A STUDENT'S TEXT-BOOK OF ZOOLOGY CORRIGENDA Page s). Line 7. For " chordate " read " non-chordate." „ 16. Line 6. Omit " mesodermal." „ 17. In the description of flg. 6, for " spina " read " spinal." „ 117. Line 5. For " branchia " read " branchial." „ 125. Line 8. For " Myliobates " read " Myliobatis." „ 146. Line 8. For " somactido " read " somactids." „ 152. Line 13. The genus Odontaspis is not extinct, see line 1 of "the same page. „ 210. Line 15 from bottom. For " Oastrosteus " read " Gasterosteus." „ 219. Line 4. For " vulggaris " read " vulgaris." „ 219. Line 20. For " E. Ind. Oreinus ; McClell " read " E. Ind. ; Oreinus McClell, „ 220. Line 12. For " RHODENIA " read " RHODEINA." „ 229. Line 26. For " trumpetor " read " trumpet or." „ 230. Line 13. Insert a comma after " JRaf." „ 231. Line 6 from bottom. Omit " Seriolella." „ 235. Line 18. For " specie " read " species." „ 242. Line 27. Insert comma after Elanura, and another after Melletes. „ 242. Line 11-12 from bottom. The 5th letter of the word Stellerina'js indistinct. „ 244. Line 23. For " Thallasso- " read " Thalasso-". „ 255. Line 14 from bottom. For " succus " read " saccus." „ 273. In the second footnote for " Paratoids " read " Parotoids." „ 276. Line 13 from bottom. Insert after the words " in other Amphibia." " In Rancr the ductus endolymphaticus which passes off from the saccule enters the cranial cavity through the foramen endolymphaticum and there dilates into the saccus endolymphaticus. This extends back into the neural canal of the vertebral column, where it lies in close apposition to its fellow along the dorsal side of the spinal cord and gives off transverse diverticula which end in small dilata- tions over the posterior root ganglia. The fluid contained in the saccus and its extensions is milk-white, the milkiness being due to the presence of crystals of carbonate of lime (otoliths). The CALCAREOUS GIAXDS are the terminal dilatations of the transverse diverticula above referred to (see Gaup, o NEW YORK: THE MACMILLAN CO. 1905 ^ L- PREFACE IN presenting the second part of my work on Zoology to the public I must apologise for the delay in its appearance and for the fact that I am not keeping to the undertaking which I gave in the preface to my first volume that the work would be completed in two volumes The delay in publication has been caused in part at least by the fact that the Verte- brata compel a lengthier and more detailed treatment than the other groups. Not only is more known about them, but they excite greater interest, and their palaeontological history has been more completely worked out than in the case of any other phylum. The result has been the present bulky volume which deals only with them and with Amphi- oxus. Embryology is of course excluded, except in the case of Amphioxus, but I have endeavoured to deal fairly fully with anatomy, habits, and classification. In the systematic portions I have probably been too ambitious, but the usefulness of a book of this kind depends largely upon ita completeness in this respect, and in cases of doubt I have generally included rather than excluded. In Aves alone have I made a selection ; for there are many excellent works devoted to them and it would be impossible to give anything like a complete list of their genera. In judging the anatomical portions I would ask the reader to remember that this is not exclusively a work on Compara- tive Anatomy, and limitations of space forbid an exhaustive treatment. This branch of the subject has not, however, been neglected, and questions of general interest have usually been at least touched upon. Considerable space has been given to extinct forms. VI PREFACE As in the first volume I have endeavoured in the index to refer the reader to the page on which technical terms are defined and most of the abbreviations are explained in the same place. To Mr. J. J. Lister, F.R.S., I am again under very great obligations. He has looked through all the proof sheets and has given me the full benefit of his wide knowledge and great critical powers. I am also indebted to his pencil for the excellent illustrations on p. 341 and p. 532. My thanks are also due to Professor Newton, Professor Kay Lankester, Mr. Boulenger, Professor MacBride, Pro- fessor Graham Kerr, Dr. Chalmers Mitchell, Dr. Andrews, Mr. Walter Heape, Mr. Assheton, Dr. Gaskell, Dr. Marett Tims and others for the assistance they have given me in different parts of the work. My principal sources of information are acknowledged in the footnotes, but I must not omit to mention here works from which I have obtained special help ; these are Gadow's Amphibia and Reptiles and Evans' Aves in the Cambridge, Natural History, Flower and Lydekker's work on Mammalia Living and Extinct, Smith Woodward's Outlines of Vertebrate Palaeontology and Zittel's Grundriss der Palaeontologie, and Gunther's Introduction to the Study of Fishes. The excel- lent volume on fishes in the Cambridge Natural History and Weber's great work on the Saugethiere appeared too late to be utilised. Of the illustrations about fifty are new ; of the remainder a considerable number are from Claus' Lehrbuch, but some, of which I have been permitted to make use by the courtesy of the author and publisher, are from Smith Woodward's Vertebrate Palaeontology, Reynold's Vertebrate Skeleton, Flower and Lydekker's Mammalia Living and Extinct, Flower's Osteology of Mammalia, Huxley's Anatomy of Vertebrate Animals, Shipley and MacBride's Zoology, Zittel's Grundzuge der Palaeontologie, Korschelt and Heider's Text Book of Embryology, Gegenbaur's V ergleichende Anatomie der Wirbelthiere Wiedersheim's Grundriss der Anatomie der Wirbelth^ r , Terrier's Traite de Zoologie, PREFACE VII Balfour's Comparative Embryology, Gadow's Amphibia and Reptiles. The third volume, which is in the press, will deal with the Tunicata, Enteropneusta, Echinodermata and Arthro- poda. It has been pointed out to me by friends who have read the proofs that I have made statements which without a fuller treatment may give rise to the view that I am unortho- dox on the great question of organic evolution. This is not the place to give that fuller treatment, but in order to pre- vent misunderstandings'! may say that any such view would be erroneous. I am and always have been a convinced evolutionist. I hold, that is to say, that matter is constantly undergoing change, and that natural selection, taking advan- age of its endless diversity in form and properties, has played and is playing an important part in determining what form of it (whether living or non-living) shall exist and what shall cease to exist. I hold further that the forms of living matter, as well as those of non-living matter, owe their existence and their properties to the operation of natural laws, though here we are treading on more uncertain ground, for we know nothing of the origin of living matter or of the sources of its properties. The chemist has made many forms of matter which have, at present at least, no existence in nature apart from organisms, but he has not yet succeeded in making living matter. Whether he will ever be able to do so is a question which may fairly be asked, but is one which cannot now be answered. The view that living matter arose in response to the operations of natural laws cannot be either proved or disproved. It must remain a matter of belief for which there is much to be said. As to the origin of the manifold properties of living matter we know nothing. The Darwinian theory did not account for properties ; it left their origin to an imperfectly understood interaction between the organism and the environment, and further than this we cannot at present go. It may, however, be pointed out that there are two ways in which Vlll PREFACE this great question may be attacked. One of these is by the method of experiment — a method which is being pursued with increasing vigour by more than one school of Biologists ; the other is the careful and thorough exam- ination of living and extinct organisms, particularly in their relations to one another. It is the second of these methods which comes under our notice in the three volumes of the first part of this work dealing with systematic zoology. It is unquestionable that this study does shed light, if only a dim light, on the course of organic evolution and indirectly on the origin of the properties of living matter, and it is most important that the light so obtained should be brought to bear upon the problem. To discover this we must approach the subject with unbiassed minds, for it is one of immense complexity and it is extremely unlikely that any particular solution which commends itself to us will turn out to be final. I would therefore ask for lenient judgment if in some pages of this work I have seemed to take up an unduly critical position with regard to views widely prevalent at the present time on some aspects of organic evolution. That does not mean that I am unsound on the great question itself, but only that I am sceptical as to the value of some hypotheses widely held as to the course of organic evolution. It is true that working hypotheses are necessary in constructive work, but in a subject of the complexity of the present one, they can only be provisional and as such are legitimately open to criticism. It may be urged that I have said too much or too little, that I ought not to have touched upon the matter unless I was prepared to state fully my own views. While allowing that there would be some justice in such a criticism, I do not admit its complete validity. In deference to it, how- ever, I have materially altered in proof what I had written in manuscript, but it was not possible to remove all refer- ence to the subject. It was necessary to note the facts in passing. In the final volume on the Principles of Zoology which I yet hope to write, I shall return to it and endeavour to justify, in the fuller treatment which will there be pos- PREFACE IX sible, the criticisms which are only hinted at here. At the same time I cannot hope to build. That is the task of the great band of workers in many departments of Biology, who, undeterred by failure and urged on by the fire, enthusiasm, and generous aspirations of youth, return time after time, generation after generation, to the assault of the fortresses of nature well knowing that their material reward will be small, that defeat means the world's neglect and that success, except the greatest, brings but a pittance of its esteem. To them I inscribe this book in the hope that it may serve if only to a small extent to smooth over the difficulties of part of the road which at first they have to traverse. A. SEDGWICK. TRINITY COLLEGE, CAMBRIDGE, February, 1905. TABLE OF CONTENTS An asterisk signifies that the group is extinct. PAGE THE CHORDATA . . . . 1 PHYLUM CEPHALOCHORDA . 10 PHYLUM VERTEBRA! A . . 45 Class I. PISCES .... 51 Sub-class 1. MARSIPOBRANCHII 95 ., 2. ELASMOBKANCHII 118 . Order I. Pleuropterygii* . 145 „ 2. Aeanthodii* . . 146 „ 3. lehthyotomi* . . 147 „ 4. Selaehii (Plagio- stomi) . . .148 Sub-order 1. Notidani . . 149 2. Squali . . 150 3. Raji . . .153 Order 5. Holoeephali . . 155 Sub-class 3. GANOIDEI . .159 Order 1. Chondrostei . .167 Order 2. Crossopterygii . .171 Sub-order 1. Osteolepida* . 175 2. Cladistia . .176 Order 3. Lepidostei . . .176 4. Amioidei . 180 Sub-class 4. TELEOSTEI 183 Sub-order 1. Malacopterygii ( Salmonidupei- formes) . .213 „ 2. Ostariophysi (Cyprinisiluri- formes] . .216 Sub-class TEIEOSTEI — contd. PAGE Sub-order 3. Symbranchii (Symbranchi- formes) . . 222 ,, 4. Apodes (An- guilllformes} 223 „ 5. Haplomi (Esociformes) 225 „ 6. Heteromi(Der- cetifwmes) . 227 „ 7. Cateostomi (Gastrostei- formes] . . 228 Tribe A. Selenichthyes . . 228 „ B. Hemibranchii . . 228 „ C. Lophobranchii . . 229 „ D. Hypostomides . . 230 Sub-order 8. Percesoces (Mugiliformes] 230 „ 9. Anacantliini (Gadiformes) 232 „ 10. Acanthoptery- gii . . .233 Tribe A. Perciformes . . 233 B. Scombriformes . . 238 C. Zeorhombi . . 239 D. Kurtiformes . . 240 E. Gobiiformes . .241 F. Discocephali . . 241 G. Scleroparei . . 241 H. Jugulares . . 243 „ I. Taeniosomi . . 244 Sub-order 11. Opisthomi . 245 12. Pediculati (Lopkii formes) 245 13. Plcctognathi (Bqlistiformcs) 246 Tribe A. Sclerodermi . . 246 „ [B. Gymnodontes . . 247 Xll TABLE OF CONTENTS PAGE Sub-class 5. DIPNOI . . .248 A. Ctenodipterini* . 259 B. Sirenoidei . . 259 ARTHRODIRA* . 260 A. Heterostraci* . . 261 B. Osteostraci* . : 261 C. Antiarcha* . . 261 Icthyodorulites*. 262 Conodonts* . . 262 Class II. AMPHIBIA . . .263 Order 1. Gymnophiona (Apoda) 300 „ 2. Urodela (Caudata) . 304 „ 3. Anura (Batraehia) . 307 Sub- order 1. Aglossa . . 309 „ 2. Phaneroglossa 310 Order 4. Stegocephali* . .313 Sub- order 1. Branchiosauri* 315 „ 2. Aistopoda* .315 ,, 3. Labyriniho- dontia* 315 Microsauria* 315 Class III. REPTILIA ... 316 Sub-class 1. RHYNCOCEPHALIA 329 „ 2. LEPIDOSAURIA . 334 Order 1. Dolichosauria* . 334 „ 2. Mosasauria* . . 334 „ 3. Lacertilia . . .335 Sub- order 1. Lacerlilia vera 348 „ 2. Rhiptoglossa . 354 Order 4. Ophidia . . . 355 372 381 Sub-class 3. CROCODILIA . . Order 1. Parasuchia* . . „ 2. Pseudosuchia* . . 382 „ 3. Eusuchia . . .382 Sub-class 4. DINOSAURIA* . 383 Order 1. Theropoda* . . 384 „ 2. Sauropoda* . . 385 3. Predentata* . - . 386 Sub-class DINOSAURIA* — contd. PAGE Tribe 1. Ornithopoda* . . 386 „ 2. Stegosauria* . . 387 3. Ceratopsia* . . 387 Sub-class 5. PTEROSAURIA* . 388 „ 6. ICHTHYOSAURIA* 391 7. PLESIOSAURIA* . 395 8. ANOMODONTIA* 398 Order 1. Pareiasauria* . . 399 „ 2. Theriodontia* . . 400 3. Dicynodontia* . .401 Sub-class 9. CHELONIA . . 402 Sub-order 1. Athecae . . 412 „ 2. Thecophora . 412 A. Cryptodira . . 412 B. Pleurodira . . 414 C. Trionychoidea . 415 Class IV AVES . . . 416- Order 1. Archaeornithes* . 454 „ 2. Neornithes . . . 456 Sub-order 1. Ratitae. . . 456 2. Odontolcae* . 458 „ 3. Carinatae . 460 Tribe 1. Ichthyornithes* . 466 „ 2. Colymbiformes . 466 3. Sphenisciformes . 466 „ 4. Procellariiformes . 461 „ 5. Ciconiiformes . . 461 „ 6. Anseriformes . . 462 „ 7. Falconiformes . . 463 ,, 8. Tinamiformes . . 464 „ 9. Galliformes . 464 „ 10. Gruiformes . 466 „ 11. Charadriiformes . 466 ,, 12. Cuculiformes . 469 „ 13. Coraciiformes . . 471 „ 14. Passeriformes. . . 474 Class V. MAMMALIA. . . 479 Order 1. Monotremata . . 524 „ 2. Marsupialia . . 529 Sub-order 1. Diprotodontia 534 „ 2. Polyprotodontia538 3. Allotheria* . 541 TABLE OF CONTEXTS •Kill PAGE PAGE Order 3. Edentata . . . 542 Order 12. Tillodontia* . . 607 Xenarthra . 543 „ 13. Ancylopoda* . . 609 Nomarthra 548 „ 14. Condylarthra* . . 609 ,. 4. Sirenia . . . . 549 „ 15. Creodonta* . 611 „ 5. Cetacea . . . 553 „ 16. Carnivora . . . 612 Sub-order 1. Mystacoccti . 560 Aeluroidea . 618 „ 2. Odontoceti 561 Cynoidea . . 621 » 3. Zeuglodonta* . 564 Arctoidea . . 622 Order 6. Hyracoidea . . 565 „ 17. Pinnipedia . . 624 „ 7. Proboseidea . . 567 „ 18. Rodentia . . 627 „ 8. Ungulata . . . 573 Simplicidentata . 632 Sub-order 1. Artiodactyla . 576 Duplicidentata . 636 „ 2. Perissodactyla 592 „ 19. Insectivora . . . 636 » 3. Lipoterna* 602 „ 20. Chiroptera . . . 641 Order 9. Amblypoda* . . 603 „ 21. Prosimiae . . . 649 „ 10. Toxodontia* . . 605 00 ,, _-. Primates . . . 653 „ 11. Typotheria* . . 606 TABLE OF GEOLOGICAL PERIODS AND FORMATIONS TERTIARY OR CAINOZOIC PERIOD. Pleistocene Pliocene . . Miocene . . Oligocene . . Eocene Recent Deposits. 1 Valley and Cave Deposits I Glacial Deposits. r Cromer Beds. j Norwich Crag. -[ Red Crag. I Coralline Crag. I In India, the Siwalik Formation. Not known in Britain but wide-spread on the Continent of Europe. It Loup Fork. Lacustrine deposits of North America] Deep River. I John Day. The Santa Cruz Beds of Patagonia are referred to the Miocene. I Hamstead and' Bembridge Beds. -J Headon and Osborne Beds. I In N. America the freshwater White River Beds belong ^ to the Oligocene. Upper : Barton Beds ; In N. America, Uinta Group. Middle: Bracklesham Beds. InN. America, BridgerGroup. Lower : Bagshot Sands, London Clay, Woolwich and Reading Beds. ('1. Wind River Group. 2. Wasatch Group. a "j 3. Torrejon Group. [4. Puerco Group. xiv TABLE OF GEOLOGICAL PERIODS AND FORMATIONS XV Cretaceous Jurassic Triassic MMdle SECONDARY OR MESOZOIC PERIOD. Chalk (with flints), le Chalk (with few flints), •j Lower Chalk. Upper Greensand. VGault. ( Lower Greensand. Upper Lower \Wealden. ( /Purbeck and Portland Beds. Upper -I Kimmeridge Clay (Solenhofen Slates in Bavaria. ) vCorallian and Oxford Clay. ,,.-,, (Great Oolite. le (Inferior OoliteJ(Stonesfield Slate belongs here). ( Upper Lias. Lower j Middle Lias. ^ Lower Lias. I \Keuper Marls and Sandstone. Middle: Muschelkalk, absent in Britain. Lower : Bunter Sandstone, Pebble Beds, Africa. Karoo of S Permian . Carboniferous . Devonian Silurian PRIMARY OR PALAEOZOIC PERIOD. / Upper : Magnesian Limestone. ^ Lower : Red Sandstones. ( TT ( Coal Measures. Upper |M}llstone Grit> ( Lower : Carboniferous Limestone and Shales. ( Upper Old Red Sandstone. ' -c Limestones (marine). Lower Old Red Sandstone. f Ludlow. J Wenlock. I May Hill Sandstone (Llandovery). ( Bala. Upper Cambrian* J Llandilo, Lower Cambrian . Tremadoc Series. Lingula Flags. Menevian Series. Harlech Series. Olenellus Beds. Precambrian L. Silurian of Murchison, Ordovician of Lapworth. CHAPTER I. CHORDATA. Animals with a notochord, a hollow dorsally placed nervous system, and a pharynx opening to the exterior by lateral passages. The group Chordata is a division of the animal kingdom superior to a phylum. It includes four phyla and is to be compared in its rank to such groups as the Metazoa and Coelo- mata, both of which are phylum -including divisions. The four phyla into which the group Chordata is divided are, stating them in the order in which they are dealt with in this work, the Cephalochorda, which includes but a single genus, Amphioxu3 ; the Vertebrata, which is by far the largest and most important division of the group ; the Tunicata, which includes a con- siderable number of marine forms of low organization ; and lastly the Enteropneusta, which has but a small number of genera mainly of vermiform appearance and is the most out- lying phylum of the group. Indeed, by some highly competent authorities the Enteropneusta are placed altogether outside the Chordata, largely on account of their early development, which differs in important particulars from that of other Chordata and approaches that of Echinodermata ; and because it is not certain that they possess that typically chordate organ, the notochord. While not presuming to pronounce an opinion on the latter point beyond saying that if the notochord is present in Enteropneusta, its development, structure and relations to other organs differ considerably from those of the notochord in the other phyla, we desire to emphasise quite distinctly our opinion that the Enteropneusta are Chordates. They present most clearly the other characteristic features of that group, viz., the hollow central nervous system and the perforated pharyngeal wall — features of organization found in no other group of the CHORD ATA. animal kingdom ; and in the arrangement of their coelom they come close to the Cephalochorda and Vertebrata. The notochord itself is a rod-like structure in all cases developed from the dorsomedian endoderm of the embryonic enteron.^ This streak of tissue undergoes a modification of structure almost identical with that presented by the axial endo- derm of the tentacles of many Coelenterata. The modification, which may be described as being of a skeletal nature, consists in both cases of a a vacuolisation of the protoplasm of the endodermal tissue (Fig. 1) and of a considerable development within this tissue of cuticular structures (vide Vol. i, p. 101). Indeed the function of the notochord, like that of the ten- tacular endoderm referred to, is a sup- porting one : it supports the axis of the body and particularly the central nervous system beneath which it lies. In the Cephalochorda this function is discharged by the notochord during the whole of life ; in the Vertebrata and Tunicata however it is purely embryonic or larval in its duration. In the Vertebrata the notochord, though it may in some forms, e.g. Pisces, persist throughout the whole of life, becomes surrounded by a stiff sheath, which takes over its function of axial support and becomes, especially in those forms in which the endoskeleton acquires rigid texture, divided up into segments corresponding with those of the embryonic muscular system. The central nervous system develops from the ectoderm of what is usually called the dorsal surface, and at first nearly always has the form of a groove, which, excepting in the Enter o- pneusta, extends along the whole of the dorsal surface and closes completely to form a canal — the central canal of the nervous system. It is characteristic of Cephalochorda, Vertebrata and Tunicata that this canal opens in the embryo for a shorter or longer pericd unto the enteron (neurenteric canal). This neurenteric communication is however never maintained in FlG.r>' 1. — Transverse section through notochord and spinal cord of the larva of Bombina- tor igneus (after Gotte, jrom Claus). ChS notochordal sheath ; Ch notochord ; Sk skeletogenous layer ; N spinal cord. MOUTH AND ANUS. 3 the adult, and its transitory existence is a highly remarkable fact for which no satisfactory explanation has ever been offered. In the Enteropneusta alone is the central nervous system confined to a short portion only of the dorsal surface (so-called collar region), and in them alone does the central canal remain permanently open and never acquire a communication with the enteron. We have said that the central nervous system arises on the dorsal surface. Now it is quite clear that this surface corresponds to the ventral surface of other Coelomata, so that it would be convenient to exchange the term dorsal for a term which would include the same surface in all Coelomata. Such a term is afforded by the term neural surface, which implies, and cor- rectly implies, that the central nervous system is developed upon it. Another term, blastoporal, having reference to the position of the embryonic blastopore might also be used. In all the Coelomata the blastopore is not only placed on the neural surface of the body, but actually perforates the embryonic rudiment of the central nervous system. This is seen most clearly in the embryonic history of the Cephalochorda, the Verte- brata, the Annelida, Arthropoda and Mollusca. In the Entero- pneusta and Echinodermata this relation is masked, and by many morphologists would be held not to occur at all. But that it does exist we are convinced, and is a most important morpho- logical fact appertaining to all Coelomata. Now in some Coelo- mata it has been definitely proved that the mouth and anus of the adult animal are directly derived from the embryonic blastopore, and it becomes a question whether this derivation, though not embryonically manifested in all forms, does not also hold throughout the Coelomata. Believing as we do in the homology of the mouth and anus, at least in the phyla Annelida, Arthropoda and Mollusca, it follows that this relation holds for them. In Peripatus the mouth and anus are not only derived from the elongated blasto- pore by its constriction into two openings, but remain throughout life included within the nerve ring derived from the neural rudiments of the embryo.* If in other * Sedgwick, " Monograph of the Development of Peripatus capensis" Studies from the Morphological Laboratory of the University of Cambridge, 4, 1889, p. 1. 4 CHORDATA. Arthropoda, in Annelida and in Mollusca we find, as we do, that the nerve ring referred to is, in the adult, incomplete behind the anus, and that the mouth and anus, though obviously referable to the blastopore, are not actually both derived from it, must we on this account deny this most obvious relation and maintain that the mouth or anus, as the case may be, in these forms is not homologous with that of Peripatus ? To maintain such a position appears to us impossible, and we entirely accept the doctrine that the mouth and anus of the Annelida, Arthropoda and Mollusca are both perfor- ations of the embryonic neural surface and are specialisations of parts of one original opening which is represented in most embryos by the blastopore. When however we come to apply this doctrine to the Chordata we stand upon more debatable ground. Placing the Enteropneusta on one side as not ob- viously conforming to our plan, we find that it is a fact of observation that in the Chordata the blastopore perforates the embryonic nerve rudiment, and that in some of them the anus is directly derived from it (many Pisces, some Amphibia, e.g. newt), whereas in others, not at all remote from these, the blastopore closes entirely and the anus is a new formation (some Pisces and Amphibia, e.g. frog, Amniota). Here also we think it may fairly be maintained that notwithstanding the diversity in the mode of development of the anus it is, in all Vertebrata at least, a derivate of the blastopore. The non- inclusion of the anus within the nerve rudiment in the adult, and its shift on to the ventral surface, cannot be brought against this view, because these facts apply both to animals in which the anus is a persistent part of the blastopore, as weh1 as to those in which it is a new formation. Here again, as in the invertebrate phyla already dealt with in this connection, the anus escapes in the adult from the embryonic nerve rudiment ; or to put it in another way the part of the nerve rudiment behind the anus never attains full development, but early undergoes atrophy.* So far then all is plain sailing, in the Vertebrata at least : the anus is a persistent portion — not the whole, as is clear from a consideration of the development of Elasmobranchs and some Amphibia — of the blastopore, as it is in the invertebrate * See especially Lepidosiren, in which the medullary folds of the embryo include the blastopore which becomes the anus. MOUTH. 5 Coelomata ; and as in most of the latter the part of the nerve rudiment behind it (in the primitive position, anterior in the position which the anus secondarily acquires on the ventral surface) undergoes atrophy. We now come to the question of the chordate mouth, a much vexed question, and one about which much of a highly specu- lative character has been written. We may at once concede 3-- Fm. 2. — Heads of young Elasmobranch embryos (Scyllium canicida) (after Sedgwick). A. Ventral view of head of embryo, 7 mm. in length, with two open pharyngeal clefts. The mouth is present as a longitudinal groove in the ectoderm of the buccal depression. B. Same view of a slightly older embryo ; the buccal groove has become a longitudinal slit. C. Side view of head of embryo, 9 mm. in length, with three open slits. D. Side view of head of embryo, 11 mm. in length ; rudiments of external gills have appeared on the hyoid and on the first and second branchial arches. E. Side view of head of embryo of 16 mm. ; external gills have appeared on mandibular arch and the angle of the jaw is marked. 1 mandibular arch ; 2 angle of jaw ; 3 second pharyngeal cleft ; 4 nasal pit ; 5 eye ; 6, midbrain ; 7, auditory sac ; 8 hyoid arch ; 9 spiracle. the point that the chordate mouth has never been brought into developmental relation with the blastopore. Even if it be allowed that the chordate blastopore really extends to the front end of the nerve rudiment (medullary plate), which is in itself a disputed point, no morphologist has ever brought to light any embryological fact which is at all in favour of the view that the mouth was originally within the nerve rudiment, 6 CHORDATA. and that its present position outside it and on the ventral surface is a secondary one, due to shifting and to atrophy of the part of the nerve rudiment in front of it. There are a number of features of vertebrate morphology of the highest interest in connection with this point : such are the cranial flexure, the close relation of the infundibulum, which there is good reason to believe is the real front end of the nerve axis, to the anterior end of the mouth, the slit-like form * (Fig. 2), which at first characterizes the buccal opening, and its extension into the rudi- ment of the pituitary body ; but there is no actually ascertained fact which tends to show that the mouth is a derivate of the blastopore, as it must be conceded to be in most other coelomate phyla, t The last chordate character to be considered is the posses- sion of lateral pharyngeal apertures. These are often used for respiration and are in consequence generally termed gill- slits. They are not however always respiratory in function — indeed in the majority of the Vertebrata in which they form a very conspicuous feature they are not respiratory at all, but are entirely functionless, being found only in the embryo. An attempt has been made in some quarters to refer the chordate mouth to a modification of a pair of these structures. We can see no fact in favour of such a view, and we are not prepared to give up the homology of the chordate mouth with that of other Coelomata. We have already stated the case with regard to its relation to the blastopore, and we have seen that there is no good embryological evidence in favour of its being so related, but we do not consider that this absence of evidence is sufficient to put out of court the view that it is the homologue of the mouth of other Coelomata. In many of these, too, no relation can be shown between the mouth and the blastopore in development, but yet we well know that in them the mouth is homologous with the mouth of forms in which it is directly derived from a part of the blastopore. Finally there is one point in the morphology of the Chordata, * Sedgwick, "Notes on Elasmobranch Development," Q.J.M.S., 33, 1892, p. 559. f For a fuller discussion of these questions, the reader is referred to the article "Embryology" in the recently issued supplement of the Encyclo- paedia Britannica. COELOM. which though not referred to in the definition, is of considerable importance, and must be shortly dealt with here, and that is the form and development of the Coelom. In the Cephalochorda and Enteropneusta the coelom originates as outgrowths of the primitive gut (archenteron). In Vertebrata, though there is no actual outgrowth of the enteron, the walls of the coelom originate from tissue which is derived from the wall of the enteric space, and there can be but little doubt that the mode of development is referable to the enterocoelic type, found in the two other phyla, and is indeed a modification of it. If then we leave out of consideration the Tunicata, which in this respect cannot at present be brought into line with the other chordate phyla, we may assert that an enterocoelic origin of the coelom, or a modification of it, is characteristic of the Chordata. Ou t s i d e the Chordata a similar mode of origin of the coelom is found in the A Chaetognatha (vol. i, B p. 590), in the Bra- chiopoda (vol. i, p. 580), probably in the Phoronidea (vol. i, p. 546), and as will be shown further on in the Echinodermata. But the coelomic resemblances between these animals go farther than this. In the Enteropneusta the archenteric outgrowths are five in number — two pairs and an anterior unpaired out- growth (Fig. 3). These, following the nomenclature of Bateson, have been named according to the position they occupy in the adult : the anterior unpaired sac is called the proboscis cavity ; the sacs of the anterior pair are the collar cavities ; and the posterior sacs are the trunk cavities. In the Enteropneusta they undergo no further division, but remaining in the parts of the body indicated by their names, they give rise to the coelomic spaces of the adult. In Amphioxus the anterior unpaired sac is called the preoral —4 FIG. 3. — Diagrams showing the origin and primitive relations of the coelomic sacs A in Balanoglossus, B in Amphioxus (after MacBride). 1 proboscis cavity in A, preoral cavity in B ; 2 collar cavity ; 3 anterior somite of trunk ; 4 trunk cavity. CHORDATA. or head cavity ; it obviously corresponds to the proboscis cavity and remains in the head region of the animal. The sacs of the second pair are called the collar cavities, because they correspond to those cavities of the Enteropneusta. They are in reality the anterior pair of somites, and give rise dorsally to the first pair of myotomes. Their exact disposition in the adult is not quite certain, but they appear to get some backward extension. The posterior sacs which come off as one pair from the enteron and correspond to the trunk cavities of Enteropneusta undergo in subsequent growth a segmentation and give rise to the whole of the mesoblastic somites of the trunk from the second pair back- wards.* In the development and arrangement of its coelomic sacs Amphloxus resembles in a remarkable manner the Entero- pneusta, the difference between them consisting in the segment- ation which the trunk cavities undergo in Amphioxus. In the Vertebrata, though it is not possible to point to such close resemblances as those which we have just described, there is a remarkable similarity in the embryonic arrangement. The first coelomic sac is preoral and unpaired ; the second is paired and large, extending backwards in the mandibular arch, so as to overlap the following somites. These mandibular cavities are clearly homologous with the collar somites of the other types. Following them we find on each side one large cavity, the dorsal parts of which are divided up into segments and become the myocoeles, and their walls the myotomes of the later embryo. These posterior cavities clearly correspond to the trunk cavities of the other types : as in them they are extensive, and occupy the whole trunk region, and as in Amphioxus they are metamerically segmented. In Amphioxus it is said that the preoral somite does not give rise to striated muscles ; in Vertebrata it gives rise to a considerable number of the eye muscles. With regard to the nonchordate phyla with enterocoelic coelonij we have only space to say this, that in the Echinoderms the Enteropneust plan of an unpaired anterior cavity and two pairs of posterior cavities can, according, to MacBride's researches, generally be made out ; that in the Chaetognatha there is an ap- proximation to the Enteropneust arrangement, but the unpaired * MacBride, Q.J.M.S., 40, 1898, p. 589. CHORD AT A. 9 cavity is at the hind end ; in the Phoronidea there are indications that the Enteropneust arrangement or a modification of it exists, but the indications are not very clear ; while in the Brachiopoda according to our present knowledge no resemblance to the Enteropneust plan exists save in the enterocoelic origin of the coelom. The formation of the coelom in the other chordate phyla, the Annelida, the Mollusca, and the Arthropoda, must be regarded as a modification of the enterocoelic method, but it is never possible in them to trace the arrangement into an unpaired chamber and two pairs of chambers which is so characteristic of the Chordata. CHAPTER II. PHYLUM CEPHALOCHORDA.* With dorsal tubular nerve-cord, and persistent notochord extend- ing forwards in front of the nerve-cord. The muscular system and gonads are segmented, and the pharynx possesses a large number of branchial slits which open into an atrial cavity and are provided with tongue bars. Without paired fins, jaws, brain, vertebrae and generative ducts. The larval life is prolonged and the larva is remarkably asymmetrical. The phylum Cephalochorda contains but the single genus, Amphioxus Yarrell. It was discovered by Pallas in 1778, who took it for a slug and named it Limax lanceolatus. Its true position in the animal kingdom was first recognized in 1834 by Costa, by whom it was named Branchiostoma. Two years later it was described by Yarrell, who called it Amphioxus, by which * J. Miiller, Ueber den Bau und die Lebenserscheinungen des Branchi- ostoma lubricum (Amphioxus lanceolatus), Berlin, 1844. Quatrefages, " Sur le systeme nerveux et sur 1'histologie du Branchiostome," Ann. des Sci. Nat. (3), 2, 1845. Kowalevsky, " Entwick, v. Amphioxus lanceolatus," Mem. Acad. Imp Sc., St. Petersbourg (7), 2, 1867. Id. " Weitere Studien, etc.," Arch. f. mic. Anat. 13, 1877. Stieda, " Ueb. d. Amphioxus lanceo- latus," Mem. Acad. Imp. Sc., Petersbourg, (7), 19, 1873. Rolph, " Ueb. d. Bau d. Amphioxus," Morph. Jahrb. 2, 1876. Langerhans, " Zur Anat. d. Amphioxus," Arch. mic. Anat., 12, 1876. A. Schneider, Anat. u. Ent- wick. der Wirbelthiere, Berlin, 1879. Hatschek, " Ueb. d. Entwick. Am- phioxus," Arb. a. d. Zool. Inst. Wien, 4, 1881, also Zool. Anz., 7, 1884, p. 517, and Anat. Anz., 3, 1888, p. 662. Rohon, " Ueb. Amphioxus lanceo- latus," Denksch. k. Akad. d. Wissenschaft, Wien, 45, 1882. Lankester, " Contributions to the knowledge of Amphioxus lanceolatus," Q.J.M.S., 29, 1889, p. 364. MacBride, " The Early Development of Amphioxus," Q.J.M.S., 40, 1898, p. 589. A. Willey, Amphioxus, etc., New York, 1894, and Q.J.M.S., 31, 1890, p. 445, and 32, 1891, p. 183. v. Wijhe, " Beitr. z. Anat. des Kopf region des Amphioxus," Petrus Camper, 1901, and Anat. Anz., 8, 1893. C. F. Cooper, " Cephalochorda," in J. S. Gardiner's Fauna etc. of Maldive and Laccadive Archipelago, 1, 1903, p. 347. R. C. Punnett, " Meristic Variation in Cephalochorda," Ibid., p. 361. HABITS. 11 name it has since been known. According to the strict rules of zoological nomenclature this is incorrect, the generic nam: Branchiostoma having two years' precedence over Amphioxus. But, as so often happens in human affairs, the unwritten law has triumphed over the written, and the almost universal custom of zoologists has been to call the genus Amphioxus. From this custom we shall not venture to depart in this work. Having thus entered our protest against a breach of con- ventional rule which is made knowingly, we had almost said wantonly, by all zoologists, we may proceed to consider the actual position in the system of this remarkable creature. Here fortunately there is no conflict between preaching and practice,, between a pedantic conformity to rule and a lawless adhesion to custom. For law and custom alike agree that the position of an animal in the system shall be determined by its natural affinities as revealed by a study of its structure and development. Judged by this test there can be no question that Amphioxus is closely allied to the Vertebrata and must be placed either within that group or in close juxtaposition to it. As our readers know we have adopted the latter 'course and have placed Amphioxus in a special phylum of its own, equal in morphological importance but very inferior in the number of its members to the great phylum Vertebrata, and have applied to it, out of a number of claimants,* the name Cephalochorda, in allusion to the extension of its notochord into the anterior part of the cephalic region. Amphioxus'f is a small, semi-transparent, colourless animal. Its body is elongated, laterally compressed, and pointed at each end ; and it may attain a length of two inches. . It is entirely marine, and is found at moderate depths in many parts of the world. It has a remarkable power of moving in sand, in which it is usually partially buried, its mouth alone protruding. But it is capable of swimming, and when removed from the sand bends its body with great activity from side to side. The mouth is an elongated oval aperture on the ventral surface immediately behind the anterior end of the body. It is sur- rounded by a number of delicate ciliated processes, the oral cirri. * Pharyngobranchii, Acrania, Leptocardii, etc. (• The anatomical description refers, unless otherwise stated, to A. !atit8. 12 PHYLUM CEPHALOCHORDA. The anus is also ventral and is placed slightly to the left of the middle line at some little distance from the posterior end of the body (Fig. 5). Extending from the mouth backwards along the ventral surface for about two-thirds of the length of the animal is a wide median groove, bounded by lateral folds and perforated at its hind end by a pore (Fig. 5). The folds are called the metapleural folds and the pore the atrial pore. There are no paired fins, but there is a continuous median fin consisting of a fold of skin extending along the whole length of the dorsal surface (dorsal fin), and round the hind end of the body on to the ventral surf ace 'as far forwards as the ventral groove (Fig. 4). Anteriorly it is also continued on the ventral surface, reaching as far as the mouth, with the right side of which it is continuous (Fig. 5). The portion between the ventral groove and the anus may be called the anal fin, and that between the anus and the hind end of the body the ventral part of the caudal fin. Amphioxus is a segmented animal. The segmentation is marked externally by a number of V-shaped grooves, placed one behind the other on each side of the body, the apex of the V being directed forwards (Fig. 4). These markings are caused by the insertion into the skin of a number of transverse septa of connective tissue, which divide the great lateral longitudinal muscles of the body into a series of successive segments, placed one behind the other and called myotomes. The grooves of the two sides of the body alternate with one another. The seg- mentation is also exhibited by the gonads which consist of a series of saccular bodies extending throughout the greater part of the pharyngeal region as far back as the atrial pore (Fig. 4). They correspond in number with the myotomes of that part of the body in which they occur and alternate with those of the opposite side of the body. The body of Amphioxus is traversed throughout almost its entire length by a flexible skeletal rod — the notochord. The notochord is pointed at either end and is placed in the centre of the body, but nearer to the dorsal than the ventral surface (Fig. 4). Lying immediately on the dorsal side of the notochord is a cord of nervous matter which may be called the cerebrospinal cord and constitutes the central nervous system. Behind, this nervous cord tapers and ends in a point, or a small PHYLUM CEPHALOCHORDA. 13 e< 14 PHYLUM CEPHALOCHORUA. swelling, immediately over the hind end of the notochord ; in front it tapers very slightly, and possesses a somewhat blunt termination placed some little distance behind the front end of the notochord. On the ventral side of the notochord is the alimentary canal, which has the form of a straight tube extending between the mouth and the anus. The central nervous system therefore, lies entirely dorsal to and the alimentary canal entirely ventral to the notochord. The alimentary canal consists of three parts : — 1. The buccal cavity. This is a short chamber opening to the exterior by the mouth and behind by a somewhat constricted opening into the pharynx. 2. The pharynx is the widest and longest portion of the alimentary canal, extending nearly half the length of the body. Its walls are perforated on each side by a number of obliquely directed slits (from above and in front ventralwards and back- wards) which place its cavity in communication with a space lying immediately outside it, and called the atrial or peripharyn- geal chamber. The atrial chamber entirely surrounds the pharynx except along the dorsal middle line (vide Fig. 10 and explanation). It opens to the exterior by the atrial pore — already mentioned — which is found at the hind end of the ventral groove (Fig. 5). The pharynx is mainly a respiratory organ, inasmuch as the blood which circulates in its walls and in the walls of the atrial cavity is aerated by the water which is con- tinually being taken in by the mouth and driven by the action of cilia through the pharyngeal slits — or gill slits as they may be called — into the atrial cavity. 3. The intestine which extends as a straight tube from the hind end of the pharynx to the anus. The anterior part of the intestine is slightly dilated and receives ventrally a simple caecalsac, which, pushing the body wall before it, extends for- wards in the atrial cavity on the right side of the pharynx and is called the liver. Detailed Description of the Organs. The ectoderm consists of a single layer of columnar or in some places cubical cells, which cover the whole external surface of the animal, are prolonged for a short distance into the buccal cavity and line the whole of the atrial cavity (Fig. 10). They SKELETON. 15 are without cilia except on the cirri, in the mouth and in the atrial cavity, and their outer surface is covered by a porous cuticle. Immediately beneath the ectoderm is a layer of fibril- lated tissue called the cutis. Beneath this comes the subcu- taneous tissue which consists of a gelatinous matrix containing sinuous fibres. The tissue within this has a similar form and extends between the myotomes, as the inter- muscular septa, to become continuous with the sheath of the notochord. In fact all the connective tissues of the body may be said to form a continuous framework which supports the organs and is on the whole of very similar structure throughout. In some parts it is firmer than in others and in some places it contains fibres, but it never presents a modification of a cartilaginous or osseous nature and never, except at the ventral ends of the primary pharyngeal bars, contains cells other than the epithelial cells which bound the spaces contained within it. These spaces are in some cases vascular and in others coelomic, but it is not possible in every case to be certain as to which of these two organs they belong. This absence of what we may call mesenchymatous elements from the connective and supporting tissues is one of the most remarkable peculiarities of Amphioxus. As skeletal tissue we may rank the notochord, the supporting tissue of the buccal ring and the axial tissue of the buccal cirri, possibly also the axial tissue of the pharyngeal bars. The notochord is made up of a number of discs placed verti- cally, and transversely to the long axis, and consisting of gelatinous tissue. It is surrounded by a tough sheath of connective tissue, which is continuous with the rest of the connective tissue framework of the body. Nuclei are present on the dorsal and ventral sides in the neighbourhood of two spaces, the so-called dorsal and ventral lymph canals of the notochord. The edges of the mouth contain a ring of skeletal tissue the buccal ring, resembling the notochord in structure. It consists of about twelve pieces on each side, and each piece gives attach- ment to a rod of the same substance, which occupies the axis and forms the support of one of the oral cirri. The tissue of this buccal skeleton consists of a number of gelatinous discs surrounded by a fibrous sheath. By some observers it is claimed as cartilage, each disc being a cell and the surrounding membrane the 16 PHYLUM CEPHALOCHORDA. cartilaginous matrix. The tissue of the axial rods of the pharyngeal bars is sometimes described as skeletal. It consists of a clear chitin-like substance devoid of cellular structures except in the ventral bifurcated parts of the primary bars, which contain branched cells. Excluding the nuclei of the notochord, which is an endodermal structure, this is the only instance of a mesodermal tissue containing mesenchymatous elements. The nerve-cord is surrounded by a tough sheath, which is continuous with the sheath of the notochord. The fin-rays are found in the greater part of the dorsal fin in a single series, and in the anal fin in a double series (Figs. 4 and 5). They are absent from the cephalic fin, and from the anterior and posterior portions of the dorsal fin, and from the ventral parfc of the caudal fin. They consist of small cubical pieces of a tough fibrous connective tissue, which in the dorsal fin are continuous with the fibrous investment of the nerve- cord. They are more numerous than the metameres, there being four or five to each muscle segment. The fins contain a longitudinal canal lined by epithelium and divided by septa into compartments. In the region of the fin-rays each of these compartments contains a fin-ray which pushes in its ventral wall and projects into it. The nature of these fin spaces is not known, but it is stated by Hatschek and others that they are, in the dorsal fin at least, coelomic in origin, being derived from the coelom of the muscle-plate (see below). The muscular system consists of striated and unstriated muscular tissue. The striated muscles are composed of fibrillated rhombic plates, and are devoid of sarcolemma. They con- stitute the lateral muscles, the transverse muscles, the muscles of the lips and cirri, and the sphincter muscle of the velum and anus. The great lateral muscle of the body is divided up into a number of successive segments, the myotomes (myomeres) by septa of connective tissue. These septa have a peculiar V-shaped course, and their insertion into the skin causes the V-shaped external markings already referred to (Fig. 4). The myotomes of opposite sides alternate with one another, i.e. the intermuscular septa of one side are opposite the middle of the myotomes of the other side. The full number of myotcmes is laid down in the embryo. In Amphioxus lanceolatus there are about sixty-two on each side. The plate-like n, res of which these muscles are composed extend the whole length NERVE CORD. 17 of the myotome from septum to septum. The other striated muscles are very similar to the lateral muscle in structure, but the cross-striation is less marked, and they are not segmentally arranged, nor derived from the myotomes of the embryo. The transverse muscles extend from the ventral end of the lateral muscles to the middle line of the floor of the ventral groove, where they are inserted into a median connective tissue septum (Fig. 10). The unstriped muscle confers contractility on the walls of the intestine and larger blood-vessels. It is exceedingly inconspicuous and thin. It is doubtful indeed if it really exists as a distinct tissue. The nervous system. The nerve-cord (cerebro - spinal cord) contains in its ventral portion a small circular cen- tral canal, which extends as a narrow fissure to the dorsal summit of the cord (Fig. 6). This canal is lined by colum- nar epithelial cells, some of which are continued into the substance of the cord as sup- porting fibres, while others may have the form of nerve cells. The cells lining the dorsal part of the canal are in close contact, so that the cavity here is virtual. The nerve cells are for the most part placed in the central part of the cord, and some of them are of giant size,* and extend right across the fissure-like part of the central canal. On the ventral side of the canal, at short (metameric) intervals along the whole length of the cord behind the second myotome, are small groups of black pigmented cells, f These structures are probably sensitive to light. There is also a pigment spot, commonly called the eye, and placed at the front end of the cord in the anterior wall of the cerebral vesicle. * For the arrangement of these giant nerve-cells and of the giant fibres which issue from them, we refer the reader to Rohde, in Zool. Beitrage, 2, 1888, p. 169. f Hesse, Z. f. w. Z., 63, 1898, p. 456. Z. ii. C FIG. 6. — Transverse section of the spina I cord of Amphioxus (after Rohde). gz nerve cells ; stf supporting fibres ; ck cen- tral canal. 18 , PHYLUM CEPHALOCHORDA. Anteriorly, though the cord itself tapers slightly, the central canal widens out into a spacious vesicle, the cerebral vesicle. This is the only representative of the brain of the Vertebrata. It gives off from its front dorsal wall a small hollow diverticulum, which ends blindly against the inner side of a small asymmetrical pit of ectoderm called the olfactory pit. This pit is placed on the left side of the body, is lined by ciliated cells, and is supposed to be olfactory in function. In the larva the process of the cerebral vesicle and the pit are in communication by a pore, wln'ch appears to be the persistent neuropore of the embryo. The cerebral vesicle is also said to possess a ventral diverticulum in the hinder part of its floor, which has been com- pared with the infundibulum of the vertebrate brain, but it is doubtful if it is always present. Posteriorly the cord tapers considerably and ends, usually in a small swelling, just short of the hind end of the notochord. In some cases it is said to extend as a filament round the hind end of the notochord on to the ventral surface. The nerves issuing from this cord, except in the case of the first two, are dorsal and ventral in their origin and correspond with the dorsal and ventral roots of the spinal nerves of the Vertebrata. In Amphioxus they do not join and there are no ganglia on the dorsal roots. The dorsal nerves arise from the dorsal part of the cord. The nerves of the first two pairs (so-called cranial nerves) arise opposite one another, altogether in front of the myotomes. They are entirely sensory, and supply the preoral part of the body. Their finer branches possess nerve cells not far from their terminations. The remainder of the dorsal nerves alternate in their origin on the two sides, and contain motor as well as sensory fibres. They pass out behind the myotome to which they belong, and divide in the subcutaneous connective tissue into dorsal and ventral branches. The ventral of these supplies nerves to the transverse muscle of the sub-atrial floor. The dorsal nerves also supply the muscles of the lips and velum. The sides of the mouth (oral hood) and the cirri are supplied by the third to seventh dorsal nerves, and the nerves to the inner side of the lips of both sides are derived from the dorsal nerves which arise on the left side of the cord (reminiscence of larval con- dition, see below). The velar supply comes from dorsal nerves four to seven. ALIMENTARY CANAL. 19 The ventral nerves (Fig. 7) are not united in bundles nor surrounded by a sheath, but issue as linear groups of fine fibres, which pass immediately to the adjacent myotome. They arise slightly anterior to the dorsal nerve of the segment to which they belong, and are exclusively motor for the muscles of the myo- tomes. Sense organs, Scattered amongst the ecto- derm cells are cells bearing short hairs. They are specially numerous at the front end of the body and round the mouth. They may be supposed to be tactile organs. They are also found in the mouth and on the velum. Organs which are supposed to be visual and olfactory have already been described. The most striking peculiarities of the ner- vous system of Amphioxus as compared with that of the Vertebrata are the absence of an anterior cerebral enlargement ; the peculiar form of the central canal, the dorsal fissure-like por- tion of which is probably represented by the FIG. ?.— Anterior en i dorsal fissure of the vertebrate spinal cord ; the absence of a junction between the dorsal and ventral nerve roots and of a ganglion on the dorsal roots ; and, lastly, the imperfect condition of the organs of special sense. IV is a ches of the second nerve arises indepen- dently from the cord in the preparation after ys f 'anc2" It leads into a spacious fJn(J secz°A/d/n ™ The alimentary canal. The mouth large, somewhat circular opening, placed on the ventral side of the body, a little distance ts0crhyne from the front end. cavity, the buccal cavity, which is bounded by a fold of the integument called the oral hood (Fig. 4). The free edge of the oral hood, which may be called the lip, contains the skeletal framework already described, and bears a number (from twenty to thirty, increasing with age) of delicate ciliated processes, the oral cirri. The right side of this oral hood is, as has already been mentioned, continuous with the preoral ventral part of the median fin, which is in accordance with what might be expected 20 PHYLUM CEPHALOCHORDA. from the development of the parts (see below). The buccal cavity is bounded behind, at the level of the anterior angle of the seventh myotome, by a muscular membrane, the velum, which separates it from the pharynx and is perforated by an aperture, which has been sometimes called the true mouth and is the actual opening which formed the mouth in the larva. The edges of the velum around this opening bear twelve delicate tentacles, the velar tentacles, which project backwards into the pharynx. The buccal cavity is lined by a ciliated epithelium, and bears on its roof slightly to the right of the median line a pit called Hatschek's pit, or simply the preoral pit (Figs. 8, 9). This organ is lined by a columnar epithelium, the cells of which bear stiff sensory hairs. It opens in a groove of columnar finely ciliated epithelium, which extends a short distance in front of it in the roof of the mouth and behind bifurcates into two ciliated grooves ; these pass obliquely backwards and outwards to the velum, and then pass ventrally along the junction of the velum and sides of the mouth. They terminate either by simply coming to an end, or by running into one another on the floor of the buccal cavity. These two diverging ciliated furrows give off access- ory furrows, which are formed by folds of the antero- lateral walls of the main fur- rows. The ac- cessory ciliated furrows pass for- wards for a short distance on the roof and sides of the ion12 mouth. The FIG. 8.— View of the roof of the buccal cavity of Amphioxus m^^\^ lanceolatus from below (after v. Wijhe). 1 posterior wall WnO of buccal cavity ; 2 ciliated groove of left side ; 3 left side «^n0f,'f«f«0 +^~ of preoral hood ; 5 notochord ; 6 second myotome ; 7 Hatschek's (preoral) pit ; 8 ciliated pit ; 9 right side of « ,T/uQ~i ~__ „ j» preoral hood ; 10 third myotome ; 11 ciliated groove of Wile organ right side; 12 accessory ciliated groove; 13 sphincter ^f r Tvn;il™ T* muscle of velum ; 14 velar tentacles. ot J • Mllller. -14 PHARYNX. 21 is an organ for creating currents in the mouth back to the pharynx. The preoral pit which opens into its front part has been variously interpreted as a sensory organ and a gland. If the preoral (Hatschek's) pit be regarded as a gland, it has been sug- gested by v. Wijhe that it is comparable to the neural gland of the Tunicata, and that the ciliated groove is comparable to the opening of that gland, the dorsal tubercle, the edges of which are frequently drawn out in a manner --20 — 17 FIG. 9. — Transverse section through the middle of the buccal cavity of Amphioxus lanceolatus to show the preoral (Hatschek's) pit and the ciliated furrow (after v. Wijhe). >1 Hatschek's pit ; 2 right side of preoral hood ; 3 outer lip-cavity ; 4 inner lip cavity (3 and 4 arejparts of the left collar somite of the embryo) ; 5 labial nerve ; 6 coelom (dorsal buccal, a portion of the collar somite of the embryo) ; 7 contorted blood vessel (continuation of right aorta) ; 9 aorta ; 10 Hatschek's nephridium (a process from the pharynx) ; 11 buccal cavity ;| 12 internal labial muscle ; 13 skeleton of cirrus ; 14 cavity (lymph space) of cirrus ; 15 exter- nal labial muscle ; 17 left side of preoral hood ; 18 second, 19 third, 20 fourth myotome. very similar to the course of the ciliated groove of Amphioxus. Against this interpretation we must set the fact that the preoral pit, whatever its origin (see below), has no relation to the central nervous system. Moreover the interpretation of it as a gland is a very doubtful one. Van Wijhe describes a special part of the ciliated furrow just behind and in close connection with Hatschek's pit as the ciliated pit (Fig, 8, 8). The pharynx is a large chamber tapering slightly posteriorly on account of the dorsal inclination of its ventral wall. Its side 22 PHYLUM CEPHALOCHORDA. walls are perforated throughout their whole dorso- ventral extent (except for a short distance behind the velum) by a number of vertically directed slits, which have a slight inclination ventralwards and backwards (more marked in the preserved than in the living animal), and which open into the atrial cavity. The dorsal and ventral parts of the pharyngeal wall are not perforated and constitute the hyperpharyngeal groove and the endostyle (hypopharyngeal groove) respectively. These termi- nate independently of one another posteriorly, but anteriorly they are connected by the peripharyngeal ciliated bands which arch round the pharynx immediately in front of the gill -slits. In front .of the peripharyngeal bands there is a small portion of the pharnyx adjacent to the velum without gill-slits. The pharynx is lined by a ciliated epithelium, which is con- tinuous through the slits with the ectodermal epithelium lining the atrial cavity. 'Along the endostyle there are four bands of specially glandular cells, which, like the remaining pharyngeal cells, bear cilia, and secrete the mucus, which passing forwards along the endostyle is driven upwards by the peripharyngeal bands into the front end of the hyperpharyngeal groove. Along this it is carried by ciliary action into the stomach. The food, consisting of smaU floating organic bodies brought into the pharynx by the ciliary currents, is entangled in this mucus and so separated from the water which passes through the gill-slits into the atrial cavity and out by the atrial pore. New gill-slits continue to be formed long after the development has ceased, during the growth of the animal. They are conse- quently more numerous in large than in small specimens. In large specimens there may be as many as 180 secondary (see below) or 90 primary gill-slits on each side. The new slits are formed at the hind end of the pharynx close to the junction with the stomach, as small circular perforations (see Fig. 24 in the account of the development). These soon become partly divided into two by the growth ventralwards of a process from the dorsal wall of the aperture (Fig. 25). This downward projection, from its resemblance to the tongue of a Jew's harp, is called the tongue bar. It eventually fuses with the ventral wall of the slit, so that the primary slit becomes completely divided into two secondary slits separated by the tongue bar. In correspondence with this we may call the parts of the pharyn- GILL-SLITS. 23 geal wall intervening between two primary slits the primary $. li ci "\f /fltU FIG. 10. — Diagram of a transverse section through Ampkioxus in the hinder part of the pharyngeal region showing the brown canals (after Lankester, from Perrier) . The division of the more ventral portion of the myotomes into two groups of fibres separated by a connective fascia is shown ; a atrium ; ad root of dorsal aorta ; b primary bar ; ft1 secondary bar ; c liver (caecum) ; ca lymph space ; cd notochord ; e^'tube of atriocoelomic funnel ; en connective tissue framework of the body ; c p SQmatic wall covering caecum ; ct t skeletal plate of endostyte ; e ectoderm ; / branchial slit ; g$ hyper- pharyngeal groove ; gv endostyle •, li ventral canal of notocljorol ; Is dorsal canal of the notochord ; m transverse muscle of atrial floor ; M myotomes ; mt metapleure ; AT spinal cord ; na ventral, np dorsal root of a spinal nerve ; o go*n.ad ; p portion of a t Id in dorsal wall of atrium ; ph pharynx ; r dorsal fin raj' ; s' so-called lymph-'spaces. i the case of the hepatic caecum, s points to the blood-vessels. 24 PHYLUM CEPHALOCHORDA. bar. The primary gill-slits of opposite sides alternate with one another, as do the myotomes. In the adult they are more numerous than the myotomes, though when they first make their appearance, they correspond with them. The anterior primary slit on each side is not divided by a tongue bar. On account of the obliquity of their direction a great many, both of primary and tongue bars, are cut in transverse section (Fig. 10). The primary bars differ from the tongue bars in structure (Fig. 11, A, B). In both there is an axial rod of a clear chi tin-like substance (insoluble in potash), placed nearer the atrial than the pharyngeal side of the bar. In the primary bar this rod is double and without a cavity, while in the tongue bar it is single and has a cavity. These rods are continued inwards towards the pharynx as a thin membrane called the septal membrane (Fig. 11). The character of epithelium covering the bars may be gathered by an inspection of Fig. 11. On the outer side the ectoderm of the atrium is found. It is said to be non-ciliated and to be separated from the more extensive pharyn- geal endoderm by some pigmented cells. The cilia of the endo- derm vary in character on different portions of the bar, recalling the condition found in the Lamellibranch gill. The primary bar con tains between the ectoderm and the skeletal rod a chamber which is a portion of the coelom, being continued dorsally into the dorso-pharyngeal coelom, and ventrally into the endostylar coelom. In addition three blood-vessels can be seen in the primary bar in the positions and with the names indicated in the figure. In the tongue-bar the visceral (10) and somatic (7) vessels are present, and in addition there is a space in the skeletal rod (3). This is interpreted by Lankester and Benham as coelomic, by others as vascular. According to Benham this space contains a blood-vessel (omitted in the figure) corresponding to the external blood-vessel of the primary bar (4). Successive primary bars are connected by transverse bars (synapticuld), which thus pass across the primary slits, internally to the tongue bars, with which, however, they are connected (Fig. 12,5). The primary bars may thus be distinguished from the tongue bars in transverse section ; and they may be distinguished by in- spection of the pharynx as a whole for the skeletal rod of the primary bars bifurcates ventrally on reaching the level of the en- dostyle, whereas thai of the tongue bars does not bifurcate. Dor- INTESTINE. 25 sally the skeletal rods of both bars bifurcate and arch over the clefts to join the branches of the adjacent bars. Below the endostyle are some flat plates of skeletal tissue, which partially overlap one another. They correspond in number with the primary slits at the lower end of which they are placed, FIG. 11. — Transverse sections at right angles to the length of the pharyngeal bars of Amphioxus lanceolatm, A of a tongue bar, B of a primary bar (after Benham slightly altered). The relative sizes of the two bars is maintained. 1 endoderm epithelium on the pharyngeal end of a bar ; 2 pigment cells ; 3 coelom of the primary bar in B, coelom or external blood- vessel contained in the skeletal rod of the tongue bar in A (coelom according to Lankester and Benham) ; 4 external blood-vessel of the primary bar f'5 atrial epithelium ; 6 skeletal rod ; 7 somatic blood-vessel ; 8, 9 septal membrane ; 10 visceral blood-vessel. The pharynx opens posteriorly into the intestine which' runs straight back to open by the anus placed a little to the left of the median line at the level of the septum between the 51st and 52nd myotome. The anus is provided with a sphincter of striated muscular fibres. The intestine is lined by a columnar ciliated epithelium. The anterior part of the intestine is slightly dilated and called the stomach. It gives off a forwardly directed 26 PHYLUM CEPHALOCHORDA. diverticulum, the liver, which pushes before itself the ventral body wall, the whole projection lying in the atrial cavity along- side the pharynx on its right side (Fig. 10). This process is attached to the dorsal wall of the atrium in front. The lining cells of the hepatic caecum are coloured green in the living animal, as are the cells of the stomach from which it is given off. Outside the intestinal epithelium is a thin connective tissue layer which may contain unstriated muscular fibres. The atrial cavity is a space lined by ectoderm and surround- ing the pharynx and anterior part of the intestine ventrally and laterally. It opens to the exterior at the hind end of the ventral groove at the level of the 36th myotome. It extends back on the right side behind the atrial pore almost as far as the anus. Its lining cells are in part ciliated and pigmented with a brown pigment. The dorsal wall of the atrium is folded, in consequence of the fact that it is reflected on to each of the primary bars at a point considerably ventral to that at which it joins the secondary bars (Fig. 12, Id). This gives rise to somewhat puzzling features in transverse sections in which the dorsal regions of the primary bars appear to be connected to the side walls of the atrium by strands of tissue traversing the atrium (Fig. 10). The ligarnentum denticulatum of J. Miiller is this folded roof of the atrium cut longitudinally. It follows from this arrangement that the atrium is prolonged further dorsally in the region of the tongue bars than it is over the primary bars. The brown canals are two tubes lined by a pigmented epithe- lium, and projecting into the dorso-pharyngeal coelom (Fig. 10). They lie parallel to the long axis of the body and probably end blindly in front at the 27th myotome. Posteriorly at the level of the junction of the pharynx and intestine they open by funnel-shaped apertures into the dorsal part of the atrial cavity, one on each side. They are to be regarded as forwardly directed diverticula of the atrium, of unknown function. These structures are often called the atrio-coelomic funnels. They were discovered by Lankester. It is doubtful whether they end blindly in front or open into the dorso-pharyngeal coelom. There is a well- developed coelom with which the gonads are in relation. A complete comprehension of the coelom cannot . EXCRETORY ORGANS. 27 be obtained until the development is studied, but the main features in its topographical arrangement seem fairly clear, and may be described at this point. In the adult there are many spaces in the tissues the exact nature and origin of which is not understood. Such will be referred to by the general term lymph spaces. It has been stated by some observers (Schneider, Lankester and others) that the coelom and vascular system are in certain parts of the body continuous. But having regard to the doubtful character of many of the body-spaces above referred to and to the difficulties to investigation presented by the vascular system, this statement cannot be accepted without further evidence. In the region of the intestine there is a perivisceral cavity which is coelomic. It entirely surrounds the intestine except dorsally, where it is interrupted by the mesentery. In the region of the pharynx the same cavity is found, but it is broken up by the gill-slits into a number of parts all continuous with each other. There are two dorsolateral chambers, one on each side of the hyperpharyngeal groove. These extend a little way outward in the lateral walls of the atrium, and dip down into the folds of its roof along the primary bars. They constitute the dorso- pharyngeal coelom, and are continuous through the above- mentioned folds with the coelom present on the outer side of each primary bar. Ventrally the coelom of the primary bars opens into a median ventral chamber below the endostyle, called the endostylar coelom. The arrangement of the coelom about the mouth, which has been described by v. Wijhe, is too complicated for description in this work : it is, however, referred to in the section on development. Excretory organs. In the dorsal wall of the atrium lying between the atrial epithelium and the epithelium of the dorso-pharyngeal coelom are a number of tubes with a ciliated lining, which have been supposed to be renal in function (Figs. 12, 13). They correspond in number and position with the primary gill-slits and do not extend behind the region of the pharynx. They open into the atrium opposite the dorsal end 'of a tongue bar and at the summit of one of the dorsal pouches of the atrium found at that point (p. 26). They divide into two canals close to the opening ; one of these passes forward and then turns round to travel for a short distance ventral- wards ; the other passes backwards. They possess a variable number of branches (from 1-5), the number being least at the two ends of the series and greatest in the middle of it. Curious fibres ending in small knobs, each of which contains a nucleus, pass off from the ends of these branches and project into the coelom. These fibres are fine tubes ending blindly internally and opening into the secretory tube. They contain a long 28 PHYLUM CEPHALOCHORDA. vibratile flagellum, arising from the protoplasm around the nucleus at the internal knobbed end and extending along their whole length so as to project into the excretory tube (Fig. 13). They have been termed soleno- cytes from their resemblance to the fibres found on the excretory tubes of some invertebrates (e.g. polychaetous annelids). The tubes them- selves are lined by a ciliated epithelium and a tuft of specially long cilia projects through the renal opening into the atrium. They receive a special vascular supply from the dorsal ends of the pharyngeal vessels, the blood being returned into the adjacent aorta. That these organs are excretory is inferred from their structure, which, as stated, closely resembles that of the excretory organs of some polychaetous annelids, etc., among the invertebrates, and on account of Weiss' experiments. He fed the 7ns ^ ----nc Jr^ — nfc ^T |^ 11 ' FIG. 12. — Dorsal portion of the left pharyngeal wall of Amphionus, showing three rena canals, on one of which the solenocytes are shown ; seen from the side, diagrammatic (from Korschelt and Heider after Boveri). Id optical section of the folded ventral wall of the dorso-pharyngeal coelom (ligamentum denticulatum) ; m myotome ; ms intermuscular septum ; nc termination of the branches of the renal tube with the solenocyte? removed ; nk renal canal ; np opening of the renal canal into the atrium ; s synapticulum ; I primary gill-bar ; II tongue bar. animal with finely- divided carmine, and then found that the cells of the main tubes contained carmine particles. But he also found carmine in the lining cells of the atrium including those of Miiller's papillae. It is possible, of course, that Weiss' interpretation of these facts is correcf, and that the carmine found in these cells was in the act of being excreted from the system into which it had been taken by the intestinal epithelium, but on the other hand it may well be that the carmine entered the cells concerned from the atrial cavity directly, in the same way that, according to Weiss' view it must have entered the intestinal walls.* * Vide Weiss, Q.J.M.S., 31, 1890, p. 497 ; and Boveri, Zool. Jahrb., 5, 1892, p. 429. Goodrich, Q.J.M.S., 45, 1902, p. 493. VASCULAR SYSTEM. 29 The vascular system. There is no heart, but the larger vessels are peristaltically contractile. Lying on the ventral side of the endostyle in the pharyngeal wall there is a longitudinal sub -pharyngeal vessel, which corresponds to the heart and ventral aorta of the Vertebrata. Anteriorly it terminates by branching to the lips. Laterally all along its course it gives off branches, which have on their bases small contractile swellings called bul- bils, to the primary bars. These ascend dorsalwards and open into the aortic root of their own side. The secondary bars receive their blood supply from the primary through the transverse bars. The roots of the aorta lie on each side of the hyperpharyn- geal groove (Fig. 10). In the intestinal region they unite to form the single dorsal aorta, which gives off branches to the intestine and lateral body walls. Both the aortic roots are con- tinued forward as the carotid arteries. These are joined to- gether by a transverse anasto- mosis, and the right vessel gives off a large much convoluted branch which passes ventral- wards at the level of the velum and ends blindly (Fig. 9). The anterior part o*f the right carotid has the form of a plexus, and gives off branches to the oral cirri of both sides. The intestinal blood is collected into a sub-intestinal vein lying in the ventral wall of the intestine. This vein is not a simple vessel, but consists of a plexus of vessels, frequently communi- cating and lying side by side. Anteriorly the sub -intestinal vein appears to break up into a capillary system on the wall of the hepatic caecum. The blood of the caecum is collected into the FIG. 13. — Portion of the excretory canal of a young Amphioxus with its soleno- cytes, from the living animal (after Goodrich). 1 solenocyte ; 2 tube of solenocyte ; 3 excretory canal ; 4 flagellum of solenocyte. 30 PHYLUM CEPHALOCHORDA. vein, the hepatic vein, which like the sub-intestinal vein consists of a plexus of vessels (Fig. 10). These commence in front and behind unite to form a single vein which is continued into the hind end of the sub-pharyngeal vessel. The hepatic vessels are said to communicate at the front end of the caecum with the dorsopharyngeal coelom, but this statement must be accepted with caution. There is undoubtedly a connection between the anterior end of the caecum and the lateral wall of the atrium. The blood is colourless ; it contains amoeboid cells, and accord- ing to some observers a few red oval corpuscles. In the larva the hind end of the subintestinal vein is continued directly into the sub-pharyngeal vessel. The direction of the flow of blood is forward in the subintestinal vein and subpharyngeal vessels, both of which are, according to J. Miiller, contractile. It follows from this that the flow must be dorsalwards in the pharyngeal bars, backwards in the dorsal aorta and ventralwards in the peri-intestinal vessels. There is a considerable vascular development in the lateral walls of the atrium and a longitudinal vessel runs along the line of the gonads, but how these and the body- wall vessels generally are related to the main trunks described above is not known. The other spaces of the body may be classed as lymph-spaces. They are lined by an epithelium and contain a coagulable fluid. Their origin and relations are not certainly known. Some of them, e.g. the lymph canals in the fins and certain spaces within the myotomes are said to be coelomic and derived from the mesoblastic somites of the embryo. Others may possibly be purely vascular ; e.g. the large canal found in each meta- pleure — the metapleural lymph canals. Generative organs. The sexes are separate. There are no external sexual differences. Generative ducts are absent, and the generative organs, which are segmented in cor- respondence with the myotomes and are placed in the lateral wall of the atrium at the ventral ends of the myo- tomes, dehisce their products into the atrial cavity by rupture of their walls. From the atrium the generative products pass to the exterior usually through the atrial pore, but in some cases, according to the observations of Kowalevsky and Hatschek on the living, and of Marshall on the preserved animal, they occasionally pass from the atrium through the gill-slits into the pharynx and are spawned by the mouth. Spawning takes place at sundown only and fertilization is effected either in the sea or in the atrium. In a fully-developed Amphioxus lanceolatus the gonads are somewhat cubical bodies, twenty-six in number on each side. THE EGG. 31 The first of them appears to be placed at the ventral end of myotome 10, and the last at the ventral end of myotome 35 just in front of the atriopore. They are contained in coelomic sacs, which in development are derived from the ventral ends of the myotomes, and to the wall of which they are attached. The phylum possesses only a single genus* Amphioxus Yarrell (Branch- iostoma Costa). It is found in all seas. About ten species are known. They differ in the number of myotomes, the presence or absence of a caudal expansion of the median fin, the presence of gonadial sacs on one or on both sides of the body, the continuity of the right metapleur with the anal fin or the cessation of both right and left metapleur behind the atrial pore ; the presence or absence of fin-rays and fin-ray spaces in the anal fin. There appears to be a considerable range of meristic variation in some of the species (Punnett). A. lanceolatus Pallas, Europe and most seas ; A. bassanum Giinther, right metapleur continuous with anal fin, gonads on right side only, anal fin with fin-rays and fin-ray spaces, Bass Straits ; A. cultellum Peters, like the last, anal fin with chambers but without rays, Torres Straits ; A. lucayanum Andrews, like the last, but without caudal fin, hind end of body being a urostyle-like process without myotomes, and fin-rays and chambers absent from anal fin, Bahamas. It has been suggested that one or two species are pelagic, but this is uncertain. Development. The development of Amphioxus presents some remarkable features, and contrary to our usual custom we have decided to give a full account of it in this work. Though strangely similar in many of its features to the type of develop- ment found in the Vertebrata, it presents some very marked features of difference. Of these we may at once mention the small size of the ovum, the archenteric origin of the coelom, the absence of any nephridial apparatus comparable to that of the Vertebrata, the origin of the gonads from the myocoel, and the extraordinary asymmetry of the larva. Our account is based upon the important researches of Hatschek and Willy, who, as is well known, worked at the species found in the Pantano at Faro in Sicily. The egg is small ('1 mm. in diameter), is surrounded by a vitel- 1'ne membrane and contains but a small quantity of yolk, which is uniformly distributed. Only one polar body is attached to the ovum after deposition ; it is probable that this is the second, * By some zoologists the phylum has been broken up into genera and rkaldy, Q.J.M.S., 37, 1895, p. 303), but this, considering er of species and points of difference, seems hardly necessary 32 PHYLUM CEPHALOCHORDA. the first having been formed in the ovary and rubbed off during the dehiscence. The segmentation is complete and almost equal , the segments of the lower pole being slightly larger than those of the upper. It leads to the formation of a hollow blast o- FIG. 14. — Cleavage of Amphioxus (after Salensky from Korschelt and Heider). A egg before cleavage, with polar body ; B two-cell stage ; C four-cell stage ; D the same seen from the upper surface ; E eight-cell stage ; F sixteen-cell stage ; G stage showing more rapid division at the animal pole ; H the same in section ; / surface view of blastosphere. sphere (Fig. 14). This becomes invaginated to form a cup- shaped gastrula (Fig. 15). The blastopore, at first wide, snrm narrows to a small opening placed at the hind end of the future dorsal surface. The embryo now elongates in the direction of the antero-posterior axis (Fig. 16), and the ectoderm of the dorsal MEDULLARY CANAL. 33 surface becomes more columnar to form the medullary plate (Fig. 17). As the medullary plate extends to the hind end of the dorsal surface, the blastopore is included in it. The lateral part of the ectoderm now be- comes detached from the medullary plate, and grows over it (Fig. 17). This overgrowth begins at the hind end of the medullary plate, so that the blastopore is covered over and comes to open into the space between the overgrown ectoderm and the medullary plate (Fig. 18). I Later the medullary plate curves over dorsally (Fig. 17), and by the junction of its two lateral edges forms the walls of the medullary canal (Fig. 26). From what has been said it is clear that the medullary canal, which is gradually developed from behind forwards, opens posteriorly into the archenteron by the the medullary canal becomes the central canal of the nervous system, the blastopore is henceforward known as Hhe neur enteric canal. It closes soon after the commencement of larval life. The anterior neuropore persists throughout the blastopore and anteriorly to exterior by the neuropore. As 0 FIG. 15.— Formation of the gastrula of Amphioxus (from Claus, after Hatschek). A blas- tosphere ; B commencing in- vagination of the lower surface of the blastosphere to form the endoderm ; C later gas- trula ; all in optical section ; 0 anterior lip of blastopore. The cilia of the ectoderm are omitted. larval life and only closes on the attainment of the adult state. It marks the site of the olfactory pit. Meanwhile two pairs of dorsb- lateral outgrowths of the archen- teron are formed (Figs. 17, 18). The anterior of these retain their communication with the archen- •hdtrfanTHeidiear)SChek teron for some 'time and give rise D 34 PHYLUM CEPHALOCHORDA. to the somites of the first pair. They are the " collar " somites of MacBride (Fig. 19, 6'). The posterior outgrowths constitute the coelomic grooves of MacBride. They remain open to the gut for a considerable time behind, but as growth progresses they are continually constricting off sacs anteriorly (Fig. 19, 5). Eventually when about fourteen pairs of somites have been formed they become separate from the endoderm of the arch- enteron and form a solid plate on each side from which the remainder of the somites are successively developed. In addition to these two pairs of archenteric outgrowths there is a median anterior outgrowth (Fig. 19, /). This grows back on each side FIG. 17. — Transverse section through two embryos of AmpMoxus to show the enclosure of the medullary plate and the formation of the coelcmic pouches. A section through an em- bryo with the rudiment of one pcuch and of the notcchord. £ section through a slightly older embryo, showing the complete separation of a coelomic poueh from the archenteroii (from Korschelt and Heider alter Batscfcek) ; ok ectodeim ; ch rudiment of r.otochord ; hb lateral ectoderm growing over the medullary plate ; ik, U endoderm*; Ih entejon Ik future coelcm ; ntk coelcmic rcuch (future scmite) ; np medullary plate. and becomes separated from the gut as a single cavity with two backwardly projecting horns. This unpaired anterior sac gives rise to the head cavities ; it becomes divided into two, of which the right shifts ventrally, becomes thin-walled, and forms the cavity of the snout in the larva (Fig. 20). It becomes largely obliterated in the adult. The left head-cavity, on the other hand, becomes transversely placed beneath the notochord and opens to the exterior on the left side in front of the mouth (Fig. 20, w). It becomes the preoral (Hatschek's) pit, and gives rise by the extension of its lips to the wheel organ (p. 20). The collar-somites (so-called first pair) and the somites developed from the coelomic groove give rise to the mesoderm, bo HATCHING. nip and generative organs in a manner de- scribed below. The collar-somites send forward a process dorsal to the head cavities on each side, the walls of which give rise to the mesodermal structures of the preoral region. "While these changes have been tak- ing place the notochord is developed. It arises as a groove of the dorso- median endoderm (Fig. 17 B], .which is constricted off so as to form, a solid rod of cells lying between the dorsal endoderm and the medullary plate. It develops from before backwards, ex- cept the front portion, which separates from the endoderm later than the rest and extends to the anterior end of the body. The embryo becomes hatched and begins to swim freely in the sea by means of the cilia of the ectoderm cells at about the stage of Fig. 18, about twelve hours after fertilization. But it remains opaque and, being incapable of taking in food, is generally spoken of as an embryo until about the thirty-sixth hour, when the yolk is sufficiently absorbed to leave the tissues transparent, and the mouth, anus, and first gill-slit are formed (Fig. 20). The mouth is formed on the left side ; the first gill -slit on the ventral FIG. 18. — Longitudinal- vertical section through an embryo of Amphioxus with the rudi- ments of two somites (after Hatschek). mp pole-cells, the existence of which is now denied ; mr uncovered part of medullary plate ; mr1 space between the medullary plate plate and overgrown ectoderm (future medullary canal ); us', us" mesoblastic somites. moving on to the right side, and the anus at the hind end of FIG. 19. — Diagrammatic longitudinal section through an embryo «"® body Slightly of Amphioxus, to show the formation of the coelomic sacs. f_ fr_ .. i-ff _f fu _ The figure combines features which in reality would only be u shown by two sections. 1 neuropore ; 2 neural canal ; 3 neuren- ^ j^lp linp teric canal ; 4 coelomic groove ; 5 somite constricted off from ""V**1 Line, anterior end of coelomic gro<5ve ; 6 cavity of first mesoblastic T'Tio IQT»TTC»I somite (so-called collar cavity) ;_"? head cavity ; 8 archenteron. L * L v ** L 36 PHYLUM CEPHALOCHORDA. stage lasts for about three months, during which the larva swims freely usually at a considerable depth by the contrac- tion of its body, and acquires the adult form and habits. The principal changes which take place relate to the formation of the gill -slits, of the preoral hood, and symmetrical adult mouth, and of the atrial cavity. The mouth acquires a gigantic size and forms a most con- spicuous object on the left side of the body (Figs. 21, 22). The gill-slits are formed successively on the ventral middle line to the number of fourteen.* All of these except the last few shift soon after their formation on to the right side. They corre- spond in number to the myotomes in the part of the body FIG. 20. — Ampfiioxus larva of about thirty-six hours from the left side, when the preoral pit- mouth, first gill-slit and anus are established (from Korschelt and Heider after Hatschek); c larval caudal fin ; ch notochord ; en neurenteric canal ; d alimentary canal ; h right preoral sac ; k club-shaped gland, which has acquired an opening to the exterior on the left side ventral to the mouth ; ks first gill-slit ; m mouth ; mr nerve tube ; np neuropore ; sv sub- intestinal vein ; w preoral pit (left head cavity). in which they occur, but this relation is ultimately lost, when they become more closely packed and the myotomes increase in size. Of these fourteen first-formed gill- slits the first and the last five close up, so that eight only are left. atrial folds arise as ridges of the skin. Posteriorly they lie close together in the middle line enclosing between them a small groove (Fig. 23). Anteriorly they pass on to the right side, one on either side of the gill-clefts. By the union of these ridges, which begins in their posterior region and gradually extends forwards, the groove becomes converted into a canal, * These first formed gill-slits are -often called the primary gill-slits, in contradistinction to the later formed secondary slits which, when the animal becomes symmetrical, are placed on the right side, the primary slits having passed over to the left side. Forster Cooper (op, cit.) describes a larva taken in the open ocean in the Maldive Archipelago containing as many as thirty-one primary slits. fs| 37 38 PHYLUM CEPHALOCHORDA. the atrial cavity (Fig. 23). At their hind end they remain separate, thus giving rise to the atriopore. At first the atrial cavity is a small canal restricted to the ventral side of the pharynx. Later it becomes larger and acquires the adult relations. Meanwhile a row of eight or nine gill -clefts appear on the right side of the body dorsal to those first formed (Figs. 24, 25). Both sets of gill-clefts acquire the U-shaped form, the tongue bar being developed (except in the first cleft (Fig. 25) which remains simple). The first-formed clef ts then shift on to the left side of the body, and become the gill-clefts of the left side. At the same time the mouth shifts to the middle line, and the preoral ZT FIG. 23. — Ventral view of three larval stages of Amphioxus (after Lankester and Willey* from Korschelt and Heider) ; A, the atrium is still entirely open ; B, the atrium is partially closed behind ; C, the atrium is almost completely closed ; ap atriopore ; k gill-slits ; If left atrial fold ; m mouth ; rf right atrial fold ; w preoral pit. hood which had made its first appearance while the mouth was still on the left side becomes developed. The apertures of the club-shaped gland (see below) and of the ciliated pit are both enclosed by the preoral hood. The original mouth opening shifts to the back of the oral cavity and persists as the aperture in the velum. The principal phases of the development are now accom- plished and the larva, in the main symmetrical, assumes the sand-burrowing habits characteristic of the adult. The mesoblastic somites, after their separation from the archenteron, which after that event is termed simply enteron, extend ventralwards on each side (Fig. 26), till they meet on the SOMITES. 39 ventral side of the alimentary canal. The septum between them (ventral mesentery) breaks down and the somites of the two sides become continuous. FIG. 24. — Two larval stages of Amphioxus from the right side, showing the origin of the gill- clefts of the right side of the adult (from Korschelt and Heider, after Willey). 1. 2, 9, 12, 14 first, etc., to fourteenth gill-cleft of future left side • /, VII first and seventh of the later formed set which eventually belong to the right side ; au eye-spot ; ch noto- chord ; dr club-shaped gland ; es rudiment of endostyle ; fh dorsal, /&' ventral fin spaces ; k rudiments of the later formed gill-clefts ; m margin of mouth ; mf edge of right metapleure ; n nerve cord ; p atrial cavity ; si sub-pharyngeal vessel ; ^v velum ; w preoral pit. FIG. 25. — Ventral view of Amphioxus larva, rather later than Fig. 24 (from Korschelt and Heider, after Willey). 2 second, 12 vestige of twelfth cleft of the first-formed row, now passed on to the left side ; I, VIII first and eighth of the later formed clefts of the right side ; be buccal cirri ; ch notochord ; [es endostyle ; m mouth (larval) ; v velum. ; At the same time a septum is formed, dividing off the dorsal part of the somites from the ventral. The dorsal part becomes 40 PHYLUM CEPHALOCHORDA. the myotome ; it retains its original segmentation and the septa separating the successive somites here persist as the myosepta (intermuscular septa). In the ventral portion (comparable to the lateral mesoblastic plate of the Vertebrata) the transverse septa, formed by the adjacent walls of the successive somites of the same side, break down and the cavities of the somite become continuous to form the splanchnocoele or body-cavity of the adult. The cells of the inner wall of the myotome become con- verted into muscles, and consti- tute the lateral muscles of the adult, while the outer wall which is applied to the ectoderm mk" FIG. 26. — Transverse section through the middle of the body of an Amphi- oxus embryo with 11 somites. On the right side the section has managed to cut two somites (from Korschelt and Heider) . ak ectoderm ; ch notochord ; dh enteron ; ik endoderm ; Ih coelom ; mk' somatic, mk" splanchnic layer of mesoderm ; n nerve tube ; us meso- blastic somite. FlO. 27. — Transverse section of the middle of the body of an ^ Amphioxus larva with five ^ gill-slits, diagrammatic (from Korschelt and Heider). 1 j outer wall of myotome (cutis i layer) ; 2 inner or muscular wall of myotome ; 3 com- mencement of the sclerotome ; 4 septum between myocoele and splanchnocoele : 5 somatic mesoderm ; 6 splanchnic mesoderm ; I myocoele ; // splanchnocoele. The sub-in- testinal vein is shown in the splanchnic mesoderm. remains thin (Fig. 27). The sclerot'ome is an outgrowth from the ventral and inner wall of the myocoele (Fig. 27). It acquires a considerable development extending dorsalwards between the muscles and the notochord and spinal chord (Fig. 28). Its inner wall gives rise to the sheath of the notochord and of the nerve cord, while its outer wall forms the so-called fascia-layer, or internal sheath (Fig. 29). The dorsal part of the myocoele is said to give rise to the dorsal fin canal ; and a ventral extension of the same space to the ventral fin canal. The myocoele appears to abort in the adult, but the sclerocoele probably in part persists as the lymph spaces on the internal sides of the lateral muscles. GOXADS. 41 In the region of the pharynx, the atrial cavity extends dorsal- wards between the splanchnocoele and the ventral extension of the myocoele (Fig. 29), and the splanchnocoele becomes cut up by the gill- clefts into the sections of the coelom already described as occurring in the primary pharyngeal bars (p. 24). The dorsal and ventral regions of this part of the splanchnocoele furnish the dorso-pharyngeal (sc) and endostylar coelom (ec) respect- ively. The first rnyotome is developed from the somites of the anterior pair, the so- called collar somites (p. 34), which retain their communication with the enteron longer than the others. This communi- cation on the left side becomes elongated and gives rise to Hatschek's nephridium, (Fig. 9,20). These somites send back ven- tral extensions which lie in the developing atrial folds. It has been suggested that these give rise to the metapleural lymph canals, but this has been denied. The origin of the metapleural lymph canals is not certainly known. Van Wijhe in his recent important paper (op. cit.) states that the walls of the collar cavity give rise to several myotomes. It is not quite clear to us whether this statement is based on em- bryological study or not. He further states that the second myotome of the body is the anterior of these myotomes which come from the collar-somite ; thus implying that the walls of the head cavities (preoral somite) give rise to a myotome. So far as we know, the head cavities do not give rise to myotomes. The collar-coelom appears to give rise to the stomocoel and cavum epipterygium of van Wijhe, who states that the cavum epipte the metapleural canals. This confirms MacBri metapleural lymph canals are parts of the coll epipterygium is also stated to communicate "with the endostylar cafbin. through the coelom in the first branchial arch, which, being part m the splanchnocoele, it might reasonably be expected to do. The arrangement of the parts of the coelom about the mouth as explained by van Wijhe is complicated. The gonads are segmented in their origin. The generative cells are first seen as the thickenings of the coelomic epithelium at the ventral ends of the myotomes, on the anterior wall of the FlG. 28. — Transverse section through a young Amphioxus immediately after metamorphosis, between the atriopore and the anus, diagram- matic (from Korschelt and Heider, after Hatschek). 1 outer wall of myotome (cutis layer) ; 2 muscles ; 3 fascia layer (outer wall of sclero- toj»e ; 4 skeletogenous (inner) w» of sclerotome ; 5 u 6 ventral continuation of skeletogenous layer and somatic wall of splanch- nocoele ; 7 splanchnic ditto ; / kj /' dorsal /* ventral fln- ! splanchnocoele. communicates wit statement that coelom. The ca 42 PHYLUM CEPHALOCHORDA. myocoele. They soon come to project into the cavity of the somite in front, pushing the myosepta before them (Fig. 30). FIG. 29.— Transverse section through the branchial region of Amphioxus showing on^the left the condition of a secondary, and on the right that of a primary gill-bar, diagrammatic (after Boveri and Hatschek, from Korschelt and Heiderj. ao aorta ; c cutis layer of myocoele ; ec endostylar coelom ; / fascia layer ; fh dorsal fin canal ; p genital sac ; gl renal vessels ; k vessel in primary bar ; kd pharynx ; Id ligamentum denticulatum ; mfmuscle- plate ; mt transverse muscle ; n renal canal"; of metapleural lymph space ; p atrial ca\ :ty ; sc dorsopharyngeal coelom ; si sub-pharyngeal vessel ; sk skeletogenous layer of sclerotome ; «/ lymph canals of the atrial floor. They lie therefore as small sacs in the preceding myocoele attached to its hind wall by a pedicle (Figs. 30 (7, 31). Later, the part of the myotome in which they lie becomes separated from ENDOSTYLE. 43 the rest and forms the perigonadial coelom which lies in the outer wall of the atrium (Fig. 10). The club-shaped gland and endostyle. The club-shaped gland B FIG. 30. — A, B, C. — Three side views of the ventral end of a niyotome of a young Amphipxus, showing the development of ihf genital rudiment and its projection into the cavity of the preceding somite (from Korschelt and Heider, after Boveri). is developed as a transverse groove on the floor of the pharynx and continued on to the right and left walls. It becomes con- stricted off from the pharynx, and acquires an opening to the exterior on the left side of the body just in front of the mouth. Later on the right end of it acquires an opening into the pharynx. The club-shaped gland is of unknown function and eventually atrophies. It has been suggested without any obvious justification that it is the metamorphosed an- terior gill of the right side. The endostyle is a ciliated tract of columnar epithelium just an- terior to the club-shaped gland on the right side (Fig. 24). It subsequently becomes bent on itself in a V-shaped manner (Fig. 25), and grows backwards between the two rows of gill- slits. It is at first, therefore,' on the right side of the body, but when the larva becomes sym- metrical, it takes up its position in the ventral middle line. From the above account it is clear that in the young larva -bni FIG. 31. — Transverse section through the genital rudiment of a young Amphioxus showing the separation of the perigonadial coelom from the myocoele. bm ventral muscle; g blood-vessel ; gd gonad ; w fold separating myocoele from peri- gonadial coelom. 44 PHYLUM CEPHALOCHORDA. the future ventral middle line of the pharyngeal region is on the right side of the body as shown by the first trace of the subpharyngeal vessel, the endostyle and the gill- slits ; whereas in the buccal region later median structures are on the left side. It results from this that, at the so-called metamorphosis when the larva becomes symmetrical, the buccal region of the body and the pharyngeal regions must rotate, so to speak, in opposite directions. No satisfactory explanation of this extraordinary larval asymmetry has even been suggested. Though largely rectified in the adult a 'trace of it persists in the slightly asymmetrical position of the olfactory pit, the anus, and the continuity of the cephalic fin with the right side of the preoral hood, and in the innervation of the preoral hood (p. 18). CHAPTER III. PHYLUM VERTEBRATA* (CRANIATA). Chordata in which the dorsal nerve cord extends some distance in front of the notochord, and is expanded at its anterior end into a brain. The axial skeleton is divided into an unsegmented cranial portion, which surrounds the brain, and a segmented vertebral portion which forms the axis of the body and protects the spinal cord. The various animals included in this phylum were first put together by Aristotle, who called them " animals with blood " ; he also recognized the possession of a bony or cartilaginous skeletal axis as a common characteristic. But it was Lamarck who first adduced the presence of a vertebral column, as a most important character, and introduced before Cuvier the name of Vertebrata into the science. This term, however, is not entirely appropriate, for in some Pisces the sheath of the noto- chord is not segmented, and there are no vertebrae (Marsi- pobranchii, Dipnoi, some Ganoidei). Nevertheless, the term may fairly be retained, for not only has it the sanction of long usage, but the cases in which the vertebral column is not jointed are few in number and unimportant in character. As already pointed out, the segmentation of the vertebral column is corre- * Stannius, Handbook der Anatomic ' der Wirbelthiere, 2nd ed., Berlin, 1854. Rathke, Beitrdge zur Bildungs und Entwickelungsgeschichte des Menschen und der Thiere, Leipzig, 1833. Owen, The Anatomy of Vertebrates, 3 vols., London, 1866-68. Huxley, A Manual of the Anatomy of Verte.brated Animals, London, 1871. Gegenbaur, Vergleichende Anatomic der Wirbelthiere, Leipzig, 1898, 1901. Zittel, Handbuch der Palaeontologie, Munich, vols. iii., iv., 1887-93 ; and Grundziige der Palaeontologie, Munich, 1895. (English translation, Macmillan and Co., 1900). A. S. Woodward, Out- lines of Vertebrate Palaeontology, Cambridge, 1898. Balfour, Comparative Embryology, vol. ii., 1882. C. S. Minot, Human Embryology, New York, 1892. 40 PHYLUM VERTEBRATA (CRANIATA). lated with its rigidity, and is therefore best developed in those animals which have to support the body on land. The integument consists of two distinct layers, the epidermis externally and the cutis internally. The epidermis is composed of many layers of cells, of which the upper and older layers are worn off, while the lower layer (stratum malpighi) is actively growing, and serves as a matrix for the continual renewal of the upper layers, and sometimes contains pigment. The cutis is principally formed of fibrous connective tissue, with which muscular elements — striped and unstriped — come into relation without however forming a dermo-muscular envelope, as in the Annelids. Some of the appendages of the skin are epidermal structures (hairs and feathers). Some are derived from ossi- fications of dermal papillae, which sometimes may even give rise to a hard and complete dermal armour (scales of fishes and reptiles, carapace of armadillos and tortoises). The epidermis is derived from the ectoderm of the embryo, the cutis or dermis being mainly a mesodermal product. The endo-skeletal tissue of the lower Vertebrata and of all vertebrate embryos consists solely of cartilage (Marsipobranchii, Elasmobranchii), but in most groups osseous tissue, supple- menting or, in the higher forms, largely replacing the cartilage, is present in the adult. The muscular tissue may be divided into two categories : They are (1) the somatic or myotome muscles, which are derived from the epithelial wall of the myotomes or dorsal segmented parts of the mesoblast of the embryo, and (2) the mesenchyma- tous * muscles, which are developed from the ventral part of the mesoblast (wall of the splanchnocoel). The myotome muscles are innervated exclusively by the ventral roots of the spinal nerves, and by the third, fourth and sixth cranial nerves, which are the only ventral nerve roots found in the brain. The mesenchymatous (visceral) muscles, which appear to be derived from mesenchymatous mesoderm, are innervated by the ventral roots in the trunk, but in the head by the fifth, seventh, ninth and tenth cranial nerves, which are usually regarded as dorsal nerve-roots, and which contain also afferent nerve fibres (see account of nerves under Pisces). The somatic muscles are cross- * Sometimes called visceral, but this is a bad name, as many of them lie in the body-wall. NERVOUS SYSTEM. 47 striped and voluntary. The mesenchymatous muscles for the most part consist of unstriped fibres, but some of them are cross-striped, and even voluntary. The muscles of the heart and oesophagus are examples of cross-striped mesenchymatous muscles in the trunk ; they are not under the control of the will. In the head many of the mesenchymatous muscles are cross- striped and voluntary ; e.g. the facial muscles, the mandibular and the branchial muscles. The eye-muscles are myotome muscles, and supplied, as stated above, by ventral roots ; but they differ from the muscles of the great lateral sheet of myotome trunk muscles in the fact that their fibres are directed dorso- ventrally, and not longitudinally, as in the latter. The dorsal nerve-cord extends in front of the notochord, and is enlarged in front to form the brain, which is constructed on the same fundamental plan in all classes. The posterior part constitutes the spinal cord. The skeletal investment of the brain is unsegmented, and constitutes the skull, while the spinal cord lies in a tube of the vertebral column, which always shows some sign of segmentation and is usually completely segmented. The spinal nerves, which are segmentally arranged, possess two roots, a dorsal and a ventral, which join. The dorsal of these roots carries a ganglion and contains afferent nerve fibres ; the ventral contains efferent fibres only. The brain possesses ten pairs of nerves, which are very similarly arranged in all Vertebrata. They differ from the spinal nerves in the fact that except in the case of three of them, they have the dorsal roots only. The third and sixth nerves may be regarded as the ventral roots of the fifth and seventh nerves respectively, and the fourth nerve must also be regarded as a ventral root, though it arises from the dorsal surface of the brain. The ninth and tenth nerves appear to be altogether without ventral roots. In the higher Vertebrata there are two additional pairs of cranial nerves, the eleventh and twelfth. Of the cranial nerves, the fifth, seventh, eighth, ninth and tenth are usually regarded as being serially homologous with the posterior roots of spinal nerves, and are supposed to be related to a vanished segmenta- tion of this part of the body.* They resemble these in having * A short account of the modern views on the nature of cranial nerves and of nerves in general, and of cranial segmentation is given in the chapter on Pisces. 48 PHYLUM VERTEBRATA (CRANIATA). a ganglion, but, with the exception of the eighth, they differ from them in containing a fair proportion of efferent nerve fibres. The first and second cranial nerves, which supply the special sense organs of smell and sight respectively, appear to differ fundamentally from the other cranial nerves. In all Vertebrata there are three organs of special sense on the head, the olfactory and visual, the innervation of which has just been referred to, and the auditory, which is supplied by the eighth cranial nerve. The connection of these organs with the head has profoundly modified the structure of both skull and brain. A special visceral nervous system, known as the sym- pathetic, is nearly always present. The alimentary canal presents very similar features through- out the series. It consists of a stomodaeum, pharynx, voeso- phagus, stomach, intestine. The stomodaeum contains the teeth and the openings of the salivary glands if present, and passes without any line of demarcation into the pharynx, which in all Vertebrata is at some time of life connected with the ex- terior by lateral apertures, the pharyngeal apertures or visceral clefts. These are never more (usually less) than eight in number on each side.* In Fishes and Amphibia the first of these (spiracle) is always smaller than the others and may be completely absent ; in the Amniota (Reptilia, Aves, and Mammalia) it is always present, and not smaller than the others. In Fishes and Amphibia the visceral clefts are used for respiratory purposes as in other Chcrdata, but they are never put to this use in the Amniota, where they appear to have no function at all. In such cases the respiratory organ of the adult is the lung, which is developed as a median outgrowth of the ventral wall of the pharynx. There are two other nearly constant features of the vertebrate alimentary canal, viz. (1) the connection of two large glands, the liver and pancreas, with the anterior part of the intestine, and (2) the connection of the generative and renal organs with the hind end of the intestine, which is com- monly called the cloaca. The junction of the endoderm and ectoderm appears to take place at the anus, and there is prac- tically no proctodaeum in the Vertebrata. * Except in some Marsipobranchs; PHYLUM VERTEBRATA (CRANIATA). 49 The vascular system is well developed, consists of arteries, capillaries and veins, and contains a red blood. There is a median ventral subpharyngeal vessel, the hinder end of which is especially muscular and contractile and differentiated as the heart. The lymphatic system consists of vein-like vessels containing a colourless fluid — the lymph — in which float numerous amoeboid cells (lymph corpuscles). .These vessels commence by blindly-ending fine tubes or sinuses in the tissues, which gradually unite with one another to form the main lymph vessels, which open into the venous system. Special gland-like bodies, the so-called lymphatic glands, in which the lymph corpuscles are produced, are inserted in the course of the lymphatic vessels. The lymphatic system is a draining system, for the purpose of carrying away from the tissues the fluid which has exuded into them through the walls of the blood-capillaries, and is undoubtedly a specially differ- entiated part of the vascular system. The body-cavity is a coelom, and has the usual relations of that organ to the urinary and reproductive organs. It is laid down early, making its appearance as a split (schizocoel) in the mesoblast ; and in the Elasmobranchii, at any rate, a certain resemblance between it at its first appearance and that of Amphioxus can be detected (p. 33). But it differs from that of Amphioxus in that the ventral portions of the trunk somites are never distinct from one another, but form from the first a continuous splanchnocoel. In the adult the body-cavity is always divided more or less completely into a pericardial division in front and a peritoneal division or general body-cavity behind. In the mammals the latter is still further subdivided, in that two anterior horns are cut off from it to form the pleural cavities. There is no coelom in the head of adult vertebrata. The urinary organs consist typically, in their origin at least, of segmentally arranged nephridia, which open internally into the body-cavity. Externally they open into a longitudinal duct which leads into the hind end of the intestine in almost all cases. Both nephridia and ducts develop as special portions of the coelom. The generative organs develop from the lining of the unsegmented ventral part of the coelom (splanchnocoel), and z — TI E .XTTiTAl f^^ OF THE I UNIVERSITY 50 PHYLUM VERTEBRATA (CRANIATA). never present any trace of segmentation.* In the female they retain this relation throughout life (except in Teleostei), but in the male the generative part of the coelomic epithelium always (except in Marsipobranchii) loses its relation with the general body-cavity in the adult. The ovum varies considerably in character in the different classes. In Pisces (except Elasmobranchii) and in Amphibia it is comparatively small and holoblastic,f and the young are always hatched out in an immature condition as larvae. In Elasmobranchii, Reptilia, and Aves the ovum is large and meroblastic, and the young when hatched resemble the adult, a larval stage being absent. In Mammalia the egg is smaller than in any other Vertebrate, and except in one class under- goes almost the whole of its development in the oviduct, the young being born in a condition closely resembling the adult. An amnion, allantois and primitive streak are found in the embryos of all Reptiles, Birds and Mammals, but are absent from all Pisces and Amphibia. The division of the Vertebrata into the four classes — Pisces, Amphibia, Aves, and Mammalia was first established by Linnaeus, though it had already been indicated in the system of Aristotle. The Pisces and Amphibia are cold-blooded animals (i.e. animals with a varying temperature) ; Aves and Mammalia are warm-blooded. Since Linnaeus' day, his group Amphibia has been split up into the naked Amphibia and into the scaly animals or Reptilia. Pisces and naked Amphibia have many characters in common, e.g. the branchial respiration, the frequent persistence of the notochord, the absence of an amnion and allantois, etc. On these grounds and in con- sideration of the many relations between Reptiles and Birds, Huxley has distinguished three principal groups of Vertebrata —the Ichthyopsida (Pisces and Amphibia), the Sauropsida (Reptilia and Aves), and the Mammalia. * See note on p. 88. f In Teleostei the ovum though small is meroblastic. CHAPTER IV. CLASS PISCES.* Aquatic vertebrata which breathe by means of pharyngeal gills and possess typically two pairs of appendages which never present any trace of a pentadactyle structure. Median fins supported, except in Marsipobranchs, by dermal fin-rays (dermotrichia) are always present. There are ten pairs of cranial nerves and paired posterior cardinal veins. Fishes are sharply marked off from all the other classes of Vertebrata by the form of their pectoral and pelvic appendages. These, which must be regarded as homologous with the limbs of the higher groups, are cutaneous expansions supported by skeletal structures, which, though presenting in their fan-like arrangement some distant resemblance to the skeletal structures of the pentadactyle limb, are yet never arranged on the penta- * Lacepede, Histoire naturelle des Poissons, 6 vols.r Paris, 1798-1803. G. Ouvier et Valenciennes, Histoire naturelle des Poissons, 22 vols., Paris, 1828-1849. Baer, Entwickelungsgeschichte der Fische, Leipzig, 1835. J. Miiller, Vergleichende Anatomie der Myxinoiden, Berlin, 1835- 45. Id. Ueber den Bau u. die Grenzen der Ganoiden u. d. natiirliche System der Fische, Abh. d. Berlin, Akad, 1846. L. Agassiz, Recherche* sur les Poissons fossiles, 5 vols. Neuchatel, 1833-44. Stannius, op. cit. Heckel and Kner, Die Susswasserfische von der osterreischischen Monarchic, Leipzig, 1858. A Dumeril, Ichthyologie, etc., 2 vols., Paris, 1866. Siebold, Die Susswasserfische von Mitteleuropa, Leipzig, 1863. Blanchard, Les Poissons des eaux douces de la France, Paris, 1866. Cope, " Classification of Fishes," Trans. Amer. Phil. Soc., 1870, and Proc. Amer. Ass. for Adv. of Science, 1871. A. Gunther, Introduction to the Study of Fishes, Edinburgh, 1880; and Catalogue of Fishes in the British Museum, vols. i-viii., 1859-1870. A. S. Woodward, Catalogue of Fossil Fishes in the British Museum, 3 vols. London, 1889-95. F. Day, The Fishes of Great Britain and Ireland, London, 1881-83. Jordan and Ever- man, Fishes of North and Middle America, Bull. U. S. National Museum, no. 47, Pts. 1-4, 1896-1900. T. W. Bridge, Fishes in the Cam- bridge Natural History, 1904. G. A. Boulenger, Poissons du Bassin du Congo, 1901. 52 CLASS PISCES. dactyle plan. The ichthyopterygium, though clearly homolo- gous with the cheiropterygium, is sharply marked off from it, and there are no intermediate forms connecting the two. Though the structure of the limb skeleton, and possibly the possession throughout life of the paired posterior cardinal veins, are the only absolutely characteristic features of fishes, there is a number of features which, while not distinctive, are highly characteristic. We may enumerate some of these :— 1. Median unpaired folds of the integument, constituting the unpaired fins and supported by dermal structures called fin-rays or dermotrichia, would be absolutely characteristic were it not for the fact that the fin-rays are absent from the median fins of Marsipobranchii. 2. The absence of an internal opening of the nasal sac would be absolutely characteristic of the class were it not for the presence of internal narial openings in Dipnoi and of the pharyngo- nasal duct in Myxinoids. 3. Respiration by means of lateral pharyngeal apertures and gills is found in the adult in no other group excepting in a few genera of the Amphibia. 4. The absence of a cloacal bladder might perhaps be cited as a distinctive character, for it is present at least in the embryos in all the higher Vertebrata. 5. Excepting in the Dipnoi the auricle is undivided, and the ventral aorta is a tube of considerable extent. 6. Excepting in the Dipnoi there is nothing corresponding to the median inferior vena cava of the other classes. 7. The absence of a tympanic cavity and membrane, and of anything corresponding to the auditory ossicles of the higher types may, we think, be cited as a distinctive character, for these structures are very rarely, if ever, completely absent in the other classes. 8. The presence of the peculiar sense organ known as the lateral line might almost be regarded as a piscine character, were it not for the fact that it is not clearly present in all Marsipobranchii, and that it is found in some Amphibia. 9. The permanent division of the great lateral longitudinal muscles of the body into segments (myotomes) corresponding in number to the vertebral segments is only found outside Pisces in some Amphibia, and in the tail of some Repbelia. SCALES. 53 The epidermis contains large mucous cells which discharge their contents on the surface. It may also contain pigment cells and leucocytes. In many fishes the slime which is excreted by the skin is poisonous. The skin is seldom completely without skeletal structures (Marsipobranchii). As a rule scales, formed as ossifications of dermal papillae which are typically completely covered by the epidermis, are embedded in it. Fish scales * are of three principal kinds : (1) Placoid scales which consist of small plates of bone in the dermis carrying an upstanding spine which projects freely, and is formed of dentine capped with enamel. These are found in Elasmobranchii and some Ganoids. (2) FIG. 32.— Perca ftuviatttis (RSgne animal). Ganoid scales are bony plates covered with a smooth layer of a substance called ganoin. Ganoin f is a dermal product allied to vitro-dentine. These scales are entirely dermal, and if their surface is exposed, it is owing to the fact that the epidermis has been rubbed off. Such scales are found in most Ganoids. (3) Scales of varying thickness consisting of bone only, without ganoin. They are found in most Teleosteans, and are there called cycloid and ctenoid scales according to the nature of their edges. The unpaired fins arise as a continuous fold of skin extending * A fuller account of the scales is given with the accounts of the orders. For principal recent literature, see Klaatsch, Morph. Jahrb. 16, p. 258, and 21, 1894, p. 153 ; F. Maurer Die Epidermis u. Hire Abkommlinge, Leip- zig, 1895. 6. Hertwig, Morph. Jahrb., ii. and vii. C. Rose, Anai. Anz., 14, 1897 ; pp. 21 and 33. Nickerson, "The development of the scales of Lepidosteus," Bull. Mus. Harvard, 24, 1893. f It was formerly supposed to be enamel and to be epidermal in origin, but this has been proved to be erroneous. 54 CLASS PISCES. along the middle dorsal line of the trunk, and continued round the tail on to the ventral surface as far as the anus. It may persist in this form, but as a rule it becomes broken up into a variable number of dorsal fins, a caudal fin which consists of a dorsal and ventral part, and an anal fin between the ventro- caudal and the anus (Fig. 32). The unpaired fins are almost always supported by the so-called fin-rays or dermotrichia. These are horny fibres of the derm is (Elasmobranchii), or bony rods (Teleostei, Ganoids, Dipnoi) which may be segmented, and more or less soft and flexible (Malacopterygians) or stout and un- segmented(Acanthopterygians). These dermotrichia are absent only in Marsipobranchii. They are composed of two closely approximated halves, and are carried except in the case of the ventro-caudal fin, by the somactids or radialia. These are carti- laginous or bony rods, placed as a rule in the basal part of the fin-fold, and between the muscles of the back. They do not necessarily correspond in number with the vertebrae. They are usually segmented into two or three pieces, to the distal of which the dermotrichia are attached. The basal piece is some- times called the axonost ; in the Teleostei it is known as the interspinous bone, because it occurs between the spines (neural or haemal) of the vertebrae. The second piece is sometimes known as the baseost. In a few fishes (e.g. Dipnoi), the somactids articulate with the spines of the vertebrae. The strong spine-like anterior fin-ray often found in Teleosteans and bony Ganoids is formed of bone. In Elasmobranchii, the strong spines which are sometimes found in connection with the fins are tooth -like in structure. The dermotrichia are of three kinds. * 1. In Elasmobranchii and Holocephali they are unjointed, occasionally branched, fibrous rays of a horny consistency and without osseous tissue : these are called ceratotrichia. Similar dermotrichia are found in the larval fins and at the edges of the adult fins of Teleosteans and Ganoids : in this case they are called actinotrichia. They are more numerous than the somactids. 2. In adult Teleosteans and Ganoids the fins have jointed, branched, bony dermotrichia developed between the actinotrichia and the skin. They are supposed to be modified scales, which they sometimes resemble, and are called lepidotrichia. They correspond in number with the somactids except in the cartilaginous Ganoids, in which they are more numerous. 3. In the Dipnoi the dermotrichia have been called kamptotrichia * Goodrich, " Dermal fin-rays of Fishes," Q. J. M. S., 1904. CAUDAL FIN. 55 They are branched, jointed bony rays, and appear to be merely modified lepidotrichia. They are more numerous than the somactids. In all fishes the ventral part of the caudal differs from the other median fins in the fact that the dermotrichia (fin-rays) are supported directly by the haemal arches. These are fre- quently imperfectly segmented from one another, and may, in the adult, have the form of two or three bony plates, or even of a single plate. They are sometimes called, when ossified, the hypural bones. A few Teleostei (eel-like forms, some Gadidae, etc.) and the Dipnoi constitute apparent exceptions to this rule as to the structure of the caudal fin, but in the former of these it is probable that the anal fin has fused with the ventral part of the caudal fin, for in a small por- tion of the ventro-caudal fin a few dermotrichia are carried directly by haemal arches. In the Dipnoi on the other hand, and in some Teleostei, the caudal fin appears to be entirely unrepresented, for the tail gradually tapers to its termination. What appears at first sight to be the ventro- caudal fin is really the anal fin, and in no part of it are the dermotrichia supported directly by the haemal arches. In the crossopterygian Ganoids, in which there is a diphycercal tail, the ventral dermotrichia of the caudal fin are clearly supported by haemal arches and not by somactids. Considerable importance has been attached to the form of the tail and to the structure of the caudal fin in fishes. In the simplest cases the vertebral column is continued straight to its termination, and the dorsal and ventral part of the caudal fin are equal and symmetrical with each other. This type of caudal fin is called diphycercal or protocercal. In many fishes, however, the posterior part of the vertebral column is bent dorsalwards, and a special enlargement of the ventro-caudal fin is formed at a short distance from the end of the tail. This type of caudal fin is called heterocercal (Fig. 64), and is character- istic of Elasmobranchs and chondrostean Ganoids. In such fishes the tail may be said to be bifid, presenting a dorsal lobe and a ventral lobe. The dorsal lobe consists of the real hind end of the tail with the dorso-caudal (if present at all) and part of the ventro-caudal fin, while the ventral lobe is the specially enlarged part of the ventro-caudal fin above described. In Teleostei, and bony Ganoids, and some Selachians, the dorsal lobe thus defined shrinks and almost disappears relatively to the greatly enlarged ventral lobe, which now forms the whole of the tail fin, and becomes svmmetrical in itself. Such a tail 56 CLASS PISCES. is called homocercal. In many Teleostei the tail fin of the larva begins diphycercal then becomes heterocercal and finally assumes the homocercal form. This correspondence between the develop- mental history of the tail and the three forms of tail fin found in living fishes is supposed to be highly significant from an evolutionary point of view, for it is supposed that the diphy- cercal tail is the most primitive, and that the homocercal is the most specialised, the heterocercal tail intervening between the two. This supposition is to a certain extent borne out by palaeontology, which seems to show that Teleosteans are the most modern group of fishes. Unfortunately for the theory, however, the oldest fishes known to us had heterocercal tails and not diphycercal, as the theory requires. In addition to these three types of tail-fin, intermediate conditions have been named. For instance, the term heterodiphycercal has been applied to slightly heterocercal tails in which the fin is much less developed on the dorsal side than on the ventral (some Crossopterygians, Fig. 104), while tails, in which the tail fin is externally symmetrical, but the hind end of the vertebral column is bent and extends some way into the dorsal lobe of the fin (Amia, etc.), are called hemiheterocercal (Figs. 107, 109). The true homocercal tail is distinguished from the hemiheterocercal by the fact that the vertebral column, which is bent dorsalwards, does not extend into the fin, but terminates in front of it. As has already been mentioried, the dermotrichia of the ventro-caudal fin of all Pisces are attached directly to the haemal arches (for apparent exceptions to this see p. 55). In the homocercal tail of the Teleostean these haemal arches are called the hypvral bones and are frequently fused together to form a single broad plate of bone. In Ganoids with hetero- cercal tails, when the upper lobe of the caudal fin (dorso-caudal) disappears it is replaced by a series of ridge scales : the " fulcra " of palaeontologists : in Elasmobranchs, when absent, it leaves no trace. The pectoral and pelvic fins also possess dermotrichia (fin-rays) and somactids (radialia). A certain number of the somactids are directly articulated to the limb girdles, and are then called basalia. There are usually three of these, which are then called pro- meso- and meta-pterygium, but their number varies con- siderably. The important point to notice is the arrangement of the peripheral somactids.* In Cladoselache, a Palaeozoic fish, these are parallel to one another (Fig. 83), and the fin-skeleton may be termed orthostichous.^ In most fishes, and notably in * Wiedersheim, Das Gliedmassenskelet der Wirbelthiere, Jena, 1892. f The same feature is found in the pelvic fins of the Ganoid Psephurus (Regan, Ann. and Mag. Nat. Hist (7), 13, 1904, p. 333.) ORIGIN OF LIMBS. 57 Elasmobranchs they are arranged in a fan-like manner, and the fin may be described as rhipidostichous. In Dipnoi the somactids of the fin are represented by a basal piece, followed by a row of them occupying the axis of the fin, with or without pre- and post- axial pieces placed like the barbs of a feather (Fig. 138). Such an arrangement may be termed rachiostichous and mesorachic. In some sharks and in the extinct Pleuracanthidae there is a succession of somactids forming a rachis, but the rachis is placed on one side of the fin and carries peripheral somactids mainly on that side. * Such an arrange- ment may be called rachiostichous and pleurorachic (Fig. 76). A similar reduction in the number of" basal somactids is sometimes found in the median fins of extinct fishes, e.g. in the anal fins of Pleura- canthus (Fig. 87).. In the discussion of the vexed question of the origin of the vertebrate paired limbs, much attention is paid to the arrange- ment of these somactids (radialia). According to Gegenbaur the Ceratodus type (Fig. 138) of fin skeleton (rachiostichous) is the most primitive, and this fin constitutes what he calls the archipterygium. On this view the skeleton of the paired fin^ and their girdles have originated from a branchial arch and its branchial rays ; the girdle being derived from the branchial arch and the somactids from the branchial rays. One important objection (among others) to this view is that the branchial arches are in the gut-wall, whereas the limb girdles lie in the body wall. On another and perhaps more acceptable view, if any view on these insoluble questions can be regarded as acceptable, the pectoral and pelvic fins are to be regarded as local specialisations of a once continuous lateral fold of the body wall, containing skeletal structures comparable to those of the unpaired fins, viz. basal segmented somactids (radialia) and peripheral dermo- trichia. This view was first suggested by Balfour. According to it the fin-skeleton of Cladoselache would be appealed to with its parallel somactids as being an obvious local specialisation of a once continuous fold, with parallel somactids all along its course. It would be useless to study the skeleton except in detail, * There appears to be some dispute as to whether this side is pre- or post- axial. According to Wiodersheim and Fritsch, the side on which the majority of the rays are placed is post-axial, but according to the more generally received opinion it is pre-axial, the limb in the specimen from which Fig. 87 is taken having been displaced. 58 CLASS PISCES. not and for that we refer the reader to the account of the different sub-classes. We must content ourselves here with an account of its more general features, to which it is desirable to call attention. The notochord forms the basis of the axial skeleton. It always persists in the adult, though it is generally consider- ably reduced. Its longitudinal extent is from the pituitary fossa of the skull in front to the end of the vertebral column behind. The notochord itself rarely forms an important element of the axial skeleton of the adult. Its supporting function — so conspicuous a feature in Amphioxus — is taken over by its sheath and by cartilaginous structures de- veloped around its sheatn. In the trunk these structures are, with few exceptions, seg- mented, and constitute the vertebral column ; in the skull they are not segmented, and are known in the embryo as the parachordal cartilages. ': has FIG. 33. — Transverse Section through the vertebral column of an advanced embryo of Scyllium in the caudal region ; na skeletogenous tissue of neural arch, ha of haemal arch ; ch notochord ; sh notochordal sheath, which has acquired nuclei (elastica inter na) ; ne outer chordal sheath (elastica externa) (After Balfour). At an early stage in the embryo a well defined structureless sheath is formed round the notochord. This is called the merribrana elastica interna. A little later a second sheath is formed round this. This outer sheath, also structureless, is known as the membrana elastica externa. Cells of the skeletogenous mesoblastic layer, which surrounds the notochord, appear now to penetrate the elastica externa and invade the elastica interna (Fig. 33), which thus becomes nucleated. The chordal sheath sometimes remains as a continuous structure (Sturgeon, Dipnoi, etc.), but as a rule becomes seg- VERTEBRAL COLUMN. 59 mented ; in other words it becomes differentiated into alter- nately short fibrous and longer cartilaginous portions (Fig. 34). The fibrous portions become the intervertebral ligaments, the cartilaginous portions the bodies or centra of the verte- brae. The sheath thickens in the central part of the verte- bral regions and constricts the notochord, so that the noto- chord assumes a beaded form, being narrowest in the middle of the vertebral regions and widest in the intervertebral (Fig. 34). In this way a bi- concave or amphicoelous cen- trum — a form eminently characteristic of fishes — arises. When the centra are formed entirely or mainly from the chordal sheath they are called chordo-centrous (Dipnoi, Elas- mobranchii). But it frequently happens that they are rein- forced by cartilage derived from the arch- tissue. The arch- tissue arises from the meso- blastic tissue (skeletogenous layer), which surrounds the noto- chord, and is continued dorsally round the spinal cord. Four special concentrations of this tissue are formed adjoining the notochord, two dor- sal and two ventral. In these the chondrifications which give rise to the neural and haemal arches begin. The neural arches do not always completely enclose the spinal canal, but are supple- mented by the intercalated pieces (Fig. FIG. 35. — Three pos- 0~\ i • i i -, •• terior trunk-verte- <*D) which are placed between them, i.e. brae of CentTophorus • , , -i -,-, -m , (after Hasse from mtervertebrally. The haemal arches may Gegenbaur). n neural i r i -, -, arch with foramen for also be supplemented by intercalated anterior root ; in in- • mi , .1 tercaiated piece with pieces. These cartilaginous arches may haemafarch1 * spread out round the notochord outside FIG. 34. — a Diagram of a longitudinal section of the vertebral column of a Teleostean with vertebral constriction of the notochord (from Claus). b vertebrae of a bony fish, ch notochord ; D neural spine ; Dl haemal spine ; J interverte- bral ligament ; k body of vertebra ; 06 neural arch : R rib : Wk vertebral body. 60 CLASS PISCES. the membrana elastica externa and unite with each other, and so reinforce the vertebral centra. When the chordal sheath is inconspicuous, and the centra appear to be mainly derived from the arch tissue, the vertebral column is said to be arci- centrous (Ganoidei, Tdeostei).* In this way a segmented vertebral column is established. In Teleosteans and bony Ganoids a further complication is added in the replacement of the cartilage by osseous tissue. The centra in the trunk (Fig. 34) carry as a rule short transverse processes, which may be called haemal arches, though they do not meet ventrally except in the caudal region, where they en- close a space containing the caudal artery and vein. The ribs are never more than short pieces of cartilage or bone attached to the outer ends of the transverse processes in the trunk region. There is no sternum in fishes. The primitive craniumf consists of a cylinder of continuous cartilage, to which are attached anteriorly the nasal capsules widely open below, and posteriorly the auditory capsules. It is thus divided into four regions ; the occipital surrounding the foramen magnum, the wide auditory region, the narrow sphen- oidal or interorbital region, and the wide nasal or ethmoidal region. The junction of the parachordal with the trabecular region of the skull is marked externally by the foramen in the median floor, which transmits the internal carotid arteries (Fig. 36, 13), and internally by the posterior clinoid ridge which forms the hinder wall of the fossa for the lodgment of the pituitary body (Fig. 36, £). Yin the embryo two elongated cartilages — the para- chordal cartilages — are developed on each side of the cranial part of the notochord. They unite with each other around the notochord and form the basilar plate which gives rise to the occipital and part of the sphenoid regions. The auditory capsules which are developed round the membranous labyrinth become fused with this part of the skull. The anterior end of the skull in front of the pituitary fossa is formed by a second pair of embryonic cartilages, the trabeculae. To the front end of these the nasal capsules become attached, thus giving rise to the ethmoidal region. / * Vide Gadow, Phil Trans., 186, 1895, p. 165. t C. Gegenbaur, Untersuchunqen z. vergl. Anat. d. Wirbelthiere, Heft 3. Leipzig, 1872. Id., "Ueb. d. Occipitalregion, etc. der Fische." Kolliker's Festschrift, Leipzig, 1887. CRANIUM. 61 The hinder ends of the trabeculae embrace the front end of the notochord, so that the posterior clinoid ridge must be regarded as being formed by the hind end of the trabeculae. Moreover it must not be forgotten that the internal carotid artery enters the skull in the embryo through the space between the trabeculae before they fuse, so that the carotid canal also belongs to the hind end of the trabecular region. Though the roof of the cranium is largely cartilaginous in fishes even when membrane bones are present on it, there is always a considerable fontanelle in which cartilage is absent. The carti- laginous cranium so constituted becomes in the Teleosts, Ganoids and Dipnoi replaced by bone to a varying extent, and reinforced by the development of osseous tissue in the adjacent connective tissue. The membrane bones, formed in the latter 'is 14 13 FIG. 36. — Median section of the cranium of Hexanchus, inner view (after Gegenbaur). 1 Fora- men for vagus, 2 glossopharyngeal, 3 auditory, 4 facial, 5 trigeminal nerves ; 6 posterior clinoid ridge ; 7 foramen for oculomotor, 8 trochlear, 9 optic nerves ; 10 fontanelle ; 11 rostrum ; 12 lateral process of ethmoid region ; 13 foramen for carotid ; 14 transverse canal in skull base ; 15 notochord ; 16 foramina for spino-occipital nerves ; 17 neural arch of the first vertebra with nerve foramina. manner, apply themselves to the subjacent cartilage and help in forming the cranial wall. The occipital region of the cranium is attached to the anterior end of the vertebral column, usually without any special articulation (except in Batoidei and Chima- era, etc.), the basioccipital region having the conical depression and form of a vertebral body. The cranial part of the noto- chord persists in the adult in some forms, but it more usually undergoes atrophy. It occasionally happens, as will be men- tioned in the special accounts, that a few of the anterior vertebrae are fused with the occipital region of the cranium. Visceral Skeleton. — The walls of the anterior part of the alimentary canal (mouth and pharynx) are supported and 62 CLASS PISCES. strengthened by incomplete cartilaginous rings, analogous to the cartilaginous rings found in the trachea of the higher Vertebrata. Like the tracheal rings, they serve to keep open the tube — in this case the mouth and pharynx — through which the respiratory medium passes to the respiratory organ, but they differ from them in being divided up into segments which are movable upon one another by means of muscles. The Visceral Arches, as these structures are called, are placed in the splanchnic meso- derm, as shown by embryology, and therefore have nothing to do with ribs, with which they have sometimes erroneously been compared. They may be described as consisting of a series of cartilaginous rods on each side, joining one another ventrally, but usually (except in the case of the first two) ending freely dorsally without connection with other skeletal structures. The first arch is called the Mandibular ; its skeleton lies near the lips and constitutes the jaws. The second is called the Hyoid Arch : it lies in the pharynx wall between the spiracle and the first branchial cleft. The remainder ^ of which there are usually five,- are the Branchial Arches : they lie in the pharynx wall between the branchial clefts, the last always occurring behind the last cleft. In Heptanchus there are seven branchial arches and seven clefts ; in Hexanchus and Chlamydoselachus there are six. The mandibular arch becomes closely associated with the cranium. It always becomes divided into two pieces : of these the dorsal piece forms the skeleton of the upper jaw and is called the palato- quadrate bar, while the ventral piece constitutes the cartilage of Meckel. The dorsal piece is longitudinally directed beneath the skull from the auditory to the ethmoidal region ; it gives articulation at its posterior (quadrate) end to Meckel's cartilage. This upper segment of the mandibular arch presents two principal arrangements in fishes. In the one of these, that which is generally called the hyostylic, its hind end is not attached to the skull directly but is held up by the stout dorsal segment of the hyoid arch, which is for this reason called the hyo-mandib- ular. This is the arrangement found in most fishes. In Chimaera and Dipnoi, however, a different arrangement is found. In these and in some extinct fishes the palato- quadrate bar is fused with the skull not only posteriorly in the auditory region, but anteriorly in the ethmoid region and in the intermediate ALIMENTARY CANAL. 63 sphenoid region. It is indeed fused all along with the side of the cranium and has the form of a laterally projecting tri- angular shelf, the projecting angle of which is the quadrate region and gives articulation to Meckel's cartilage. This arrangement is called Autostylic ; * the hyoid arch taking no part in the suspension of the upper jaw.f It has been proposed to divide fishesi nto two great groups based upon the condition of the primitive upper jaw skeleton — the Autostylici and the Hyostylici ; but as we shall explain in the sequel, there appear to be good reasons for adopting a different arrangement. In the Teleostei, Ganoidei and Dipnoi cartilage bones are developed in both the mandibular and hyoid arches to a varying extent and membrane bones may come to overlie them, largely supplementing them and even replacing them. The digestive organs vary much in structure. The mouth, which is placed at, or near, the anterior end of the head, usually has the form of a transverse slit, and can sometimes be extended forward by means of the movable supporting bones of the upper and lower jaws. The buccal cavity is distinguished by its width, and by the great number of teeth it contains, which are developed from the papillae of the mucous membrane by dentinal ossifica- tion. There are often two curved parallel rows of teeth on the upper jaw ; an outer row on the premaxilla, and an inner row on the palatine, and there may also be a median unpaired row on the vomer. On the lower jaw there is only one curved row of teeth. There may also be teeth on the* hyoid arch and on the maxillae, pterygoids, and parasphenoid, and, as a rule, on the branchial arches also, especially on the upper and lower pharyn- geal bones. The teeth may be distinguished according to their shape into pointed conical prehensile teeth, and grinding teeth. They are developed in the mucous membrane and are attached to the skeletal structures by ligament or by ankylosis. In a few cases only are they implanted in sockets. .— A small, hardly movable tongue is developed on the floor of the * The so-called amphistylic arrangement which is found in a few Elasmo- branchs (see below) would seem to be a variety of the autostylic. t The suspension of the mandibular arch, found in the skulls of Am- phibia and Sauropsida, in which the palato-quadrate bar is not attached along its whole length, but only in the auditory and ethmoid region, must be regarded as a more typical form of the autostylic arrangement than that found in Chimaera and the Dipnoi . 64 CLASS PISCES. buccal cavity, and the lateral walls of the pharynx are pierced by the gill-slits. Following the pharyngeal cavity, there is a usually short funnel-shaped oesophagus, and a large stomach, which is frequently drawn out into a caecum of considerable size (Fig. 37). Caecal appendages (pyloric appendages) are. not unfrequently met with at the entrance of the lower mid-gut (small intestine) which is marked off by a valve ; they probably serve the purpose of increasing the extent of the secreting surface of the alimentary canal. The intestine is usually several times coiled, and its internal surface is remarkable for the longitudinal folds of the mucous membrane ; villi such as are found in the higher Verte- brates are only rarely present ; but in the Selachians, Ganoids, and Dipnoi there is a peculiar spirally-coiled longitudinal fold — I'd FIG. 37. — Alimentary canal and generative organs of Clupea harengus (after Brandt). A anus ; Ap pyloric appendages ; Br gills ; D intestine ; Dp pneumatic duct ; Gp genital pore ; Oe oesophagus ; S spleen ; T testis ; V stomach ; Vd vas deferens ; Vn swimming bladder. the so-called spiral valve — which contributes essentially to the enlargement of the absorbent surfaces. A rectum is not always clearly marked off, and when present is always short, and in the Selachians it is furnished with a caecal appendage. The anus is usually situated far back, and is always ventral, and in front of the urinary and generative openings, when the latter do not lead unto the rectum (cloaca). In fishes with jugular pelvic fins, and in some Teleosteans without pelvic fins, it is situated very far forward, and may even be on the throat. Salivary glands are absent in fishes, but there is a large liver which is rich in fat and is usually provided with a gall-bladder ; there is also usually a pancreas, which is by no means replaced in Teleosteans by the pyloric appendages as was formerly be- lieved. AIR-BLADDER. RESPIRATION. 65 In many fishes the swimming bladder, an organ which by its mode of origin corresponds to the lungs, is developed as a diver- ticulum of the alimentary canal : it is sometimes closed, but sometimes remains in communication with the interior of the alimentary canal by the pneumatic duct (Physostomi) (Fig. 37 r Dp}. Its walls are formed of an external elastic membrane which is sometimes invested with muscles, and an internal mucous membrane. Glandular structures are sometimes pre- sent in the internal coat, and these may exert an influence on the enclosed air. The internal surface is usually smooth r but is sometimes provided with reticulated projections which lead to the origin of cellular cavities (some Ganoidei). Physio- logically the swimming bladder is a hydrostatic apparatus, the function of which seems to consist essentially in rendering the specific weight of the fish variable. When it is present the fish must have the power of compressing it, partly by the muscles in its walls and partly by the muscles of the body, thus rendering the body specifically heavier so that it sinks. When the compression of the muscles is removed the compressed air will again expand, the specific gravity di- minish and the fish will rise. If the anterior and posterior parts are separated and the pressure on them is unequal, then that half of the fish which is rendered specifically heavier will sink. Still more complicated relations, however, seem to exist.* Respiration is in all cases effected by gills, which may be supplemented by other structures, e.g. the lungs in the Dipnoi, and in Teleostei by vascular folds found in cavities (Fig. 38) in connection with the gill passages themselves or with the cavity beneath the operculum into which the gill-slits open. For a description of these we refer the reader to the special account of the Teleostei. The gills themselves are folds, con- taining many blood-vessels, of the mucous membrane of the passages which lead outward between the branchial arches to- open on the side of the head. These passages, which may be short and slit-like, or long and tubular, open either directly to the exterior (Elasmobranchii] or their outer openings are covered by a fold of skin generally containing cartilaginous or bony supports and called the operculum. In this case they may be- said to open into a branchial cavity which itself opens to the * See account of Teleostei. F 66 CLASS PISCES. exterior. The gills are either lamelliform (Elasmobranchii) and attached along their whole length to the interbranchial septa, or filiform and projecting (so-called pectinate gills of Teleostei, etc.). They are arranged in a row on each side of the branchial arch, so that each branchial arch carries two rows of gills (holo- branch), one on its anterior and one on its posterior face. Sometimes there is only one row (hemibranch), and some- times gills are absent on each side, or present only as a vestigial structure called a pseudobranch. The general arrangement is as follows. The branchial passage between the mandibular and hyoid arches is called the spiracle. It is always reduced and is present only in most Elasmobranchs and some Ganoids. Behind this there follow typically five branchial passages or gill- clefts. The mandibular arch never carries more than a vestige of a gill, which is called the mandibular pseudobranch v-^v \SNHJBHOHB I \ or pseudobranch of the spir- acle. The hyoid arch never carries more than a demibranch and that on its posterior face. The first four branchial arches carry typically holobranchs, while the last branchial arch is always without a gill. It thus happens that, if the hyoid arch carries a hemibranch on its hinder surface, the first four branchial passages have gills on both anterior and posterior walls, while the fifth branchial passage has a gill only on its anterior wall, the fifth branchial arch being always gill-less. In most fishes, however, the hyoid demibranch is reduced to a vestige, and is then known as the hyoidean or opercular (because the hyoid arch carries the operculum) pseudobranch. Externally projecting gills are found in the embryos of Elasmobranchs and a few Teleostei. They are not true external gills, but are much elongated internal gills. In the young Polypterus and some Dipnoi there appear to be true external gills. The brain of fishes is small and does not fill the cranial cavity. It presents all the parts of the typical vertebrate brain. BRAIX. 67 It is perhaps chiefly characterised by the small development of the cerebral or prosencephalic part of the fore-brain. The anterior end of the medullary tube becomes at an early embryonic stage, when its walls are still epithelial, differentiated by two constric- tions into three vesicles, the fore, mid, and hind cerebral vesicles. Of these the posterior vesicle or hind-brain gradually tapers behind into the spinal cord, and the portion of the medullary canal contained in it gives rise to the fourth ventricle of the adult. Its walls become transformed into the medulla oblongata, which is a development of the floor and sides of the hind-brain and is frequently called the myelencephalon. The cerebellum (meten- cephalon) is a special development of the anterior part of the dorsal wall of the hind-brain. The greater part of the dorsal wall remains throughout life at the epithelial stage and never develops nervous tissue. The mesoblast (pia mater) overlying this permanently epithelial wall becomes especially vascular and gives rise to the choroid plexus of the fourth ventricle. The middle vesicle or mid-brain (mesencephalon) gives rise by its roof and sides to the optic lobes or corpora bigemina, and by its floor to a stout nervous mass consisting largely of strong bundles of nerve fibres which in the mammalian brain constitute the crura cerebri or peduncles of the cerebrum. The portion of the medullary canal in the mid-brain is the iter a tertio ad quartum ventriculum or aqueductus sylvii. The anterior vesicle or fore-brain becomes early differentiated into three parts ; a posterior part, the thalamencephalon, the central canal of which constitutes the third ventricle ; a ventral part usually described as part of the thalamencephalon, the infundibulum ; and an anterior part the proscncephalon or cere- brum, the ventricle of which is the second ventricle. The cere- brum is usually divided into a right and left lobe by a longitu- dinal vertical constriction, but this happens rarely (Dipnoi, Marsipobranchii) in fishes, though there are sometimes indica- tions of this division in the form of a longitudinal surface groove, and in Elasmobranchs the contained ventricle is ac- tually divided into a right and left ventricle which open be- hind into the third ventricle by the foramen of Munro and are termed the lateral ventricles. The anterior end of the cerebrum is always marked off as two lobes of varying size and shape into which the second ventricle 68 CLASS PISCES. is continued (as the first ventricle) and which give off from their anterior ends the olfactory nerve fibres : they are called the olfactory lobes or the rhinencephala. In the embryonic brain before the differentiation into thalamencephalon and prosen- cephalon has been effected, the fore-brain becomes bent ventrally, forming an angle with the posterior part of the basi-cerebral axis. This bend in the cerebral axis constitutes the cranial flexure ; it takes place, roughly speaking, at the junction of the fore-brain and mid-brain and a short distance behind the front end of the notochord, the anterior end of which is under the pos- terior part of the fore-brain. The notochord is therefore in- volved in the cranial flexure and its front end becomes hook- shaped. The other organs of the anterior end of the head are also affected as is shown by the somewhat longi- tudinal disposition of the anterior gill-slits and arches (Fig. 39) as compared with the transverse disposition of those behind. In consequence of this bend in the nerve axis the anterior end of the neural tube becomes eventually directed ventralwards and, by a posterior outgrowth, backwards beneath the floor of the mid- brain : it constitutes the infundibulum, which has already been mentioned as one of the three divisions into which the fore-brain is differentiated. In addition to the infundibulum, a fourth divi- sion of the fore- brain has to be distinguished. At an early em- bryonic stage, before the prosencephalon is marked off, the anterior cerebral vesicle gives off a right and left lateral outgrowth : these are the optic vesicles. They at once become applied externally to the lateral skin of the head, and their con- nection with the brain becomes constricted to form a stalk-like structure which eventually becomes solid and forms the optic nerve. At the same time the cavity of the optic vesicle becomes obliterated by the invagination of its outer wall next the skin upon the inner wall on the brain side. This collapse of the optic vesicle, if not caused by, takes place in connection with the formation of the lens from the outer ectoderm at the point where the optic vesicle before its collapse touched the skin ; it gives rise to the formation of a two-walled optic cup, the mouth of which is occupied by the lens and the double wall of which becomes the retina of the eye. To return to the fore-brain. When the cranial flexure is established, the posterior part of its dorsal wall looks forward. CRANIAL FLEXURE. FORE-BRAIX. 69 This part becomes greatly developed and produced forwards into a large vesicle, the front part of which soon becomes marked off as the rudiment of the cerebrum or prosencephalon. This forward growth is shown by the subsequent relations of the optic nerve to have taken place behind (in the original position of the parts) the attachment of that structure (optic chiasma) to the cerebral roof (Fig. 113). If this interpretation of the complex em- bryonic growths is correct, it would appear that the cerebrum is derived from a dorsal extension of the original fore -brain just behind the point of origin of the optic nerves, and that the olfactory nerves which are developed from the front end of the cerebrum and are usually described as the first pair of cranial nerves, are in reality the second, the optic nerves being anterior to them in position. The optic nerves then are attached to the roof of the original fore-brain at about halfway between its front and hind ends. A ganglionic mass is formed on each side at this point in the side wralls of the fore-brain ; these great ganglia are the optic thalami and lie in the adult brain at the side of the third ven- tricle, constituting the chief bulk of the thalamencephalon. Immediately in front of the optic thalami two great ganglionic developments are formed in the ventral wall of the cerebral out- growth : these are the corpora striata found in the adult on the floor of the second ventricle or, if it is divided, of the lateral ventricles. The front end of the cerebral outgrowth also gives rise to nervous tissue of the olfactory lobes. We now come to the roof of the reconstituted fore-brain, after the cerebral outgrowth has been formed.* This roof is divided into two parts by the velum transversum (see below). Of these the posterior part, the part overlying the third ventricle, remains in all Vertebrates almost entirely in an epithelial condition. It gives rise by its posterior part to the epiphysis or pineal body. To this we shall return shortly. The anterior part, the part belonging to the cerebral rudiment, is called the pallium. It is marked off from the posterior part by a transversely directed fold of the epithelial roof. This fold dips down into the ventricle at the junction of the thalamencephalon and prosencephalon and encloses between its two laminae a vascular development of the pia mater, which is always present and gives rise in the * Mino; , American Journal of Anatomy, 1, 1901, p. 81. 70 CLASS PISCES. higher brains to the choroid plexuses of the lateral ventricles. It is called the velum transversum. This folded-in part of the roof is generally regarded as belong- ing to the cerebrum : in all Vertebrata it retains throughout life its epithelial condition. In front of it, the roof of the brain (pallium) behaves in a different way in different animals. In Elasmobranchs, Marsipobranchs, Dipnoi, and in all Vertebrata above fishes, it loses its simple epithelial condition and develops nervous tissue, forming the dorso -lateral part of the cerebral hemispheres above the lateral ventricles. In most other fishes, Teleosteans, and Ganoids, the pallium retains its epithelial condition throughout life, so that in these groups the roof of the lateral ventricle remains permanently thin and epithelial, as does the roof of the third ventricle and that of the posterior part of the fourth ventricle, and in Lampreys of the aqueductus sylvii as well. Curiously enough — for what reason it is difficult to under- stand— the condition in which the cerebral pallium consists of a thin epithelial layer is regarded as secondary. By all the or- dinary tests which are applied in speculations of this kind — viz., embryonic development and general diffusion of the character in the lower Vertebrata and absence in the higher, it should surely be regarded as a primitive character. Indeed, if we may be allowed to indulge in a little speculation of this kind, it would appear from development that the whole medullary canal at one time had purely epithelial walls, and there appears to be a tendency to the retention of this character along the middle dorsal line throughout life in all Vertebrata. To return to the pineal body. It is developed as a divert- iculum of the hinder part of the roof of the thalamencephalon. Its terminal portion becomes the pineal body or epiphysis ; its proximal part is the pineal stalk. The terminal part sometimes gives rise on its anterior wall to an outgrowth which is called the parietal organ. The parietal organ may be developed separately from the brain roof just in front of or by the side of the epiphysis.* It is not always formed, and usually van- ishes with later growth ; but in Lampreys and Lizards it * On account of this fact it has been suggested that the epiphysis is really a paired organ, one of the pair becoming the actual epiphysis (pineal body) of the adult, and the other either degenerating or becoming PARIETAL ORGAN. 71 persists and assumes a peculiar structure resembling that of the retina of the eye. For this reason it has been called the pineal eye. In Lampreys the pineal body also assumes the same structure. A great deal of significance has been attached to the curious eye-like structure which is assumed by the parietal organ. It has been regarded as the vestige of an unpaired eye. In our opinion the resemblance to an eye is accidental, but for a discussion of the question we refer the reader to the account of the parietal organ in the section devoted to Reptilia. Fia. 39. — Heads of young Elasmobranch embryos (Scyllium canicula) (after Sedgwick). A. Ventral view of head of embryo, 7 mm. in length, with two open pharyngeal clefts. The mouth is present as a longitudinal groove in the ectoderm of the buccal depression. B. Same view of a slightly older embryo ; the buccal groove has become a longitudinal slit. C. Side view of head of embryo, 9 mm. in length, with three open slits. D. Side view of head of embryo, 11 mm. in length ; rudiments of external gills have appeared on the hyoid and on the first and second branchial arches. E. Side view of head of embryo of 16 mm. ; external gills have appeared on mandibular arch and the angle of the jaw is marked. 1 mandibular arch ; 2 angle of jaw ; 3 second pharyngeal cleft ; 4 nasal pit ; 5 eye ; 6, midbrain ; 7, auditory sac ; 8 hyoid arch ; 9 spiracle. The superior commissure is a small nervous development in the otherwise epithelial roof of the third ventricle just in front of the attachment of the pineal stalk. The posterior commissure, the so-called parietal organ (pineal eye). See Dendy, Q. J. M. S., 42, 1899, p. 111. This view is supported to a certain extent by the arrange- ment in the lamprey (see p. 106). 72 CLASS PISCES. which probably belongs to the mesencephalon, is just behind the attachment of the pineal stalk. The paraphysis is the recess in the roof of the cerebrum caused by and just in front of the velum transversum. By some mor- phologists it is regarded as a special glandular organ, the secretion of which passes into the ventricle. It is not always present as a distinct structure. The pituitary body or hypophysis develops as an evagination of the front part of the buccal cavity. It is indeed the anterior part of this cavity. In Elasmobranchs the original buccal slit — for the vertebrate mouth perforation has at first the form of a longitudinally extended slit (Fig. 39) — is continued into it. It is applied to the infundibulum and eventually becomes cut off from the mouth, except in Polypterus and Calamoichthys in which the buccal opening is retained throughout life. At the point in the embryo where the pituitary rudiment meets the infundibulum there is a close approximation and partial fusion of three other organs, viz., the front end of the gut, the anterior end of the notochord, and the median part of the premandibular somite (preoral coelom). The lobi inferior es and saccus vas- culosus are parts of the infundibulum, very generally present in fishes. The former are lateral diverticula or thickenings of the infundibulum ; while the saccus vasculosus, or infundibular gland, is a glandular dilation of its end, where it is in contact with the pituitary body. With regard to the cranial nerves, it ought to be noticed that they all arise from the walls of the mid- and hind-brain, except the olfactory and optic nerves. These come off from the pre- notochordal part of the fore-brain, and it is doubtless to them that the fore-brain owes its relatively enormous development in all Vertebrata. The other cranial nerves, from the third nerve onwards, probably all belong to the series of nerves which is continued along the spinal cord as the spinal nerves. Indications of this are shown by a careful study of the early stages of their growth, particularly in Elasmobranch embryos, in which they appear to be associated with the cephalic segments of the coelom. These segments, for a knowledge of which we are indebted to Balfour* and to the later researches of Van Wijhe.f are as follows : The first cranial * A Monograph of the Development of Elasmobranch Fishes, London, 1878. f " Ueber die Mesodermsegmente und die Entwickelung der Nerven des Solachierkopfes," Verhandl. der k.Acad. d. Wissensch. zu Amsterdam, 1882. CRANIAL SEGMENTS. 73 segment is represented by the premandibular somite* — an unpaired sac with epithelial walls, immediately in front of the notochord (preoral head cavity, vide p. 8). The walls of this sac give risa to all the eye- muscles except the superior oblique and external rectus, and to mesen- chyme. Its cavity vanishes, as do the cavities of all the cranial segments. The nerves are the ram us ophthalmicus profundus, which develops from the nerve crest immediately in front of the trigerninal and represents the dorsal root, and the third nerve which represents the ventral root. These two roots are both connected to the ciliary ganglion (see accoimtof cranial nerves under Ela.smobranchii). The second segment is the mandibular somite which is dilated in its dorsal muscle-plate region and extends ventrally to the lower end of the mandibular arch (collar-somite, vide p. 7). The walls of this sac give rise dorsally to the superior oblique muscle and ventrally to the mesen- chyme and muscles of the mandibular arch. The nerves are the trigerninal and the fourth, the latter being regarded as an abnormally situated anterior root. The first two cranial somites were discovered by Balfour. The third and following segments are represented only by their dorsal muscle-plate sections, the ventral portions being merged in the continuous splanchnocoel (pericardial division). These segments may be regarded as the anterior of the trunk series of Amphioxus. They do not apparently communicate with the ventral splanchnocoel, which in this region under- goes a pseudo-segmentation in consequence of the formation of the gill- pouches. These pseudo-segments, or hyoid- and branchial- arch cavities, open ventrally into the pericardium, of which they are a part, and were taken by Balfour for the posterior cranial segments. The first of these muscle-plates, which may be called, from its position, the hyoid myotome, is better developed than the rest and gives rise to the external rectus muscle. It was observed by Balfour. The next three, which were discovered by v. Wijhe, are very faintly marked and transient and give rise to no muscles, in correspondence with which fact may be noted the absence of ventral roots. The next three segments (seventh-ninth) are represented by well-developed muscle plates which persist and give rise to muscles. The nerves of the hyoid segment (third) are the facial (dorsal root) and sixth (ventral root). The nerves of the next three segments are supposed to be represented by the auditory, glossopharyngeal and vagus and are without ventral roots in correspondence with the absence of myotome muscles. In the last three (or sometimes more) cranial seg- ments, dorsal roots are present only in the embryo for a short time, but ventral roots are developed, supplying presumably the myotome muscles of this region and called the occipito-spinal nerves. These were mistaken by Gegenbaur, who did not study the embryo and observe the transient dorsal roots belonging to them, for ventral roots of the vagus. The following table represents in brief the view of cranial segmentation which has just been described. f * There is in some forms a pair of head cavities in front of the pre- mandibular somite. These are sometimes in communication with and developed as diverticula of the premandibular somite, but in Acanthias they are said to be independent of it (J. B. Platt, Journal of Morvholoav . 5, 1891, p. 79). t The view here given takes no account of the scheme given on p. 77, according to which the cranial nerves were originally tripartite, consisting of dorsal, lateral and ventral roots. It was formulated before the modern 74 CLASS PISCES. Coelomic Sac Nerve Dorsal Root Ventral Root Cranial segment 1 Premandibular so- Ramus ophthal- Third nerve. mite. Its walls micus profun- , give rise to all the dus. eye muscles sup- plied by the third nerve. 2 Mandibular somite. Trigeminal. Fourth nerve. The walls of its dorsal part give rise to the superior oblique muscle. 3 Hyoid muscle plate Facial. Sixth nerve. giving rise to ex- ternal rectus mus- cle. 4 Muscle plate (tran- Auditory. None. sient). 5 Muscle plate (tran- Glossopharyn- None. sient). geal. 6 Muscle plate (tran- Vagus. None. sient). 7 ^ Muscle plate per- 'j (Spino-occipital 8 ^ sistent and giving -None. nerves (so- 9 j rise to muscles. IJ called ventral and sometimes more J vagus roots). The spinal nerves have two roots, which unite, and the dorsal of which has a ganglion. The ganglion may, however, be placed at the junction of the two roots. The cranial nerves are in ten pairs, but there are often some small nerves — the spino- occipital nerves (miscalled ventral vagus roots) — arising from the ventral side at the hind end of the medulla. They pass out through foramina in the skull, but are perhaps better regarded as anterior spinal nerves (see above) the dorsal roots of which are not developed. analysis of the cranial nerves, which is due to Gaskell and is there referred to, was fully developed. The later work, which is still incomplete, may very possibly necessitate a new scheme of cranial segmentation, and the groups of cranial nerves expressed by the terms fifth, seventh, ninth and tenth may possibly be found to be connected with the ventral segmenta- tion (pseudo-segmentation of the text) of the branchial pouches, and to be independent of the mesoblastic segmentation which is so conspicuous a feature in the trunk. CRANIAL NERVES. SYMPATHEMIC. 75 For an account of a typical arrangement of the cranial nerves the reader is referred to the section on Elasmobranchii. Their arrangement in fishes differs from that in higher types, mainly on account of the presence of the lateral line sense organs. The nerves to these appear to arise from a special part of the brain, the tuberculum acusticum, from which the auditory nerves also arise. They are associated in their course to the periphery with the seventh and tenth nerves, and constitute the acustico-lateralis system. The fibres of this system which run with the fifth and ninth are derived from these two nerves. The nerves which pass from the facial roots to the fifth nerve cause an intermingling of the roots of these two nerves, which is not easy to unravel, and which is characteristic of fishes. A sympathetic nervous system appears to be present. In Marsipobranchii, in which all the nerves are without a medullary sheath, it cannot be fully traced, but the spinal nerves give off branches which pass to the viscera, where small ganglia are found. In other Pisces there is a series of sympathetic ganglia which develop as outgrowths of the spinal nerves, becoming detached from the rudiments of the spinal ganglia at an early stage. These ganglia are usually connected by longitudinal commissures, but though regularly developed, their arrange- ment is not easy to trace in the adult. In Elasmobranchii * the system tends to take a plexiform structure, and lies in the neighbourhood of the cardinal veins. There is an especially large ganglion at about the level of the ductus cuvieri ; this is supplied by a number of spinal nerves, and gives off several branches, which are distributed to the viscera with the coeliac artery. The system appears not to extend into the head. In Teleostei there is a definite chain of small ganglia on each side of the vertebral column. In these forms it is continued into the head, where it is connected with the trigeminal nerve and ciliary ganglion, and into the tail, where it runs in the caudal canal. The analysis of the nerves, which is the outcome of the recent work f of morphologists and physiologists, is beyond the scope of this work, but the following points may be noted here : * R. Chevrel, " Surranatomiedusystemenerveuxgrandesympathetique des Elasmobranches et des poissons osseux." Arch. Zool. Exp. (2) 5, supplement. | W. H. Gaskell, "The structure and function of the nerves which inner. iO CLASS PISCES. Five kinds of nerve fibres, characterised by their structure, function and distribution, may be distinguished. 1. The system, of the somatic sensory (afferent) fibres. These include the largest heavily medullated fibres which terminate in the skin and myotome muscles. They pass out by the dorsal roots in the cord, and by the roots of the trigeminal in the brain ; * their ganglia being the posterior root ganglia (spinal), and the gasserian. 2. The somatic motor (efferent) system. The fibres of this system are also large and heavily medullated : they terminate in the myotome (somatic) striated muscles ; i.e. the muscles derived from the muscle- plate?, including those derived from the dorsal part of the mandibular and from the premandibular somites. They pass out by the anterior roots in the cord, and by the third, fourth, and sixth cranial nerves, and are without peripheral ganglia. 3. The visceral sensory (afferent) system (communis system}. Tho fibres of this system are smaller, and they are distributed to the internal mucous surfaces. They leave the cord by the posterior roots, their ganglia here being posterior root ganglia. The cranial fibres of this system are present in the roots of the fifth, seventh, ninth and tenth nerves ; the ganglia being the gasserian (in part), geniculate, glossopharyngeal and jugular ganglia. Their destination is mainly the mucous surfaces of the anterior part of the alimentary canal. 4. The visceral motor (efferent) system. The fibres of this system may be subdivided into (a) those which innervate the striated voluntary muscles (mesenchymatous) of the anterior part of the alimentary canal (mandibular, branchial, and facial muscles), are fairly large, and are non- ganglionated, and (6) those which supply the unstriped muscles through- out the body (blood-vessels, gut-wall, skin, etc.). The latter are small fibres, all of which pass through peripheral ganglia. The true motor- fibres to the muscles of the small intestine and anterior part of the alimen- tary canal and its appendages (lungs, etc.) are derived from cranial nerves, whereas the fibres to blood vessels, skin, walls of Miillerian and kidney ducts come from the spinal cord by anterior roots. The cranial nerves contain no vaso-motor fibres. 5. The acustico-lateral system (see p. 75). This system consists of large fibres and passes out exclusively in the roots of the seventh, eighth, tenth and possibly ninth cranial nerve. It is absent in the trunk and in higher Vertebrates, except in the auditory nerve, and is distributed only to the membranous labyrinth and the lateral line sense organs (lateral line, ampullae and pit organs). From this account it will be gathered that in the head the visceral motor (efferent) fibres travel out with the visceral sensory fibres and in the case of the fifth with the somatic sensory fibres as well, the somatic motor fibres being distinct ; whereas in the trunk they leave the cord with the somatic motor fibres. vate the visceral and vascular systems." Journal of Physiology, 1, 1886; and "On the cranial nerves" in Journal of Physiology, 10, 1889. O. S. Strong, " The cranial nerves of Amphibia," Journal of Morphology, 10, 1895; C. J. Hernck, " The cranial and first spinal nerves of Amphibia," Journal of Comp. Neurology, 9, 1899, p. 157. * The third nerve appears to contain fibres belonging to this system (muscular sense), and it is possible that a few of them may be contained in the vagus (Arnold's nerve). SENSE-ORGANS. 77 To summarise the matter Gaskell has suggested that in the primitive condition, both in brain and spinal cord, there were three rows of nerve roots : (1) a dorsal containing somatic sensory fibres, (2) a ventral with somatic motor fibres, and (3) a lateral row containing both visceral sensory and visceral motor. This condition is modified in all existing forms in the cord by the splitting of the lateral roots in such a way that the visceral sensory roots have joined (1) and the visceral motor roots (2) ; whereas in the brain the roots of the lateral row have persisted and the somatic sen- sory roots (restricted to one) have joined them, the somatic motor roots (three in number) remaining distinct. This scheme does not, however, take account of the acustico -lateral system. The Eyes have a flat cornea, and a large almost spherical lens, the anterior part of which projects far out of the pupil. Movable eyelids are present in ElasmobranchiL but are absent from most other fishes. There are no lacrymal glands. The usual eye muscles are present. There is frequently a rete mirabile, the choroid gland, on the ophthalmic artery as it enters the eye near the entrance of the optic nerve. The processus falciformis and campanula halleri are described below under Teleostei. The eyes are much reduced and functionless in most adult Marsipobranchs and some cave-dwelling and ; abyssal Teleosteans. The Auditory Organ consists of the otocyst or membranous labyrinth, which is embedded in the side walls of the auditory region of the skull. It lies in a cavity which is closed from the cranial cavity in most Elasmobranchs, but communicates with the latter in Chimaera, and Teleosteans, Ganoids and Dipnoi. It consists (Fig. 40) of a central chamber, the vestibule, and of three semicircular canals opening into the vestibule. The vestibule is divided into two parts by a constriction ; of these the upper is the utricle, the lower the saccule. The semicircular canals open into the utricle while the saccule in some fishes gives off from its posterior end a process called the lagena, which is an incipient cochlea. In Chimaera and the Squall the ductus endolymphaticus which is given off by the saccule opens on the surface of the head.* Both saccule and utricle contain a chalky mass of otoliths. When the lagena is well marked its papilla acustica (pi) becomes separate from the macula acustica sacculi and receives a separate nerve (vide account of membranous labyrinth under Teleostei}. The membranous labyrinth enters into peculiar relations with * This is the remains of the aperture of invagination of the embryo. 78 CLASS PISCES. -ha, B the air bladder in some Teleostei, which are fully described in the account of that sub-class. The Olfactory Organs are a pair of simple pits or sacs, in the lining of which the fibres of the olfactory nerve terminate. In the Marsipobranchii the olfactory organ is partly single and presents peculiar re- lations (see account of Marsipobranchii ) . In other fishes each sac usually has two openings, both of which are external except in Dipnoi. In Elasmobra n c h s there is usually only one opening. The internal surface of the sacs is generally increased by folds of the mucous mem brane. We know practi- cally nothing about FIG. 40. — Right membranous labyrinth of Chimaera, seen f,hp <;f»nsf» of taste from the median side (from Wiedersheim, after Retzius). aa ampulla of anterior vertical canal ; etc auditory nerve ; The tactile Sense is ade opening of ductus endolymphaticus ; ae ampulla of " horizontal canal ; ap ampulla of posterior vertical canal ; J^Q doubt Specially ass process of the sinus utriculi ; ca, anterior, c-p posterior vertical canal ; ce horizontal canal ; cr crista acustica served bv the lips ampullae ; de ductus endolymphaticus, which opens at ade •* through the skin ha ; mn macula acustica neglecta ; ms and their appendages, macula acustica sacculi (the macula ac. utr. rec is on the other side and not properly visible) ; pi papilla acustica and bv special parts lagenae (the lagena, however, can hardly be said to be J present in this form) ; branches of auditory nerve rota to of the appendages anterior ampulla, rae to horizontal ampulla, rap to pos- terior ampulla, rec to macula acustici utriculi, rs to which are riclllv macula sacculi and lagenae, ru to macula recessus utriculi ; s saccule ; sp sinus utriculi posterior, ss sinus utriculi innervated (e.g. superior ; u utricle. Trigla.) The system of embedded epidermal* sense organs which are found in all fishes requires a detailed description. * Levdig, Lehrbuch d. Histologie des Menschen u. d. Thiere, 1857. Solger, rec rs LATERAL LINE. 79 In Elasmobranchs four kinds of organs are included under this head : (1) the lateral line proper, or mucous canals, with its cephalic ramifications ; (2) the ampullary canals, or Lorenzini's ampullae ; (3) Savi's vesicles ; (4) pit organs. The essential C$0 sot om Flo. 41. — Diagram illustrating the distribution of the dorsal branches of the cranial nerves and of the lateral line canals, and the position of the groups of ampullae in an Elasmobranch (after Ewart,.from Gegenbaur). A auditory nerve with labyrinth ; it also points to the groups of Lorenzini's ampullae ; Bu buccal branch of facial ; bu inner branch to part of infraorbital canal, and to the inner buccal group of ampullae, bu' its outer branch which supplies part of the infraorbital canal, and the outer buccal group of ampullae ; ch post- branchial branch of facial to mucous membrane, and giving off motor branches to some jaw muscles ; CSO, CSO supraorbital canal ; CJO, CJO infraorbital canal ; Fa, Fa' roots of facial nerve ; Gp glossopharyngeal, arising under cover of the lateralis branch of the vagus nerve ; Hm hyomandibular canal arising from the infraorbital, and giving off the mandibular canal, the mandibular group of ampullae is in the angle between these two ; Hm' branch of the hyomandibular nerve to the hyoid group of ampullae ; in intestinal branch of vagus with ganglion, where it separates from fourth branchial branch ; In lateralis branch of vagus nerve ; m mouth ; N nasal sac ; om deep branch of oculomotor giving off short root of ciliary ganglion (shown, but not marked), the long root is also shown, as are the short ciliary nerves to the eye ; opr root of ophthalmicus profundus ; opv dorsal branch of same, giving off long ciliary nerves ; pol second branch of lateralis supplying some lateral line sense organs and a row of pit organs, the first branch supplies the commissure connecting the two lateral canals, and some sense organs of the main canal ; sof ophthalmicus superficialis facialis, which supplies the supraqrbital canal, and the superficial ophthalmic group of ampullae ; sot ophthalmicus superficialis trigemini ; it arises from the gasserian ganglion ; sp spiracle ; Tr trigeminus ; V1, V2, V3 the first three branchial branches of the vagus nerve, each with a ganglion and with pharyngeal, prebranchial and post-branchial branches ; V* the united fourth branchial branch of vagus and intestinal branch ; 1-5 gill-slits. part of these organs seems in all cases to be sensory patches of the epidermis, consisting of sensory cells, bearing short sensory hairs, and of supporting cells. Neue Untersuchungen zur Anatomie der Seitenorgane der Fische, Arch. f. mic. Anat., 1879-80. Allis, Anatomy and Development of Lateral Line system in Amia, Journ. Morphology, 2, 1889. Fritsch, Die electrischen Fische, Leipzig, 1890. Ewart, The sensory canals of Laemargus, Trans. Roy. Soc. Edinburgh, 37, p. 59, 1891 ; and The sensory canals of the skate, Ibid. Pollard, The lateral line system in Siluroids, Zool. Jahrb. 5, 1892. Cole, On the cranial nerves and lateral sense organs of fishes, Trans. Linnean Soc., 1898. 80 CLASS PISCES. In the lateral line system these sensory patches are modifi- cations of the lining epithelium of a canal, which extends the whole length of the body and on to the head, where it branches in a somewhat complicated manner. The canals lie in the dermis or deeper in the subcutaneous tissue, and their walls contain either stiff connective tissue or cartilage (skates) for the purpose of keeping them permanently open. They communicate at intervals with the exterior by tubules. The trunk section of the canal usually lies at the junction of the dorsal and ventral divisions of the lateral muscles. The sense organs and the tubules seem to be usually metamerically arranged in the trunk, and the sense organs and tubules correspond, but in the head the metameric arrangement is of course out of the question, and the sense organs appear to be more numerous than the tubules. That this system has originated from a skin groove is indicated by its development and by the fact that in some Elasmobranchs it has the form of an open groove throughout life. In Chlamy- doselachus it has the form of a groove guarded by overlapping scales. In Chimaera it is also an open groove, though in the head the lips of the groove tend to approximate over the sense organs (Fig. 42.) In Heptartchus it is a groove in the greater part of the trunk, but closes into a canal in front and on the head. The course of the cephalic portion in a typical case is shown in Fig. 41. The lateral 'canal on reaching the head is connected with its fellow of the opposite side by a cross canal — the commissural canal — which may pass in front of or behind the openings of the otocysts. A short distance in front of this it branches into a canal passing above the eye — the supraorbital canal (CSO] and one passing below the eye, the infraorbital canal (CJO). The supraorbital canal extends to the front end of the snout and then passes back to join the infraorbital canaL The infraorbital canal gives off a branch back to the hyoid region, called the hyomandibular canal (Hm), which itself gives off a branch to the mandible. In Chimaera (Fig. 42) the arrangement is very similar. In skates the hyomandibular canal is enormously extended backwards in a loop which lies partly on the dorsal and partly on the ventral surface of the pectoral fin, and communicates with the exterior by rather long tubules. In the same animal the lateral line canal near the head gives off LORENZINl'S AMPULLAE. 81 two long canals which pass backwards and outwards on the dorsal side o the fin ; the anterior of these anastomoses with the dorsal part of the above described extension of the hyomandibular canal. The whole of this system of canals is in Elasmobranchs supplied by the facial nerve and the lateralis branch of the vagus, which probably belongs to the facial system (see account of cranial nerves under Elasmobranchii and Fig. 41). The ampullary canals or Lorenzini's ampullae, are un- branched canals (Fig. 43), opening, usually in groups, on the surface of the head and ending internally in vesicles — the ampullae — which are beset with radial dilatations (Fig. 44). The ampullae are placed in groups, the position of which in a typical case is shown in Fig. 41. FIG. 42. — Cephalic lateral line of Chimaera (from Gegenbaur). a lateral groove of trunk, b, e" infraorbital, c supraorbital groove, c' supraorbital grooves ; passing back to join infraorbital ; x frontal appendage. . The sensoiy epithelium is confined to the ampulla to which the nerves, in all cases branches of the facial, are distributed. The tubes and ampullae contain a gelatinous matter. Savi's vesicles are found in Torpedo round the electrical organs They are completely closed. The pit organs, found in many Elasmobranchs are sense organs sunk in pits on different parts of the head and trunk, and are supplied by the facial nerve, the lateralis of the vagus, and the trigeminal. In Teleosteans, Ganoids and Dipnoi the lateral line system and the pit organs alone are present. The lateral line has an arrangement very similar to that described for Elasmobranchs, z — ii G 82 CLASS PISCES. FIG. 43. — A portion of the snout of Scyllium in section, show- ing ampullary tubes (from Gegenbaur). N nerve ; a ampullae ; c epidermis ; t tubes ; c' dermis ; o openings of the tubes ; a' passage of a tube through the dermis. but the canal wall is sometimes ossified, especially on the head, and the ossifications may be fused with the dermal and cranial bones. Very often the canal traverses the scales and bones, and the sense organs are contained in the osseous tissue. In such cases the lateral tubules, which are in some cases branched, their openings forming so-called cluster pores, per- forate the bone, as does the nerve going to the sense organs. In this way certain scales on the body and bones of the head may acquire a special relation to these organs. Pit organs are present both on the trunk and head and often lie along the course of the main canals. In a few cases (Esox, Gobius, Liparis, etc.) the cephalic canals are alone present, the sense organs in the trunk being isolated and not connected by a longitudinal canal. In a few cases the longitudinal canal may have the form of a groove for a part of its extent. The openings of the lateral tubules may occur between the scales as well as upon them. In addition to the innervation found in Elasmobranchs the glossopharyngeal frequently sends a branch to a few of the posterior cephalic sense organs ; and it has been stated that the ophthalmicus super- ficialis trigemini also takes part, but this must be regarded as doubtful. In any case the nerves innervating this system of lateral line sense organs can always be FIQ traced to the special centre in the brain from which the auditory nerve arises. The pit organs are innervated by the trigemi- nal as well as by the facial and lateralis of the vagus. 44. — Lorenzini's am- pulla. A from the side with nerve n and portion of tube c ; B in section (from Gegenbaup). ELECTRICAL ORGANS. 8B EO Electrical organs,* the function of which is to develop con- siderable quantities of electricity, are found in some fishes. They occur in different parts of the body and in fishes belong- ing to quite different groups (e.g. Torpedo and Hypnos among Elasmobranchs, and Gymnotus and Malapterurus among Teleoste a n s ) . They differ both in struc- ture and po- sition in the body, but they always consist o f peculiarly modified cross- striped muscu- lar tissue. In Torpedo (Fig. 45) they are placed be- tween the bran- chial pouches and the anterior cartilages of the pectoral fins, and occupy the whole space be- tween the dor- sal and ventral integument. They consist of vertically ar- ran g e d col- umns, sup- ported by walls of connective tissue, and divided by horizontal septa of the same material into a number of compartments placed one FIG. 45. — Torpedo with electric organ EO and brain exposed (after Gegenbaur), dorsal view. On the right side the dorsal surface only of the organ is exposed ; on the left the nerves which supply it are shown, br branchial sacs ; Gr sensory canal tubes of the skin ; Le electric lobe of the brain ; o eye ; Tr trigeminal nerve ; V vagus nerve. * Fritsch, Die electr. Fische, Abt. 1 and 2, Leipzig, 1890. Ballowitz, Electr. Organ v. Torpedo, Arch. f. mic. Anat., 42, 1893. Sanderson and Gotch, Elect. Organ of Skate, Journ. Physiology, 10, 1889. Ewart, Electric Organ of Skate, Phil. Trans., 1888 * 1892. 84 CLASS PISCES. above the other. Each compartment is filled with gelatin- ous tissue, through the middle of which runs a horizontal plate composed of a finely granular nucleated substance and of numer- ous nerve-endings. This is the electrical plate. The electrical plates correspond to a certain degree to the copper and zinc ele- ments of the voltaic pile, the gelatinous matter representing the moist intermediate layers ; while the connective tissue frame- work serves to hold the parts together and to carry the blood- vessels and nerves. The face of the plate on which the nerves ramify is the same in all the columns of the same organ, and is always electro -negative, the other surface being positive. In Torpedo the nerve enters on the lower surface of the plates, the upper surface is therefore electro- positive. The organ is supplied by five strong nerves, of which the anterior is a branch of the facial, the four posterior being branches of the vagus group. In the electric Teleostei the electric organs are placed in the trunk and tail, and are supplied by spinal nerves. They are simi- larly constituted, but the col- umns are horizontally placed. In Malapterurus they lie along the body beneath the skin, and the posterior surface of the plates, the surface on which the nerve enters, is electro-positive. This apparent exception is explained by the fact that the nerves pass through the plate and are distributed on the anterior surface, which is electro-nega- tive. In the electric eel (Gymnotus electricus) the electric organ lies at the side of the tail, and consists of long horizontal columns (Fig. 46). The so-called pseudo-electric organs found in the tail of Raja and of Mormyrus have a similar structure, but manifest only feeble electric phenomena. They constitute a very good example of an organ which is practically of no use to its possessor, and which we should entirely fail to understand the meaning of were it not for the cases in which the electric organ is fully developed. VASCULAR SYSTEM. — The blood is generally red ; it is white FIG. 46. — Longitudinal section through two columns of the electric organ of Gymnotus. a horizontal partition ; I transverse partition walls, convex headwards ; e electric plates (from Gegenbaur, after Max Schultze). VASCULAR SYSTEM. 85 only in the Leptocephalidae (larvae of the eels) ; it circulates in a closed vascular system, in which a muscular pulsating region or heart is present. The heart (Fig. 47) is placed far forward on the throat, ventral to the branchial framework, and is enclosed in a pericardium, the cavity of which communicates with the body cavity in some Plagiostomes, Chimaera, Acipenser, etc. It is a simple venous branchial heart, and is composed of a sinus venosus, a thin-walled large auricle and a very powerful muscular ventricle. The sinus venosus receives the venous blood returning from the body, and the ventricle forces it through the ventral aorta to the respiratory organs. The aorta begins with a bulbous swelling (bulbus arteriosus} which in the Ganoids, Plagiostomes and Dipnoi is replaced by an inde- pendently pulsating part of the heart, with rows of semi-lunar valves (conus arteriosus}. While the fishes with a simple non-mus- cular bulbus arteriosus have but two semi-lunar valves at its origin, the above-mentioned orders usually have two or four, or rarely five, rows of three, four, or more valves each, in the conus The aorta at once divides into a number of paired vascular arches, corresponding to the embryonic aortic arches. These are the bran- chial arteries ; they pass into the branchial arches and give off branches to form the capillary net- works of the gills. From the capillary networks small vessels pass out which unite to form a larger vessel in each branchial arch (epibranchial or efferent branchial artery). The arrange- arf-Arin«n« Fm- 47.— Diagram of the circulation of US< a Teleostean. Ab arterial arches; Ao aorta descendens into which the epibranchial arteries passing out from the gills unite ; Ba ventral aorta with the arterial arches which carry the blood to the gills ; D intestine ; Lk portal circulation ; N kidneys ; V ventricle. The branchial capillary system is omitted. 86 CLASS PISCES. ment of these vessels corresponds to that of the afferent branchial arteries ; they unite to form the large aorta descendens or dorsal aorta. Before they unite the cephalic arteries pass off from the efferent vessels of the anterior arch. The arrangement of the principal venous trunks in fishes is most nearly related to the embryonic condition. Corresponding to the four cardinal veins of the embryo, two anterior and two posterior cardinal veins bring back the blood from the anterior and posterior part of the body respectively. These veins unite on each side to form two transverse veins — the ductus Cuvieri — which enter the sinus venosus of the heart. The course of the returning venous blood is complicated by the insertion of a double portal circulation. The caudal vein does not pass directly into the posterior cardinal veins, but breaks up into capillaries in the kidneys, from which the blood passes into the posterior cardinal veins. There is thus a renal-portal circulation. For the hepatic portal circulation on the other hand the venous blood of the intestine is used ; this blood after passing through the capillaries of the liver is returned to the heart by one or more veins, which open into the sinus venosus between the two ductus Cuvieri. Such capillary systems must be a considerable hindrance to the circulation of the blood and explain the development of the so- called accessory hearts on the caudal vein of the eel and on the portal vein of Myxine. The urinogenital organs are described under the different sub- classes. With regard to them the following general remarks may be made. A pronephros is present and functional in the larva of all fishes except Elasmobranchii, in which there is no larval stage. It has been maintained, and a great deal has been written on the subject, that there is a vestige of a pronephros in the embryos of Elasmobranchs ; but if there is it is very feebly developed and never possesses a glomerulus. The pronephros is the anterior and first developed portion of a longitudinal gland, which extends, in the embryo at least, the whole length of the body cavity from the pericardium to the hind end. This extended excretory organ consists of nephridia, which in Elasmobranchs are developed, as was first shown by Sedgwick,* from the portions of the body cavity which connect the lower ends of the muscle plate cavities with the general body * Q. J. M. S., 20, 1880, p. 164. EXCRETORY ORGAXS. 87 cavity. These portions of the body cavity are called nephro- tomes, and accurately correspond at first with the segments of the embryonic muscular system. The pronephros is the anterior end of this excretory organ, which is developed before the rest to meet larval needs. The serial homology between the prone- phros and the hinder part of the excretory system was for many years denied, partly because of a certain difference in structure and partly because there is usually a gap between it and the front end of the rest of the organ. But the differences in struc- ture are very small, in some cases indeed (e.g. Lepidosteus) do not exist ; and vestigial nephridia have been found in the gap between the two organs (cf. especially Price's researches on the development of the excretory organs of Bdellostoma). Finally, Brauer's recent researches on the development of the excretory organs of Gymnophiona remove all doubt on the point. The hinder part of the excretory system differs from the pronephros mainly in the fact that the glomerulus — the vas- cular tuft which secretes the fluid part of the urinary excre- tion— is segmented into portions, one for each nephridium (or kidney tubule), instead of forming a continuous structure as in the pronephros ; and the portion of the body cavity containing each of these is partly shut off from the rest to form the malpighian body of the kidney tubule. This malpighian capsule, however, frequently, though not always, retains its connection with the rest of the body cavity by the so-called nephrostome. The internal opening is retained in most Elasmobranchs, but is lost in Teleostei, Dipnoi, Ganoidei and Marsipobranchii. This hinder part of the excretory system becomes in Elasmobranchii much reduced in front and largely developed behind. In consequence of this it is described as consisting of the mesonephros in front and the metanephros behind ; but this differentiation is not found in other fishes. As in all Vertebrates, the longitudinal duct (archinephric duct) is the first part of the excretory organs to appear. The prone- phros is developed in connection with the front end of this duct, so that the duct is at first the pronephric duct. The Miillerian duct is found in all Pisces with the probable exception of the Marsipobranchii and the Teleostei, and in all cases it becomes the oviduct in the female, but is reduced in the adult male. In Ganoids the longitudinal duct joins the oviduct (Miillerian) 88 CLASS PISCES. before opening externally, but in Elasmobranchii the two ducts open separately into the cloaca. The development of the Miil- lerian duct is known only in Elasmobranchii. It there arises in connection with the first establishment of the longitudinal duct as an evagination of the parietal mesoderm of one of the anterior nephrotomes, so that it at first consists simply of a funnel-shaped opening of the longitudinal duct into the body cavity. It soon, however, by a process of gradual shifting, comes to open further and further back into that duct until it acquires an independent opening into the cloaca. The female genital glands, which are, as is usual in Vertebrates, specialised patches of the lining of the coelom, and of the unseg- mented * portion of it called the splanchnocoel which persists as the general body-cavity, dehisce their ova into the body cavity, whence they escape by the Miillerian ducts — except in Marsipobranchii and Teleostei. These exceptions however are doubtful. In Marsipobranchs the genital pores by which they escape may be Miillerian ducts, though it must be confessed that there is not much to be said for so regarding them. In Teleostei the ovaries are generally saccular and continued directly into their ducts, but in some families they discharge into the body cavity and the eggs are taken up by two funnel-shaped structures which join each other and open behind the anus. It is quite possible, though not definitely proved, that these funnels are short Miillerian ducts, and that the ducts in the more usual con- dition, in which they are continuous with the walls of the ovary, are also Miillerian ducts, which have spread round the ovary or fused with the edges of a peritoneal recess into which the ovary has sunk. The male gonads also are specialised patches of the coelo- mic lining, but the Marsipobranchii alone retain the primitive condition of testis dehiscing into the general body cavity, escape being made by genital pores of unknown homology. In all others the testis is continuous with its duct. In Teleos- teans this continuity is very like the continuity found in the female between the ovary and its duct, and the homology of the male duct in these animals is not understood. It may be a per- * The contention which has been put forward in some quarters that the gonads of Elasmobranchs arise from the segmented part of the coelom cannot be seriously maintained. GENERATIVE ORGANS. 89 sistent Miillerian duct which has fused with the testis or it may be something else. In other Pisces, with the apparent exception of the Ganoid Polypterus, the testis has come to consist of tubules which are connected by means of a network of tubes, called the testicular network, with some of the kidney tubules. The con- nection may take place along the greater part of the length^of the kidney, as in Lepidosteus and Acipenser, or it may be confined to- the anterior region (mesonephros), as in Elasmobranchs, or finally, as in Dipnoi, it may only occur through the hind end of the kidney. The connection is usually through the malpighian bodies of the renal tubules, but in Amia the tubes from the testis join the renal tubes beyond the malpighian bodies. In Poly- pterus alone is there no connection with the kidney, the testis duct passing directly back from the testis to join the longitudinal duct near th? cloaca. This condition in Polypterus is not understood any more than is the condition of the male Teleostean, though theories have been put forward to account for it. It may be that in these forms the Miillerian duct has acquired a connection with the male gonad and persisted ; or it may be that the con- nection is .really effected through a part of the kidney which has lost all kidney structure, as has happened in some male Amphibia and in the higher Vertebrata. A study of development can only settle the question, and that has not yet successfully been made. To return to the longitudinal duct. This, as explained above, is called at first the pronephric duct, except in Elasmobranchs, in which it is called the segmental duct, there being no functional pronephros. Later, when the kidney is formed and the prone- phros has atrophied, it becomes the kidney duct. In Elasmo- branchs, in which the kidney differentiates into meso- and meta- nephros distinguished, not by any break in continuity, but by size and by the course of the so-called collecting tubes of the nephridia (see below), it is called the mesonephric duct, because it appears to be related more especially to the mesonephric portion of the kidney. Inasmuch as in the male Elasmobranch this meso- nephric duct is chiefly concerned with carrying off the spermatozoa which pass, as has been described above, through a part of the mesonephros, it is also called the vas deferens. In the higher classes of Vertebrata the mesonephric duct is called the Wolfnan duct in the embryo, and persists in the male adult as the vas 90 CLASS PISCES. deferens, but disappears or is reduced to a slight vestige in the female. In Elasmobranchs the longitudinal duct is at first called the segmental duct on the view that the Miillerian duct is segmented off it. As we have seen, this is not a good description of what happens, and the name is not a happy one. After the Miillerian duct has become distinct from it, it becomes the duct of the per- sisting kidney, and eventually, owing to the shifting back of the point of opening of the metanephric tubules, the mesonephric duct. The nephridia typically open directly by the so-called collect- ing tubules into the part of the longitudinal duct opposite to them, but with the differentiation of the metanephros the collect- ing tubes of the posterior nephridia shift their point of opening into the longitudinal duct backwards, so that they all come to open close together into the longitudinal duct — now called mesonephric duct — close to the cloaca. They are usually re- ferred to as ureters. The development of the nephridia of the part of the kidney behind the pronephros. as direct transformations of a portion of the coelom occurs only in Elasmobranchs. In other fishes the development of these tubules is delayed until the myotomes and adjacent tissues have become functional, and have lost their primitive relations. The consequence is that the development is modified and the nephridia (except of the pronephros) are de- veloped from small nodules of growing tissue, which make their appearance during larval life in the proper positions. Abdominal pores, as distinct from generative pores, are pre- sent in most Elasmobranchs, some Teleostei and in Ganoids, but they are strangely variable in their occurrence. They never act as generative outlets, and their function would appear to be for the outlet of excretory substances of the body-cavity itself. As Bles has pointed out, they are rarely present in forms in which the nephrostomes of the kidneys are persistent. Generative Organs. — Excepting in certain forms, such as Serranus and Chrysophrys, which are hermaphrodite, fishes are of separate sexes ; the two sexes sometimes present external differences. The male and female reproductive organs often resemble one another so closely in form and position that it is necessary to investigate their contents in order to dis- HABITS. 91 tinguish the sex, especially as external sexual differences are frequently absent. Copulatory organs are only found in male Elasmobranchs, in the form of long grooved cartilaginous appendages (claspers) of the pelvic fins. Most fishes are oviparous ; only a few Teleosteans, as Ditrema, Zoarces, the Cyprinodonta, etc., and a great number of the sharks, bear living offspring, which for the most part undergo their em- bryonic development in a dilated part of the oviduct, which serves as an uterus. Reproduction usually takes place only once in the year, most frequently in spring, more rarely in the summer, and exceptionally, as in many of the Salmonidae, in winter. Many fishes, especially the males, undergo changes of colour and develop growths of skin at the spawning time. The two sexes often assemble in great shoals and seek out shallow places near the banks of rivers or near the sea coast (Herrings) for spawn- ing. Some make more extended migrations and pass in large shoals over great distances along the sea-coast (Tunny-fish). Others leave the sea and pass up the mouths of rivers, and over- coming great obstacles (Salmon leaps) make their way up into the smaller streams, in which they deposit their spawn in sheltered places where the food is plentiful (anadromous, as the Salmon, Sturgeon, etc.). The Eels, on the other hand, migrate from the rivers into the sea, and in the following spring the young Eels enter the fresh waters by millions and pass up stream (kata- dromous). The spawn is as a rule fertilized in the water, and thus artificial fertilization and pisciculture are rendered possible. In the viviparous fish, and in the Rays, Chimaera, and Dogfishes, which lay large eggs enclosed in a horny shell, a true copulation, and an internal fertilization of the egg takes place. It is worthy of note that in a few exceptional cases the male undertakes the charge of the brood (Hippocampus, Coitus, Gasterosteus). The embryonic development of fishes is principally distin- guished from that of most higher Vertebrates by the fact that neither amnion nor allantois are developed. Both the small eggs of the Teleosteans, which are provided with a micropyle, and the large eggs of the Elasmobranchs, which are surrounded by a hard horny case, contain a large quantity of food yolk, and undergo a partial segmentation. The eggs of Cyclostomes, Ganoids and Dipnoi, however, undergo a total segmentation. As a rule 92 CLASS PISCES. the young fishes leave the egg- membranes tolerably early, with more or less distinct remains of the yolk-sac, which still projects externally, like a hernia. Although the body form of the just- hatched fish differs essentially from that of the adult animal, no sudden metamorphosis takes place save in a few exceptional cases. Most fishes live in the sea, and the number of their species and genera increases as we approach the equator. But they are not all exclusively confined to fresh or salt water. Many, as the Plagiostomes, live almost entirely in the sea ; others, as the Cyprinidae and Esocidae, are confined to fresh water, but there are also fish which periodically change their habitat, especially at spawning time. Some fish live in subterranean waters, and are blind, like the inhabitants of caves (Amblyopsis spelaeus). Few fish are able to live any length of time out of water ; as a rule the wider the gill-slits the quicker does the fish die on dry land. Fishes with narrow gill-slits (Eels) possess an uncommon tenacity of life out of water. According to Hancock, a species of Doras migrates in great shoals over the surface of the ground from one piece of water to another, Except the Dipnoi, certain East Indian freshwater fish, the upper pharyngeal bones of which are hollowed out into the form of a labyrinth (Fig. 38) and form a multicellular reservoir for water, are capable of living the longest time out of water (Anabas scandens). There are even fishes which can float through the air (Exocoetus, Dactylopterus) . Marine fishes may be distinguished into shore fishes, pelagic fishes, and deep-sea fishes, which, as in the case of marine and fresh-water fishes, graduate into one another. Shore fishes live near the surface, and do not descend to any great depth ; they are comparatively restricted in range. Pelagic fishes inhabit the surface waters of the ocean, where they usually spawn, though some visit the shores for this purpose ; they are usually strong swimmers and wide ranging, but a few (e.g. Hippocampus, Antennarius, etc.) are poor swimmers, and infest floating sea- weed, or drift on the surface. Some pelagic fishes come to the surface at night only, descending in the daytime to a consider- able depth (Brama, Sternoptychidae, Scopelus, Astronesthes). The largest fishes belong to the pelagic fauna, e.g. Rhinodon, Selache, Carcharodon, Myliobatidae, Thynnus, Xiphiidae, Ortha- goriscus. The features of deep-sea fishes are referred to below. DEEP-SEA FISHES. 93 Owing to the uniformity of the conditions of life in the abyss in different parts of the world, they are probably for the most part wide ranging. The greatest depth at which fishes are known to exist is 2,900 fathoms. Many littoral fish descend periodically within the limits of the deep-sea fauna, but these are not conspicuously modified. Fishes which habitually live at a depth of 80-120 fathoms, have a black lining to the pharynx and large eyes. Fishes which belong to the real deep-sea fauna all present very similar characters, those from 300 fathoms being as much modified as fishes from 2,000 fathoms. The principal changes in external conditions to which deep-sea fishes are subjected as compared with surface forms are (1) absence of light, (2) still- ness of the water, (3) constant low temperature, and (4) increase of pressure. With regard to the latter, it may be stated that pressure increases by one ton on the square inch for every 1,000 fathoms of depth. The principal bodily characters are as follows : The eyes are largely developed and luminous organs, or, to speak more correctly, organs the function of which is probably to supply luminosity are present. When the supposed luminous organs are not present the lateral line canals are much dilated, sometimes into wide cavities, and full of mucus. The eyes are, however, in some cases reduced or absent. The osseous and muscular systems are feebly developed ; the bones being light and provided with little calcareous matter, and the muscles thin. When the fish are brought to the surface the bones are found to be but loosely bound together, and the body easily falls to pieces. This is probably due to the expansion of gases within the body. The air-bladder presents no special modifica- tions, and appears to be always without a pneumatic duct, even in Physostomous forms. It is generally ruptured in fishes brought up from the deep sea, and in fishes from 80 fathoms it is much distended, and the eyes protrude and the stomach is everted. Deep-sea fishes are sometimes found floating on the surface in a dead or dying condition, and often with the stomach distended with recently swallowed prey. It is conjectured that such fishes have accidentally ascended too far above their normal depth, possibly during the struggle of swallowing their prey which may be as large as themselves, and then owing to the expansion of gases consequent on the diminished pressure have been 94 CLASS PISCES. carried to the surface. Sharks, Rays and flat-fish (with one exception in each case) cease below 500 fathoms. Twelve hundred fathoms is the limit for Holocephali. The eggs of some deep-sea fishes ascend to and develop at the surface, but in other cases the development undoubtedly takes place in the abyss. Fishes are of great importance to our knowledge of the develop- ment of animal life on the earth, owing to the frequent appearance of their fossil remains in all geological periods. In the palaeozoic formations very singular fish-forms, as the Cephalaspidae (Cephalaspis, Coccosteus, Pterichthys), constitute the oldest representatives of the Vertebrata. From the palaeozoic forma- tions to the chalk we find almost exclusively cartilaginous fishes and Ganoids, amongst which the forms with persistent notochord and cartilaginous skull predominate. Ganoids, with a fully developed bony skeleton, round scales and an externally homo- cereal caudal fin, appear for the first time in the Jurassic rocks, where we also find the first Teleosteans. From the chalk onwards, in the more recent formations, the Teleosteans increase in number and variety of forms the nearer we approach to the fauna of the present time. The class Pisces is divided into the five sub-classes, Marsipobranchii, Elasmobranchii, Ganoidei, Dipnoi, and Teleostei. CHAPTER V. SUB-CLASS MARSIPOBRANCHII (CYOLOSTOMATA).* Vermiform fishes with smooth scaleless skin, cartilaginous skeleton and persistent notochord ; with suctorial mouth, single nasal organ, and straight intestine ; without jaws, paired appen- dages, generative ducts, sympathetic system, and conus arteriosus. The unpaired fins are without actinotrichia (dermal fin-rays). The Marsipobranchii are vermiform in appearance, varying in length from two feet or more (Bdellostoma) to a few inches (Petromyzon fluviatilis). The skin is smooth and without scales, and the skeleton is cartilaginous and notochordal. They are without paired fins, but possess an unpaired caudal fin (Myxi- nidae, Fig. 62), to which may be added a dorsal unpaired fin in the posterior region (Petromyzontidae, Fig. 48). In the * J. Miiller, Vergleichende Anatomie der Myxinoiden, Berlin, 1835-45. A. Giinther, Catalogue of the fishes in the British Museum, London, 1870. C. Kupffer, Die Entwickelung der Petromyzon planeri, Arch. f. mic. Anat., 35, 1890. P. Furbringer, Unters z. vergl. Anat. d. Muskulatnr. d. Kopfskelets d. Cyclostomen, Jena. Zeitsch, 9, 1S75. W. K. Parker, Skeleton of Petromyzon and Myxine, Phil. Trans., 1883, p. 373. P. Langerhans, Unters. ub. Petromyzon planeri, Ber. d. naturf. Gesellsch. zu Freiburg, 1873. A Schneider, Beitrdge, z. vergl. Anat. etc. d. Wirbe.l- thiere, Berlin, 1879. T. H. Huxley " On the Cranio-facial apparatus of Petromyzon." Journal of Anat. and Physiology, 10, 1876, pp. 412-28. F. Ahlborn, Das Gehirn v. Petromyzonten. Z. f. w. Z., 39, 1883, pp. 191-295, and Hirnnerven v. Petromyzon Z. f. w. Z., 40, 1884, pp. 286-308. G. C. Price, Ontogenie d. Myxinoiden Bdellostoma stouti, Sitz. ber. Math.-phys. Idasse d. k. bayer. Akadl d. Wiss, 26, 1896, Munich. Id. Development of excretory organs of Bdellostoma stouti, Zool. Jahrbnch. Anat., 10, 1897, p. 207. W. F. Pv. Weldon, The head-kidney of Bdellos- toma, Q. J. M. S., 24, 1884, p. 171-182. J. W. Spengel, Die Excretions- organe von Myxine, Anat. Anz., 13, 1897, p. 49-60. F. C. Studnicka, Sur les organes parietaux de Petromyzon vlaneri, Vestnik Ceske Spol. Nauk Prag, p. 1-50, 1893. J. D. Ogilby, "A Monograph of the Australian Marsipobranchii," Proc. Lin. Soc. N. S. W., 21, 1896, p. 388-426. 96 SUB-CLASS MARSIPOBRAXCHII (CYCLOSTOMATA). Petromyzontidae and in the tail of myxinoids the fins are sup- ported by cartilaginous somactids. They possess a suctorial mouth, which is without jaws, but is provided with horny teeth. By means of it, with the assistance of a suctorial tongue-like structure they attach themselves to and suck their prey. Myxine indeed bores its way into the body cavity of other fishes, and is truly parasitic. The nasal aperture is single, and leads into an unpaired nasal sac. In this and in other features of their anatomy, which will be described later, they are unique amongst Vertebrates. Nevertheless, we shall not follow the example of some zoologists who have established the Marsipo- branchii as a separate class of the Vertebrate, distinct from the class Pisces. We hold them, in spite of the remarkable and unique features of their organization to be true Pisces, not only FIG. 48. — a. Petromyzon fluviatilis (after Heckel and Kner). b, c, d, stages in the transforms' tion of Ammocoetes branchialis into Petromyzon planeri (after v. Siebold) ; b head of an eye- less larva, side view ; c the same, ventral view ; d later stage with small eyes, side view. by their aquatic habit of life, but by the characters of their respiratory and vascular organs. They possess a simple tubular heart, which distributes the blood by means of a ventral aorta to the walls of the gill pouches ; and these open to the exterior on the sides of the body in the ordinary piscine manner. In the structure of their mouth parts they present some resemblance to the larvae of anurous Amphibia, but the resemblance is too vague to permit of any definite approximation to that group in classification. The Marsipobranchii fall into two main groups which present marked points of difference from one another. These are the Petromyzontidae or lampreys, and the Myxinidae or hag-fishes. In the Petromyzontidae the nasal sac does not communicate with the mouth, the eyes are normally developed, and possess eye- muscles with their corresponding nerves, the pericardium does LATERAL LINE. MUSCLES. 97 not communicate in the adult with the abdominal part of the body cavity, and the pronephros does not persist. In the Myxinidae the nasal sac does communicate with the alimentary canal by an aperture which perforates the roof of the mouth, the eyes are much reduced and without the muscles and the cor- responding cranial nerves, the pericardium communicates with the general body cavity by a wide opening on the right side, and the pronephros is persistent in the adult. Moreover, the Myxinidae possess a contractile dilatation on the portal vein (portal heart) which is not present in the lampreys. The skin is slimy, and has the usual vertebrate structure. It possesses unicellular glands which secrete the mucus. In the Myxinidae there is in addition on each side of the body and embedded in the subcutaneous tissue, a row of segmentally arranged slime-glands, which open on the surface and pour out a mucus containing an immense, number of threads. These threads arise in special cells of the gland and unwind them- selves when the mucus is discharged. They were discovered by Retzius and described and figured by Mfiller. Nothing of the nature of lateral line sense-organs has been observed in FIG. 49.— Thread- n/r • "j i> •j.'ui n alutinosa with unwinding Myxinidae, but in the lampreys small thread (after Muiier). sensory eminences, partially sunk in pits, are found on the head and in two double rows on the body.* The great lateral muscles are divided up by septa, which have a zig-zag course, into myomeres of the usual piscine type. The myomeres extend on to the head to just behind the eyes. In the Myxinidae there is in addition a ventral sheet of obliquely directed muscle-fibres which is unsegmented. There is a complicated system of muscular bands connected with the mouth, tongue, and pharynx. In Petromyzon the portion of the lateral muscles dorsal to the gill-sacs is continued to just behind the eye and contains a greater number of seg- ments than the corresponding ventral portion. The ventral part reaches to just in front of the first gill opening. In Ammocoetes there is one myomere anterior to the first gill aperture ; this in the adult divides up into nine or ten myomeres (Schneider). * Langerhans, op. cit. z — II H 98 SUB-CLASS MARSIPOBRANCHII (CYCLOSTOMATA). The skeleton consists of cartilaginous, notochordal, and mem- branous tissue ; there is no bone. There are two kinds of cartilage at least, the hard with considerable intercellular matrix, and the soft with but little. The vertebral column consists of a persistent notochord with a tough sheath, which is formed of two layers, an inner somewhat fibrillated chordal sheath (or membrana elastica internet,) and an outer thin elastic coat (membrana elastica externa) ; both are devoid of nuclei. The notochord so constituted is surrounded by a nucleated membranous sheath (the so-called skeletogenous tissue, or membrana reuniens), which extends dorsalwards on each side so as to enclose the spinal cord. Small cartilaginous FIG. 50. — Cartilaginous skeleton of the anterior part of the body of Petromyzon fluviattiis ; side view (after A. Schneider). 1 Foramen for sensory, 2 for motor root of spinal nerve ; 3 eleventh dorsal arcualium ; 4 first '.dorsal arcualium, pierced by the first anterior root, which passes into the ligament between the fourth and fifth myomere ; 5 foramen for vagus ; 6 auditory capsule ; 7 foramen for trigeminal ; 8 foramen for optic ; 9 nasal capsule ; 10 posterior dorsal cartilage ; 11 anterior dorsal cartilage ; 12 annular lip carti- lage ; 13 anterior lateral cartilage : 14 styliform cartilage of 12 ; 15 unpaired lingual cartilage; 16 posterior lateral cartilage; 17 subocular arch; 18 styloid process; 19 cornual cartilage ; 20 branchial basket-work ; 21 the seventh gill aperture (the first and sixth gill apertures are omitted) ; 22 pericardial cartilage. pieces are developed on the membrana reuniens on each side : these, the dorsal arcualia (dorsalia), are roughly segmentally arranged. In Petromyzon there are in the branchial and trunk region two pairs of arcualia in each segment, while in the tail they are fused to form on each side a continuous ridge, with which the cartilaginous fin -rays (somactids) here present are continuous. Ventralia, fused to a continuous ridge, are also present in the caudal region, and are continuous ventrally with ventral somactids. In myxinoids cartilaginous elements are restricted to dorsalia and somactids in the caudal region, and to some somactids in the trunk. Anteriorly the notochord extends into the base of the skull, ending just behind the pituitary body. The skull consists of cartilage and membrane. The roof is entirely membranous in myxinoids, but in Petromyzon there is a narrow bar of cartilage passing across the posterior part of the SKULL. 99 otherwise membranous roof. In the floor there is a basi-cranial fontanelle (Fig. 52, 9} in the anterior (trabecular) region just in front of the anterior end of the notochord ; this transmits the pituitary pouch or posterior nasal canal, which, passing back from the nasal capsule, ends blindly in the Petromyzontidae (Fig. 51, 1], but opens into the mouth in the Myxinidae. This canal lies between the basilar plate and the roof of the mouth. The olfactory capsule is single and attached to the anterior part of the cranium by fibrous tissue (Fig. 50, 9}- The auditory 12 11 10 9 14 13 15 FIG. 51. — Longitudinal vertical section through the anterior end of Petromyzon fluviatUis (after Huxley). 1 Blind end of posterior nasal canal; 2 hinder margin of hard palate (inter- trabecula) ; 3 cartilaginous roof of skull ; 4 brain ; 7 nasal capsule ; 8 posterior wall of nasal capsule ; 9 the anterior portion of the subocular arcade ; 10 postero-lateral carti- lage : 11 postero-dorsal cartilage (ethmovomerine plate) ; 12 tongue ; 13 anterolateral cartilage ; 14 anterodorsal cartilage ; 15 annular lip cartilage ; 16 median ventral carti- lage ; 17 lingual cartilage ; 18 ventral division of fifth nerve coming through the sub- ocular foramen; 19 cornual cartilage; 20 posterior part of subocular arch ; 21 styloid process (hyoid) ; opening of tube ; 27 oesophagus ; 28 notochord ; 29 spinal cord. yoid) ; 22 tentaculate branchial valve ; 23 pharyngeal velum ; 24 internal first branchial pouch ; 25 ditto of second branchial pouch ; 26 suboesophageal capsules (Fig. 50, 6) are attached laterally in the posterior region on each side of the basilar plate. The subocular arch (Fig. 50, 17) is a ventro-lateral continuation of the basilar plate and trabe- cular region on each side ; it contains a fenestra and is supposed to correspond to the subocular or palato-quadrate arcade of the Amphibian skull. At the point where the posterior part of this arch joins the basilar plate, there is given off ventral wards the styloid process (Fig. 50, I'S, and Fig. 51, 21), the end of which in Petromyzon extends horizontally as the cornual cartilage (Fig. 50, 19). The styloid process and cornual cartilage have been 100 SUB-CLASS MARSIPOBRANCHII (CYCLOSTOMATA). compared to the hyoid arch of the higher forms. In the Petromyzontidae the following additional cartilages are present. Attached to the anterior end of the base of the skull is a large median plate of cartilage — the posterior dorsal cartilage (Fig. 50, 10, and Fig. 51, 11}. Immediately in front of this, and overlapped by it, is the anterior dorsal cartilage (Fig. 50, 11, and Fig. 51, 14)- Just below the latter there is an annular cartilage (Fig. 50, 12, and Fig. 51, 15), which lies within the lips, and from which there projects back on each side the so-called styli- form cartilages (Fig. 50, 14). The anterior lateral cartilages are paired rods (Figs. 50 and 51, 13} in relation with the anterior dorsal cartilage, and the posterior lateral (Fig. 50, 16, and Fig. 51, 10} are similarly in relation with the posterior dorsal. Finally, in the tongue in the floor of the mouth there is a median i1-T,fT11Ql oartilarm /TTifr ^O 7o nnH T^icr lingual Cartilage (JJ Ig. OU, lO, and KI 1?\ QTirl TTanfral +n +Vii« tVif* U), ana Ventral no]]^ rnorlian Tr^ntral r>artilao-^ fTTio- called median ventral cartilage ^ig. 61, ^; not shown in Fig. 50). The lingual cartilage is also present in Myxinoids, in which it is very largely 8 FIG. 52.— Ventral view of skull of Petromyzon marinus (after W, K. Parker). 1 posterior dorsal cartilage (ethmoyomerine plate), formed by union of cornua trabecuiae, passing behind into (formed by union of parachorial cartilages), showing contained notochord ; 9 basi-cranial fonta- Snfuesed°rpo8sPtaedor ^pSn *? It is probable that the posterior dorsal trabecuiae ; 10 cartilage formed cartilage is derived from the fused anterior &£& t-SSVa^aVg end of the trabeoula* and it has been of J. Miiller) ; it lies between suggested that the posterior lateral cartil- SS*SM?torfirSSS& ages are the homologues of Meckel's cartil intertrabecula of myxinoids. age. The anterior dorsal> and the anterior laterals, and the annular cartilages are generally regarded as labials, while the lingual cartilage has been compared to the basi-hyal ; it is connected with the styloid process (supposed hyoid) in Myxinoids (Fig. 53). The foramen for the optic nerve is in the side wall of the skull above the subocular arch (Fig. 50, 8), that for the fifth nerve just in front of the auditory capsule (Fig. 50, 7), while the seventh SKULL OF MYXIXE. 101 fs. CD 01 01 "S ° s p,-"^ a °^» -a«ilaiil 3 p. 4illlii ilftr-in - S|i«tl£!* « g-s s ..a &• f-*r'STO . _ i ,— --sits O ~ 102 SUB-CLASS MARSIPOBRANCHII (CYCLOSTOMATA). nerve passes out through the auditory capsules. The vagus nerve passes out behind the skull, between it and the first dorsal piece of the vertebral column (Fig. 50, 5}. In its general features the chondro -cranium of Maraipdbranchii resembles that of other fishes, but is more largely supplemented by membranous structures, nearly the whole of the roof, the entire occipital region and the basicranial fontanelle being membranous. From what is known of the development of the lamprey's skull, it would appear that the basilar plate is formed by two parachordals between which the notochord lies (Fig. 52), that these become continuous above and below the notochord (below only in Myxine), and that to their outer sides the auditory capsules become attached. In front of the parachordals and continuous with them are the trabeculae cranii which always remain separate posteriorly, leaving the basicranial fontanelle, but unite in front to form in Petromyzon (Fig. 52) the hard palate, the ethmoid and the posterior dorsal plate (ethmovomerine plate), and in Myxine (Fig. 53, 13) a small median piece in front of the large basicranial fontanelle. In front of the latter there is in Myxinoids a pair of cartilaginous horns (Fig. 53, 17) which may be regarded as the homologues of the posterior lateral cartilages of the lamprey. Moreover in Myxinoids there are two median cartilages, called intertrabecular car- tilages not present as separate structures in Petromyzon ; one of these — the posterior intertrabecular — is a spoon-shaped cartilage lying in the basi- cranial fontanelle and underlying the naso-palatine canal ; the other, the anterior intertrabecular, extends in front of this and lies beneath the nasal canal. In all Marsipobranchs there is a ventro -lateral process of the hinder part of the trabecular region (anterior lateral process) which meets and fuses with a corresponding process of the anterior end of the basilar plate or auditory cartilage (posterior lateral process, said to be comparable to the pedicle of the Amphibian suspensorium, Fig. 53, 7) ; thus forming the so-called subocular arcade. The subocular arcade closely resembles the same structure in the Amphibian skull, but against the comparison of the two is the fact that the ventral division of the fifth nerve passes dorsal to the arcade in Amphibians, whereas in Marsipo- branchii it passes through the fenestra (Fig. 51, 18)- In Myxinoids the hinder part of the subocular arcade contains two fenestrae (Fig. 53,6, 21), which are not present in lampreys. The supposed hyoid arch arises in all Marsipobranchs from the hind end of the subocular arcade (Fig. 50, 18 ; Fig. 53, 23)- In lampreys it ends in the expansion of the cornual cartil- age (Fig. 50, 19) and is not connected with the lingual (15) which is the supposed median element of the hyoid arch. In Myxinoids it joins the great lingual cartilage (Fig. 53, 18) which consists of several parts. In the same group the hinder part of the subocular arcade also gives off close to the point of origin of the hyoid a bar of cartilage which passes ventral wards just within the hypoblastic epithelium to join the subocular arch lower down (Fig. 53, 24). This structure is not represented in lampreys and is supposed to be the first branchial arch. The velum in Myxine is supported by some pieces of cartilage which are in connection with the upper end of this supposed branchial arch (Fig. 53, 2, 3). In Myxinoids the brain lies entirely above the cartilaginous skull, which is a mere floor, the side walls and roof being entirely formed of membrane. Moreover in Myxinoids the angle of the subocular arcade is posterior (Fig. BRANCHIAL SKELETON. 103 53), i.e. the palatine process of the trabecular region is directed backwards and not merely outwards as in lampreys (Fig. 50). Further the labial cartilages of the lamprey, viz. the anterior dorsal, the anterior laterals and the annular are not present in Myxinoids. On the other hand the oral barbules of the Myxinoids contain a cartilaginous axis and the lingual cartil- age is enormously developed (Fig. 53), and connected with the styloid process (supposed hyoid arch). In the Petromyzontidae there is a branchial basketwork of cartilage placed superficially, near the skin and supporting the outer parts of the branchial passages. It consists (Fig. 50) of eight irregularly curved bars of cartilage placed between the successive gill sacs.* They are connected dorsally by a longi- tudinal band of cartilage, which lies along the notochord sheath and is continuous with the hind end of the skull, and by three other longitudinal bands, of which two are lateral, one being above and one below the branchial apertures, and one is ventral and partly fused with its fellow. The branchial basket is con- tinuous posteriorly with a cartilaginous cup which supports the wall of the pericardium (Fig. 50, 22). The first branchial aper- ture is behind the first bar, and the seventh or posterior in front of the last bar. The branchial basketwork is supposed to be developed in the somato- pleure and not to be homologous with the branchial arches of other fishes which lie in the gutwall. The only representative of these structures in Myxinoids (in addition to the supposed traces of mandibular and hyoid arches) is the structure described above as the first branchial arch (Fig. 53, 24). The alimentary canal. The mouth, or buccal funnel (Fig. 54), is suctorial and armed with horny epidermic teeth ; in the lamprey it is surrounded by a lip carrying short papillae, in Myxinidae by eight barbules (Fig. 61). On the ventral side of the mouth is the tongue, which, moving backwards and forwards like a piston, enables the animal to attach itself by its mouth as by a sucker. The tongue carries teeth (two rows in the myxinoids on the supralingual cartilages), which enable it to inflict con- siderable wounds upon its prey. The buccal funnel leads into a tube, which is supposed to be stomodeal, and may be called the buccal cavity. This is continued behind into the oesophagus, * There appears to be some variation in the details of the arrangement of the branchial basketwork in the different species (cf. W. K. Parker's account with Schneider's figure). 104 SUB-CLASS MARSIPOBRANCHII (CYCLOSTOMATA). which passes back through the pericardial cavity to become continuous by a valvular aperture with the straight intestine, which opens at the anus. The anterior end of the intestine is slightly dilated, and receives the opening of the bile-duct, and in Petromyzon the intestine is provided with a longitudinal fold or valve, which takes a slightly spiral course. In Myxinoids the naso-palatine canal (pituitary pouch) opens into the hinder part of the buccal cavity. This may be taken to mark the junction of mouth and pharynx. The opening is guarded by an epiglottis-like valve and directly behind there is a velar membrane of a peculiar form hanging from the dorsal wall and supported by the so-called pharyngo- branchial cartilages (Fig. 53, 2, 3). The part of the alimentary canal immediately succeeding the mouth and receiving the internal gill aper- tures should be called pharynx, though it is commonly termed oesophagus. In Petromyzon a velar fold marks the junction of thesuboeso- phageal tube or bronchus with the hinder part of the mouth. There is a gall bladder, but in the adult Petromyzon the bile-ducts and gall- bladder atrophy and the liver cells be- come filled with fat (Schneider). A pancreas and spleen appear to be absent, and the mesentery is very imperfectly developed. The anus is placed in the median ventral line in a shallow pit im- mediately in front of the urogenital FIG. 54. — Head of Petromyzon rmcminn- marinus, seen from below u-Hei showing the lip and horny Thp rpsniratnrv nrtranc r>nn«i«t r»f teeth of the buccal funnel. ICbpirdlOry orgdllb C number of branchial sacs, in which the branchial lamellae are contained. In Petromyzon there are seven pairs of these sacs, and each of them opens independ- ently to the exterior in the anterior region of the body, by a short external branchial passage (Fig. 48), but internally they open into a longitudinal suboesophageal tube (bronchus), which opens into the buccal cavity in front (Fig. 51), and ends blindly behind. The anterior opening of the suboesophageal tube is guarded by a membranous velar fold. In Myxine there are six pairs of branchial sacs (seven have been observed in rare cases). In Bdellostoma the number is more variable ; seven pairs appear to be the usual number, but RESPIRATION. 105 there may be six pairs, or seven on one side and six on the other (heterotrema), or there may be more than seven (up to fourteen pairs). In both genera the sacs are connected directly with the oesophagus (pharynx) by internal branchial tubes, and with the exterior by external branchial tubes, but whereas in Bdellostoma each of the external branchial passages opens separ- ately by a small aperture (Fig. 61), in Myxine all the external branchial tubes of the same side are directed backwards, and unite together before opening to the exterior by a common opening at the hind end of the branchial region. In both genera there is on the left side a tube, the oesophageo-cutaneous duct, which leads directly from the oesophagus behind the last gill sac, to open with the external branchial tube of the last left gill sac in Bdellostoma, and with the left gill aperture of Myxine. The oesophageo-cutaneous duct is much wider than the external branchial tubes. In all Marsipobranchs respiration can be effected while the animal is adhering to foreign objects by the suctorial mouth. In Petrom'yzon respiration is effected by taking in water through the external branchial openings into the branchial sacs, and then expelling it again by the same way. In Myxinoids the water is said to enter through the nasal tube, which communi- cates with the mouth through the posterior nasal passage, and passes out by the branchial sacs ; but the nasal passage is a narrow one, and perhaps hardly sufficient to supply all the respiratory water. Moreover, it would not be available when the animal's head is buried in the tissues of its prey. It would appear more probable that some at least of the inspiratory water enters through the oesphageo- cutaneous duct. In Petromyzon the branchial basket plays an important part in respiration. In expiration it is compressed by the trans- verse muscles ; in inspiration it recovers by its own elasticity. In Myxinidae it is possible that the huge lingual apparatus may play some part in bringing about inspiratory dilatation of the pharynx, but it has been asserted that the inflow of water in these animals is effected through the nasal canal by ciliary action. In the respiration of Ammocoetes water is taken in through the mouth and passed out by the clefts. The expulsion of the water is effected by muscular constriction of the branchial region ; the 106 SUB-CLASS MARSIPOBRANCHII (CYCLOSTOMATA). inspiration by the dilatation of the branchial region caused by the elasticity of the branchial basketwork. The double valve (velar fold) at the junction of the mouth and branchial portion of the alimentary canal prevents the regurgltation of water in expiration. In Petromyzon it has been ob- served that in every inspiratory and expiratory movement of the muscles of the branchial region water is at the same time taken in and expelled from the nasal opening. The central nervous system is constructed on the usual vertebrate type. The brain of Petromyzon (Fig. 55) is unique amongst Vertebrate, for the fact that the median part of the roof of the sylvian aqueduct (iter) is epithelial and covered by a choroid plexus. Moreover, the cerebellum is very small, and the thalamencephalon of some length. The third ventricle di- vides in front into a right and left Canal> each °f which, passing FIG. 55,-Dorsal view of the brain of Petromyzon fluviatUis (after Ahlborn). 1 olfactory nerves : 2 left ganglion habenulae (the two pineal bodies have been removed) ; 3 continuation off n Wflrrj« into War< anrl of 2 along the roof of the third ven- forwards inrn flip nlfaofn™- tricle; 4 swollen termination of 3 Jliaciory which is connected with the ventral of the two pineal bodies ; 5 fourth ventricle; 6 edge of thin roof of fourth ventricle; 7 cerebellum; 8 optic lobes ; 9 edge of thin roof of iter ; 10 posterior commissure ; 11 right ganglion habenulae ; 12 cerebral hemisphere ; 13 olfactory lobe. pineal body, or consists of two the one dorsal to the other, over the anterior part of the thala- mencephalon. The larger dorsal vesicle is the pineal body proper, the ventral smaller one being called the parietal organ. The dorsal vesicle lies close beneath the skull-wall, and is the so-called pineal eye. The ventral part of its walls contains a white, sometimes a black, pigment, and presents a structure which recalls that of the retina. It is connected by a solid stalk (pineal stalk) containing nerve- BRAIN. CRANIAL NERVES. 107 fibres with the right of two thickenings on the superior com- missure, called the right ganglion habenulae. The ventral and smaller vesicle also presents in its lower wall, though not so markedly, features which recall retinal structure. It is hollow and is connected with a small solid body on which it lies, and which is the anterior part of the small left ganglion habenulae (Fig. 55, 3). These two bodies, though in contact, are apparently not connected. The pineal stalk is connected to the roof of the brain just in front of the posterior commissure, while the parietal organ is attached just anterior to the superior com- missure. The pineal body lies close to the roof of the skull, and the skin above it is not pigmented (Fig. 56). The hypophysis or pituitary body is dorso- ventrally flattened and follicular in structure. It lies beneath the infundibulum. In Myxinoids the corpora bigemina are present in the normal form and there is no thin place in the roof of the iter. The thalamencephalon is not nearly so prominent, the optic lobes being approximated to the cerebral lobes. The anterior part of the brain is solid, the central canal not extending beyond the mid- brain. A small space in the region of the thalamen- cephalon may be made out but it is quite isolated from the iter. A pineal body, has, so far, not been found in the myxinoid brain. -No,. FIG. 56.— Dorsal view of the head of Pe- tromyzon planeri (after Ahlborn). Na external nasal aper- ture ; Ep position of the epiphysis (the non-pigmented char- acter of the skin at this spot is not clearly shown). The cranial nerves are fairly normal in their arrangement. There are ten pairs, but in Myxinoids the third, fourth and sixth appear to be entirely absent, in correspond- ence with the absence of eye- muscles. The optic nerves, which are very small in Myxinoids, appear not to cross, the chiasma being hidden in the substance of the brain. In Petromyzon the sixth nerve arises close to the fifth, and supplies the inferior rectus as well as the external rectus. There is said to be no lateral line branch of the vagus in Myxinoids, but in the lam- preys this nerve is well developed and reaches to the end of the tail. It is, however, in an unusual position, being placed far from the skin close to the neural sheath of the vertebral column, and it appears to be connected with the posterior roots of the spinal nerves dorsal to which it passes. 108 SUB-CLASS MARISPOBRANCHII (CYCLOSTOMATA). The fifth nerve divides into two branches, the ophthalmic * which is purely sensory and passes to the skin on the head, and the ventral branch, which is both motor anci sensory. The ventral branch divides into an ex- ternal and internal branch, which do not correspond to the superior and inferior maxillary branches of other fishes, for they both supply muscles which in Selachians are supplied by the inferior maxillary branch, t The seventh nerve is in Petromy- zon a purely sensory nerve ; in Myxinidae it is mainly sensory. The vagus arises by eight roots, of which the four an- terior group themselves to- gether as a nerve which by its distribution to the tissue between the first and second branchial pouches must ob- viously be compared to the glossopharyngeal (Fig. 57, Br^). The other four vagus roots unite in a ganglion which is joined by a com- missural branch from the seventh nerve (vii.-x.) and gives off dorsally the lateralis (lateral line branch ) and ven- trally the visceral branch. The latter supplies the bran- chial region and the whole length of the intestine, in the wall of which it lies. In the branchial region in Petromy- zon, and possibly in the intes- tinal region as well, the vis- ceral branch of the vagus is connected with the posterior roots of the spinal nerves. As suggested by Miiller, it very possibly represents the sympathetic which is other- wise absent in Marsipo- branchs. FIG. 57.— Diagrammatic dorsal view of the pos- terior cranial nerves of Petromyzon (after Ahlborn). a Sensory root of hypoglpssal from the glossopharyngeal ; Au auditory capsule; Brl glossopharyngeal; 6. sp. 1, O.sp 2 etc., ganglia on postrior roots of spinal nerves ; Lot, lateral branch of vagus ; N.X. vagus roots (including those of the glosso- pharyngeal) ; N. sp. spinal nerves ; Ophth. ophthalmic branch of trigeminal ; Pn vagus ; rd dorsal, r.v ventral ramus of ventral root of spinal nerve ; vm motor, vs sensory root of trigeminal ; w.d. dorsal, w.v. ventral roots of spinal nerves ; VII facial nerve ; VII-X branch connecting facial and vagus ; VIII auditory nerve ; XII hypoglossal ; XII rd. dorsal branch of XII to muscles of head. Some ventral roots arise behind the vagus (so-called ventral vagus roots) and unite to form a nerve * A motor branch is mentioned by some authors. f Furbringer, foe. cit. SPINAL CORD. NOSE. EYES. 109 (xii.) which supplies the tongue muscles. It is called the hypoglossal, and gives off near its origin a branch which supplies the anterior part of the dorsal muscles. The spinal cord is flattened ; it has neither dorsal nor ventral fissure, though traces of a dorsal fissure may be represented by a fine tract of connective tissue which passes from the dorsal side of the central canal to the dorsal surface of the cord. The spinal nerves have dorsal and ventral roots which unite in Myxinoids, but not in Petromyzon. The posterior roots possess a ganglion, which lies just outside the skeletogenous wall of the neural canal. All the nerves are without the medul- lary sheath and the motor fibres are larger than the sensory. In Petromyzon the dorsal root of the first spinal nerve enters the septum between the fourth and fifth myomeres,the ventral root divides and supplies the fourth and fifth myomeres. The motor root of the second spinal nerve supplies the fifth and sixth myomeres, while the third and subse- quent spinal nerves each supply one myomere only. Sense Organs. The external nostril and the nasal sac are single and median, though the olfactory nerves are double. From the ventral side of the nasal sac a tube — the nasopalatine canal — is continued backwards between the brain and the skull floor, passes through the basicranial fontanelle and ends blindly on the ventral side of the anterior end of the notochord in Petromyzontidae, whereas in Myxinidae the same tube opens posteriorly into the mouth. The palatal opening of this canal has nothing to do with the posterior nares of higher vertebrates. It appears to be derived from the pituitary invagination of the embryo, which arises in Marsipobranohs, not as in most Vertebrates from the mouth, but as an ectodermal invagination in front of the mouth, which secondarily becomes connected with the nasal pit. It is for this reason sometimes called the pituitary pouch. The eyes are normal in Petromyzon, and possess the usual eye muscles. In the Myxinoids they are extremely reduced and without eye-muscles. In Bdellostoma they are embedded in a spherical fatty mass, and placed beneath the skin which is without pigment immediately over them. In Myxine they lie deeper within the muscles close to the skull wall. 110 SUB-CLASS MARSIPOBRANCHII (CYCLOSTOMATA). The eye in Myxine is without any trace of lens and appears to have but little if any pigment.* It consists of little more than the much degener- ated optic cup. According to Miiller the optic nerve in Myxine passes dorsal to the ophthalmic branch of the fifth. The auditory organ differs from that of other Vertebrata. In Petromyzontidae it consists of a vestibule and two semi-circular canals, in Myxinidae of a single semicircular canal only. The body cavity is divided into two parts, the pericardial cavity and the general body cavity. These two cavities com- municate by a wide opening on the right hand side in Myxinidae ; in Petromyzontidae they communicate in the larva, but not in the adult. The general body cavity opens into the urinogenital sinus by two genital pores (one in Myxine}, through which the generative products escape. In the anterior part of the body there is a system of spacious venous sinuses. These are specially developed round the branchial sacs, and in Myxinoids round the ventral aorta, thus constituting a kind of haemocoelic body cavity for these parts. The vascular system is arranged essentially in the manner usually found in fishes. The heart consists of sinus venosus, large auricle, and ventricle. There are only two valves at the junction of the ventricle and ventral aorta, and the base of the ventral aorta (bulbus) is much swollen owing to the large amount of elastic tissue in its walls. This bulbus is without any muscular tissue. There is no conus arteriosus. The branches of the ventral aorta are distributed one to each branchial sac, and the efferent branchial vessels are collected into two aortic roots which are continued forward as carotids, and unite behind to form the dorsal aorta. The dorsal aorta is also continued forward in the middle line as a carotid. In Myxinidae the genital vein and some veins from the anterior part of the body wall fall into the portal. Moreover, in Myxinidae the portal vein is dilated into a contractile sinus, which, contracting about as rapidly as the heart, drives the blood through the liver. It is a remarkable fact that no muscular fibres can be found in the walls of this portal sinus (J. Miiller). The portal vein extends for some distance in the intestinal wall and has been called the subintestinal vein. * C. Kohl, " Rudimentare Wirbelthieraugen," Bibl. Zool. (Leuckart & Chun.), 4, heft 13, 1892. KIDNEYS. Ill The kidneys are not divided into meso- and meta-nephros. They are placed along the dorsal side of the body cavity for the middle of its length, being absent at the anterior and posterior ends. The longitudinal ducts (pronephric ducts) extend along their whole length and join together posteriorly in Petromyzon to open by a single opening into the urogenital sinus. The urogenital sinus which also receives the two genital pores opens at the end of a papilla just be- hind the anus into a depression of the skin into which the anus also opens. In Ammocoetes the kidney ducts open into the hind end of the intestine (cloaca). The separation of the uro- genital sinus and the formation of the genital pores takes place just before the metamorphosis. In Myxinidae * there is a shallow cloaca which receives the opening of the intestine in front, the wide genital pores (single in Myxine) on the dorsal side of the anus, and the two urinary ducts, opening close together at the end of a papilla behind. In Myxine the kidneys have an ex- ceedingly simple structure. The longi- tudinal ducts give off at segmental intervals short lateral tubes which open into large malpighian bodies. The glomeruli are multipolar, i.e., are con- nected at several places with the wall of the capsule. In Petromyzon the structure is very similar but more complicated, though the malpighian capsules of successive tubules are separate, the vascular tissue of the glomeruli is continuous. The tubules of the kidney do not open into the body cavity by nephrostomata. B repr A highly magnified ; a pro- nephric (longitudinal) duct ; b kidney tubule ; c glome- rulus ; d afferent, e efferent artery. * R. H. Burne, " Porus genitatis in Myxinidae" Journ. Linn. Soc., 26, 1898, p. 487. 112 SUB-CLASS MARSIPOBRANCHII (CYCLOSTOMATA). Abdominal pores as distinct from the genital pores appear to be absent. In Myxinidae the pronephros persists as a lobulated organ in the pericardial cavity of the adult, and was called by Miiller the suprarenal body. In Petromyzon it is quite absent in the adult through present in the larva. The pronephros of Petromyzon is developed in the embryo and has three or four body-cavity openings and a continuous glomerulus. It is in rela- tion with the pericardium and atrophies during the Ammocoetes stage. The kidney is developed in the young larva posteriorly, its anterior end being a short distance behind the pronephros. The tubules arise as excavations in the mesoblastic tissues. This larval kidney of the Ammocoetes atrophies after the metamorphosis and is replaced by an exactly similar structure placed further back. The pronephros of the adult Myxinoid consists of a large number of nephrostomes which end blindly internally in a mass of lymphoid tissue, or possibly in some cases, perhaps in young specimens, open internally into an isolated anterior portion of the pronephric (segmental) duct. There is also at the hind end of this organ a glomerulus of some size projecting into an open recess of the pericardial cavity. The whole excretory system of Bdellostoma * appears to develop in the same way in which the pronephros does in other types, that is to say the longitudinal duct (pronephric) and the excretory tubules (segmental tubules) arise in continuity with each other from the body-cavity epithe- lium. The parts of the body-cavity into which the segmental tubes open soon become separate from the rest and form a series of small vesicles each communicating with a segmental tube. These vesicles become the malpi- ghian bodies and the segmental tubes become the renal tubules. It is not known how the pronephric part of the system acquires the peculiar structure which it exhibits in the adult. As stated above the persistent kidneys of Marsipobranchs show no differentiation into meso- and meta-nephros and the testes are not con- nected with them. The generative organs are unpaired. They are attached to the dorsal wall of the body cavity by a broad membrane, and the generative products, both male and female, are shed into the body cavity, whence they escape by the genital pores. Myxine is hermaphrodite, and the reproductive gland produces sper- matozoa before ova f (protandrous). Petromyzon is dioecious, but ova have been observed in the testis. * Price, op. cit. f J. T. Cunningham, "Reproductive elements in Myxine glutinosa" (Q. J. M. S., 27, 1887), and " Spermatogenesis in M yxine " (Q. J. M. S., 33, 1891). The hermaphroditism of Myxine is denied by Dean (Journ. Coll. Sci., Tokyo, 19, 1904). DEVELOPMENT The development has been fully worked out in Petromyzon planeri.* The egg is small, about 1 mm. in diameter. It is enclosed in a membrane formed of an inner perforated and an outer structureless layer. Outside there is a mucous envelope which causes the egg to adhere to foreign objects. The male adheres to the female during oviposition and the ova are depos- ited in a hole previously made and subsequently covered up, the fish moving stones for this purpose by means of their suctorial mouths. The eggs are laid in April and May. Segmentation is unequal but complete ; the gastrula is formed by a combination of invagination and epibole, and the blasto- pore persists as the anus. The central nervous system is formed by a solid keel-like ectodermal ingrowth, in which the central canal arises by excava- tion, and there is a solid cord of cells connecting the hind end of it with the dorsal hypoblast. The pituitary body is formed from an invagination of the ectoderm and subsequently becomes connected with the nasal pitr with which its proximal part remains in communication throughout life,, as the naso -palatine canal or pituitary pouch. The young are hatched as larvae which soon become Ammocoetes. These live for three or four years, developing and increasing in size until they become as large as or larger than the adult. They then undergo a sudden (in three or four days) metamorphosis (from August to January) and become adult (Fig. 48). The adult possesses ripe generative organs and spawns in April-May. After spawning the lamprey (in the case of P. ftuvi- atilis at least) dies. Ammocoetes was formerly regarded as a distinct genus of animal, and separate species were distinguished. The fact that it becomes metamorphosed into Petromyzon was discovered two hundred years ago by L. Baldner, a fisherman of Strasburg, and rediscovered by Aug. Miiller.t In Ammocoetes the buccal cavity is without the annular lipr but possesses a semicircular upper lip (Fig. 59), and a small separate lower lip. There are no teeth, but several fringed barbels surround the mouth. The eyes are imperfect and hidden beneath the skin. The gill openings are placed in a groove on each side. The median fin extends all along the back, as a continuous structure. The branchial pouches open into the pharynx directly, and there is no suboesophageal tube or bronchus distinct from the pharynx. In Ammocoetes there is a gall bladder and bile duct, which opens into the intestine. In the lamprey both these structures are absorbed, and the intestine itself undergoes partial atrophy. The eye in Ammo- * F. M. Balfour, A Treatise on Comparative Embryology, vol. ii. 1881 (with literature to date). A. E. Shipley, Q. J. M. S., 27, 1887. C. Kupffer, Arch. mic. Anat., 35, 1890. P. Bujor, " Metamorphose de 1' Ammocoetes branchialis," Rev. Biol. Nord. France, iv. 1891, p. 41. t Miiller's Arch., 1856, p. 325. 114 SUB-CLASS MARSIPOBRANCHII (CYCLOSTOMATA). coetes is only partly developed and lies be- neath the skin. In the lamprey it becomes fully developed and travels to the surface. The pericardium of Ammocoetes opens into the general body cavity, but the two become completely separate in the adult. The an- terior part of the kidney which had been developed in the Ammocoetes disappears, and a fresh posterior part is formed. The prone - phros had already begun to disappear dur- ing the Ammocoetes stage, and the portion of the cloaca into which the urinary ducts open becomes separated off as a urogenital sinus shortly before the metamorphosis. The skeleton undergoes very considerable change at the metamorphosis. The Ammo- coetes is without the cartilaginous dorsalia or neural arches in the trunk region. These appear at the metamorphosis as do cartilages of the mouth, and the side walls and roof of the skull. The spinal cord, which is nearly round in section in Ammocoetes, becomes flattened at the metamorphosis. The head muscles of Ammocoetes are entirely destroyed and reformed at the metamorphosis. The thyroid body arises in the embryo as a groove in the branchial region of the g pharynx. The opening soon becomes narrowed to a pore placed between the second and third permanent* branch- ial pouches (Fig. 60). In Am- mocoetes the tube so formed becomes divided and assumes FlQ/59. — A. Ammo- coetes of Petromy- zon planeri, 2 ins. long, side view (after W. K. Park- ,er) brl first, br7 seventh branchial B Ventral view of * There is said to be a trace of an eighth pouch, in front of the first permanent one, in the embryo and young larva. It is supposed to repre- sent the hyomandibular cleft of other fishes, but it never acquires gill folds or an external opening. aperture ; na na- sal aperture; £eye. the head of the same larva. HABITS. AFFINITIES. 115 a very complicated glandular form. After the metamorphosis it is less conspicuous and appears to be without the opening into the throat. In Myxinidae the eggs are very much larger (19 mm. x 7 mm, in Myxine, 31 mm. x 9-5 mm. in Bdellostoma] and contain a considerable quantity of yolk. They are enclosed in a horny case with hooked processes proceeding in tufts from each end. The egg case appears to be the vitelline membrane. The Ammocoete lives buried in mud and sand and likes dark places. It lives on small aquatic organisms (Infusoria; Daphnia, Rotifers, etc.). The marine lamprey ascends rivers at spawning time, sometimes carried by the salmon or shad (Alausa vulgaris), to spawn. They eat worms and small aquatic animals. The Myxinidae live exclusively on other fishes. They are able by their t,Y N formidable dental armature and powerful lingual muscles to make 0 KS^—*^^ K ° their prey, in which FIG. 60.— Diagrammatic longitudinal s-ction through the head of a larva of Petromyzon (after Balfour). Ab optic they are Sometimes vesicle; C heart.; Ch notochord ; H thyroid involution ; Ks branchial pouches ; N nervous system ; 0 mouth ; lOUnd em beaded. Ol olfactory pit; Ot auditory vesicle (represented .as _ _ . visible) ; Ve velum. The Marsipo- branchii are sometimes spoken of as a degenerate group. We do not think that there is any evidence of degenera- tion. The most important points in which they differ from other fishes relate to the skeleton, and to the nasal organ. But these are precisely the organs which show the greatest amount of variation within the group as at present constituted. This seems to point to the fact that they separated off from other fishes at a time when these two organs were in a highly indeter- minate condition, and had not attained to that fixity of structure which characterises on the whole the general arrangement of the skeleton, and nasal and pituitary sacs in other fishes. The condition of the eyes in Myxinidae might be held to be evidence of degeneration, but we should rather be inclined to regard it again as the survival from a time when the visual organ was more variable and had not obtained that fixity of character which it has at the present day. No J16 SUB-CLASS MARSIPOBRAXCHII (CYCLOSTOMATA). doubt many individuals were then produced with imperfect visual organs. Most of these naturally died out in competition with their more highly endowed brethren, but in some cases compensating advantages in other organs enabled them to hold their own in spite of their defective sight. To hold that a free-living animal, and a myxinoid must after all be regarded as such, can lose its eyes through disuse would seem to be an impossible position. The absence cannot be considered as other than a disadvantage to it. Fam. 1. Petromyzontidae (Hy- peroartia), lampreys, nine-eyes. With seven external gill apertures on each side which lead into a sub- pharyngeal tube opening anteriorly into the pharynx and ending blindly behind, and with a com- plete branchial basket-work. The suctorial mouth is surrounded by a circular lip and is provided writh horny, simple or multicuspid teeth, without barbules. The single nasal opening is in the middle of the up- per side of the head, and the nasal duct (pituitary sac) ends blindly behind. Eyes are present. With two dorsal fins, and a spiral valve in the intestine ; gall-bladder absent. The eggs are small, and there is a prolonged larval stage in which the young are known as Ammo- coetes. Fresh waters and coasts of the temperate regions of both hemispheres. Petromyzon Art. coasts and fresh-waters of the northern hemisphere ; P. marinus L., sea-lamprey ; P. ftuviatilis L., river-lamprey. Ichthyomyzon Gir- ard, west coast of North America ; Mordacia Gray, without gular pouch, coasts of Chile, south-east Australia and Tasmania, entering fresh-water to breed ; Geotria, Gray, with gular pouch, rivers of Chile, mal). south and south-east Australia and New Zealand ; Velasia Gray, with- out gular pouch, is an immature stage of Geotria. Fam. 2. Myxinidae (Hyperotreta). Nasal aperture single, at the anterior end of the body ; the nasal duct (pituitary sac) opens posteriorly into the pharynx and has cartilaginous rings ; mouth suctorial, without lips, with barbules, with one median palatal tooth and two rows of lingual teeth ; branchial openings far behind the head, lead directly into pharynx ; branch- ial basket-work not present ; a series of mucous sacB on each side of the body ; eyes hidden under the skin, and very imperfect, without lens or d.KC FIG. 61. — Ventral view of anterior end of Bdellostoma forsteri (after W. K. Parker) br p. external aper- tures of branchial sacs ; d. oe c opening of ductus oesopha- geus cutaneus. FIG. 62. — Myx- MYXIXIDAE. 117 muscles ; intestine without spiral valve ; gall bladder present ; egg large with horny case provided with threads for adhesion ; marine in temperate regions of both hemispheres. Myxine L., hag-fish, with six pairs of bran- chial sacs opening by one external opening on each side. Bdellostoma J. Mull. (Homea Fleming), with six or more (up to fourteen) branchial FIG. 63. — Palaeospondylus gunni ventral view of head and side view of vertebral column. tp trabeculo-palatine part of skull ; pa, parachordal part of skull ; d c, Ic, vc oral cirri ; a, b, c, n markings of more uncertain significance ; x post occipital plates (from S. Woodward). apertures on each side, each leading to a branchial sac ; the number of branchial sacs may be different on the two sides of the body ; southern hemisphere. Palaeospondylus Traquair, from the old Red Sandstone of Scotland, is supposed to be a fossil Marsipobranch with calcified cartila- ginous endoskeleton (Fig. 63). The notochordal sheath appears to have contained rings. CHAPTER VI. SUB-CLASS ELASMOBRANCHIL* Fishes with a cartilaginous endoskehton, placoid scales, and abdominal pelvic fins -provided with claspers in the male. There is a conus arteriosus, an optic chiasma and a spiral valve in the intestine. There is no air-bladder. The eggs are large, and, except in Laemargus, provided with a horny case. In the embryo the gills ^woject^from the gill clefts as filaments. The Elasmobranchii or cartilaginous fishes include the sharks and rays. With the exception of one or two sharks and a few rays they are entirely marine forms. They are remarkable for possessing more features which are embryonic in the higher purely terrestrial Vertebrata than any^ other group of fishes. Of such may be mentioned the oro-nasal groove, the opening between the membranous labyrinth and the exterior, the un- covered gill apertures, the open spiracle, the cartilaginous skeleton, the opening between the pericardium and the body cavity. Lastly they are the only fishes which possess eggs containing so much yolk that the whole development is embryonic. Only two species of shark are known to be exclusively inhabitants of fresh-water (Carcharias nicaraguensis and gan- geticus), but several ascend large rivers, e.g., the Tigris and Ganges, to a considerable distance. Most Selachians are pelagic or shore forms, and some descend to great depths (Scyllium has been taken at 700, Chlamydoselachus at 100-150, Centroscyllium at 245, Pnstiurus at 500, and Centrophorus at 345-500 * See Giinther, Day, Jordan and Evermann, Bridge, Boulenger, cited under Pisces ; J. Miiller and J. Henle, Systematische Beschreibung der Plagiostomen, Berlin, 1839. F. M. Balfour, Development of Elasmobranch Fishes, London, 1878. EXTERNAL FEATURES. 119 fathoms). Their flesh is not usually esteemed as food, but some of them are eaten by poor people. The body is elongated and spindle-shaped in the Squall^. the anterior part being somewhat broad and depressed dorso-ven- trally as compared with the narrower posterior region ; in the Raji it is strongly compressed dorso-ventrally. In some forms, mostly in the Raji, the snout is prolonged to a greater or less extent. This is most markedly the case in the saw-fish shark and in the saw-fish Pristis. In the hammer- heads the anterior part of the head is_elongatedjtransversely, the eyes being placed at the ends of the prolongations. The\ median fins are typically two dorsal, a _caudal, the ventral part of which is divided by a notch into two parts, and^an^anal/ placed between the caudal and the anus. The paired fins are, well-developed : the pelvic being smaller than the pectoral,^ FIG. 64. — Acantffl&s vulgaris. spl spiracle ; ks gill slits (from Claus). and abdominal in position. In the males the pelvic fins are provided, each with a copulatory appendage — the clasper (pterygopodmm, mixipterygium), which is grooved on its dorsal side, the groove leading into a cavity at the base of the appendage. In the Raji the pectoral fins are very large and their line of attachment to the body has a considerable antero-posterior extension. The muscular system is on the usual piscine type. The great lateral muscle is divided into a dorsal and ventral half, the myomeres of which alternate. There is the usual system of branchial and mandibular muscles. The gill-clefts are tubes usually five on each side (in Chlamy- doselachus and Hexanchus there are six, in Heptanchus seven), and their external openings which are placed laterally in Squall, ventrally in Raji, are not covered by an operculum (Fig. 64). Internally they open into the pharynx and their walls are pro- 120 SUB-CLASS ELASMOBRAXCHII. vided with a number of Jamellate folds of the mucous mem- brane, which are placed on their anterior and posterior walls (except in the last tube, which has no branchial lamellae on its posterior wall), and are attached through their whole length (Fig. 118), not projecting freely as do the pectinate gill processes of Teleostci. In addition there is usually an anterior tube leading outwards from the pharynx and opening exter- nally on the dorsal surface close behind the eye. This opening is called the spiracle and the tube itself must be regarded as belonging to the series of branchial tubes of which it is the first. It differs from these, however, in never possessing branchial lamellae, though it often, has traces of these as a few small folds of the lining of its anterior wall, which constitute the pseudo- branch or mandibular gill of these fishes. In the embryo long filaments — the so-called external gills — project from all these openings including the spiracle ; they are in reality externally projecting internal gills. |F IirRaji the spiracle is much larger than in Squali and it doubtless allows of the entrance of water into the pharynx when the animal is lying flat upon the ground or partly buried in sand. In Squali, in which it is very variable, being sometimes absent and nearly always small, its function is not clear. In some species in which it is very small it may be present or absent in •different individuals. It is sometimes present in embryos of forms in which it is absent in the adult (Carcharias), but whether this is always the case is not known. From the fact that it is smaller than the posterior branchial apertures even at its first appearance (which is subsequent to that of the others) it may be presumed that it is usually absent in such cases, but the matter wants looking into. When it is absent in the adult and present in the embryo, it is without projecting gill filaments in the embryo (Miiller); In Scyllium, Pristiurus, Mustelus, etc., the spiracle gives off a diverticulum to the auditory cartilage of the skull. The nasaLapertures and mouth are almost always placed on the ventral surface of the head (in Chlamydoselachus the mouth is anterior and the nasal apertures are dorsal), usually at a considerable distance from the front end. The nasal apertures are frequently connected with the mouth by a groove, the orc- nasaLETOove, and sometimes they are so close that their open- ings are confluent with the mouth. In other cases they are at some distance from the mouth and there is no oro-nasal groove. The anus (cloacal opening) is placed between the pelvic fins, and there are two abdominal pores, one on either side of the anal opening, which lead into the body- cavity. SENSE ORGAXS. SKIX. 121 »v Abdominal pores vary considerably.* In some species they are absent altogether, in others they are present in the adult, while in yet others they are present in some individuals and not in others. Their external opening is always on an ectodermal surface, either just outside the cloacal boundary, or into a cloacal pouch, which is a diverticulum of the proctodeal part of the cloaca. The eyes are usually provided with upper and lower cutaneous folds which represent eyelids, and in some forms there is a third inner eyelid or nictitating membrane which can be drawn over the eye. The otocysts retain their communication with the exterior by means of a canal, the aqueductv^ vestibuli, which opens on tJie dorsal surface throughout life (p. 77). The latejral line is a canal which extends in the skin from the very hind end of the body to the head, where it branches out to different parts in the usual piscine manner (p. 80). It opens to the exterior at intervals. In addition to the system of the lateral line, there are the openings of the so-called ampullary canals. These are placed in groups in the head (Lorenzini's ampul- lae, p. 81). Luminous or- gans | irregularly scattered over the body are found in many pelagic msmbsrs of the Spinacidae (e.g., Spinax, Laemargus, Isistius) in the form of minute cutaneous patches which probably secrete a luminous mucus. The skin is tough and rough owing to the presence of a vast number of placoid scales. These are rhombic bony plates embedded in the cutis and carrying a small spine, which * Bles, " Correlated Distribution of Abdominal Pores and Xephrostoni33," Journ. Anat. and Phys., 32, 1898, p. 484. t B. Burckhardt, "Luminous Organs of Selachians, Ann. and Mag. Nat. Hist. (7) 6, 1900, p. 558-568. FIG. 65.— Placoid scales of an adult Scyllium in surface view (after Klaatsch). The anterior end of the figure is uppermost. The spines are omitted from some of the scales. Ck the central canal (pulp cavity) of the spine as it perforates the basal plate Sb of the scale ; Sa spine of the scale. 122 SUB-CLASS ELASMOBRANCHII. projects freely on the surface in a backward direction and consists_of__dentine containing a^pulp_£aYijty and capped by enamel. The placoicLacales though numerous are not in con- tact, and fresh scales are continually being developed between them, to replace those worn off. The presence of these spines enables the skin of Plagiostomes to be used by polishers (shagreen). The spines are sometimes much enlarged, e.g. the peculiar spines on male skates, the caudal spine of the sting ray (Trygon), the large spines often present on the dorsal fins, etc. The fa^th in T,V>A mouth are special modifi- cations of placoid scales. The endoskeleton is entirely cartilaginous, but the cartilage is frequently more or less calcified. It is possible that t _^~t^ perichondriai ossifi- :BRANCHII. cula seminalis (Fig. 81, vs). The two vesiculse seminales join to form the urogenital sinus and just before their union each receives the opening of the seminal bladder (ss), which is a pouch lying on the ventral side of the vesicula seminalis, and the openings of the ureters, usually four or five in number in Scyllium canicula. The urinary sinus which in the male is common to the urinary and generative organs opens into the cloaca through its dorsal wall by a median papilla — the urino- genital papilla (ug.p.). In the female the hind end of the longitudinal duct is dilated (ub) and receives not far from its union with its fellow the ureters (ur) by one or 5 more openings. The oviducts open close to- gether through the dorsal wall of the cloaca in front of the urinary papilla. They extend for- ward to the front end of the body ca v ity into which they open close together on the ventral side of the anterior end of the liver (fl.tf). Not far from their front end their walls are much thickened owing to the presence of glandular tissue constituting the oviducal gland (od.g). Traces of the oviducts are often present in the male, particu- larly near the abdominal openings. The oviduct in the viviparous forms presents a uterine dilata- tion and the oviducal gland is much reduced. In many Selachii the nephrostomes of a certain number of the primary kidney tubules persist into the adult as ciliated openings. These are minute in Scyllium, but in some forms they attain a considerable size. The egg is large and heavily yolked. It receives a coat of FlQ. 80. — Testis and anterior part of niesonephros (Wolffian body) of an embryo of Squatina vulgaris (after Balfour) : to show the testicular network. There are five vasa efferentia connecting the longitudinal canal in the base of the testis with a longitudinal canal in the mesonephros. From the latter there pass off four ducts to as many malpighian bodies. 1 vasa efferentia ; 2 malpighian bodies ; 3 mesonephros ; 4 longitudinal or mesonephric duct ; 5 longitudinal canal of the testis ; 6 testis. PLEUROPTERYGII. 145 The Elasmobranchii are the most ancient of all known fishes. They make their appearance in the Upper Silurian. They are almost entirely active, carnivorous, predatory fishes and with very few exceptions exclusively marine. The following is the classification adopted in this work : — Order 1. PLEUROPTERYGII (extinct). 2. ACANTHODI 3. ICHTHYOTOMI 4. SELACHII (PLAGIOSTOMI) . Suborder l. Notidani. 2. Squall. 3. Eaji (Batoidei). 5. HOLOCEPHALI. FIG. 83.—Cladoselache. A pectoral, B pelvic fins x $ ; B basal somactids within the body- wall, D dermal fin membrane, R peripheral somactids. Left border preaxial (after Dean, from Woodward). Order 1. PLEUROPTERYGII.* With unconstricted notochord and heterocercal caudal fin. Paired fins with unsegmented parallel radials, reaching to the edge of the fin. Eyes with a circle of thin dermal plates. Male without claspers on the pelvic fins. The skull is unknown, but the jaws are suspended by a slender hyo- mandibular. The teeth have a principal cusp and several accessory lateral cusps. They resemble teeth which have long been known from the Carboniferous under the generic name Cladodus. There were certainly * B. Dean, Contributions to the Morphology of Cladoselache, Journ. Morph., 9, 1894. Jaekel, Ueber Cladodus, Sitzungsb. d. Gesellsch. naturf. Freunde, Berlin, 1892. R. Traquair, Geol. Magazine, 1888, p. 83. Z — II L 146 SUB-CLASS ELASMOBRANCHII. five gill-arches but there may have been more. Two dorsal fins have been seen (one only shown in Fig. 84), but no anal. The caudal fin is strongly heterocercal ; the neural arches are continued to the end of the tail, and carry stout somactids which extend to the edge of the fin. The paired fins are horizontal expansions of the integument. The peripheral somactids are parallel, unsegmented, and extend to the margin ; between their distal ends are slender cartilages which are pos- sibly displaced somactido. Thebasals are also parallel, and are contained FIG. 84. — Restoration of Cladoselache newberryi Dean (from Woodward, after Dean). ii in the body wall. The skin is covered by minute denticles, not enam- elled. Cladoselache Dean, Lower Carboniferous of Ohio ; Cladodus Ag. for some tune known only by teeth ; Devonian, Carboniferous, and Per- mian. Order 2. ACANTHODII.* With dermal calcareous plates on the skull and pectoral arch, and with a mosaic of quadratic dermal scales on the body. All the fins except the caudal, with a powerful dentine spine on their anterior margin. Without claspers. There are no cranial bones, nor membrane bones connecting the pectoral arch with the cranium. This group, which was formerly placed with the Ganoids, is now placed with the Elasmobranchs. The endoskeleton contains granular calcifica- tions, and the dermal plates placed on the head, body and pectoral girdle FIG. 85.— Acanthodes Wardi x i (after Woodward), gill-frills are hypothetical. The orbit is made too small and the seem to have consisted of vaso-dentine or of structureless lamellae without bone-cells. The most marked characteristic of the group is the large * Huxley, Geological Survey of the United Kingdom, 10, 1861. Fritsch, Fauna der Gaskohle in Bohmen, 2, 1889. Reis, Zur Kenntniss des Skelets der Acanthodinen, Geognost. Jahreshefte,'Munchen, 1890, 1894. Traquair, Geol. Mag., 1888, p. 511 ; 1889, p. 17. ICHTHYOTOMI. 147 spines on the front of the fins. These appear to have consisted of dentine, and are doubtless comparable to the spines found in similar positions in the fins of Elasmobranchs. It is probable that a number of isolated spines which have received special generic names (Onchus, Byssacanthus , Homacanthus, etc.) may have belonged to fishes of this group, and that a number of quadratic scales, e.g. Thelolepis, Coelolepis, etc., from the Upper Silurian, were part of the dermal armature of similar fishes. Pec- toral and pelvic fins are always found. FIG. 86.—Ciimatius scutiger, outline of fish with spines shaded. The pectoral fins pines pc are the two large spines next the head ; then follows a double row of smaller spines i, the last of which are the pelvic fin-spines plv. The large fin with spine a between the paired spines and the caudal fin is the anal ; dl, d* dorsal fins. The eye is surrounded by dermal plates ; the notochord must have been persistent ; the supports of the fins are not preserved ; the tail is hetero- cerca-1 and the caudal fin without any trace of upper lobe. Comparatively small fishes. Acanthodes Ag. (Fig. 85), Lower Devonian to Lower Permian ; Diplacanthus Ag., and Climatius Ag., Upper Silurian and Lower Devonian, without teeth, with four or five pairs of spines between the pectoral and pelvic fins (Fig. 86). Most of the sub-order do not show teeth, but there is a powerful dental armature in Ischnacanthus. Order 3. ICHTHYOTOMI.* The cartilaginous endoskeleton is permeated by granular calcifications ; notochord unconstricted, with slight calcifications in its sheath ; neural and haemal arches calcified, with long spinous processes ; tail diphycercal ; pectoral and pelvic fins with long segmented axis and biserial radii ; pelvic fins with claspers in the male. The teeth have two large lateral cusps, with one small median cusp. Lower Carboniferous to Lower Permian. Pleuracanthus Ag. (Xenacanthus Beyr.) (Fig. 87). Body elongated, to half a metre ; skin probably naked, with a long 3pine attached to the occipital region of the cranium ; five, possibly seven, branchial arches ; all the fins with dermotrichia ; pectoral girdle arch-like, united with its fellow, composed of two pieces ; pectoral fin (Fig. 76) with segmented axis ; rachiostichous and pleurorachic, or nearly so (p. 57) ; pelvic fir- similar, but with no postaxial somactids, with claspers ; pelvic girdle of two separate arches. The somactids of the long dorsal fin are segmented * Fritsch, loc. cit. Davis, On the Fossil Fish Remains of the Coal Measures in the Brit. Islands, I. Pleuracanthidae, Trans. Roy. Dublin Soc.t 4, 1892. 148 SUB-CLASS ELASMOBRANCHII. FIG. 87. — Pleuracanthus decheni, restored by A. Fritsch, x i ; Lower Permian, Bohemia (from Woodward). A1, A'2 anal fins; C dorsal part of caudal fin ; Z> dorsal fin. The specimen from which this figure was taken has been crushed in such a way that the paired fins appear to have their postaxial sides turned forward. into three pieces, and are twice as numer- ous as the neural arches, those of the dorsal part of the caudal fin are similar, but equal in number to the neural arches ; the anal fin is double, and its supports are partly fused, and branch peripherally. Didymodus Cope, Permian of Texas, skull shows symmetrical fissuring, to which the name of the sub-order is due. Order 4. SELACHII (PLAGIOSTOMI), Elasmobranchii with characters of the soft parts as defined for the sub- class, with hyostylic skull (except Notidanidae) and heterocercal tail. The notochordal sheath is always segmented, though sometimes imper- fectly ; the pectoral fins with three basal cartilages, and the pelvic fins of the male with claspers. The mouth is placed on the under surface of the head, except in Chlamydoselachus in which it is sub- terminal, and Rhinodon and Rhina in which it is terminal. The skin has detached placoid scales only. The body is either fusiform or flat- tened dorsoventrally, and there is usually a spiracle, but the pseu- dobranch is absent in the Scym- nidae, Lamnidae, Myliobatis, Trygon, etc. In Carcharias and Zygaena, in which the spiracles are absent, a pseudobranch is present buried in the flesh or placed on the front wall of a recess of the mouth. They are almost all marine, but a few ascend American and Asiatic rivers, and a few are confined to freshwater (some Trygons, two species of Carcharias). They have existed since Palaeozoic times XOTIDANI. 149 Sub-order 1. NOTIDANI. With six or seven branchial apertures and a small spiracle with one dorsal fin without spine. Vertebral col- umn imperfectly segmented. Cau- dal fin without a pit at its root ; without labial fold and nictitat- ing membrane. Fam. 1. Chlamy- d o selac hidae.* Body eel-like; mouth anterior ; nasal opening divided and on side of head ; lateral line as an open groove on the body, but closed (with openings left) on the head ; with six gill openings and six branchial arches ; opercular fold (first gill- FIG. 88. — A. Chlamydoselachus angui- neus Garman ; B. side view of the head of the same (after Garman). x f cover) free across the isthmus ; the palatoquadrate is not articulated with the post- orbital process of the skull and there is a large hyomandibular ; notochord unconstricted pos- teriorly ; teeth similar in both jaws, each with three slender, curved cusps separated by a pair of rudimentary denticles on a broad base; vivi- parous. Chlamydoselachus Garman, from the deep sea, 5-6 feet. Japanese seas, Atlantic and Arctic. * Garman, Bull. Mus. Zool. Harvard College, 12, 1885. 150 SUB-CLASS ELASMOBRANCHII. Fam. 2. Notidanidae. Mouth sub-inferior; nostrils on lower side, nearer snout than mouth ; dentition unequal in the two jaws ; in the upper jaw one or two pairs of awl-shaped teeth, the next six teeth broader and each provided with several cusps, one of which is the strongest ; lower jaw with six large, comb-like teeth on each side, beside the smaller pos- terior teeth ; viviparous ; sometimes reach a large size. The palato- quadrate articulates with the postorbital process of the skull and the hyo- mandibular is comparatively slender ; each segment of the vertebral column of the middle part of the trunk region carries two neural arches and corresponds to two pairs of spinal nerves (p. 126) ; the pseudobranch is very large and has several M'ell-developed laminae. Temperate and warm seas. Hexanchua Raf., with six pairs gill-apertures ; vertebrae without calcification ; H. griseus Gmelin, 8-26 feet, Mediterranean, W. coast Scotland ; Heptanchua Raf., with seven pairs gill-apertures, vertebrae asterospondylous. Sub-order 2. SQUALL Vertebral column well segmented, vertebrae amphicoelous with a double cone of calcified cartilage, outside which and springing from it there may be radiating calcareous lamellae (asterospondylous) or additional concentric calcified rings (cyclo- spondylous) ; with two dorsal fins, and with or without anal fins. With five gill apertures laterally placed, spiracle present or absent, never large. The palatoquadrate is not articulated directly to the skull except in Cestraciontidae. This sub-order includes the great body of living sharks, and has existed since the Silurian period. Some of the living genera have existed since early times ; e.g. Cestracion, Upper Jurassic ; Scyllium, and Scapanorhynchus (Mitsukurina) Cretaceous ; Pristiurus, Upper Jurassic. Most of them are active predatory creatures, and some attain a considerable size. The largest are however harmless creatures, which like the whalebone whales exist on small marine organisms which are detained on their prolonged gill-rakers (Selache, Rhinodon). Fam. 3. Cestraciontidae. Bull-head sharks; asterospondylous; the palato-quadrate articulates by an extensive surface with the preorbital region of the skull ; two dorsal fins with spines, the first dorsal opposite the space between pectorals and pelvics, the second in advance of the anal ; upper lip divided into seven lobes, the lower with fold ; spiracle small, below posterior part of eye ; without nictitating membrane ; den tition similar in both jaws, viz. small obtuse teeth in front, pointed and provided in young individuals with three to five cusps ; lateral teeth large, pad-like, twice as long as broad ; Pacific and East Indian Archi- pelago ; size small ; oviparous, egg-case spirally twisted. Cestracion Cuv. (Heterodontus Blainv.), C. phillipi Blainv., Port- Jackson shark. SQUALL 151 Extinct genera are Orodus Ag., Campodus de Kon., Sphenacanthus Ag. , Carboniferous Limestone ; Hybodus Ag., Trias to Cretaceous ; Palaec- spinax Eg., Lias ; Acrodus Ag., Trias to Cretaceous, etc. Fam. 4. Scylliidae. Dog-fishes ; asterospondylous, dorsal fins without spine ; first dorsal above or behind the pelvic ; an anal fin ; no membrana nictitans ; spiracle distinct ; mouth inferior ; teeth small ; nostrils near the mouth, sometimes confluent with it, sometimes with cirri. Scyllium Cuv. (Scylliorhinus Blainv.), upper edge of caudal fin smooth ; Sc. canicula Cuv. , small-spotted dog, single nasal flap, pelvic fins separated ; eggs laid in April, hatched in December ; Sc. catulus Cuv., large-spotted dog, nurse hound, nasal flap divided, pelvic fins almost conjoined ; Pris- tiurus Bon., snout much produced, small flat spines along upper edge of caudal fin, P. melanostomus Bon. ; Ginglymostoma M. and H., large sharks of the warm seas, nostrils confluent with mouth ; Stegostoma M. and H., tiger-shark to 15 feet, India ; Parascyllium Gill, Tasmania, 2| feet ; Chiloscyllium M. and H., nasal and buccal cavities confluent, Indian and Australian, 2^ feet ; Crossorhinus M. and H., Australia and Japan to 10 feet, mouth nearly anterior, nasal and buccal cavities confluent, Extinct genera : Palaeoscyllium Wagn., Upper Jurassic ; Mesiteia Kramb., Cretaceous. Fam. 5. Carchariidae. Asterospondylous ; first dorsal opposite space between pectoral and pelvic, without spine ; an anal ; with nictitat- ing membrane ; mouth crescentic, inferior ; spiracles present or absent. Carcharias M. and H. (sub-genera, Scoliodon M. and H., Physodon M. and H., Aprionodon Gill, Hypoprion M. and H., Prionodon M. and H.) com- prises the true sharks ; no spiracle ; teeth with a single cusp, snout pro- duced, pit at root of tail ; temperate and tropical ; C. glaucus Cuv., blue-shark, 25 ft. ; C. gangeticus of the Ganges and inland lakes of the Fiji Islands, and C. nicaraguensis G. and B., Lake Nicaragua, only fresh- water sharks known, 7 ft. ; Hemigaleus Bleek, East-Ind. Archipelago ; Loxo- don M. and H., Ind. Ocean ; Galeocerdo M. and H., arctic, temp, and trop. seas ; Thalassorhinus M. and H., Med. and Atl. ; Galeus Cuv., small spiracle, teeth with single cusp, snout elongated, no pit at root of tail, temp, and trop., viviparous ; G. canis Rond., tope, whithound, penny dog, miller's dog, 7 ft. ; Zygaena Cuv. (Sphyrna Raf.), temp, and trop., no spiracles, hammer-headed sharks, eyes at extremity of head lobes ; Z. malleus Shaw, hammer-shaped head, vivi- parous, balance fish ; Triaenodon M. and H., no spiracles, Red Sea, Ind. Ocean ; Leptocarcharias Smith, no spiracles, S. Africa ; Triads M. and H., Ind. and Pac. Oceans ; Mustelus Cuv., viviparous, teeth flat and paved, temp, and trop. seas, spiracles small, no pit at root of tail, bottom-fish ; M. laevis Risso, embryo attached to uterus by placenta ; M. vulgaris M. and H., no placenta, 6 ft., smooth hound, skate-toothed dog. Scylliogaleiis Blgr., Natal. Fam. 6. Lamnidae. Mackerel sharks. Large sharks with large teeth ; first dorsal opposite space between pectoral and pelvic, without spine ; asterospondylous ;. an anal fin ; no nictitating membrane ; mouth crescentric, inferior ; spiracles absent or minute, varying even in the same species ; pelagic, attain large size. Lamna Cuvier, Porbeagles, teeth lanceolate, large, with smooth edges, keel at side of tail, temperate and tropical seas ; L. cornubica Gmelin, Porbeagle (porpoise and beagle) or Beaumaris shark, viviparous, 10 ft. ; Carcharodon M. and H., C. rondeletii M. and H., great blue-shark, man-eater, spiracle minute or absent, Med. to 152 SUB-CLASS ELASMOBRANCHII. i Australia, 36 ft. ; Odontaspis L. Ag., temp, and trop. seas ; Alopecias (Alopias) M. and H., teeth triangular, flat, smooth edges, caudal fin long, no keel at side of tail, temp, and trop. seas ; A. vulpes Gmelin, thrasher-shark, fox ; Selache Cuv. (Cetorhinus Blainv.), teeth small, numerous, conical and smooth, keel at side of tail, whalebone-like gill- rakers consisting of dentine on the gill-arches, Arctic to Med. ; S. maxi- mus Gunner, basking-shark, sun-fish, one of the largest of living fishes, to 40 ft., large gill apertures, vertebrae appear to be tectospondylous, owing to presence of a number of concentric lamellae in the adult ; in- offensive, of great strength, has been known to tow a 70-ton boat against a fresh gale ; a large fish yields 1£ tons of oil ; Pseudotriacis Capello. Mitsukurina Jordan (Scapanorhynchus S. Wood), Japan. Extinct genera : Orthacodus S. Wood, Jurassic and Cretaceous ; Odontaspis Ag., Upper Cretaceous. Fam. 7. Rhinodontidae. Whale-sharks ; asterospondylous ; origin of first dorsai fin in front of pelvic ; the second small, opposite the anal, both without spines ; a pit at root of caudal ; side of tail with keel ; spiracle small ; membrana nic titans absent ; mouth and nostril near the front of snout ; teeth small ; gill openings wide, with gill-rakers, Cape of Good Hope, Seychelles, Japan. Rhinodon Smith ; a gigantic shark known to exceed 50 ft., said to attain 70 ft. Fam. 8. Spinacidae. Cyclospondylous ; spiracles present ; gill openings narrow ; without nictitating membrane ; a deep groove along either side of the mouth ; a spine on front side of each dorsal fin ; without anal fin. Centrina Cuv. (Oxynotus Raf.), body somewhat three-sided with a fold of skin at each angle, teeth in lower jaw triangular, erect and with finely serrated edges, no membrana met., Med. and adjacent At- lantic ; C. Salviani Risso, attains to 4-5 ft. ; Acanthias Risso (Squalus), teeth rather small, their points placed so obliquely that their inner margin which is smooth forms the cutting edge, no membrana nict., temp, seas of both hemispheres ; A. vulgaris Risso, picked dog-fish, spur-, spear-, or bone-dog, hoe, skittle-dog ; viviparous ; Centrophorus M. and H. , Eur. seas, Moluccas ; some species live at a great depth (400-500 fathoms) ; Scymnodon Boc and Cap. ; Spinax Cuv. (Etmopterus Raf.), Eur. Seas, W. Indies ; Centroscyllium M. and H., Greenland, has been taken 300-400 fathoms. Fam. 9. Scymnidae. Like Spinacidae, but no spines on dorsal fins ; Scymnus Cuv., Med. and Atlantic ; Laemargus M. and H. (Somniosus Lie Sueur), teeth in upper jaw small and conical, those in lower jaw in several rows, their points placed so obliquely that their inner margin, which is smooth, forms the cutting edge ; L. microcephalies Kroyer (borealis M. and H.), Greenland shark; notochordal sheath imperfectly segmented, un- calcified (calcified and segmented in L. rostratus) ; attains 25 ft., bites pieces out of whales ; with two pyloric caeca ; eggs large, soft, globular, without shell, dropped in the ooze on the sea bottom, said to be fertilised externally; they breed at considerable depths (100 fathoms); Euproto- micrus Gill, Ind. Ocean ; Echinorhimis Blainv., skin with irregularly placed round osseous tubercles, teeth large, oblique, with several small cusps on each side of the main one, Med. and Atl., E. spinosus Blainv., to 8 ft. ; Isistius Gill. Fam. 10. Rhinidae. Ray-like sharks. Tectospoiidylous ; spiracles large, gill openings wide, lateral, and partly concealed from above by pectoral fins ; body flat ; mouth anterior ; nostrils at front end of snout BATOIDEI. 153 with skinny valvular coverings ; pectoral fins large, laterally expanded,, but not attached to head ; dorsal fins spineless, in caudal region, no anal ; males with claspers ; temperate and tropical seas ; intermediate between the sharks and rays. Rhina Klein (Squatina Dum.), angel-fish, monk- fish, viviparous, to 8 ft. Fam. 11. Pristiophoridae probably here. Shark saw-fish. Rostral cartilage produced into long flat lamina, armed along each edge with teeth ; Pristiophoms M. and H., Japan, Australia. Sub-order 3. RAJI (BATOIDEI). Gill openings ventral, five in number : spiracle always present, without anal fin ; dorsal fins, if present, on the tail ; vertebrae tectospondylous. Skates and Rays. The body is much flattened dorso-ventrally and the pectoral fins are enormously expanded in an antero-posterior direction. The five branchial apertures are entirely on the ventral surface of the body. The spiracles are dorsally placed behind the eyes ; they are wide and can sometimes be closed by a valve. It i& probable that they are used for the intake of th3 respiratory water when the fish is lying on the ground. The caudal region is usually slender, and in some forms very much so. There is no anal fin, and the dorsal fins when present are placed on the tail. The Pristidae and Rhinobatidae, which have a well-developed caudal region and are intermediate in the form of their body between the sharks and rays, are powerful swimmers, but most of the Raji lead a more sedentary life on the bottom, rarely coming to the surface. They feed chiefly on Mollusca and Crustacea. A few deep-water forms are known, but they are rarely taken below 100 fathoms. Most are shore-forms. The Myliobatidae, which include the largest forms in the sub- order, are however met with in the open sea. Some species are confined to fresh water. They are for the most part oviparous. The flesh of many of the species is eaten. Some of the living families have existed since the Jurassic. Fam. 1. Pristidae. Saw-fishes. Snout much produced (rostral process of cranium) with lateral saw-like teeth ; body somewhat shark- like, the disc-like body gradually passing into the tail, which is com- paratively thick, with two dorsal fins and a caudal fin, without serrated caudal spine. Pristis Latham, tropical and sub-tropical, attain a con- siderable size, with a saw 6 ft. long and 1 ft. broad at base. Fam. 2. Rhinobatidae. Tail long and strong with two dorsal fins, a caudal and a longitudinal fold on each side, without serrated caudal spine ; rayed portion of pectoral fins not continued to snout ; no electric 154 SUB-CLASS ELASMOBRANCHII. •organ ; viviparous. Rhinobatus Bl. and Schn., guitar-fishes, tropical and sub-tropical, and fossil from the upper Jurassic ; Rhynchobatus M. and H., Ind. Ocean to China ; Trygonorhina M. and H., Australia ; Zapteryx Jor. and Gilb., Peru ; Platyrhinoidis Gar., California. Extinct genera : Asterodermus Ag., Belemnobatis Thiol., Upper Jurassic. Fam. 3. Torpedinidae. Trunk a broad, smooth disc ; tail with rayed dorsal (absent in Temera) and caudal fins and a longitudinal fold on each side, without serrated dorsal spine ; anterior nasal valves con- fluent into a quadrangular lobe. An electric organ between the pectoral fins and the head. Eocene to the present time. Torpedo Dum. (Narco- batis Blainv.), large specimens (width from 2 to 3 ft.), can disable a man. Med., Atl., Ind. Oceans ; T. nobiliana Bon., spiracles not fringed at their margins ; on flat sands or mud, 40-50 fathoms ; T. mar- morata Risso, spiracles fringed ; Narcine Henle, trop. and sub-trop. ; Hypnos Dum., Australian ; Discopyge Tschudi, Peru ; Astrape M. and H., Ind., S. Afr. ; Temera Gray, E. Ind. Fam. 4. Rajidae. Skates. Disc broad, rhombic, generally with asperities or spines ; tail with longitudinal fold on each side ; pectorals extend to snout ; no electric organ or serrated caudal spine ; oviparous ; sexual differences are frequently observable, in colour, form of teeth and arrangement of spines. Cretaceous to the present tune. Raja Art., tail distinct from disc, pectoral fins not extended to front end of snout, caudal fin rudimentary ; may attain width of 6-7 ft. ; in some species the teeth of the male are sharper than in the female, and in all species the males are armed with patches of claw-like retractile spines on the upper side of the pectoral fin ; seas of both hemispheres ; R. batis L., skate, oviposi- tion from May to September, to 6-7 ft. ; R. macrorhynchus Raf., flapper skate ; R. alba Lacep., white skate, to 8 ft. ; R. oxyrhynchus L., long- nosed skate ; R. fullonica L., Fuller's ray, shagreen skate ; R. clavata L., thornback ; R. maculata Montagu, homelyn ray, spotted ray ; R. micro- cellata Montagu ; R. radiata, starry ray ; R. circularis Couch, sandy ray. Psammobatis Giinth., South America ; Sympterygia M. and H. ; Platy- rhina M. and H. Extinct genus : Cyclobatis Eg., Cretaceous. Fam. 5. Trygonidae. Sting-Rays. Pectorals continued to and confluent at end of snout ; tail long and slender, without lateral folds ; vertical fins none or imperfect, often replaced by strong serrated spine. Tertiaries to present time. Urogymnus M. and H., Ind. Oc. ; Ellipe- surtis Schomburgk ; Trygon Adanson (Dasyatis Raf.), tail with long serrated spine, temp, and trop. ; T. pastinaca, sting ray, sandy ground near land, caudal spine causes severe wounds ; Taeniura M. and H., Indian seas, fresh-waters of trop. America ; Urolophus M. and H., Australian and Caribbean seas ; Pteroplatea M. and H., temp, and trop. seas. Fam. 6. Myliobatidae. Eagle-rays. Disc broad with large pectoral fins which are not present at the sides of the head, but reappear as at the extremity of the snout as a pair of detached fins ; teeth hexagonal, flat, tesselated ; tail long, thin, whip-like ; viviparous. Myliobatis Cuv., snout with a soft prolongation with fin rays ; temp, and trop. seas ; M. aquila L., mill-skate, whip-ray, eagle-ray ; Aetobatis M. and H., tropical seas ; Rhinoptera Kuhl, trop. and sub-trop. seas ; Dicero- batis Blainv. (Aodon Lac.) (Cephaloptera Dum.), head with a forwardly- pointing horn-like projection on either side ; attain great size ; temp, and trop. seas ; D. giornae. Ceratoptera M. and H. (Mania Bancroft), HOLOCEPHALI. 155 temp, and trop. seas, attain great size, 20 ft. wide. Last two genera often called sea-devils. Extinct genera : Ptychodus Ag., Cretaceous ; Promyliobatis Jaekel, Eocene. The following extinct Palaeozoic families are placed here : — COCHLIODONTIDAE, with several genera, from the Carboniferous Lime- stone ; PSAMMODONTIDAE, from the Carboniferous Limestone ; and PETALODONTIDAE, also from the Carboniferous. Order 5 — HOLOCEPHALI.* Without spiracle, with four clefts covered by an opercular fold which contains a cartilaginous plate. The skull is autostylic, and the notochordal sheath is unsegmented. There are two dorsal fins and an anal. The Holocephali differ from the Plagiostomi in the fact that there are only four gill clefts (though there are five branchial FIG. 8Q.—Chimaera momtrosa (Rdgne Animal). arches). Moreover the gill apertures are covered by an opercular fold, and the palatoquadrate bar is continuous with the skull in its whole extent. They have a cartilaginous skeleton and claspers on the pelvic fins of the male. The mouth is small, ventral, and bounded by lip -like folds supported by labial cartilages. The nostrils are confluent with the mouth. The urogenital part of the cloaca is separate from the rectum and opens behind the anus. The anterior dorsal fin has a strong spine, on its front border, which is attached to the fused neural spines of the anterior vertebrae. The tail is hetero- oercal and prolonged in Chimaera into a long filament. There * G. Good and T. H. Bean, Oceanic Ichthyology, Memoirs of the Museum of Comp. Anat. Harvard College, 22, 1896. A. A. W. Hubrecht, Kent- •niss des Kopfskelet d. Holocephalen, Niederldnd Arch. ZooL, 3, 1877. S. Garman, The Chimeroids, Bull. Mus. Harvard Coll., 43, 1904. 156 SUB-CLASS ELASMOBRANCHII. is on each side in the male a peculiar structure consisting of a plate carrying teeth and sunk in a pit just in front of the pelvic fin. The heapLof the male is provided with an erectile hook-like process projecting forwards over a groove and armed on its lower surface with small spines. The skin is usually naked except in the young, in which small placoid spines are found principally in a double row on the back. The lateraljine may be an open groove (Chimaera) or a closed canal (Callorhynchus). In Chimaera the lips of the groove are approximated on the head (Fig. 42), but remain apart at intervals giving the appear- ance of openings. Ampullary canals are present as in Plagio- stomes. The eyes are without lids. The notochordal sheath is thick, cartilaginous and unsegmented, It contains in Chimaera numerous calcified rings (four or five to each segment). The arch tissue is segmented, except in the front and in the whip-like tail, and does not meet round the notochord sheath, except again at the front end and in the tail. The neural arches consist of two pairs of pieces and a dorsal piece for each pair of spinal nerves. The neural spine of the anterior fused arch tissue is large and carries the anterior dorsal fin, the basals of which are fused into one piece. The skull is autostylic (p. 63), has rostral continuations and is without the prefrontal fontanelle. It has a well marked mem- branous interorbital septum, which is placed dorsal to the brain. It articulates with the vertebral column by two condyles. The auditory capsule is incomplete internally so that the space for the membranous labyrinth is open to the cranial cavity. There are three pairs of labial cartilages and the hyoid arch which carries branchial rays is attached by ligament to the skull. There are five branchial arches. The hyoid arch carries a demibranch (uniserial), the first three branchials each have a holobranch (biserial), and the fourth branchial carries a demi- branch. There is no gill-cleft between the fourth and fifth, branchial arches. The gill filaments are attached as in Plagio- s tomes and do not project. The paired fins and their girdles are formed on the Plagiostome type, except that the two halves of the pelvic girdle are united only by ligament. The tgetharelg/rge and few in number consisting of strong HOLOCEPHALI. 157 plates with cutting edges. (There are two pairs in the upper and one pair in the lower jaw. ) The intestine has a spiral valve and the anus isjnjfront of the urogenital aperture. There is a conus arteriosus with three rows of valves. The brain is characterised by the great length of the thala- mencephalon, and the cerebrum is very small. The olfactory peduncles are long and there is an optic chiasma. The pineal body is as hi Plagiostomes, and there is an extracranial part of the pituitary body lodged in a pit on the base of the skull. The cranial nerves* are arranged in the usual manner ; the roots of the fifth and seventh are more distinct than in most fishes. There is a pericardio-peritoneal canal. The urogenital organs appear to be similar to those of Plagio- stomes. In the female the shell gland is large. In the male there is a large vesicula seminalis, and the miillerian ducts are complete tubes opening at each end. Thex-are, oviparous andJiayjpi la,rge_eggs. The eggshell is covered with hair-like processes, and may attain a great size. They are probably laid in deep water, where the young are for the most part found. They have existed since the Jurassic period. In many points of structure these animals depart from other Elasmobranchs, and they have by some authorities been removed from that group and raised to the rank of a sub-class with affinities to the Dipnoi by the characters of their skull and teeth. There is much to be said for this view, for they present affinities to more than one piscine sub-class : to Elasmobranchs by their placoid scales, cartilaginous skeleton, absence of membrane bones, their gill-laminae, the open otocyst, the ampullary canals, the form of the brain, the structure of the urinogenital organs, their fin skeleton and claspers, and by their large eggs and develop- ment : to the Ganoids by the separation between the urinogenital sinus, and the alimentary canal, and by the incomplete internal wall of the auditory capsule : and to the Dipnoi by the last- named feature, by their autostylic skull, their peculiar teeth, and their vertebral column. They differ from Elasmobranchs and resemble Ganoids and Dipnoi in having an operculum, but they stand by themselves in having only four branchial clefts and a * F. J. Cole, Cranial Nerves of Chimaera. Trans. Eoy. Soc. Edinburgh, 38, 1896, p. 631. 158 SUB-CLASS ELASMOBRANCHII. demibranch on the fourth branchial arch. It is clear from this that the Elasmobranch characters strongly predominate, and in our opinion they may fairly be retained as an order of that subclass. Chimaera L., snout soft, prominent, without appendage ; tail produced into a fine filament ; deep water (200 to 1,200 fathoms) of coasts of Europe, N. Pacific, Cape ; C. monstrosa L., King of the Herrings ; attains 3 or 4 ft. Hydrolagus Gill, like Chimaera, but three dorsals and caudal, and tail without filament ; surface waters, N. Pacific. Callorhynchus Gronov. Snout with a cartilaginous prominence ending in a cutaneous flap ; S. Pacific, Cape ; egg-case 9 or 10 in. x 3 in. Hariotta Goode and Bean. Snout elongated, no frontal clasper, anal fin as cutaneous fold, deep water 700 to 1,200 fathoms. Extinct genera*: Ischyodus Eg., Upper Jurassic and Lower Cretaceous ; Ganodus Ag. ; Edaphodon Buckl., Cretaceous, Eocene, Oligocene ; Pachy- mylus Smith, Upper Jurassic, etc. The extinct families PTYCTODONTIDAE, known by teeth from the De- vonian ; SQUALOBAJIDAE, known by its skeleton, from the Lias ; MYRIA- CANTHIDAE, also known by skeletons from the Lias and Upper Jurassic, are placed here. * E. T. Newton, Chimaeroid Fishes of the British Cretaceous Rocks, Mem. Geol Soc. U. Kingdom, 1878. CHAPTER VII. SUB-CLASS GANOIDEI. * Fishes with a conus arteriosus, optic chiasma, free pectinate gills and an operculum, abdominal pelvic fins, a spiral valve in the intes- tine, an air-bladder, and without a processus falciformis and choroid gland. The oviducts and urinary ducts unite and open by a common urogenital aperture behind the anus. The skull is hyo- stylic and is without a supraoccipital bone. The segmentation of the ovum is complete, a pronephros is present in the larva and abdominal pores are always found. Very few of the characters mentioned in the definition are pecu- liar to Ganoids : they are almost all found in Teleostei or Elas- mobranchii. This fact coupled with the great variations of structure found in the group points to the conclusion that the sub-classes, Elasmobranchii, Teleostei, Ganoidei and Dipnoi are the survival of a once great and continuous group of animals, a large number of which have become extinct, leaving three groups, Elasmobranchii, Teleostei and Dipnoi, each fairly compact and showing but little variety of organization, and one, the Ganoideir loose and heterogeneous with large gaps between the individual members. Although it may fairly be held that by such forms as Lepidosteus and Amia the Ganoids more nearly approach the Teleostei than the Chondrostei do the Elasmobranchii ; we cannot * J. Miiller, Ueb. d. Bau u. d. Grenzen der Ganoiden, Abh. d. Berlin Akad. d. Wiss. 1844; id. Myxinoiden, op.cit. J. Hyrtl, Ueb. den Harn- werkzeuge bei den Ganoiden, Wien Denkschriften, 8, 1855. Liitken, Ueb. d. Begrenzung und Eintheilung d. Ganoiden, Palaeontographica, 22, 1872, Huxley, ' The systematic arrangement of the fishes of the Devonian. Epoch,' Mem. Geolog. Survey. London, 10, 1861 ; and 12, 1866. Tra- quair, Ganoid fishes of the British Carboniferous Formations, Palaeontogr. Soc. 1877. V. Wijhe, Visceral Skel. u. Nerven des Kopfes der Ganoiden u. Ceratodus, Nied. Arch. Zool., 5, 1882, p. 207. Zittel, Grundzuge d. Palaeon- tologie, Leipzig, 1895 (English Translation, Macmillan, 1902). Smith- Woodward, Outlines of Vert. Palaeontology, Cambridge, 1898. 160 SUB-CLASS GANOIDEI. agree with those zoologists who wish to unite the Ganoids and Teleosteans into a single group, distinct from other piscine orders. In this opinion we are in company with two of the greatest ana- tomists of the last century, J. Miiller and F. M. Balfour. The latter has expressed his view in words, with which we are in entire agreement and which we quote here, because they appear to express in the most judicial form the state of the question. He says, " We do not recommend such an arrangement (union of the Ganoids and Teleosteans) which in view of the great pre- ponderance of the Teleostei amongst living fishes would be highly inconvenient, but the step from Amia to the Teleostei is certainly not so great as that from the Chondrostei to Amia, and is un- doubtedly less than that from the Selachii to the Holocephali." The scales present some variation in arrangement and struc- ture. In the living Chondrostei they may be almost absent on the body as in Polyodon, or arranged in rows as in Acipenser ; not, however, forming a continuous cuirass except in the caudal region. They frequently carry bony spines, which are without any enamel cap. In some extinct Chondrostei they form a con- tinuous cuirass, and have often the rhombic form typical of the order. In Crossopterygii they form a continuous armour and are either rhombic or cycloidal ; in the living forms it can be shown that they are coated externally by ganoin and that in some cases they carry spines which consist of cones of dentine capped with enamel. In Lepidostei there is also a continuous armour of rhombic or cycloid scales, and in the living Lepidosteus it has been shown that these scales are coated with ganoin (yavos sheen) and may, especially in the young state, carry one or a number of small spines having exactly the structure of the spine of a placoid scale. In Amia the scales, which form a complete armour, are Teleostean in character and consist of bony plates without ganoin. Moreover it has been shown from a study of living forms that the scales save for the toothlike projections which occur in Poly- pterus and Lepidosteus are purely mesodermal products, and that the ganoin which was formerly thought to be enamel is really of the nature of vitrodentin and is formed by the scleroblasts * of the dermis. * Scleroblasts are cells which secrete a hard skeletal substance. VERTEBRAL COLUMX. 161 3 Although the scales are dermal structures, they are said to be frequently exposed on the surface in the adult. This, if true, must be due to the fact that the overlying epidermis and dermis have been rubbed off. This may frequently happen in the hand- ling of animals during their preservation. Vertebral column. In the Chondrostei the notochord is persistent and its sheath is stout but unsegmented and unossified. The neural and haemal arches on the other hand are segmented ; they are attached to the sheath but do not extend round it. In other living Ganoids (for the vertebral column of extinct genera the reader is referred to the special accounts), vertebral bodies are formed by the extension of the arch tissue round the sheath, its chondrification, segmentation and ossification. In Poly- pterus and Amia the vertebral bodies are amphicoelous as in Teleosteans ; in Lepi- dosteus they are con- cave behind and con- vex in front (opistho- coelous). In some extinct Ganoids (Fig. 90A) bony plates are found, corresponding to each arch, on the ventral side of the noto- chord. They are called hypoeentra, and carry the haemal arches. A corresponding plate, which may be composed of two pieces, is found on the dorsal side, and called pleurocentrum. Such incipient vertebrae are called half -vertebrae. They may, each of them, extend completely round the notochord (Fig. 90s), as in the tail of the living Amia, in which case one of them only carries the neural and haemal arches. Ring-vertebra is the term used when the pleurocent- rum and hypocentrum are joined to form a ring round the noto- chord, as in the amphicoelous vertebrae of Polypterus, Teleostei, etc., in which the bony ring has thickened so as to constrict the notochord in the middle of each vertebra. In Lepidosteus the arches are continuous with the bony centrum. In all other Ganoids with bony centra the arches are separated from the centra by persistent cartilage. The caudal fin is diphy cereal in Polypterus. In other living genera it is heterocercal. In Amia it is hemiheterocercal (ex- ternally homocercal, internally heterocercal), and the dorsal lobe of the caudal fin is reduced to the covering of fulcra. z.-n. M A B FIG. 90.— A. Vertebra of Caiurv.s furcatus. B. Caudal vertebrae of Enrycormus speciosus (after Zittel). 1 neural arch ; 2 the bifurcated neural spine ;3 hypo- centrum ; 4 pleurocentrum ; 5 rib. 162 SUB-CLASS GANOIDEI. The structure of the skull is very varied, but the cartila- ginous cranium always contains cartilage bones and is invested by membrane bones. The dorsal membrane bones are dermal structures and are frequently in the adult exposed on the surface in consequence of the superjacent epidermis and dermis having been rubbed off (see p. 161). The suspensorium is always hyostylic. The shoulder girdle and pectoral fin skeleton presents great variation, but on the whole inclines to the Teleostean type of structure. In the Chondrostei and in Amia the pectoral girdle is cartilaginous. In the other orders it is ossified very much as in Teleosts. In the Chondrostei and Crossopterygii there are three separate membrane bones in relation with the shoulder -girdle, the supra- clavicle, clavicle and infra-clavicle,* while in the other orders the infra-clavicle is absent. In the pectoral fin the skeleton of Polypterus is on the Elasmobranch type (rhipidostichous), in extinct Crossopterygians on the Dipnoan type (rachiostichous and mesorachic). In other orders the arrangement on the whole is Teleostean. The other anatomical characters of the Ganoids are (1) the possession of a conus arteriosus with more than one row of valves ; (2) the very general presence of a gill on the hyoid arch supplied by a branch of the ventral aorta, or of a spiracle (see below) ; (3) the presence of an air-bladder with pneumatic duct ; (4) the union of the urinary and Mullerian ducts and their opening by a median urogenital pore behind the anus ; (5) the testis appears to be connected except in Polypterus with the kidney by a testicular network ; (6) a spiral valve is present in the intestine (small in Lepidosteus and Amia) ; (7) the optic nerves form a chiasma ; (8) the openings of the nasal pits are double as in Teleosteans ; (9) the processus f alcif ormis and choroid glands are absent in Polypterus, Lepidosteus, and possibly in other Ganoids ; (10) the scales are bony plates embedded in the dermis, and frequently covered by a layer of peculiar substance called ganoin. Ganoin is probably vitro-dentin : it used to be * These bones are now often called supracleithrum, clei thrum and clavicle respectively, on the view that the last-named alone is homo- logous with the clavicle in the Amphibia and Amniota, . the cleithral elements not being represented in these groups. BRAIX. GILLS. 163 regarded as enamel. These scales in the young state and some- times throughout life bear spines which project through the epidermis and are formed of dentine capped by enamel (Poly- pterus, Lepidosteus}. (11) The fins frequently, not always, possess a single or double row of spine-like scales, called fulcra, on their anterior edge (absent in Polypterus, Polyodon and Amia). The dermotrichia are soft and segmented, and the pelvic fins are abdominal. (12) They lay relatively small eggs which undergo complete cleavage, and the young are hatched out as larvae which differ in many respects from the adult and possess a prone- phros as in Teleosteans. In Lepidosteus the medullary canal arises as a solid keel-like projection of the ectoderm which sub sequently becomes hol- low. The brain is on the whole Teleos- tean - like . There is a thin pallium (Fig. 91) and a val vul a cerebelli and the cerebrum is undivided ; though it may be grooved externally so as to suggest a division into two lobes. The branchial apparatus presents remarkable variation. In Spatularia (Planirostra edentula) there is no hyoid gill, but a pseudobranch in the spiracle as in Acipenser. In Polypterus there is no hyoid gill, nor spiracular pseudobranch. The spiracle is present in Acipenser, Polyodon and Polypterus, but absent in Scaphirhynchus, Lepidosteus and Amia. The following table summarizes the matter :— Hyoid gill, pseudobranch and spiracle . Acipenser. Hyoid gill, pseudobranch* but no spiracle Lepidosteus. * It is doubtful if the structure identified by Muller as pseudobranch in Lepidosteus is really such. FIG. 91. — Median section through the brain of Acipenser ruthenus (from Gegenbaur, after Goronowitsch). Cb cerebellum ; ch optic chiasira ; cp pineal body ; hy pituitary body ; pi pal- lium with choroid plexus projecting in between the" third ven- tricle z and the ventricle of the cerebrum V ; pi' roof of fourth ventricle ; Sv saccus vasculosus. 1G4 SUB-CLASS GAXOIDEI. Hyoid gill, but no pseudobranch or spiracle Scaphirhynchus. Hyoid gill absent, pseudobranch and spiracle present Polyodon (Spatularia). Hyoid gill and pseudobranch absent, spiracle present ..... Polypterus. Hyoid gill absent ; pseudobranch and spiracle also absent (4 double gills) . Amia. In those forms (Polypterus and Polyodon) without hyoid gill the ventral aorta still gives a branch to the hyoid arch. Abdominal pores* are present in all living Ganoids. Their external openings are placed on each side of the anus and they open internally into the body cavity. Pericardio-peritoneal canals are present in the Sturgeons as single unbranched tubes ; they appear to be absent in other Ganoids. Of these characters which on the whole suggest Elasmobranch more than Teleostean affinities, it will be convenient to treat the urinogenital organs more fully at this point. The urinogenital organs present many features of interest, and as they differ in important points in the different members of the group it will be necessary to describe them at some length. In all of those, the development of which is known, there is a pronephros in the larva, constructed on the Teleostean type, i.e. it consists of one large malpighian body (Fig. 92, v) connected by a convoluted tube (pr.n) with the anterior end of the longi- tudinal duct (sd) of the future kidney. In Lepidosteus and possibly in others, this body is connected with the peritoneal cavity by a ciliated canal (/). In none of them does the kidney show any differentiation into a meso- and meta-nephros, and in all the oviduct (mullerian) joins the kidney duct of its side posteriorly. Moreover mullerian ducts appear to be in all classes present, generally in both sexes, and they present the peculiarity of being short and opening into the body cavity at a point much nearer the anus than in other groups in which they are found. We have full knowledge of the urinogenital organs in two * Bles, "Correlated distribution of Abdominal Pores and Nephro- stomes in Fishes," Journ. Anat. and Phys., 32, 1898, p. 484. URINOGEXITAL ORGANS. 165 Ganoids ; one of these is Lepidosteus of which we have a description from the pen of Balfour,* and the other Polypterus in which their structure has been recently elucidated by Budgett.f In Lepidosteus the kidneys extend forward from the anus about three-fifths of the length of the body cavity (Fig. 93, Jc). Anteriorly they are continuous with a band of lymphatic tissue of a very similar appearance. The ureters (sg), which lie on their ventral side and receive the openings of numerous collecting tubes, enlarge posteriorly, and approaching each other coalesce to form an unpaired vesicle (bl) which opens by a median pore FIG. 92. — Diagrammatic view of the pronephros of Lepidosteus (from Balfour), isolated and seen from the side, pr.n Coiled tube ; sd longitudinal duct of kidney ; v malpighian body of pronephros ; / tube leading from v to the body cavity (peri- toneal funnel) ; bv blood vessel of glomsrulus oi v. on a papilla (ug) behind the anus. The ovary is a hollow sac, attached about its middle to the oviduct and continued back- wards and forwards from this attachment into a blind process. The oviduct is a thin- walled tube, continuous in front with the ovarian sac and opening behind (od) into the dilated part of the kidney duct of its side. In the male the testis is lobulated and the vasa efferentia pass in the mesorchium to the kidney, where they open into a longitudinal canal. From the longitudinal canal pass off tubules which open into the kidney tubules them- selves. Near the testis the vasa efferentia are united into an * Phil. Trans., 1882. f Trans. Zool. Soc., 15, 1901, p. 323. 16(3 SUB-CLASS GAXOIDEI. irregular network. No trace of the oviducts has been observed in the male. In Amia* there is a testicular network, and the posterior part of the kidney is provided with nephrostomes. In the female the oviduct opens into the body cavity. In Polypterus the oviduct is short and opens anteriorly into the body cavity, and posteriorly into the kidney duct just before its union with its fellow. In the male the kidney duct dilates behind and joins its fellow to form a considerable sinus which receives the testis duct of each side and opens to the exterior behind the anus. The testis is an elongated body almost as long as the kidney, to the ventral side of which it is attached. Two parts may be distinguished in it — an anterior dilated portion which forms spermatozoa and is the functional testis, and a posterior streak-like portion — the testis-ridge. — in which the ug „ bl FlG. 93 — Diagram of the urinogenital organs of an adult female of Lepidosteus (from Balf our and Parker), od Right oviduct ; the od which is placed on the upper side of the figure points a little to the "right of the opening of the oviduct into the dilated lower end of the kidney duct ; ov ovary ; bl urogenital sinus : uq urogenital aperture ; sg kidney duct ; the reference line goes a little too far in the figure ; k kidney, ly lymphatic organ. The organs of the right side only are shown. testicular tubules are few in number and do not form sperma- tozoa. The testicular tubules, both of testis and testis-ridge, open by numerous short ducts into the testis duct which is a longitudinal canal extending the whole length of the organ close to the ureter. In Polypterus then the testis tubules do not communicate with the kidney tubules, but there is a special testis duct which opens behind into the kidney duct at the point at which the oviduct opens in the female. No trace of oviducts has been seen in the male and there are no nephrostomes in the adult. In the Sturgeon there is a testicular network in the male, as was first shown by Rathke, and there are short mullerian ducts in both sexes open- ing widely into the body cavity and behind into the kidney duct. In the female the miillerian duct is of course the oviduct. Open mullerian ducts have been seen in both sexes of Scaphirhynchus and Polyodon, but a testicular network has not yet been seen in those genera. * Jungersen, Zool. Anzeiger, 23, 1900, p. 328. ' CHONDROSTEI. 167 The Ganoids reached their greatest development in the Palae- ozoic, Triassic and Jurassic periods. From the Cretaceous epoch onwards they have gradually become less numerous, until at the present day they are represented by a few widely scattered, extremely isolated, and for the most freshwater genera. The Ganoids are here divided into four orders — Chondrostei, Cross- opterygii, Lepidostei and Amioidei. Order 1. CHONDROSTEI. Endo-skeleton largely cartilaginous ; head covered with bony plates. Body naked or with rows of bony plates, or with rhomboid, rarely cycloid scales. Operculum weakly developed, branchio- stegals numerous, few, or absent; teeth small or absent. Infra- clavicle present. Caudal fin usually heterocercal. One dorsal and one anal fin, with fulcra. Paired fins non-lobate. Pelvic fins with a series of basal cartilages. Lower Devonian to present time. So far as the facts which can be made out from a study of the extinct forms are concerned, this order agrees with the Crossopterygians in the feebleness of the ossification round the notochordal sheath (except Polypterus), and in the presence of infra- clavicles ; it differs from them in the absence or absorption into the body of the basal lobe of the paired fins, in the heterocercal tail, and in the absence of jugular plates between the rami of the mandible. There are two living families, the Acipenseridae and theSpa- tulariidae. The following remarks apply to them : — The tail is heterocercal. The notochord is persistent and un- constricted. Its sheath is thick, unsegmented, and entirely within the elastica externa. The neural and haemal arches are seg- mented (except in front) and placed upon the sheath, but are not continuous with one another round it (fig. 94). There are inter- calated pieces both in the dorsal and ventral arches, and there is a longitudinal ligament on the neural arches. The haemal arches of the two sides meet ventrally and enclose the dorsal aorta. In addition they carry lateral projections which must be regarded as the transverse processes. These meet ventrally in the tail to enclose the caudal vein, so that in the caudal region there are two canals, one for the dorsal aorta and a ventral one for the caudal vein. The anterior part of the column is continuous with the skull, and here the neural and haemal arches are not 168 SUB-CLASS GANOIDEI. segmented. The notochord is continued into the base of the skull to the pituitary fossa in Polyodon, but not in Acipenser. In the Acipenseridae there are five rows of large scales with projecting spines, * one dorso-median and two on each side (Fig. 96). Between these are numerous small scale-like plates, which frequently carry one or more spines. In the caudal region these smaller plates, also carrying spines, are rhombic, in contact, and regularly arranged. In Polyodon (Spatularia), which is often described as naked, there are also small spine- bearing scale-like structures. The scales are placed in the sub- cutaneous tissue and are composed of bone (with bone corpuscles -and branched processes, and in the larger scales haversian canals). The spines are composed of a similar substance in layers, either with bone cells in their lower portions or without them. There is no enamel and appar- ently no ganoin. The cranium^ is massive and largely cartilaginous, there being but slight traces of ossification (in the periotic region). As in other Ganoids and in Teleostei the inner wall of the otic capsule is not developed. The roof con- sists of almost continuous cartilage, but the fontanelles though not conspicuous are not altogether absent. It is completely covered by membrane bones which lie in the dermis. There is a large parasphenoid and a single or paired vomer in the floor. There is a partially ossified hyomandibular and a symplectic cartilage (Fig. 95) which suspend the mandibular arch. The anterior ends of the pterygoids meet in a symphysis. The visceral arches are partially ossified but the meckelian cartilages are unossified. The hyomandibular carries an opercular plate and the hyoid arch a branchiostegal ray. Membrane bones (a * H. Klaatsch, Morph. Jahrb., 16, 1890, p. 146. O. Hertwig, ibid., 2, 1876, p. 373, f Bridge, Osteology of Polyodon. Phil. Trans. 1878, p. 683. FlQ. 04.— Lcteral view of the vertebral column of Spatularia (from Wieders- heim). Cs sheath of notochord ; Ic intercalated piece ; Ob neural arch ; Ub haemal arch ; Ps neural spine. CHONDROSTEI. 169 maxilla and dentary) are developed in connection with the jaws. Teeth are entirely absent in the adult, but small teeth are present in the young Polyodon and possibly other genera. There are five branchial arches. In the unpaired fins the fin rays (dermotrichia) are more numerous than the interspinous supports (somactids), and the ca-udal fin has fulcra in a single row. The pectoral arch is un- ossified and the^two halves are not united ventrally. They are covered by membrane bones; supra-clavicles, clavicles, and infra- clavicles being present. The pelvic arch is absent, its place being taken by enlarged basal cartilages of the fins as in Teleostei. The paired fins are without a basal lobe (non-lobate), their whole free portion being supported by dermal fin-rays. These are carried by a row of cartilages of which the posterior is possibly FIG. 95. — Cephalic skeleton of the sturgeon (from Claus after Wiedersheim) . Ro Rostrum; i_ Cn nasal pit ; 0 orbit ; H m hyomandibular ; S symplectic ; Pq palatoquadrate ; M d lower jaw ; Hy hyoid ; V foramen for vagus ; R ribs. a basal somactid (metapterygium) and the others peripheral somactids which have become articulated to the pectoral arch. The anterior dermal fin- ray is enlarged and directly articulates with the pectoral girdle. The branchial apparatus presents remarkable variability. In Acipenser there is a spiracle with a pseudobranch, and a hyoid demibranch ; in Scaphirhynchus there is no spiracle or pseudo- branch, but a hyoid demibranch is present. In Polyodon spiracle and pseudobranch are present, but there is no hyoid gill. Each of the first four branchial arches carries a double row of gill- filaments and there is a cleft behind the fourth gill arch. In Acipenser the inter-branchial septa are fairly broad. In the conus there are three longitudinal rows of valves with three or four valves in each row. 170 SUB-CLASS GANOIDEI. Fam. 1. Acipenseridae. Sturgeons. With five rows of keeled plates in the skin, elongated snout, small mouth without teeth, with four barbels in front of the mouth, opercular and four double gills. Stomach without blind sac. With fulcra on the dorsal lobe of the caudal fin. Dorsal and anal fin with two rows of supports (axonosts and baseosts). Large fishes in the seas and rivers of the northern hemisphere, feeding on small animals and plants. Most are migratory (anadromous). Caviare is the ovary of the sturgeon. Acipenser L. with spiracle. Scaphirhynchus Heckel, without spiracles. Fossil remains rare ; an Eocene species (A. toliapicus Ag.) is known, and isolated remains of scales from the Upper Lias of Whitby (Gyrosteus) and from the Upper Chalk of Kent (Pholidurus). Fam. 2. Spatulariidae. Skin with small isolated scales, tail scaled as in sturgeons, snout prolonged into a thin flat blade ; without barbels ; with spiracle, without hyoid gill ; gills, 4 ; gill-rakers long, in a double series on each arch, except on the fifth, which has only one series. Air-bladder cellular. Jaws with fine teeth in young individuals. Fresh- FiG. 96. — Aciyenser ruthenus (after Heckel and Kner). waters of N. America and China. Polyodon Lac., Mississippi ; Pse- phurus Giinther, Chinese rivers ; a fossil genus Crossopholis Cope, from the Eocene of Wyoming. EXTINCT FAMILIES. Fam. 1. Chondrosteidae. Parietals and frontals paired ; near the parietal a great squamosal. Jaw edentulous. Branchiostegal rays present. Operculum small, sub-operculum large. Body naked ; the dorsal lobe of the caudal fin with fulcra and covered with rhombic ganoid scales. Chondrosteus Egerton, Lower Lias, England. Fam. 2. Belonorhynchidae. Trias, Lias, may be placed here. Fam. 3. Palaeoniscidae. Body elongated with rhombic, rarely cycloid scales ; with operculum, sub-operculum and branchiostegal rays. In no single genus are the osteological characters well known. Devonian to Jurassic. Cheirolepis Ag., Devonian ; Palaeoniscus Blv., Upper Per- mian ; Coccolepis Ag., Jurassic. Fam. 4. Platysomidae. Deep-bodied fishes from the Carboniferous to the Permian ; very similar to the Palaeoniscidae. Cheirodus M'Coy, Eurynotus Ag., Platysomus Ag., Benedeniits Traquair. The Catopteridae, Triassic, may be placed here. CROSSOPTERYGII. Order 2. CROSSOFTERYGII.* Vertebral column ossified, or unossi- fied. Tail diphycercal or hetero-diphy- cercal (upper lobe of caudal fin weaker than the lower lobe, Fig. 103). Paired fins with scaled basal lobe which is either pointed or rounded. In the former case it is unibasal, rachiostichous and mesorachic, in the latter case, found only in living forms, the pectoral fin is tribasal and rhipidostichous . In place of the branchiostegal rays are two large jugular plates between the rami of the lower jaw, near which there are in many palaeozoic forms a number of smaller lateral plates and a median anterior plate. Ganoid scales, rhombic or cycloidal, cover the whole body and tail. Dorsal fin either two in number ; or if single, long or multifid. Devonian to present time. This order which was established by Huxley f in 1861, is mainly based on the form of the fin, which is fringed with dermotrichia on both sides. It is difficult to state any other charac- ters peculiar to the living and extinct forms, unless it be the two jugular plates between the rami of the lower jaw. The following remarks apply to the living Polypteridae :— The scales of Polypterus are very similar to those of Lepidosteus (see p. 177). 171 FIG. 97.— Cheirolepis Irailli ; re- stored (except margin of fins) by Traquair (from Snrith- Woodwara), quarter natural size. They are " Traquair, Cranial Osteology of Polypterus, Journ. Anat. and Physiology, 4, 1871. Pollard, " Anatomy of Polypterus," Zool. Jahrb.,5, t Preliminary Essay upon the systematic arrangement of the fishes of the Devonian Epoch, Mem. Geol. Survey United Kingdom, 1861. 172 SUB-CLASS GANOIDEI. attached to one another in much the same way and are arranged in oblique rows. They are rhombic in form and similarly covered by ganoin. The canals in the deeper part of them, containing blood vessels, open on the surface, but differ from those of Lepidosteus in being branched. They are without spines and the remains of spines, except at the bases and on the hinder surface of the pectoral fins and on the plates overlying the shoulder girdle. They are covered by soft skin in the adult, but this is frequently rubbed off in pre- served specimens, leading to the view that the scales are freely exposed in the adult. The vertebral column is completely ossified ; the vertebrae are amphi- coelous, as in Teleosteans, and carry in the body two pairs of ribs, of which the shorter and ventral pass between the muscles and the peritoneum and correspond to the ribs of other forms, while the dorsal and larger pair lie between the dorsal and ventral lateral muscles. The body of the first verte- bra is united to the skull, but the arch and ribs are separate. The cranium consists mainly of persistent cartilage (Fig. 98) with a large fon- tanelle in the roof and in the floor (pituitary) and a few cartilage bones, of which the most important are exoc- cipitals which completely surround the foramen magnum, opisthothic (~), sphenoid (? sphenethmoid), postfrontal (8), prefrontal (9) and a median ethmoid (^). There is a slender arcade of car- tilage passing forward from the postfrontal to the anterior part of the sphenoid cartilage. The whole is invested dorsally by membrane bones, while ventrally there is a parasphenoid which extends back beneath the body of the first vertebra, and two vomers (as in Lepidos- teus). The mandibular arch is very like that of Teleosteans : in the upper jaw arcade (Fig. 100) are palatine, ectopterygoid, -7 FIG. 98. — Dorsal view of the u cartilaginous cranium of Polypterus with the mem- brane bones removed (after Traquair).- 1 Ethmoid, 2 nasal opening, 3 sphenoid, 4 optic foramen, 5 occipital bone, 6 foramen magnum, 7 opisthotic, 8 post-frontal, 9 pre-frontal. CROSSOPTERfGII. 173 entopterygoid, metapterygoid and quadrate ; the lower jaw has an articular and mento-meckelian cartilage bone and an angular, splenial and dentary, the two latter being dentigerous. On the upper jaw 1C. 'I A 1312 n : /\ ,• 10 16 15 24 25 19 FIG. 99. — Side view of skull of Polypterus with membrane bones (after Traquair). 1 Nasal opening, 2 premaxUla, 3 ethmoid, 4 accessory nasal, 5 os terminale, 6 nasal, 7 anterior suborbital, 8 posterior suborbital, .9 postfrontal, 10 frontal, 11 prespira- cular, 12 spiraoular bones, 13 parietal, 14 post-spiracular, 15 supratemporal, 16 post-temporal, 17 operculum, 18 sub-oper- culum ; 19 preoperculum, 20 quadrate, 21 jugals, 22. articular, 23 angular, 24 dentary, 25 maxilla. are premax- illae, max- illae and jugals (two). The hyo- m a n dibular is ossified and bent ; there is no symplectic. There is an ope rculum, s u b o p e r- culum and a large bone called preoperculum. There are also circumorbital bones and pre- and post-spiracular bones (Fig. 99), as well as two bones (12) — the spiracular bones — in the flap which covers up the spiracle. There are also three supra- temporals (15) and a post-temporal (16) on each side. The membrane bones are in the skin and their outer surfaces are smooth and tuberculated, and not exposed in uninjured speci- mens. The pre- maxillae and maxillae bear teeth, and the vomers, ptery- goid and para- sphenoid are covered with fine closely set teeth. The cartilaginous shoulder girdle is small and contains a scapular and coracoid ossification ; it is overlaid by a series of clavicular bones, of which the supraclavicle is attached to the pos t- temporal; FIG. 100. — Side view of palatoquadrate, hyoman dibular and opercular bones of Polypterus (after Traquair). 1 vomer, 2 palatine (small and not appearing in palate), 3 ectoptery- goid, 4 entopterygoid, 5 quadrate, 6 subopcrculum, Jf operculum, 8 hyomandibular, 9 metapterygoid. 174 SUB-CLASS GANOIDEI. there is an infraclavicle (clavicle, see note, p. 1G2). The skeleton of the pectoral fin is tribasal as in Elasmobranchs ; it consists of three basal somactids, carrying two rows of peripheral somactids (r, r"} to which the dermotrichia are attached (Fig. 101). The pro- and meta- pterygium are ossified, the meso- pterygium is mainly cartilaginous but contains an ossification (o). The pelvic girdle is represented by a small piece of cartilage (Fig. 102) to which the large ossified basal somactid is attached. The latter carries a row of ossified peri- pheral somactids to which the der- motrichia are attached. There are no fulcra. There is a number of dorsal fins in Polypterus, each with an anterior spine ; and the tail is diphycercal. In Polypterus there are three main longi- ,* tudinal rows of valves in the conus arteriosus with nine valves in each row, and between these there are three incom- plete rows of smaller valves (about forty-five in all). Both ductus Cuvieri open into the auricle. The spiracle is present, but the hyoid gill and pseudobranch are absent. The ventral aorta sends off a branch to the hyoid arch which supplies the external gill of the larva. The fourth branchial arch bears only one row of filaments and has no slit behind it. The stomach has a caecum and there is one pyloric appendage. The air bladder is double and cel- lular and its duct opens into the ventral wall of the pharynx ; its blood supply is from the last efferent branchial vessel and its vein joins the hepatic. According to Budgett Polypterus is capable of FIG. 101.— Skeleton of pectoral fin of Polypterus (from Gegenbaur). R propterygium, R' metaptery- gium ; o ossification in meso- pterygium ; r', r primary radials (basal somactids) ; r" secondary radials ; s fin rays. Fia. 102. — Pelvic girdle p and skele- ton of pelvic fin of Poly- pterus (from Gegenbaur, after v. Davidoff). 6 basal somac- tid ; r peri- pheral somac- tids ; ft trace of pelvic girdle. CROSSOPTERYGII. 175 breathing air. The pituitary body retains its opening into the buccal cavity through life.* The young Polypterus has an external gill, which is attached to the operculum. Classification of Crossopterygii, living and extinct. Sub-order 1. OSTEOLEPIDA. Notochord more or less persistent. Pectoral fins rounded or pointed, unibasal. Nares on the lower surface of snout. Fam. 1. Tarrasiidae. Axonosts and baseosts of median fins in simple regular series, much fewer in number than the dermotrichia. Tarrasius Traquair, Calciferous Sandstones (Lower Carboniferous) of Dumfriesshire. Fam. 2. Holoptychiidae. Body covered with overlapping cycloid ganoid scales ; vertebral column unossified ; pectoral fin with long pointed scaled axis (mesorachic) ; tail hetero-diphycercal. Axonosts of each of the dorsal and anal fins fused into a single piece with a broad distal end, bearing three to six rod-like baseosts, which are much fewer than and overlapped by the dermotrichia in all the median fins. Lateral jugular Fio. 103. — Holoptychius flemingi, Devonian, Scotland, after restoration by Traquair, from Woodward. plates, clavicles and infra-clavicles present ; the teeth have a compli- cated folded structure. Holoptychius Ag. (Glyptolepis, Platygnathus Ag.) Devonian ; isolated teeth have been described as Dendrodus, Lamnodus, Apedodus. Fam. 3. Rhizodontidae. Like the preceding but with shorter pec- toral and pelvic fins. Devonian, Carboniferous. Ehizodus. Strepsodus Rhizodopsis, etc. Fam. 4. Osteolepidae. Ring vertebrae in the caudal region, Devonian. Osteolepis, Thursius, Diplopterus, Megalichthys. Fam. 5. Onychodontidae, known only by fragments from the Devonian, is placed here. Fam. 6. Coelacanthidae. Notochord persistent, vertebral column un- ossified. Axonosts of each of the dorsal and anal fins fused into a single piece ; a series of axonosts, equal in number to the neural and haemal spines present in the caudal fin above and below, each axonost directly connected with a single dermal fin ray. Caudal fin diphycercal ; air-bladder with ossified walls ; paired fins with short, obtuse axis ; only one opercular bone. Lower Carboniferous to the Upper Chalk. Coela- canthus Ag., Undina Miinst., Macropoma Ag., etc. * Bickford, Anat. Anz., 10, 1895. 176 SUB-CLASS GANOIDEI. Sub-order 2. CLADISTIA. Notochord more or less constricted and replaced by ossified vertebrae. Baseosts in median fins rudimentary or absent ; axonosts of dorsal fins equal in number to the apposed dermotrichia. Pectoral fins di- or tri- basal. Nares on the upper surface of the snout. Fam. 7. Polypteridae. Body covered with rhombic ganoid scales ; vertebral column ossified ; tail diphycercal ; pectoral fins tribasal ; two jugular plates only ; dorsal fins numerous ; pulp cavity of teeth simple. Polypterus Geoffr. (Fig. 104), with pelvic fins, rivers of North and Equatorial West Africa ; Calamoichthys Smith, elongated and without pelvic fins, rivers of Old Calabar and the Cameroons. FIG. 104.— Polypterus bichir (from Claus). Order 3. LEPIDOSTEI. Body covered with rhombic or rhomboidal scales arranged in oblique rows and articulated together. Caudal fin hemi-hetero- cercal. Vertebral column in the most different degrees of ossification. Unpaired, and sometimes paired fins with fulcra. Branchiostegal rays numerous, and often a median jugular plate Always four opercular bones ; between preoperculum and orbit at least one row of postorbitals. Infraclavicle absent. Somactids of the unpaired fins as numerous as the dermotrichia. Teeth pointed or conical. The pelvic fins are without baseosts. This order, which appears to be closely allied to the Palaeonis- cidae on the one hand and to the Amioidei on the other, in- cludes the living genus Lepidosteus. With the exception of the Permian genus Acentrophorus, the extinct members are found in the Lias, Jurassic, Lower Cretaceous, and Tertiaries. The following remarks apply to the living Lepidosteidae. The body is covered by rhombic scales articulated together (see p. 177). The tail is hemi-heterocercal. The paired fins are non-lobate, and all the fins bear paired fulcra. The jaws are much elongated, forming a snout ; the premaxillae form most of the upper jaw. Both jaws bear teeth, small and large, and there are fine close-set teeth on the palatines and vomers. The vertebrae are well developed and ossified, and have opisthocoelous centra. The chondrocranium is large and cartilaginous with LEPIDOSTEI. 177 embedded cartilage bones much as in the salmon, and is com- pletely invested by membrane bones. The vomer is double. There is a symplectic in the suspensorum, several bony elements are present in the mandible, and the maxilla is divided trans- versely into many bones. The pectoral girdle is as in Teleosteans, with scapula and coracoid ossifications and overlying clavicle. The pectoral fin has one row of basal elements carrying the dermotrichia. The pelvic girdle is absent. In Lepidosteus * the scales of the trunk are arranged in oblique rows which pass from above and in front backwards and ventral- wards. The scales of a row are more closely connected with each other than with those of neighbouring rows, in consequence of a peg and socket articulation. Each scale has its anterior and dorsal angle produced into a pro- cess which fits into an excavation under the next dorsal scale of the same row, where it is attached by a ligamentous band. The centre of the scale is bored by one or several canals which pass Fm lp5 * 4 portion of th*armour of right through it and transmit blood vessels. The scale con- sists in its deeper portions of bone with bone corpuscles and vertically directed fibres, which are prolongations of large scleroblast cells at the surface (so- called odontoblasts, they do not become enclosed) and of a superficial layer of structureless dentine or ganoin, as it was called by Williamson, which is only found on the exposed part of the scale, and was formerly taken for enamel. In some parts of the body (ventral side of head, investing bones of skull and shoulder girdle, the fin scales) the scales carry one or a number of small teeth, which consist of enamel, dentine with fibres, and a pulp-cavity. In the young animal all the scales carry teeth, with the dentine of which the ganoin of the scale is continuous. The scales are developed as plates in the dermis, * O. Hertwig and Klaatsch, op. cit. ; see also W. C. Williamson, On the microscop. structure of scales, etc., of fishes, Phil. Trans., 1849 and 1851. Nickerson, Scales of Lepidosteus, Bull. Mus. Comp. Zoology r Harvard, 24, 1893. z. — II. N 178 SUB-CLASS GANOIDEI the scleroblasts of which become successively included in the scale to form the bone corpuscles. The ganoin is formed later (in a fish of fifty-two months) by the cells on the upper surface of the plate. The teeth are developed as in Selachians on a papilla of dermis projecting into the epidermis. The cells of the papilla form the dentine and those of the epidermis the small cap of enamel. Some of the scales possess processes which indicate where teeth were formerly attached. .....Olf There is no spiracle in the embryo or adult, though there is a pouch-like trace of one in the embryo ; but the hyoid bears a gill, dorsal to which is a structure called by J. Miiller the pseudobranch. This so-called pseudobranch is not found in the larva. There are four double gills, and a cleft behind the fourth branchial arch. The conus arteriosus possesses eight * equally developed longi- tudinal rows of valves, five in each row. The ventral aorta gives off on each side posteriorly two branches, one to the third and fourth branchial arches, and one to the second ; anteriorly it also gives off two, the posterior of which goes to the first branchial and the anterior to the hyoid. The so-called pseudobranch ap- pears to be supplied by arterial blood from the efferent vessel of the first branchial arch, as in other fishes. The left ductus Cuvieri opens into the sinus, the right into the auricle. The stomach is without a blind sac and there is a number of pyloric caeca and a small but distinct pancreas. The intestine is coiled and has a spiral valve pos- * Boas states, Morph. Jahrb. 6, p. 323, that the valves of four of these rows are smaller than those of the other four. There seems to be some variation in the valves of the conus of Ganoids. — ce — op. — -cb m o FlQ.\106. — Dorsal view of the brain of Lepidosteus (after Balfour and Parker), cb cerebellum ; ce, ce an- terior and posterior lobe of cere- brum ; TO o medulla oblongata ; olf olfactory lobes ; op,l optic lobes ; v,th vesicle of thalamen- cephalon. LEPIDOSTEI. 179 teriorly. The air-bladder is single and opens into the pharynx dorsally. It has a continuous central cavity opening on each side into lateral chambers which are placed in the thickness of the wall. In Lepidosteus the cerebrum (Fig. 106) is divided into two parts, an anterior and a posterior. The anterior part tapers in front into the olfactory lobes and is double ; the posterior part is single and its ventricle possesses a thin roof like that of the thalamencephalon, with wilich it is continuous. The dorsal part of its side walls, where they pass into the roof are everted and thickened and form the prominent posterior cerebral lobes (Fig. 106). The anterior part of the roof of the thalamen- cephalon is produced dorsalwards into a large thin-walled vesicle which projects just in front of the pineal body. The cerebellum is of medium size and has a forwardly projecting lobe. FIG. 107. — Dapediits politus, restored with scales removed, quarter natural size, Lower 'Lias (from British Museum Catalogue). Fam. 1. Lepidosteidae. Body covered with thick, rhombic, ganoid scales, vertebral column completely ossified, vertebrae opisthocoelous ; tail heterocercal ; the snout is much elongated. Lepidosteus Lac., gar- pikes. Fresh-waters of N. America and Cuba, one species is known from China. Sluggish habit, voracious, valueless as food. Extinct families. Fam. 1. Stylodontidae. All fins with fulcra. Jaws and vomer with several rows of teeth. Vertebral column composed of half vertebrae or of ring-vertebrae. Upper Permian to Cretaceous. Acentrophorus Tra- quair. Semionotus Ag., Dapedius de la Beche, a deep-bodied fish (Fig. 107). 180 SUB-CLASS GANOIDEI. Fam. 2. Sphaerodontidae. Very similar to preceding ; with obtusely conical or chisel-shaped teeth. Trias to Chalk. Colobodus Ag., Lepi- dotus Ag. (Fig. 108). Fam. 3. Eugnathidae. Very similar to preceding ; elongated bodies ; caudal fin homocercal or hemi-heterocercal ; Trias to Cretaceous. Eugnathus Ag., Ptycholepis Ag., Caturus Ag., Strobilodus Way. Fam. 4. Macrosemiidae. Macrosemius Ag., Ophiopsis Ag., Prop- terus Ag., Notagogus Ag. Fam. 5. Pholidophoridae. Pholidopleurus Bronn. Fam. 6. Pycnodontidae. Body laterally compressed, high, oval ; covered by rhomboidal scales, articulated together and strengthened by a vertical ridge ; scales sometimes absent in the caudal region, rarely absent altogether. Notochord persistent, without ossifications in its sheath, but ribs, arches and spinous processes ossified. Caudal fin inter- nally hemi-heterocercal ; fulcra absent ; pelvics small, anal and dorsal Jong ; fulcra absent ; somactids of the unpaired fins equal in number to FIG. 108. — Lepidotus minor, restored, one-fifth natural size, Purbeck'JBeds (from British Museum Catalogue). the segmented dermotrichia ; opercular apparatus reduced, often only one opercular bone present. Dentition of oval, crushing teeth. The chondrocranium often well ossified. Infraclavicles absent. Jurassic, Cretaceous, Eocene. Gyrodus Ag., Mesturus Wagn., Mesodon Wagn. (Fig. 109), Pycnodus Ag., etc. Fam. 7. Aspidorhynchidae. Very similar to preceding, but with ring-shaped or complete and biconcave vertebrae (pleurocentra and hypo- centra never distinct). Caudal fin homocercal ; snout elongated and pointed ; scales rhomboidal, unequal ; lower jaw with movable pre- mandibular ; fulcr.a weak. Lower Oolite to Upper Chalk. Aspidorhynchus Ag., Belonostomus Ag. Fam. 8. Lepidosteidae. See p. 179. Order 4. AMIOIDEI. Body covered with thin cycloid or rhombic scales, overlapping, but not articulating, without ganoin. Caudal fin internally heterocercal, externally symmetrical (hemi-heterocercal). Vertebral column with either half -vertebrae or completely ossified amphi- coclous vertebrae, or without vertebrae, ossification extending from AMIOIDET. 181 the arches into the notochordal sheath. Fulcra present or absent. Branchiostegal rays flattened ; a median jugular plate. Teeth pointed, conical. From the Lias onwards. This order, which is close to the Lepidostei, contains the single living genus Amia. It approaches more closely to the Teleostei than any other ganoid fish, but it differs from them FIG. ] 09.— Mesofon macropterus, restored, with cheek-plates removed, two-thirds natural size. Upper Jurassic (after Smith Woodward), fr frontal, meth mesethmoid, md mandible, op operculum, orb orbit, p.op prcoperculum, pa parietal, pmx premaxilla, socc supra occipital, sq squamosal, v vomer. (The caudal region is destitute of scales in this fish.) in important features of the urinogenital organs and structure of the heart, and alimentary canal. Fam. 1. Pachycormidae. Extinct. Ethmoid region more or less produced in front of the mouth ; vertebral column variable as in the order. Fulcra present, branchiostegal rays numerous. Lias to Lower Cretaceous. Pachycormus Ag., Euihynotus Wagn., Hypsocormus Wagn. Fam. 2. Amiidae. Vertebral column ossified with complete amphi- coelous vertebrae. The caudal region consisting of vertically divided 182 SUB-CLASS GANOIDEI. half -vertebrae, of which one half (variously described as the anterior and posterior) carries the arches. Fulcra absent. Tail somewhat heterocercal, more so in the young. The scales of Amia, which are thin, elastic and cycloidal, consist of a superficial layer containing bone corpuscles, and a deeper layer containing fibres. They thus closely resemble the scales of Teleostei, and are without a superficial layer of ganoin (Klaatsch, op. cit., p. 179). The skull* is decidedly Teleostean in its structure. The chondrocranium is well developed and contains a few widely separate cartilage bones. As in Lepidosteus, it is without fenestrae in the roof. From the fact that two neural arches are attached to the basioccipital bone, it would appear that the centra of the two first vertebrae are fused to the skull. The investing bones of the dorsal surface, though closely applied to the carti- lage, are dermal structures, and are for the most part not covered externally by soft skin in the adult. The vomer is double and bears teeth. The suspensorial apparatus is almost exactly similar to that typical of Teleos- teans. The premaxillae, maxillae, palatines, and pterygoids also bear teeth. There are four opercular bones, numerous branchiostegal rays, and a median jugular plate, The shoulder girdle is cartilaginous and is over- laid by the large clavicle. It carries one basal cartilage (metapterygium), to one side of which the somactids, which carry the dermotrichia, are attached. Spiracle, pseudobranch and hyoid gill are absent. There are four double gills and a slit behind the fourth arch. The conus arteriosus has three transverse rows of valves with four or five valves in each row. The base of the ventral aorta is slightly swollen into a kind of bulbus. The air-bladder is cellular and lung-like and opens by a duct into the dorsal wall of the pharynx. Stomach with a blind sac, pyloric caeca absent. The intestine contains a spiral valve. For urinogenital organs, see p. 166. A single living genus Amia L., with one species, in the fresh waters of the United States ; flesh valueless as food ; species are known in the Lower Tertiaries of Europe and N. America. Megalurus Ag., with fulcra, from the Upper Jurassic may be placed here. Fam. 3. Oligopleuridae. Upper Jurassic to Upper Cretaceous. Oenos- copu» Costa, Spathiurus Davis. * Bridge, Cranial Osteology of Amia calva, Journ. Anat. and Phys., ., 1877, p. 605. CHAPTER VIII. SUB-CLASS (AND ORDER) TELEOSTEL* Fishes with a bony endoskeleton, distinct, usually amphicoelous, vertebrae, a supraoccipital bone, pectinate gills, a branchial oper- culum and usually a hyoidean pseudobranch. Without spiracle, conus arteriosus, optic chiasma, and intestinal spiral valve. An air-bladder is often present. The gonads are usually continuous with their ducts, and the testes are not connected with the kidney. The eggs are heavily yolked, but usually small, and the young are, with a few exceptions, hatched in an immature condition and undergo a larval development. The Teleostei include the vast majority of living fishes. They are found in freshwater as well as in the sea, and in some cases they possess organs which enable them to exist in or to breathe air (Anabas, Periophthalmus, fishes which live in foul or muddy water, e.g. many marsh fishes). The form of the body is exceedingly variable, most often it is typically piscine, but it may be elongated and snake-like as in the eels, strongly compressed laterally in the ribbon-fishes (Trichiurus, etc.) and in the flat-fishes (Pleuronectidae), or the vertical axis may equal or even exceed the longitudinal in length (Orthagoriscus). The tail which is usually the principal organ of locomotion is in the last-named modification so much reduced that it appears to be absent, and in the sea-horses (Hippocampus) it is without caudal fin and is used as a pre- hensile organ. The body is divided into head, trunk, and tail ; the gill- opening usually marks the boundary between the head and * See, besides the works of J. Miiller, Gxinther, Day, Jordan and Ever- mann, Boulenger, Bridge, already cited, G. B. Goode and T. H. Bean, " Oceanic Ichthyology," Memoirs of the Museum of Comparative Anatomy at Harvard College, 22, 1896. 184 SUB-CLASS (AND ORDER) TELEOSTEI. trunk, and the vent that between the trunk and tail. The mouth is usually at the anterior end of the head,but it varies considerably in form and position ; it may be anterior or superior (on the upper surface), or inferior or lateral (extending along each side of the head). The upper jaw may be formed by the premaxillaries (intermaxillaries) and maxillaries or by the premaxillaries only, and both jaws may be provided with tactile appendages called barbels. The nostrils are usually double on each side, and in a few cases (some eels) the anterior or lower opening perforates the upper lip. The eyes are large, are usually without lids, and the cornea is flat. In cave-fishes and in abyssal forms they may be much reduced and even hidden beneath the skin, but in some deep-sea forms they are enormously enlarged. The branchial slits are hidden by the gill-cover, which is a fold of skin, con- taining dorsally four flat bones, the opercular bones or opercles, and ventrally a number of bony rods called the branchiostegal rays. The branchial aperture or aperture of the space bounded by the operculum is usually a slit of considerable dorso- ventral extent, but in some cases it is much restricted, and in Sym- branchus, the openings of the two sides coalesce in the middle line. The space on the throat between the gill apertures is called the isthmus. There is no spiracle. The actual gill-openings themselves, i.e. the openings leading from the pharynx into the space below the gill-cover are not tubular but simply slits, separated by the rod-like branchial arches. These carry the gills which are usually filiform and project freely. There are usually two rows of them (holobranchs) on each branchial arch, projecting like the prongs of a comb ; hence they are said to be pectinate (Fig. 118). There are usually five gill-slits, but the last is always smaller than the others, and is sometimes absent, in which case the fourth branchial arch bears only one row of gill-filaments (hemibranch). They may be still further reduced. There are five branchial arches, but the last never bears filaments. The hyoid never carries true gill-filaments, but it generally carries them in a reduced form as a pseudobranch. In some cases the pseudobranch is without a trace of filaments and has the form of a red glandular patch. The vent is usually at the junction of the trunk and tail, but it may be shifted far back on the tail or far forward on the trunk. It consists of at least two openings placed close together, the foremost being the anus FINS. 185 and the hinder one the urinogenital aperture. The genital aperture may, however, in some forms be separate from and in front of the urinary. The median fins are subject to considerable variation. There are usually two dorsal fins, a caudal and an anal fin, but there may be a single dorsal fin continuous round the tail with the anal, or there may be a series of finlets in place of the dorsal fin, and the anal may be multiple or absent. The paired fins are in two pairs, the pectoral and pelvic (sometimes called ventral}. Of these the pel vies are small, vary considerably in position and are sometimes absent (fishes which live in mud) ; they may be behind the pectoral on the abdominal surface in which case they are said to be abdominal in position, or they may be below the pectorals (thoracic] or in front of them (jugular}. In some cases they coalesce and form a suctorial organ (Gobies) and in some Blennies they are adapted for walking. The pectorals are usually close behind the gill-opening and vary much in size. They assist the fish in balancing and enable it to execute back- ward movements. In some forms (Periophthalmus, Trigla9 Lophius, etc.) the pectoral fins are used for walking, and in the flying fish (Exocoetus) they are enlarged to form parachute-like organs. All the fins are supported by osseous fin-rays (dermotrichia) which are for the most part jointed and flexible (sometimes they branch peripherally), but in some cases (Acanthopterygian fishes) more or fewer of the anterior der- motrichia in the dorsal (anterior dorsal if there are two), anal, and pelvic fins are unjoin ted and generally stiffened. Such dermotrichia are called spines. The caudal dermotrichia are always jointed, as are all the dermotrichia in the Malacoptery- gian fishes. In some Malacopterygian fishes the posterior dorsal fin is without jointed dermotrichia, but contains adi- pose tissue only and delicate unjointed horny dermotrichia resembling the embryonic dermotrichia ( actinotrichia) : this is the adipose fin of certain Siluridae, Salmonidae, etc. The dermotrichia of the ventral part of the caudal fin are carried by the haemal arches, those of the other unpaired fins by special skeletal rods — the somactids (p. 54) — lying in the median fibrous septum separating the dorsal parts of the muscles. These somactids of the Teleostean unpaired fin are often called inter spinous bones. 186 SUB-CLASS (AND ORDER) TELEOSTEI. The spines of the dorsal fin of Acanthopterygians can be raised or de- pressed at will. In the depressed position the spines may cover one another completely (homacanth), or they may be turned slightly to one side or the other alternately (heteracanth). The skin usually contains pigment and may be very brightly coloured. Many fishes possess the power of changing their colour in a protective manner according to their surroundings, and in almost all the dorsal surface is darker in colour than the ventral. The body is usually covered with scales, overlapping one another in such a way that the posterior part of the scale is free and covers the anterior part of the next scale behind. The scales are thin plates of bone imbedded in the dermis, and are more frequently absent from the head and fins than elsewhere. They are absent in most eels and in fishes with electric organs. The epidermis is soft and contains many mucous and pigment cells. The scales have a concentric striation and are of two kinds, cycloid and ctenoid. In ctenoid scales the posterior free margin possesses denti- culations which may extend on to the surface, whereas in cycloid scales such denticulations are absent. In some cases the scales are enlarged into great scutes (Siluridae, Lophobranchii, Plectognathi) and in Balis- tidae they have the form of ossified papillae which project in a shagreen- like manner. In Siluridae the scales may carry movably articulated dermal teeth. The scales of most Teleostean fishes are thin calcified lamellae, without bone corpuscles, lying in a dermal sac. They are surrounded by sclero- blasts which, as in Lepidosteus, form them and add to them during growth. The outer portion of the scales is structureless, the inner contains fibres. In some forms, scleroblasts of the upper surface are included and become bone corpuscles, so that the scales consist of an outer part showing ordin- ary bone structure and an inner consisting of connective tissue impreg- nated with calcareous matter by the scleroblasts. The spines which are present on the scales of some Teleostei, e.g. some Acanthopterygians and Plectognaths, frequently contain a cavity which suggests a pulp cavity. They may possibly be regarded as homologous with the spines of Elasmobranchs, but without the enamel cap. The spines of the hinder part of the ctenoid scales are due to the sculpture of the surface and are not denticles. According to the recent and as yet (March 1904) unpublished researches of Marett Tims, the scales of Gadus consist of a number of small plates of structureless bone (vitrodentine) lying close together on a fibrous basis. In the young state each plate carries a small spine which does not reach the epidermis. The sense organs * (nerve eminences) of the lateral line are * F. Leydig, Integument u. Hautsinnesorgane der Knochenfische, ZooL Jahrb. Anat., 8, 1894, p. 1-152. W. Collinge, Sensory Canal System of Teleostei, Proc. ZooL Soc., 1895, pp. 274-299. R. McDonnell, Lateral line in Fishes, Trans. P. Irish Acad., 129, 1862, pp. 161-187. PHOSPHORESCENT ORGAN'S. 187 usually enclosed in canals, but they may be unenclosed and only protected by flaps of skin (Batrachus, Lophius, etc). Their distribution, whether they are enclosed or not, resembles that found in Elasmobranchs and is indicated on Fig. 42. When they are enclosed in canals, the tubes which leave the canal and open on the surface are simple, ending in a single external pore ; or they branch considerably before reaching the surface and open by several pores. The trunk part of the lateral line canal is placed in the dermis and is usually without osseous supports, but on the head the canals are either enclosed in small bones which lie outside the skull bones, or they burrow in and are protected by the bones of the skull and visceral arches themselves. There may be accessory lateral lines in the tru?ik, placed on the sides of the body near the dorsal and ventral middle lines. There are no ampullary canals. Many deep-sea fishes possess numerous shining bodies in the skin resembling in their general features eyes.* It appears probable that they are phosphorescent organs. They are found either on the head near the eyes, on the lower jaw, at the end of barbels, under the gill-cover, or in rows in which they may be segmentally arranged along the sides of the body, and sometimes in connection with the lateral line. They vary considerably in structure from being simply glandular patches of the skin which are supposed to secrete a phosphorescent mucus to a state in which they are more eyelike in appearance and possess a lens-like body which, it is suggested, acts like the lens of a bull's-eye lantern in concentrating the rays proceeding from the internal parts of the organ. In the latter case a kind of tapetum can often be made out at the back of the organ which appears to act as a reflector. They are probably in all cases modified skin glands. It appears probable that in many cases these organs are for the purpose of enabling their possessors, which are generally provided with large eyes, to see in the dark abysses of the ocean, but in some cases no doubt they act as lures (when placed at the end of barbels or far back on the body). In the case of Ipnops in which * F. Leydig, Die augendhnlichen Organe der Fische. Bonn, 1881. M. Ussow, Ueber d. Bau der sogennanten augenahnlichen Flecken etc., Bull. Soc. Nat. de Moscou, 1879. A. Gunther, Deep-Sea Fishes, in Chal- lenger Reports, 22, 1887, and appendices by H. N. Moseley and R. v. Lendenfeld. 188 SUB-CLASS (AND ORDER) TELEOSTEI. there is no trace of eyes, optic nerve or olfactory nerve, and in which the supposed luminous organs have the form of two broad laminae on the upper surface of the head, and in other deep sea forms in which the eyes are imperfect (e.g. the Pediculati) they can only be of use as lures. In the endoskeleton the primitive cartilage is largely replaced by bone, but some cartilage, varying in amount in the different forms, may persist. The vertebral column is usually completely ossified and consists of amphicoelous vertebrae (Fig. 34). The vertebrae are con- nected by articulating processes placed on the neural arches. In the trunk the centra carry transverse processes, which are directed outwards, and ribs which are articulated to the centra or to the base of the transverse processes (Fig. 34). In the tail the centra carry complete haemal arches, which enclose a canal containing the caudal artery and vein and are prolonged like the neural arches into a median spine. In some forms a pair of small bony rods — the inter-muscular bones, are attached to the centra near the neural arches. The vertebrae are arcicentrous, the notochordal sheath remaining thin, but the skeletogenous tissue develops very little cartilage being rapidly replaced by membrane bone in the centra as well as in the arches. In most Teleosteans the end of the vertebral column is bent dorsalwards, and is unsegmented, though the notpchordal sheath is ossified to form a bony urostyle. The haemal arches of this part of the vertebral column persist in a modified form as the hypural bones, which carry the dermotrichia of the ven- tral part of the caudal fin (p. 55). In such cases the tail though symmetrical externally, is internally asymmetrical, and is said to be homocercal (see pp. 55, 56). In a few forms (Gadidae, etc.), the end of the vertebral column is not bent dorsalwards, and the tail-fin is symmetrical internally as well as externally (diphycercal). In these fishes the dermotrichia of the ventral part of the caudal fin are carried by interspinous bones, and it seems highly probable that the true tail-fin has atrophied completely, as it has in some Heteromi, Syngnathidae, etc., in which the tail tapers to a point and is without any trace of a caudal fin, and has been secondarily replaced by a backward extension of the dorsal and anal fins (p. 55). Such tails are therefore secondarily diphycercal. SKULL. 189 The skull is always hyostylic and possesses both mem- brane and cartilage bones. It differs considerably in the extent to which the primitive cartilage persists. In many forms, e.g. the salmon, pike (particularly in the less specialised, more ganoid-like fishes), a considerable amount of cartilage persists and the cartilage bones are separated Ztkl Ethi Sse FIG. 110. — Cephalic skeleton of Perca fluviatilis (Kegne animal). Ac post-clavicles ; '-A alisphenoid ; An angular ; Ar articular ; Brs branchio-stegal rays ; Cl clavicle ; Cvr cora- coid ; D dentary ; Ekp ectopterygoid ; Enp entopterygoid ; Ethi median ethmoid :,} EtM lateral ethmoid (prefrontal) ; Fr frontal ; Frp posti'rontal (sphenotic) ; Hm hyomandibular ; Htf hyoid arch ; Jm premaxilla ; JOp interpperculum ; Mt-p metapterygoid ; MX maxilla ; Oex epiotic ; Op operculum ; Os supraoccipital ; Pal palatine ; Par parietal ; POp pre- operculum ; PrO prootic ; Ps parasphenoid ; Q quadrate ; S symplectic ; Sc scapula ; SOp suboperculum ; Sq pterotic ; Ssc supraclavicle ; Vo vorner. by wide tracts of intervening cartilage. In others, e.g. the cod, the cartilage is almost entirely ossified. The cartilage is usually deficient in the roof of the skull except in the occipital region, in which a basi-, two ex- and a supra-occipital are developed. The auditory region usually presents five separate cartilage bones, the epiotic, opisthotic, prootic, the pterotic and the sphenotic (postfrontal). The sphenoid region is feebly ossified : there is always a small basisphenoid and sometimes an alisphenoid 190 SUB-CLASS (AND ORDER) TELEOSTEI. and orbit osphenoid, and the anterior part of it generally acquires a considerable vertical extension forming an interorbital septum (absent in Siluroids, Cyprinoids, etc.). The ethmoid region remains unossified, or at most has two bones — the lateral eth- moids or prefrontals (Fig. 110 EM}. The membrane bones of the roof are parietals (Par], large f rentals (Fr) and a bone over the ethmoid (supraethmoidal or median ethmoid, Ethi). There may be other bones, called nasals, over the ethmoidal region. The parietals may touch in the middle line between the f rentals and supraoccipital, or be pushed apart and separated by the junction of the f rentals and the supraoccipital. In the floor there is a large and important parasphenoid strengthening the base of the skull and a vomer (Vo) underlying the ethmoid region. The orbit is surrounded by a ring of circumorbital membrane bones (not shown in Fig, 110), of which the anterior is called the lacrymal. Premaxillae (Jm) and maxillae (Mx) are present, and there may be a jugal, but the maxillae are usually toothless and frequently take no part in the formation of the edge of the mouth. The palatine bar of the mandibular arch always presents osseous palatine, pterygoid and quadrate elements : in front is the palatine (Pal) often dentigerous ; then follows the pterygoid region usually presenting three elements — the pterygoid (ectoptery- goid), the mesopterygoid (entopterygoid Enp] and the meta- pterygoid (Mtp) ; lastly comes the quadrate (Q) which gives articulation to the lower jaw. The front (palatine) end of this bar is attached to the ethmoid region, while the quadrate is not attached to the cranium directly, but is supported by the strong dorsal element (hyomandibular) of the hyoid arch. In the lower jaws Meckel's cartilage persists, being ossified proxi- mally as the articular (Ar), and ensheathed by the dentary bone (D) distally : in addition there is often an angular element (An). In the hyoid arch there is a powerful dorsal hyomandi- bular elemen,t which presents two bones, the hyomandibular (Hm) and the symplectic ($). The hyomandibular bone articulates with the auditory region of the cranium and passes ventralwards behind the metapterygoid, while the symplectic lies distally and is closely applied to the quadrate. The rest of the hyoid arch consists of three ossified pieces on each side (Fig. Ill) — the epihyal (c), ceratohyal (6) and hypohyal. The VISCERAL SKELETOX. 191 epihyal is joined to the cartilage interval in the hyomandibular element between the hyomandibular and the symplectic bones by a small osseous piece the interhyal (d), while ventrally the hypohyal joins the large median, sometimes toothed glossohyal. In connection with the hyoid arch is a number of membrane bones — the opercular bones supporting the opercu- lum, and the branchiostegal rays in the branchiostegal mem- brane. The opercular bones are in connection with the dorsal, hyomandibular element and consist of the operculum (Fig. 110, Op) and preoperculum (POp), and sometimes of a suboperculum FIG. 111. Hyoid apparatus and branchial arches of Perca ftunatUis (R£gne animal). a, b c, d segments of the branchial arches ; the upper joints Ops are the superior pharyngeal bones (pharyngobranchials) ; VI, Opi the inferior pharyngeal bones (reduced fifth branchials) ; Cop median pieces (copulae) ; Kb branchiostegal rays; I = Zbg hyoid arch ; // — V branchial arches. (SOp) and interoperculum (JOp). The branchiostegal rays (Brs) are attached to the lower part of the hyoid arch, partly to the inner and partly to the outer side (Fig. Ill Rb). There are five pairs of branchial arches. Of these the anterior four are usually segmented into four pieces (Fig 111), the pharyngo- (Ops=d), epi-(c), cerato-(6) andhypo-(a) branchials. More or fewer of the pharyngo-branchials, which are not joined to the skull or vertebral column, are united with one another to form the so-called superior pharyngeal bone which generally bears 192 SUB-CLASS (AND ORDER) TELEOSTEI, teeth (Ops). The hypobranchials, which may be wanting in the fourth arch, are attached to a varying number of median elements, the copulae or basibranchials (Cop). The fifth bran- chial arch is reduced to a single rod on each side which is usually strongly toothed, and the pair are called the inferior pharyngeal bones (Opi) ; they are sometimes ankylosed to form a single bone. The four anterior branchial arches bear small tooth-like projections, in one or two rows, which act as strainers ; these are the gill-rakers. Pectoral and pelvic* paired fins are present, but one or both of them may be absent. In the pectoral girdle, which is usually present even when the fin is absent, the primitive cartilaginous c o r a c o- scapular elements are but slightly developed and relatively > unim- portant, while the membrane bones (clavi- cles t) are largely de- veloped. Flo. 112. — Bight pectoral girdle and fin off Gadus (after Gegenbaur). c clavicle (cleithrum) ; b supraclavicle (supracleithrum) ; a post-temporal ; d post-clavicle ; / scapula ; e coracoid ; g basal somactids of the fin ; h bony dermotrichia. The coraco-scapular arches do not join ventrally and are attached to the inner sides of the clavicles. They present two bony elements — the scapula and coracoid (by some regarded as precoracoid) with persistent intervening cartilage. The scapula usually has a foramen, and there is sometimes a third bony element placed dorsal to the coracoid and in front of the scapula and called the mesocoracoid. The clavicle (cleithrum) is a large membrane bone meeting its fellow ventrally under the throat. To its dorsal end there is usually attached a smaller supraclavicle (supracleithrum) which is connected dorsally with a forked bone the post-temporal. This bone is attached to the auditory region of the skull, by one prong to the epiotic and by the other to the pterotic bone. Projecting back from the * The pelvic paired fins are usually called ventrals. •j- Called cleithra by some anatomists, on the view that they are not homologous with the clavicles of higher Vertebrates (see notes, p. 162). BRAIN. 193 upper end of the clavicle is a bony rod, the post-clavicle. There is no infraclavicle. The skeleton of the fin consists of usually five basal ossified somactids which are articulated with the corac o-scapula, and of a row of small cartilaginous pieces representing distal somactids. These are followed by the dermotrichia, the anterior of which is continuous with the an- terior of the basal somactids. The pelvic girdle is always absent, its place being taken by a large osseous basal somactid, commonly called the basiptery- gium ; to this are attached a few small, partly bony, distal somactids, which carry the dermotrichia. The brain * of Teleosteans presents the following features. FIG. 113. — Median longitudinal vertical section through the biain of the trout (from Gegen- haur, after Rab-Riickhard). Aq aqueductus sylyii ; Bo olfactory lobe ; Cbl cere- bellum ; Cc central canal of spinal cord ; Ccn anterior commissure ; Cho. optic nerves ; Ci inferior commissure ; Glp pineal body ; Hy, Hy' hypophysis ; J infundibulum ; Not olfactory nerve ; Pa pallium ; pf velum transversum ; Sv saccus vasculosus ; Too pia mater on the dorsal side of the mid-brain ; Tl dorsal wall of mid-brain between the two optic lobes ; tr crossing of the fibres of the fourth nerve ; Vc valvula cerebelli ; Vcm ven- tricle of the cerebrum ; Vq fourth ventricle ; Vt third ventricle. (1) The olfactory lobes are usually much elongated and slender ; they are swollen at their free ends against the nasal capsules and at their origin. The cerebral ventricle is continued into the swollen base.f * Rabl-Ruckhard, Das Grosshirn der Knochenfische, etc., Arch. f. Anat. und Phys., Anat. Abt., 1883, p. 279-322. E. Baudelot, Rtcherches sur le sy steme nerveux des Poissons, Paris, 1883. A. Schafer, Die morphol. u. hist. Entwick. d. Kleinhirns der Teleostei, Morph. Jahrb., 21. f There seems to be some difference of opinion as to whether these basal swellings alone constitute the olfactory lobes, the slender prolongations being only olfactory nerves. It has been suggested that in some fishes, e.g. Salmon, this is the case, whereas in others, e.g. Cyprinoids, the whole elongated structure is also part of the olfactory lobe. Z — II O 194 SUB-CLASS (AND ORDER) TELEOSTEI. (2) The cerebrum is not clearly marked off from the thalamen- cephalon, and its roof is entirely composed of a thin gallium. The latero-ventral parts are thickened into the great corpora striata, which were formerly taken for the cerebral hemispheres themselves. The ventricle of the cerebrum is entirely un- divided, and the pallium or dorsal wall of it is not marked by a groove (Fig. 114). (3) The thalamencephalon is very inconspicuous and the optic thalami are hardly if at all developed. The anterior part of the thin roof is folded inwards in the usual way (Fig. 113 pf.). The pineal body lies in the skull over the pallium ; it has folded walls and appears to open by a narrow pore at its point of attachment to the roof just in front of the posterior commissure. There is no parietal organ. The floor presents the usual structure ; in front is the thickening caused by the optic nerves, which sim- ply cross after leaving the brain and do not form a chiasma. The infuridibulum possesses lateral lobi inferiores, and is provided posteriorly with a glandular sac, the saccus vas- culosus, opening into it by a minute pore. The pituitary body is solid and is attached to the infundibulum in front of the saccus vasculosus. (4) The mid-brain presents the two optic lobes (corpora bigemina) dorsally and contains a projection formed by the wall of the brain at the junction of the optic lobes and cerebellum. This is the valvula cerebelli (Fc), or fornix of Gottsche. (5) The hind-brain is in the usual form, the cerebellum being large and containing a prolongation of the fourth ventricle. It projects back over the medulla oblongata. In some Teleosteans (e.g. Gymnarchus, Mormyridae) the brain attains a very large size, the cerebellum and sometimes the optic lobes being especially well developed. The spinal cord frequently ends in an oval or spherical swell- ing. In some forms (Plectognathi) it is much reduced in length ; FlO. 114.— Transverse section through the cerebrum of the trout, through the line xx in Fig. 113. Vcm and Vt Ventricle of the cerebrum ; Cst corpus striatum ; Sm pallium ; Gp pineal body ; PI choroid plexus (after Rabl-Ruckhard from Gegenbaur) . NERVES. SENSE ORGANS. 195 e.g. in Orthagoriscus mola it is barely as long as the brain and hardly reaches beyond the skull. In such cases it ends in a slender filum terminale. The central canal usually contains a fibre * (Reissner's fibre), which extends from the anterior end of the optic lobes, with which it is continuous, backwards along the whole length of the central canal. It consists of a bundle of nerve fibres, and communicates with the tissues of the spinal cord throughout its course. It appears to be absent in blind fishes. In some Teleosteans (Ctenoldbrus, Pleuronectes, etc.) giant nerve cells f are found in the posterior fissure of the spinal cord, their neurites passing into the substance of the cord. The cranial nerves i resemble in their general arrangement those of other fishes. The ramus ophthalmicus profundus if present is much reduced. There is a dorsal branch of the fifth nerve — the ramus lateralis accessorius or r. recurrens trigemini — which receiving branches from the facial and vagus passes dorsalwards in the cranium to perforate the frontal bone. It then travels backwards near the skin to supply the skin of the trunk near the dorsal fin (sense-buds and pit-organs). It appears to be composed partly of so-called communis (afferent- visceral) fibres (ramus lateralis accessorius), and partly FioT 115. — Hori- of lateralis fibres supplying pit-organs and derived from through the^eye the facial. of Esox Indus (from Claus). Co cornea ; L lens ; Sense organs. The olfactorv organ is Pf rrocessus fai- ciformis ; CH usuallv provided with two openings. campanula Hai- leri ; No optic The eve is distinguished bv the possession nerve; se ossm- . . , cations in the of a flat cornea and a sclerotic which is sclerotic, frequently more or less ossified. The lens is closely approximated to the cornea, the anterior chamber of the eye being small. Traversing the vitreous humour, some- what on the lower side of the eyeball, and extending from the entrance of the optic nerve to the eye, is a band of tissue (a process of the choroid coat) containing blood vessels and smooth muscular fibres ; this is the processus falciformis. At its point of attachment to the lens it is swollen into the so-called cam- panula halleri. One of the functions of this structure is said to be that it assists in accommodation for vision of distant objects, * Sargant, Anat. Anzeiger, 17, 1900, p. 33. A similar fibre has been described in Petromyzon. t Sargant, Anat. Anzeiger, 15, 1898, p. 212. J Stannius, op. cit. Cole, Gadus, Trans. Lin. Soc., (2), 7. Herrick, Menidia, Journ. Comp. N enrol., 9, 1899. 196 SUB-CLASS (AND ORDER) TELEOSTEI. by drawing the lens nearer to the retina. The eyes of fishes when at rest are accommodated for vision of near objects, i.e. the opposite of the condition in the eyes of the terrestrial Vertebrates. There is a layer of tissue between the choroid and sclerotic which contains crystals, the argentea ; and round the entrance of the optic nerve between the same coats there is in many Teleosts (those with a pseudobranch) a vascular plexus of unknown function, called the choroid gland. ass cp The eye-muscles are the usual four recti and two obliqui ; the former often arise from a subcranial bony canal the floor of which is formed by the parasphenoid. Movable eyelids are not present, though there may be a circular fold of skin round the eye. In Anableps the cornea is crossed by a hori- zontal stripe which divides the pupil, so that there appear to be two pupils one above the other. The auditory organ * consists of the membra- FlG. 116.— Membranous labyrinth of Perca fluvia- noUS labyrinth and is tilis, inner view (from Wiedersheim). aa anterior .., ampulla; ac auditory nerve; ae external, ap WltllOUt accessory StrUC- posterior ampulla ; ass apex sinus superioris ; ca anterior, ce horizontal, cp posterior semi- circular canal ; de ductus cndolyni.phaticus ; I lagena cochleaeti mn macula acustica neglecta ; ms macula acusca sacculi ; mu macula ac. re- cessus utriculi ; o otoliths of the recessus utriculi, the saccule and the lagena ; pi papilla acustica lapenae ; raa, rap, rl, rs, nerves to the am- pullae of the anterior, and posterior semi- circular canals, to the lagena, and to the saccule ; rec recessus utriculi ; s saccule ; ss sinus utriculi superior ; u utricle. tures except in those forms (Ostariophysi) in which a chain of small bones connects it with the air bladder (see p. 202). The membra- nous labyrinth is con- tained in the auditory region of the skull wall, but the cavity in which it is placed is not shut off from the cranal cavity by bone or cartilage. It is constructed on the usual plan, con- sisting of a central chamber or vestibule and three semicircular canals (Fig. 116). The vestibule is divided by a constriction into two parts, an upper, the utricle (u) into which the semi- * G. Retzius, Das Gehororgan der Wirbelthiere, Bd. 1., Stockholm, 1881. ALIMENTARY CANAL. 197 circular canals open, and a lower the saccule (s), which possesses a small posteriorly directed process (lagena) representing the cochlea. The otoliths vary much in form in different fishes. They occur in the utricle, saccule and lagena. That in the saccule (the sagitta) is generally the largest and of a crystalline structure. There is another small one (asteriscus) in the lagena, and a third in the utricle close to the ampullae of the anterior and horizontal canals (o). The ductus endolymphaticus is present as a process of the saccule, but does not open externally. The lateral line has already been described (p. 187). Its sense organs are probably innervated by branches of the facial and lateral line branch of the vagus as in other fishes, but this has not been shown in all cases. The alimentary canal is distinguished by the very general presence of an air bladder, which must be regarded as an appen- dage of the oesophagus, though it is often in the adult separate from this ; by the presence of more or less numerous appendages — the pyloric caeca, opening into the first part of the intestine ; by the inconspicuous character of the pancreas, which in some cases is even said to be absent ; by the absence of a spiral valve in the intestine ; and by the absence of a cloaca common to the alimentary canal and urinogenital organs. The teeth * are usually well -developed, but in some cases are al- together absent (some Lophobranchii, Coregonus). They may be borne by the maxillary, premaxillary, palatine, vomer, dentary bones, by the glosso-hyal and by the branchial arches, and rarely by the pterygoid and parasphenoid. The maxilla is usually with- out teeth, and does not always form part of the edge of the mouth. The teeth are generally ankylosed to the subjacent bony structures, but in some cases there is a ligamentous connection of such a kind that they can be bent inwards when food is being swallowed, but not in the reverse direction (some Gadidae, Lophius, Esox). In a few cases they are implanted in sockets (Sphyraena, etc.). As a rule teeth continue to be developed throughout life from new germs (not from pre-existing germs), placed behind the functional teeth. These come into function and position as the old teeth are worn down and cast off. When the teeth are implanted in sockets, the new tooth, though * R. Owen, Odontography, London, 1840-15. C. S. Tomes, Dental Anatomy, London, 1898. 198 SUB-CLASS (AND ORDER) TELEOSTEI. developed behind the old one, comes to lie beneath it, so that the succession is vertical. The teeth are generally conical, and may be minute, slender and sharp- pointed (villiform, e.g. Perch), or longer but very fine (ciliiform, setiform, as in Chaetodonts). Larger conical teeth are termed card-like (rasp-teeth, raduliform). In Goniodonts the teeth are bent on themselves like a tenterhook. They may vary in shape in different parts of the mouth, the anterior teeth being conical, and the posterior broad and molar-like for grinding the food, as in the wolf -fish (Fig. 117), and many Sparidae. In Sargus indeed the anterior teeth are incisor-like, and in Dentex there are canine-like teeth. Small molar-like teeth are called granular, In Labrus crushing teeth are borne by the upper and lower pharyngeal bones. Compound teeth, which are found in the Gymnodonts and the Scari,* are made up of a number of teeth which are developed successively but are joined by cement when full grown and functional. They thus present the appearance of teeth which continue to grow throughout life. Pharyngeal teeth may be present on the superior and inferior pharyn- geal bones. In the Cyprinoids the mouth is edentulous, and teeth are only found on the inferior pharyngeal bones, which bite against a tubercle on the basi-occipital. They are also present on the edges of the branchial arches, but except in a few cases, e.g. Orthagoriscus, in which they are long and sharp, these gill- teeth are little more than horny excrescences, which however are sometimes elon- gated into setiform horny processes — the gill-rakers. FIG. 117.— Teeth of the Wolf-flsh, Anarrhichas lupux (after Gunther). There is usually on the floor of the mouth a small non-muscular elevation which represents the tongue ; it is supported by the glosso-hyal, and is sometimes toothed. The oesophagus is a wide tube, hardly if at all marked off from the stomach. The stomach, which varies in form, is usually but slightly dilated and is either U-shaped as in Elasmobranchs, or is provided with a caecal prolongation of its cardiac portion and a short pyloric region placed near the junction with the oesophagus (Fig. 37). The intestine is usually slightly convoluted, is without a spiral valve (though a trace of one may be made out in some genera), and opens to the exterior by the anus. In some forms (e.g. Tinea, Cobitis] the striped muscles of the oesophagus are con- tinued over the stomach and intestine outside the smooth * J. E. Boas, Die Zahne der Scaroiden, Z. /. w. Z., 32, 1879, p. [189-215. RESPIRATORY ORGANS. 199 muscles. The pyloric caeca are tubular structures opening into the first part of the intestine just beyond the pylorus ; they vary in number from one to two hundred, and are very generally present. The liver is provided with a gall bladder which opens just beyond the pylorus. It generally contains much oily matter, but in some forms the oil occurs in all the tissues of the body and is not only a feature of the liver. The pancreas,* the presence of which in Teleostei used to be denied, is not usually a conspicuous structure, though functionally of great importance, especially in those Teleosts in which the stomach is without gastric glands. It is either embedded in the liver or diffused in the mesentery, and its duct opens in close connection with the hepatic duct. In some forms (Salmo, Gadus, Perca, Platessa, Brama, etc.) there is a small gland opening with the bile duct, which is possibly pancreatic. The Thyroid t body is represented by small reddish masses lying ventral to the ventral aorta, and the thymus f is placed at the dorsal ends of the last pair of branchial arches. The respiratory organs. — There is no spiracle. The branchial apertures are narrow slits and the tissue between them has the form of an arch, not of a sep- tum. In consequence of this the gills themselves are filamentous, not lamellar (Fig. 118). The external openings of the clefts are covered by the operculum. The gill-filaments are borne in a double row (holobranch) by the four anterior branchial arches, the last gill-aperture is smaller than the others, and the fifth branchial arch never bears gill-filaments. E. Laguesse, Structure, etc., du pancreas d'apres les travaux recents, Journ. Anat. Phys., Paris, 30, 1894, p. 591 and 731. E. Goeppert, Die Entwick. d. Pancreas Teleost., Morph. Jahrb., 20, 1893. f F. Maurer, Schilddriise u. Thymus in Teleostier, Morph. Jahrb., xi., 1886, p, 128-75. FIG. 118. — Transverse sec- tion through a branchial arch of Zygaena (right hand) and of Gadus (left) (after B. Hertwig, from Wiedersheim). a and v afferent and efferent blood vessels, b skeletal branchial arch, bll and fro^posterior and anterior demibranch together con- stituting a holobranch, h septum between two branchial apertures in Zygaena, r cartilaginous branchial ray supporting the same, z small tooth- like tubercle (sometimes elongated as a gill-raker) in a double row on the branchial arch of Gadus. 200 SUB-CLASS (AXD ORDER) TELEOSTEI. A pseudobranch * is generally present on the posterior side of the hyoid arch. It contains a rete mirabile and usually has the form of short filaments or ridges. In some cases it is con- cealed below the mucous membrane, and the organ may have the form of a red, lobed swelling (so-called glandular pseudo- branch or vaso-ganglion). Sometimes it lies so far from the surface that it is quite hidden : indeed it may be covered by fat and muscles and even by bone. It is sometimes absent, and its absence appears to be very generally correlated with that of the choroid gland. The function of the pseudobranch is unknown ; it lies in the course of the greater part of the blood supply to the eye (see below), and it is generally regarded as a vestige of a hyoid gill. Certain fishes, e.g. eels, can exist for some time out of water, but those with large gill-apertures usually perish rapidly. In some active fishes, e.g. the Scombridae, the temperature of the blood is considerably higher than that of the water ; probably it is always slightly higher, but we know very little on this subject. The Teleosts have usually five gill-clefts, but the fifth is always smaller than the rest and is sometimes absent. In this case the fourth branchial arch bears one row of filaments only (demibranch) or may be gill-less. In some forms the gill-apparatus, both arches and gills, may be still more reduced (Symbranchidae, Malthe, etc. ; in Amphipnous cuchia the second branchial arch alone bears gill-filaments). In some Lophobranchii the gills have the form of curious tufted processes. Accessory respiratory structures are met with, especially in cases in which the gill-filaments are reduced. Thus in Amphipnous there is a lung-like vascular-lined sac, opening into the first gill-cleft, for breathing air. In Saccobranchus there is a very similar sac. In Anabas scandens, in which the gill-apparatus is complete, the superior pharyngeal bones are honeycombed so as to form a laminated organ covered with vascular mucous membrane to enable it to breathe out of water (Fig. 38). Accessory respiratory organs are also found in the Ophiocephalidae, certain Siluridae (Clarias, Heterobranchus, Heterotis], and in Chanos. The air-bladder is present in most but not in all Teleostei. It presents great variety of structure throughout the group. In the Malacopterygii, Ostariophysi, Apodes, and Haplomi it usually, but not always, opens into the alimentary canal (usually into the oesophagus) on its dorsal side by the pneumatic duct (laterally in JErythrinus), which may however be partly or * Joh. Miiller, Vergl. Anat. der Myxinoiden, loc. cit. Maurer, Morph. Jahrbuch, 9, 1883, p. 229. AIR-BLADDER. 201 entirely occluded. In other Teleostei there is rarely * a ductus pneumaticus in the adult, though such may be present in the young form ; and the air-bladder is a closed sac. Inasmuch as it develops in the embryo as a diverticulum of the oesophagus, this closed condition must be regarded as secondary. In the Clupeidae the ductus pneumaticus opens into the fundus of the stomach and in the Herring there is a second duct opening to the exterior on the left side of the reproductive aperture. The air-bladder always contains gas which consists of nitrogen, oxygen and a trace of carbonic acid. It lies dorsal to the ali- mentary canal, and is usually closely adherent to the ventral surface of the kidneys, lying between those organs and the peritoneum, and in many Siluroids it is partly enclosed in osseous capsules formed by the vertebrae. In some cases however it projects into the body cavity, lying more or less loosely. Its walls consist of connective and elastic tissue and yield isinglass. Tufts of blood vessels in the form of retia mirabilia (red bodies) covered by a glandular epithelium are often present on its walls, and sometimes project into it, like huge vascular glomeruli. These vascular bodies are absent from the Ostariophysi and from most fishes which have a pneumatic duct. They are however present in some of the latter, e.g. the eel. The air-bladder may be coextensive in length with the body-cavity, but it frequently extends as a single or double prolongation some distance beyond into the caudal region beneath the caudal vertebrae (Gymnotus, Ophiocephalus, etc.), or forwards into (see below) or towards the head (Gadus, etc.). Sometimes it is much restricted (some Siluridae, Pediculati, Plectognathi, etc.). In some cases it is partially or completely divided transversely into two or even more compartments (Cyprinidae, Characinidae) ; more rarely it is divided longi- tudinally (Arius}. In some forms it is so much reduced in size that it almost escapes notice (some Siluridae and Cyprinidae). It frequently gives off diverticula, which in the Sciaenidae and Polynemidae are numerous and branched. As a general rule the internal cavity is unbroken except for the partitions complete or * An open pneumatic duct is said to be present in Holocentrum, Pria- canthus, Caesio, etc. See Kner, Einiges iib. die Thymus bei Fischen u. d. Schwimmblase der Stachelflosser, Sitz. Mat. Nat. Classe Akad. Wiss., Wien, 49, 1864, p. 455. 202 SUB-CLASS (AND ORDER) TELEOSTEI. incomplete just referred to, and the lining may be smooth or cellular (Clupeidae, etc.), but sometimes it is much broken up and has spongy, lung-like structure (Heterotis, Gymnarchus, and other forms). In the Ostariophysi the anterior end of the air-bladder is connected by a chain of small bones, which are probably de- tached portions of some of the anterior vertebrae and are called the Weberian ossicles,* with the wall of a chamber in the skull wall enclosing a diverticulum of the membranous labyrinth ; and in other cases anterior prolongations of the air- bladder reach the skull and come into immediate contact with the wall of the space in which the membranous labyrinth is contained. In the simplest cases (Myripristis, Holocentrum, Sparus, Sargus, etc.) the two anterior horns of the air-bladder apply themselves to membranous spaces in the bony wall of the occipital region containing the membranous labyrinth. In many Clupeidae the slender anterior end of the air-bladder enters a canal in the basi-occipital and divides into two narrow branches. Each of these dilates within the bone and divides again into two, each of which forms a spherical swelling. A process from the vestibule (utricle) of the membranous labyrinth comes into contact with these vesicles ; moreover, the vestibules of the two sides are connected by a transverse canal. In the Ostariophysi, in which the connection is effected by a chain of ossicles, a few of the anterior vertebrae are ankylosed together and modified in certain of their parts, some of which are partially detached to form the auditory or Weberian ossicles. These are four in number, of which three, the tripus (malleus, Fig. 119, 11), intercalarium (incus, 9) and scaphium (stapes 8) form a chain connecting the air-bladder with the membranous labyrinth. In addition there is a fourth — the claus- trum — partly dorsal to and partly in front of the scaphium, which, how- ever does not form part of the chain, but simply lies in the wall of the atrium (see below). The first vertebra is represented by the centrum only which is distinct from but firmly connected to the skull and the next centrum. The second centrum (10) although it shows no marked sign of being composite, con- sists of the completely fused centra of vertebrae -2, 3 and 4-. It is therefore called the complex centrum. To the hind end of the complex centrum the three next centra may be united, but these remain distinguish- able. The saccule (15) of the membranous labyrinth gives off a process, the ductus endolymphaticus (4), which unites with its fellow in the middle line, and gives off posteriorly from the point of union a median sac, the saccus endolymphaticus (5). The saccule and saccus lie in excavations of the basioccipital bone, called respectively the foveae sacculi (10) and the cavum sinus imparis (14). These are partly separated from one * E. H. Weber, De aureetauditu hominis etanimalium,Fars I.,De aure animalium aquatilum, Leipzig, 1820. T. W. Bridge and A. C. Haddon, The air-bladder and Weberian ossicles in the Siluroid Fishes, Phil. Trans., 184, 1893, p. 65-333. AIR-BLADDER. 203 another and from the cranial cavity by bony plates, but they open into the cranial cavity in front. The foveae sacculi (16) end blindly behind, but the cavum sinus imparis opens behind into two laterally-placed chambers the floor of which is formed by the basioccipital, the sides and roof by thick fibrous walls. These chambers are the atria sinus imparis (13). To the outer wall cf these is attached a process of the scaphium (S), the anterior of the three ossicles. Of these, the posterior or tripus (11) is by far the largest, and is inserted behind into the fibres of the anterior chamber of the air-bladder ; laterally it has a process articulating with the complex centrum, and anteriorly it is connected by a strong ligament — the interossicular ligament (12) — to the scaphium. In this ligament and between the tripus and scaph- ium, is the inter- calarium (8). The tripus and inter- calarium are partly enclosed in a thin- walled fibrous sac, containing a deli- cate fibrous net- work and called the saccus para- vertebralis. The air-bladder is divided into a n anterior and pos- terior chamber, of which the anterior is usually especially distensible, and in the Siluridae comes into close contact with the skin on each side. The pos- terior division is generally divided into two by a longitudinal septum which frequently gives off incomplete transverse septa, into the anterior chamber. FIG. 119. — A view from above of the cranial floor and anterior vertebrae of Macrones ntmurus, semidiagrammatic (after Bridge and Haddon). The brain has been removed, and the bone cut awav, so as to expose more completely the mem- branous labyrinth. 1 sphenotic, 2 prootic, 3 pterotic, 4 ductus endolymphaticus, 5 saccus endolymphaticus, 6 epiotic, 7 exoccipital, 8 scaphium, 9 intercalarium, 10 complex centrum, 11 tripus, 12 interossicular ligament, 13 atrium sinus imparis, 14 cavum sinus imparis, 15 sacculus, 16 fovea sacculi, 17 ductus sacculo-utricularis, 18 utricle. The ductus pneumaticus opens Striped muscles are frequently present in the wall of the air- bladder (species of Trigla, Batrachus, Pogonias, Zeus, and others). Sometimes there are extrinsic muscles passing from the ventral surface of the vertebral column on to the air-bladder (species of Gadus, Diodon, Tetrodon, etc.). The blood supply of the air-bladder is arterial from the system of the dorsal aorta, either from the efferent branchial vessels, from the coeliac artery or from the dorsal aorta. In Gym- 204 SUB-CLASS (AND ORDER) TELEOSTEI. narchus the efferent vessels of the third and fourth branchial arches go exclusively to the air bladder (Hyrtl). The veins join either the system of the posterior cardinal, or the hepatic, or the portal. The air-bladder is extraordinarily variable in its occurrence. It is entirely absent in some families, e.g. Blennidae, Pleuro- nectidae, Symbranchidae. It may be present in some genera of a family and absent in others, or even in different species of the same genus.* Several functions f have been ascribed to. the air-bladder ; it has been said to be hydrostatic, a resonator, sound producing, and respiratory. There can be but little doubt that it is a hydrostatic organ : J its function appears to be to keep the weight of the fish equal to the weight of the volume of water it displaces. Thus if the fish sinks, its body is compressed and the specific gravity is increased. To meet this the air-bladder slowly secretes gas, which distends the bladder and so restores the specific gravity of the fish to its former point. Further, when the fish rises, its air-bladder becomes distended and its specific gravity diminishes. The fish consequently has some difficulty in preventing its body rising to the surface. To meet this, the superfluous gas is slowly absorbed and the air-bladder becomes reduced in bulk so that the specific gravity of the fish returns to its normal point. In the Ostariophysi the reduction of pressure causes the fish to expel the gas through its pneu- matic duct, but this does not always occur in other fishes with pneumatic duct, though possibly it does so in some cases. In fishes without pneumatic duct the only way in which the super- fluous gas can be removed is by absorption. With regard to the process of secretion, it takes place so slowly that it would not be worth while for a fish to change its depth unless it meant the change to be of some duration. Moreover, Biot,§ and more recently Moreau, have shown that the gas secreted is mostly oxygen. The gas in the air-bladder of fishes taken near the surface contains nitrogen, oxygen, and a trace of carbonic acid (not more than 1 or 2 per cent.). The nitrogen in such cases is * See Stannius, Handbuch, 2nd edit. loc. cit., p. 221. f Vide W. Sorensen, in Journal of Anatomy and Physiology, 29, 1895, p. 109, et seq. J A. Moreau, Recherches exp. s. 1. functions de la vessie natatoire, Ann. d. Sci. Nat., 4, 1876. § Memoir es d. 1. Societe d1 Arcueil, 1, 1807. AIR-BLADDER. 205 considerably in excess of the oxygen, which amounts to from 9 to 20 per cent. In fishes taken from deeper water the per- centage of oxygen increases, to as much as 87 to 90 per cent, in fishes taken from a great depth. Further, if the air-bladder be artificially emptied, the fish sinks to the bottom, but it slowly recovers by gas-secretion ; the gas so secreted is richer in oxygen than air. At the same time nitrogen must also be secreted, and sometimes appears to be the only gas secreted.* Recently Bohr f has shown that section of the vagus prevents the secretion of gas into the air-bladder. The evidence that it acts as a respiratory organ is very slight. Many fishes swallow air, but there is nothing to show that the air is taken into the air-bladder. In some cases however it has been shown (e.g. by Budge tt, op. cit., in young Gymnarchus) that the fish dies if it is prevented from coming to the surface to take in air. It has however been suggested with more probability that the oxygen secreted into the bladder may serve as a store when the fish enters water in which the supply of oxygen is too small. In some fishes, e.g. the Ostariophysi, in which the posterior chamber is non-distensible and often enclosed by bone, while the anterior chamber is distensible and connected by ossicles to the membranous labyrinth, and in other fishes (see above) in which the air-bladder is in connection with the ear, it has been surmised, though not in any way proved, that the air-bladder acts as a re- sonator in intensifying sound vibrations and transmitting them to the auditory apparatus. On this view the Weberian apparatus may be of use in increasing the acuteness of hearing. It has also been suggested that it acts as a sound-producing organ, as a consequence of the incomplete septa and membranes which project into it being set in vibration by a movement of the con- tained air, caused by contraction of the extrinsic and intrinsic muscles which are contained in its walls. This suggestion rests on observation, for many fishes possess the power of producing sound (grunting, drumming, hissing, etc.), as many Sciaenidae, some Siluridae (Doras, Platystoma, etc.), Trigla gurnardus, Dactylopterus volitans, Malapterurus electricus, and many others ; and in some cases the sound has actually been detected * See Hiifner, in Arch. f. Anat. und Physiologic, 1892, p. 54. t C. Bohr, Influence of section of vagus on gas secretion in air-bladder, Journ. of Physiology, 15, 1894, p. 494-500. 206 SUB-CLASS (AND ORDER) TELEOSTEI. proceeding from the air-bladder in fishes just removed from the water and opened. It appears that sounds may be produced in some fishes without special air-bladder muscles by the activity of muscles, with the fascia of which the air-bladder is in close connection (Peristedion cataphractum, Trigla lyra, Sciaena aquila, etc.). Vascular System. The heart is without a conus arteriosus, and is usually separated from the ventral aorta by two semilunar valves only, though there is sometimes a small third valve. In some Clupeids it is said * that a trace of a small conus provided with striped muscles may be made out, and in Butirinus (Albula) there are actually two rows of valves (two large and two small in the proximal, and two large in the distal row). The ventral aorta presents at its ventricular end a swelling, the bulbus arteriosus, due to the thickness (elastic tissue and smooth muscular fibres) of its walls at this point. It gives off branches to the four anterior branchial arches, which usually bear gills. The ventral aorta does not as a rule give off branches to arches which are without gills, but in some cases with deficient posterior gills (Sym- branchus, Amphipnous, etc.) the afferent vessel from the ventral aorta is present and passes round directly into the efferent vessel, so that venous blood is conveyed into the dorsal aorta. As an example, we may mention Amphipnous, the first branchial of which has no gills ; the second has a few filaments ; the third has a transparent fringed membrane, and the fourth has no gills. The breathing organs are two sacs filled with atmospheric air and placed over the upper ends of the branchial arches ; they open into the branchial cavity between the dorsal end of the hyoid and first branchial arches. The ventral aorta gives off a branch on each side, which passes to the fourth branchial arch and joins its fellow to form the dorsal aorta ; it then gives off small branches to the second and third branchial arches and to the air-cavities, the blood from which is returned in two trunks which join the dorsal aorta. The blood after passing through the gills is collected by the efferent branchial arteries, of which one leaves each gill-bearing arch. These fall into the two roots of the dorsal aorta, which anastomose in front dorsal to the parasphenoid bone and con- stitute the so-called circulus cephalicus (Fig. 120, cc). The circulus cephalicus gives off anteriorly the internal (at a) and external (b) carotid arteries, and receives an anastomosing branch from the hyoidean artery (vh) ; posteriorly it gives off the * J. E. V. Boas, Morph. Jahrb., vi., 1880, p. 527. VASCULAR SYSTEM. 207 two subclavian arteries, and on the right side the large coeliac artery (r). The hyoidean artery (vh) is a continuation of the ventral end of the efferent vessel of the first branchial arch on to the hyoid, up which it runs in a dorsal direction to supply the pseudobranch (/&). The efferent vessel of the pseudobranch 208 SUB-CLASS (AND ORDER) TELEOSTEI. (ef) anastomoses with its fellow of the opposite side and then passes round the external carotid to the choroid gland, a rete mirabile in the choroid coat round the entrance of the optic nerve into the eye. When the pseudobranch is not present, there is no choroid gland. In Gymnarchus the efferent vessels of the third and fourth branchial arches do not join the dorsal aorta but pass to the air-bladder. The blood of the choroid gland supplies the choroid coat. The iris and sclerotic are supplied by the external carotid. The efferent bran- chial vessels give off small vessels for the nutrition of the branchial arch tissues, and near their ventral ends they give off vessels for the ventra part of the body, the neighbouring parts of the head, and even in some cases the heart. The hyoidean artery which supplies the pseudobranch is an example of this system of arteries. The dorsal aorta is frequently closely adherent to the ventral wall of the vertebral column, so that the latter appears to form part of its wall. It may be swollen at intervals, and in some forms (Esox, Clupea, Salmo, Silurus, etc.) a fibrous elastic band projects into its cavity. The principal branches are subclavian, which may come off from the circulus cephalicus, the coeliaco-mesenteric which frequently gives off the air-bladder vessel and a posterior mesenteric. The veins are arranged in the usual piscine fashion. The left posterior cardinal is often smaller than the right and appears as a small branch of the latter coming from the anterior part of the kidney ; or the right vein may alone be present, lying almost in the middle line and receiving branchlets from each side. A renal-portal system appears to be present in most Teleosteans. The hepatic-portal vein may receive tributaries from the air- bladder, the gonads (Perca, Bhnnius, Cyprinus, Osmerus, etc.), and the ventral body wall (Salmo, Alosa, Clupea, etc.), though these veins more generally open into the posterior cardinals. The body-cavity has the usual piscine arrangement. The pericardium is completely separated from the general body- cavity. Paired abdominal pores opening at the sides of the anus are absent in most Teleosteans, but they are found in the Salmonidae and Mormyridae, though not universally. They must not be confused with the pore-like oviducts of female Salmonidae, etc. (see below). The urinary organs are paired streaks of kidney substance * M. Weber, Morph. Jahrbuch, 12, 1886, p. 336. Jungersen, Arb. a, d. Zool. Inst. Wurzburg, 9, p. 93. URINOGENITAL ORGANS. 209 D R placed oil the ventral side of the vertebral column between it and the air-bladder. They have a great longitudinal extension, frequently reaching from the head to the end of the body-cavity, or even in some cases extending into the caudal region. Their front ends are enlarged into the so-called head-kidneys. The head-kidney, as was shown by Balfour, consists of lymphatic tissue which occupies the place of the pronephros of the larva. There are two longitudinal urinary ducts which unite posteriorly to form the single ureter. This structure, which frequently has a bladder-like dilatation, passes ventral- wards on one side of the air-bladder to open externally behind the anus, or into the rectum, into which the genera- tive duct may also open (some Sym- branchii, Plectognatki, Pediculati), or, in the Pleuronectidae, on a papilla placed asymmetrically on the coloured side of the body. Nephrostomata are never present. The generative and urinary open- ings, whether separate or united, fre- quently open on a papilla which may be of some length (Blenniidae, Gobiidae, etc.). In Rhodeus the opening of the oviduct is prolonged in the breeding season into a tube, by means of which the female deposits her ova in the shells of living bivalves (Fig. 123). The ovaries are usually double, rarely single, saccular bodies the walls of which are continued into the short median oviduct which opens between the anus and the urinary opening, or with the latter. In some Teleosteans the ovaries are separate from their ducts, and the ova are dehisced into the body- cavity whence they escape by two funnels which join to form a short tube which opens to the exterior usually between the anus and the ureter (Salmonidae, Muraenidae, etc.). In viviparous forms development takes place in the ovaries or in the oviduct. The testes are paired saccular bodies, and are apparently always, z-n. p 7e TIG. 121.— Kidneys of- Salmo- iario (after Hyrtl). D ductus Cuvieri; R kidneys; U ureter ; Ur efferent duct of bladder ; Vs bladder-like dilation ; Vs subclaviau vein. 210 SUB-CLASS (AND ORDER) TELEOSTEI. continuous with the short duct which either opens in the same position as in the female, or joins the ureter, so that there is a median porus urogenitolis behind the anus. In the viviparous forms fertilisation is effected by an in- tromittent organ, which is usually formed by the urogenital papilla. A few Teleosteans (Serranus, etc.) are hermaphrodite. The ova fall into the body-cavity and escape by porelike oviducts in the Galaxiidae, Hyodontidae, Nolopteridae, Muraenidae and Salmonidae. In Fierasfer there is said to be a pronephros in the adult, and the pos- terior part of the kidney is not developed. The ova are always provided with soft shells and vary con- siderably in size ; amongst the largest are those of Gymnarchus (10 mm.) and of Arms (5 to 18 mm.). They may be deposited singly (salmon, trout, etc.), or they may be agglutinated to- gether by a substance secreted by the walls of the oviduct. In freshwater forms they either adhere to some foreign body or are deposited in nests ; in marine forms they are either attached to foreign bodies or float freely in the surface waters of the ocean, or sink to the bottom as in the herring. Most fishes breed once a year at a definite period, but some breed more than once, and in some the breeding period is much prolonged — as in the cod and herring. Care of the brood by the female is often found (Aspredo, Solenostoma, Cichlidae, see systematic part) ; in the male it is more frequent (by nests in Gymnarchus,* Heterotis, Coitus, Gafttrosteus, Cyclopterus, Antennarius, Ophiocephalus, Callichthys, etc. ; in Arius the eggs are carried in the pharynx of the male, in Lophobranchs in a pouch on the abdomen). The segmentation is meroblastic and the germinal layers arise by delamination. The cerebro-spinal cord is formed as a solid keel-like thickening of the ectoderm,which subsequently becomes hollow. The young are hatched at an early stage arid undergo the remainder of their development as larvae. The larvae have a pronephros and considerable remains of the yolk-sac. The Teleostean pronephros is characteristic, in that the portion of the body-cavity containing the glomerulus is quite cut off from the rest and is in relation with the pronephric duct by one body- cavity opening only. * Budgett, Breeding Habits of some W. African Fishes, etc., Trans. Zool. Soc., 16, 1901, p 115. HABITS. 211 Teleosteans frequently undergo remarkable changes of form in their growth. This is a marked feature of the group, and leads to some difficulty in the recognition of species. As ex- amples may be mentioned the Pleuronectidae, Cyttidae, Muraenidae, Xiphiidae, Plectognathi. In many cases the young are so different that they have been described as distinct genera. Moreover, Teleostean fishes are often highly variable under the influence of changed conditions (variation in acquired char- acters), particularly with regard to colour, both of skin and flesh. The change in the colour of the skin is due to the pigment cells ( chroma tophors). Secondary sexual differences are usually present. The male is generally smaller than the female, and some of its fin-rays or fins may be prolonged. The male is often more brightly coloured in the breeding season, or its skin may become warty. Hybridism is also known to occur (Serramcs, Pleuronectidae, Cyprinidae, Salmonidae, etc.). Some fishes are very long-lived (carp and pike to beyond 100 years), and growth frequently appears to be somewhat indefinite and to continue for a long time. Fishes which rapidly a,ttain to their full size (e.g. sticklebacks) are said to be shortlived. A few fishes have been domesticated and transported to different parts of the globe (carp, Crucian carp, tench, goramy), and certain species of salmon and trout have been acclimatised in countries in which they are not indigenous (see accounts of families). Marine fishes are usually extremely sensitive to changes of temperature, freshwater fishes much less so. It is said that the carp will survive after being frozen in a block of ice. A modified hibernation has been observed in some Cyprinoids and Muraen- oids in cold weather, and many tropical fishes (Siluroids, Labyrinthici, Ophiocephaloids, etc.) pass the dry season in a torpid state in hardened mud. The flesh of many fishes is poisonous, and in unknown waters, especially in tropical seas, great care must be exercised in select- ing fish for food. The wounds caused by the spines of many fishes are poisonous. This is generally due to the poisonous nature of the mucus which covers the body, but it may be caused by special poison glands, as in Synanceia, Thalassophryne. 212 SUB-CLASS (AND ORDER) TELEOSTEI. There is a considerable number of marine fishes which occasionally wander into freshwater and ascend rivers (e.g. Sciaenidae, Pleuronectes, species of Clupeidae), and a smaller number of freshwater fishes which occasionally descend into the sea (some species of Salmo, of Siluroids, of Coregonus, and pre-eminently the Gastrosteidae and Cyprinodontidae) ; but most of these are inhabitants of the brackish water. They must be dis- tinguished from those fishes which migrate for the purpose of spawning. Such are of two kinds ; there are the anadromous fishes which ascend rivers to spawn in freshwater, as the salmon and the salmon-trout, some Clupeids, etc., and katadromous fishes, like the freshwater eel, which descend to the sea to spawn. There are many clear cases of marine fish which by geological changes have been retained in freshwater basins ; such are Coitus quadricornis, in the large lakes of Scandinavia ; species of Gobius, Blennius and Atherina in the lakes of N. Italy ; Comephorus in the depths of Lake Baikal. The classification of the Teleostei adopted here is essentially that of Mr. G. A Boulenger, F.R.S., to whom I am greatly in- debted for having allowed me to see proofs of his work before its publication. It is as follows :*— Sub-order 1. MALACOPTERYGII (SALMONICLUPEIFORMES).| 2. OSTARIOPHYSI (CYPRINISILURIFORMES). „ 3. SYMBRANCHII (SYMBRANCHIFORMES). „ 4. APODES (ANGUILLIFORMES). „ 5. HAPLOMI (ESOCIFORMES). .. 6. HETEROMI (DERCETIFORMES). ,: 7. CATOSTEOMI (GASTROSTEIFORMES). Tribe A. Selenichthyes. ,, B. Hemibranchii. „ C. Lophdbranchii. „ D. Hypostomides. „ 8. PERCESOCES (MUGILIFORMES). „ 9. ANACANTHENI (GADIFORMES). 10. ACANTHOPTERYGII. Tribe A. Perciformes. „ B. Scombriformes. „ C. Zeorhombi (Zeirhombi formes). „ D. Kurtiformes. . „ E. Gobiiformes. ,, F. Discocephali (Echinei former). „ G. Scleroparii (Trigliformes). ,, H. Jugulares (Blennii formes}. * Boulenger, Ann. and Mag. Nat. Hist. (7), 13, 1904, p. 161. t The names in brackets are those used in the fish-gallery of the British Museum. M ALACOPTERYGII. 213 Tribe I. Taeniosomi (Lophotiformes). Sub-order 11. OPISTHOMI. „ 12. PEDICULATI (LOPHIIFORMES). „ 13. PLECTOGNATHI (BALISTIFORMES). Tribe A. Sclerodermi. ,, B. Gymnodontes. The old group Physostomi (with a ductus pneumaticus to the air- bladder), which is sometimes referred to in the preceding pages, included, roughly speaking, the Malacopterygii, Oslariophysi, Symbranchii. Apodes, Haplomi, Heteromi (in part), and Perce- soces (in part), of the above classification. Sub-order 1. MALACOPTERYGII (SALMONI-CLUPEIFORMES)- Soft-rayed fishes with the anterior vertebrae simple, unmodi- fied, and without auditory ossicles ; symplectic present or absent ; opercular bones distinct ; pharyngeal bones simple above and below, the lower not falciform. Pectoral arch suspended from the skull ; mesocoracoid always well developed. Maxillary bone forming part of margin of upper jaw ; no barbels. Supra- occipital sometimes separated from the frontals by the parietals. Gills 4, a slit behind the fourth. Air-bladder if present with a pneumatic duct. Dorsal and anal fins without true spines. Pelvic fins abdominal, sometimes absent ; scales usually cycloid, sometimes ctenoid ; occasionally absent. No developed photo- phores. Adipose fin present or absent. This sub-order of Teleostei is nearest to the Ganoids. Fam. 1. Leptolepidae. Extinct. Upper Lias to Lower Cretaceous; vertebral centra nearly complete, pierced by the notochord ; without fulcra ; scales cycloid. Leptolepis Ag., Thrissops Ag. The Pholidophoridae (p. 180), Oligopleuridae (p. 182), and the Archaeo- maenidae, all extinct, are placed here by Smith Woodward and Boulenger. Fam. 2. Mormyridae.* Body and tail scaly ; head scaleless ; upper jaw formed by the two premaxillaries which are fused, and by the maxil- laries. Sub- and very small inter-operculum present ; supraoccipital separated from frontals by parietals. On each side of the skull there is a large cavity leading into the interior and covered by a thin bony lamella. They are without pharyngeal teeth. All the fins are well developed in Mormyriis, caudal, anal and pelvic fins are absent in Gym- narchus. No adipose fin. Pectorals directed upwards. Pseudobranch absent, gill-apertures reduced to a short slit. Air-bladder simple, com- municating with the ear. Two pyloric caeca. A series of pores along * Kolliker, Bericht v. d. zootom. Anstalt zu Wurzburg, 1849. Hyrtl, Denkschr. Akad. Wiss. Wien, 1856, xii. p. 1. ErdL, Munchner Gelehrte Anzeigen. Boulenger, Poissons du Bassin du Congo, 1901. 214 SUB-CLASS (AND ORDER) TELEOSTEI. the base of the dorsal and anal (if present) fins. f. w. of tropical Afr. They possess an electric organ on each side of the tail with feeble electric functions, consisting of modified muscle- tissue. The snout fre- quently of strange shape ; eyes often reduced. The brain is remarkable for its size ; 10 genera. Mormyrus L., teeth in rows along the middle of the palate and the tongue ; M. oxyrhynchus Geoff., venerated by the ancient Egyptians ; Hyperopisus, Mormyrops ; Gymnarchus Cuv., Nile and W. Afr. rivers, eel-like, each jaw with incisor-like teeth ; air-bladder cellular, very extensible, duct with sphincter at oesophageal opening ; lays very large eggs ; the gills of the embryo project beyond the gill-openings. Gnathonemus. Fam. 3. Clupeidae. Body covered with scales, head naked ; supra- occipital in contact with frontal. Abdomen frequently compressed into a serrated edge. Maxillaries (of three pieces) and premaxillaries both enter into upper jaw. Opercular apparatus complete. Adipose fin absent, dorsal not elongate, anal sometimes very long. Stomach with blind sac, pyloric caeca numerous. Gill openings usually wide. Pseudo- branch usually present. Air-bladder simple, large, communicating with the ear. Principally coast fishes ; none from the deep sea ; may enter f. ws. communicating with sea ; temp, and trop. zones. Many fossil forms. Engraulis C. et V., anchovies, upper jaw prominent ; mouth with a very deep cleft ; eyes covered by skin ; E. encrasicholus L. (Anchovia J. and E.), the anchovy, abundant in Med., also taken in E. Channel ; Cetengraulis Gthr. ; Stolephorus Lac. ; Coilia Gray ; Dussumieria ; Etrumeus ; Chatoessus C. et V. (Dorosoma Raf.% C. Amer., Aust., E. Ind., Japan. Clupea Cuv., herrings, upper jaw not projecting, eyes with free lateral adipose lids, more than 60 species, most used as food, but some trop. species poisonous ; C. harengus L., the herring, incredibly prolific, whitebait consists chiefly of the young of the herring (and sprat), the air-bladder opens into the stomach, and also on the left side near the anus,* the eggs are attached to stones, etc.; C. pilchardus Walb. (Clupanodon Lac.), the pilchard (the young is the sardine), equally abundant in Brit. Channel, on coast off Portugal and in Med. ; C. sprattus L., the sprat, in Norfolk sold as anchovies ; C. alosa L. (Alausa, Alosa), the shad or allice-shad, coasts of Eur. ascending rivers ; C. finta, the twaite-shad. Other Clupeoid genera are Clupeoides, Pellonula, Clupeichthys, Pellona, (Ilisha), Pristigaster, Chirocentrodon, Pomolobus, Sardinella, Opisthonema, Brevoortia, Opisthopterus, Odontognathus, Pristigaster ; Chanos Lacep. wi£h accessory branchial organf in a cavity behind the gill-cavity, Indo- Pac., 4 ft., edible. The following genera may be placed here : Elops, Megalops (M. atlan- ticus, the tarpon), Albula (Butirinus) with a trace of the conus (with two rows of valves) in the heart ; Pterothrissus (Bathythrissa), deep sea, Japan. Fam. 3a. Hyodontidae (moon eyes), f. w. fish of N. America, no oviducts ; Hiodon, Le Su. Fam. 4. ' Alepocephalidae. Deep-sea fishes approaching the Sal- manoids ; without adipose fin or air-bladder. Phosphorescent spots none or small. Stomach curved, without blind sac ; pyloric caeca in moderate number. Pseudobranch present. Alepocephalus, Mitchillina, Bathytroctes, Talismania, Conocara, Platytroctes, Aleposomus. * Weber, " De aure et auditu," 1, 1820, vii., 63. f J. Miiller, Bau u. Grenzen d. Ganoiden, p. 75. MALACOPTERYGII. 215 Fam. 5. Notopteridae, with one genus Notopterus, f. w. of E. Ind. and W. Afr. Fam. (5. Osteoglossidae. Body covered with large mosaic-like scales ; head scaleless, its integument confluent with the bone ; dorsal fin on tail and opposite anal ; gill openings wide, pseudobranch absent ; air- bladder simple or cellular, stomach without caecal sac, pyloric append- ages two. Eggs fall into body cavity. Large f. w. fishes of the tropics. 4 genera. Osteoglossum Vandelli, S: Amer. ; Ara- paitna Mull., Brazil and Guyanas ; Heterotis Ehr., trop. Afr.; Sclero- pages, Australia, E. Ind. Arch. Excluding the E. Ind. Archipelago, the distribution of this family is the same as that of the Dipnoi. Heterotis niloticus forms a nest and the young larvae have external gills. Fam. 7. Pantodontidae. One genus, f. w.,W. Afr., pectorals very large. Fam. 8. Ctenothrissidae. Extinct, Cretaceous. Fam. 9. Phractolaemidae. One genus, W. Afr. Fam. 10. Saurodontidae (Ichthyodectidae). Extinct, Cretaceous ; Portheus, Ichthyodectes. Fam. 11. Chirocentridae. One genus, Ind. Ocean and Seas of China and Japan. Fam. 12. Salmonidae. Body generally covered with scales, head scaleless ; margin of upper jaw formed by maxillaries and premaxillaries , a small adipose fin behind the dorsal ; pyloric caeca generally present and numerous ; air-bladder large and simple with a pneumatic duct ; pseudobranch present ; no oviducts. Inhabitants of sea and f. w. ; most of the mar. genera are from the deep sea ; most of the f. w. forms are peculiar to the temperate and arctic region of the Northern Hemisphere, one occurring in New Zealand ; many f. w. species are anadromous ; no fossils of f. w. species known. Osmerus and other genera from the Miocene. Salmo Art., trout, salmon and charr, inhabitants of f. w., many species descending to the sea after spawning (anadromous), the young of all are barred, the bars vanishing in adult except in small varieties ; many of the species are highly variable and capable of considerable adapta- tion to their surroundings ; the marine forms usually silvery with or without black spots, the f. w. forms more or less speckled with black and red. Some individuals of full size are sterile, but this is pro- bably only a temporary condition ; overgrown individuals are sterile ; anadromous fish, generally return to their native river. River and sea trout have been acclimatised in Tasmania and N. Zealand, and appar- ently in India. As these species are highly variable in response to change of condition the observation of these acclimatised races will afford an extremely interesting study. S. salar L., salmon, N. Hemisphere between latitudes 45° and 75° ; does not occur in rivers opening into Med.,* the last Thames salmon was caught in 1833 ; a marine fish ascending rivers to spawn (Sep. to Jan. in Britain), the nest or redd is dug out by the female in gravel and the eggs are buried ; young salmon of the first and second year are called parr or pink (4 to 6 in.) ; they then become smolts, which descend to sea, and reascend the rivers as grilse, which having spawned go to the sea and return as salmon ; a salmon which has spawned is a kelt, kelts go to the sea, and probably reascend next year ; kipper is a male kelt, or a salmon which has been detained in f. w. and got * Not even in those of Macedonia, notwithstanding Fluellen (Henry V., Act 4, Sc. 7) ! 216 SUB-CLASS (AND ORDER) TELEOSTEI. lean ; male parrs may become sexually mature and fertilise the eggs of a full-grown female ; will hybridise (artificially) with the trout and charr ; not a highly variable species and change of conditions is fatal. The various kinds of British trout are probably all varieties of the same species, as they will freely cross, but three species may for convenience be dis- tinguished : S. trutta Flem., sea- or salmon-trout, phinok, sewin, a migra- tory species ; S. jario L., the brook trout ; and S, levenensis Walker, Loch Leven trout. To these the bull-trout, S. eriox, and great lake- trout, S. ferox, may possibly be added. S. alpinus L., the charr, breed Nov. to Dec. ; redd usually in gravelly shallows in the lakes (the Windermere charr is known as S. willughbii) ; the charrs are migratory or non-migratory and inhabit the deep waters of lakes ; there appears to be one British species with several varieties. S. fontinalis Mitchell, f. w. of Brit. N. Amer., acclimatised in Britain. Oncorhynchus Suckley, anadromous fish in American and Asiatic rivers flowing into the Pacific ; O. tschawytscha, the Calif ornian salmon ; Brachymystax Gthr., Siberian rivers ; Lucio- trutta Gthr. (Stenodus Rich.), Arctic N. Amer. ; Plecoglossus Schley., f. w. of Jap. and Formosa. Osmerus L., smelts, migratory, ascending rivers to spawn and frequently becoming resident in them, Atl. coasts of N. Eur. and N. Amer. ; O. eperlanus L., the smelt, sparling, irregular in its migrations, spawning in rivers near high-water mark, generally found in rivers from Aug. to May, spawns about March or April, when fresh exceedingly good eating, but deteriorate in a few hours ; allied genera are Hypomesus and Thaleichthys from the Pacific coast of N. Amer., the latter, known as Oulachan, has so much oil it will burn like a candle ; Mallotus Cuv. ; Coregonus Art. (Argy- rosomtis Ag.), whitefish, mostly lacustrine, a few anadromous, northern parts of temp. Eur., Asia, and N. Amer. ; C. oxyrhynchus L., houting, marine entering f. w. Holland, Germany, Denmark ; C. clupeoides Lac., schelly, f. w. of Lake District and Wales ; C. vande- sius Rich., vendace, f. w. lochs of Scotland ; C. pollan, Thomps., pollan, f. w. of Ireland ; Thymallus Cuv., graylings, clear streams of Eur., Asia, and N. Amer.; Th. vulgaris Nilss., grayling, flesh good, in best condition Oct. and Nov. ; Salanx Cuv. ; deep- sea genera are Argentina, Microstoma, Bathylagus. Fam. 13. Stomiatidae. Scales absent or thin ; a hyoid barbel ; eyes large ; luminous spots more or less developed ; no pseudobranch ; ovi- ducts present. Deep-sea fishes descending to the greatest depths and distinguished by their barbel and formidable dentition. Astronesthes Rich., Stomias Cuv., Echiostoma Lowe, Grammatostomias G. and B., Photonectes Gthr., Malacosteus Ayres, Bathyophis, Chauliodus Bloch and Schneider, Bathylaco, Maurolicus, Sternoptyx, etc. Fam. 14. Gonorhynehidae, one genus, Aust. and Japanese seas. Fam. 15. Cromeriidae, with one genus Cromeria, recently discovered in the White Nile. Sub-order 2. OSTARIOPHYSI (CYPRINI SILURIFORMES). The anterior vertebrae are co-ossified and have some of their lateral elements detached to form a chain of small bones, the Weberian ossicles, which connect the air-bladder with the ear (p. 202). The air-bladder is probably always present though OSTARIOPH7SI. 217 it may be very small. When well developed it has a pneumatic duct. Pectoral arch suspended from the skull, mesocoracoid present. The great majority of freshwater fishes are included in this sub- order. Fam. 16. Characinidae. Body scaly, head naked; barbels absent, margin of upper jaw usually formed by the premaxillaries and maxillaries, rarely by the premaxillaries only ; jaws usually toothed ; parietals distinct from supraoccipital ; symplectic present ; generally a small adipose fin behind the dorsal ; pelvics abdominal ; pyloric appendages more or less numerous ; air-bladder divided into two portions ; pseudo- branch absent or much reduced. Freshwaters of Africa and of tropical America. In America they replace the Cyprinoids ; unknown as fossils. Erythrinina. Adipose fin absent ; trop. Amer. The genera are Macrodon, Erythrinus, Lebiasina, Nannostomus, Pyrrhulina, Corynopoma. Curimatina. A short dorsal and an adipose fin ; dentition im- perfect ; trop. Amer. ; Prochilodus, Caenotropus, Hemiodus, Saccodon, Parodon. Citharinina. A rather long dorsal and an adipose fin ; minute labial teeth ; trop. Afr. Citharinus Cuv., attaining to 3 ft. Anastomatina. Short dorsal and an adipose fin ; teeth in both jaws well developed ; the gill membranes grown to the isthmus ; nasal openings remote from each other ; trop. Amer. Lepo- rinus, Anastomus, Rhytiodus, Nannocharacina. Like the last, except that incisors are notched, and nostrils close together. Nannocharax. Tetragonopterina. Short dorsal and an adipose fin; teeth well developed, notched or denticulated ; gill-membranes free from the isthmus ; nasal openings close together ; S. Amer. and trop. Afr. Alestes M. and T., trop. Afr. ; Tetragonopterus Cuv., trop. Amer. Of the other genera Nannaethiops and Bryconae- thiops are African, the rest are S. American, viz., Chirodon, Megalobrycon, Gastropelecus, Piabucina, Scissor, Pseudochalceus, Aphyocharax, Chalceus, Brycon, Chalcinopsis, Bryconops, Creagrutus, Chalcinus, Piabuca, Paragoniates, Agoniates. Hydroeyonina. Short dorsal and adipose fin ; teeth well developed and conical ; gill-membranes free from the isthmus ; nasal openings close together. S. Amer. and trop. Afr. Fishes of prey. Hydrocyon Cuv., trop. Afr., and Cynodon Spix., S. Amer., both to 4 ft. Except Sarcodaces from W. Afr. the other genera are trop. Amer., e.g., Anacyrtus, Hystricodon, Salminus, Oligo- sarcus, Xiphorhamphus, Xiphostoma, etc. , Distichodontina. Dorsal fin rather elongate, adipose fin present ; gill-membranes attached to the isthmus ; belly rounded. Trop. Afr. Distichodus M. and T. Ichthyborina. An adipose fin ; dorsal rays 12 to 17 ; gill-mem- branes free from the isthmus ; belly rounded ; canine teeth ; trop. Afr. Ichthyborus Giinth., Nile ; Eugnathichihys , Phago, W. Afr. Crenuchina. Dorsal fin rather elongate, an adipose fin ; gill- membranes free from the isthmus ; belly rounded ; without canine teeth. Crenuchus Giinth., Brit. Guiana; Xenocharax, W. Afr. 218 SUB-CLASS (AND ORDER) TELEOSTEI. Serrasalmonina. The caribe. Dorsal fin rather elongate ; an adipose fin ; gill-membranes free from the isthmus ; belly serrated ; trop. Amer. ; exceedingly voracious, they assail persons entering the water. Mylesinus, Serrasalmo, Myletes, Catoprion. Fam. 17. Gymnotidae (yvfwds naked, vuros back). Head scale- less ; barbels none ; body eel-shaped ; scales small or absent ; margin of upper jaw formed by premaxillaries and by maxillaries ; an- terior vertebrae united, modified, with Weberian ossicles ; dorsal fin absent or reduced to adipose strip, caudal generally absent ; tail ending in point, can be regenerated ; anal long, pelvics absent ; anus on or near the throat ; shoulder girdle attached to skull ; ribs well developed ; gill-openings narrow ; air-bladder double ; stomach with caecal sac ; pyloric caeca present ; ovaries with oviducts. Eel-like f. w. fishes from S. America. Sternarchus Cuv., Rhamphichthys M. and T., Sternopygus M. and T., Carapus M. and T., Gymnotus Cuv. (Electro- phorus), electric eel, Brazil and Guyanas, electric organ along each side of the tail ; Giton Kaup, Eigenmannia J. and E. Fam. 18. Cyprinidae. Body generally covered with scales ; head naked. Anterior 4 vertebrae modified and joined, margin of the upper jaw formed by the premaxillaries. Belly rounded or if trenchant without ossifi- cations. No adipose fin ; a dorsal and anal fin, pelvic fins abdominal. Stomach without blind sac. Pyloric appendages absent. Mouth toothless ; lower pharyn- geal bones well developed, falciform, sub-parallel to the branchial arches, pro- vided with teeth in one, two, or three series. Air-bladder large, divided into an anterior and posterior portion by a con- Fio. 122.-Lower pharyngeal bones striction> or into a right and left portion, of a carp (after Heckel and Uner, enclosed in an osseous capsule. Ovarian from Claus). gacs closed. About 200 genera and 1,200 species ; freshwaters of the Old World and N. America. The fossil forms can be referred mostly to living genera. I. Catastomina. Pharyngeal teeth in a single series, numerous ; dorsal fin long, anal short ; barbels none. Lakes and rivers of N. Amer., 2 spec, from N.-E. Asia, generally known as suckers. Ictiobius Raf., Carpiodes Raf., Cycleptus Raf., Panto- steus Cope, Catostomus Le Sueur, Chamistes Jordan, Xyrauchen Eig. and Kirsch, Erimyzon Jordan, Minytrema Jordan, Moxo- stoma Raf., Placopharynx Cope, Lagochila Jord. and Bray. II. Cyprinina. Anal fin short with not more than 5 or 6, rarely 7, branched rays. Abdomen not much compressed. Barbels often present, never more than 4. Three branchiostegals. Air- bladder without osseous covering. Cyprinus Art., carps ; large scales ; dorsal fin long with its last un- divided ray osseous and serrated ; pharyngeal teeth in three rows, molar-like (Fig. 122); four barbels. C. carpio the carp, indigenous in Persia and China, introduced into Europe (known 1258 A.D.), into England (known 1496); food vegetable and animal ; bury themselves in mud in winter, will live for some time out of water, may attain a large OSTARIOPHYSI. 219 size (20-50 lb.), and great age (50-100 years, Gesner, Buffon), very prolific, spawn on weeds about May, said to form hybrids with the Crucian carp with the tench and the bream. Carassius Nilsson, without barbels ; C. vulggaris Nilss., the Crucian carp, Prussian carp, Eur. and Siberia ; C. auralus L., gold-fish, China and Japan, introduced into Eur. and Amer. as an aquarium fish and natural- ised in many streams ; very variable under domestication in colour and otherwise, brilliancy generally decreases when turned into the open, in the wild state greenish ; so-called telescope-fish is a variety ; breeds in May and June. Catla C. and V., E. Ind. ; Labeo Cuv., Afr. and E. Ind. ; Discognathus Heck., Ind., Ceylon, S.-W. Asia, Afr. ; Capoeta C. and V., W. Asia ; Barbus Cuv., barbels, 200 species, Eur., Asia, Afr., dorsal fin with the (third) longest simple ray sometimes enlarged and serrated only exceptionally with more than nine branched rays commencing opposite or nearly opposite the root of the pelvic fin ; eyes without adipose eyelid ; mouth arched without inner folds ; lips without horny covering, barbels 4, 2 or 0 ; B. vulgaris Fleming, Europe, to 50 lb., as food coarse, roe scmetimes poisonous. Thynn- •>'chthys Bleek, E. Ind. Oreinus ; McClell, Himalayas ; from same region Ptychobarbus , Gymnocypris, Schizopygopsis, Diptychus ; Gobio Cuv., Eur. a small maxillary barbel; dorsal fin with few rays, without spine; G. fluviatilis Flem., the gudgeon. Allied are Ladislavia and Pseudogobio , E. Asia; Ceratichthys Baird and Gerard, N. Amer., called chub in the U. S. ; similar genera of N. Amer., and generally called "minnows," are Pimephales (black head), Hyborhynchus, Hybognathus, Campostoma (stone- lugger), Ericymba, Cochlognathus, Exoglossum (stone-toter or cut- lips), Rhinichthys (long-nosed dace). Other Old World genera are Cirrhina, Dangila, Osteochilus, Barynotus, Tylognathus, Abrostomus, Crossochilus , Epalzeorhynchus, Barbichthys, Amblyrhynchichthys, Albulichthys, Aulopyge, Bungia, Pseudorasbora. III. Rohteichthyina. Anal fin very short, with not more than six branched rays ; dorsal fin behind pelvic ; abdomen compressed ; no barbels ; pharyngeal teeth in triple series. Rohteichthys Bleek, East Ind. Arch. IV. Leptobarbina. Anal fin as in last ; dorsal opposite pelvic ; abd. not compressed ; barbels present, not more than 4 ; phar. teeth in triple series. Leptobarbus Bleek, E. Ind. Arch. V. Rasborina with Rasbora, from E. Ind. Cont. and Arch, and E. Afr. ; Amblypharyngodon, Luciosoma, Nuria and Aphyocypris, from E. Ind. Cont. VI. Semiplotina with Cyprinion from Syria, Persia, Semi- plotus from Assam. VII. Xenocypridina with Xenocypris and Paracanthobrama from China, Mystacoleucus from Sumatra. VIII. Leuciscina. Anal fin of short or moderate length, with 8-11 branched rays, not extending forwards below the dorsal, which is short and without osseous ray. Barbels generally 0 ; pharyngeal teeth in a single or double series. Leuciscus Klein, white-fish, north temperate zone of both hemispheres ; species found in England are L. rutilus Flem., the roach, said to form hybrids with the bream and ludd ; L. cephalus Flem., the chub; L. vul- 220 SUB-CLASS (AND ORDER) TELEOSTEI. garis Flem., the dace ; L. erythrophthalmus Flem., the rudd or red- eye ; L. phoxinus Flem., the minnow ; L. idus, the id or nerfling, found in Europe, is domesticated in Germany, assuming the golden hue of semialbinism like a goldfish. Tinea Cuv., Eur. and Asia Minor, has been acclimatised in India, T. vulgaris Cuv., the tench, golden tench as a variety, due to albinism, as in the id and gold- fish. Leucosomus Heck., N. Amer., L. pulchellus (fall-fish, dace or roach), L. corporalis (chub), Chondrostoma Ag., Eur. and W. Asia; other Old World genera are Myloleucus, Ctenopharyngodon, Paraphoxinus ; N. American are Mylopharodon, Meda, Orthodon, Acrochilus. IX. Rhodenia with genera Achilognathus, Acanthorhodeus, Rhodeus, Pseudoperilampus, roach-like fishes in East. Asia and Japan ; in the females a long external urogenital tube is developed externally in the breeding season ; this deposits the large eggs into the mantle cavity of the pond mussel where they develop. Rh. amarus, the bitterling (Fig. 123), extends into Europe. X. Daninina. Small fish from E. Ind. Cont., Ceylon, E. Asiatic Islands, and a few from Afr. rivers ; Danio, Pteropsarion, Aspidoparia, Barilius, Bola, Scharca, Opsariichihys, Squaliobarbus, Ochetobius. XI. Hypo- phthalmichthyina. With H y p o- phthalmic hthy s from China. XII. Abra- midina. Anal fin elongate ; abdo- men or part of the abdomen FIG. 12S. — Rhodeus amarus, female (after comnressed A- v. Siebold, from Claus). oramis Cuv., the breams, temper- ate parts of both northern hemispheres; A. abr amis Flem., A.blicca Ag., both in Britain and Europe, hybrids between these two species and even other cyprinoids are not rare (Giinther) ; A. ballerus L., the zope, A. vimba L., the zarthe, Europe ; A. crysoleucas Mitchill, shiner, bream, United States. Aspius Ag., E. Eur. to China ; Alburnus Heckel, bleak, Eur., W. Asia ; A. lucidus Heck. u. Kner, Britain, Eur. north of Alps, absent in Scotland and Ireland ; other genera are Leucaspius and Pelecus, Europe ; Pelotrophus, E. Afri. ; and the rest, Rasborichthys, Elopichthys, Acanihobrama, Osteobrama, Chanodichthys, Hemiculter, Smiliogaster, Toxabramis, Culter, Eustira, Chela, Pseudolabuca, Cachius, from E. Ind. or temp. Asia. XIII. Homalopterina. Air-bladder absent, hill streams in E. Ind., genera Homaloptera, Gastromyzon, Crossostoma, Psilorhynchus. XIV. Cobitidina. Loaches. Barbels 6 or more ; dorsal fin short or of moderate length, anal fin short ; scales small or absent ; pharyngeal teeth in single series ; air-bladder partly or entirely enclosed in bony capsule ; pseudobranch absent. Misgurnus Lacep., Eur. and Asia, M. (Cobitis) fossilis Lacep., largest European loach ; Nemacheilus v. Hass., Eur., Asia, Abyssinia, without spine OSTARIOPHYSI. 221 near orbit ; N. barbatula Giinth., groundling, stone-loach, etc., Britain and ^Europe; Cobitis Artedi, Eur., E. Ind. ; C. taenia L., spined loach, with preorbital spine, Britain (rare) and Europe ; Botia Gray, E. Ind. ; from tropical India are Lepidocephalichthys, Acanihopsis, Oreonectes, Paramisgurnus, Lepidocephaltis, Acanth- ophthalmus, Apua. >.v*»:j Fam. 19. Siluridae. Cat-fishes. Skin naked or with osseous scutes, without scales. The 4 anterior vertebrae joined. Barbels always pre- sent ; maxillary bone small, almost always forming a support to a maxillary barbel. Margin of the upper jaw formed by the premaxillaries and maxillaries or by the premaxillaries only. Parietal bones confluent with the supra-occipital. Sub-operculum absent. Adipose fin present or absent. Pyloric appendages absent. Mostly inhabitants of the fresh waters of all the temperate and tropical regions, some entering the salt water, but keeping near the coast ; some are said to be able to cross land in search of other waters (Callichihys, Clarias, etc.). Over 100 genera and upwards of 1,000 species known. Clarias Gronov., Africa and S. Asia, muddy and marshy waters, an accessory branchial organ is attached to the convex side of the second and fourth branchial arches ; Nilotic species known as Carmoot. Heterobranchus G. St. Hil., Afr. and E. Ind. Arch., ace. gills as in Clarias ; Plotosus Lacep., brackish waters of Indian Ocean and Aust., brackish waters of Aust. ; Copidoglanis Giinth., Cnidoglanis Giinth, Chaca C. and V., East Indies ; Saccobranchus C. and V., E. Ind., gill-cavity with accessory posterior sac with contractile walls, vessels from last branchial artery and delivering into aorta. Silurus Art., temperate palaearctic rivers, S. glanis L., the wels, Europ. rivers east of the Rhine, to 300-400 Ib. African genera are Schilbe, Eutropius ; E. Indian are Silurichthys, Wallago, Belodontichthys, Eutropiichthys, Cryptopterus, Callichrous, Hemisilurus, Siluranodon, Ailia Schilbichthys, Lais, Pseudeutropius, Pangasius, Helicophagus, Silondia. Hypophthalmus C. and V., S. Amer., eye behind and below angle of mouth, Helogenes Giinth., Bagrus C. and V., Nile, B. bayad; African genera are Chry- sichthys, Clarotes ; E. Indian are Macrones, Pseudobagrus, Liocassis, Bagroides, Bagrichthys, Rita, Acrochordonichthys, Akysis. Amiurus (Ameiurus) Raf., horned pout, cat-fishes of N. Amer., one sp. in China; from N. Amer. also are Hopladelus, Noturus. Platystoma Ag., S. Amer. snout long, spatulate ; allied are Sorubim, Hemisorubim, Platystomatichthys, Phractocephalus, Piramutana, Platynematichthys , Piratinga, Bagropsis, Sciades, all from S. Amer. Pimelodus Lacep., 40 S. Amer. sp., 2 W. Afr. sp. ; allied are Pirinampus, Conorhynchus, Notoglanis, Callophysus, Lophiosilurus, all from S. Amer. Aucheno- glanis, trop. Afr. ; Arius C. and V., 70 sp., in all trop. countries and seas ; allied are Galeichthys, S. Afr. and Amer., mar. ; Genidens, Paradiplomystax Brazil ; Diplomystax Chili ; Aelurichthys C. and S. Amer. ; Hemipi- melodus, Ketengn^, Osteogeniosus, Batrachocephalus E. Ind. ; Atopochilus W. Afr. Bagarius Bleek, E. Ind. ; Euglyptosternum Bleek, Syria ; Glypto- sternum, Hara, Ambliceps E. Ind. Doras C. and V., Oxydoras Kner, Rhinodoras Kner, these three genera travel over land in the dry season in search of a pond of greater capacity, they make nests and both sexes tend the eggs, tropical S. Amer. in rivers flowing into the Atlantic ; the following also are S. American, Ageniosus, Tetranematichthys, Euanemus, Auchenipterus, 222 SUB-CLASS (AND ORDER) TELEOSTEI. Glanidium, Centromochlus, Trachelyopterus, Cetopsis, Astrophysus ; Syno- dontis C. and V., trop. Afr. Callomystax Giinth., Bengal, Mochocus Joannis. Rhinoglanis Giinth., Upper Nile. Malapterurus Lac., electric cat-fish, trop. Afr., electric organ extends over whole body beneath the skin.* Stygogenes, Arges, Brontes and Astroplebus in the lakes and torrents of the Andes, Humboldt thought they lived in subterranean waters and were ejected by volcanoes ; Callichthys, similar in dist. and habits to Doras (p. 221) ; Chaetostomus with the allied Plecostomus, Liposarcus, Pterygoplichthys ; Rhinelepis, Acanthicus, Xenomystus, from f. w. of S. Amer. ; Hypoptopoma ; Loricaria L., trop. Amer., Acestra Kn., Brazil, Surinam ; Sisor, N. India ; Erethistes M. and T., Assam ; Pseudecheneis Blyth, Himalayas ; Exostoma Blyth, E. Ind. continent. Aspredo L., Guiana, the female attaches the eggs to the spongy integu- ment of its belly by pressing against them ; Bunocephalus, Bunocephal- ichthys and Harttia from trop. Amer. Heptapterus, Nematogenys, Trichomycterus, Eremophilus, Pariodon are small S. American forms from f . w. of high altitudes to 14,000 ft.; they resemble the loaches of the N. Hemisphere in appearance and habits. Stegophilus Rein., and Vandellia C. and B., small fishes from Brazil, the latter are said to ascend urethra of persons bathing, but there is no doubt that they enter the gill-cavity of larger fishes. Cathorops Jordan and Gilbert, Panama ; Ictalurus Raf., f. w. of N. Amer. Sub-order 3. SYMBRANCHII. Body eel-shaped. Shoulder-girdle usually joined to the skull ; no mesocoracoid. Scales minute or absent. No paired fins. Unpaired fins reduced. Anus far from head. No air-bladder ; gill-openings confluent in a single slit. Stomach without cae- cum and pyloric caeca. Ovaries with oviducts. Widely dis- tributed in warm seas and fresh waters. Fam. 20. Symbranchidae. Eel-like, without paired fins, scales minute or absent ; gill-openings confluent into one slit on the ventral surface ; anus far from head ; no air-bladder, stomach caecum or pyloric caeca ; with oviducts ; f. w. and brackish w. of trop. Amer. ; 3 genera, and one marine genus (Chilobranchtis) from Australia. Amphipnous Mull, Bengal, 3 branchial arches of which the second alone possesses gills, and narrow slits, with a lung-like branchial sac on each side opening between hyoid and first branchial arches and supplied by branchial arteries ; A. cuchia ; Monopterus Lacep., 3 branch, arches and small gills, no branch, sac, East Ind. Arch, and Cont. ; Symbranchus Bl., 4 branch, arches and large gills, trop. Amer. and E. Ind. * Ballowitz, Das elect. Organ des afrikanischen Zitterwelses- Jena 1899. APODES. 223 Sub-order 4. APODES (ANGUILLIFORMES). The Eels. The premaxillaries small or absent, the maxillaries lateral, the body eel-like and without pelvics. Symplectic absent ; operculum and palatine arch reduced ; scales absent or feeble ; pectoral arch not attached to skull ; fins with- out spines, median fins if present confluent ; no pseudobranch ; tail protocercal ; no pyloric caeca ; no generative ducts. Air- bladder, when present, with a ductus pneumaticus. The eels are spread over the f. ws. and seas of the trop. and temp, zones ; some descend to the greatest depths. The young of some have a limited existence and are known as Leptocephalus (see below). Fossil Anguilla in chalk of Aix and Oeningen, Anguilla, Sphagebranchus, Ophichihys at Monte Bolca and Urenchelys S. Wood., with homocercal tail, from the chalk. The breeding * of the common eel was until a short time ago a mystery. During their sojourn in freshwater they do not develop reproductive organs, and it was not known how they originated. Aristotle thought that they came from the " entrails of the earth." It is now known, thanks to the researches of Grassi and Calandruccio, that they breed in the depths of the sea, that the eggs float but remain near the bottom, and that they hatch out as a larva, which soon becomes transformed into a ribbon- shaped, transparent creature, which has long been known and called Leptocephalus. There are several kinds of Leptocephalus. That of the common eel is L. brevirostris. It appears to remain at the bottom, pro- bably hiding under stones or burrowing in sand and mud until it meta- morphoses into the elver. Elvers (see below) are the young of eels which ascend rivers in great numbers. The Italian naturalists worked at Catania in the Straits of Messina, where specimens of the Leptocephalus brevirostris are common in certain years at the surface, and at all times in the stomach of Orthagoriscus mola, a deep-sea fish, and they showed that this particular kind is the larva of the common eel. That it should be taken here and nowhere else is a curious fact, considering that the common eel is widely distributed. The probable explanation is that it is brought to the surface by the currents and whirl- pools which abound in this locality, while elsewhere it has escaped observa- tion by lurking at considerable depths (300 fms.) in mud and under stones. Several species of Leptocephalus, which doubtless belong to different Muraenidae, are known as pelagic forms, especially in the tropics, so that it is probable that all Leptocephali are not confined to deep water during their development. Speaking generally it appears that female Murae- noids cannot mature their ova except in deep water, while the male can * B. Grassi and S. Calandruccio, Ulteriori recerche sulle metamorfosi dei Murenoidi, Rend. Ace. Lincei (5), vi., p. 43, 1897 ; also Q. J. M. S.,39, 1897, p. 371, and Proc. Roy. Soc., 1896. Cunningham, Journal of Marine Biological Assoc. (2), 3, 1895, p. 278, and (2), 1, 1891, p. 16. 224 SUB-CLASS (AND ORDER) TELEOSTEI. arrive at maturity at a less depth, but has to migrate to a greater depth to fertilise the eggs. The eggs float, but at a considerable depth and only exceptionally mount to the surface. The characteristics of typical Lepto- cephali are the transparent ribbon-shaped body with colourless blood, vent near the tip of the tail, small head, and large eyes. It has long been suspected that certain Leptocephali were the larvae of the conger, but many held that they were abnormal overgrown larvae incapable of further development, on the ground that they attained a size larger than that of the youngest conger, and because of the great variability of their form and dentition. The first naturalist who definitely observed the metamorphosis of a Leptocephalus into a young conger was Delage in 1886 (Comptes Bendus, 103, 1886, p. 698). In this metamor- phosis the skin became pigmented, the blood coloured, the air-bladder developed, and the body cylindrical and shorter. Grassi has shown that L. stenops (in part), L. morrisii and punctatus belong to the life-cycle of Conger vulgaris ; that L. haeckeli, yarrelli, bibroni, gegenbaurii, kdllikeri, stenops (in part) belong to Congromuraena mystax ; L. taenia, inornatus, and diaphanus to Congromuraena balearica, etc. Fam. 21. Derichthyidae. Body eel-like, from the abysses of the Atlantic ; Derichthys Gill. Fam. 22. Muraenidae, with the characters of the sub-order. (This family is now usually divided into several.) Group 1. Eels in which the branchial openings in the pharynx are wide slits. Nemichthys Rich., jaws produced into long slender bill, eyes large, with Serrivomer, Spinivomer, Avocettina, Labichthys are deep-sea (500-2,500 fms.) forms. Anguilla Cuv., eels, small scales imbedded in the skin, upper jaw not projecting beyond the lower ; gill-openings narrow, at the base of the pectoral fins ; dorsal fin some distance from head ; they freely ascend rivers, descending to the sea for purposes of reproduction ; f . w. and coasts of temp. and trop. zones, not yet found in S. Amer., W. coast of N. Amer. and W. Afr. A. anguilla L., the common Eur. and Brit, species, they descend rivers in the autumn and spawn in the deep sea ; the larva is known as Leptocephalus brevirostris ; the young eels are called elvers, and ascend rivers in incredible numbers in spring (April and May), overcoming all obstacles and even crossing land ; such migra- tions are known as eel-fares (of which elver may be a corruption), they bury themselves in mud and become torpid in winter, do not develop their generative products in freshwater ; the adult eels are said not to re- ascend rivers and to die soon after spawning. In eels migrating down the rivers to the sea the reproductive organs are enlarged, and the skin has a silver colouration. The eyes also are enlarged. All these peculiarities are observed in the sexually mature forms taken from the deep water. Young elvers are not known of a less size than 5 cm., while the larva, L. brevirostris attains a length of 8 cm. Simenchelys Gill, and Ilyophis Gilbert, are deep-sea eels ; Synapho- branchus Johns., gill-openings united into a longitudinal slit, deep-sea congers ; Conger Kaup., congers or marine eels, scaleless (Leptocephalus, Oxyurus, Helmictis, are all said to have priority over Conger), C. conger L., prefers deep waters with rocky bottom, attains to 8 ft., almost cosmo- politan ; allied genera are Poeciloconger, Congromuraena, Uroconger, Heteroconger ; Muraenesox McClell., scaleless, trop. seas ; Nettastoma Kaf., scaleless, deep-sea, the leptocephalid form is Hyoprorus ; Sauren- HAPLOMI. 225 chelys (Chlopsis), Oxyconger, Hoplunnis, Neoconger, all with superior or lateral nostrils, and Myrus, Ahlia, Myrophis, Paramyrus, Chilorhinus, Muraenichthys with nostrils in the upper lip, may be placed here. Oph- ichthys Gthr. (Ophichthus), nostrils labial, extremity of tail free, more than eighty species known, very numerous in trop. seas, formidable den- tition in jaws and palate ; Sphagebranchus, Verma, Letharchus, Myr- ichthys, Pisoodonophis, Callechelys, Bascanichthys, Quassiremus, Mystri- ophis, Scytalichthys, Brachysomophis, are other allied genera ; Moringua Gray, E. Ind., Fiji, Japan. Group 2. Eels in which the branchial openings in the pharynx are narrow slits. Muraena Gthr., scaleless ; teeth well developed ; pectoral fins absent, are as abundantly represented in tropical and sub-tropical waters as is Ophichthys ; more than eighty species ; most of them with formidable teeth, attain a length of 8 ft. and attack man, most are highly coloured. M. helena L., the muraena of the ancient Romans, can be domesticated, will live in fresh-water, Mediterranean, etc. ; other genera are Gymno- muraena, Myroconger, Enchelycore, Pythonichthys, Rabula, Lycodontis, Echidna Forster 1778, Uropterygius, Channomuraena. The family Saecopharyngidae may be placed here. They are eel-like deep-sea (Atlantic) fishes with feeble muscular system, but little earthy matter in their bones, and branchial arches far behind the skull, without palato-pterygoid bar, narrow tail ending in filament, and with pedunculated appendages in place of the lateral line. Sacco pharynx, Gastrostomus, Eury- pharynx. Sub-order 5. HAPLOMI (ESOCIFORMES). Pike-like fishes. Soft-rayed fishes with the mesocoracoid wanting, the cora- coids normally developed, and the post-temporal normally at- tached to the cranium. Parietal bones separated by the supra- occipital. Symplectic present, opercular bones well developed. Anterior vertebrae unmodified. Air-bladder with duct ; pelvic fins abdominal, rarely absent. First ray of dorsal fin occasion- ally stiffened and spine-like ; no adipose fin. Chiefly f . w. Fam. 24. Galaxiidae. Xaked, without barbels ; margin of upper jaw chiefly formed by premaxillaries. Dorsal fin opposite anal ; pseudo- branch absent. Without adipose fin ; with air-bladder. Ova dehisced into abdomen F. w. and seas of temperate parts of S. hemisphere (S. Afr., Patagonia, N. Zealand, Tasmania), some are katadromous ; Galaxias Cuv. Fam. 25. Haplochitonidae, representing the salmonoids in the S. hemisphere. Haplochiwn Jen. ; Prototroctes Gthr. Fam. 26. Enehodontidae. Extinct, Cretaceous. Enchodus Ag., etc. Fam. 27. Esocidae. Body covered with scales ; margin of upper jaw formed by premaxillaries and toothless maxillaries ; barbels and adipose fin absent ; unpaired fins far back ; stomach without blind sac ; pyloric caeca absent ; pseudobranch glandular hidden ; air-bladder simple ; gill-opening very wide ; noted for their voracity. Esox (L. ) Cuv. (Lucius Raf.), the pikes, f. w. of temp. Eur., Asia and Amer. z. — ii. Q 226 SUB-CLASS (AND ORDER) TELEOSTEI. E. lucius L., common pike, pickerel, jack, luce, hake, Eur., N. Asia and northern parts of North America ; extremely voracious, does not refuse frogs, voles, house rats, puppies, kittens, weasels, foxes, ducks, geese, has been found with a human infant in its stomach, has been known to lay hold of a swan, a tame cormorant, and to attack otters, dogs, asses, mules, oxen, horses, men, and to catch swallows, dislikes sticklebacks. Umbra Kramer, Austria, Hungary (Hundsfisch), and United States (mud-minnow). Fam. 28. Dalliidae, f. w. fishes from Alaska and Siberia. Fam. 29. Scopelidae. Naked or scaly. Margin of upper jaw formed by premaxillary only ; opercular bones thin but complete. Barbels none. Pseudobranchs usually well developed. Air-bladder small or none. Adipose fin present. The eggs are enclosed in the sacs of the ovary and are extruded by oviducts. Intestine short. All marine, mostly inhabiting shore waters, some descending to the deep sea. The following fossil forms are probably allied here : Hemisaurida, Parascopeltis, Anapterus. Saurus Cuv., Med., trop. Atl. and Pac. ; Bathysaurus Giinth., deep sea, Pac., 1,100-2,400 fms. ; Harpodon Les., Ind. and China Seas, H. nehereus, Bombay duck ; Scopelus Cuv., lantern-fishes, luminous spots along sides of body, pelagic fishes, taken at any depth to 2,500 fms. ; Ipnops Giinth., 1,600 to 2,150 fms., phosphorescent organs extending along the median line of the snout, have been regarded as modified eyes, which are otherwise absent, pseudobranch absent ; Paralepis Risso, small, pelagic, from Med. and Atl. ; Sudis Raf. ; Plagyodus Pall. (Alepidosaurus or Alepisaurus Lowe), one of the largest deep-sea fishes ; other genera are Aulopus, Chlorophihalmus, Scopelosaurus, Odontostomus, Nannobrachium, Bathypterois ; Trachino- cephalus Gill, Synodus Bloch and Sch., Benthosaurus Goode and Bean ; Myctophum Raf., pelagic fishes coming to surface at night, taken at any depth to 2,000 fms. Fam. 30. Cetomimidae. Rondeletia Goode and Bean, deep sea ; Ceto- mimus, Goode and Bean, deep sea. Fam. 30 a Chirothricidae. extinct. Fam. 31. Kneriidae. Small loach-like fishes from f. w. of trop. Africa. Kneria. Fam 32. Cyprinodontidae (Poeciliidae J. and E.). Head and body covered with scales ; barbels absent. Margin of upper jaw formed by premaxil- laries only. Teeth in both jaws ; upper and lower pharyngeals with cardiform teeth. Adipose fin absent ; dorsal fin on the hinder half of the body. Stomach without blind sac ; pyloric appendages absent. Pseudo- branch absent ; air-bladder simple. Sexes usually unlike, the fins being larger in the males, which however are often much smaller in size than the females ; mostly viviparous, the young being well developed at birth. The anal fin of the male is frequently modified as a copulatory organ. Freshwater fishes of S. Eur., Asia, Afr. and Amer., some of them occurring in arms of the sea. Some are carnivorous and some live on organic substances in mud. Fossil remains in tertiary strata. I. Carnivorae. Bones of each ramus of the mandible firmly united, intestine short or but little convoluted ; carnivorous. Cyprin- odon Lacep., in Mediterranean region and N. Amer., are able to live in brine pools, e.g., of Dead Sea and Sahara, and at high tem- peratures ; sometimes lose their ventral fins and then known as Tellia ; oviparous. Allied are Fitzroyia from Monte Video, and HETEROMI : 227 Characodon from Central Amer. Haphlochilus M'CL, E. Ind. trop. Afr., temp, and tropical Amer. Fundulus C. et V., killifish, abundant in New World, one in Spain and one in E. Afr. ; allied are the South American Limnurgus, Lucania, Rivulus and Cynolebias ; Orestias C. et V., East Peru and Bolivia, at an elevation of 13,000 to 14,000 ft. ; Jenynsia Gthr., Madonado ; Gambusia Poey, W. Indies, and S. Amer. ; allied are Pseudoxiphophorus and Belonesox of Cent. Amer. ; Andbleps Art., four-eyed fishes, iris with two pupils, swims with part of head out of water, trop. Amer. II. Limnophagae. Mandibular bones but loosely joined, intestine convoluted, sexes differentiated, mud-eating, trop. Amer. Poecilia, Mollienesia, Platypoecilus, Girardinus. Fam. 33. Amblyopsidae (Heteropygii). Head naked, body with very small scales, barbels absent. Villiform teeth in jaws and on palate. Adipose fin absent. Pelvic fins small or absent. Vent in front of pectorals. Stomach caecal ; pyloric caeca present. Pseudobranch absent (con- cealed). Fishes of small size living in the swamps and subterranean streams of the United States. Aniblyopsis De Kay, the blind fish of the Mammoth Cave of Kentucky, colourless, 5 inches, eyes and optic nerve very imperfect * ; viviparous ; allied species without pelvic fins are known as Typhlichthys Gerard. Chologaster Ag., with normal eyes and coloured ; swamps and entering caves. Fam. 34. Stephanoberycidae, deep sea. Fam. 35. Pereopsidae. F. w. of N. Amer. ; adipose fin present ; dorsal and anal with a few spines, pelvics abdominal with more than five soft rays, with a trace of pneumatic duct and with pseudo-branch ; body covered with ctenoid scales. Percopsis Ag., Columbia Eigenm. Sub-order 6. HETEROMI (DERCETIFORMES). Air-bladder without open duct ; parietals separating the frontals from the supraoccipital ; no mesocoracoid. Pelvics abdominal if present. Fam. 36. Dercetidae. Eel-shaped fishes without ordinary scales. Body generally with four series of sub triangular scutes and intermediate scale-like smaller scutes. Head long and jaws produced. Extinct, Cre- taceous. Dercetis Ag., Pelargorhynchus v. d. Marck. Fam. 37. Halosauridae, deep-sea forms. Halosaurus, Aldrovandia. Fam. 38. Notacanthidae, deep-sea, pelvics abdominal, air-bladder with duct. Notacanthus, Macdonaldia. Fam. 39. Lipogenyidae. Deep-sea. Fam. 40. Fierasferidae.f Without pelvic fins, vent at the throat ; eel-like, small, shore-fishes of tropical seas, often living as lodgers in cavities of other animals, e.g. Holothurians, starfishes and bivalve molluscs ; often commensal with the pearl oyster ; are harmless to their hosts. Fierasfer Cuv., Encheliophis. Lycodapus Gilbert may be placed near here. * Eigenmann, Arch. f. Entwick. Mech., 8, 1899, p. 545. t Emery, Fauna und Flora d. Golf. v. Neapel, 1880. 228 SUB-CLASS (AND ORDER) TELEOSTEL Sub-order 7. CATOSTEOMI (GASTROSTEIFORMES). Air-bladder, if present, without open duct. Parietals, if pre- sent, separated by supraoccipital. No mesocoracoid. Ventral fins abdominal if present. Mouth bordered by the premax- illaries or by them and a small portion of the FIO. 124;— Gasterosteus aculeatus (after Hockel and Kner, maxillarieS. A. SELENICHTHYES Preoperculum and symplectic distinct ; branchial apparatus fully developed ; mouth terminal, toothless ; post-temporal forked, free ; pelvic bones connected with the scapular arch ; pelvics with fifteen to seventeen rays ; ribs long, sessile ; fins without spines. Fam. 41. Lamprididae. Body short and deep, with minute scales. Lampris Retzius ; L. luna Gmelin, the opah or king-fish, to 4 ft., N. Atl. and Med. B. HEMIBRANCHH. Gills pectinate. Post- temporal furcate. Superior pharyngeal bones reduced in number, the bones of the gill-arches also reduced ex- cept in Gasterosteidae ; in- ferior pharyngeal bones pre- sent, not united. Pelvic fins abdominal. Mouth bounded above by premax- illaries only. Basis of cra- nium simple and without tube. Mouth small, at the end of the snout which is usually produced. Fam. 42. Gasterosteidae. Sticklebacks. Body elon- gate, compressed, cleft of the mouth oblique, villiform teeth in the jaws. Opercular bones not armed. Scales none, but generally large scutes along the side FIG. 125. — Nest of Gasterosteus punQitius (from Glaus after Landois). Isolated spines in CATOSTEOMI. 229 front of the soft dorsal fin. Pelvics abdominal, joined to the scapular arch. Branchiostegals 3. Pseudobranch and air-bladder present. Small fishes inhabiting f. w. and arms of the sea in Eur., As. and Amer. ; noted for their pugnacity ; they are very destructive to the spawn and fry of other fishes. In many species the males build nests for the eggs with blades of grass, etc., cemented together by cutaneous mucus ; the male defends the eggs. They are extremely variable and susceptible to change of conditions. Gasterosteus Artedi, probably only 3 Brit, species, though many varieties have been described as such, G. aculeatus, the 3-spined, f . w., G. pungitius, the 9-spined, f . w., and G. spinachia, the marine stickleback ; Eucalia Jordan, Pygosteus Brevoort, Apeltes De Kay ; Aulorhynchus Gill. Fam. 43. Protosyngnathidae. Extinct. Fam. 44. Fistulariidae. Gigantic marine sticklebacks, flute-mouths, pipe-fishes, trop. and sub-trop. Atl. and Indo-Pac. Fistu- laria, Aulostoma, Auliscops. Fam. 45. Maerorhamphosidae. Bones of the skull much prolonged anteriorly forming a long tube which bears the short jaws at its end ; two dorsal fins, the spinous short ; pelvics truly abdominal, imperfectly devel- oped ; the 4 anterior vertebrae much elon- gated. Macrorhamphosus Lac. (Centriscus Cuv.), snipe fishes, M. scolopax L., the trumpeter bellows-fish, rarely occurs on S. coast of England ; Amphisile Klein, body so thin as to be semi-transparent, trunk part of vertebral column composed of 6 verte- brae, and four times as long as the caudal, which consists of 14, with a dorsal cuirass formed by portions of the skeleton C. LOPHOBRANCHII. Fl