes single, sometimes five in number, are found in various regions of the lower interbrachial spaces. Pedicellaria-like processes are sometimes present. Fam. 1. Astro phytidae. With simple arms. Astrotoma Lym.* ; Astronyx M. andT.* ; Astro- chele Vll.* ; Astrogomphus Lym.f ; Astroporpa Orst. and Ltk.f; Ophiocreas Lym.| ; Astrocherna Oerst. and Ltk.J; Astro- ceras Lym.f. Brit. sp. : Astronyx loveni to 350 fms. Fam. 2. Trichasteridae. The arms branch a few times near their free ends. Trichaster L. Ag. (Fig. 146) ; Astroclon Lym.; Astrocnida Lym. Fam. 3. Euryalidae. The arms branch much and from near their base. Euryale Lmk. ; Gorgonocephalus Leach (Fig. 147) ; Astropliyton Linck. Brit, sp. : Gorgonocephalus lincM and eucnemis. The genus Astrophiura Sladen though undoubtedly an Ophi- uran presents some Asteroid features. The family Astro phiu- ridae has been created for its reception. The disc is penta- gonal and the greater part of the arms are included in it. The free portion of the arms is short, reduced and without tube-feet. There are no teeth and the buc- cal armature is simple and im- perfect. Under arm-plates are present and the cavities for the retracted feet are spacious. The madreporite appears to be ven- FIG. 147. Young Gorgonocephalus agassizi, ventral view (after Lyman). tral. Madagascar and neighbouring islands. the Disc large. Disc moderate (about one-tenth of length of arms). { Disc small. 208 PHYLUM ECHINODERMATA. Class ECHINOIDEA * Spherical, oval, or discoidal Echinoderms with a shell composed of calcareous plates usually closely fitting and carrying movable spines. The mouth is on the under surface, and the anus either within the apical system or between the apical system and the mouth. The five ambulacra are indicated by rows of pores and usually extend almost to the aboral pole. There are typically five inter- radial gonads. The body of an Echinoid is typically spherical, but is often heart-shaped, oval, or flattened. There are no arm-like pro- longations, but there is a five-rayed symmetry in the build of the body as is shown by the disposition of the water-vascular, nervous and other systems of organs. The dermal skeletal plates, which are pentagonal or hexagonal in shape, are con- nected together so as to form, with rare exceptions (Echino- thuridae, which see, and certain Palaeechinoidea, plates of posterior ambulacrum of some Spatangids, peristomial plates of Cidaroida), a firm immovable skeleton, the test or shell. The mouth is on the lower surface of the body, generally in the centre, but sometimes shifted towards what is called the anterior end ; the apical system is on the upper surface and the anus is either within it (Endocyclica) or outside it (Exocyclica}. * A. Agassiz, Revision of the Echini, Cambridge U.S.A., 1872-4. Id. Report on the Echinoidea, Challenger Reports, 1881. Id. Panamic deep- sea Echini, Mem. Mus. Harvard College, 31, 1904. Cotteau, Echinides, Paleontologie Francaise, 7, 9, 10, Paris 1862-79. P. Martin Duncan, A revision of the genera and groups of the Echinoidea. Journ. Linnean Society, 23, 1891. S. Loven, Etudes sur les Echinoides, Kongl. Svenska Vetenskaps-Akademiens Handlingar, 11, 1872; a translation in Ann. and Mag. Nat, History (4), 10, 1872. C. F. and P. B. Sarasin, Ub. d. Anat. d. Ecliinothuriden etc. Ergebnisse Nat. Forschungen Ceylon, 1, 1888. Id., Die Auge etc. der Diadeniatiden, ibid. 1, 1887. J. Muller, L. Cuenot, O. Hamann, Delage Op. cit. Prouho, Recherches sur lo Dorocidaris papillata, etc., Arch. Zool. Exp. et gen. (2), 5, 1888, p. 213. O. Hamann, Echinoidea in Bronn's Thierreich, 1901-1905. Th. Mortensen, Echinoidea (Pt. l),in the Danish Ingolf -Expedition, I, vol. 4, 1903. F. Leipoldt, Das angebliche Excretions-organ der Seeigel, Z.f.w.Z., 55, 1893, p. 585. S. Loven, on Pourtalesia, K. Svenska Vet. Akad. Handl. 19, 1884. Id., Echinologica, Bih. Svenska Vet. Akad. ForhdL, 13, 1887. Id., Echinologica, Bih. Svenska Vet. Akad. Handl., 18, 1892. H. Theel, On the development of Echinocyamus pusillus, Nova Acta R. Soc. Sci. Upsala, 1892. Id., Prel. ace. of the devel. of Echinus miliaris, Bih. Svenska Akad. Handl. ,28, 1902. E. W. MacBride, Dev. of Echinus esculentus etc., Phil. Trans., 1903, p. 285. ECHINOIDEA. 209 When the anus is outside the apical system, it always lies in what is called the posterior interradius (V. I). The outline of an Echinoid shell when viewed from the apical pole is called the ambitus. The plates of the skeleton are covered by the ciliated epidermis and lie entirely superficial to the nervous and water- vascular systems ; they are perforated by apertures for the passage of the tube-feet, and bear prominences and tubercles to which the variously shaped spines are movably articulated. The apical system is very limited in extent and takes up the whole of the abambulacral surface of the body (Fig. 149). The XOJ-* ^-^ft^^y^''-- -'- ^ftit^^fS ^t/i^^-^S~ : FIG. 148. Test of a young regular sea-urchin Strong -jlocentrotus droebachiensis (from Claus). a from the aboral side ; b from the oral side. PR rows of pores in the anterior radius. The peristomial membrane contains the mouth with 5 teeth in the centre and 5 pairs of plates perforated by pores for the oral tube-feet. madreporite is on the upper surface and is one of the apical plates, generally the basal of what is called the right anterior interradius (II, III). For purposes of description, there may be said to be two principal kinds of Echinoids, the regular forms, in which the body is more or less spherical and the anus is within the apical system (Endocyclica, Regularia), and the irregular forms, in which the body is oval, or heart-shaped, and more or less flattened in the principal axis, and in which the anus is outside the apical system in the posterior interambulacrum (Ectocyclica, Irregularia). The test consists of the plates of (1) the apical system, and (2) the corona or rest of the shell ; z m P 210 PHYLUM ECHINODERMATA. to these may be added the peristomial plates, i.e. the plates in the peristomial membrane in the middle of which the mouth is placed (Fig. 148 b). In the regular forms or Endocyclica (Diadematoida, Cidaroida, most Palaeechinoidea), the apical system (Fig. 149) consists of (1) the periproct area containing a number of small plates, the periproct plates, amongst which, towards the right posterior radius (No. I) or in- terradius (I. II) lies the amis (an) ; (2) the five interradially placed basals (genitals), which are usually per- forated each by a genital opening (go), and one of which (m) is perforated by the water - pores of the m a d r e p orite. With this system, though not belonging to it, must be mentioned the five radially placed ocular plates (radials) which are perforated for the small terminal tentacles of the water- vascular system : these plates are really the terminals and belong to the ambulacral surface. Some or all of the radials may be wedged in between the basals and assist in forming the boundary of the periproct. Occasionally, as a specific or individual character, more than one genital opening is present on each genital plate. Thus two pores have been found in Cidaris perornata, Arbacia punctulata, etc., five on the madreporite of Echinus acutus. In short, among both fossil and recent species there is some variability in this character.* In the young of the Saleniidae the centre of the apical system is occupied by a central, on one side of which, in the right posterior interradius, the anus is placed. This condition is repeated in most young sea-urchins (Fig. 150), but in adults a number of small plates the periproct or Flo. 149. Apical plates of young Strongylocentrotus droe- bachiensis (from Lang, after Loven), an anus among the periproct plates ; go genital openings on the basal or genital plates ; m madreporite (right anterior basal) ; r radial plates (ocular). * In the cretaceous and tertiary genus Goniopygus (Arbaciidae) the genital openings are outside the apical system. ECHINOIDEA. 211 anal plates are added and the central becomes indistinguishable. The anus is always excentric, being displaced towards the right posterior interradius (as in Asterids). The test or corona consists of ten double meridional rows of plates passing round from the apical to the oral pole (Fig. 148). Five of these double rows are radial in position and constitute the ambulacra, while five are in- terradial and are called the inter - ambulacral plates. This statement is true of all Euechi- noidea* In the Palaeechinoidea, how- ever, the number of rows of plates in the interambulacra is either one (Bothrioci- daris) or more than two (from 3 to 11). The plates are pentagonal in F ig. iso. Apical system of a young e i i -, ,1 ,i Echinus (from Delage, after Loven). lorm aild are SO placed that the An anus; c central] gtx one of the v i ,1 genital plates. I. to V. indicates the median SUture between the tWO enumeration of the radii adopted in rOWS of an ambulacrum is zigzag ^rdTinteiradiS No^AlVand the /-TV i