FOR THE PEOPLE FOR EDVCATION FOR SCIENCE LIBRARY OF THE AMERICAN MUSEUM OF NATURAL HISTORY ru p- <=o CD' a m CD a THE ANATOMY OF VERTEBRATES, VOL. I. Works by the same Author LECTURES on the COMPARATIVE ANATOMY and PHYSIOLOGY of the INVERTEBRATE ANIMALS, delivered at the Royal College of Surgeons. Second Edition (1855), illustrated by numerous "Woodcuts 8vo. 21s. On the CLASSIFICATION and GEOGRAPHICAL DIS- TRIBUTION of MAMMALIA, being the Lecture on Sir R. Reade's Foundation, delivered before the University of Cambridge, in the Senate House, May 1859. To which is added an Appendix on the Gorilla, and on the Extinction and Transmutation of Species. With 9 Figures on Wood 8vo. 55. INSTANCES of the POWER of GOD as manifested in His Animal Creation : a Lecture delivered before the Young Men's Christian Association, November 1863. With 11 Figures Crown 8vo. 1*. London : LONGMANS, GREEN, and CO. < c: ox Tin. fir ANATOMY OF VERTEBRATES. ^ -- VOL. I. " * " A v> * FISHES AND EEPTILES. BY RICHARD OWEN, F.R.S. RUPEItlSTKNDBNT OF THE NATURAL HISTORY DEPARTMENTS 0>" TIIK 1UUTISH MUSEUM. FOREIGN ASSOCIATE OF THE INSTITUTE OF FRANCE, ETC. LONDON : LONGMANS, GREEN, AND CO. 1866. LONDON BY SPOTTISVVOODB AX D CO. NEW-STREET SQUARE PREFACE. THE PRESENT WORK completes the outline of the Organisation of the Animal Kingdom which was begun in that on the Inverte- brates. 1 They may be regarded as parts of a whole, having the same general aim, and, together, form a condensed summary of the subjects of my ( Lectures on Comparative Anatomy and Physiology, according to the Classes of Animals,' delivered in the Theatre of the Royal College of Surgeons of England in the years 1852, 1853, and 1854. In the choice of facts, as then and since acquired by science, I have been guided by their authenticity and their applicability to general principles. In the first, regard has been had to the agreement of several observers, or to the nature of the fact as making it acceptable on the testimony of a single expert. Appearances that require helps to vision are those that call for multiplied concurring testimony, and on such alone are offered the descriptions and illustrations of the microscopical characters of ( tissues ' premised to most of the chapters. In the second aim, the parts and organs, severally the subjects of these chapters, are exemplified by instances selected with a view to guide or help to the power of apprehending the unity which underlies the diversity of animal structures ; to show in these structures the evidence of a predetermining Will, producing 1 Lectures on the Comparative Anatomy and Physiology of the Invertebrate Animals, 8vo. 1843; 2nd eel. 8vo. 18oo. vi PREFACE. them in reference to a final purpose ; and to indicate the direction and degrees in which organisation, in subserving such Will, rises from the general to the particular. Anatomy, or the ( Science of the Structure of Organised Bodies,' is primarily divided into ' Phytotomy,' or that of plants, and ' Zootomy,' or that of animals. When particular provinces, classes, or species of animals have monopolised the time and skill of anatomists, such special knowledges have received particular denominations : such as ( Malacotomy,' or anatomy of mollusca ; ' Entomotomy,' or anatomy of insects ; ( Ichthyotomy,' or anatomy of fishes ; ( Ornithotomy,' or anatomy of birds, &c. An animal may be dissected in order to a knowledge of its structure, absolutely, without reference to or comparison with any other, its species being regarded as standing alone in creation. The knowledge so gained, from the very limitation of the field of enquiry, may be most accurate and minute, most valuable in its application to the repair of accident, the remedy of injury and decay, and the cure of disease. Such, e.g., is f Anthropotomy,' or the anatomy of man, and ' Hippotomy,' or that of the horse. Besides the numerous and excellent works on these special sub- jects, I may cite the ' Traite Anatomique de la Chenille du Saule,' 4to., 1762, by LYONNET ; the ( Anatome Testudinis Europaeae,' fol., 1819, by BOJANUS; the ( Anatomic Descriptive du Melolon- tha vulgaris? 4to., 1828, and the ( Anatomic Descriptive du Chat,' 4to., 2 vols., by STRAUS-DURCKHEIM ; also, in application of the science to art, ' The Camel, its Anatomy, Proportions, &c.,' fol., 1865, by ELIJAH WALTON ; as unsurpassed examples of this monographical kind of anatomical science. As applied to Man it is commonly called ' Human Anatomy,' and is, in strictness of speech, one of the manifold ways of human work. But the anatomist may apply himself to a particular organ instead of a particular species, either exhaustively in one animal, or by tracing such organ or system throughout the animal king- dom. The ( Neurotomies ' and ' Neurographies ' to which JOSEPH SWAN, e.g., has devoted a laborious life, the ' Osteographie ' of PREFACE. vii DE BLAINVILLE, and my own ' Odontography/ are examples of this way of anatomy. JOHJST HUNTER assembled the evidences of his labours, in the unique and grand department of his Museum illustrative of anatomy properly so called, in series according to the organ, beginning with the simplest form, followed in succession by the progressively more complex conditions of the same organ, the series culminating, in most cases, with that which exists in the human frame. The mechanism of the organ is here unfolded, and its gradations were compared, to discover its mode of work- ing ; and, as ( Physiology ' mainly consists in such determinations of functions or final aim, this kind of investigation of organic structures might be termed ' Physiological Anatomy.' 1 ( Homological Anatomy ' seeks in the characters of an organ and part those, chiefly of relative position and connections, that guide to a conclusion manifested by applying the same name to such part or organ, so far as the determination of the namesakeism or homologv has been carried out in the animal kingdom. This o./ o aim of anatomy concerns itself little, if at all, with function, and has led to generalisations of high import, beyond the reach of one who rests on final causes. It has been termed, grandiloquently, ' Transcendental ' and ( Philosophical ;' but every kind of anatomy ought to be so pursued as to deserve the latter epithet. A fourth way of anatomy is that which takes a particular species in the course of individual development, from the impreg- nated ovum, tracing each organ step by step in its evolution up to the adult condition. It is called ' Embryology ' and ' Develop- mental Anatomy.' A fifth way of anatomy is that which investigates the structure of an animal in its totality, with the view of learning how the form or state of one part or organ is necessitated by its functional connections with another, and how the co-ordination of organs is adapted to the habits and sphere of life of the species ; but does 1 See ' Descriptive and Illustrated Catalogue of the Physiological Series of Com- parative Anatomy in the Museum of the Royal College of Surgeons,' 4to. 5 vols. 1832-1840; 2nd ed. vol. i. 18.V2. viii T'REFACE. not stop here, having for its main end the comparison of these associated modifications and iiiterdependencies of organs in all the species of animals. As their degrees of affinity and the characters and circumscription of natural groups are hereby illustrated, this way may be termed ' Zoological Anatomy.' In the hands of the anatomist the microscope has been mainly applied to the constituent parts of an organ, called * tissues ;' and the results of such research, combined with those of chemical tests, constitute a sixth sort of anatomy called ' Histology.' It has been termed ( Microscopical Anatomy,' but this is essentially only a more refined method of the scrutiny of organic parts. In so far, however, as ( Histology ' treats of structure according to the proximate tissues common to different organs, it corresponds with the branch of the science which BICHAT, its founder, called, loosely, e Anatomie Generale.' l Finally, a seventh way in which the highest generalisations in biological science may be aimed at is that which is taken when we pursue investigations of form and structure beyond the animals that are to those that have been. Here, however, the anatomist is limited, as a rule, to such tissues and organs as are petrifiable, e.g., corals, shells, crusts, scales, scutes, bones, and teeth ; but he has been stimulated to a degree of minuteness and accuracy of observation in this field of research to which few of the other ways and aims would have led him. In applying the results of such researches to the restoration of extinct species, physiology has benefited by the study of the relations of structure to function requisite to obtain an insight into the food, habits, and sphere of life of such species ; and zoology has gained an immense accession of subjects through such determinations, with improved systems of classification due to the expanded survey of organic nature opened out by ' Paleontology.' The word ' Anatomy ' is still commonly used to signify ( Anthro- potomy,' or ( Human Anatomy.' Almost all begin the study of the science, as medical students, with the dissection of the human 1 Anatomie Generale, appliqiu'e a la Physiologie, &c., Svo. 1801. PREFACE. ix body, and most end there ; but no special anatomy can be rightly and fully understood save on the basis of the general science of which it is an integral part. The reason lies in the diversities of organic structure being subordinated to a principle of unity. Of this principle, apprehended as an ( idea ' or truth of reason, the understanding receives evidences in number and comprehen- sibility differing in different natural groups of the animal king- dom. Illustrations of the ' idea ' will be found in the chapters on the Articulate Province and other parts of the ' Lectures on In- vertebrates,' and, in accordance with the present phase of ana- tomical science, more abundantly in the present Work. True it is that in the first steps to organisation we seem to see a tendency to disintegration, to a reduction of the primary whole to the sub- ordinate characters of a part. The first centre of sarcode, or undifferenced organic matter, however originated, yet with de- finite tendencies to formal character and course of growth (as in a Foraminifer, e.g.), buds forth a second centre of identical nature; this a third, and so on, until a group of such exists as an assem- blage of coherent homogeneous or like parts. These, if clothed by a delicate crust of characteristic structure, constitute a chambered shell, straight, bent, or spiral, each chamber occupied by the same vital sarcode, outshooting filamentary food-getting processes through the shell-pores ; in which seeming complexity the inci- pient unity or ( whole ' is reduced to the ' part ' called innermost chamber, or is conceivable as a lesser whole in the larger one. The Annelides offer a familiar example of such repetitions of a primal complexly organised whole, by successive buddings in a linear direction ; the nerve-centre, the muscles, the skeleton-seg- ment, perhaps heart and gills, being regularly repeated in each, and thereby reducing the original whole to one of many parts of a segmented unity. Almost every organ in the progressively differenced organism initiates itself under a similar character of irrelative repetition, suggestive of operance akin to that of inorganic polar growths, as in a group of crystals, wherein each exemplifies the characters x PREFACE. of the mineral or crystalline species, but is subject, like vital growths, to occasional malformation. As this principle of growth by multiplication of like parts is manifested more commonly and extensively in plants, it is illus- trated in the ( Invertebrate Volume ' 1 under the term ( Vegetative o Repetition.' In the vertebrate series it is exemplified by the hundreds of similar teeth in the jaws of many of that low class (Pisces^ in which true dentinal teeth first appear in the animal kingdom. The numerous and similar many-jointed terminal divi- sions of the pectoral limbs of the fishes thence called e Kays,' the multiplied similar endoskeletal segments of the vertebral column of these and other cartilaginous fishes, of murasnoids and serpents, are likewise lingering exemplifications of the low irrelative principle of development. In the vertebral embryo the first appearance of the parts of the skeleton, in gristle or bone, is a segmental one ; in fishes the mus- cular system shows much, and in all Vertebrates a little, of the like segmental constitution of the trunk ; the same idea is suggested by the symmetrical and parial origins of the nerves, and phy- siologists have mentally recognised a corresponding segmental condition of the myelon or spinal chord, which is visibly exem- plified in certain fishes. But these appearances are concealed by the general tegument ; not exposed, as in the Articulates, in which the segmented skeleton is at the same time tegument. A Verte- brate may be defined as a clothed sum of segments. But in this highest province of the animal kingdom growth by repetition of parts rapidly gives place to the higher mode of development by their differentiation and correlation for definite acts and complex functions. Nevertheless, I am constrained by evidence to affirm that in the vertebrate as in the invertebrate series there is mani- fested a principle of development through polar relations, work- ing by repetition of act and by multiplication of like parts, con- trolled by an opposite tendency to diversify the construction and enrich it with all possible forms, proportions, and modifications of 1 Op. cit. 2nd ed. p. 541. PREFACE. xi parts, conducive to the fulfilment of a pre-ordained purpose and a final aim : these opposite yet reciprocally complemental factors co-operating to the ultimate result, with different degrees of dis- turbance, yet without destruction, of the evidences of the typical unity. 1 Thus, the dentition of Vertebrates will be seen to pass from irrelative sameness and multitude to the state in which the teeth in the same jaw are classed according to diversities of form and function, and where each tooth has its own character, bears its own name and symbol. In like manner may be traced the gradations by which the terminal divisions of the limb ascend from the multitude of many jointed rays, swathed in a common sheath of integument, to indi- vidual freedom, with reduction of number and of joints, and with a special form and action ; according to which each digit in the human hand, e.g., has its special name and symbol, and can be combined in action with any other digit for a particular purpose. The same principle, through reduction of number with differen- cing of the parts, is exemplified by the fact that the competent anthropotomist will distinguish and name each of the four and twenty ' true vertebra' of the human skeleton. In the Mammalian class there are four muscular pulsatile cavities concerned in the propulsion of blood ; but they differ from those cavities in the Annelide, in each having its own special structure and powers, and being in such relation with another cavity that the whole can combine to effect two complete but mutually related systems of circulation, the four pulsatile cavities constituting one complex and perfect ' heart.' The ox has four bags for the digestion of food ; but they differ from those cavities of the Polyyastria, not only in their minor number and more de- finite structure as bags, but by each performing a distinct part in the process of digestion, and combining with the rest, in mutual 1 This idea will be found more fully exemplified in my work, ' Principes d'Osteolo- gie Comparee,' Sro. (Paris) 1855, p. 366, et seq. xii PREFACE. subserviency, to tlie completion of the most perfect act of that function, the conversion, namely, of grass into flesh. Thus, in tracing through the animal series this course of parts and organs, we pass from the many and the like to the few and the diverse. A * homologue ' is a part or organ in one organism so answer- ing to that in another as to require the same name. Prior to 1843 the term had been in use, but vaguely or wrongly. 1 ( Analogue ' and 'analogy' were more commonly current in anatomical works to signify what is now definitely meant by ( homology.' But f analogy ' strictly signifies the resemblance of two things in their relation to a third ; it im- plies a likeness of ratios. An ( analogue ' is a part or organ in one animal which has the same function as a part or organ in another animal. A ' homologue ' is the same part or organ in different animals under every variety of form and function. In the Draco volans (Vol. I. fig. 163) the fore-limbs are ' homologous' with the wings of the bird (Vol. II. fig. 1); the parachute is ( analogous ' to them. Relations of homolosry are of three kinds ; the first is that o/ above defined. When the ( basilar process of the occipital bone ' in Man is shown to answer to the distinct bone called f basioc- cipital ' in the fish, the special liomology of that anthropotomical process is determined ; as such homologies are multiplied, the evidence grows that man and fish are constructed on a common type. A wider relation of homology is that in which a part or series of parts stands to such general type. When the ( basilar process of the occipital bone ' is determined to be the l centrum ' of the last cranial vertebra, its general liomology is enunciated. The archetype skeleton represents the idea of a series of essentially similar segments succeeding each other in the axis 1 ' Les organes des sens sont Jwmologues, comme s'exprimerait la philosopliie allemaude ; c'cst-a-dire, qu'ils sont analogues dans leur mode de deYeloppement.'- Geoffroy St. Hilaire, Annales des Sciences Nat., torn. xii. 182o, p. 34-1. PKEFACE. xiii of the body ; such segments being composed of parts similar in number and arrangement. Accordingly, a given part or appen- dage in one segment is repeated in another, just as one bone is represented in the skeletons of different Vertebrates ; and this representative relation in the segments of the same skeleton is ( serial homology.' As, however, the parts can be namesakes only in a general sense, as e centrums,' ' ribs,' &c., and since they must be distinguished by special names according to their special modifications, as tf basioccipital,' ' mandible,' f coracoid,' ' humerus,' c., I have called such serially related or repeated parts ( homo- types.' The basioccipital is the homotype of the basisphenoid, and the humerus is the homotype of the femur : when the basi- occipital is shown to repeat in its f vertebra ' the element which the ( odontoid process ' represents in the succeeding vertebra, or the basisphenoid in the preceding one, its f serial homology ' is indicated. The extent to which serial homologies can be determined shows the degree in which vegetative repetition prevails in the organisation of an animal. The study of homologies is comparatively recent ; much of this field of research remains for future cultivators, especially in regard to the muscular and nervous systems. When engaged in the f third way ' of anatomy, and in making known the results of such labour as applied to the skeleton, 1 I found a great impediment in the want of names of bones. For these, when first studied, had been mostly described under phrases suggested by forms, proportions, or likeness to some familiar object, which they present in the human body. A reform of this nomenclature was an essential first step, and it is gradually making its way against the usual impediments. The best workman uses the best tools. Terms are the tools of the teacher ; and only an inferior hand persists in toiling with a clumsy instrument when a better one lies within his reach. But * he has been used to the other.' No doubt ; and some extra 1 On the Archetype and Homologies of the Vertebrate Skeleton, 8vo. 1848; and On the Nature of Limbs, 8vo. 1849. xiv PREFACE. practice is necessary to acquire the knack of applying the new tool. But in this acquisition a small capital of trouble will have been invested with a sure return of large profits. A single sub- stantive term is a better instrument of thought than a paraphrase. 1 But the substitution of such terms for definitions is still more advantageous when they are susceptible of becoming adjectives by inflection. Thus the term ( notochord ' for ( chorda dorsalis ' or ' dorsal chord ' enables one to predicate of species or groups of vertebrates as being ' notochordal ; ' that single epithet implying that the embryonal body in question is, in them, persistent. A like advantage cleaves to ( myelon ' for ( chorda spinalis ' or ' spinal chord ;' the Physiologist, e.g., can then speak of f myelonal functions,' and the Pathologist of ' myelonal' disease, with the cer- tainty of being understood to signify properties and affections of the s spinal chord ;' not, as in ' spinal disease,' that of its case, or e spinal column/ In regard to the part so called and its con- stituent ' vertebras,' their modifications are so many, so charac- teristic, so important, especially in the application of Anatomy to Paleontology, that I was early compelled in the latter kind of labour to substitute single pliable terms for the phrases ( trans- verse process,' ( oblique ' or ' articular process,' 6 body of the vertebra,' ' vertebral lamina,' ( vertebral rib,' ' sternal rib,' &c., by which the parts of the ' vertebra ' were then designated. But the single names of parts and constituents of the skeletal segment called ' osteocomma ' or f vertebra ' have not merely the advantage above illustrated, as where the adjective ' neurapo- physial ' can be applied to a f ridge,' notch, or ( foramen,' in the vertebral lamina (neurapophysis) ; the vertebral terminology in use in the present Work indicates a profound truth which is hidden by the language of anthropotomy. The terms ( pleur- apophysis' and ( hajmapophy sis ' imply parts of the segment corre- 1 ' Superoccipital,' e.g., for ' pars occipitalis stride sic dicta partis occipitalis ossis spheno-occipitalis ' of the eminent anthropotomist SOEMMERING. (See TABLE OF SYNONYMS, &c., appended to Vol. II.) Similarly, in the present Work, I use the word { Vertebrate ' as a substantive. We do not speak of a ' Confederate ' animal, and the added word is as unnecessary in regard to the ' Vertebrate ' one. TREFACE. xv lative in independency of development and elemental grade with the ( neurapophysis,'- -a fact of high generalisation not only ignored but impliedly contradicted bv the reckoning of the O J. / */ O ' vertebral rib ' and the ' sternal rib,' or ' rib-cartilage,' as bones distinct from, and countable with, that which the anthropotomist equally holds to be a single bone under the name e vertebra.' Furthermore, as each distinctly recognisable part or thing must have its verbal sign, for the purposes of intelligible predication of its nature and qualities, the course of knowledge of the vertebral column would have enforced the origination of such signs irre- o o spective of the abstract need of improving the mental tools of anatomy. When it came to be discovered that ( the transverse process of a cervical vertebra ' was other, and more than, as well as formally different from, the ' transverse process of a dorsal ver- tebra,' and that this process was a different thing from the ' transverse process ' of a l lumbar ' or ( sacral ' vertebra, the re- sults of such analysis necessitated the creation of a correspondent nomenclature. ' Transverse processes,' as such, are, as JOHANNES MULLER first pointed out, of two kinds ; they are, in relation to hori- zontally disposed vertebrates, f upper ' and ' lower ' -in our no- menclature, ( diapophyses ' and f parapophyses.' Both kinds exist in the ' transverse processes ' of the neck from the crocodile upwards ; and the seeming unity of the outstanding part in birds and mammals is caused by the soldering thereto of a third element the ' cervical rib ' of the herpetotomist, the ( styloid process ' of the ornithotomist. ] Referring to the ( Introductory Chapter ' of the ( Archetype of the Vertebrate Skeleton,' 2 for further illustrations of the advantage of single well-defined terms, I will here only show how such advantage may be affected by reason of an unsettled definition. The anatomical term { organ ' has diverse significations. The Macartney. Art. ' Birds,' Rees' Cyclopaedia. * Op. cit. xvi PREFACE. chief constituent idea is ' work for a special end :' thus, the heart is the ' organ ' of circulation ; the lungs, the ( organ ' of respira- tion ; the liver, the c organ' of bilification, &c. But also, incipient stages in the development or formation of parts are called the 6 organs ' of such; e.g., the periosteum is the ' organ of bone,' the pulp is the ' dentine organ;' other parts of the growing complex tooth are the ( enamel organ,' f cement organ,' &c. The parts in which independent cells, with special powers, originate, are also called the ( organs' of such; as, e.g., the ovary is the ' organ of ovulation ;' the testis the ' organ of semination.' It is obvious, however, that the part which the more or less con- densed cellular basis, or ( stroma,' of the ovary or the testis may take in the production of the gerrn-cells or sperm-cells and sperma- tozoa is very different from that which the heart performs in the motion of the blood, or the lungs in the mutation of the air inspired. Zoological anatomy is now an indispensable instrument to the classifier, if not to the determiner, of the species of animals. The anatomist properly so called, but commonly qualified as the f comparative ' one, makes known the results and applications of his comparisons of structure in zoological as well as homologi- cal or anatomical works. The ( Regne Animal ' and the f Lepons d' Anatomic Comparee ' of CUVIER exemplify these different appli- cations and ways of exposition of his science. As a zoologist or classifier, the anatomist avails himself of the definite modification and full development of a part or organ, in- dicating, and predicating of such conditions by special terms, for the required characters. The ( fin,' the ( hoof,' the ( paw,' the 6 foot,' the ' hand,' are to him so many kinds of limbs, the presence or absence of which serve to differentiate his groups ; anthropo- tomical terms of parts of the brain reaching their full and cha- racteristic development in Mammals or in Man, e.g., f fornix,' 6 corpus callosum,' ( hippocampus minor,' l posterior cerebral lobe,' &c., serve and are used, absolutely, for the same end ; so likewise PREFACE. xvii with regard to special forms and proportions of teeth indicated by the terms ' canine/ f carnassial,' ' tusk,' &c. The absolute way in which the things or characters so desig- nated are affirmed or denied in zoological definitions is essential to their purpose. Amongst the characters by which CUVIER differentiated the hoofed quadrupeds which he had restored from their frag- mentary fossil remains in the building-stone of Paris, the most important in his estimation was 'the presence of canines' in one (Paloeotheriwn), their absence in the other (Anoplotherium). 1 Nevertheless, Homological Anatomy easily indicates in the series of nine teeth in the ( morceau de conviction,' 2 on which the character was founded, the teeth answerable to those which, because their pointed crowns projected beyond their neighbours in the Palasothere, were called and characterised as ' canines.' Now here was a temptation to an aspirant to scientific notoriety ( to meet ' the great anatomist ( by a flat contradiction,' and 'affirm that the Anoplotherium possessed canine teeth.' I allude to such abuse because, of late, a practice has been creeping in, to the opprobrium of some of our English zootomists, of repre- senting a zoological definition of a part which an anatomist may have given in a classificatory work, as the exponent of his homo- logical knowledge and descriptions of such part in its various modifications and grades of development. CUVIER, in his characters of the order Bimana, affirms that Man is the only animal possessing ( hands ' and ( feet : ' ( L'homme est le seul animal vraiment Mmane et bipede.' 3 The Quadruma?ia are distinguished as having ' hands' instead of ' feet,' a ' hand' being defined as having the thumb opposable ' le pouce libre et opposable aux autres doigts, qui sont longs et flexibles.' 4 The aim of the author in the zoological work above cited was to impart obvious and easily apprehended differential characters 1 ' Le plus important fut celui qui m'apprit que cette espece n'a point de dents canines.' Rechcrches sur les Ossemens Fossiles, 4to. 1822, torn. iii. p. 14. 2 Ibid. 3 Regne Animal, torn. i. p. 70. 1829. 4 Ibid. p. 85. VOL. I. a xviii PREFACE. of the organ which observation had shown to define the groups. The naturalist, thus enabled to place his subject in its proper class or order, is not concerned, as such, in knowing the homo- logical or transcendental relations of the part or character which has afforded him the means of effecting what he wished to do. LINN/EUS, to whom mainly is due the discernment of the power- ful instrument of well-defined terms in acquiring a systematic Science of Nature, and to whom we owe our best knowledge of its use, so named the guiding parts of plants and animals, for such arbitrary or special application, in botany and zoology : to this end he differentiates the ( bract,' the ' spath,' the ( sepal,' the { petal,' from the ' leaf,' as things distinct. What would be thought of the botanical critic who, quoting the definition of the flowers of Cyperaceous plants, as consisting, for example, of s glumes,' should meet the statement by affirming that they were ( nothing but little bracts,' and who, then, with a show of profounder research, should proceed to expound the ' bract ' as being the first step by which the common leaf is changed into a floral organ ? The answer is obvious. But what next might be said, if it were pointed out that the objector had obtained this very notion from the ( Prolepsis Plantarum,' or other homological writings of the author criticised, where such philosophical con- siderations, foreign to the classificatory work, were the proper aim and object ? So, with regard to the zoological definitions and characters of CUVIER. Those which I have cited are open to the opposite averment that, ' The " hind hands " of the Quadrumana are nothing but " feet ; '' ' and the contradictor might then proceed to demonstrate, with much show of original research, the homology of the ' astragalus,' ' calcaneum,' f cuboides,' ( cuneiform bones,' &c., in order to establish his discovery that a hand and foot are all one. It is true that if the homological descriptions in the ' Lemons d' Anatomic Comparee ' had been quoted, as well as the zoological definitions in the e E-egne Animal,' the immortal author of the latter work would be shown to have had previous possession of the pretended discovery. Moreover, in the ' Cinquieme Lepon, PREFACE. xix Articles VJI.-IX. " Des os clu pied," ' l the frame of the hind feet of Man, Ape, Lion, Seal, Elephant, &c., is shown to consist of homologous bones. Nevertheless the great zootomist, in his labour and character as zoologist, does not hesitate to define and differentiate the ' foot,' the ' hand,' the ' paw,' the ' fin,' and the ( hoof,' respectively : nor does he deem the demonstration of the unity underlying the diversity to make the f man ' an f elephant ' or a ' seal,' any more than it makes him a f dog ' or an ' ape ' ! The ' corpus callosum ' is defined as ' a horizontal mass of trans- verse fibres covering the lateral ventricles, and exposed by divari- cating the cerebral hemispheres.' If a group of mammals want such commissural fibres, and another group possess them, the classifier will avail himself of a well-defined term expressing such difference, without prejudice to his reception of any homological determination of the parts, or their rudiments, 2 in anatomical works of the applier of the term. Only by ignoring such indication of the e rudiniental com- mencement of the corpus callosum,' may a semblance of superior knowledge be assumed by him who asserts, as an antagonistic proposition to an affirmation of its absence as a zoological cha- racter, that the Marsupialia, e.g., do possess the ( great com- missure,' or ( corpus callosum.' 3 So likewise with other well-defined parts of the human brain, the homologues of which may not be traceable to the same extent QJ V down the mammalian series. KUHL, e.g., in Ateles Belzebutli* TIEDEMAXN in the Macaque 5 and Orang, 6 VAN DER KOLK and VROLIK in the Chimpanzee, 7 and myself in the Gorilla, 8 had 1 Le9ons d'Anat. Comparee, vol. i. 1799. 2 As given in the 'Philosophical Transactions' for 1837, p. 41. 3 Proceedings of the Royal Society, No. 72, and March 23, 1865. 4 Beitrage zur Zoologie tind vergleichenden Anatomic, 4to. 1820, zweite Abtheilung, p. 70, Taf. vii. 5 Icones cerebri Simiamm, fol. 1821, p. 14, fig. iii. 2. 6 Treviranns, Zeitschrift fiir Physiologie, Bd. ii. S. 25, Taf. iv. 7 Nieuwe Yerhandliugeu der erste Klasse van het Koningl. Nederlandsche Instituut. Amsterdam, 1849. 8 Fullerian Lectures on Physiology, Royal Institution (March 18, 1861); reported, with copies of diagrams, in ' Athenpeum/ March 23rd, 1861, p. 395. a 2 xx PREFACE. severally shown all the homologous parts of the human cerebral organ to exist, under modified forms and grades of development, in Quadrumana. But because the presence or absence of the ( ergot,' or ( pes hippocampi minor,' as defined by TIEDEMAXN (see Vol. II. p. 273 of the present Work), had been used as a zoological character, the anatomical world has been deluged, since the date of the last under-cited work, with descriptions and figures of the homologous part in the Orang and other Quadrumana, as a new discovery mainly serviceable as a battery of contradictory affirmations. Nevertheless the distinctive characters of the human brain, &uch as the manifold and complex convolutions of the cerebral hemi- spheres, their extension in advance of the olfactory lobes and farther back than the cerebellum, thereby defining a posterior lobe, with the corresponding ( horn of the lateral ventricle ' and ' hippocampus minor,' are as available to the zoologist in classifi- cation as are the equally peculiar and distinctive characters of the calcaneum, hallux, and other structures of the foot. So much, in connection with the ( fifth way ' and application of anatomy, I regret to find myself compelled to state, in order to expose and stigmatise procedures which consist in representing the homological knowledge and opinions of an author by his de- finitions in a purely zoological work, and in suppressing all re- ference to the descriptions and statements in the anatomical writings of the same author, where his actual knowledge and o O opinions on the nature and homology of parts are given, and where alone they can be expected to be found. Somewhat analogous to the course of observation pursued through the animal kingdom, from the lowest to the highest O O o species, is that which traces each organ through the several phases of its development in the same species. The right use of sense, in both ways, stores the understanding, empirically, with a series of facts, as the raw material for reasoning up to their principles. But Embryology has this inferiority, that PREFACE. xxi every species is such db initio, and takes its own course to the full manifestation of its specific characters, agreeably with the nature originally impressed upon the germ. A perch, a newt, a dog, a man, does not begin to be such only when the embryologist may discern the dawnings of their respect- ive specific characters. The embryo derived its nature, and the potency of self-development according to the specific pattern, from the moment of the impregnation ; and each step of development moves to that consummation as its end and aim. This truth has been masked to some apprehensions by the course of the developmental steps from the general to the par- ticular ; the initial ones, more especially, offering likenesses or analogies to finished lower species exemplifying degrees of organi- sation in the animal kingdom. Each step differs in degree of difference from the analogous grade at which a lower species rests, and inversely as the advance of such species. Accordingly, the less the degree of difference, and the wider the resemblance or analogy spreads between the embryonal phase and the parallel grade in the series of species. The formation of the germ-mass (Vol. I. figs. 1-4, 422,452) the first step after impregnation is a general phenomenon in animality (Vol. ' On Invertebrates,' figs. 48-56, 73, 74, 80-84, 181, 209-212, 232) ; thereat and thereby the man resembles and behaves like the monad. 1 But, the germ-mass completed, the vertebrate at once circumscribes itself or withdraws into its vertebrality. The proteine substance is the seat of a chemical differencing, leading to excess of albumen along one tract, balanced by excess of gelatine along a parallel tract. Thus are laid down the bases of the myelencephalon and vertebral axis. The s notochord ' is soon followed by the protovertebral specks in double parallel series (Vol. I. fig. 5; Vol.11, fig. 133): the embryonal trace is established, and it is one of a vertebrate. The formation of neural and hnsmal arches next follows ; and 1 Compare the above-cited figures with fig. 17, 'Lectures on Invertebrates,' 2nd *d. p. 29. xxii PREFACE. the phenomenon of the appearance of the latter, in which the blastemal is accompanied by a vascular arch, with clefts inter- vening between contiguous arches, especially at the fore part of the embryo, has led to the idea that a reptile, bird, or mammal, is a fish before it becomes what it is tending to. True it is, that o the embryos of these air-breathers float in fluid, and not any of them breathe the air until birth or exclusion, or near to exclusion ; but they do not breathe water : the oviparous air-breather has one kind of temporary lung, the mammiferous embryo another kind, each alike special to the class. From the vascular loops accom- panying the haemal arteries branchiae are not developed ; one only of the interhaemal fissures is deepened on each side, brought into communication with the pharynx, and straightway converted into the ( eustachian tube,' according to the precocious rate of growth and development characteristic of the special organs of sense and their appendages. No true branchial or piscine breathing apparatus is at any time, or in any degree, manifested in the embryo of an air-breathing vertebrate. The deepening and open- ing of several interhremal fissures in the embryo of a perch, and the subsequent course of development therewith of gill-arches arid gills, with their subservient mechanism of branchiostegal rays and the opercular lid or door, are as distinctive manifestations of the original nature of the fish, as is the vascular lining; of the egs: that O ' O oO of the bird, or the vascular arrangement for borrowing breath from the mother that of the foetal mammal. At the incipient stages of these provisional and deciduous respiratory conditions the circulation in the embryo lizard, fowl, beast, is like that of a fish in its simplicity ; but, as TREVIRANUS 1 rightly remarked, it is far from being identical ; there are, indeed, characters of the circulating organs at this grade of simplicity, which not only distinguish the embryo of the air-breather from / O ' that of the water-breather, but also the embryo of the mammal from that of the bird or reptile ; so soon is the course of deve- lopment affected by the specific taint ! 1 Gr. R. in ' Zeitschrift fiir Physiologie/ vol. iv. ; and 'Edinburgh New Philosophical Journal.' ]83'2. vol. xiii. p. 75-86. PREFACE. xxiii Marked deviations from the archetype characterising existing species are directly approached in the progress of development. If, as, e.g., in a thoracic or jugular fish, the position of the pelvic limbs departs from the typical one, these limbs bud out in the embryo in that special and anomalous place. When a higher species departs from type by a thoracic position of the scapular or occipital limbs, they likewise bud out in such special position. In both cases the haemal arch, sustaining such appendages, is libe- rated from the rest of its segment for the special needs of the species, and the embryo of such never shows it fixed. At most, perhaps, the general character and typical connections may be indicated by the closer contiguity of the detached scapular arch to the rest of its proper occipital segment; as, e.g., in the embryo of birds and long-necked ruminants, to be removed to a distance determined by the later growth of the series of vertebras inter- vening between head and chest. To infer from such developmental phenomena that the throat- fins of the cod are not the displaced homologues of the hind legs or pelvic limbs of air-breathers, and that the fore-legs of such are not the homologues of the typically situated and connected scapu- lar limbs of fishes, is an abuse or misuse of the empirical facts ascertained by observation of embryonal phenomena. V V In like manner the developmental phenomena of the skull of an avian and mammalian species, succeeding those that broadly and intelligibly mark out the four pairs of neurapophyses and corresponding haemal arches, plainly indicating the seginental or vertebral type of the skull, depart therefrom to attain the par- ticular character of the face and mouth of the species. After the first budding indications of the halves of the maxillary (fore- most cranial haemal) arch, the development of it, as upper jaw, with that of the palate, pterygoid, and zygomatic appendages, obeys the impress of impregnation, and proceeds directly to es- tablish the specific characters of such jaw in the particular bird or beast ; the points of ossification, their deposit in membrane or gristle, and subsequent growth, having no other or deeper signifi- xxiv PREFACE. cation. If a species be gifted with acute hearing, and the move- ments of the ear-drum require several ossicles, these, like the labyrinth, grow to full size in the embryo, appropriating the blastema of the contiguous hremal arch, and proportionally re- ducing, by arresting the development of, the pleurapophysis of such arch. The inherited tendency to a special or specific form which thus influences early developments and growth of parts has misled some who have mistaken such for homological or archetypal characters. But the determination of these characters is arrived at by other routes of research ; and, so reached, such determination serves to explain many of the phenomena of development which otherwise would remain as mere empirical facts. Embryology, e.g., shows that in the human foetus the sternum is developed from a series of ossific centres (Vol. II. p. 555, fig. 364), whilst the co-articulated clavicle as long a bone is de- veloped from a single ossific centre, and a contiguous rib, though of greater length, is also hardened from a single ossific centre ; but embryology affords no explanation of the reason of such dif- ference. That is afforded by a knowledge of the archetype skeleton, which teaches that the sternum reckoned as a single o bone in anthropotomy consists of a series of vertebral elements, but that the rib and the clavicle are single elements. Embryology shows that the canon-bone of a ruminant, re- garded as a single bone by the veterinarian, is developed from five ossific centres ; two on the same transverse line near the middle, one on the upper, and a pair which soon coalesce at the lower end. But no clue is afforded to the signification of these several cen- tres : embryology is no criterion of their homologies ; these are determinable on other grounds or f ways of anatomy.' A knowledge of the f Nature of Limbs,' derived from homolo- o-ical studies leading to a recognition of the archetype, could alone determine that two only out of those five centres represent dis- tinct bones in the typical pentadactyle foot of the mammal; the rest having no such signification, but serving to perfect the ulti- mate growth as ' epiphyses.' So likewise with the collar-bone PREFACE, xxv and rib. At a period long subsequent to the deposition of the first centre of bone, a second appears at the sternal end of the human clavicle, and two are added to complete the head and tubercle of the lib, the shaft of which had been ossified by growth from a single centre. Recognition of the archetype skeleton elucidates the empirical facts of embryology, and teaches us to distinguish between the points of ossification of a bone in a higher vertebrate which sig- nify or answer to bones that retain their distinctness in lower vertebrates, and the points of ossification which merely help out the growth or have their final purpose in the exigencies of the vouno- animal. A lamb or foal, e.o\, can stand on its fore legs J O shortly after it is born, and soon begins to run and bound. The / shock to the limbs themselves is broken at this tender a;e by the c^ > cushions of cartilage at the ends of the shafts, and which continue for some time to be interposed between the ' epiphyses' and * dia- physis.' The jar that might affect the large and pulpy brain 01 the immature man is similarly diffused and intercepted by the ' epiphysial ' extremities of the vertebral centrums. Such final purpose in the several centres of ossification of the vertebral bodies and the long bones of the limbs of mammals does not apply to those of reptiles ; and no epiphyses with interposed cartilage attend the growth of the limb-bones of saurians and tortoises. But, when the reptile moves by leaps, ossification of the long limb-bones by distinct centres again prevails; the ex- tremities of the humeri and femora are ' epiphyses ' in the frog. Embryology affords no criterion between the ossific centres that have a ( homological ' and those that have a ' ideological ' signifi- cation. A knowledge of the archetype skeleton is requisite to teach how many and which of the separate centres that appear and coalesce in the human, mammalian, or avian skeleton, re- present and are to be reckoned as distinct bones, or elements of the archetype vertebra. For the want of this guide great and estimable anatomists have gone astray. Thus CUVIER, comment - - on the arbitrary enumeration of the single bones in the human o xxvi PFxEFACE. skeleton, affirmed that to learn their true number in any given species we must go to the first osseous centres as these are manifested in the foetus ; ! and GEOFFROY ST. HiLAiRE 2 concurred in this view. In the cartilages called ' epiphysial,' that eke out the ends and margins of bones, ossification begins later than does that of the bone itself. The times of appearance of the osseous nucleus in the coracoid process and acromion of the human scapula well exemplify this difference ; in the coracoid, e.g., at the first year, in the acromion at the fifteenth year. Embryology teaches the facts but affords not the reason. Special homology shows that the coracoid is a distinct bone, the acromion a mere process, in the vertebrate series. General ho- mology gives the ground of the distinctness the coracoid being the hasmapophysis of the ha3mal arch of which the scapula proper is the pleurapophysis. In most mammals this haamapophysis is stunted and terminates freely, like that of the last (floating) rib. In Monotrenies it attains and articulates with its haemal spine, as in the ' true rib,' and keeps this normal extent and condition through all the lower vertebrates. It is the typical state of the coracoid, which is departed from in all vertebrates above Mono- tremes : but such typical state is not passed through in the course of their development. As in that of other modified hasmal arches, the maxillary, e.g., so in the scapular arch, the special con- dition of the aborted haamapophysis is gained directly, not through any intervening transitory manifestation of the general character. So far is embryology from being a criterion of homology. In regard to what I have reckoned a ( seventh way of ana- 1 ' Pour avoir le veritable nombre des os de chaque espece, il faut remonter jusqu'aux premiers noyaux osseux tels qu'ils se montrent dans le foetus.' Lemons d' Anatomic Com- paree, 8vo. ed. 1835, torn. i. p. 120. 2 ' Ayant imagine de compter aufant d'os qu'il y a de centres d'ossification distinct?, et ayant essaye de suite cette mauiere de faire, j'ai eu bien d'apprecier la justesse de cette idee.' Annales du Museum, torn. x. p. 344. See, however, the remarks on this point in my ' Lectures on the Comparative Anatomy of the Vertebrate Animals/ Svo. 1846, p. 37, et seq. PREFACE. xxvii tomy,' I would remark that the existing kinds of vertebrates constitute part only, perhaps but a small proportion, of those which have lived. Two large primary groups of fishes have almost wholly passed away ; but the Polypterus, Lepidosteus, and sturgeon yield the anatomist some insight into the structural modifications of the Ganoidei of AGASSIZ ; whilst the shark, the skate, and the cestracion give a fuller knowledge of those of the Placoidei. Present reptiles form a mere fragmentary remnant of the great and varied class of cold-blooded air-breathing vertebrates which prevailed in the mesozoic age. More than half of the ordinal groups of the class, indicated by osteal and dental characters, have perished ; and it is only by petrified faeces or casts of the intestinal canal, by casts of the brain-case, or by correlative deductions from characters of the petrifiable remains, that we are enabled to gain any glimpse of the anatomical conditions of the soft parts of such extinct species : by such light some of the perishable structures of these animals are indicated in the text. As vertebrates rise in the scale and the adaptive principle pre- dominates, the law of correlation, as enunciated by CuviER, 1 be- comes more operative. In the jaws of the lion, e.g., .there are large laniaries or canines, formed to pierce, lacerate, and retain its prey. There are also compressed trenchant flesh-cutting teeth, which play upon each other like scissor-blades in the movement of the lower upon the upper jaw. The lower jaw is short and strong ; it articulates to the skull by a transversely extended convexity or condyle, received into a corresponding concavity, forming a close- fitting joint, which gives a firm attachment to the jaw, but almost restricts it to the movements of opening and closing the mouth. 1 ' Tout etre organise forme un ensemble, un systeme unique et clos, dont les parties se correspondent mutuellement, et concourent a la meme action definitive par une re- action reciproque. Aucune de ces parties ne peut changer sans que les autres changent aussi ; et par consequent chacune d'elles, prise separement, indique et donne toutes les autres.' Discours siir hs Revolutions de la Surface du Globe. 4to. 1826, p. 47. In this definition Cuvier apprehended, exclusively, the operance of the differencing and adapting pole, and the law become^ limited in its application accordingly. xxviii PREFACE. The jaw of the Carnivore developes a plate of bone of breadth and height adequate for the implantation of muscles, with power to inflict a deadly bite. These muscles require a large extent of surface for their origin from the cranium, with concomitant strength and curvature of the zygomatic arch, and are associated with a strong occipital crest and lofty dorsal spines for vigorous uplifting and retraction of the head when the prey has been griped. The limbs are armed with short claws, and endued with the requisite power, extent, and freedom of motion, for the wield- ing of these weapons. These and other structures of the highly- organised Carnivore are so co-ordinated as to justify CUVIER in asserting that ( the form of the tooth gives that of the condyle, of the blade-bone, and of the claws, just as the equation of a curve evolves all its properties ; and exactly as, in taking each property by itself as the base of a particular equation, one discovers both the ordinary equation and all its properties, so the claw, the blade-bone, the condyle, the femur, and all the other bones in- dividually, give the teeth, or are given thereby reciprocally ; and in commencing by any of these, whoever possesses rationally the laws of the organic economy will be able to reconstruct the entire animal.' l . The law of correlation receives as striking illustrations from the structure of the herbivorous mammal. A limb may termi- nate in a thick horny hoof. Such a foot serves chiefly, almost exclusively, for locomotion. It may f paw the ground,' it may rub a part of the animal's hide, it may strike or kick ; but it cannot grasp, seize, or tear another animal. The terminal ungu- late phalanx gives, as CUVIER declares, the modifications of all the bones that relate to the absence of a rotation of the fore-leg, and those of the jaw and skull that relate to the mastication offered by broad-crowned complex molars. But there are certain associated structures for the coincidence of which the physiological law is unknown. f I doubt,' writes 1 Op. cit. p. 49. PREFACE. xxix CUVIER, ' whether I should have ever divined, if observation had not taught it me, that the ruminant hoofed beasts should all have the cloven-foot, and be the only beasts with horns on the frontal bone.' l I may add that we know as little why horns should be in one or two pairs in those ungulates only which have hoofs in one or two pairs ; whilst in the horned ungulates with three hoofs there should be either one horn, or two odd horns placed one be- hind the other, in the middle line of the skull ; or why the ungu- lates with one or three hoofs on the hind foot should have three trochanters on the femur, whilst those with two or four hoofs on the hind foot should have only two trochanters. 2 ' However,' continues CUVIER, ( since these relations are con- stant, they must have a sufficing cause ; but as we are ignorant of it, we must supply the want of the theory by means of observa- tion. This will serve to establish empirical laws if adequately pursued, as sure in their application as rational ones.' 3 'That there are secret reasons for all these relations observation mav / convince us, independently of general philosophy.' f The con- stancy between such a form of such organ and such another form of another organ is not merely specific, but one of class with a corresponding gradation in the development of the two organs.' 4 ' For example, the dentary system of non-ruminant ungulates is generally more perfect than that of the bisulcates ; inasmuch as the former have almost always both incisors and canines in the upper as well as the lower jaw ; the structure of their feet is in general more complex, inasmuch as they have more digits or hoofs less completely enveloping the phalanges, or more bones distinct 1 Op. cit. 50. - Quarterly Journal of the Geological Society, p. 138. 1847. 3 ' Puisque ces rapports sont constants, il faut bien qu'ils aient une cause suffisante, maiscorame nous ne la connaissons pas, nous devons suppleer au defaut de la theorie par le moyen de 1'observation.' Op. cit. p. 50. 4 ' En effet, quand on forine un tableau de ces rapports, on y remarque non seulement une consistance specifique, si 1'on peut s'exprimer ainsi, entre telle forme de tel organe et telle autre forme d'nn organ different ; mais Ton apergoit aussi une Constance classique et une gradation correspondante dans le deyeloppement de ces deux organes, qui montrent, presque aussi bien qu'un raisonnement effectif, leur influence mutuelle.' Op. cit. p. 51. xxx PREFACE. in the metacarpus and metatarsus, or more numerous tarsal bones ; or a more distinct and better developed fibula ; or a concurrence of all these modifications. It is impossible to assign a reason for these relations ; but, in proof that it is not an affair of chance, we find that whenever a bisulcate animal shows in its dentition any tendency to approach the non-ruminant ungulates, it also manifests a similar tendency in the conformation of its feet.' After citino- similar instances of such constant relations, CUVIER as^ain O O declares that the palaeontologist f must avail himself of the method of observation' as a supplementary instrument when the reason or law of such relations is undiscovered ; and that he is most suc- cessful in the reconstruction of a whole from a part, who applies to the task ' efficacious comparison,' guided by ' tact (adresse) in discerning likeness.' l As we descend in the scale of life from the grade illustrative of ( Cuvier's Law,' the method of empirical observation becomes more and more essential, the tact with which it is applied being, however, in the ratio of the discernment of the correlations of structures. The results of the combined methods of interpreting fossil remains are leading to views of life transcending the gains to zoology as a record of well-classed species, or to physiology as illustrative of final purpose. A progress from more generalised to more specialised structures, analogous to that exemplified in existing grades of animal life and in successive phases of individual development, is appreciable in the series of species which have succeeded one another upon our planet. Certain structures which are transitory or rudimental in exist- ing species are persistent and developed in extinct. The caudal vertebrae are laid down in a gradually decreasing series of cartilaginous nuclei, in the embryo of modern bony fishes ; but in the course of ossification they become massed and blended together to form the base of a vertically extended symmetrical tail-fin. In all palaeozoic fishes the initial embryo-state persists, and the tail-fin, through the length of the upper lobe retaining the 1 Op. eit. p. 52. PREFACE. xxxi terminal series of vertebras, is unsyrnmetrical. The process of differencing which leads to the ( homocercal ' type begins in the mesozoic period and prevails in the neozoic. (See TABLE OF STRATA, &c., p. xxxviii.) A corresponding modification of the caudal vertebra? prevails in neozoic birds ; but the embryos of the existing species show the terminal vertebra distinct, in a tapering series, before they are massed into the 'ploughshare bone;' and such, doubtless, was the law of development in all the extinct species which have left tertiary ornitholites. But the earliest and as yet sole evidence > / of the fossil skeleton of a mesozoic bird shows the retention of the embryo condition, with ordinary growth of the vertebra?. 1 Modern ruminants are hornless when born, and have the me- tnpodials supporting the phalanges of the cloven foot distinct ; at an earlier foetal period rudiments of upper fore-teeth start in the gum but do not get beyond it. The eocene mammal that first indicates the ruminant type retained the transitory, and developed the aborted, characters of its successors. The metacarpals and metatarsals never coalesced to form a ' canon-bone ;' the upper canines and incisors were functional, but small and equal-sized ; and, as horns never sprouted, CUVIER called the extinct beast ( weaponless' (Anoplotherium). In modern horses the digit on each side the one supporting the hoof is undeveloped, and is represented by a concealed rudiment of the metapodial called ( splint-bone.' In the miocene horses these metapodials reached their full length and supported hoofed digits, but of small size, like the ' spurious hoofs ' of the ox. The eocene mammal initia- ting the type had these hoofs so developed as to form a functional tridactyle fool. Moreover, in the Palaotherium, certain teeth (symbolised in the present Work as p 1 ) which are rudimental and deciduous in the horse, were persistent and functional. The mesozoic marsupials manifested a lower or less differenced state of dentition, either by the degree of sameness of form (Phascolothere), or by the superior number (Thylacothere) of the molar series of teeth. 1 Philosophical Transactions, 1863, pp. 33, 45, pis. I. and III. xxxii PREFACE. The ( rudiments ' of parts and organs which are retained un- developed, or do not acquire the state capable of acting, or ( per- forming the function ' done by them in other species, are of two kinds : one exhibits the totality of the organ in miniature, as, e.g., lacteal glands and nipples of the male mammal; the other is a part of an organ, as, e.g., the few concealed caudal vertebra in the sloth, to which other vertebra are added, with concomitant growth, to make the organ perfect for its function, as in the tail of the Megathere. Some rudiments show beginnings of parts which rise to perfection in higher species of the existing series ; others are remnants of organs that were fully developed and func- tional in extinct species. TIEDEMANN'S ( scrobiculus parvus in loco cornu posterioris ' in the brain of Macacus, 1 and the part which VROLIK believed himself entitled to regard as an indication of the 'hippocampus minor' in the brain of Troglodytes,- are beginnings of structures which show their full development in the human brain, and merit the nomenclature assigned to them in anthropotomy. The filamentary limb of Protopterus (Vol. I. fig. 101, A), the didactyle limb in Ampliiuma (Ib. B), the tridactyle homologue in Proteus, are bemnnino-s of organs which attain full functional de- ~ o o velopment in higher vertebrates. The styliform metacarpals and metatarsals in Equus, on the other hand, are remnants of parts of digits which were entire in Hipparion, and were functionally developed in Palceotherium. Ruminants which habitually frequent heated arid plains or deserts, as the giraffes and camels, e.g., have lost the digits (ii and v, Vol. II. fig. 193, ox) that add to the resistance of the hoof on swampy ground, as in the bison, elk, and reindeer (Ib. fig. 311). The visual organ degenerates in species inhabiting dark caves or recesses (Amblyopsis (Vol. I. fig. 175), Heteropygii, Proteus, 1 Icones cerebri Simiarum, fol. p. 1 1, fig. iii. 2. 2 Versl. en Mededeel. der Kon. Ak;id.,xiii. 1862, p. 7. PREFACE. xxxiii the craw-fish of the ( Mammoth Cave,' and numerous insects and arachnidans). Lepidopus, Trichiurus, Stromateus, exemplify fishes which lose the ventral fins entirely with age ; they are rudimental in Gem- pylus, Psettus and Centronotus ; Soleotalpa has only the right ventral developed and the left rudimental ; the pectoral fins are rudimental in many pleuronectoids, either on both sides, as in Buglossus and Achirus., or on the blind side only, as in Monochir and many species of Synaptura. The ( adipose fin ' of certain Siluroid and Salmon old fishes is a rudimental dorsal, sometimes showing traces of rays. The prevalence of birds in New Zealand without wings (Dinor- nis\ or too feebly developed for the purpose of flight (Apteryx, Brachypteryx, Notornis, &c.), is associated with the absence in those islands of any higher form of life exercising destructive mastery of organisation, until the immigration of the human race. The wings of such birds, like the eyes of the cavern fishes and crustaceans, would seem to have degenerated for want of use ; their legs, by which locomotion was exclusively exercised, to have gained in size and strength. LAMAKCK, 1 adverting to observed ranges of variation in certain species, affirmed that such variations would proceed and keep pace with the continued operation of the causes producing them ; that such changes of form and structure would induce corresponding changes in actions, and that a change of actions, when habitual, became another cause of altered structure ; that the more frequent employment of certain parts or organs leads to a proportional increase of development of such parts, and that as the increased exercise of one part is usually accompanied by a corresponding disuse of another part, this very disuse, by inducing a proportional degree of atrophy, becomes an added element in the progressive mutation of organic forms. Concomitant changes of climate, and other conditions of a coun- 1 Philosophic Zoologique, torn. i. chaps, iii. vi. vii. VOL, I. b xxxiv PKEFACE. try affecting the sustenance or well-being of its indigenous animals, may lead not only to their modification but to their destruction. I have, in another work, pointed out the characters in the animals themselves calculated to render them most obnoxious to such ex- tirpating influences ; and have applied the remarks to the expla- nation of so many of the larger species of particular groups of animals having become extinct, whilst smaller species of equal antiquity have remained. ( In proportion to its bulk is the difficulty of the contest which, as a living organised whole, the individual of such species has to maintain against the surrounding agencies that are ever tending to dissolve the vital bond and subjugate the living matter to the ordinary chemical and physical forces. Any changes, therefore, in such external agencies as a species may have been originally adapted to exist in, will militate against that existence in a degree proportionate, perhaps in a geometrical ratio, to the bulk of the species. If a dry season be gradually prolonged, the large mam- mal will suffer from the drought sooner than the small one ; if such alteration of climate affect the quantity of vegetable food, the bulky Herbivore will first feel the effects of stinted nourish- ment ; if new enemies are introduced, the large and conspicuous quadruped or bird will fall a prey, whilst the smaller species con- ceal themselves and escape. Smaller animals are usually, also, more prolific than larger ones.' l The actual presence, therefore, of small species of animals in countries where larger species of the same natural families for- merly existed, is not the consequence of any gradual diminution of the size of such species, but is the result of circumstances, which may be illustrated by the fable of the ' Oak and the Reed ; ' the smaller and feebler animals have bent and accommodated themselves to changes which have destroyed the larger species. They have fared better in the ' battle of life.' Accepting this explanation of the extirpation of species as true, 1 On the Genus Dinornis (Part iv.), Zool. Trans., vol. iv. p. 15 (February 1850). PREFACE. xxxv MR. WALLACE J has applied it to the extirpation of varieties ; and as these do arise in a wild species, he shows how such deviations from type may either tend to the destruction of a variety, or to adapt a variety to some changes in surrounding conditions, under which it is better calculated to exist, than the type-form from which it deviated. IS T o doubt the type-form of any species is that which is best adapted to the conditions under which such species at the time exists ; and as long as those conditions remain unchanged, so long will the type remain ; all varieties departing therefrom being in the same ratio less adapted to the environing conditions of existence. But if those conditions change, then the variety of the species at an antecedent date and state of things may become the type-form of the species at a later date, and in an altered state of things. In his work ( On the Origin of Species by Natural Selection,' 2 MR. DARWIN more fully exemplifies, conjecturally, the reciprocal influence of external conditions and inherent tendencies to variety, in carrying on, as he believes, the deviations from type to specific and higher degrees of difference. All these, however, are conceptions of what may have, not observations of what have, originated a species. Applied to the structures which differentiate Troglodytes from Homo? or Chiromys from Lemur* they are powerless to explain them : and the structural differences in these instances are greater than in many other species maintaining their distinction by sexual in- capacity to produce fertile hybrids. An innate tendency or susceptibility in an offspring to differ from a parent is a fact of observation : when carried beyond a certain point the issue is called, from its rarity, a f monster.' But this tendency and its results are independent of internal volitions and external influences. 1 Proceedings of the Linnean Society, August 1858, p. 57. - Svo. 1859. 3 On the Classification and Geographical Distribution of the Mammalia, Svo. 1859, p. 92. 4 Transactions of the Zoological Society, vol. v. p. 86. b2 xxxvi PREFACE. Therefore, with every disposition to acquire information and receive instruction as to how species become such, I am still com- pelled, as in 1849, to confess ignorance of the mode of operation of the natural law or secondary cause of their succession on the earth. But that it is an ( orderly succession,' or according to law, 1 and also ' progressive ' or in the ascending course, is evident from actual knowledge of extinct species. The inductive basis of belief in the operation of natural law or ' secondary cause ' in the succession and progression of organised species, was laid by the demonstration of the unity of plan under- lying the diversity of animal structures, as exemplified by the determinations of special and general homology ; by the discovery of the law of ' Irrelative repetition ; ' by observation of the ana- logies of transitory embryonal stages in a higher animal to the matured forms of lower animals ; and by the evidence that in the scale of existing nature, as in the development of the individual, and in the succession of species in time, there is exemplified an ascent from the general or lower to the particular or higher con- dition of organism. The most intelligible idea of homologous parts in such series is that they are due to inheritance. How inherited, or what may be the manner of operance of the secondary cause in the pro- duction of species, remains in the hypothetical state exemplified by the guess-endeavours of LAMARCK, DARWIN, WALLACE, and others. In the lapse of ages, hypothetically invoked for the mutation of specific distinctions, I would remark that Man is not likely to preserve his longer than, contemporary species theirs. Seeing the greater variety of influences to which he is subject, the present characters of the human kind are likely to be sooner changed than those of lower existing species. And, with such 1 BADEN POTVELL, quoting from my Work ' On the Nature of Limbs,' 8vo. 1849, p. 86, writes: ' To what actual or secondary cause' ('Essays on the Unity of Worlds,' 1855, p. 401), instead of, 'To what natural laws or secondary cause the orderly suc- cession and progression of species may have been committed, we are, as yet, ignorant.' PREFACE. xxxrii change of specific character, especially if it should be in the ascensive direction, there might be associated powers of pene- trating the problems of zoology so far transcending those of our present condition, as to be equivalent to a different and higher phase of intellectual action, resulting in what might be termed another species of zoological science. With the present psychical and structural characteristics of the human species, it may be reasonably concluded that those of other existing species, especially of the distinctly marked vertebrate classes, will be, at least, concurrent and co-enduring ; and, in that sense, we may accept the dictum of the French zoologist:- - f La stabilite des especes est une condition necessaire a Fexistence de la science d'Histoire Naturelle.' At the same time, indulging with LAMARCK in hypothetical views of transmutative and selective influences during aeons transcending the periods allotted to the existence of ourselves and our contemporaries, as we now are, we may also say: ( La nature n'offre que des individus qui se suc- cedent les uns aux autres par voie de generation, et qui provien- nent les uns des autres. Les especes parmi eux ne sont que relatives, et ne le sont que temporairement.' Pleistocene lauie oj Virata ana uraer oj appearance of Vertebrate Life upon the Earth. 3 a c3 d i VI r-t m * VI 4^> 0) H MAN by Eemains. TERTIARY or NEOZOIC Turbary. Shell Marl. Glacial Drift, i g Brick Earth. JjJ \ by Weapons. Norwich "j Eed r Crag. Coralline / Pliocene Mammals under present geographical distribution. o* Faluns. Molasse. Miocene Euminantia. ^ Quadrumaua. & Proboscidia. ,v o** Gyps. London ) ,. h Clays. Plastic J J [Eocene ^ ^ ^ Eodentia. ^cS" Ungulata. Carnivora. SECONDARY or MESOZOIC Maestricht. Upper Chalk. Lower Chalk. Upper Greensand. Lower Greensand. Cretaceous Cycloid. ] Mosasaurus. Ctenoid, j FISHES< Polyptychodon. BIKDS, by Bones. Procoelian Crocodilia. Pterodactyles. Weald Clay. Hastings Sand. d "S Iguanodon. >^^lVT:iT'iiil'j -~Cliplomfl, bv Bonp! Kimmeridgian. Oxfordian. p Kellovian. Forest Marble. Bath-Stone. Stonesfield Slate. Great Oolite. ,4 Lias. o Pliosaurus. Birds by Bones and Feathers. g' Marsupials. ^ ^ ^ g Ichthyopterygia. "V O /3y ^ MAMMALIA .,. U. New Eed Sandstone. Muschelkalk. Bunter. fi ^ /ganoid. ^ '3 placo-ganoid. ^ PISCES! placoid. ^ Wenlock. I %' Llandeilo. ^ \^X^-^ c ^VV\^% cV Lingula Flags. 02 ^ ^ ^^ &<3& ^ Cambrian. Fucoicls. Zoophytes. CONTENTS, OR SYSTEMATIC INDEX. CHAPTER I. CHARACTERS OF VERTEBRATES. SECTION PAGE 1. Developmental characters . . 1 2. Structural characters .... . . 3 3. Piscine modification .... ..... 4 4. Reptilian modification ... . 5 o. Avian modification .... ... 6 6. Mammalian modification . 6 7. Genetic and thermal distinctions . 6 8. Sub-classes of Haematocrya, or Cold-blooded Vertebrates ... 7 9. Orders of Hsematocrya 9 CHAPTER II. OSSEOUS SYSTEM OF H.EMATOCRYA. 10. Composition of Bone ... .... 19 11. Development of Bone , . . 21 12. Growth of Bone ...... 23 13. Classes of Bone ........... 26 14. Type-segment, or Vertebra .... .27 15. Archetype Skeleton . . 29 16. Development of Vertebrae ......... 30 17. Vertebral column of Fishes .... .... 34 18. Vertebral column of Batrachia ........ 46 19. Vertebral column of Ichthyopterygia ....... 50 20. Vertebral column of Sauropterygi a . . . . . . . 51 21. Vertebral column of Ophidia ........ 53 22. Vertebral column of Lacertia ........ 57 23. Vertebral column of Chelonia ........ 60 24. Vertebral column of Crocodilia ........ 65 25. Vertebral column of Pterosauria ........ 70 26. Development of the Skull 71 27. Skull of Plagiostomi 76 28. Skull of Protopteri .......... 82 xl CONTENTS. SECTION PAGE 29. Skull of Batraclria .... 30. Skull of Osseous Fishes .... 31. Skull of Chelonia 126 32. Skull of Crocodilia 135 33. Skull of Ophidia H6 34. Skull of Lacertilia 154 35. Skull of lehthyopterygia 158 36. Skull of Dicynodontia 159 37. Skull of Pterosauria . . . . . . . . . . 161 38. Scapular arch and appendages of Hsematocrya 161 39. Pectoral limb of Fishes 163 40. Pectoral limb of Reptiles 169 41. Pelvic arch and limb of Fishes 179 42. Pelvic arch and limb of Reptiles . . . . . . . . 181 43. Dermoskeleton of Fishes ......... 193 44. Dermoskeleton of Eeptiles ......... 198 CHAPTER III. MUSCULAR SYSTEM OP H^EMATOCRYA. 45. Structure of Muscle 200 46. Myology of Fishes 202 47. Myology of Reptiles .......... 215 48. Locomotion of Fishes .......... 243 49. Locomotion of Serpents ......... 259 50. Locomotion of Limbed Reptiles ........ 262 CHAPTER IV. NERVOUS SYSTEM OF KLEMATOCRYA. 51. Nervous Tissues ......... 266 52. Myelencephalon of Fishes ......... 268 53. Myelencephalon of Reptiles 290 54. Myelencephalic membranes in Hgematocrya 296 55. Nerves of Fishes ........... 297 56. Nerves of Reptiles .......... 309 57. Sympathetic nervous system . . . . . . . . . 318 58. Sympathetic nervous system of Fishes ....... 320 59. Sympathetic nervous system of Reptiles . . . . . . 321 60. Appendages of the Nervous System ....... 323 61. Organ of Touch in Hsematocrya ........ 325 62. Organ of Taste in Reptiles . . . . . . . . . 327 63. Organ of Smell in Hamatocrya ........ 328 64. Organ of Sight in Fishes ......... 331 65. Organ of Sight in Reptiles ......... 337 66. Organ of Hearing in Fishes ......... 342 67. Organ of Hearing in Reptiles .... ... 347 68. Electric Organs of Fishes 350 CONTENTS. xli CHAPTER V. DIGESTIVE SYSTEM OF H^EMATOCRYA. SECTION PAGE 69. Dental Tissues 359 70. Teeth of Fishes 368 71. Teeth of Eeptiles 385 72. Alimentary canal of Fishes .... . . 409 73. Liver of Fishes 425 74. Pyloric appendages and pancreas of Fishes ...... 428 75. Alimentary canal of Reptiles ... . 433 76. Liver of Reptiles ....... 448 77. Pancreas of Reptiles ....... 453 CHAPTER VI. ABSORBENT SYSTEM OF HJEMATOCRYA. 78. Connective tissue : serosity ..... 455 79. Absorbents of Fishes .... 456 80. Absorbents of Reptiles ...... . . 458 CHAPTER VII. CIRCULATING AND RESPIRATORY SYSTEMS OF ILEMATOCRYA. 81. Blood of Fishes . . .... 463 82. Veins of Fishes . .... 464 83. Heart of Fishes ..... 470 84. Gills of Fishes .... . 475 85. Arteries of Fishes ...... . 488 86. Air-bladder of Fishes . 491 87. Blood of Reptiles . 500 88. Veins of Reptiles 501 89. Heart of Reptiles 505 90. Gills of Batrachians ... . 512 91. Arteries of Reptiles .... 516 92. Lungs of Reptiles 521 93. Larynx of Reptiles .... 527 94. Respiratory actions of Reptiles . . 530 CHAPTER VIII. URINARY SYSTEM OF HJ5MATOCYRA. 95. Kidneys of Fishes ... . 533 96. Kidneys of Reptiles . . . 537 9". Adrenals of Hsematocrya ... 542 YOL. I. C xlii CONTENTS. CHArTEK IX. TEGUMENTARY SYSTEM OF II^MATOCHYA. SECTION PAGE 98. Composition of Tegument ,"> i.> 99. Teguments of Fishes 546 100. Teguments of Keptiles 550 CHAPTER X. PECULIAR AND DUCTLESS GLANDS AND REPRODUCIBLE PARTS. 101. Scent-glands of Eeptiles 562 102. Poison-glands of Reptiles ......... 563 103. Thyroid body or gland of Hsematocrya ....... 564 104. Thymus body or gland of Reptiles ....... 565 105. Reproducible parts in Hsematocrya ....... 566 CHAPTER XI. GENERATIVE SYSTEM OF HJEMATOCRYA. 106. Male organs of Fishes 568 107. Female organs of Fishes 571 108. Male organs of Batrachians 576 109. Male organs of Reptiles 579 110. Female organs of Batrachians and Reptiles ...... 583 CHAPTER XII. GENERATIVE PRODUCTS AND DEVELOPMENT OF ILEMATOCRYA. 111. Semination of Hsematocrya 589 112. Ovulation in osseous Fishes and Batrachians ..... 592 113. Ovulation in cartilaginous Fishes and scaled Reptiles . . . . 597 114. Fecundation in Fishes 599 115. Development of Fishes .......... 601 116. Growth and nests of Fishes ......... 611 117. Fecundation in Reptiles ......... 614 118. Oviposition in Reptiles ......... 616 119. Development of Batrachians . . . . . . . . . 619 120. Development of Reptiles 630 ERRATA. Below Cuts 134, 135, 136, 137, for ' xxxm.,' read ' xxiu ' Page 396, eight lines from top, /or ' premaxillary,' read ' vomerine.' ,, 448, eighteen lines from top, /or ' timical,' read ' tumid.' ,, 512, eleven lines from bottom, transpose ' fig. 399, R ' to tenth line, after ' ventricle.' 6'25, four lines from top, for ' fig. 435,' read ' fig. 424, &.' 630, five lines from bottom, for ' bodies,' read ' borders.' THE ANATOMY OF VERTEBRATES. CHAPTER I. CHARACTERS OF VERTEBRATES. 1. Developmental characters. Vertebrates, like lower animals, begin in a semifluid nitrogenous substance called ( plasma,' fig. 1 , A, a ; primarily differentiating into albumen, fibrine, lemma, ib. , c l , nuclei and cells ; in winch lat- ter form the individuality of the new organism first dawns as a nucleated ( germ-cell ' or $ germinal vesicle, ib. d. By the evolution of albumi- nous granules and oil-particles plasma becomes f yolk,' fig. 1, B,C ; the germinal vesicle may be obscured by endogenous multiplication of granules, gra- nular cells and oil-globules, which combine with those of the yolk to form its germinal part : an outer layer of ( lem- ma,' D, ch, completes the un-. impregnated vertebrate egg. For further developement another principle is needed, viz. the hyaline nucleus or * Stages of derelopement of the ovarian egg of a vertebrate prodllCt Of the Sperm-Cell, fig. animal (Gasterosteus). CLXXVI. 2, called ( spermatozoon.' Its reception by the egg, as at a, I, fig. 3, is followed by the formation of a germ-mass. This mass is due 1 Gr. lemma, skin ; also called 'primary' or 'basement' membrane ; distinguished, through its relations, as * ncurilemma, sarcolemma, adcnolemma ' or the limitary membrane of gland-follicles, c. ANATOMY OF VERTEBKATES. to a series of self-splittings of the impregnated centre, which ' fissiparous' progeny assimi- late or incorporate more or less of the yolk. In fig. 4, A, d is the impregnated germ- yolk ; c the fluid between it and the zona,^/; f is albumen from which the chorion, cho, arises. In B, fig. 4, is shown the first division or segmen- tation of the germ-yolk ; c shows the second division; and D, a later stage in which the properties of the impregnated Stages of developement of the ovarian egg of a verte- haV6 brate animal (cowled). CLXXVI. and distributed by fissiparous multiplication amongst the countless nucleated cells which form the germ-mass. Thus far the vertebrate germ resembles in form, structure, and 2 behaviour, the infusorial monad and the germ- stage of invertebrates. The next step impresses upon the nascent being its ' vertebrate ' type. Linear rows of the nucleated cells coalesce and become converted into the nervous axis, which under the form or appearance of a double chord, fig. 5, ch, marks the dorsal or f neural ' aspect with three of the embryonal rudiment. The nutritive organs spermatoa, and their . . ., ., , i nucleus the 'spermato- grow irom the opposite side. Along the mter- zoon ' (Cock). CLXXVII. i i j i i P-I 11 space is laid the basis ot the skeleton, as a gelatinous cylinder, in a membranous sheath, called f notochord,' * 3 which developes a pair of plates ( neurad ' 2 to enclose the nervous axis, and a pair of plates ( haamad ' 3 to enclose the vascular axis and organs of vegetative life. Flesh and skin coextend with the enclosing plates. This formation of two distinct parallel cavities ( neural ' and e haemal ' -under symmetrical guidance in the vertical or { neuro-haamal ' direction, with a repeti- tion of parts 011 the right and left sides, transverse or ( bi-lateral ' I Vertebrate egg, impregnated by the spermatozoa (Rabbit). CLXXVI. 1 The ' chorda dorsalis' of embryologists. 2 Backward in man, upward in beasts. 3 Forward in man, downward in beasts. ANATOMY OF VERTEBRATES. 3 Stages of developement of a vertebrate germ (Rabbit), evil. symmetry, constitutes the chief developmental characteristic of the vertebrate animal. The twofold symmetry is shown in the bone-segment, fig. 7 ; also in the flesh-segment surrounding the skeletal one in fig. 6, in which the mid point 4 marks the f noto- chord ; ' with the neu- ral canal above, the haBinal canal below ; both surrounded by the two neural and two haemal masses of muscles on each side. The lancelet, Branckiostoma, fig. 23, superinduces its distinc- tive characters upon this stage. Aponeurotic septa accompany the pairs of nerves and divide the longitudinal muscular masses into segments. At the next rise segmentation is shown by the develop- ment of cartilage, forming pairs of plates, fig. 5, v, commonly corresponding with the pairs of nerves sent off from the neural axis, and with the pairs of vessels from the haemal axis. As these plates ossify, ossification commonly also begins at cor- responding points of the notochord, dividing it into as many central parts as there are peri- pheral plates or arches, and constituting skeletal segments or ' vertebrae ; ' according, or reducible to, the type, fig. 7. 2. Structural characters. The series of ( vertebrae/ under their several modifications, as the neural or haemal organs may predominate, constitute the vertebral column. The neural axis consists of ' encephalon' or brain, and of ( myelon' or spinal chord. The organs of the five senses - touch, taste, smell, hearing, and sight - - are usually present. The blood-discs, fig. 8, speedily acquire the red colour which, by their number and minuteness, they * J Section of impart to the whole blood. The heart is a compact mus- Corm of a Rabbit (Barry) OF cular organ, of two or more cavities, propelling the blood, ment;tauofa through a closed system of arteries and veins, directly to the breathing-organ, and, in most vertebrates, directly also to the body. The breathing-organ communicates with the pharynx. The alimentary canal has distinct receptive and expellent apertures, usually at opposite ends of the trunk. The mouth is provided with two jaws, placed one above or before the other, working in the direction of the axis of the body. The muscles surround B 2 ANATOMY OF VERTEBRATES. the bony or gristly levers on which they act. The limbs do not exceed two pairs. The sexes are distinct, and the individual is developed directly from an impregnated ovum. Under the vertebrate plan of structure animals grow to a greater size and live a longer time, than under any of the invertebrate plans. ^^^^1 3. Piscine modification. All pleurapophysis vertebrates, during more or less of their developmental life-period, float in a liquid of similar specific gravity to themselves. A large proportion, constituting the lowest organised and ~ ~ first developed forms of the pro- vince, exist and breathe in water, and are called ( fishes.' Of these a few retain the primitive vermiform condition and develope no limbs : in the rest they are ( fins,' of simple form, moving by one joint upon the body, rarely adapted for any other function than the impulse or guidance 8 ||neural spine zygapophysis J||% neurapophysis cliapophysis _.f^ ^_ 2~~rc | LI " '''"''vor*- parapopliysis ^i^_^,_x3 ijJiff hremapophysis zygapophysis :;F O ll hajmal spine Ideal typical vertebra. CXLV. d g Blood-discs, each magnified 300 diameters linear, a, Man ; &, Musk-deer ; c, Goose ; d, Crocodile ; e, Frog ; /, Siren ; g, Cod-flsh ; 7i, Skate. CXLV. of the body through the water. The shape of the body is usually such as is adapted for moving with least resistance through a liquid medium. The surface of the body is either smooth and lubricous, or is smoothly covered by overlapping scales, is rarely defended by bony plates or roughened by tubercles, still more rarely armed with spines. The neural axis presents but one local enlargement, at the fore end, forming the ( encephalon ; ' it is small, and consists of a suc- cession of simple ganglionic masses, most of which are appro- priated to the function of a nerve of special sense. Touch is feebly exercised, and an organ for that sense rarely developed. ANATOMY OF VERTEBRATES. 5 The tongue, as an organ of taste, is hardly conspicuous; the framework supporting it relates chiefly to the mechanism of swal- lowing and breathing, and is suspended to a pedicle common to it and the mandible. Of the organ of hearing there is no outward sign ; but the essential internal part or ( labyrinth ' is present, and its semicircular canals are, in most fishes, largely developed. The labyrinth is devoid of a ( cochlea,' and is rarely provided with a proper chamber, but is lodged, in common with the brain, in the cranial cavity. The eyes are usually large, seldom defended by eyelids, and never served by a lacrymal apparatus. The ali- mentary canal is commonly short and simple, with the divisions less clearly marked than in higher vertebrates; the short and wide gullet being hardly distinguishable from the stomach. The pancreatic function appears to be performed by commonly more or fewer crecal appendages to the duodenum. The heart consists essentially of one auricle receiving the venous blood, and one ventricle propelling it to the gills, or organs submitting that blood in a state of minute subdivision to the action of aerated water. From the gills the arterial blood is carried over the entire body by vessels, the circulation being aided by the contraction of the surrounding muscles. The blood is cold, or with a temperature rarely above that of the surrounding medium. The coloured discs are, in some fishes, subcircular, fig. 8, g\ in others, subelliptical, ib. h, or elliptical ; comparatively large, but not the largest amongst vertebrates. The primordial renal glands (corpora Wolffiana) are persistent, and secrete the urine from venous blood. Such are the leading anatomical characters of the class Pisces Fishes. 4. Reptilian modification. Many fishes have a bladder of air between the digestive canal and kidneys, which, in some, com- municates by an air-duct with the gullet ; but its office is chiefly hydrostatic. When, in the rise of structure, this air-bladder begins to assume the vascular and pharyngeal relations, with the form and cellular structure of lungs, the limbs acquire the character of feet; at first, as in Lepidosiren, fig. 41, 99, thread- like and many-jointed - - then bifurcate, or two-fingered, with the ordinary elbow and wrist-joints of land-limbs (Amplmima), fig. 100, B, D, next, three-fingered, as in Proteus, or four-fingered, but reduced to the pectoral pair, as in Siren. From these gill- retaining transitional forms, up to and including crocodiles, all cold-blooded vertebrates, with lungs, breathing air directly, are called Reptiles (Reptilia, Cuv.). The heart has two auricles; the ventricle, in most, is imperfectly divided, and more or less of 6 ANATOMY OF VERTEBRATES. the venous blood is mixed with the arterial blood which circulates over the body. The lungs retain the form of bags, with cellulo- vascular walls, varying as to thickness, and are situated, with the other organs of vegetative life, in a common thoracic-abdominal cavity. 5. Avian modification.- -When the lungs become spongy, and the cavity of the air-bag is obliterated by the multiplication of vascular cellules, and when a four-chambered heart transmits the venous blood to the lungs, and pure arterial blood to the body, the temperature is raised, and is maintained at from 90 to 105 Fahr., whatever may be that of the surrounding medium. Of these hot-blooded vertebrates, one class has the lungs fixed, and communicating with air-cells extending into the abdomen, and o ~ usually other parts of the body ; this class is oviparous, is clothed with feathers, and has the pectoral limbs modified as wings ; it is called Aves Birds. 6. Mammalian modification. In the other class of warm- blooded animals, the spongy lungs are freely suspended and confined to a thoracic cavity, defined by a midriff from the abdomen ; the class is hair-clad, viviparous, and suckles the young, whence it is called Mammalia - - Mammals. 7. Genetic and thermal distinctions.- The broad and well- marked characters afforded by the respiratory system will probably give permanence to the division, so convenient for most purposes, of the vertebrate province into the four great classes above defined, viz. Pisces., Reptilia, Aves, Mammalia. But many important relations and affinities are thereby masked. Although the last two classes agree, as ' hot-blooded vertebrates,' in their higher cerebral developement, and in the more complex heart and lungs, birds, by genetic and developmental characters, as well as by the general plan of their organisation, are more intimately and naturally allied to the oviparous saurian s than to the viviparous mammals. In their generation and development, modern batrachians differ from other cold-blooded air-breathers, and agree with fishes. Present knowledge of extinct forms more clearly exposes the artificial nature of the primary groups of the oviparous vertebrates. An important link, the Pterosanria, or flying reptiles, with wings and air-sacs, fig. 108, more closely connecting birds with the actual remnant of the reptilian class, has passed away. Other extinct orders (Ganocepliala and Laln/rintho- dontia) have demonstrated the artificial nature of the distinction between fishes and reptiles, and the close transitions that connect together all the cold-blooded vertebrates. ANATOMY OF VERTEBRATES. 7 Thus vertebrates might be binarily divided into oviparous, I. II. in., and viviparous, iv. ; into aiiallantoic or branchiate and allantoic or abranchiate ; into H&matothermal, 1 having the four- chambered heart, spongy lungs, hot blood, and Hcematocryal? having less perfect breathing organs, less complex heart, with cold blood ; and each of such divisions are artificial and convenient. It suits my present purpose to adopt the latter. 8. Subclasses of Hcematocrya. - With the best insight peering into the dark vistas of the remote past - - that one can command into the nature of the strange forms which then o perished, and combining with pakeontological research the results of anatomical and developmental scrutiny of existing vertebrates, the following seem to be the best defined cold-blooded groups, each with such characters in common as leads to their beino; called O ' natural,' and of a value which may be expressed by the term 6 sub-class.' I. DERMOPTERI. III. PLAGIOSTOMI. II. TELEOSTOMI. IV. DIPNOA. V. MONOFNOA. Subclass I. DERMOPTERI. Body vermiform, limbless; endo- skeleton membrano- cartilaginous and notochordal, 3 ribless ; skin scaleless, lubricous ; a vertical fin-fold bordering the hind part of the body, without fin-rays ; myelon opaline, ductile, elastic ; no sympathetic nerve ; organ of smell single ; eyes wanting, or very small ; "optic nerves not crossing each other ; auditory labyrinth of one or two semicircular canals ; mouth jawless, or suctorial ; alimentary canal straight, simple, without crecal appendages, pancreas, or spleen. Branchial function independent of the mouth ; heart, without ' bulbus arteriosus ; ' a pulsatile portal sinus ; no swim-bladder ; testes and ovaria elongated plates without ducts ; generative outlet peritoneal ; ova numerous, small, simultaneously developed, and impregnated externally ; cleavage of yolk entire ; no amnios or allantois ; a metamorphosis, as, e. g. from Ammo- ccetes to Petromyzon, after the third year from the egg. Subclass II. TELEOSTOMI. 4 - -Body pisciform, with medial and 1 Gr. /Kiima, blood ; thermos, hot. 2 Gr. haima, blood; cruos, cold. Retaining the notochord or primitive basis of the vertebral column. This word (from Gr. telos, end or completion ; fifonui, mouth ;) refers to the com- pletion of the mouth by opposing upper and lower jaws, and also to its terminal position, opening at the fore end of the head. 8 ANATOMY OF VERTEBRATES. parial fins, supported by rays ; endoskeleton in most, more or less ossified ; hyoid arch attached to tympanic pedicle ; scapular arch attached to the occiput ; no sternum : skin defended by scales or plates ; brain with predominant mesencephalon ; myelon opake, inelastic ; a sympathetic nerve ; organ of smell double ; eyes usually large, with bony sclerotic ; auditory labyrinth with three semi- circular canals, in the cranial cavity ; mouth formed by upper and lower jaws, opening at the fore part of the head, and admitting the respiratory currents ; intestine, in most, with pyloric appendages and spleen ; anus in front of urethra ; air-bladder in most ; gills, free ; branchial outlet single on each side, defended by a bran- chiostegal flap, with one or more rays ; testes (' milt ') and ovaries (* roe ') large, with continuous ducts in most ; ova very numerous and small, simultaneously developed, and impregnated, usually, externally ; no amnios or external allantois. Subclass III. PLAGIOSTOMI. Endoskeleton cartilaginous, or partially ossified; scapular arch detached from occiput; exo- skeleton as osseous granules or tubercles ; body with medial and parial fins, the hinder pair pelvic in position ; caudal-fin with produced upper lobe ; brain with the prosencephalon predominant ; auditory labyrinth in a special chamber ; mouth, in most, a wide transverse slit, opening below the head; intestine with a spiral valve, pancreas, and spleen ; no air-bladder ; bulbus arteriosus with numerous rows of valves ; gills, in most, fixed, and with several branchial outlets on each side ; testes of moderate size, with sperm-duct and copulatory apparatus ; ovaries with few and large ova, successively developed and conveyed away by a detached oviduct ; ova impregnated and, in some, developed in- ternally ; embryo without amnios or allantois, and with deciduous external gills. Subclass IV. DIPNOA. Endoskeleton more or less ossified ; ribs wanting, or short and free; parial members as legs ; brain with predominant prosencephalon ; optic nerves not decussating ; audi- tory labyrinth in a special chamber, but with only the ( fenestra vestibuli ; ' nostrils communicating with the mouth ; intestine, with pancreas and spleen ; air-bladder as a pair of lungs, com- municating by a duct and glottis with the ha3mal side of the pharynx; heart, in most, with one ventricle and two auricles. Testes of moderate size, with sperm-ducts, but no intromittent organs or claspers ; ovaries with detached oviducts ; ova simulta- neously developed, and, in most, impregnated externally. Embryo without amnios or allantois, and with external gills. ANATOMY OF VERTEBRATES. 9' Subclass V. MONOPNOA. Encloskeleton ossified ; exoskeleton in most as horny scales, in some as bony scutes ; one occipital conclyle ; vomer usually single ; trunk-ribs long and curved. Brain with predominant prosencephalou. Labyrinth with both fenestra vestibuli and fenestra rotunda ; a tympanum in most ; lungs ; heart with two auricles, and with the ventricle more or less completely divided. Testes with ducts and intromittent organ. Ovaria with detached oviducts. Ova successively developed, impregnated with copulation. An ainnios and allantois. No metamorphosis. 9. Orders of H^EMATOCKYA. Subclass I. Order I. CIKROSTOMI. Body compressed ; mouth a longitudinal fissure with sub-rigid cirri on each side. Pulsating vessels or sinuses in place of heart. Blood pale ; free pharyngeal branchial filaments, and a branchial dilatation of the ossophagus. Gen. Branchiostoma. Ex. Lancelet. Order II. CYCLOSTOMI. Body cylindrical; heart distinct; branchial artery without bulb ; branchia3 sacciform, with external spiracles, six or seven on each side, blood red. Mouth subcircular, suctorial, but longitu- dinal when closed. Olfactory sac communicating with, or produced into, a canal. Gen. Mijxine. Ex. Hag-fish. Petromyzon. Ex. Lamprey. Subclass II. A. Arterial bulb with one pair of valves ; optic nerves decussating ; vertebras biconcave. Order III. MALACOPTERI. Skin, in most with cycloid scales, in a few with ganoid plates ; rarely naked. Fins supported by rays, all of which (save the first in the dorsal and pectoral, in some) are ( soft,' or many-jointed ; a swim-bladder and air-duct ; peritoneal outlets in many. 10 ANATOMY OF VERTEBRATES. Suborder I. APODES. Fam. 1. SymbranchidcB. Ex. Cuchia. 2. Mur&nidcR. Ex. Eel. 3. Gymnotidcs. Ex. Gymnotus. Suborder II. ABDOMINALES. Fam. 1. Heteropygii. Ex. Amblyopsis. 2. Clapeidce. Ex. Herring. 3. Salmonidcs. Ex. Salmon. 4. Scopelid(R. Ex. Saurus. 5. Characinidce. Ex. Myletes. 6. Galaxidce. Ex. Galaxias. 7. Esocidce. Ex. Pike. 8. MormyridcE. Ex. Mormyrus. 9. CyprinodontiddB. Ex. Umber. 10. Cyprinidce. Ex. Carp. 11. Siluridce. Ex. Sheat-fish. 12. Alepisauridce. Ex. Marine Sheat-fish. Suborder III. PHARYNGOGNATHI. Fam. 1, Scomber-esocidce. Ex. Saury-Pike. Order IV. ANACANTHINI. Endoskeleton ossified ; exoskeleton in some as cycloid, in others as ctenoid scales ; fins supported by flexible many-jointed rays ; ventrals beneath or in advance of the pectorals, or wanting ; swim- bladder, when present, without a duct. Fam. 1. Ophididce. Ex. Ophidium. 2. Gadidce. Ex. Cod. 3. Pleuronectidce. Ex. Plaice. Order V, ACANTHOPTEKI. Endoskeleton ossified ; exoskeleton, in most, as ctenoid scales ; fins with one or more of the first rays unjointed or inflexible spines ; ventrals, in most, beneath or in advance of the pectorals ; duct of swim-bladder obliterated. ANATOMY OF VERTEBRATES. 11 Suborder I. PPIARYNGOGNATHI. Fam. 1. Chromidce. Ex. Chromis. 2. Cyclo-labridce. Ex. AVrasse. 3. Cteno -lab r idee. Ex. Pomacentrus. Suborder II. ACANTHOPTERI VERI. Fam. 1. Prrcidce. Ex. Perch. 2. Sqiiammipennes. Ex. Chastodon. 3. SparidcB. Ex. Sea-bream, Gilthead. 4. Scicenidce. Ex. Maigre. 5. Ldbyrinthobranchii. Ex. Anabas or Tree-climber. 6. Muyilidce. Ex. Mullet. 7. Atlierinidce. Ex. Sand-smelt. 8. SplujrainidcR (cycloid). Ex. Barracuda. 9. ScombcridfB (cycloid). Ex. Mackerel. 10. Sclerof/t'iildtf. Ex. Gurnard, Miller's thumb. 11. Taitioidci. Ex. Riband-fish. 12. Teuthyidce. Ex. Lancet-fish, 13. Fistnl& *r pl modifications of which constitute the organs of special sense. That of smell, 4, 19. is 1 . Cervical segment or situated in advance of its proper (nasal) seg- vertebra meiit, which becomes variously modified to enclose and protect it. The organ of sight, lodged in a cavity or 'orbit' be- tween its own (the frontal) and .the nasal segment, is here drawn above that interspace. The nerve of taste perforates the neurapophysis of the third segment, 6, or passes by a notch between this and the frontal segment, to expand in the sense-organ, or ' tongue,' which is supported by the haemal spine, 41, 42, of its own (parietal) segment. The fourth is the organ of hearing, 16, indicated above the interspace between the neura- pophysis of its own (occipital) and that of the antecedent (parietal) vertebra, in which it is always lodged ; the surrounding vertebral elements being modified to form the cavity for its reception, which is called f otocrane.' The jaws are the modified haemal arches of the first two seg- ments. The mouth opens at the interspace between these haemal arches ; the position of the vent varies (in fishes), but always opens behind the pelvic arch, s, 62, 63, p, when this is ossified. Outlines of the chief developements of the dermoskeleton, in different vertebrates, which are usually more or less ossified, are added to the neuroskeletal archetype ; as, e. g. the median horn supported by the nasal spine, is, in the rhinoceros ; the pair of lateral horns developed from the frontal spine, n, in most rumi- nants ; the median folds, D i, D 11, above the neural spines, one or more in number, constituting the ' dorsal ' fin or fins in fishes and cetaceans, and the dorsal hump or humps in the buffaloes and camels ; similar folds are sometimes developed at the end of the tail, forming a ' caudal ' fin, c, and beneath the haemal spines^ constituting the ( anal ' fin or fins, A, of fishes. 30 ANATOMY OF VERTEBRATES. The different elements of the pri- mary segments are distinguished by peculiar markings : The neurapophyses by diago- nal lines, thus The diapophyses by vertical lines- The parapophyses by horizon- tal lines- The centrum by decussating horizontal and vertical lines The pleurapophyses by diago- nal lines The appendages by dots- -'.'..'.. The neural spines and haemal spines are left blank. In certain segments the elements are also specified by the initials of their names :- ns is the neural spine. n is the neurapophysis. pi is the pleurapophysis. c is the centrum. h is the haemapophysis, also indi- cated by the numbers 21, 29, 44, 52, 58, 63, 64. l hs is the haemal spine. a is the appendage. The centrum is the most constant vertebral element as to its existence, but not as to its ossification. There are some living fishes, and formerly there were many, now extinct, in which, whilst the peripheral elements of the vertebra become ossified, the central one remains unossified ; and here a few words are requisite as to the developement of vertebrae. 16. Developement of vertebras. The central basis of the neuroskeleton is laid down in the embryo of every vertebrate animal as a more or less 1 See 'TABLE OF SYNONYMS, Special Homologies,' for the names of the bones indicated by numbers. ANATOMY OF VERTEBRATES. 31 cylindrical fibrous sheath, filled with simple cells containing jelly. The centrums, or 6 bodies of the vertebras,' are developed in and from the notochord. The bases of the other elements are laid down in fibrous bands, diverging from the notochord, and giving the first indication of the segrnental character of the skeleton. In Dermopteri the neu- adipose substance inner layer outer layer - of fibrous capsule neural canal fibrous band, or basis of gelatinous chorda ral and haemal canals are formed by a separation of the layers of the outer division of the sheath of , i i -I n or A Transverse vertical section of vertebral column of Myxine. xxi. the notochord, ng. 22. A transverse partition divides the larger portion of the neural canal, lodging the myelon, from a smaller portion above containing adipose tissue. In the Lancelet the substance of the noto- chord, fig. 23, ch, consists of a number of circular discoid or flattened vesicles, pressed one upon another within the sheath, like a pile of coins in a purse ; the sheath is strength- ened by a longitudinal filamentary ligament above and below. Aponeurotic septa pass off, with each pair of nerves, to the interspaces of the muscular segments, giving attachments to the fibres. A median vertical membrane rises from the neural 23 Diagram of anatomy of the Lancelet, Branchiostoma sheath, and beyond the abdominal cavity descends from the hremal sheath, passing between the right and left series of myocommata. The dermo-neural and dermo-haemal spines are indicated by short linear series of firmly adhering flattened cylindrical cells. The next step in the skeletal tissues is shown in a pair of jointed cartilaginous filaments, fi ;. 23, h 9 which bound or strengthen the borders of the longitudinal oral slit, each cartilage supporting on conical prominences the oral cirri (ib. f, f) : numerous carti- laginous filaments strengthen the sides of the branchial cavity, ib. a, with intervening fissures, not opening upon the skin. In the Lamprey cartilaginous neurapophyses, fig. 24, n, n, strengthen the sides of the neural canal, In the Sturgeon, fig. 25, the inner 32 ANATOMY OF VERTEBRATES. layer of the notochordal capsule has assumed the texture of tough hyaline cartilage; and not only are firm opakc cartilaginous neurapophyses present, but also parapophyses, pleurapophyses, Fore part of skeleton, Lamprey (Pctromyzon) and neural spines. The part of the iieurapophysis bounding the true neural canal is usually distinct from that bounding the fat- filled fissure above. The parapophyses are united by a con- tinuous plate of cartilage forming an inverted arch beneath the aorta, in the trunk, ana- logous to that formed by bone in the lower neck- vertebra of birds, fig. 20. iuterneural cartilage ueural spine filiro-adipose canal neural canal gelatinous chorda inner layer of fibrous capsule as hyaline cartilage O Pleurapophysis In the ChimcBra der subossified ngs parapophysis - interhitmal cartilage hanual canal Abdominal vertebra, Sturgeon in the o nous sheath of the noto- chord, which are more numerous than the neu- ral arches. These, where unconflueiit with each other, are distinct also from the parapophyses, which in the tail bend down to form the hoemal arches. In the Mediterranean Grey Shark (Notidanus cinereus) the vertebral centres are still feebly and irrelatively marked out by numerous slender rings of hard cartilage in the notochordal capsule, the number of vertebra? being more definitely indicated by the neurapophyses and parapophyses ; but these remain cartilaginous. In the Lepidosiren the peripheral vertebral elements, fig. 41, ??, ns, p, hs, are ossified, but the notochord, ch, with a thicker and condensed capsule, remains. In the Piked Dog-fish (^Acanthias) the vertebral centres coincide in number with the neural arches, and are defined by a thin plate of bone, shaped like an hour- glass, and forming the conical cavity at each end of the centrum : the rest of which is cartilaginous external to the ' hour-glass,' and subgelatinous within its terminal cavities. In the Spotted Dog- fish (Scyllium) the two thin bony cones of each centrum are con- ANATOMY OF VERTEBRATES. 33 Vertical transverse section of centrum of Selache maxima fluent at their apices, which are perforated, and the notochord, reduced to a beaded form, is continued through them : the exterior of the bony cones is occupied by a clear cartilage. In the Porbeagle Shark (JLamna corniibica) further ossification of the conical plate has reduced the central communication to a minute foramen. Os- seous plates have also been developed in the exterior clear cartilage : these plates are triangular, parallel with the axis of the vertebra, their apices converging towards the centre : the interspaces are filled by cartilage. In the great Basking Shark [Selache maxima) fig. 26, the longitudinal bony laminae are more numerous and shorter than in Lamna, are peripheral in position, and extend about one-third of the way towards the centre of the interspace between the terminal cones, the rest being occupied by a series of concentric cylinders of bone, interrupted by four conical converging cavities, filled by cartilage ; of these, two, n, n, are closed by the bases of the neurapophyses, and two, p, p, by those of the parapophyses. There is a transition from the cylin- drical to the longitudinally lamellar structure, the exterior and largest of the cylinders sending out processes which join the in- ternal margins of the converging lamellae In the Monk-fish (Squatina) the osseous part of the centrum between the termi- nal cones is entirely in the form of concentric layers, few in number, and decreasing in breadth as they approach the centre. In the Cestracion there are no concentric cylinders, but only longitudinal Iamella3, radiating from the centre to the circumference, and giving off short lateral plates as they diverge. In the Topes ( Galeus), the Blue Sharks ( Carcharias), and in most sharks which possess the nictitating eyelid, may be seen the most advanced stao-e of ossification in the cartilaginous fishes : o o the entire centrum, save at the four cavities closed by the neur- and par-apophyses, is occupied by a coarse bone, more compact where it forms the smooth exterior surface and that of the ter- minal articular cavities. In osseous fishes (most Teleostomi) the neur- and par-apophysial cavities are obliterated by bone, and the neur- and par-apophyses are confluent, or suturally joined, with the centrum ; but they retain a greater proportion, than in higher classes, of the primitive gelatinous basis, which fills up the deep cone or cup at each end of the centrum, fig. 27, c c. Only in the ganoid Lepidosteus, among fishes, does ossification so extend VOL. i. D 34 ANATOMY OF VERTEBRATES. Scants Lepidosteus as to obliterate the front cavity, and protrude into the hind cavity of the preceding vertebra, fig. 28 ; thus establishing a cup-and- ball articulation on the ' opisthocoelian ' plan. The cup-and-ball structure prevails throughout the air-breathing, land-seeking, or terrestrial, Hcematocrya. So interlocked, the vertebra? are better fitted to support the body in air, and transfer its weight to legs. Sometimes the cup is behind, as in the land-salamander, the Surinam toad (JPipa), and some extinct crocodiles, thence called Streptospondylus ; but, as a general rule, existing reptiles have ( procoelian ' vertebra, or with the cup in front. In many extinct reptiles {Sauropterygia 9 Dinosauria) ossi- fication was so advanced as to leave no cavity at either end of the centrum ; and these parts were coarticulated by flattened or almost flat- tened surfaces, as in mammals. Finally, both extinct and recent Keptilia aiford instances in which the parts or elements of the vertebra have coalesced into one bone. The progressive stages in the developement of a vertebra, which have been illustrated by the chief of those at which it is arrested in the cold-blooded series, bear a close analogy to those by which it reaches the coalesced condition as a single bone in the warm-blooded classes. The principal secondary and adaptive modifications will next be pointed out which mark with special characters the collective trunk-vertebras in H&matocrya. 17. Vertebral column of Fishes.- -In the Sturgeon (Aci- penser), fig. 29, the first five or six neural arches are confluent with each other and with the parapophyses, forming a continuous sheath of firm cartilage (fig. 62), inclosing the fore part of the notochord, ib. , and myelon, and perforated for the exit of the nerves. The tapering end of the notochord is continued forward into the fused basal elements of the cranial vertebras, ib. g, g", and backward into the base and upper lobe of the tail-fin, fig. 29, c. The vertebras are represented by their peripheral elements, and principally by the neural and haemal arches. The pleurapophyses are limited to about twelve of the anterior trunk-vertebras, are articulated by simple heads to parapophyses, fig. 62, p, and rapidly shorten in the two or three hinder pairs ; the large ones sometimes consist of two or three pieces joined end on end, like the modified occipital rib, called f scapula.' Vegetative repetition of perivertebral parts ANATOMY OF VERTEBRATES. 35 not only manifests itself in the double pleur- and neur-apophyses on each side, but in small interneural and interhaemal cartilages, fig. 25. These peripheral cartilages are more feebly developed in Spatularia. In the Chimseroicls (Holocephali) the bases of the neur- and par-apophyses of about ten of the anterior trunk-vertebras coalesce and form a continuous accessary cartilaginous / o covering of the fore part of the notochord ; and the confluent neural spines here form a broad and high compressed plate. Between the neurapophyses are wedged accessory in- terneural cartilages. In NotidanuSyAcanthias, Centrina,and. Scym- nus, the interneurals, fig. 30, z, resemble the neurapophyses, ib. n, inverted, and are in- terposed, like wedges, between them, with the apices reaching the centrum. In Scyttium, Mustelus, Sphyrna, and Carcharias, the in- terneurals resemble the neurapophyses in size and shape, but occupy a position above the intervertebral joint. In Galeus the ( vegeta- tive repetition ' is further exemplified by four stellate points of ossification, one of which is intervertebral ; and above these are rudiments of neural spines. The spinal nerve directly perforates the neurapophysis ; or, when the two roots escape separately, one also per- forates the interneural. The pleurapophy- ses are short and simple cartilages, either wedged into the interspaces of the parapo- physes (Notidanus, Carcharias, Scymnus), or attached to the ends of the parapophyses ( Galeus) of, say, the twenty-six anterior verte- brae. In Acanthias there may be forty pairs of such riblets, fig. 30, pi. In the flat Plagiostomes (Skates, fig. 64, Rays, Torpedos) vegetative repetition mani- fests itself in the multiplication of vertebrae, and especially of the central elements ; which, as indicated by their rudimentary ossification in Chim&ra, are commonly more numerous D 2 \ o Skeleton of Sturgeon, (Acipenser Sturio). cxiv. 36 ANATOMY OF VERTEBRATES. than the neural arches ; nor are interneural and interhaemal pieces wanting. In Raia clavata these ( ossa intercalaria ' constitute the chief part of the neural arch, at the anterior part of the vertebral column ; whilst the neurapophyses resume their ordinary share in its formation at the posterior part of the column. In Zygcena there are interspinal cartilages. In Rhinobatus a single spine answers to two vertebral bodies, and we may well suppose this mul- tiplication of central pieces to have been carried still farther in the pri- maeval fossil Ray (Spinachorhinus) from the lower Lias. In the anchylosed cervical verte- brae of the Skate the short centrums are indicated by transverse bars along the middle of the under part. In the Monk-fish (^Squatina) the body of the atlas is confluent with the basioccipital, but the neural arch re- mains distinct. The parapophyses in most Rays pass forward, and then backward, the angle of one fitting, like an articular process, into the notch of the para- pophysis in advance : they do not support pleurapophyses ; they gradu- ally bend down behind the pelvic arch, and complete the haemal canal about six vertebra? beyond it; the ha3inal spines become flattened in the tail of some Rays. In osseous fishes a trunk-vertebra consists of a biconcave body, fig. 27, c, of a pair of neurapophyses, fig. 31, n, usually develop- ing a spine, ib. ns, from their point of coalescence above the neural canal ; and of a pair of parapophyses, ib. p ; to which are added in the abdominal region in most fishes, and also in the caudal region of some, a pair of pleurapophyses^ pi, figs. 31, 32. Ossification usually commences in the bases of the neur- and par-apophyses, and in the terminal cones of the centrum ; it may proceed to blend the six points into one bone, and fill Forepart of skeleton, Piked Dog-fish (Acanthias}. XLIII. ANATOMY OF VERTEBRATES. 37 31 32 Abdominal vertebras (Hugil) Abdominal vertebras, Pike (Esox) up the hollow outside the cones, as indicated by the dotted tract in the section, fig. 27. But, in some, a communicating aperture is left between the terminal cones, as indicated by the dotted line in fig. 31. In many fishes the plates by winch the bone attains the periphery of the centrum leave interspaces permanent- ly occupied by cartilage, forming cavities in the dried or fossil bone, or giving a reticulate surface to the sides of the centrum. The bases of the neur- and par-apophyses sometimes expand so as to wholly inclose the centrum before coalescing therewith ; as, for example, in the Tunny, where the line of demarcation may be seen at the border of the articu- lar concavity. In the Pike the neurapophyses seldom, in the Polypterus and Amia, never, coalesce with the centrum : the letter s shows the neurapophysial suture in fig. 32. In the Salmonidce the neur- apophyses remain distinct from both the centrum and from each other, in the anterior vertebra ; where each developes a long and slender spine. 1 The parapophyses remain for some time distinct from the body of the vertebra, as well as from the ribs. In the anterior vertebra of the Carp the neurapophyses remain distinct, as they do in the atlas of many other fishes, and a suture is ob- servable between the parapophyses and centrum in embryo Cypri- noids. In each vertebra the summits of the two neurapophyses usually become anchylosed together, and to their spine ; but in the Lepidosiren, fig. 41, the spine retains its character as a distinct element, and is always attached by ligament to the top of the neurapophysis, as it is in the Sturgeon, fig. 25. In the anterior abdominal vertebrae of the Tetrodon, each of the neurapophyses, though they coalesce in the interspace of the two spines to form the roof of the neural canal, sends up its own broad truncated spine ; and these are not much-developed oblique processes, but gradually approximate and blend together, to form the single normal spine at the fifth abdominal vertebra. 2 In the Barbel the neural arches also support two spines, but one is placed behind the other. 1 XLIV. vol. i. p. 16, No. 46. Ib. vol. i. p. 81. 38 ANATOMY OF VERTEBRATES. Terminal caudal vertebra?, Sword-fish, xxm. The interspaces of the neural arches are occupied by a fibrous aponeurosis the remains of the primitive covering of the neural axis : but in most fishes the arches are ad- j /^ /^> ditionally con- nected toge- ther by articu- lar or oblique process e s (zygapophy- ses ) : in the Pike the ante- rior one, fig. 32, 2, is present, which barely touches the neural arch in advance ; in Polypterus it overlaps that part. In the Perch a posterior zygapophysis projects to receive the overlapping anterior one, the relative positions being the reverse of those in most air- breathing vertebrates. But, in some fishes, a second pair of zygapophyses are developed, which resemble the normal pair in higher vertebrates in relative co-adaptation, but seem to grow as exogenous processes, from the centrum itself, fig. 31, z. It is also peculiar to fishes to have articular processes developed from the parapophyses, as, e. g. in the abdominal region of the Rays, and from the caudal vertebra? of the Sword-fish, fig. 33, z. In the Tunny these pi'ocesses are branched, and form a network about the haemal canal. In Loricaria peculiar accessory processes are sent out from the neural arch of the seven anterior ver- tebra3 which abut against the lateral shields of the dermo-skeleton. The parapophyses are short in some fishes (Sahno, Clupcea., Amia), of moderate size in many, and longest in the Cod-tribe, fig. 34, p, where they expand in the abdominal region and sustain the air-bladder which adheres to their under surface. In one species of Gadus, the bladder sends processes into deeper cavities of the parapophyses, foreshowing, as it were, the pneumatic bones of birds. The parapophyses gradually bend lower down as they approach the tail, where, in many fishes, they unite to form the haemal canal. In Lepidosteus the canal is formed by the pleura- pophyses : whilst these, in Amia., Thynnus, and some others, are appended to the parapophysial inverted arches, like haemal spines. In Lepidosiren the elements p, fig. 41, which in the abdomen represent either pleurapophyses or long parapophyses, bend down in the tail to form the haemal arch. Not until we reach the Batrachia in the ascensive comparison do we find true ' ha3ma- ANATOMY OF VERTEBRATES. 39 pophyses,' fig. 43, h, forming the haemal arch in the tail, and coexisting there with par- and pleur-apophyses, ib. p, and pi. The pleurapophyses of fishes correspond to what are termed in Comparative Anatomy, ( vertebral ribs,' and in Human Anatomy ( false or floating ribs : ' for, with few exceptions, of which the Herring is one, fig. 37, their distal ends are not connected with any bones analogous to sternal ribs or sternum ; i. e. the abdomen is unclosed below by the osseous parts completing the haemal arch. The true homologues of sternal ribs and sternum retain the primitive aponeurotic texture, and may be well seen in the Bream, ex- tendino: from the ends of the vertebral ribs. These elements, or o pleurapophyses, figs. 31, 32, pi, are usually appended to the extremities of the parapophyses, p, the articulation frequently pre- senting a reciprocal notch in each. But, in some bony fishes, as Platax, the ribs articulate with the bodies of the vertebrae, in de- pressions behind the parapophyses ; and in Polypterus beneath the parapophyses, as in the cartilaginous Heptanchus, Carcharias, and Alopias. Between the floating ribs extends an aponeurosis, the remains or homologue of the primitive fibrous investment of the abdomen in the Lancelet and Lamprey. In the Salmon and Dory the ribs continue to be attached to some of the parapophyses after they are bent down, as in the Amia and Tunny, to form the haemal canal V J and spine in the tail. The costal appendages of the first vertebra of the trunk are usually larger than the rest, and detached from the centrum ; at least if we regard as such the styliform bones which project from the inner side of the scapula?, and which have been described as coracoids (Cuvier), and sometimes as displaced iliac bones (Carus) : by the muscles attached to these styliform bones the succeeding ribs are drawn forward and the abdomen expanded in the Cyprinoids. Pleurapophyses are entirely absent in the Sun-fish, Globe-fish (Diodon), the Tetrodon, the Pipe-fish (Fistu- laria and Syngnafhus), the Lump-fish and the Angler. Of all osseous, or rather semi-osseous, fishes, Lophius presents the simplest vertebral column : the abdominal vertebrae are not only devoid of ribs, but have the feeblest rudiments of parapophyses. The bodies of the vertebra? interlock at their lower and lateral parts by a short angular process fitting into a notch in the next vertebra ; the lower border of this notch represents the lower transverse process in other fishes : it is obsolete in the anterior abdominal vertebra? ; begins to appear about the middle ones ; shows its true character in the tenth; and elongates, bending downward, backward, and inward, to coalesce with its fellow, and form the haemal arch at the twelfth or thirteenth vertebra, from which the haemal spine is 40 ANATOMY OF VERTEBRATES. developed. The interlocking process of the anterior vertebra dis- appears as the true inferior transverse process is increased. The side of the neural arch is perforated for the nerve, and that of the haemal arch for the blood-vessel. The anterior abdominal vertebra 3 , of the Tetrodon are firmly clamped to- gether by the para- pophyses. A vegetative same- ness of form prevails in fishes throughout the vertebral column of the trunk, fig. 34, which is made up of only two kinds of ver- tebras, characterised by the direction of the parapophyses, p : these in the abdomi- nal region are lateral, usually stand out and support ribs : but in the caudal region bend down to form, either by direct co- alescence or by the ribs that continue to be attached to them in a vertical position, the hremal arch. The atlas is usu- ally distinguished by some modification of the anterior articular end of the centrum, by the persistent suture of the neural arch, or by the ab- sence or detachment of its pleurapophy- ses. Peculiar pro- cesses are sometimes Skeleton of the Haddock (Gadus cegleflnus) SCnt off from the ANATOMY OF VERTEBRATES. 41 ' under part of the centrum, as, e. g. the two which articulate with the basioccipital in the Arapaima gig as. As the centrum of the atlas retains its normal relations to the other elements, and the ordinary mode of articulation with the body of the second verte- bra, this shows no ' odontoid process ' in fishes. The number of vertebrae varies greatly in the different osseous fishes : the Plectognathi (Diodon., Tetrodon) have the fewest and largest: the apodal fishes (Eels, Gyninotes) have the most and smallest, in proportion to their size. It is not easy to determine the precise number, on account of the coalescence of some of the vertebra, or at least of their central elements, in particular parts of the column. In- stances of anchylosis of some of the anterior vertebra?, analogous to that noticed in the cartilaginous Sturgeons, o o o y Chimaerae, Rhinobates, and some Sharks, occur also amongst the osseous fishes, as in many Siluroid and Cy- prinoid species, in Loricaria and Dactylopterus. Fig. 35 represents the four singularly elongated anchylosed ante- rior vertebrae in the Tobacco-pipe fish (Fistularia tabac- caria). A coalescence of several vertebrae is more con- stant at the opposite end of the column in osseous fishes, in order to form the base of the caudal fin, when this is symmetrical in form, as in fig. 33, and in most existing species of Teleostomi. But this modification is arrested at different stages in the piscine class. In Cyclostomi the gristly parts of the vertebrae continue distinct, with gradual reduction in size to the taper end of the long tail : in Protopteri the bony representatives of the caudal ver- tebrae behave in the same way: the notochord persists in both orders. In Mur&nidce, where it is changed into cen- trums, these also gradually diminish in size, and remain distinct to the tail-end. The continuous vertical fold of skin bordering the compressed, long, and slender termination of the vertebral column is not specialised as a caudal fin. 1 In Plagiostomi, Holo- cephali, Sturionidce, and many Ganoidei, the caudal fin, fig. 29, c, is formed chiefly by the haemal spines and appendages, developed to support a lower 6 lobe ;' the vertebrae continue distinct to the end of the tail, which bending upward, seems to form an upper lobe longer than the lower : to this unsymme- trical tail-fin the term ( heterocercal ' is applied. By decreased 1 This primitive embryonal basis of the piscine tail-fin is not to be confounded, because it is symmetrical as to shape, with the extreme stage of developemental modi- fication constituting the true ' homocercal ' type of most existing fishes. 42 ANATOMY OF VERTEBRATES. number, with progressive confluence, of the caudal vertebrae, the e upper lobe ' becomes gradually reduced in length, until the symmetrical shape is attained. But this coexists in the Salmon, Perch, and many extinct Ganoids with an unsymmetrical bend of the coalesced caudal vertebra? into the base of the upper lobe. In true ( homocercals ' the terminal bodies of the caudal vertebras are not separately established in the primitive notochord, but are continuously ossified to form a common, compressed, vertically extended, and often bifurcated bony plate, fig. 33, n'h', from which the neural and haemal arches and their spines o f* radiate : from these elements alone can the number of vertebras of such caudal fin be estimated ; normal de- velopement proceeding here in the peripheral elements, as throughout the vertebral column in Lepidosiren, whilst it is arrested in the central parts of the vertebrae. In the Sun-fish ( Orthagoriscus mola) it would seem as if a row of rudimental vertebra? had been blended together at right angles to the rest of the column, in order to support the rays of the short, but very deep caudal fin, which terminates the suddenly truncated body of this oddly shaped fish. It is rare to find anchylosis save at the ends of the vertebral series in fishes : sometimes, however, in the PleuronectidcB, a kind of sacrum is formed by such bony union of the bodies, c, and ha?mal spines, hs } of the first two of the caudal series, as in fig. 36 ; ! in which the broad and deep haemal spines are concave forwards, and form a sort of pelvic posterior wall of the abdomen. In the Halibut (JHippoglossus) the parapophyses of the corresponding vertebra? with those of the last abdominal are similarly united, though the bodies remain distinct. In Loricaria both the upper and lower arches of a con- siderable part of the caudal region are blended together into an inflexible sacrum ; but, as a general rule, there exists no such impediment to the lateral inflections of the tail in the present class. The number of trunk- vertebra? is a useful specific character in Ichthyology ; and in counting them the coalesced caudals are usually reckoned as ( one.' In the Sun-fish ( Orthagoriscus} I find but 8 abdominal and 8 caudal vertebra? by distinct bodies. In a Globe-fish ^ Tetrodon) there are 7 abdominal and 10 caudal vertebra? : 1 Osteol. Collection, Mus. Coll. Chir. No. 188 ; xuv. i, p. 50. ANATOMY OF VERTEBRATES. 43 total, 17. 1 In the Conger there are 162 vertebras ; in the Ophidium, 204 ; in the Gymnotus, 236 ; and even this number is surpassed in some Plagiostomes. Although the vertebras maintain a considerable sameness of form in the same fish,, they vary much in different species. The bodies are commonly subcylindrical ; as deep, but not so broad, as they are long ; more or less constricted in the middle, in some to such a degree as to present an hour-glass figure. In Spina- chorhinus they are extremely short ; in Fistularia extremely long ; in Tetrodon 2 they are much compressed ; in Platyceplialus they are more depressed ; in the tail of the Tunny the entire ver- tebra is cubical, 3 with the ends hollowed as usual, but the four other sides flat, the upper and lower ones being formed, in the connected series, by the neural and haemal arches of the vertebra in advance, flattened down and, as it were, pressed into cavities on the upper and under surfaces, of the centrum of the next vertebra; so that the series is naturally locked together in the dried skeleton ; and these arches cover not the neural and haamal canals of their own, but of the succeeding, centrum. The principle of vegetative repetition is manifested, in osseous fishes, by the numerous centres of ossification, from which shoot out bony rays affording ad- ditional strength to many of the intermuscular aponeuroses. In this system of bones may be ranked those spines which are attached to, or near to, the heads of the ribs, and extend upward, outward, and backward, between the dorsal and lateral masses of muscles, fig. 32, i p, fig. 21, pi, a. These ' scleral ' spines are termed, according to the vertebral element they may adhere to, ( epineurals,' ' epicen- trals,' and ( epipleurals ' ; though each may shift its place, rising or falling gradually along the series of vertebra?. All three kinds are present in the herring, fig. 37, in which n a is the ( epineural,' p a the ( epicentral,' pi a the epipleural spines. The latter have been called ' upper ribs,' and in Polypterus are stronger than the ('under') ribs themselves. In Esox and Thymallus the epineural and epicentral spines are present: in Cyprinus the epineural and epipleural ones : in Perca and Gadus the middle series only is found, passing gradually from the Abdominal vertebra, Herring (flupea) 1 Osteol. Collection, Mus. Coll. Chir. No. 357, p. 81. 2 Ib. No. 357. 3 Ib. No. 247. XLIV. i, p. 62. 44 ANATOMY OF VERTEBRATES. par- to the pleur-apophyses : in Salmo only the upper series exists, developed from the second to the antepenultimate abdo- minal neurapophysis, in S. Eriox. 1 There are, however, gristly representatives of epipleurals. In Gtyphysodon the epipleurals are anchylosed to the ribs, foreshowing their normal condition in the bird's thorax. According to the seat of their develope- ment they belong to the ( scleroskeleton : ' by their attachments to bone they are ( vertebral appendages.' The vertical folds of skin from the middle line, constituting the azygos fins, are the seat of ossifications in most fishes, develop- ing a second row of spines, figs. 34, 38, dn, dn, above the neural, n, and a corresponding row, dh, dk, below the haemal, h, spines. Some of these dermal bones, in certain fishes, project as hard enamelled weapons from the surface of the body. From the bases of the dermal spines, other spines (fig 34, in, Hi) usually shoot downward into the intervals of the neural and haemal spines. In deep-bodied fishes they are broad and strong, as e. g. in the Cock-fish, fig. 38 ; in the flat-fishes they are double, figs. 39 and 40 ; and these modifications are usually repeated above and below. Both interneural and interha3inal spines are commonly shaped like daggers, plunged in the flesh to the hilt, which is re- 38 Aryyrciosus seiipinnis presented by the part to which the fin-ray (dermoneural or dermohamial spine) is attached. In the plaice tribe (Pleuro- nectidce) these superadded dermal ossifications are developed above the cranial as well as the corporal vertebra? (fig. 39, dn), 1 XLIV. i, p. 16. CLIII. ANATOMY OF VERTEBRATES. 45 and along the whole haemal region of the trunk, from the head to the tail. This want of correspondence with the number of the true segments of the endoskeleton, and the seat of developement of the inter- and dermo-neurals and inter- and dermo-hsemals, with some minor considerations, led me, in 1845, to substitute for the views and illustration of the typical vertebra} proposed by GeofFroy St.-Hilaire, l and then accepted and taught by Professor R. E. Grant 2 in this country and by others abroad, the interpretation of the supposed type-exemplar, which is contrasted with Geoffroy's in fig. 40. The names applied by the French philosophical anatomist to the several parts of the combined endo- and exo-skeletal segment 39 Pleuronectes Solea. XLIII. are opposite the left hand of the reader : those applied to them in my ' Archetype of the Skeleton ' are opposite the right hand. The small exogenous process standing out from the sides of the centrum is a dismemberment of the parapophysis ; in the first caudal vertebra it is given off from the base of the parapophysis, increases in length in the second caudal, rises upon the side of the centrum in the third, and becomes distinct from the parapophysis in the fourth : it diminishes and disappears in the ninth and tenth caudal vertebra. In Polypterus and Murae- noids a transverse process coexists, from the same cause, with the parapophysis. This, in the twenty-fifth trunk-vertebra of Murcena Helena,* bifurcates, and in the following vertebrae the 1 Memoires du Mus. 4to, is. 1822, p. 119, pi. v. 2 Lectures on Comp. Anat. p. 58. 3 XLIV. p. 14. No. 37- 46 ANATOMY OF VERTEBRATES. o g i to a 6 ^ o p fissure deepens and the fork elongates, until at the seventy-third the lower prong descends at a right angle to the upper one, and, meeting its fellow, forms the hamial arch. There are no true 40 hrcmapophyses in the tail of fishes : the elements there composing the luemal arch are parapophyses, pleura- pophyses, or both combined. In the abdomen only are hremapophyses represented by the supporting bones of the ventral fins, fig. 41, 64. The slender ossicles along its under part in the Herring, fig. 37, dh, are dermal bones, which, like the scutes of serpents, are connected with the lower ends of the ribs, pi. In the subclass Protopteri the notochord, fig. 41, ch, persists : the neural arches, n, ns, are ossified : the haemal arches in the abdomen are represented by parial bones, p, attached to the notochordal sheath, and curving outward, like the long parapophyses in the Cod, and the short pleurapophyses in the Amia and Salamander, with which they, more probably, are homologous. These riblets bend down and o meet at the beginning of the tail, p, to form the ha3inal arch and support the ha3inal spines, hs, along that region. As in fishes, the Lepidosiren also cle- velopes in the continuous vertical fin-fold the acces- sory ossicles marked in, ih, in the cut. 18. Vertebral column of Batrachia. Neither inter- nor dermo-neurals are present in any gano- cephalan or batrachian. In the former amphibious order the notochord persists, but with beginnings of the ossification of centrums : } in Batrachia it is con- verted into a series of separate centrums. These in the Ichthyomorphs are biconical, and deeply cupped at both ends, through the same arrest of ossification as in fishes : the developement of the vertebra goes through the same piscine stage in the Iarva3 of the Theriomorphs, as indicated by the dotted lines d, fig. 42 ; in the mature quadrupedal stage of these Ba- trachia, ossification converts one terminal cup into a ball ; which may be the front one, as in Pipa, or the hind one, as in Rana, and most frogs and toads. In the Land- Salamander, also, ossification goes to this - and exo-ske- stage, with the ball in front. lotal elements of a caudal vertebra of the Plaice, n. 1 Lectures on Comp. Anat. p. 194, Fig. 84. a o ft ANATOMY OE VERTEBRATES. 47 The Siren lacertina has between eighty and ninety trunk-vertebrae. They have many longitudinal ridges, the neural arch has coalesced with the centrum, the neural spine forms the highest ridge and bifurcates posteriorly to terminate upon the zygapophysis. A 41 /is -m Skeleton of Lepidosiren anncctcns. xxxin. hypapophysial ridge forms, by defect of ossification on each side, the under part of the centrum. A parapophysial ridge extends from a short anterior parapophysis to the longer parapophysial part of the posterior transverse process. A diapophysial ridge extends above, and nearly parallel with the former, from the anterior zygapophysis to the diapophysial part of the posterior transverse process. Thence a third short ridge is continued to the posterior zygapophysis. The vacuities between these several ridges resemble those in the vertebras of some fishes. The body of the atlas extends forward like a short odontoid process : short par- arid di-apophysial plates are developed from each side of the atlas, which has also the posterior zygapophyses. In the second vertebra the par- and di-apophysial plates have united to form a compound 42 30 Skeleton of Tadpole of Itana csculenta transverse process, which supports a short straight pleurapophysis. These elements are similarly developed from six or seven succes- sive vertebrae. In the tail the vertebra is compressed and vertically extended by the bending down of the parapophysial plates to form two vertical walls, intercepting a haamal canal. In the Proteus, which has about sixty trunk-vertebras, the third to the ninth in- clusive support short ribs, attached to the lower (parapophysial) 48 ANATOMY OF VERTEBRATES. half of the transverse process : they are wanting in the twenty-one following vertebrae, and re- appear, well developed, in the thirty-first, where they form with cartilaginous haemapophyses, a pelvic arch. In the Menopome, fig. 43, the second to the nineteenth vertebrae support short straight pleurapophyses, articulated to the ends of transverse processes formed by par- and di- apophyses, which intercept by their terminal confluence an arterial canal. These processes, t, are enlarged in the twentieth vertebra, s, and a second rib-like piece, 62, the homotype of the second part of the scapula in fishes, is articulated to the short and thick rudimental rib, pi; the inferior or haemal arch 63, 64, beincf cartilaginous. * o o The segment thus completed by the haemal arch, represents a so-called ( sacral : vertebra : the second division of its rib answers to the ' ilium,' 62, and the haemal cartilage to the ( ischium,' or c pubis.' Transverse processes t, progressively decreasing in length are developed from the six succeeding vertebrae. Bony pleurapophyses pi, are attached to the first of these, and cartila- ginous rudiments of the same element to the three following. Haemal arches are anchylosed to the under part of the centrum of the second to the twelfth caudal vertebra inclusive, and these become more compressed to the end of the tail, for the support of a vertical fin. The neural arches are broad, depressed, anchylosed to the centrum : they are complete to the fourteenth caudal vertebra. The body of the atlas pre- sents an odontoid process between the two arti- cular surfaces for the occipital condyles ; it is deeply cupped behind, as are the succeeding vertebrae at both ends. This vertebra has neither di- nor pleur-apophyses. The skeleton of the Newt ( Triton) resembles that of the Menopome in its general characters ; the neural and haemal spines are more produced in the long tail, supporting there the chief swimming skeleton of the Menopome organ of this aquatic batrachian.^ In one kind are more developed, occasioning the sub- 8 m or Protonopsis ANATOMY OF VERTEBRATES. 49 Gl f" genus called Pleurodeles. In the land Salamander the backbone is strengthened by the ball-and-socket articulation of the trunk- vertebrae. Cuvier notices a curious inconstancy in the place of attachment of the pelvic arch, sometimes to the fifteenth,, some- times to the sixteenth, and in one instance sus- pended by the right pier to the sixteenth, by the left to the seventeenth, vertebra, in Salaman- dra atra. 1 The ophiomorphous batrachia are remark- able for the multiplicity, the theriomorphous for the paucity, of distinct vertebra? in the trunk ; these latter have the ball-and-socket articula- tion. The frog, fig. 44, A, has nine vertebrae and the coccygeal style c ; but by coalescence of this with the sacrum, and of the atlas with the second vertebra, in the Surinam toad (Pipa), the number of distinct trunk- segments is in that species reduced to seven. In Rana boans the atlas has no diapophy- ses ; but they are present and of great length in the succeeding vertebrae to the sacrum inclusive, where they are thick and support by their truncate ends two long rib-like bones, ib. A, 62, which expand at their distal ends, and unite there to two partially anchylosed bony plates, 64, which complete the haemal arch of the ninth segment of the trunk. The superior developement of this arch relates to the great size and strength of its appendages 1 CLI. torn. v. pt. ii. p. 413, E SC' TVii Skeleton of frog, A ; vertebra B and carpus c of toad. 50 ANATOMY OF VERTEBRATES. the hinder extremities- -in the tailless order, especially the frogs. In the seven vertebrae between the atlas and sacrum, two zyga- pophyses looking upward are developed from the fore part, and two looking downward from the back part of the neural arch ; there is also a short spine. In the Toad (Bufo vulgaris) the number of trunk-vertebrae, fig. 44, B, is the same as in the Frogs, but the diapophyses of the third and fourth vertebra? are relatively longer, those of the sacral vertebra, s, relatively shorter, broader, and expanded so as to over- lap the ilia, which are shorter and more arched. In Cystignathus pachypus the sacral diapophyses are subcylindrical. In Pipa the diapophyses of the second and third vertebra are of unusual length, and support semi-ossified, short, flattened pleurapophyses. The diapophyses of the four succeeding vertebras are short and slender ; those of the sacrum are more expanded than in the toad, and rest upon the anterior halves of the iliac bones. The coccy- geal style shows, in most anourans, a simple anchylosed neural canal, and also a haemal canal, as at h, D, fig. 44. In the Ophiomorphs ( Ccecilice) the vertebras, besides being very numerous, are biconcave. 19. Vertebral column of Ichthyopterygia. In an extinct order (Ichthyopterygia) of Dipnoal Reptiles, modified for marine life, but breathing air, the trunk-vertebras were very numerous, very short, and biconcave ; the centrums remained distinct from the neural and haemal arches, and were ligamentously, not sutur- ally, united thereto. In the Ichthyosaurus communis, fig. 105, there are about 140 vertebras ; in the anterior sixteen a short parapophysis is developed from the side of the centrum, and a diapophysis from the base of the neural arch ; but this soon begins to project from the neurapophysial border of the centrum, and then from the side of the centrum below that border. It continues gradually to sink in position until, at about the fortieth vertebra, it blends with the parapophysis, which alone continues to represent a transverse process, as far as at about the eightieth vertebra, 1 where it disappears and the succeeding centrums become compressed, indicating the vertical position of the dermal tail-fin which they supported. The atlas and axis centrums become anchylosed by flat surfaces ; but each supports its own neural arch. Between the lower part of the atlas and the occipital condyle is a wedge-shaped hypapophysis, representing the part called ' body of the atlas ' in anthropotomy : a similar bone is 1 A dislocation or fracture commonly occurred at this part between the death and final imbedding of the decomposing animal ; CLXI. ANATOMY OF VERTEBRATES. 51 wedged between the atlas and axis, a third between this and the third vertebra ; all tending to strengthen and stiffen, the part of the vertebral column sustaining the skull, and adding to its power of displacing the water in the agile movements of this ancient predatory aquatic animal. 1 As in Fishes, also, the continuity of the broad occiput with the trunk was uninterrupted by any cervical constriction. The ribs commence at the second vertebra, but by a bifurcate head ; and so continue, articulating with both par- and di-apophyses until the confluence of those processes, when they become single-headed. The ribs rapidly increase in length, which is greatest at the middle of the thoracic-abdominal cavity, and then gradually diminish to short and straight appendages, resem- bling detached transverse processes, in the tail. The longer ribs are grooved longitudinally ; their lower ends are united to ha^m- apophyses, subdivided into tAVO or three overlapping slender portions, the lowest articulating with a median transverse style, pointed at each end, representing the haemal spine, and completing the lijemal arch in the abdomen. In the tail the haemapophyses are simple, and attached by ligament, above to the centrum, and below to one another. 20. Vertebral column of Sauropterygia. In this extinct order of aquatic Reptiles the vertebral bodies had their terminal articular surfaces either flat or slightly concave, or with the middle of such cavity a little convex. In certain genera the neck-vertebra3 were uncommonly numerous ; this was remarkably so in the Plesiosaurus, fig. 45, in which those vertebra? consist of centrum, neural arch, and pleurapophyses. The latter are wanting in the first vertebra ; but both this and the second have the hypapophyses. The cervical ribs are short, and expand at their free end. They articulate by a simple head to a shallow pit, which is rarely supported on a process, on the side of the centrum. The body of the atlas articulates with a large hypapophysis below, with the neurapophysis above, with the body of the axis behind, and with part of the occipital condyle in front ; and all the articulations, save the last, may become obliterated by anchylosis. The hypapophysis forms the lower two-thirds, the neurapophysis contributes the upper and lateral parts, and the centrum forms the middle or bottom of the cup for the occipital condyle. The second hypapophysis becomes ^confluent with the inferior interspace between the bodies of the atlas and axis. 2 As the cervical vertebra? approach the dorsal, the costal pit gradually 1 CLXV. 2 CLXVI. E 2 52 ANATOMY OF VERTEBRATES. 45 df Skeleton of Plesiosaurus. CLXIII. rises from the centrum to the neurapophysis. This takes place at the fortieth vertebra in the Plesiosaurus homalo- spondylus of the Whitby Lias, but, in the PL doliclwdeirus, fig. 45, of the Dorsetshire Lias, at about the thirtieth, c. The dorsal region is arbitrarily commenced by the vertebra in which the costal surface begins to be supported on a diapo- physis ; this progressively in- creases in length in the second and third dorsal, continues as a transverse process to near the end of the trunk, and on the vertebra, s, between the iliac bones, 62, it subsides to the level of the neurapophysis. In the caudal vertebras the costal surface gradually descends from the neurapophysis upon the side of the centrum ; it is never divided by the longitu- dinal groove which, in most Plesiosauri, indents that sur- face in the cervical vertebras. The neural arches are com- monly unanchylosed with the centrum. The long and large spinous processes, in contact along the trunk and base of the neck, must have restricted the bending movements chiefly to the lateral directions. The pleurapophyses gain in length, and lose in terminal breadth, in the hinder cervicals ; and become long and slender ribs in the dorsal region, curving outward and downward so as to encompass the upper two- ANATOMY OF VERTEBRATES. 53 thirds of the thoracic-abdominal cavity. They decrease in length and curvature as they approach the tail, where they are reduced to short straight pieces, as in the neck, but are not terminally expanded ; they cease to be developed near the end of the tail. The hsemapophyses in the abdominal region, are subdivided, and with the haemal spine or median piece, form a kind of 6 plastron' of transversely extended, slightly bent, median and lateral, overlapping bony bars, occupying the subabdominal space between the scapular, 52, and pelvic, 64, arches. In the tail the hasmapophyses are short and straight, and remain, as in the Ichthyosaurus, ununited both above and below. One Sauro- pterygian genus, Tanystropheus, had the centrum, in certain vertebras, so long and hollow as to simulate a limb-bone. In another genus, (Pliosaurus) they were as short, in the cervical region, as in the Ichthyosaurus. In a third genus (Nothosaurus} two vertebras are recognised as sacral by their thick, straight, and convergent pleurapophyses, of which the first overlaps the second. In a fourth genus the wedge-shaped hypapophyses occur at the lower interspaces of the dorsal and lumbar vertebras, whence its name, Sphenosaurus. 21. Vertebral column of Ophidia. Amongst existing Reptiles, the Serpents ( Ophidia) surpass all others in the vast number of their vertebras, which, with incomplete hasmal arches, compose the skeleton of the long, slender, limbless trunk, fig. 46. In all these vertebras the autogenous elements, except the pleurapophyses, fig. 46, pi, coalesce with one another, and the pleurapophyses become anchylosed to the diapophyses in the tail. There is no trace of suture between the neural arch, fig. 47, ns, z, and centrum, c. The outer substance of the vertebra is compact, with a smooth or polished surface. The vertebras are ( proccelian ; ' that is, they are articulated together by ball-and- socket joints, the socket being on the fore part of the centrum, fig. 47 A, where it forms a deep cup with its rim sharply defined ; the cavity looking not directly forward, but a little downward, from the greater prominence of the upper border : the well-turned prominent ball terminates the back part of the centrum rather more obliquely, its aspect being backward and upward, fig. 47, c. The hypapophysis, h, is developed in different proportions from different vertebras, but throughout the greater part of the trunk presents a considerable size in the cobra, 46, hy, and crotalus, figs. 47, 47 A, h : it is shorter in the python and boa. A vascular canal perforates the under surface of the centrum, and there are sometimes two or even three smaller foramina. In the 54 ANATOMY OF VERTEBRATES. 46 23 python a large, vertically oblong, but short diapophysis extends from the fore part of the side of the centrum obliquely backward : it is covered by the ar- ticular surface for the rib, is convex lengthwise and convex vertically at its upper half, but slightly concave at its lower half. In the rattlesnake the diapo- physis developes a small, circumscribed, articular, tubercle, ib. ^/,-for the ( vertebral rib ' or pleur- apophysis, pl\ a parapo- physis, d! ', extends down- ward and forward below the level of the cen- trum ; the anterior zy- gapophysis, z, is sup- ported by a process, ib. d" 9 from the upper end of the diapophysis. The base of the neural arch swells outward from its confluence with the centrum, and de- velopes from each angle a transversely-elongated zygapophy sis ; that from the anterior angle, z, looking upward, that from the posterior angle, z' ', downward ; both sur- faces being flat, and almost horizontal, as in the Batrachians. The neural canal is narrow ; the neural spine, ns, is of moderate height, about Skeleton of the cobra (Naija tripudians} equal to its terior extent ; it is compressed and truncate. A wedge-shaped process, ' zygosphene,' zs, is developed from the fore part of the ANATOMY OF VERTEBRATES. 55 47 ns Two vertebra? of the Rattlesnake (Grotalus) base of the spine ; the lower apex of the wedge being, as it were, cut off, and its sloping sides presenting two smooth, flat, articular surfaces. This wedge is received into a cavity, the ' zygantrum,' excavated in the posterior expansion of the neural arch, and having two smooth articular surfaces to which the zygosphenal surfaces are adapted. Thus the vertebrae of serpents articulate with each other by eight joints in addition to those of the cup and ball on the centrum ; and interlock by parts reciprocally receiving and entering one another, like the joints called tenon-and-mortice in carpentry, fig. 47. In the caudal vertebras, the hypapophysis is double, the transition being effected by its progressive bifurcation in the posterior abdominal vertebrae. The diapophyses be- come much longer in the caudal vertebrae, and support in the anterior ones short ribs which usually become anchylosed to their extremities. The pleurapophyses or vertebral ribs have an oblong articular surface, concave above and almost flat below in the Python, with a tubercle developed from the upper part, and a rough surface excavated on the fore part of the ex- panded head for the insertion of the precostal ligament. They have a large medullary cavity, with dense but thin walls, and a fine cancellous structure at their articular ends. Their lower end supports a short cartilaginous hasmapophysis, which is attached to the broad and stiff abdominal scute. These scutes, alternately raised and depressed by muscles attached to the ribs and integument, aid in the glid- ing movements of serpents ; and the ribs, like the legs in the centipede, subserve locomotion ; but they have also accessory functions in relation to breathing and con- striction. The anterior ribs in the cobra, fig. 46, pi, are unusually long, and are slightly bent ; they can be folded back one upon another, and can be drawn forward, or erected, when they sustain a fold of integument, peculiarly coloured in some spectacled cobra - - and which has the effect of making this venomous snake more conspicuous at the moment when it is about to inflict its deadly bite. The ribs commence in the cobra, as in other serpents, at the third vertebra from the head. Front view of a vertebra, Rattlesnake species e.g., the 56 ANATOMY OF VERTEBRATES. The centrum of the first vertebra coalesces with that of the second, and its place is taken by an autogenous hypapophysis : this, in the python, is articulated by suture to the neurapophyses ; it also presents a concave articular surface anteriorly for the lower part of the basioccipital tubercle, and a similar surface behind for the detached central part of the body of the atlas, or ' odontoid process of the axis.' The base of each neurapophysis has an antero-internal articular surface for the exoccipital tubercle, the middle one for the hypapophysis, and a postero-internal surface for the upper and lateral parts of the odontoid ; they thus rest on both the separated parts of their proper centrum. The neura- pophyses expand and arch over the neural canal, but meet without coalescing. There is no neural spine. Each neura- pophysis developes from its upper and hinder border a short zygapophysis, and from its side a still shorter diapophysis. In the second vertebra, the odontoid presents a convex tubercle anteriorly, which fills up the articular cavity in the atlas for the occipital tubercle ; below this is the surface for the hypapophysial part of the atlas, and above and behind it are the two surfaces for the atlantal neurapophyses. The whole posterior surface of the odontoid is anchylosed to the proper centrum of the axis, and in part to its hypapophysis. The neural arch of the axis de- velopes a short ribless diapophysis from each side of its base ; a thick sub-bifid zygapophysis from each side of the posterior margin ; and a moderately long bent-back spine from its upper part. The centrum terminates in a ball behind, and below this sends downward and backward a long hypapophysis. At the opposite extreme of the elongated body, two or three much simplified vertebrae are usually found blended together ; they support the horny rings forming the warning rattle of the Cro- talus. There is no sternum in true Ophidia. The skeleton of the Python (P. tigris) 1 has 291 vertebras, of which the 3rd to the 251st support movable ribs. The 74 anterior vertebrae develope hypapophyses. The skeleton of the Boa constrictor* has 305 vertebras, a hypapophysis being developed from the 60 anterior ones. In the skeleton of a Rattle-snake (Crotalus horridus^ with 194 vertebras, 168 support movable ribs, and all these develope hypapophyses, fig. 47, h, as long as the neural spines, ns. In the Naja, fig. 46, as many vertebrae have the lower process, but of less length. In the Rough Tree-snake (Deirodon sealer)* with 256 vertebras, a hyj apophysis projects 1 XLIV. vol. i. No. 602, p. 123. 2 Ib., No. G30, p. 132. 3 Ib., No. 640, p. 135. 4 Ib., No. 638, p. 134. ANATOMY OF VERTEBRATES. 57 48 from the 32 anterior ones,, directed backward in the first ten, and forward in the last ten, where they are unusually long, and tipped with a coat of hard dentine ; these perforate the oesophagus, and serve as teeth. The jaws are merely roughened by rudiments of teeth. The relation of this singular condition of the cervical hypapophyses and the modification of the dental system to the food of the Deirodon will be explained in the chapter on teeth. 22. Vertebral column of Lacertia. - The anguine or snake- like reptiles, with fixed upper-jaws and a scapular arch, pass gradually, by other forms with rudiments of limbs (Pseudofms), to the slender-bodied lon^-tailed lacertians. The dis- o tinction is effected through the establishment of a costal arch in the trunk, completed by the addition of a hasmal spine (sternum) and haama- pophyses (sternal ribs) to the pleurapophyses or vertebral ribs, which are alone ossified in Ophidia. The vertebra of the trunk have the same procoelian character, i. e., with the cup anterior and the ball behind, fig. 48 ; the latter, c, being usually less prominent, more oblique, and more trans- versely oval than in serpents. The vertebras also are commonly larger, and always fewer in number than in the typical Ophidia. Those of the Iguanas retain the superadded articular surfaces of the zygosphene, fig. 48, zs, and zygantrum ; but I have not met with these superadded processes in other lacertians. In the 49 Trimk vertebra, Iguana Fore part of skeleton of a Lizard Geckos the vertebra are, exceptionally, biconcave. ! The ribs do not begin to be developed so near the head as in Ophidia, Not only the atlas and dentata, but the third vertebra, fig. 49, and sometimes, as in the Monitor ( Varanus)^ the four following verte- bras, are devoid of pleurapophyses : when these first appear they See those of the subgenus Rliynchocephalus, XLIV. vol. i. No. 662, p. 142. 58 ANATOMY OF VERTEBRATES. 50 are short, as at a ; but elongate in succeeding vertebras, b to e ; and usually at the eighth, or ninth, fig. 49, /, 6, (Lacertci), from the head or tenth ( Varanus), they are joined through the medium of ossified haema- pophyses to the ster- num. Two( Varanus),i\\TQQ ( Chameleo, Iguana), or four (Cyclodus), following ver- tebrae are similarly com- pleted; and then the haama- pophyses are either united below without intervening sternum ( Chameleo), or two or three of them are joined by a common cartilage to the cartilaginous end of the sternum. The haama- pophyses afterwards pro- ject freely, and are reduced to short appendages to the pleurapophyses. These also shorten, and sometimes suddenly, as, e. g., after the eighteenth vertebra in the Monitors ( Varanus), in which they end at the twenty-eighth vertebra, as they began, viz., in the form of short straight ap- pendages to the diapo- physes. The Draco volans, fig. 50, is so called on account of the wing-like expansions from the sides of its body, supported, like the hood of the cobra, by slender elon- gated ribs. In this little Skeleton of Draco volans j^^ ^^ ^ twenty vertebra? supporting movable ribs, which commence apparently ANATOMY OF VEKTEBKATES. 59 at the fifth. Those of the eighth vertebra first join the sternum, as do those of the ninth and tenth ; the plenrapophyses of the eleventh vertebra suddenly acquire extreme length ; those of the four following vertebras are also long and slender ; they extend outward and backward, and support the parachute formed by the broad lateral fold of the abdominal integuments. The pleurapo- physes of the succeeding vertebra? rapidly shorten. The sacrum consists of two vertebrae. There are about fifty caudal vertebra?. The semi-ossified sternum in the Iguana has a median groove and fissure, and readily separates into two lateral moieties. The long stem of the episternum covers the outer part of the groove, where it represents the c keel ' of the sternum in birds. The two sacral vertebra? retain, in most Lacertians, the cup- and-ball joints ; and in the Scincks, where they coalesce, the second presents a ball to the first caudal. Haamapophyses are wanting in the first caudal, commence in the second, but are displaced to the interval between this and the third ; they are confluent at their distal ends, and there produced into a spine : these ' chevron bones ' are continued usually along two-thirds of the tail. In most of the caudal vertebra? the anterior third of the centrum is marked off by a line, just anterior to the dia- pophyses, where the tail snaps off, when a lizard escapes, leaving the part that has been seized in the hands of the baffled pursuer. The ossification of the centrum from two points, and their in- complete anchylosis has prospective relation to the liability of lizards to be caught by their long tail, and lends itself to their escape. The epiphysial line does not extend through the thin and brittle neural arch, which readily snaps when the two parts of the centrum to which it is anchylosed are separated. Lizards reproduce the lost tail ; but the vertebral axis is never ossified in the new-formed part. In the slow-worm (Anyuis) there are 111 vertebra?, 61 of which, beginning at the fourth, support free ribs. The transverse pro- cesses of the tail are formed by short anchylosed pleurapophyses, which are bifurcate in the second and third caudals. The hypa- pophyses are, also, anchylosed to the centrum ; but, instead of remaining distinct, as in true Ophidia, they unite at their lower ends and complete the haamal arch. The vertebra? of the Amphis- bcena have no neural spine. The lacertian modifications of the atlas and axis 1 agree in the 1 XLIV. vol. i. pp. 139 149. GO ANATOMY OF VERTEBRATES. main with those in the Python. In Istiurus the cervical hypa- pophyses are compressed, distinct, and articulated to the inter- space between their own vertebra and the one in advance. The caudal vertebras are remarkable for the great length of their neural spines. In the Chameleon the ribs commence at the fourth vertebra, and those of the sixth are articulated by semiossified cartilages to the sternum, as are the three following pairs ; in the next eight or ten pairs the long and slender cartilages meet and unite to- gether at their extremities. There are two lumbar and three sacral vertebras ; the tail is long and prehensile. In the Iguana tuberculata twenty-one vertebras, commencing with the fifth, support free ribs, and those of the ninth first join the sternum. 23. Vertebral column of Chelonia.- -This column is most ex- 51 Skeleton of Emijs Enroptca. xxxvin. traordinary in those Reptilia to which, in the manifold modifica- tions of the organic framework, has been given a portable abode, in compensation for inferior powers of locomotion and the want of defensive weapons. ANATOMY OF VERTEBRATES. 61 The expanded thoracic-abdominal case, fig. 51, K, K, into which, in most Chelonians, the head, the tail, and the four ex- tremities can be withdrawn, and in some of the species be there shut up by movable doors closely fitting both the anterior and posterior apertures- -as, e.g., in the box-tortoises (Cino- sternon, Cistudo)- -}uas been the subject of many investigations; and not the least interesting result has been the discovery that this seemingly special and anomalous superaddition to the ordinary vertebrate structure is due, in a great degree to the modification of form and size, and, in a less degree, to a change of relative position, of ordinary elements of the vertebrate skeleton. The natural dwelling-chamber of the Ckelonia consists chiefly, and in the marine species (Clielone) and soft-turtles (Trionyx) solely, of the floor and the roof: side-walls of variable extent are added in the fresh-water species (Emys) and land-tortoises ( Tes- tudo). The whole consists chiefly of osseous ' plates ' with superincumbent horny ( scutes,' except in Trionyx said. Sphargis, in which these latter are 52 wanting. ffi7 10 Carapace of the Loggerhead Turtle (Chelone caouannci) The roof, or ( carapace,' fig. 52, consists of a ' median ' series of symmetrical plates, ch, s i to 511, and of two ' lateral ' series forming a pair, pi i to pi 8, the whole being surrounded by a circle of ' marginal ' pieces, in i to py, completed anteriorly by ch, the first of the median series. Of the median series eight, s i to s 8, are attached to the spines of eight subjacent vertebra? : the lateral or parial plates, pi i to pi 8, are attached to, and more or less blended with, the ribs of the same vertebra? ; and the ends of these ribs usually articulate by gomphosis with a corresponding number of the marginal pieces, of which, however, there may be from twenty-four to twenty-six, including the two median and symme- trical ones, ch and py. That these marginal pieces are the least essential parts of the carapace is shown, not only by their incon- stant number, but by their partial or total absence in some of the soft-turtles (Gymnopus, Sphargis). 62 ANATOMY OF VERTEBRATES. The median pieces, s i to s 11, are called the f neural' plates; the lateral pieces, pi i to pi 8, the ' costal ' plates ; the term 6 marginal ' is restricted to those peripheral pieces which form pairs, m i to in 12 ; the anterior symmetrical piece, ch, constant in all Chclonia, is called the ' nuchal ' plate ; the posterior symmetrical piece, py, which is wanting in all the Trionycidce, is the f pygal ' plate. The neural arch, connate with the first neural plate, s i, is supported partly by the centrum of the vertebra to which the first pair of free ribs is articulated, and which, therefore, is reckoned as the first dorsal vertebra : these ribs are small and slender, attached at both their extremities, the outer end abutting against ' ~ C? the under part of the first pair of costal plates, which they help to sustain. The second to the ninth dorsal vertebrae inclusive, being those which are more immediately connected with the neural and costal plates, are the ( vertebra of the carapace : ' their characters, though not less artificial than those which distinguish the ( dorsal ' or ( lumbar ' vertebrae of other reptiles, are much more marked and constant. The eighth vertebra of the cara- pace is succeeded by one, winch in some species (e. g. Clielone caouanna) supports a pair of short ribs, in others ( Trionyx) none, and which is therefore reckoned a ' lumbar ' vertebra ; this is followed by two other vertebra?, with short and thickened ribs, abutting against the iliac bones and representing the e sacrum,' fig. 51, G : as these three vertebras are not immediately united with the ninth, tenth, and eleventh ( neural plates,' they have less claim than the first dorsal vertebra to be regarded as entering into the composition of the carapace. The ' plastron,' fig. 53, or floor of the thoracic-abdominal chamber, consists, in all recent Chelonia, of nine pieces. The median and symmetrical piece, s, is the ( entosternal ; ' the four pairs, counted from before backward, are respectively, the ( episternals ' (es), ' hyosternals ' (hs,) ( hyposternals ' (ps), and ( xiphisternals ' (xs). In all the Chelonians, save the coriaceous (Sphargis) and soft turtles ( Trionycidcs), the outer surface of the carapace is impressed by the horny scutes, commonly called ( tortoise-shell ; ' and these epidermal productions have received definite names in Zoological Treatises, their modifications being found of great use in charac- terising species. In fig. 52, v\ is placed on the first c vertebral scute ' close to its union with the first and second ( costal scutes ; ' and v 2 to v 5 indicate the succeeding vertebral scutes, the outer angles of which are similarly wedged between the adjoining pairs of ( costal scutes : ' beyond the costal scutes are a series of ( mar- ANATOMY OF VERTEBRATES. 63 53 ginal scutes,' supported by the marginal plates, and crossing their sutures. In the Trionycidce the exterior surface of the carapace and plastron is remarkable for its rough vermicular or punctate sculpturing. The median bony pieces of the carapace, fig. 52, ch, s i to s 11, have been regarded as lateral expansions of the summits of the neural spines ; the medio-lateral pieces, ib. pi i to pi 8, as similar developements of the ribs ; and the marginal pieces ib. m\ to m 13, as the homologues of the sternal ribs. But the developement of the carapace shows that ossification begins independently in a fibro-cartilagmous matrix of the corium in the first, ch, and some of the last, s 9 to s 11, median plates, and extends from the summits of the neural spines into only eight of the in- tervening plates, s i to s 8 : ossification also extends into the contiguous lateral plates, pi i to pi 8, in some Chelonia, not from the corresponding part of the subjacent ribs, but from points alter- nately nearer and farther from their heads, 1 showing that such extension of ossification into the corium is not a developement of the tubercle of the rib, as has been supposed. Ossification commences independently in the corium for all the marginal plates, in i to py ; these never coalesce with the bones uniting the sternum with the vertebral ribs, are often more numerous, sometimes less numerous, than those ribs, and in a few species are wanting. Whence it is to be inferred that the ex- panded bones of the carapace, which are supported and impressed by the thick epidermal scutes called 'tortoise-shell,' are dermal ossifications, homologous with those which support the nuchal and dorsal epidermal scutes in the crocodile. Along the under surface of the costal plate the slender or proper portion of the rib may be traced, of its ordinary breadth to near the head, which liberates itself from the costal plate, as at I, fig. 51, to articulate to the in- terspace of the two contiguous vertebra, to the posterior of which such rib properly belongs. In the ' plastron,' fig. 53, the entosternal, s, answers to the sternum in the crocodile : the parial pieces are ( haemapophyses ' or Plastron of Cfielone caouanna 1 CLXII. p. 163, pi. xiii. fig. 4. 64 ANATOMY OF VERTEBRATES. sternal ribs, connate in a more or less complete degree with dermal bony plates. There were five pairs in the extinct Pleurosternon. In the marine Chelonians the dermal ossifications, fig. 52, pi i to 8, do not cover the whole of the intercostal spaces ; the slender ribs project beyond them. In the fresh-water and land kinds they extend to the marginal plates and complete the bony roof, as in fiir. 51. There is a similar difference in the decree of ossification O tJ of the ( plastron ' between the genus Chelone and the genera Emys and Testudo. In the Chelonia the true centrum of the atlas does not coalesce, as an ' odontoid ' process, with that of the axis, and usually supports its own neural arch : the hypapophysis is proportionally reduced. 1 All the eight cervical vertebra, fig. 51, E, are free, movable, and ribless : the fourth of these vertebras has a much elongated centrum, which is convex at both ends : the eighth is short and broad, with the anterior surface of the body divided into two transversely elongated convexities, and the posterior part of the body forming a sino-le convex surface divided into two lateral facets : the under o part of the centrum is carinate ; the neural arch, which is anchylosed to this centrum, is short, broad, obtuse, and overarched by the broad expanded nuchal plate. The first dorsal vertebra is also short and broad, with two short and thick pleurapophyses, articulated by one end to the expanded anterior part of the centrum, and united by suture at the other end to the succeeding pair of ribs. The head of each rib of the second pair is supported upon a strong trihedral neck, and articulated to the interspace of the first and second dorsal vertebras : it is connate, at the part corresponding to the tubercle, with the first broad costal plate, which articulates by suture to the lateral margin of the first neural plate, and to portions of the nuchal and third neural plates : the connate rib, which is almost lost in the substance of the costal plate, is continued with it to the anterior and outer part of the carapace, where it resumes its subcylindrical form, and articulates with the second and third marginal pieces of the cara- pace. The neural arch of the second dorsal vertebra is shifted forwards to the interspace between its own centrum and that of the first dorsal vertebra. A similar disposition of the neural arch and of the ribs prevails in the third to the ninth dorsal vertebras inclusive. The bony floor of the great abdominal box, or ( plastron,' is formed by the hasmapophyses and sternum connate with dermal osseous plates, forming, as in the turtle, nine pieces, 1 CLXVII p. 435, pi. xiii. ANATOMY OF VERTEBRATES. 65 but they are more ossified, and the hyo- and hypo-sternals unite suturally with the fourth, fifth, and sixth marginal plates, forming the side-walls of the bony chamber cut through in fig. 51. The junction between the hyo- and hypo-sternals admits of some yielding movement. The iliac bones abut against the pleurapo- physes of the tenth, eleventh, and twelfth vertebras, counting from the first dorsal vertebra. These three vertebra form the sacrum : their pleurapophyses are unanchylosed, converge, and unite at their distal extremities to form the articular surface for the ilium. Beyond these the caudal vertebrae, ib. H, thirty-five in number in Testudo elephantopus, are free, with short, straight, and thick pleurapophyses, articulated to the sides of the anterior expanded portions of the centrums. They diminish to mere tubercles in the tenth caudal vertebra, and disappear in the remainder. The neural arches of the caudal vertebra? are flat above, and without spines. 24. Vertebral column of Crocodilia. - - In this order free pleurapophyses are developed from all the cervical vertebras ; that of the atlas, fig. 54, #, is attached to the hypapophysis ; the neur- apophyses rest, in part upon this element, in part upon the proper centrum, which coalesces with that of the axis : the neural spine of the atlas remains distinct, like that of the occiput, and is broad and flat. The centrum of the axis is flat in front, and convex behind : the neural arch, as in the succeeding vertebra, is com- pleted by the connate spine. The pleurapophysis, ib. b, has a bifurcate head. "\Vith the exception of the two sacral vertebras, which are flat at one end and concave at the other, and of the first caudal vertebra, which is convex at both ends, the bodies of all the vertebras beyond the axis are concave in front and convex / behind. The procoslian centrum of the third cervical is shorter but broader than the second ; a parapophysis is developed from the side of the centrum, and a diapophysis from the base of the neural arch ; the pleurapophysis is shorter, its fixed extremity is bifid, articulating to the two above-named processes ; its free extremity expands, and its anterior angle is directed forward to abut against the inner surface of the extremity of the rib of both the axis and atlas, whilst its posterior prolongation overlaps the rib of the fourth vertebra. The same general characters and imbri- cated coadaptation of the ribs, not given in the diagram, 54, characterize the succeeding cervical vertebras to the seventh inclusive, fig. 57, p, the hypapophysis progressively though slightly increasing in size. In the eighth cervical the rib, h, becomes elongated and slender ; the anterior angle is almost or quite suppressed, and the posterior one more developed and produced VOL. I. F 66 ANATOMY OF VERTEBRATES. 54 more downward, so as to form the body of the rib, which termi- nates, however, in a free point. In the ninth cervical, the rib, ?', is increased in length, but is still what would be termed a ' false ' or l floating rib ' in anthropotomy. In the succeeding vertebra the pleurapophysis, fig. 54, k, articulates with a hrcmapophysis, and the haemal arch is completed by a haemal spine ; by which completion of the typical segment we distinguish the commencement of the series of dorsal vertebras. With regard to the so-called ' perforation of the transverse process ' this equally exists in the pre- sent vertebra, as in the cervicals ; on the other hand, the cervical vertebras equally show surfaces for the articu- lation of ribs. The typical characters of the segment, due to the completion of both neural and haemal arches, are continued in some species of Crocodilia to the sixteenth, in some (Crocodilus acutus) to the eighteenth vertebra. In the Crocodilus acutus and the Alligator lucius the hremapophysis of the eighth dorsal rib (seventeenth segment from the head) joins that of the antecedent vertebra. The pleurapophyses project freely outward, and become * floating ribs ' in the eighteenth, fig. 55, b, nineteenth, ib. c, and twentieth, ib. d, vertebrae, in which they become rapidly shorter, and in the last appear as mere appendages to the end of the long and broad diapophyses : but the haemapo- physes by no means disappear after the solution of their union with their pleurapophyses ; they are essentially independent elements of the segment, and are accordingly con- tinued, in pairs, fig. 55, 3, 4, 5, 6, 7, and 56, along the ventral sur- face of the abdomen of the Crocodilia, as far as their modified homotypes the pubic bones, ib. 8. They are more or less ossified, and are generally divided into two or three pieces. A short carti- laginous piece, an unossified part of the pleurapophysis, intervenes Diagram of anterior vertebras, Crocodile, cc. ANATOMY OF VERTEBRATES. 67 between it and the hasmapophysis. A small cartilaginous appen- dage is attached to some of the ribs. The lumbar vertebrae are those in which the diapophyses cease to support moveable pleurapophyses, although they are elongated by the coalesced rudiments of such, ib. e, f, y, h, which are distinct in the young Crocodile. The length and persistent individuality of more or fewer of these rudiniental ribs determines the number of the dorsal and lumbar vertebrae respectively, and exemplifies the purely artificial character of the distinction. The number of vertebras between the skull and the sacrum is twenty-four. In the skeleton of a Gavial, I have seen thirteen dorsal and two lumbar; in that of a Crocodilus cataphractus twelve dorsal and three lumbar ; in those of a Crocodilus acutus and Alligator lucius, eleven dorsal and four lumbar, fig. 57, which is the most com- mon number. Cuvier assigns five lumbar vertebrae to Croc. 55 Diagram of posterior trunk-vertebra, Crocodile, cc. biporcatus. But these varieties in the developement or coales- cence of the stunted pleurapophysis are of no essential moment. The coalescence of the rib with the diapophysis obliterates of course the character of the ' costal articular surface,' which we have seen to be common to both dorsal and cervical vertebrae. The lumbar zygapophyses have their articular surfaces almost horizontal, and the diapophyses, if not longer, have their antero- posterior extent somewhat increased ; they are much depressed, or flattened horizontally. The sacral vertebras, fig. 57, s, are very distinctly marked by the flatness of the coadapted ends of their centrums ; there are never more than two such vertebras in the Crocodilia, recent or extinct : in the first the anterior surface of the centrum is concave, in the second the posterior surface; the zygapophyses are not obliterated in either of these sacral vertebras, so that the aspects of F 2 68 ANATOMY OF VERTEBRATES. 56 their articular surface upward in the anterior pair, downward in the posterior pair determines at once the corresponding ex- tremity of a detached sacral vertebra. The thick and strong transverse processes form another characteristic of these vertebra? ; for a long period the suture near their base remains to show how large a proportion is formed by the pleurapophysis. This element, fig. 55, z, articulates more with the centrum than with the diapophysis developed from the neural arch ; it ter- minates by a rough, truncate, ex- panded extremity, which almost or quite joins that of the similarly but more expanded rib, ib. k, of the other sacral vertebra. Against these extremities is ap- plied a supplementary costal piece, serially homologous with the fibrous tract indicated by the dotted lines between h and 7, (/ and 6, fig. 55 ; but ossified, expanded, and interposing it- self between the pleurapophyses and ha3inapophyses of both sacral vertebras, not of one only. This intermediate pleurapophy- sial part is called the ' ilium ' fig. 57, 62 : it is short, thick, very broad, and subtriangular, the lower truncated apex form- ing with the connected extrem- o ity of the haamapophysis an arti- cular cavity for the diverging appendage, called the ( hind leg.' The hasmapophysis of the anterior sacral vertebra is called ' pubis,' fig. 55, 8, fig. 56, 5 ; it is moderately long and slender, but expanded and flattened at its lower extremity, which is directed forward toward that of its fellow, and joined to it through the intermedium of a broad, cartilaginous, hrcmal spine, ib. 10 and 11, completing the haemal canal. The hremapophysis of the second sacral, fig. 55, 9, fig. 56, 4, is broader, subdepressed, and subtriangular, expanding Diagram of the hremal arches of the trunk, viewed from above. Crocodile, cc. ANATOMY OF VERTEBRATES. 69 as it approaches its fellow to complete the second pelvic hremal arch. The size of these elements of the hsemal arch, and their distinctive shapes, have obtained for them, in anthropotomy , special names : their diverging appendage being developed into a potent locomotive member. The crocodile yields a clear view of the serial homolosnes of the haemal elements alons; the trunk. In fio-. 56, O O o ' they are sketched as seen from the dorsal aspect. The haemapo- physes extend from h i, 2, 3, to h 6, 5, 4 : the haemal spines, mostly confluent, are co-extensive from hs to 10, where they expand as a cartilage between 6 and 5. The pair of hsemapophyses, h i, are called ' coracoids,' and bear the special number 5 2: the pair, 5, are the ( pubic bones ' ; the pair, 4, the e ischia.' The haemal spine, hs, is called ( episternum,' the succeeding more or less confluent spines, 9, form the ' sternum ' : in Man their abdominal continuation, not quitting the fibrous tissue-state, is called f linea alba ' ; it be- comes cartilage in the Crocodilia, ib. 10, and partly bony in old specimens. The abdominal haemapophyses, represented by the 6 intersectiones tenclineae musculi recti abdominis ' of anthropotomy, are commonly ossified, each from two centres, in old Crocodilia. The pleurapophysis is reduced to a transverse process in the first caudal vertebra, fig. 55, / ; which, besides being biconvex, has no articular surface for the haemapophyses : these elements reappear in the succeeding segments, detached, as in the lumbar series, from their pleurapophyses, but articulated to the centrum directly, fig. 7, with a backward displacement, to the interspace between their own and the succeeding vertebra, fig. 57, //. After the fourteenth caudal vertebra the transverse processes disappear, the centrum becomes compressed, and the neural and haemal spines give adequate vertical extent to the long and strong nata- tory tail, to near its pointed termination. The characters of the trunk-vertebras of existing Crocodilia, especially their proccolian type, are those which their predecessors presented throughout all the tertiary series of deposits, 1 and by some species from cretaceous beds. 2 But in all the secondary series below the chalk, the Crocodilia present flattened or sub-con- cave vertebral surfaces; or, if the cup-and-ball structure be present, it shows reverse positions to the procoelian type, e. g. in the anterior trunk-vertebrae of the genus of oolitic Crocodilian, thence termed 4 Streptospondylus? A similar e opisthocoelian ' modification is presented by the cervical and anterior dorsal vertebrae of the more gigantic Cetiosaurus ; and, in a minor degree, i. p, 117, pis. 1 D, 3, 3 A. ^? O r O /IT ~V T TT V\ Q Q O 1 CLXIII., part iii. p, 117, pis. 1 D, 3, 5 2 Crocodilus basifissus, CLXtv. p. 380. 70 ANATOMY OF VERTEBRATES. ,5, N 57 by some of the great reptiles, with limbs more adapted for terrestrial pro- gression, called ' Dinosauria' ; favour- ing in these the flexibility of the neck, as the same ball-and-socket structure does in the large herbivorous quadru- peds of the present day. 1 The neural arch in the dorsal region of Dinosauria^ was enlarged and strengthened by a bony platform, with supporting ridges : the sacrum included from four to six vertebras, having the neural arch shifted so as to rest upon two cen- trums and bind them together. 25. Vertebral column of Ptero- sauria. - - In tracing the modifications of the skeleton from the earliest forms of extinct species, the procoelian type of vertebra appears first in the extinct group of Reptiles (Pterosaurid) adap- ted for flight ; the Pterodactyles of the Lias show it, with a confluent neural arch and a pneumatic foramen on each side of the vertebra. 2 The cervical ver- tebrae of Pterodactyles, fig. Ill, are the largest, seven or eight in number, of which the first two coalesce. The atlas has a very short discoid centrum and two slender neurapophyses. The dorsal vertebrae become smaller to the pelvis ; they may be fifteen in number, fol- lowed by two lumbar, from three to seven sacral, and a variable number of caudal vertebras. One family of Pterodactyles had a long and stiff tail ; the rest, as in fig. Ill, a short tail. The anterior free ribs have bifurcate heads ; and, as this structure is asso- ciated in modern Reptilia, with a four-chambered heart, that organ had probably reached the same stage of skeleton of Alligator perfection in the flying Reptiles, the 1 CLXIII. parts v. and vi. ; xxx. p. 285. 2 CLII. p. 161, pi. x. ANATOMY OF VERTEBRATES. 7! huge terrestrial Dinosaurs, and other extinct groups with the same costal structure. The existing Reptilia are but a remnant of a once extensive and varied class of cold-blooded vertebrates, which, since the mesozoic epoch has been on the wane. 1 26. Developement of the skull. In reviewing the modifications of this part of the vertebral column in the H&matocrya, we retrace our steps to the lowest water-breathing forms, and recommence with the Dermopterous subclass. Passing from the trunk to the head, we find in the Lancelet (Branchiostoma), fig. 23, that the cranium is not indicated by difference of size or structure of the rudimental vertebral column, but consists of the gradually contracting anterior termination of the neural canal, which retains its primitive fibro-membranous wall, 71, ob, without any superaddition of parts, and is supported by the tapering end of the notochord, ib. ch. This part extends farther forward than the cranial end of the neural canal, indicating the 7 o iion-developement of the prosencephalon and corresponding part of the cranial cavity. In fact, there is no ganglionic cerebral expansion whatever in this vermiform fish : the epencephalon or medulla oblongata is indicated by the origin of the trigeminal nerve, ib. ob, in advance of which the mesencephalic segment sends off the short optic nerve to the dark ocellus, op, and there terminates, somewhat obtusely, beneath what Dr. KoLLiKEE, 2 has described as a ciliated olfactory capsule, ib. ol. The cranium of the Lancelet, therefore, may be said to be composed of the notochord and its membranous capsule, without the superaddition of cartilaginous or osseous coverings. But, as an appendage to the skull, may be described the jointed, cartilaginous, hannal arch, ib. h, which extends from below the cranial end of the chorda dorsalis, down- ward and backward on each side of the orifice of the pharynx ; this represents the labial arch of higher Myxinoids, and supports several pairs of the jointed slender oral filaments. It is the sole chondrified part of the skeleton in the Branchiostoma. The cartilaginous tissue is superinduced upon the fibrous brain- sac in osseous fishes, in the following manner. The notochord advances as far as the pituitary sac, or f hypophysis cerebri,' where it terminates in a point ; cartilage is developed on each side, forming a thick f occipito-sphenoidal ' 3 mass, which extends out- ward, and forms the earball or acoustic capsule. The cartilage rises a little way upon the lateral walls of the cranium, and is there insensibly lost in the primitive cranial membrane. At the 1 CLXXX. p. 320. 2 xxxii. p. 32. 3 Plaque nuchale, Vogt ; Knocherne basis cranii, Miiller, xxi. Muller 72 ANATOMY OF VERTEBRATES. end of the notochord the basal cartilages, developed in continua- tions of its capsule, diverge, surround the pituitary vesicle, and meet in front of it, forming the f sphenoidal arches/ l which join, or expand into the ' vomerine plate.' 2 The immature Lamprey, called Sand-lance (Ammocoetes}, retains a like condition of the skull, fig. 58, to the second or third year. The occipital cartilages extend from the sides of the pointed end of the notochord, ib. ch, and expand into the acoustic capsules, ib. 16: the sphenoidal arches, ib. 5, encom- pass the pituitary or hypophysial space, Ay, now closed by a membrane-cartilaginous plate, and unite anteriorly to form a small vomerine plate, ib. is, in front of which is the single undivided nasal capsule, ib. 19. The now expanded cerebral end of the neural canal, fig. 59, ft, is still defended by fibrous membrane only; but is divided from the vomerine plate, ib. 13, by a Base of skull, t i i backward extension of the nasal sac, ib. 19, to the pituitary vesicle. In the Myxine the acoustic capsules are approximated at the base of the skull ; the sphenoidal arches are longer, and unite with the palatine plate and arches, from which are sent off the labial cartilaginous processes supporting the buccal tentacles homologous with those in the Lancelet. In the long hypophysial interspace of the sphenoidal arches a more or less firm cartilaginous plate is developed, from which a slender median process is continued for- ward to the vomerine or palatine plate, which supports the nasal capsule ; another side view of stun, Ammocete, process extends backward to the occipital Miiller 1 cartilage. Other processes are also sent off from the sides, which form a complex system of peculiarly Myxinoid cartilages. 3 In the mature Lamprey (Petromyzoii)., fig. 60, the occipital cartilage is continued backward, in the form of two slender processes, -i of the skuii. Larva 43, 10, perforated by the optic nerves, are ossmed in the cartilaginous basis, as are those of the fourth segment (prefrontals), figs. 42, 44, 68, u, perforated by the olfactory nerves ; whilst those of the second segment, 6 ali- sphenoids,' ib. 6, perforated by the trigeminal, longer remain gristly. All the chondrogenous elements are thick bones. From the membranous basis of the skull are developed the following bones, which are more or less lamelliform. The basi- occipito-sphenoidal plate, fig. 73, m, forms the base of the skull from the condyles to the vomerine cartilage. The mastotym- panic, fig. 43, 25, fig. 44, s, 25, fig. 68, 8, extends from the mastoid cartilage, where it is broadest, to the outside of the ANATOMY OF VERTEBRATES. 87 69 69A Hyo-branckial frame, skull, Tadpole, cxxxix. hypotympanic, fig. 43, 29. The parietals, ib., 44 and 68, 7, and afterwards the frontals, ib. ib., 11, progressively cover the 'fon- tanelle ' above, as the basioccipito-sphenoid covers the hypophysial vacuity below. An antorbital plate, fig. 72, b, extends from the frontal to the maxillary. The premaxillaries, at first beak-shaped, figs. 42, 22, and 6 9 A, n, expand transversely as the month widens to form its fore-part, fig. 71, n : external to the premaxillary pedicles begins the ossification of the turbinals. The tf pterygoid plate,' fig. 43, 24, extends to the inner side of the hypotympanic, 29, and forward to the ( palatine ' bone, and the bifid dentigerous 6 vomerine ' plate, fig. 73, Z, /. From the membrane covering ( MeckePs cartilage,' figs. 69A and 70, d, are exclusively developed the mandibular elements, the ( angular,' fig. 43, so, and f dentary,' ib. 32, being the chief; there is also a ' splenial,' which in some perennibranchiate Batrachia supports teeth. As the mandible, fig. 71, d, lengthens, the tympanic, ib. e, shortens and becomes more vertical, and the hyoid arch, ib. a, shifts its attachment to the petrosal, close behind, but distinct from, the tympanic. In the Lepidosiren the ali- and orbito-sphenoids and the hypo- tympanic remain cartilaginous ; premaxillaries are represented by their ascending or facial parts coalesced into a single plate, supporting the tw T o pre- hensile teeth. The postorbito- supertemporals, fig. 41, 12, are 6 dermal ' or scleral bones, over- lapping the fronto-parietals. They are not present in modern Batrachia. In the Axolotl (Axolotes marmoratus), the basioccipital is repre- sented by the posterior part of the common broad and flat basi- cranial bone. The exoccipitals are separated below by this process, and above by a cartilaginous representative of the superoccipital. Each exoccipital developes a small, almost flattened coiidyle, anterior to which it is perforated by the eighth pair of nerves ; it articulates above with the parietal and mastotympanic, and is separated from the alisphenoid by the large cartilaginous petrosal, to which a small discoid representative of the stapes is attached, 71 Hyo-braucliial fi-ame, skull, older Tadpole, cxxxi 88 ANATOMY OF VERTEBRATES. closing the homolosfue of the * fenestra ovalis.' The basi- o o sphenoidal portion of the basicranial plate sends out an angular process on each side, which supports the alisphenoid. The surfaces of the alisphenoid are directed forward and backward, instead of from side to side, and it constitutes chiefly the anterior parietes of the otocrane ; the inner and anterior border is notched by the great trigeminal nerve. The parietals are long and broad, divided by the sagittal suture, and impressed at the posterior and outer angle by the anterior attachment of the great dorsal trunk-muscles. The masto-tympanic is articulated to this part of the parietal and to the exoccipital ; it includes all the divisions of the pedicle save the lowest, ( hypotympanic,' which affords the articulation to the mandible. The orbitosphenoids are divided by an unossified tract of some extent from the ali- sphenoids, and articulate above with the extremity of the parietal, the frontal and prefrontal bones. There are neither paroccipitals nor postfrontals. The vomerine portion of the basicranial plate is chiefly cartilaginous. The turbinals are very small, and separated from each other by the junction of the premaxillaries with the frontals. The bone extending from the frontal to the maxillary in front of the orbit may be termed ( antorbital ; ' the ossification which extends therefrom, in higher Batrachians, takes the situation of the facial plate of the prefrontal, of the nasal, and of the lacrymal. The pedicles ( tf apophyse montante,' Cuvier,) of the premaxillaries are long and narrow. The small maxillary is attached to the antorbital, to the palatine, and to the premaxil- lary ; the end of the bone extends freely backward as in the Menopome, fig. 43, 21. The alveolar border of both premaxillaries and maxillaries supports a single row of small equal and sharp- pointed denticles. Two bones attached to the anterior and outer part of the basicranial bone, and which may be regarded either as ' vomerine or palatal, support each a narrow rasp-like group of minute denticles, which are continued backAvard upon the be- ginning of the pterygoids ; the pterygoids continued from these bones and from the sides of the basicranial bone expand as they extend backward and apply themselves to the inner side of the tympanic pedicle. The nasal meatus has its posterior termination between the beginning of the pterygoid and the end of the maxillary bones. Besides the ordinary row of denticles upon the clentary piece of the lower jaw, there is a second shorter series upon the splenial piece. In theMeiiobranch(^/eft0Z/Ymc//s later alls) the occipital condyles are transversely oblong, convex vertically, concave transversely, ANATOMY OF VERTEBRATES. 89 developed from the exoccipitals, which are separated above and below, as in the Axolotl : each exoccipital forms the posterior half of the otocrane, is perforated by the nervus vagus, and articulates above with the parietal and masto-tynipanic. The basisphenoid is very broad and flat : the alisphenoids bound the fore part of the otocrane, transmit the trigeminal nerve, and abut against the tym- panic pedicle in its course backward to the rnastoid. The parietals are divided by the sagittal suture and develope a small ridge there posteriorly : each parietal sends down a process in front of the ali- sphenoid which rests upon the pterygoid, representing the so-called 'columella' in Lizards. There are no maxillary bones. The alveolar border of the premaxillaries, which support a single row of long and slender teeth, ten in number in each bone, terminates in a point projecting freely outward and backward. The vomero-pala- tine bones unite together anteriorly, but diverge posteriorly, where they give attachment by their outer margin to the pterygoids. The two foregoing are examples of the Ichthyomorphs which retain the gills, and thence are termed ( perennibranchiate.' The Menopome, figs. 43, 72, and 73, represents a later phase of larval 72 Upper view of skull of the Menopome. cxxxis. Under view of the skull. life, the gills being absorbed and only the branchial slits re- maining. In fig. 72, e e are exoccipitals, each developing a condyle ; c, c, parietals ; g, g mastotympanics ; h hypotympanic ; a, a, frontals, b, b, antorbitals ; d, d, nasals ; n, orbitosphenoid ; k, k, premaxillaries ; i, i, maxillaries ; f, f, pterygoids. In fig. 73, m is the basioccipito-sphenoidal ; e, e, exoccipitals ; g, g, mastotympanics ; A, h, hypotympanics ; /, /, pterygoids ; /, /, vomers ; k, k, premaxillaries. In the Frog (Rana) when the metamorphosis is complete, the 90 ANATOMY OF VERTEBRATES. exoccipitals have coalesced with the superoccipital above, and with the basioccipito-sphenoidal plate below ; this latter, fig. 98 A, sends out on each side a process to form the floor of the otocrane, and its forward extension is long and narrow : the tympanic developes a frame for the large ear-drum, fig. 44, N : the stapes, now colu- melliform, stretches from that membrane to the foramen of the labyrinth. ( Meckel's cartilage,' figs. 69 and 71, d, contributes nothing to the bony conductor of sonorous vibrations which becomes subdivided into a chain of ossicles in Mammalia. The hypotympanic, fig. 44, 28, sends forward a process to the end of the maxillary, thus articulating, as in the Plagiostomes, with both upper and lower jaws. The essential or neurapophysial parts of the prefrontals encompass the prosencephalon, and coalesce to form a ring of bone, like the exoccipitals : it is the ' os en ceinture ' of Cuvier, 1 part of which appears at the upper surface of the cranium, fig. 44, u, between the frontals and antorbitals, ib. is, which here, and still more in the Toad, assume the character of nasals connate with lacrymals. Between these and the premaxillaries are the small bony parts of the olfactory sacs, usually described as ' nasal bones.' The orbital and temporal fossae form one wide common vacuity on each side the cranium : it is divided from the nostril by the junction of the maxillary, ib. 21, with the naso-lacrymal bone : the premaxillaries, ib. 22, are small bones, with a well-marked facial and buccal portion. The palatines, fig. 98, A, are transversely extended : the divided vomer is dentigerous : the pterygoid, ib. 24, sends out three rays for the sphenoidal, tympanic, and palato-maxillary connections re- spectively. The mandible is edentulous. The hyoid arch with its branchial appendages has changed its connections as well as shape. In the tadpole, with the fully-developed gills, the carti- lage representing the stylo- and cerato-hyals, figs. 69 and 6 9 A, a, is short and thick, and attached to the back of the tympanic pedicle, ib. e, to the end of which is articulated the mandible, ib. d. The ceratohyals are connected below to a median piece, ib. Z>, which may represent both the basihyal and basibranchial : it directly supports the hypobranchials c, c, to which the ceratobranchials, or branchial arches are attached. As the gills wither, the stylo-ceratohyals, figs. 70 and 71, a, lengthen, attenuate, and acquire an independent attachment to the petrosal ; the basi- and hypo-branchial s, fig. 74, c, c, coalesce into a single cartilaginous plate, with the 6 basihyal,' ib. b ; and the ceratobranchials are reduced to a single pair, which represent the so-called e posterior cornua ' of the hyoid. 1 cxxxix. torn V. pt. 2, p. 389, pis. xxiv. xxvii., well illustrate the osteology of the Batrachia. ANATOMY OF VERTEBRATES. 91 74 Hyobrancliial frame, Rana paradoxa. cxxxix. The scapular arch, fig. 42, so, 51, retrogrades, like the hyoid, from its primitive position in the larva. Cuvier, at thexonclusion of his description of the batrachian skull, remarks, ' This skull does not accord with the theory of the three, four, or seven vertebrae, or even of one (cranial) vertebra, any more than it does with that of the identity in the number of bones ' (in different animals). l At the same time he de- termines the special homo- logy of the twenty-six bones, exclusive of the mandible and hyoid apparatus, and assigns to them the same names, --and as regards the majority, correctly, which those bones bear in the rest of the vertebrate province. We have been led, therefore, to look for some higher law within which that of the special conformity may be included. In many instances of trunk-vertebra?, the neurapophyses meet below, as well as above the neural axis, their bases being extended towards each other so as to interpose between that axis and the vertebral centrum. This condition is repeated by the exoccipitals which form the neural arch of the epencephalon, and encompass it, in Batrachia, giving passage to its chief pair of nerves and de- veloping articular processes for the succeeding vertebra. The two pairs of neurapophyses in advance, retain the more ordinary rela- tions of these elements, the more expanded mes- and pros-encephala having their bony ring or arch completed by a centrum below and a spine above. One neurapophysis (alisphenoid) transmits the trigeminal nerve, the other (orbitosphenoid) the optic nerve : the fourth or anterior neural arch ( f os en ceinture ' and ( ethmoide ' of Cuvier) encompasses the foremost segment of the brain as the exoccipitals do the hindmost ; and they give passage to the olfactory nerves. Ossification of this ring of bone begins in its lateral halves : the essential relations and functions being those which characterise the bones which in bony fishes will be described as ( prefrontals.' Beneath, and supporting them, is a pair of bones which may be regarded as a mesially divided ' centrum ' (vomer) : and above is a pair of bones which may be 1 cxxxix. ' Ce crane ne s'accorde pas plus avec la theorie des trois, des quatre, ou des sept vertebres, meme avec celle d'une vertebre, qu'avec celle de 1'identite de nombre des os,' vol. v. pt. ii. p. 391. 92 ANATOMY OF VERTEBRATES. regarded as a mesially divided neural spine (nasal). Thus may be discerned four cranial segments having the essential characters and relations of the neural arch of the type vertebra. The upper, 22, and lower, so, jaws, the hyoid, 40, and scapulocoracoid, 50-52, fig. 42, constitute four inverted arches ; but their vertebral relations will be better understood in the composition of the skull in bony fishes. 30. Skull of Osseous Fishes. The head is larger in proportion to the trunk in fishes than in other vertebrate classes ; it is usually in form of a cone, figs. 34, 38, whose base is vertical, directed back- ward, and joined at once to the trunk, and whose sides are three in number, one superior, and two lateral and inferior. The cone is shorter or longer, more or less compressed or squeezed from side to side, more or less depressed or flattened from above downward, with a sharper or blunter apex, in different species. The base of the skull is perforated by the hole, called ( foramen magnum,' for the exit of the spinal marrow ; the apex is more or less widely and deeply cleft transversely by the aperture of the mouth ; the eye- sockets or 'orbits,' ib. 17, are lateral, large, and usually with a free and wide intercommunication in the skeleton ; the two vertical fissures behind are called ( gill-slits,' or branchial or oper- cular apertures ; and there is a mechanism like a door, ib., 35, 36, 37, for opening and closing them. The mouth receives not only the food, but also the streams of water for respiration, which escape by the gill-slits. The head contains not only the brain and organs of sense, but likewise the heart and breathing organs. The inferior or ' hasmal ' arches are greatly developed accordingly, and their diverging appendages support membranes that can act upon the surrounding fluid, and are more or less employed in locomotion : one pair of these appendages, ib. P, 55, 56, answers, in fact, to the fore-limbs in higher animals ; and their sustaining arch, ib. 51, 52, in many fishes, also supports the homologues of the hind-limbs, v, :o. Thus brain and sense-organs, jaws and tongue, heart and gills, arms and legs, may all belong to the head ; and the disproportionate size of the skull, and its firm attachment to the trunk, required by these functions, are precisely the conditions most favourable for facilitating the course of the fish through its native element. o It may well be conceived, then, that more bones enter into the formation of the skull in fishes than in any other animals ; and the composition of this skull has been rightly deemed the most difficult problem in Comparative Anatomy. c It is truly remark- able,' writes the gifted Oken, to whom we owe the first clue to its solution, ( what it costs to solve any one problem in Philosophical ANATOMY OF VERTEBRATES. 93 Anatomy. Without knowing the what, the how, and the u'hy, one may stand, not for hours or days, but weeks, before a fish's skull, and our contemplation will be little more than a vacant stare at its complex stalactitic form.' To show what the bones are that enter into the composition of the skull of the fish ; how, or according to what law, they are there arranged ; and why, or to what end, they are modified, so as to deviate from that law or archetype, will next be our aim. These points, rightly understood, yield the key to the composition of the skull in all vertebrata, and they cannot be omitted without detri- ment to the main end of the most elementary essay on the skeletons of animals. The comprehension of the description will be facilitated by reference to figs. 75 85 ; and still more if the reader have at hand the skull of any large fish. In the Cod(Gadus morrhua, L. fig. 75), e. g., it may be observed, in the first place, that most of the bones are, more or less, like 15 Skull of Cod (Morrliua vulgaris), Cuv. large scales; have what, in anatomy, is called the e squamous' cha- racter and mode of union, being flattened, thinned off at the edge, and overlapping one another ; and one sees that, though the skull, as a whole, has less freedom of movement on the trunk, more of the component bones enjoy independent movements. Before we proceed to pull apart the bones, it may be well to remark, that the principal cavities, formed by their coadaptation, are the ( cranium, 94 ANATOMY OF VERTEBRATES. 76 lodging the brain and the organs of hearing; the c orbital,' and f nasal' chambers ; the ( buccal ' and f branchial ' canals. Some of these cavities are not well defined. The exterior of the skull is traversed by five longitudinal crests, intercepting four channels which lodge the beginnings of the great muscles of the upper half of the trunk. The median crest is developed from the superoccipital, figs. 75, 76, 3, and sometimes also from the frontal, fig. 75, n : the lateral crest is formed by the parietal, fig. 76, 7, and paroccipital, ib. 4: the external crest by the postfrontal, ib. 12, and mastoid, ib. 8. The lower border of the orbit, fig. 75, g, g, projects freely downward. The hind border of the operculum is produced into spines in some species, fig. 82. In the analysis of the fish's skull it is best to begin at the back part ; for the segments of the skeleton de- viate most from the archetype as they recede in position toward the two ex- tremes of the body. After a little practice one succeeds in detaching the bones which form the back part or base of the conical skull, and which immediately precede and join those of the trunk ; we thus obtain a ' segment ' or 6 vertebra ' of the skull. If we next proceed to separate a little the bones composing this segment, we find those that were most closely in- terlocked to be in number and ar- rangement as follows : Two single and symmetrical bones, and two pairs of unsymmetrical bones, forming a circle ; or, if the lower symmetrical bone, which is the largest, be regarded as the base, the other five form an arch supported by it, of which the upper symmetrical bone is the key-stone, fig. 77. This answers to the f neural ' arch of the typical vertebra : the base- bone is the ( centrum,' i ; the pair of bones, which articulated with its upper surface and protected the hind division of the brain, form the ( neurapophyses,' 2 ; the smaller pair of bones, projecting outward, like transverse processes, are the ( diapophyses,' 4 ; the symmetrical bone completing the arch, and terminating above in a long crest or spine, is the f neural spine,' 3. It will be observed that the centrum is concave at that surface which articulates with Upper surface of cranium, Perch (Perca fluviatilis) ANATOMY OF VERTEBRATES. 95 Disarticulated epencephalic arch, viewed from behind : Cod (Morrhua vulgar is) the centrum of the first vertebra of the trunk : the opposite surface is also concave, but expanded and very irregular, in order to effect a much firmer union with the centrum of the next cranial segment in advance great strength and O O O fixity being required in this part of the skeleton, instead of the mobility and elas- ticity which is needed in the vertebral column of the trunk. It may be also observed that the ( neurapophyses ' are per- forated, like most of those in the trunk, for the passage of nerves ; that the diapo- physes give attachment to the bones which form the great inferior or hremal arch ; and that the neural spine retains much of the shape of the parts so called in the trunk. Nevertheless, the elements of the neural arch of this hindmost segment of the skull have undergone so much developement and modification of shape, that they have received special names, and have been enumerated as so many distinct and particular bones. The centrum, i, is called ' basioccipital ; ' the neurapophyses, 2, e exoccipitals ; ' the neural spine, 3, ' superoccipital ; ' the diapophyses, 4, ' parocci- pitals.' In the human skeleton all those parts are blended together into a mass, which is called the ' occipital bone.' In Philosophical Anatomy it is the ' epencephalic arch,' because it surrounds the hindmost segment of the brain called ( epencephalon.' The entire segment, here disarticulated, is called the ( occipital vertebra,' and in it we have next to notice the widely-expanded inferior or hnamal arch, fig. 81, 50, H. This consists of three pairs of bones. The first pair are bifurcate, and have two points of attachment to the neural arch, the lower prong, answering to what is called the ' head of the rib,' abutting upon the neura- pophysis ; the upper prong, answering to the ( tubercle of the rib,' articulating to the diapophysis. The second pair of bones are long and slender, and represent the body of the rib. The first and second piece together answer to the element called ( pleurapophysis ; ' the third pair of bones are the ' haemapophyses ; ' these support diverging appendages consisting of many bones and rays. The special names of the above elements of the haemal arch of the occipital vertebra are, from above downwards, e suprascapula,' so ; ( scapula,' 51 ; ( coracoid,' 52. The inverted arch, so formed, encompasses, supports, and protects the heart or centre of the haemal system ; it is called the ( scapular arch.' There 96 ANATOMY OF VERTEBRATES. are cold-blooded animals the gymnothorax and slow- worm, e. g.- -in which this arch supports no appendage ; there are others -Lepidosiren and Protopterus, fig. 41, 52 in which it supports an appendage in the form of a single many-jointed ray, ib. 57. In other fishes, the number of rays progressively increase, until, in those called ( rays ' par excellence, fig. 64, they exceed a hundred in number, and are of great length, forming the chief and most conspicuous parts of the fish. The more common condition of the appendage in question is that exhibited in the Cod, fig. 34, So developed, it is called in Ichthyology the f pectoral fin,' ib. P : otherwise and variously modified in higher animals, the same part becomes a fore-leg, a wing, an arm and hand. Proceeding to the next segment, in advance, in the Cod-fish's o ~ * * skull, we find that the bone which articulated with the centrum of the occipital segment is continued forward beneath a great pro- portion of the skull. In quadrupeds, however, the corresponding part of the base of the skull is occupied by two bones ; and if the single long bone in the fish be sawn across at the part where the natural suture exists in the beast, we have then little difficulty in disarticulating and bringing away with it a series of bones similar in number and arrangement to those of the occipital segment. In the skeletons of most animals the centrums of two or more segments become, in certain parts of the body, confluent, or they may be connate ; they form, in fact, one bone, like that, e. g., which human anatomists call ( sacrum.' By the term ( confluent ' is meant the cohesion or blending together of two bones which were originally separate ; by ( connate,' that the ossification of the common fibrous or cartilaginous bases of two bones proceeds from one point or centre, and so converts such bases into one bone : this is the case, e. g., in the radius and ulna of the frog, and in its tibia and fibula. In both instances they are to the eye a single bone ; but the mind, transcending the senses, recognises such single bone as being essentially two. In like manner it recognises the ( occipital bone ' of man as essentially four bones ; but these have become ' confluent,' and were not ' connate.' The centrums of the two middle segments of the fish's skull are con- nate, and the little violence above recommended is requisite to detach the penultimate segment of the skull. When detached, the bones of it are seen to be so arranged as to form a neural and a hremal arch. In the neural arch, fig. 78, the centrum, neura- pophyses, diapophyses, and neural spine are distinct: moreover, the neural spine in the Cod, and many other fishes, is bifid, or split at the median line. The centrum is called ' basiphenoid,' 5 ; ANATOMY OF VERTEBRATES. 97 the neurapophysis, f alisphenoid,' 6 ; the neural spine, f parietal,' 7 ; and the diapophysis, ' mastoid,' 8. The alisphenoids protect the sides of the optic lobes, and the rest of the penultimate segment of the brain called ( mesencephalon ; ' the mastoids project outward and backward as strong transverse pro- cesses, and give attachment to the piers of the great inverted haemal arch. Before noticing its struc- ture, I may remark that, in the recent Cod-fish, the case, partly gristly, partly bony, which Contains Disarticulated mesencephalic arch, viewed f, i . . -, from behind ; Cod (Morrhua vulgaris) the organ of hearing, is wedged between the last and penultimate neural arches of the skull. The extent to which the ear-case is ossified varies in different fishes, but the bone is always developed in the outer-wall of the case. In the Cod it is unusually large, and is called ' petrosal,' fig. 81, IG; in the Perch, fig. 84, 16, and Carp, fig. 83, 16, it is smaller : it forms no part of the segmented neuroskeleton. In the acoustic organ which it contributes to enclose, there is a body as hard as shell, like half a split almond : it is the ( otolite,' fig. 81, 16. The haemal arch consists of a pleurapophysis and a haemapo- physis on each side, and a haemal spine ; the pleurapophysis is in two parts, the upper one called ( stylohyal,' ib. 38 ; the lower one called f eplhyal,' ib. 39 ; the haemapophysis is called ' ceratohyal,' ib. 40. The haemal spine is subdivided into four stumpy bones, called collectively 'basihyal,' ib. 4i ; and which, in most fishes, support a bone directed forward, entering the substance of the tongue, called f glossohyal,' ib. 42 ; and another bone directed backward, called ' urohyal,' ib. 43. The ceratohyal part of the haemapophysis supports an appendage, or rudimental limb, called ' branchiostegal,' fig. 81, 44, answering to the pectoral fin diverging from the haemal arch, in the adjoining occipital segment. The penultimate segment of the skull above described is called the f parietal vertebra ; ' the neural arch is called ( mesencephalic ; ' and the hamial arch is called ' hyoidean ' in reference to its sup- porting and subserving the movements of the tongue. The next segment, or the second of the skull, counting back- ward, can be detached from the foremost segment without dividing any bone. It is then seen to consist, like the third and fourth VOL. I. H 98 ANATOMY OF VERTEBRATES. Disarticulated prosencephalic arch, Cod (Morrhua vulgaris) segments, of two arches and a common centre ; but the consti- tuent bones have been subject to more extreme modifications. The centrum, called ( presphenoid,' fig. 79, 9, is produced far forward, slightly expanding ; the neurapophyses, called ( orbito- sphenoids,' ib. 10, are small semi- oval plates, protecting the sides of the cerebrum ; the neural spine, or key-bone of the arch, called f frontal,' ib. 11, is enormously expanded, but in the Cod is single ; the diapophyses, called ( post-frontals,' ib. 12, project outward from the hinder angles of the frontal, and give attachment to the piers of the inverted haemal arch. The first bone of this arch is com- mon in Fishes to it and to that of the last described vertebra, being the bone called ( epitympanic,' fig. 81, 25 ; this modification is called for by the necessity of consentaneous move- ments of the two inverted arches, in connection with the deglutition and course of the streams of O water required for the branchial respiration. The haemal arch of the present segment --enormously developed is plainly divided primarily on each side into a pleurapophysis and hgema- pophysis ; for these elements are joined together by a movable articulation, whilst the bones into which they are subdivided are suturally interlocked together. The pleurapophysis is so subdivided into four pieces ; the upper one, articulating with the postfrontal and mastoid the diapophyses of the tAVO middle segments of the skull- -is called ' epitympanic,' ib. 25; the hind- most of the two middle pieces is the ' mesotympanic,' ib. 26 : the foremost of the two middle pieces is the ' pretympanic,' ib. 27 ; the lower piece is the hypotympanic, ib. 28 ; this presents a joint- surface, convex in one way, concave in the other, called a ( gingly- moid condyle,' for the hagmapophysis, or lower division of the arch. In most air-breathing vertebrates- -the Serpent, fig. 97, e.g. --the pleurapophysis resumes its normal simplicity, and is a single bone, 28, which is called the ( tympanic ; ' in the eel-tribe, as in the Batrachia, figs. 43, 72,^, h, it is in two pieces. The greater subdivision, in more actively breathing Fishes, of the tympanic pedicle, gives it additional elasticity, and by their overlapping, interlocking junction, greater resistance against fracture ; and ANATOMY OF VERTEBRATES. 99 these qualities seem to have been required in consequence of the presence of a complex and largely developed diverging appendage, which forms the framework of the principal flap or door, called ( operculum,' figs. 81, 84, 34-37, that opens and closes the branchial fissures on each side. The appendage in question consists of four bones ; the one articulated to the tympanic pedicle is called ( pre- opercular,' ib. 34 ; the other three are, counting downward, the ' opercular,' ib. 35 ; the f subopercular,' ib. 36 ; the interopercular,' ib. 37. The hremapophysis is subdivided into two, three, or more pieces, in different fishes, suturally interlocked together ; the most common division is into two subequal parts, one presenting the concavo-convex joint to the pleurapophysis, and called ( articular,' ib. 29 ; the other, bifurcated behind to receive the point of 29, and joining its fellow at the opposite end, to complete the haemal arch : it supports a number of the hard bodies called ' teeth,' and hence it has been termed the ( dentary,' ib. 32. In the Cod there is a small separate bone, below the joint of the articular, forming an angle there, and called the e angular piece,' fig. 75, 30. In consequence of this extreme modification, in relation to the offices of seizing and acting upon the food, the pair of ha3ma- pophyses of the present segment of the skull have received the name of ( lower jaw,' or ( mandible ' (mandibula). The haemal arch is, hence, called ' mandibular : ' the neural arch ( prosen- cephalic : ' the entire segment is called the ' frontal vertebra.' The first segment, forming the anterior extremity of the neuro- skeleton, like most peripheral parts, is that which has undergone the most extreme modifications. The obvious arrangement, nevertheless, of its constituent bones, when viewed from be- hind, after its detachment from the second segment, affords one of the most conclu- sive proofs of the principle of adherence to common type which governs all the segments of the neuroskeleton, whatever Offices they may be modified tO fulfil. Disarticulated rhinencephalic arch, , -I 1 r i i Cod (Morrhua vulgarly Hie neural arch, ng. 80, is plainly mani- fested, but is now reduced to its essential elements viz., the centrum, the neurapophyses, and the neural spine. The centrum is expanded anteriorly, where it usually supports some teeth on its under surface in fishes; it is called the ' vomer,' ib. 13. The neurapophyses are notched (in the Cod), or perforated (in the Sword-fish), by the crura or prolongations of the brain, which expand into its anterior division, called rhinencephalon, or H 2 100 ANATOMY OF VERTEBRATES. ( olfactory lobes' ; the special name of such neurapophysis is ( pre- frontal,' ib. 14. The neural spine is usually single, sometimes cleft alon the middle ; it is the 'nasal.' ib. 15. o The haemal arch, fig. 81, 20-^2, H, is drawn forward, so that its apex, as well as its piers, are joined to the centrum (vomer), and usually also to the neural spine (nasal), closing up anteriorly the neural canal. The pleurapophyses are simple, short, sending backward an expanded plate ; they are called ' palatines,' ib. and fig. 84, 20. The haemapophyses are simple, and their essential part, intervening between the pleurapophysis and haemal spine, is 81 .TTl Side view of cranial vertebrae and sense-capsules ; the liamial arches, H H, in outline, Cod (Morrhua vulgaris) short and thick ; but they send a long process backward ; this element is called ' maxillary,' ib. 21. The haemal spine, cleft at the middle line, sends one process upward of varying length in different fishes, and a second downward and backward, and its under surface is beset with teeth in most fishes : it is called ( premaxillary,' ib. 22. Each pleurapophysis supports a f diverg- ing appendage,' consisting commonly of two bones : the outer one, which fixes the present ha3inal arch to the succeeding one, is called f pterygoid,' figs. 75, 81, 24; the inner one is the ' ento- pterygoid,' ib. 23. The entire segment is called the f nasal vertebra ;' its neural arch is the e rhinencephalic ; ' its haemal arch, forming what is termed the upper jaw (maxilla), is called the f maxillary ' arch and appendages. On reviewing the arrangement of the bones of the foreo'oinoj O o O O segments, one cannot but be struck by the strength of the arches which protect and encompass the brain, and by the efficiency of that ANATOMY OF VERTEBRATES. 101 arrangement which provides such an arch for each primary divi- sion of the brain ; and a sentiment of admiration naturally arises on examinino; the firm interlocking of the extended sutural sur- o o faces, and especially of those uniting the proper elements of the arch with the buttresses wedged in between the piers and key- stone, and to which buttresses (diapophyses) the larger hremal arches are suspended. In addition to the parts of the neuroskeleton, the bones of the head include the ossified part of the ear-capsule, ( petrosal,' fig. 81, 16, already mentioned; an ossified part of the eye-capsule, commonly in two pieces, ' sclerotals,' ib. 17 ; and an ossified part of the capsule of the organ of smell, ' turbinal,' ib. 19. Another assemblage of splaiichnoskeletal bones support the gills, and are in the form of slender bony hoops, called ' branchial arches,' fig. 85, 48, 49. They are partly supported by the hyoidean arch. Amongst the bones of the muco-dermal system, may be noticed those that circumscribe the lower part of the orbit, fig. 75, g, g ; of which the anterior, ib. 73, is pretty constant in the vertebrate series, and is called ' lacrymal.' In fishes they are called ' subor- bitals,' and are occasionally present in great numbers, as, e. g., in the Tunny, or developed to enormous size as in the Gurnard, fio\ 82, and allied fishes, thence called f mail-cheeked.' A similar o * 82 Fore part of the skeleton of the Gurnard (Trigla Lyra) series of bones called ' supertemporals ' sometimes overarches the temporal fossa. At the outset of the study of Osteology it is essential to know well the numerous bones in the head of a fish, and to fix in the memory their arrangement and names. The latter, as we have 102 ANATOMY OF VERTEBRATES. seen, are of two kinds, as regards the bones of the neuroskeleton : the one kind is ( general,' indicative of the relation of the skull- bones to the typical segment, and which names they bear in common with the same elements in the segments of the trunk ; the other kind is ( special,' and bestowed on account of the par- ticular developement and shape of such elements, as they are modified in the head for particular functions. A great proportion of the bones in the head of a fish exist in a very similar state of connection and arrangement in the heads of other vertebrates, up to and including man himself. No method could be less con- o ducive to a true and philosophical comprehension of the vertebrate skeleton than the beginning its study in man the most modified of all vertebrate forms, and that which recedes furthest from the common pattern. Through an inevitable ignorance of that pattern, the bones in Anthropotomy are indicated only by special names more or less relating to the particular forms these bones happen to bear in man ; such names, when applied to the tallying bones in lower animals, losing that significance, and becoming arbitrary signs. Owing to the frequent modification by confluence of the human bones, collections of them, so united, have received a single name, as, e. g. ( occipital,' ( temporal,' &c. ; whilst their constituents, which are usually distinct vertebral elements, have received no names, or are defined as processes, e. g. ' condyloid process of the occipital bone,' ' styloid process of the temporal bone,' i petrous portion of the temporal bone,' &c. The classifi- cation, moreover, of the bones of the head in Human Anatomy, viz. into those of the cranium and those of the face, is artificial or special, and consequently defective. Many bones which essentially belong to the skull are wholly omitted in such classification. In regard to the archetype skeleton, fishes, which were the first forms of vertebrate life introduced into this planet, deviate the least therefrom ; and according to the foregoing analysis of the bones of the head, it follows that such bones are primarily divisible into those of- The Neuroskeleton ; The Splanchnoskeleton ; The Dermoskeletoii. The neuroskeletal bones are arranged in four segments, called The Occipital vertebra ; The Parietal vertebra ; The Frontal vertebra ; The Nasal vertebra. ANATOMY OF VERTEBRATES. 103 Each segment consists of a 'neural' and a 'hsemal' arch. (Fig. 81, N, H.) The neural arches are- N I. Epencephalic arch (bones Nos. i, 2, 3, 4) ; N II. Mesencephalic arch (5, 6, 7, s) ; N in. Prosencephalic arch (9, 10, 11, 12); N iv. Rhinencephalic arch (13, u, is). The haemal arches are H i. Scapular arch (50-52) ; H II. Hyoidean arch (33-43) ; H in. Mandibular arch (28-32) ; H iv. Maxillary arch (20-22). The diverging appendages of the haemal arches are 1. The Pectoral (54-57) ; 2. The Branchiostegal (44) ; 3. The Opercular (34-37); 4. The Pterygoid (23-24). The bones or parts of the splanchnoskeleton which are inter- calated with or attached to the arches of the true vertebral segments, are- The Petrosal (IG) or ear-capsule, with the otolite, is"; The Sclerotal(i7) or eye-capsule; The Turbinal (19) or nose-capsule ; The Branchial arches (45-49) ; The Teeth. The bones of the dermoskeleton are- The Supratemporals (74) ; The Postorbitals (72) ; The Superorbitals (71); The Suborbitals (73); The Labials (75), and others which will be pointed out in certain ganoid fishes. Such appears to be the natural classification of the parts which constitute the complex skull of Osseous Fishes. 104 ANATOMY OF VERTEBRATES. The term f cranium ' might well be applied to the four neural arches collectively, figs. 76, 83 ; but would exclude some bones called ' cranial,' and include some called ( facial,' in Human Anatomy. In a side view of the naturally connected bones of those arches, such as is shown in the Carp, fig. 83, the upper part of the cranium is formed by the neural spines called super- occipital 3, parietal 7, frontal 11, and nasal is; the lower part by the centrums called basioccipital i, basisphenoid 5, presphe- noid 9, and vomer 1 3 : the side-walls by the neurapophyses called exoccipital 2, alisphenoid 6, orbitosphenoid 10, and prefrontal 14. Between 2 and 6 is intercalated the petrosal 16 : between the fore part of 9 and 10 is the ( inter orbital is,' which is an inconstant ossification in fishes. The outstanding or transverse processes are the paroccipital 4, the mastoid 8, and the postfrontal 12. 83 Cranium of a Carp. In the Carp the parietals meet and unite upon the vertex by a ' sagittal ' suture : in most osseous fishes, as in the Cod and Perch, figs. 76, 77, they are separated by the junction of the superocci- pital, 3, with the very large frontals, 11,11. At the base of the skull may be seen, in the Perch, fig. 84, the basioccipital i, the articular processes of the exoccipitals 2, and the spine-shaped end of the superoccipital 3. The paroccipital 4, is separated below from the exoccipital by the petrosal 16. The basi-presphenoid, 5 and 9, carries forward the bodies of the vertebra? to the vomer 13*, which is expanded and dentigerous anteriorly, as the bodies of the cervical vertebras support teeth in the Deirodon (p. 57). The alisphenoids 6, the orbitosphenoids 10, and the prefrontals 14, are attached to the sides of the basal elements ; more externally are seen the frontal 11, postfrontal 12, mastoid 8, and paroccipital 4. On the left side are shown the palatine 20, the entopterygoid ANATOMY OF VERTEBRATES. 105 23, and external to it the pterygoid abutting upon the hypotym- panic, 28 d : between this and the epitympanic, 28, are the mesotympanic, 38, and the pretympanic b. The preopercular, 34, runs parallel with, strengthens, and connects together the divisions 12 11 Base of the skull with left side of mandibular arch and its orercular appendage, Perch (Perc a fluviatilis) of the tympanic pedicle : it supports the opercular, 35, the sub- opercular, 36, and the interopercular, 37. In the mandibular ramus the articular is marked 29, and the dentary 32. The free end of the maxillary is seen at 21. In fig. 85 the maxillary and mandibular arches and appendages are removed, the stylohyal, 38, having been detached from the epitympanic. It resumes its normal attachment to its segment when the special branchial apparatus becomes abrogated, as in the advanced batrachian, fig. 71, in which we saw the change of position, as contrasted with the earlier piscine condition of the larva, fig. 69 A. In the complex and ossified hyoidean arch of 106 ANATOMY OF VERTEBRATES. fishes we find, after the stylohyal 33, the epihyal 39, the cerato- hyal 40, and basihyal 41 ; to which may be articulated a glosso- hyal 42, and a urohyal 43 : this is a large compressed lamelli- form bone in the Perch. Seven branchiostegal rays, 44, are articulated to the epi- and cerato-hyals. Four branchial arches are attached to the base of the cranium. The first consists of the ceratobranchial, 47, and epibranchial, 48, elements : both of which support a series of processes, 63, directed towards the cavity of the mouth and defending the entry to the branchial fissures. The second and third arches are connected above by the pharyngo- 85 50 Hyobranchial and scapular arches, Perch (Perca fluviatilis) branchial elements, 49, to the cranium ; and these elements usually support teeth. The gills are attached to grooves on the outer side of the epi- and cerato-branchials ; the arches being closed below by the c basibranchials ' which are attached to the hyoid. The suprascapula, so, is attached by its lower branch to the basi- occipital, and by its upper one to the paroccipital, 4. The scapula, 51, supports the coracoid, 52, to which the clavicle, 58, is attached, the relative position of which to the coracoid becomes changed as the scapular arch is detached from its natural con- ANATOMY OF VERTEBRATES. 107 nection and displaced backward. The humeral segment of the fore limb is rarely developed in fishes ; the radius, 54, and ulna, 55, are directly articulated with the coracoid, and are commonly much more broad than lonsr. o Some of the special characters and modifications of the bones of the head will next be briefly noticed. The articular cup for the atlas varies from the deep conical excavation seen, fig. 77, i, in the Cod, to the almost flat surface in the Halibut ; it is rare to find, as in the Pipe-fish (Fistularia), the basioccipital presenting a convex surface for articulation with the body of the atlas ; or to find this centrum confluent with the basioccipital, as in Polypterus. In many fishes the under part of the basioccipital is expanded and excavated; in the Carp, the under part is produced into a broad triangular plate, fig. 83, i, which supports the large upper pharyngeal grinding tooth ; in the ganoid Lepidosteus, the basioccipital developes two plates from its upper and outer angles, which complete the foramen magnum and support the exoccipitals above. The exoccipitals, fig. 77, 2, are perforated for the passage of the nervi vagi, some- times for the first spinal or hypoglossal nerve ; the foramina being unusually large in the Carp tribe, fig. 83, 2, where they relate also to the connection of the air-bladder with the organ of hearing, by means of the ossicles, , b, c, d, and e. In some fishes, e.g. Perca, fig. 84, 2, the exoccipitals send backward articular processes modified to allow a slight move- ment upon the anterior articular processes of the atlas. Like the neurapophyses of the trunk in some fishes (e.g. Lepidosiren, Thynnus, Xipliias), the bases of the exoccipitals expand, and meet upon the upper surface of the basioccipital, and immediately support the medulla oblongata. The superoccipital, fig. 77, 3, usually sends upward and back- ward a strong compressed spine from the whole extent of the middle line, and a transverse ( superoccipital ' ridge outwards from each side of the base of the spine, to the external angles of the bone. In most fishes this bone advances forward and joins the frontal, pushing aside, as it were, the parietals, as in fig. 76, 3 ; in Balistes the produced part of the superoccipital is even wedged into the hinder half of the frontal suture. In the Carp, on the contrary, the anterior angle of the superoccipital is trun- cated, forming the base of the triangle, and is articulated by a lamboidal suture to the parietal bones, fig. 83, 7, which here meet at the mid-line of the skull, and the upper part of the occipital spine is low and flattened. The superoccipital is also separated 108 ANATOMY OF VERTEBRATES. from the frontal by the parietals, in the Salmonoid, Clupeoid, Mursenoid, and most ganoid fishes ; and is itself divided, in Amia and Lepidosteus, by a median suture ; these modifications tell strongly against extending the homology of the superoccipital with the supernumerary ' interparietal' bone of Mammals, beyond the anteriorly produced interparietal portion ; which, however, is not developed from a separate centre in Fishes. When the skull is much compressed the occipital spine is usually very lofty, as in the Opah-fish sm&Argyreiosus, fig. 38 : in the Light-horseman fish (Epliippus) it expands above its origin into a thick crest of bone, giving the skull the appearance of a helmet ; but in low flattened skulls the spine is much reduced, projecting merely backward, as in the Pike and Salmon, and being some- times obsolete, as in the Remora. In a few instances, the broad posterior part of the superoccipital articulates with the neural arch and spine of the atlas, and sometimes, on the other hand, e.g. in the Halibut, the entire bone is pushed by the paroccipitals upon the upper surface of the skull, where it manifests the loss of symmetry by the absence of the expanded plate on the left side of the spine. In broad and depressed skulls the par occipital, 1 fig. 76, 4, forms a strong crest, and exceeds the exoccipital in size ; in narrow and deep skulls the proportions of these bones are commonly reversed, and the paroccipitals sometimes disappear. In the Shad, the paroccipitals unite with the mastoids almost as in the Chelonia ; and in Polyprion they are connate with the exoccipitals as in batrachian and crocodilian Reptiles. In Synodus, Callichthys, and Hcterobranclius., the paroccipital is visible only at the back part, not at the upper part, of the skull. The inner surface of the paroccipital, like that of the exoccipital, is excavated for the lodgment of part of the posterior and external semicircular canal of the enormous internal oro-an of hearing in Fishes. The outer o ~ projecting process supports the upper fork of the first piece of the scapular arch ; sometimes, as in Ephippus, by a distinct arti- cular cavity. The neural parts of the occipital vertebra are those which are commonly in Fishes the most completely ossified at the expense of their primitive cartilaginous bases ; and, in Polypterus, they become anchylosed into one piece, like the occipital bone of Anthropotoiriy, the superoccipital being as little developed as in Protopterus. 1 The paroccipitals are not to be confounded with the dermal bone called ' epiotic ' by Professor Huxley, in his reproduction of Miiller's figure of the head of Polt/pterus, in the Government Publication, (CLXVIII.) p. 22, fig. 16. ANATOMY OF VERTEBRATES. 109 The basisphenoid (figs. 78 and 84, 5) is usually bifurcate poste- riorly, and more or less expanded beneath the cranial cavity ; it is then continued forward (sometimes after sending out a pair of lateral processes, as in the Perch, more commonly without such processes) along the base of the interorbital space to near the fore part of the roof of the mouth : its posterior extremity is joined by a squamose suture, as in Diodon, to the basioccipital ; or, more commonly, as in the Cod, is firmly wedged by a kind of double gomphosis into the basioccipital ; its expanded part sup- ports the petrosals and alisphenoids : the presphenoidal prolono-a- tion (figs. 83 and 84, 9) articulates with the orbitosphenoids and the ethmoid, is, when this is ossified; and it terminates forward by a cavity receiving the pointed end of the vomer, fig. 84, is. It is this portion of the basi-pre-sphenoid which manifests the loss of symmetry in the flat fishes (Pleuronectidce*), being twisted up to one side of the skull. The basi-pre-sphenoid varies in form with that of the head in general, being longest and narrowest in long and narrow skulls, and the converse. The whole of its upper surface is commonly rough for articulation with the petrosals and alisphenoids ; rarely does any portion enter into the direct formation of the cranial cavity, and then, e. g. in the Cod, a small surface may support the pituitary sac. When it enters more largely into the formation of the floor of the cranial cavity, it usually sends upward a little process on each side ; or, as in Fis- tularia, a transverse ridge. The basisphenoid is smooth below, where it is usually flattened or convex, but sometimes is pro- duced downward in the form of a median ridge, and sometimes is perforated for the lodgment of certain muscles of the eyeball. In the Polypterus both ali- and orbito-sphenoids are aiichylosed to the basi-pre-sphenoid, and the result is a bone that answers to the major part of the ' os sphenoides' of Anthropotomy. As two large and important hremal arches of the head are suspended from the parapophyses of the second and third cranial vertebrae, this seems to be the condition of the fixation and coalescence of the bodies of those vertebras in all Fishes. In some, e. g. Perch and Carp, the base of each alisphenoid rises above the basisphenoid, and then sends inward a horizontal plate, which, meeting that of the opposite alisphenoid, forms the immediate support of the mesencephalon, and at the same time the roof of a canal, excavated in the basisphenoid, and which traverses the base of the skull, below the cranial cavity, from before backwards, opening behind at the under part of the basi- occipital ; this subcranial canal exists in the Salmonoids, Sparoids, 110 ANATOMY OF VERTEBRATES. Scomberoids, and is very remarkable in most fishes with lofty compressed skulls, as the Epliippus. In them it resembles, but is not homologous with, the posterior prolongation of the nasal pas- sao-es in the Crocodiles, and it lodges some of the muscles of the O ^ eyeball. The form of the alisphenoids is influenced by that of the skull ; when this is low and flat, their antero-posterior exceeds their vertical extent ; in deep and compressed skulls they are narrow and higli plates ; in ordinary shaped skulls they present either a sub- circular form, and are perforated, as in the Carp, fig. 83, 6, or are reniform, the anterior border being deeply notched, as in the Cod, fig. 81, G; they form a more definite and fixed proportion of the lateral parietes of the skull than do the petrosals, ib. IG, which are interposed between them and the exoccipitals ; and they have their essential function in sustaining and protecting the sides of the mesencephalon, and in affording exit to the second and third divisions of the fifth pair of nerves. The alisphenoid articulates in the Cod with the petrosal posteriorly, with the orbitosphenoid anteriorly, and with the mastoid and postfrontal above. Where the alisphenoids have a greater relative size, as in the Perch, and where the less constant petrosal decreases or disappears, their connections are more extensive ; they then reach the exoccipitals, and sometimes even join a small part of the basioccipital. In the incompletely ossified skulls of some fishes, e. g. the Pike and the Salmon tribe, the basal and lateral cranial bones are lined by cartilage, which forms the medium of union between them, especially the lateral ones : in better ossified fishes, e. g. the Cod, the union of the alisphenoids is by suture, partly dentated, partly squamous. In the Cod the second and third divisions of the tri- geminal nerve pass out of the cranium by the anterior notch ; in some other fishes they escape by foramina in the alisphenoid : a part of the vestibule and the anterior semicircular canal of the acoustic labyrinth usually encroach upon its inner concavity, whence some have deemed it to be the petrous bone. The chief variety in the parietals, figs. 76 and 83 7, has been noted in con- nection with the superoccipital, ib. 3. In some fishes the parietal is perforated by the ( nervus lateralis,' which supplies the vertical fins. The left parietal is broader than the right in the Halibut and some other flat fishes (Pleuronectidcz). The process for the attachment of the great trunk-muscles is developed from the outer margin of the mastoid, figs. 83, 85, s ; the inner side of this bone is expanded, and enters slightly into the formation of the walls of the cranial, or rather of the acoustic ANATOMY OF VERTEBRATES. Ill cavity ; its inner, usually cartilaginous, surface lodging part of one of the semicircular canals. It is wedged into the interspace of the ex- and par-occipitals, the petrosal, the alisphenoid, the parietal, the frontal, and postfrontal bones. The projecting pro- cess lodges above the chief mucous canal of the head, and below affords attachment to the epitympanic, or upper piece of the bony pedicle from which the mandibular, hyoid, and opercular bones are suspended. The orbitosphenoids, figs. 83, 85, 10, are osseous plates usually of a square shape, sometimes semicircular or semielliptic, as in the Cod; larger in the Malacopteri, fig. 83, 10 ; very small in most Acantlwpteri ; and sometimes represented by a descending plate of the frontal, as in the Garpike, or by unossified cartilage, as in mail-cheeked fishes. In the Carp their bases meet, like those of the alisphenoids, above the sphenoid : when osseous matter is developed in the interorbital septum the orbitosphenoids are articulated by their under and anterior part to that bone or bones, fig. 83, lo. 1 The olfactory nerves pass forward by the superior interspace of the orbitosphenoids and the optic nerves escape by their inferior interspace, or by a direct perforation ; and the essential functions of the orbitosphenoids relate to the pro- tection of the sides of the cerebrum or prosencephalon, and to the transmission of the optic nerves. The orbitosphenoids frequently bound or complete the foramen ovale. Although the frontal always enters into the formation of the cranial cavity, its major part forms the roof of the orbits, which accessory function is the chief condition of the great expanse of this neural spine in fishes. Single, and sending up a median crest in the Cod, the Ephippus, and some other fishes, the frontal is more commonly divided along the median line, the divisions having the form of long and broad subtriangular plates, fig. 76, 11, 1 1 ; narrower in the lofty compressed skulls, smaller in those with large orbits, and becoming greatly expanded in the fishes with small and deep-set eyes. Each frontal sends up its own crest in the Tunny, 2 the interspace leading to a foramen, penetrating the cranial cavity in front of the single occipital spine : a larger fontanelle exists in the Cobitis and some Siluroids between the frontal and parietal bones. In the Salamandroid fishes (e. g. Polypterus) each frontal sends down a vertical longitudinal plate, 1 The specially developed interorbital septum, or 'cranial aethmoid' of Cuvier in the Bream and Carp, misled Bojanus into the belief that it was the body of the prosencephalic vertebra (vertebra optica). Isis, 1818, p. 502. 2 Reminding one of the double spine of the neural arch of the atlas in Tetrodon. 112 ANATOMY OF VERTEBRATES. which rests directly upon the presphenoid, and intercepts a canal along which the olfactory c crura ' arc continued forward to the prefrontals : the lateral parietes of this canal thus form not only a complete, but a double bony partition between the orbits. In the Shad a corresponding descending plate takes the place of the orbitosphenoid. In most Acanthopteri an olfac- tory groove is formed by short vertical descending plates from the under surface of the frontal. The midfrontal is single in the PleuronectidcB, but has undergone more modification than any of the preceding bones in connection with the general distor- tion and loss of symmetry of the head : in the Halibut the right posterior angle is truncated, and the rest of that side scooped out, as it were, to form the lar^e orbit of the risjlit side : the left side ? o o of the bone retains its normal form : a median crest, a continuation of that upon the superoccipital, divides the two sides. The postfrontals, figs. 75, 76, 83, 12, 12, obviously belong to the same category of vertebral pieces as the mastoids, whose promi- nent crest they partly underlie and complete, lending their aid in the formation of the single (e. g. Cod, Salmon), or double, (e. g. Pike) articular cavities for the tympanic pedicle : like the mastoids they are ossified in and from the primitive cranial cartilage ; and their inner surface is expanded, but this less frequently enters into the formation of the cranial cavity : they form the posterior boundary of the orbit ; are articulated below to the orbitosphenoid and alisphenoid, above to the frontal, and by their posterior and upper surfaces to the mastoid. The vomer, figs. 83, 84, 1.3, is wedged into the under part of the presphenoid ; its antero-lateral angles are articulated to the prefrontals; its upper surface supports the nasal bone, sometimes immediately, sometimes by an intervening ethmoidal cartilage. The palatine bones abut against the expanded anterior part of the vomer, the under side of which commonly supports teeth. The left ala of the anterior end of the vomer is chiefly developed in the Halibut and other flat fishes. In the Lepidosteus, the vomer is divided into two, as in Batrachia, by a median cleft. Although its posterior end joins obliquely to the under part of the presphenoid, it is not, therefore, less a continuation of the basicranial series than is the postsphenoid, which joins in a similar manner with the basioccipital. The prefrontals defend and support the olfactory prolongations of the cerebral axis, give passage to these so-called ( olfactory nerves,' bound the orbits anteriorly, form the surface of attachment or suspension for the palatine bones, and through these for the ANATOMY OF VERTEBRATES. 113 palato-maxillary arch : they rest below upon the presphenoid and vomer, support above the fore part of the frontal and the back part of the nasal bones, and, by their outer or facial extension, give attachment to the large antorbital or lacrymal bone. They are ossified in and from pre-existing cranial cartilage. Such are the essential characters of the bones which Cuvier has called e frontaux anterieures ' l in Fishes, and to which I apply the name of ' prefrontal ' in all classes of Vertebrate animals. In the Cyprinoids, and most Halecoids, the prefrontals form part of an interorbital septum. When anchylosis begins to prevail in the cranial bones of Fishes, the prefrontals manifest their essential relationship to the vomerine and nasal bones by becoming confluent with them : thus we recognise the prefrontals in the confluent parts of the nasal vertebra of the Conger, by the external groove conducting the olfactory nerves to the nasal capsules, and by the inferior process from which the palatine bone is suspended. 2 In the Mur&nce, also, the prefrontals are plainly confluent with the nasal, is, bone, and form the well marked articular surfaces for the palato-maxillary bone. In some fishes a process of the prefrontal circumscribes the foramen by which the olfactory ' crus ' finally emerges from the anterior prolongation of the cranio-vertebral canal. In the Carp this part of the brain traverses a deep notch on the inner side of the prefrontal, fig. 83, u. In the Cod the palatine arch is chiefly but not wholly suspended to the prefron- tals. The right prefrontal is the smallest in the unsymmetrical skulls of the flat-fishes. The nasal bone is usually single, and terminates forward in a thick obtuse extremity. The anterior end of the nasal is deepest in those Fishes which have a small maxillary arch suspended from the cranial axis by vertical palatines, and which have a large 1 ' Deux frontaux anterieures, qui donnent passage aux nerfs olfactifs, ferment les orbites en avant, s'appuyent sur le sphenoide et le vomer, et donnent attache par une facette de leur horde inferieitre aux palatins.' Legons a" Anat. Comp. ii. 1837, p. 606. Compare this enunciation of the essential characters of the anterior frontals with Cuvier's descriptions of the bones to which he applies that name in other classes, and with the variable determinations of the same bones by other anatomists le lacrymal, Geoffroy and Spix ; lamina cribrosa ossis ethmoidei of Bojanus ; seitliche re'ichbeine, Meckel, Wagner. Without at present entering into the respective merits or demerits of these determinations, I shall only state that the prefrontals, under whatever names they are described, are essentially the neurapophyses of the nasal vertebra, and that the failure in the attempt to determine the special homologies of these bones may, in every case, be traced to the non- appreciation of their true general homology. 2 In the Conger, Cuvier l recognises the prefrontals as persistent cartilages. 1 Op. cit. (xui.), ii. p- 235, VOL. I. I 114 ANATOMY OF VERTEBRATES. basicranial canal. In some fishes, as the Salmonida, the nasal is broad but not deep : in Istiophorus it is long and narrow : in the Discoboles and Lophobranchii it is a short vertical compressed plate : it is altogether absent in the Lophius, or is represented here, as in the Diodon, by a fibrous membrane, retaining the primitive histological condition of the skeleton. In the Flying Gurnard the nasal has no immediate connection with the vomer ; but this is a rare exception. In most fishes the nasal cavity is more completely divided by the nasal bone into two distinct lateral fossa) than in any other class of Vertebrates. In Amia, Lepidosteus, Polypterus, and many extinct ganoid Fishes the nasal is divided at the median line. The horn-like pro- jection from the fore part of the skull of the Naseus unicornis is formed chiefly by a process of the frontal bone, to the under part of which a small nasal is articulated. The turbinals, or osseous capsules of the nose, are situated at the sides or above the nasal : the premaxillary and the maxillary bones are usually attached to its extremity through the medium of a symmetrical cartilage which is articulated with the fore part of the nasal bone, and extends forward to the interspace of the upper ends of the premaxillaries. This ( prenasal ' cartilage often forms a septum between the two ' ossa turbinata : ' it is partially ossified in the Carp. The sense-capsules are so intercalated with the neural arches, which are modified to form cavities for their reception, that the demonstration of the skull will be best facilitated by describing them before we proceed to the hrcmal arches of the cranial verte- bras. Acoustic capsule, or petrosal, figs. 81, 83, 85, 16. We have seen that the first developed cartilage upon the primitive membranous wall of the skull forms a special protecting envelope for the labyrinth, which alone constitutes the organ of hearing in Fishes (Ammocetes, fig. 58, ie). In the progressive accumulation of cartilaginous tissue upon the base and sides of the cranium, the ear-capsule loses its individuality, and becomes buried in the common thick basilateral parietes of the cranium. It is blended with that persistent cartilaginous part of the skull in the Lepidosiren ; but, in the better ossified Fishes, when the osseous centres of the neurapophyses of the cranial vertebra? begin to be established in that cartilaginous basis, a distinct bone is likewise, in most cases, developed for the more express defence of the labyrinth. Since, however, functions are less specialised, less confined to the particular organ ultimately destined for their ANATOMY OE VERTEBRATES. 115 performance in the lower than in the higher classes, we find in Fishes several bones taking part with the special acoustic capsule in the lodgment of the labyrinth ; and it is only in the higher Vertebrates that the capsule, under the name of the f petrous bone,' entirely and exclusively envelopes the labyrinth. Its ossification commences later than that of the cranial neurapo- physes, in the series of Osseous Fishes : there are species (e. g. Pike) in which, after the exoccipitals, alisphenoids, and orbito- sphenoids have received their destined amount of ossification, the petrosal still remains in the cartilaginous state : it is small in the Carp, fig. 83, 16, and Bream; in the Perch, figs. 84, 85, 16, it is more developed ; it is somewhat larger in the flat-fish (e. g. Halibut); and in the Cod, fig. 81, 16, attains an equal size with the alisphenoid, ib. 6, which it resembles in form, except that the notched margin is posterior. Here it forms the posterior lateral wall of the cranium ; articulates below with the basioccipital i, and basisphenoid, above with the mastoid 8, and paroccipital 4, behind with the exoccipital 2, and before with the alisphenoid 6 : it sup- ports the cochlear division of the labyrinth containing the otolites. The cavity called ' otocrane ' lodging the petrosal with the rest of the ear-capsule, is formed, on each side, by the exoccipital, paroccipital 4, alisphenoid 6, mastoid 8, and postfrontal 4 : it is some- times closed externally, but opens widely into the cranial cavity. The optic capsule, or sclerotal, fig. 81, 17, like the acoustic cap- sule, is cartilaginous in Plagiostomes, and also in the semi-osseous fishes, as in most Ganoids, the Lepidosiren, the Lophius, the Lophobranchs and Plectognathes. In better ossified fishes it is bony, and commonly consists of two hollow hemispheroid pieces, each with two opposite emarginations ; the inner ones circum- scribing the hole, (analogous to the meatus internus of the petrosal), for the entry of the nerves and vessels to the essential parts of the organ of vision ; and the outer or anterior emargina- tions supporting the cornea. As this part of the skeleton of the head retains its primitive fibro-membranous condition in Man, it is called f the sclerotic coat of the eye ; ' and the osseous plates developed in it in Birds, many Reptiles, and Fishes, are termed ( sclerotic bones.' It bears, however, the same essential relation to the vascular and nervous parts of the organ of sight, which the petrous bone does to those parts of the organ of hearing, and which the turbinal bones do to the organ of smell : the per- sistent independence of the eye-capsule, which has led to its being commonly overlooked as part of the skeleton, relates to the requisite mobility and free suspension of the organ of vision. In I 2 116 ANATOMY OF VERTEBRATES. the Cartilaginous Fishes, however, it is articulated by means of a pedicle with the orbitosphenoid. The osseous cavity or ( orbit ' lodging the eyeball is formed by the presphenoid, orbitosphenoid, frontal 11, postfrontal 4, prefrontal u, and palatine 20, bones : it opens widely outwards, where it is, often, further circumscribed by the chain of suborbital scale-bones below, and, but less fre- quently, by a superorbital bone above. The bony orbits in most fishes communicate freely together, or rather with that narrow prolongation of the cranial cavity lodging the olfactory crura : but, in many Malacopteri, e. g. the Shads and Erythrinus, the Citharinus and Hydrocyon, the Synbranchus, and the genus Cyprinus, fig. 83, an osseous septum, is, divides the orbits. In the Amia, Lepidosteus and Polypterus the orbits are divided by a double septum, forming the proper walls of the olfactory prolongation of the cranium, as is the case in the Batrachia. The olfactory capsules, or turlinals, fig. 81, 19, are lodged in a cavity called ( nasal,' bounded by a variable number of bones, of which the vomer, ib. 13, the prefrontals, ib. u, and the nasals, ib. 15, are the most constant : in many bony fishes the nasal chamber is closed behind by cartilage, which partly forms the interorbital septum ; but in which, in some species, a slender symmetrical bifurcate (Perch) or subquadrate ossicle is developed ; in the Cyprinoid (fig. 83, is) and Siluroid Fishes, it articulates below to the presphenoid, behind and above to the orbitosphenoids, and above and before to the frontals and prefrontals, forming the chief part of the interorbital septum. The capsules of the terminal pituitary expansion of the organ of smell are cartilaginous in the Plagiostomes, Chimreroids, in most Ganoids, and in the Lepido- siren. They form a single tube, with interrupted cartilaginous parietes, like a trachea, in several of the Cyclostomes. The tur- binals are developed for the more immediate support of each ol- factory capsule, in osseous fishes ; they are generally thin, more or less elongated, and coiled scales ; situated at the sides of the nasal bone and of the ascending processes of the premaxillaries ; usually free, but in the Gurnards articulated with the prefrontals and nasal, and in the Cock-fish (Argyreiosus) suspended above the nasal bone, from the anterior prominence of the frontal spine. The palato-maxillary arch, fig. 81, 20, 21, 22, H, presents a simple and intelligible condition in the Lepidosiren and Plagiostomous fishes ; in all it is completed or closed at one point only, viz., where the premaxillaries meet or coalesce, fig. 67, 22. The palatine bones are the piers of this inverted arch, and their points of suspension are their attachments to the prefrontals, the vomerine, and the ANATOMY OF VERTEBRATES. 117 nasal bones. The arch is completed by the maxillary and pre- maxillary bones, the symphysis of the latter forming its apex ; and it is inclined forward, nearly or quite parallel with the base of the skull ; which, in most fishes, extends to the apex of the arch, and in some far beyond it, being usually more or less closely attached to it. In air-breathing Vertebrates the arch is more de- pendent, circumscribing below the nasal or respiratory canal. The pterygoid bones project backward and outward as the appendages of the palato-maxillary arch, ib. 23. Both maxillary and intermax- illary bones tend by their peculiar developement and independent movement in bony fishes to project freely outward, downward, and backward. We find, at least, that the general form, position, and attachments of the single and simple palato-maxillary arch, in the Lepidosiren or Cestracion, are represented in most osseous fishes, by their several detached bones, the names of which have been just mentioned. The palatine (pleurapophysis of nasal vertebra, figs. 81, 84, 20) is an inequilateral triangular bone, thick and strong at its upper part, which sends off two processes : one is the essential point of suspension of the palato-maxillary arch, and articulates with the prefrontal and vomer at their point of union ; the other is convex, and passes forward to be articulated to a concavity in the superior maxillary, to which, in all Fishes, it affords a more or less moveable joint. In the Parrot-fishes and Diodons the articulation is quite analogous to that of the mandible below with the tympanic pedicle. In the Lepidosteus, Amia, and most Ganoids, it is by a suture. In the Shad the palatine articulates with the premaxillary as well as the maxillary. In the Mormyrus the palatines meet, and unite together at the median line. The posterior border is joined to the entopterygoid, fig. 84, 23, and its outer angle to the pterygoid. The palatine contributes to form the floor of the orbit and the roof of the mouth ; in many fishes it supports teeth, but is eden- tulous in the Cod. It varies much in form in different species ; is slender and elongated in the wide-mouthed voracious fishes as o the Pike, and is short and broad in the broad-headed, small- mouthed fishes. The maxillary (haemapophysis of nasal vertebra, fig. 81, 21) is usually a small edentulous bone, 1 concealed in a fold of the skin between the palatine and premaxillary : it lies, in the Cod, fig. 75, 21, posterior to and parallel with the premaxillary, 22, which it resembles in form, but is longer and thinner in most osseous fishes : 1 The Os mystaceum of ichthyotomists. 118 ANATOMY OF VERTEBRATES. the upper, usually bifurcate, end of the maxillary, forms a socket on which the ascending or nasal process of the premaxillary glides ; a posterior tubercle at this end is attached to the palatine, and ligaments connect the same expanded end to the nasal, the turbinal, the vomer, and the premaxillary : the lower and hinder expanded end of the bone is attached by strong elastic ligament, in which a labial gristle is commonly developed, to the lower jaw. In the Salmon and Herring tribe, the Sudis, fig. 86, 21, Amia, and most Ganoids, the maxillary supports teeth. In the Plecto- gnathi (Globe-fish and File-fish), the maxillaries coalesce wholly or in part with the premaxillaries. In the Lepi- dosteus the contrary condition prevails : the premaxillary and maxillary bones constitute, indeed, a single dentigerous arch or border of the upper jaw, as in Disarticulated bo.es of paiato- % 86, but are subdivided into many maxillary arch (Arapaimagigas) bony pieces, a Condition which SCCmS tO have prevailed in some of the ancient extinct ganoid fishes. In the Polypterus the maxillary is large and undivided on each side ; it supports teeth, and sends inward a palatine plate to join the vomer and the palatine bone ; thus acquiring a fixed position and all the normal features of the bone in higher animals. The maxillary bone is very diminutive in the Siluroid fishes, and appears, with the premaxillary, to be entirely wanting in certain Eels (MurcenidcB). The premaxillary (haemal spine of nasal vertebra, figs. 75, 81, 22), one of a symmetrical pair in the Cod and most other osseous fishes, is moderately long and slender, slightly curved, expanded and notched at both extremities : the anterior end is bent upward, forming the nasal process, and is attached by lax ligaments to the nasal bone and prenasal cartilage, to the palatine, and to the anterior ends of the maxillary bones. The premaxillaries are movably connected to each other by their anterior ends ; the nasal processes are separated by the prenasal cartilage, the lower or outer branches project freely downward and outward, fig. 75, 22 : the labial border of each premaxillary is beset with teeth, whilst the maxillary bone is quite edentulous in most osseous fishes, as in the Cod, ib. 21. In Dioclon the premaxillaries and their lamellated dental apparatus coalesce and constitute a single sym- metrical beak-shaped bone : the premaxillary is also single in Mormyrus. The confluent premaxillaries constitute the sword- like anterior prolongation of the snout in Xiphias, and are firmly ANATOMY OF VERTEBRATES. 119 and immovably articulated with the prenasal and maxillary bones, in both the Sword-fish and the Garpike. The premaxillaries are commonly more extended in the transverse than in the vertical direction ; but there are many examples in Fishes where their de- velopement is equal in both directions. The vertical extension, which forms the nasal branch of the premaxillary, is of unusual length in the fishes with protractile snouts, as, for example, in the Picarels (Menidce), the Dories (Zeus), and in certain Wrasses, as Coricus, and especially the Epibulus, or Sparus insidiator of Pallas, fig. 87, 22. In this fish the nasal branch of the inter- maxillary, ib. 22', plays in a groove on the upper surface of the skull, and -^^ S 7 reaches as far back as the occiput when the mouth is retracted. The descend- ing or maxillary branch is attached by a ligament, ib. 22 ", longer than itself, to the lower end of the maxil- lary bone, ib. 21, and consequently draws forward that bone, together with the lower jaW, tO which the Same end Mechanism of protraction and retraction .,, 111 T ^ tne m o utn (Epibulus insidiator) oi the maxillary is attached by liga- ment, when the long nasal branch of the premaxillary glides forward out of the epicranial groove. The protractile action is further favoured by a peculiar modification of the hypotympanic, ib. 28, which, by its great length and movable articulation at both ends, cooperates with the long premaxillary in the sudden projection of the mouth, by which this fish seizes the small, agile, aquatic insects that constitute its prey. In the Lopltius the nasal processes of the premaxillaries enter a groove in the frontal : in the Uranoscopus they also reach the frontal, playing upon the small nasal bone and pressing it down, as it were, upon the vomer. In the Dactylopterus they penetrate between the nasal and the vomer, and play in the cavity of the rhinencephalic arch. The diverging appendage of the palato-maxillary arch consists, in Fishes, of the pterygoid and entopterygoid bones, which, as they are the least important parts of the arch, so are they the least constant : they are wanting, for example, in the Synodon, Platystacus, Hydrocyon, and Lophius ; are connate with, or indistinguishable from, the palatine in most Salmonoids and Eels ; whilst in the Muraena a single bone, the pterygoid, exists, but is disconnected with the maxillary arch. Most Fishes, however, present, as in the Cod, the two bones above named. The ento- pterygoid is edentulous in the Perch, fig. 84, 23, Cod, and most 120 ANATOMY OF VERTEBRATES. other fishes, but is richly beset with teeth in the Arapaima gigas. It principally constitutes the floor of the orbit, its breadth de- pending much upon the depth of that cavity ; it sometimes is joined by its median margin to the vomer and presphenoid, as in the Cod-tribe, Carp-tribe, and Flat-fishes ; and to the basisphenoid in Lepidosteus, Erythrinus, and Pofypterus, and then divides the orbit from the mouth ; but more commonly a vacuity here exists in the bony skull, filled up only by mucous membrane in the recent fish ; in Upeneus, Polyprion, and Cheilinis, for example, the entopterygoid does not join the basisphenoid. The pterygoid forms in the Cod, fig. 75, 24, an inequilateral triangular plate, but more elongated than the palatine, with which it is dovetailed anteriorly ; it becomes thicker towards its pos- terior end, which is truncated and firmly ingrained with the anterior border of the hypotympanic ; its lower border is smooth, thickened, and concave ; edentulous in the Cod, but more fre- quently supporting teeth, as in the Perch. The pterygoid and palatine appear to form one bone in the great Sudis, (Arapaima gigas, fig. 86, 20, 24): and they are confluent in the Eel tribe. The ten bones of which the palato-maxillary arch is composed in Osseous fishes are, in the Cod and most other species, so dis- posed, in relation to the peculiar movements of the mouth, as to appear like three parallel and independent arches, successively attached behind one another, by their keystones, to the fore part of the axis of the skull, and with their piers or crura suspended freely downward and outward, fig. 75, 22, 21, except those of the last or pterygo-palatine arch, ib. 23, 24, which abut against the tympanic pedicles. The simplification or confluence of the two first of these spurious arches is eifected in the Salmonoid Fishes, Sudis, fig. 86, &c., by the shortening of the premaxillary, and by the mode of its attachment to the maxillary, which now forms the larger part of the border of the mouth and supports teeth : the maxillaries are brought into close articulation with the pala- tines in the Plectoo-nathes, and the consolidation of the whole O ' series into its normal unity is effected in the Lepidosiren. The palatines form the true bases of the inverted arch at their points of attachment to the prefrontals ; the premaxillaries constitute the true apex, at their mutual junction or symphysis ; the approxi- mation of which to the anterior end of the axis of the skull is rendered possible in fishes, by the absence of any air-passage or nasal canal ; the pterygoids are the diverging appendages of the arch ; but are attached posteriorly to strengthen the pedicle sup- porting the lower jaw, and combine its movements with those of ANATOMY OF VERTEBRATES. 121 the upper jaw ; just as the bony appendages of one costal arch in Birds associate its movements with those of the next. Tympano-mandibular arch, fig. 81, H, 25 32. This presents its true inverted or haemal character ; its apex or key-stone formed by the symphysial junction of the lower jaw hanging downwards freely, below the vertebral axis of the skull. The piers, or points of suspension, of the arch, are formed by the epitympanics : each epi- tympanic is articulated to both the postfrontal, 12, and the mastoid, 8, and is divided artificially in fig. 81 ; its articular surface is formed in the Cod by a single elongated condyle, fig. 75, 28 ; in many other fishes by a double condyle, one for each of the above-named cranial parapophyses, fig. 84, 28. In the Diodon the upper border of the epitympanic is articulated by a deeply indented suture to the frontal, the postfrontal and mastoid bones : its posterior margin supports, as in many other fishes, a circular articular sur- face for the opercular bone, fig. 84, 35. Below the condyle, the epitympanic in the Cod, fig. 75, becomes compressed laterally, but is much expanded from before backward. The almost con- stant bifurcation of both ends of the epitympanic in osseous fishes, for articulation with two cranial parapophyses above, and suspending two inverted arches below, make it appear like a coalescence of the uppermost pieces of both those arches. In most fishes the lower end is bifid, and supports both the man- clibular and the hyoidean arches; the stylohyoid, fig. 81,38, being attached near the junction of the epitympanic with the meso- tympanic. The contiguous ribs of the Chelonia are immovably connected together to ensure fixity and strength to the carapace : the bulky apparatus suspended from the parietal and frontal ver- tebra of osseous fishes demanded the additional strength in the supporting axis which is gained by the confluence of their bodies, and also by that of the proximal pieces of the pleurapophyses by which the two haemal arches are suspended from those vertebra. The anterior division of the epitympanic piece articulates with the preopercular, fig. 75, 34, the mesotynipanic, fig. 81, 26, and pretympanic, ib. 27 ; the posterior division is again bifurcate in the Cod, supporting part of the preopercular and part of the opercular bone. A strong crest projects from its outer surface in this and many other fishes. The epitympanic is simple at both ends in the Carp tribe. The mesotympanic, figs. 81, 26, 84, 38, is a slender, compressed, slightly curved, elongated bone, articulated by its upper part or base to the epitympanic and preopercular ; by its lower end to the inner side of the hypotympanic, reaching almost to the mandibular 122 ANATOMY OF VERTEBKATES. trochlea; and by its anterior border to the pretympanic. ib. b. The mesotympanic is confluent with the epitympanic in the Siluroid, the Mursenoid, and some other fishes ; but does not join the epitympanic in the Lepidosteus, being in that fish supported by the preopercular. The pretympanic, figs. 81, 27, 84, Z, is an oblong bony scale, with the posterior margin thickened and grooved for the reception of the fore part of the mesotympanic and the upper and fore part of the hypotympanic. It is confluent with the hypotympanic in the Conger and Mursena : it does not join either this or the meso- tympanic in the Lepidosteus. The hypotympanic, figs. 81, 28, 75 and 84, 2$d, is a triangular plate of bone, like the epitympanic reversed, bearing the articular convex trochlea for the lower jaw upon its inferior apex and with a straight base. The posterior margin of the hypotympanic is grooved for the reception of part of the preopercular, ib. 34, its inner side is excavated for the insertion of the pointed end of the mesotympanic, and the anterior angle is wedged between the pretympanic and the pterygoid, 24, and is firmly united to the latter ; the trochlea is slightly concave transversely, convex in a greater degree from before backwards. The Sly-bream (Epibulus, Cuv,), presents the most remarkable modification of the hypotympanic, fig. 87, 28 ; it is much elongated and slender, carrying the lower jaw at an unusual distance from the base of the skull, and it is itself movably connected at its upper end with the mesotympanic. Thus, in the extensive protractile and retractile movements of the mouth, the under jaw swings backward and for- ward on its long pedicle, as on a pendulum ; the lower jaw being further supported or steadied in those movements by a long ligament, extending from the preoperculum to its angular piece, ib. /, so. By the confluence of the meso- and epi-tympanics, and of the pre- and hypo-tympanics, in the Eel tribe, the suspensory pedicle of the lower jaw is reduced to two pieces, as in Batrachia. In the Lepidosiren it is represented, as we have seen, by a single osseous piece ; but this I regard as the homologue of only the lower half of the pedicle in the MurcencB) viz. the confluent pre- and hypo-tympanic pieces. This progressive simplification, or diminution of the multiplied centres of ossification of the tympanic pedicle of Fishes, even within the limits of the class, has mainly weighed with me in rejecting the Cuvierian view of its special homologies ; according to which, not only the squamotemporal bone and the malar bone of higher animals, but also the ( syrnplectic ' a peculiar ichthyic bone --are superadded to the 'tympanic' or quadrate bone of Reptiles and Birds, in the formation of the ANATOMY OF VERTEBRATES. 123 suspensory pedicle of the under jaw of Fishes. Ascending to the higher generalisations of homology, we see in the tympanic pedicle a serial repetition of the palatine bone ; and, in both, the ribs or pleurapophyses of contiguous vertebras specially modified for the masticatory functions of the arches they support. The mandible, figs. 81, 84, 29, 32, is the lower portion of the arch, being articulated to the hypotympanics above, and closed by a ligamentous union or bony symphysis with its fellow at its lower end. The term f ramus ' is applied in Anthropotomy to each half of the mandible, and each ramus consists of two, three, or more pieces in different fishes. Most commonly it consists of two pieces, one (hremapophysis proper, 29,) articulated to the suspensory pedicle, and edentulous, analogous to the maxillary ; and the other (haemal spine, 32,) completing the arch, and commonly supporting teeth, like the premaxillary. In the Cod, and some other fishes, a third small piece is superadded, at the angle of the posterior piece, fig. 75, so. The dentary, 32, is deeply excavated, and receives a cylindrical cartilage, the remnant of the embryonal haemal arch, fig. 69 A, d, and the vessels and nerves of the teeth. The Sudis, fig. 88, the Polypterus, and Amia, have the splint- like plate along; the inner OO surface of the ramus, called e splenial : ' it supports teeth and developes a coronoid pro- cess. In both Sudis and Le- pidosteus there is superadded a Lower jaw ' (Arapaima gi(ja * small bony piece, ib. 29 a, answering to the surangular in Reptiles. The Diverging Appendage of the tympano-mandibular arch consists of the bones which support the gill-cover, a kind of short and broad fin, the movements of which regulate the passage of the currents through the branchial cavity, opening and closing the branchial aperture on each side of the head. The first of these 'opercular' bones is the preopercular, fig. 75, 34, which is usually the longest in the vertical direction. In the Gurnards, or ' mailed- cheeked' Fishes, fig. 82, the preopercular is articulated with the enormously developed suborbital scale bone, 73. Three bones usually constitute the second series of this appendage : the upper one is commonly the largest and of a triangular form, thin and with radiated lines like a scale : it is the opercular, figs. 75, 84, 35 : in the Cod it is principally connected with the posterior margin of the preopercular, and below with the subopercular, ib. 36 ; but it has usually, also, a partial attachment to the outer angle of the epitympanic, fig. 84 ; and is some- 124 ANATOMY OF VERTEBRATES. times (Diodon, Lopliius, Anguilla) exclusively suspended therefrom. In the Lophius piscatorius the opercular is a long and strong bone suspended vertically from the convex epitympanic condyle, and with a long and slender fin-ray proceeding from the back part of that joint. The subopercular forms the chief part of the opercular fin by its long backwardly produced lower angle. The sub- opercular bone in the Conger is soon reduced to a mere ray, which curves backwards and upwards like one of the branchio- stegals. The opercular itself, though shorter and retaining more of its laminated form, also shows plainly, by its length and curva- ture in the Eels, its essential nature as a metamorphosed ray of the tympanic fin. We have seen that all the framework of this fin had the form of rays in the Plagiostomes. In Muraena the small opercular bones articulate only to the under half of the tympanic pedicle. The subopercular is wanting in the Shad. The lowermost bone, called the interopercular, figs. 75, 84, 37, is articulated to the preopercular above, to the subopercular behind, and usually to the back part of the mandible ; it is attached, also, in the Cod, by ligament to the ceratohyoid in front. The inter- opercular and preopercular are the parts of the appendage which are most elongated in the peculiarly lengthened head of the Fistularia. The third inverted arch of the skull is the c hyoidean,'fig. 81, 38-4 1, and is suspended, in Osseous Fishes, through the medium of the epi- tympanic bone, 25, to the mastoid, s ; showing it to be the ha3mal arch of the parietal vertebra. The first portion of the arch, stylohyal, fig. 85, 38, is a slender styliform bone, which is attached at the upper end by ligament to the inner side of the epitympanic, close to its junction with the mesotympanic, and at the lower end to the apex of a triangular plate of bone, which forms the upper portion of the ( great cornu.' I apply to this second piece, which is pretty constant in fishes, the name of epihyal, ib. 39 : the third longer and stronger piece is the ceratohyal, ib. 40. The keystone or body of the inverted hyoid arch is formed by two small subcubical bones on each side, the basihyals, ib. 41. These complete the bony arch in some fishes : in most others there is a median styliform ossicle, extended forward from the basihyal symphysis into the substance of the tongue, called the glosso-hyal, ib. 42 ; and another symmetrical, but usually triangular, compressed bone, which expands as it extends backwards, in the middle line, from the basihyals ; this is the urohyal, ib. and fig. 75, 43. It is connected with the symphysis of the coracoids, which closes below the fourth of the cranial inverted arches, and it thus forms the isthmus which ANATOMY OF VERTEBRATES. 125 separates below the two branchial apertures. In the Conger the hyoidean arch is simplified by the persistent ligamentous state of the stylohyal, and by the confluence of the basihyals with the ceratohyals ; a long glossohyal is articulated to the upper part of the ligamentous symphysis, and a long compressed urohyal to the under part of the same junction of the hyoid arch. The glosso- hyal is wanting in the Mur&nophis. The Diverging Appendage of the hyoidean arch retains the form of simple, elongated, slender, slightly curved rays, articulated to depressions in the outer and posterior margins of the epi- and cerato-hyals : they are called ( branchiostegals,' or gill-cover rays, fig. 85, 44, because they support the membrane which closes externally the branchial chamber. The number of these rays varies, and their presence is not constant even in the bony Fishes : there are but three broad and flat rays in the Carp ; whilst the clupeoid Elops has more than thirty rays in each gill-cover : the most common number is seven, as in the Cod, fig. 75, 44. They are of enormous length in the Angler, and serve to support the membrane which is developed to form a great receptacle on each side of the head of that singular fish. The fourth cranial inverted arch, fig. 81, so 52, H, is that which is attached to the paroccipital ; or to the paroccipital and mastoid ; or, as in the Cod, to the paroccipital and petrosal ; or as in the Perch, fig. 85, so, and Shad, to the paroccipital and basioccipital : thus either wholly or in part to the parapophysis of the occipital vertebra, of which it is essentially the haemal arch ; it is usually termed the ( scapular arch.' In the Eel tribe, where it is very feebly developed, and sometimes devoid of any diverging appendage, it is loosely suspended behind the skull ; and in the Plagiostomes, fig. 30, si, 52, it is not directly attached to its proper vertebra, the occiput, but is removed further back, where we shall usually find it displaced in higher Vertebrates, in order to allow of greater freedom to the move- ments of the head. The superior piece of the arch, f supra-scapular,' figs. 81, 85, 50, is bifurcate in the Cod, or consists of two short columnar bones, attached anteriorly, the one to the paroccipital, the other and shorter piece to the petrosal, and coalescing posteriorly at an acute angle, to form a slightly expanded disk, from which the second piece of the arch is suspended vertically. This piece, called ' scapula,' ib. 51, is a slender, straight bone, terminating in a point below, and mortised into a groove on the upper and outer side of the lower and principal bone of the scapular arch. The 126 ANATOMY OF VERTEBRATES. supraseapula and scapula together represent the rib or pleur- apophysis of the occipital vertebra ; they are always confluent in the Siluroicls. The lower bone ' coracoid,' ib. 52, completes the arch. In the Cod its pointed upper extremity projects behind the scapula ; the middle part developes backward a broad plate giving attachment to the radiated appendage of the arch : the lower end bends inward and forward gradually decreasing to a point,, which is usually connected to that of the opposite coracoid by ligament, and also to the urohyal. In the Siluridas the coracoids expand below, and are united together by a dentated suture. In all Fishes they support and defend the heart, and form the frame or ' sill ' against which the opercular and branchiostegal doors shut in closing the branchial cavity : they also give attach- ment to the aponeurotic diaphragm dividing the pericardial from the abdominal cavity. The bones of the head being in completest number, de- parting least from the vertebral pattern, and susceptible of the most intelligible definitions in the class of Fishes, afford the best basis for determining their homologies and fixing their nomencla- ture in the higher vertebrate series. 31. Skull of Chelonia. If the back part of the skull of a Turtle (Chelone, fig. 89) be compared with that of a Cod, fig. 77, it will be seen that the lowest bone, i, offers an articular surface for the centrum of the atlas, passes for- ward, expanding, to articulate with the basisphenoid, supports the ' medulla ob- longata,' and is suturally articulated above to the pair of bones, fig. 89, 2, 2, which pro- tect the sides of the epencephalon. These, Back view of cranium, Tin-tie , , i i ^ -t moreover, transmit the nypogiossal and vagal nerves, develope each an articulation for the neurapophyses of the atlas, and converge above to support the keystone of the arch, 3. We have, thus, unmistakeable characters of the basi- ex- and super-occipitals ; there is also a bone, 4, wedged between the ex- 2, and super- 3, occipitals mesially, and joined laterally to the mastoid, 8 : excavated on its inner surface by the postero- external semicircular canal, and produced on its outer surface for the insertion of the ( biventer cervicis ' and ' complexus ' ; it is the homologue of the paroccipital (' occipitale externe,' Cuvier), and bears the same general relation to the hindmost vertebral segment of the skull which the mastoid, 8, does to the next segment in advance. ANATOMY OF VERTEBRATES. 127 In fig. 90, the centrum i, neurapophyses 2, and neural spine 3, of the epencephalic arch, are seen from their inner or cranial surface : with the increasing bulk of the brain, the spine, 3, begins to expand laterally, and take a greater share in roofing over the hinder part or epencephalon : the parapophysis, 4, is excluded in this view. The gristly capsule of the ear-organ fills up the otocrane formed by the bones, 2, 3, 5, and 6 ; and extends outward and backward to enter the basal cavity of 4, the par- 90 Section of cranium, Turtle (diclone mydas) occipital : were ossification to extend into the acoustic capsule, either from an independent centre, like 16, figs. 81, 83, 84, or by continuous growth from any of the otocranial bones, the true homologue of the ( petrosal ' or ' petrous portion of the temporal bone ' of Anthropotomy would be established. In some Emydians there is a small autogenous bony plate in the acoustic cartilage, close to the foramen caroticum. The basisphenoid, 5, continues forward the series of cranial centrums, expands beneath the cranial cavity, articulates on each side with the alisphenoid, 6, and sends out from its under and lateral surface a plate to articulate with the pterygoids, fig. 98 B, 24, and, in the Emys, with the petrosal. The alisphenoid, 6, fig. 90, protects the side of the mesencephalon (optic lobe), is widely notched anteriorly by the emerging divisions (2nd and 3rd) of the trigeminal nerve, is perforated posteriorly by a filament of the acoustic nerve, where it joins the cartilaginous petrosal ; it articulates above with the mastoid, 8, and parietal, 7, and in front with the orbitosphenoid, 11. The anterior semicircular canal is partly lodged in the cavity of the otocrane contributed 128 ANATOMY OF VERTEBHATES. by the alisphenoid. Thus in the bone 6, we have all the characters of that so numbered in figs. 81, 83, and 85, and called 'ali- sphenoid ' in the fish. The chief modification is due to the greater developement of 3, fig. 90, in Chelonia, which overlaps 6 as well as 2. The parietals, figs. 90, 91, are united, as in Cyprinoid and Ganoid fishes, by the sagittal suture, and are much expanded both transversely and longitudinally, overlapping, in the Turtle, the 91 18 Skull of Turtle (Chelone mydas) superoccipital, fig. 90, 3, and articulated with it and the mastoids, fig. 91, 8, behind; and with the frontals, ib. n, before. Each parietal, also, sends down a long vertical plate, 7', fig. 90, which unites with the alisphenoid, 6, and orbitosphenoid, 10, this ossifica- tion taking the place and function of the latter neurapophyses in fishes. The bone, figs. 89, 91,8, which articulates with the paroccipital 4, parietal 7, and postfrontal 12, which aifords the surface of attach- ment to the upper end of the tympanic 28, enters into the for- mation of the acoustic chamber in some Emydians, and projects outward and backward to give insertion to the latissimus colli and trachelomastoideus, repeats the chief and essential characters of the bone so numbered, and called ' mastoid ' in Fishes, figs. 75, 76, 83, 85,8: and forms the transverse process of the parietal vertebra. The forward continuation of the vertebral bodies from 5 remains cartilaginous : the lower half of the sides of the prosencephalon ANATOMY OF VERTEBRATES. 129 are defended partly by fibre-cartilage, partly by the exogenous descending lamellae, 7', of the parietals : there are no separate ossifi- cations answering to 9 and 10 in fishes. 1 The frontals, fio;s. 90. 91, O O ' 11, are supported like an arch between the parietals 7 and pre- frontals, 1 4 : and each sends down a longitudinal lamella, bound- ing the sides of the narrow anterior continuation of the brain- o chamber, as in Polypterus ; but continued by an unossified plate to the cartilaginous presphenoid and vomer below. The postfrontal, fig. 91, 12, extends from its connections with the frontal n, and parietal 7, downward and backward to unite with the mastoid, 8, in the Turtle, and with the malar, 2fi, and squamosal, 27, in all Chelonia. It forms the posterior boundary of the orbit, but does not contribute any share to the proper cranial walls. The median symmetrical bone, fig. 90, 13, which, like a hypa- pophysis, is developed in the lower part or production of the notochordal capsule, which underlaps the anterior end of the basi-pre-sphenoid, 5, by its narrow hinder part, - - expanding as it advances to articulate with the prefrontals, 14, having the pala- tine bones, ib. 20, abutting against the broad anterior part, and entering by its under surface into the formation of the roof of the mouth, fig. 98 B, n, repeats the essential characters of the bone so numbered and termed f vomer ' in Fishes, figs. 81, 83, 84, 85, 13 ; and, like it, represents the centrum of the foremost segment of the vertebral series. The vomer is single in Chelonia, as in most fishes. The bones, fig. 90, 14, in neurapophysial relation with the vomer, protecting the sides of the rhinencephalon or olfactory bulbs, entering into the antero-superior boundary of the orbit, forming part of the surface of attachment of the palatines, supporting the fore part of the frontals, and connected, but more commonly connate, with the nasals, ib. is, fig. 91, 14, repeat the essential characters of the prefrontals of Fishes, figs. 83, 85, 1 4. Connate, as in Chelonia they usually are, with the nasals, their outer expanded plate unites with the maxillary, fig. 91, 21, and completes the upper border of the nostril, is. The palatines, figs. 90, 98 B, 20, form the sides of the roof of the mouth, articulating medially with the vomer, is, n, and laterally with the maxillary, 21, and pterygoid, 24. The maxillary, figs. 90, 91, 21, presents a palatal, facial, and orbital plate. The palatal plate, fig. 98 B, 21, developes a masticatory ridge parallel with the sharp alveolar border. The facial plate, fig. 91, 21, shows the con- nections with the prefronto-nasal, u, the premaxillary, 2-2, and the malar, 26 ; the orbital plate is usually perforated by the lacrymal 1 xxxviir. ; Tab. xxi. fig. 89, 1, r. VOL. I. K 130 ANATOMY OF VERTEBRATES. canal, the bone so called being ossified continuously, as a process, from the maxillary. The premaxillaries, figs. 90, 91, 22, closing the arch anteriorly, are very small in all Chelonia, and the sutures marking them oft' from the maxillaries are wanting in some Mud-turtles (Tetronyx longicollis, Fitz. Trionyx Bibroni): 1 the premaxillary part of the facial profile is vertical in many Chelonia, as in fig. 91 : but in Tetronyx it extends from the nostril down- ward and backward the reverse of prognathism. The pterygoid, fig. 90, 24, diverges from the vomer and pala- tine, or from the palatine and maxillary, fig. 98 B, backward and outward : uniting, in Chelone, with its fellow below the basi- sphenoid, fig. 90, 5, and diverging outward and backward to abut, at , against the tympanic, 28. In some Soft-turtles, e.g., Trionyx ( Gymnopus) indicus, the vomer is directly continued from the basi- pre-sphenoid, and divides the pterygoids from each other. A second outer bar of bone, fixing the maxillary arch to the tympanic, is present in all Chelonia, and divided into two pieces. The proximal piece, fig. 91, 26, is articulated with the maxillary, 21, enters into the lower and back part of the orbital border, unites superiorly with the postfrontal, 12, and posteriorly with the second piece, 27. To the bone, 26, the term 'malar' is given; to the bone, 27, the name ( scjuamosal.' The latter, resembling a vertical scale or plate, articulates above with the postfrontal, 12, and mastoid, 8 ; and behind with the tympanic, 28. It completes the arch called ( zygomatic,' bounding externally the temporal fossa, which is roofed over by bone in the Turtles (figs. 89 and 91), and a few Emyds ; but is widely open above in other Chelonia. The tympanic pedicle is a single bone, fig. 91,28, expanded above, with a more or less circular border for the insertion of the mem- brana tympani ; excavated internally by the tympanic air-cells ; notched behind for the reception of the colurnellar stapes, as in the Turtles, fig. 91 ; with a narrower cleft in Tetronyx, and with the borders uniting in the Tortoises and some other Chelonia. O * reducing the stapedial passage to a foramen or canal, fig. 92, 28. The lower end of the tympanic supports a transversely extended condyle with an undulated or nearly flattened surface. The tym- panic articulates above with the paroccipital, fig. 89, 4, in some species with the alisphenoid, in others with the superoccipital, as well as with the mastoid, ib. and fig. 91, 8. The mandible consists of an 'articular' element, small, but dis- tinct in the Turtle, fig. 91, ao; connate in Emys with the f suran- 1 XLIV. No. 954, p. 185. ANATOMY OF VERTEBRATES. 131 gular,' fig. 92, 29 ; of an ( angular ' continued into a f splenial,' ib. so ; of a ' coronoid,' ib. 29' ; and of a f dentary,' ib. 32. All Chelonia are edentulous : the alveolar borders of both upper and lower jaws are sheathed with horn : but in a few species, especially the soft turtles ( Trionyx, Tetronyx} these borders are notched or pro- duced into tooth-like processes. The dentary elements coalesce at the symphysis ; which, in the Snappers, especially Chelydra ( Chelonura) Temminckii, is produced into a sharp hook. The hyoid arch consists of a basihyal, fig. 92, 41, a pair of short 92 ii Side view of cranial vertebras, Emys processes, ib. e, giving attachment to the genio- and hyo-glossi muscles : of a pair of long ceratohyals, 40, by which the arch is suspended to the mastoids ; and of a pair of hyobranchials, 47. To complete the series of skull-bones homologous with those of the fish, represented in fig. 81, it is necessary to bring forward the scapular arch which had receded a short distance from its vertebra in the Batrachia, fig. 42, 52, from a more remote position in the Chelonia : we then find that 51, fig. 92, answers to the scapula, fig. 81, 51 ; and that 52, fig. 92, answers to the coracoid, fig. 81, 52 : the diverging series of many-jointed rays in the fish, fig. 81, are now developed into the fore-limb, fig. 92, 53 58. K 2 132 ANATOMY OF VERTEBRATES. In this figure the several bones of the head of the European Box-terrapene (Emys EuropcBa, Wgl.) are represented, disarti- culated, in a side view of their vertebral relations. Beneath the Roman figure, I, are the centrum, i. neurapophysis, 2, neural spine, 3, and parapophysis 4, forming the neural (epencephalic) arch ; with the pleurapophysis, 51, and haemapophysis, 52, forming the haemal (scapular) arch, with its appendage, of the occipital vertebra. Beneath n are the centrum, 5, the neurapophysis, 6, the neural spine, 7, the parapophysis, 8, forming the neural (mesencephalic) arch : from 8 is suspended by an unossified pleurapophysis the hremapophysis, 40, the haemal spine, 41, with the appendage, 47, of the haemal (hyoidean) arch of the parietal vertebra. Under in are the neurapophysis, 10, neural spine, n, and parapophysis, 12, forming the neural (prosencephalic) arch ; with the pleurapophysis, 28, and composite haemapophysis, 29 32, forming the haemal (mandibular) arch of the frontal vertebra, of which the centrum is not an independent ossification. Beneath IV are, the centrum, 13, the connate neurapophyses and neural spines, 14, forming the neural (rhinencephalic) arch ; with the pleurapophysis, 20, haemapophysis, 21, and haemal spine, 22, forming the haemal (maxillary) arch of the nasal vertebra. The diverging appendages, for the fixation of this haemal arch are more developed than in Fishes, where it retains more of its typical mobility. Besides the appendage, 24, of the pleurapo- physis, there is now another, extending in two successive segments, 26 and 27, from the haemapophysis. The splanchnic ossicle, 16', is part of the acoustic organ : the circle of bones, 17, belong to the visual organ. Such are the ' general homologies ' of the bones of the chelonian head, in reference to the vertebrate archetype, fig. 21. Compared with bones of the piscine head, fig. 81, previously named and characterised, those of fig. 92 are : 1. Basioccipital. 2. Exoccipital. 3. Superoccipital. 4. Paroccipital. 5. Basisphenoid. 6. Alisphenoid. 7. Parietal. 8. Mastoid. 9. Presphenoid (unossified). 10. Orbitosphenoid (in great part cartilaginous), n. Frontal. 12. Postfrontal. ANATOMY OF VERTEBRATES. 133 is. Vomer. u. Prefrontal (with, is, nasal, distinct in some Clielonia). 16. (Petrosal, unossified from an independent centre); 16', a superadded ossicle, s stapes,' ( columella ' ; with a gristly represen- tative of ( malleus; ' in special relation to an organ of hearing affected by vibrations of air : superadded to all the bones developed in and from the embryonic haemal arch called ( Meckel's process.' 17. Sclerotals. 19. Turbinal (unossified). 20. Palatine. 21. Maxillary. 22. Premaxillary. 24. Pterygoid, with ossification extending into the seat of 23, ento-pterygoid. 26. Malar (not answering to the bone so numbered in fig. 81). 27. Squamosal (ib. these bones do not exist in Fishes). 28. Tympanic (here a single bone ; its subdivisions are 25 28 in fig. 81). 29. Articular with Surangular. 29'. Coronal. so. Angular with Splenial. 32. Dentary. 40. Ceratohyal. 41. Basihyal. 47. Cerato-branchial, (or ( thyrohyal ' in reference to the larynx of air-breathers, a new developement upon the vestige of the branchial apparatus of fishes). 50. Suprascapula (unossified). 51. Scapula. 52. Coracoid. 52'. Acromial process of scapula. 53. Humerus (rarely a separate ossification in Fishes). 54. Ulna. 55. Radius. 56. Carpus. 57. 58. Digital rays. The chief differences in regard to the presence and absence of bones between the Tortoise and the Fish are seen in those belonging to the category of ( diverging appendages : ' thus the < branchiostegals,' 43, and ' operculars,' 34 37, fig. 81, are sup- pressed in the Reptile ; while the ' malar,' 26, and squamosal, 27, are not developed in the fish. Some minor, but interesting, modifications of cranial structure present themselves within the 134 ANATOMY OF VERTEBRATES. limits of the Chclonian order. Figs. 90 and 91 exemplify that which prevails in the marine species (Chelone}. In them the head is proportionally larger ; and, being incapable of retraction within the carapace, is additionally protected by extension of bone into the fascia covering the temporal muscles, so as to form a complete osseus vaulted roof over the temporal fossae, due to exogenous growths from the postfrontals, fig. 91, 12, the parietals, 7, and the mastoids, 8. In the almost sole instance in which such accessory defence is afforded to a non-marine species the Brazilian Pipitu (Podo- cnemis expansa] - - the temporal roof is chiefly formed by the parietals, and is completed laterally by a larger proportion of the squamosal than of the postfrontal, which does not exceed its relative size in other Terrapenes. The present species further differs from the marine Turtles in the non-ossification of the vomer and the consequent absence of a septum in the posterior nostrils ; in the greater breadth of the pterygoids, which send out a compressed rounded process into the temporal depressions : the orbits also are much smaller, and are bounded behind by orbital processes of the postfrontal and malar bones : the mastoids and paroccipitals are more produced backward, and the entire skull is more depressed than in the Turtles. In other freshwater Tortoises (Emys, &c.), the parietal crista is continued into the occipital one without being extended over the temporal fossa? ; the fascia covering the muscular masses in these fossae undergoing no ossification. The bony hoop for the membrana tympani is incomplete behind, and the columelliform stapes passes through a notch instead of a foramen to attain the tympanic membrane. The mastoid is excavated to form a tympanic air-cell. In the true Tortoises the temporal depressions are exposed, as in the Terrapenes : the head is proportionally small and can be withdrawn beneath the protective roof of the carapace. The skull is rounder and less depressed than in the Terrapenes. The tympanic hoop is notched behind, but the columelliform stapes passes through a small foramen. The palatine processes of the maxillaries are on a plane much below that of the continuation of the basis cranii formed by the vomer and palatines. In the soft-turtles ( Trionicidce), the skull is long, depressed, triangular, the muzzle forming the obtuse apex, and the base remarkable for its four backward prolongations. The inferior of these is the shortest, and terminates in the occipital condyle ; the superior is the longest, and is formed by the superoccipital spine : the two lateral processes are developed from the paroccipitals and ANATOMY OF VERTEBRATES. 135 mastoids. The premaxillary is either wanting,, or it is very small, and represented by its alveolar border only ; the maxillaries meeting above it. The alveolar borders of both upper and lower jaws show a regular series of vascular pits or foramina, indica- tive of the primitive separate matrices, like those of teeth, winch laid the foundation in the young animal of the continuous horny coverings of the jaws. Temminck's Snapper (Chelonura Temminckii) is remarkable for the upper convexity and enormous expanse of the cranium, chiefly due to the temporal fossae, contrasted with the short and narrow face. In a fossil chelonian from the Portland stone ( Ch. planiceps) and in another from the Chalk ( Ch. pulcliriceps) the nasals were distinct from the prefrontals, which is a rare exception in existing species. 93 Nnr Side view of cranial vertebra and sense-capsules, Crocodile 32. Skull of Crocodilia. Passing next to the skull of the Crocodile, we find the first difference in the less complex condition of the epencephalic arch, fig. 93, N i, which consists of four, instead of, as in the Fish and Turtle, six bones. The basioccipital, figs. 93 and 94, i, presents, like the centrums of the trunk, a convexity at its posterior articular surface ; but its anterior one, like the hind- most centrum of the sacrum, unites with the next centrum in ad- vance, ib. 5, by a flat rough sutural surface. Like most of the cen- trums in the neck and beginning of the back, that of the occiput developes a hypapophysis, but this descending process is longer and larger. The exoccipitals, ib. 2, articulate suturally, like the neur- apophyses of the trunk, with the upper and lateral parts of their 136 ANATOMY OF VERTEBRATES. centrum; are concave mesially, fig. 94, 2, towards the brain-segment which they protect, meeting above it to support the neural spine, 3 ; they develope a petrosal plate, which meets a corresponding one from the alisphenoid ; they give exit to the vagal and hypo- glossal nerves, and send outward a strong process, fig. 93, 4, which articulates with the mastoid and tympanic. The anterior and inner part of the base of this process is excavated by part of the acoustic cavity : its outer extremity is rough for the attach- ment of muscles : it thus repeats the essential characters of the ' par- occipital ' in the Fish and Turtle ; but it is ossified, as an exogenous transverse process, from the neurapophysis (exoccipital, 2). The superoccipital, figs. 93 and 94, 3, is broad and flat, like the similarly detached neural spine of the atlas ; it advances a little forward, beyond its sustaining neurapophyses, to protect the upper surface of the cerebellum ; it is traversed by tympanic air-cells, and assists with the ex- and par-occipitals, 2, 4, in the formation of the ear-chamber. Proceeding with the neural arches of the Crocodile's skull, if we dislocate the segment in advance of the occiput, fig. 93, N 2, we bring away, in connection with the long base-bone, 5, the bone, 9. The two connate cranial centrums must be artificially divided, in order to obtain the segments distinct to which they belong. The hinder portion, 5, of the great base-bone, which is the centrum of the parietal vertebra, is the basisphenoid. It supports that part of the ( mesencephalon,' which is formed by the lobe of the third ventricle, and its upper surface is excavated for the pituitary prolongation of that cavity. The basisphenoid developes from its under surface a ( hypapophysis,' which is suturally united with the fore part of that of the basioccipital, but extends further down, and is similarly united in front to the ' pterygoids,' fig. 94, 24. These rough sutural surfaces of the long descending process of the basisphenoid are very characteristic of that centrum, when detached, in a fossil state. The neurapophyses of the parietal vertebra, 6, 6, the ali- sphenoids, protect the sides of the mesencephalon, and are notched at their anterior margin, for a conjugational foramen transmitting the trigeminal nerve. As accessory functions they contribute, like the corresponding bones in fishes, to the formation of the ear- chamber. They have, however, a little retrograded in position, resting below in part upon the occipital centrum, and supporting more of the spine of that segment, 3, than of their own, 7. The spine of the parietal vertebra (parietal, figs. 93, 94, 95, 7), is a single, depressed bone, like that of the occipital vertebra ; it completes the mesencephalic arch, as its crown or key-bone ; it is partially ANATOMY OF VERTEBRATES. 137 excavated by the tympanic air-cells, and overlaps the superoccipital. The bones, ib. 8, 8, wedged between 6 and 7, are developed from independent centres, and preserve their individuality, as in Fishes. They form no part of the inner walls of the cranium, but are partially excavated by the tympanic cavity, and send outward and backward a strong transverse process for muscular attachment. They afford a ligamentous attachment to the haemal (hyoid, fig. 93, H 2, 40) arch of their own segment, and articulate largely with the pleurapophyses, (tympanic, ib. 2s), of the antecedent hremal arch, whose more backward displacement, in comparison with its position in the fish's skull, is well illustrated in the metamorphosis of the Frog, figs. 69 A and 71. On removing the neural arch of the parietal vertebra, after the section of its confluent centrum, the elements of the corresponding arch of the frontal vertebra, fig. 93, N in, are seen to present the same arrangement. The compressed produced centrum (presphe- noid, ib. 9) has its form modified like that of the vertebral centrums at the opposite extreme of the body in many birds. The neurapo- physes (orbitosphenoids, ib. 10) articulate with the upper part of 9 ; they are expanded, and smoothly excavated on their inner surface to support the sides of the large prosencephalon, showing more plainly their archetypal character than in Chelonia ; they dismiss the optic nerves by a notch. They show the same tendency to a retrograde change of position as the neighbouring neurapophyses, 6 ; for though they support a greater proportion of their proper spine, 11, they also support part of the parietal spine, 7, and rest, in part, below upon the parietal centrum, 5. The neural spine, n, of the frontal vertebra retains its normal character as a single symmetrical bone, like the parietal spine which it partly overlaps ; it also completes the neural arch of its own segment, but is remark- ably extended forward, where it is much thickened, and assists in forming the cavities for the eyeballs; it is the ( frontal' bone. In contemplating in the skull itself, or such side view as is given in fig. 94, the relative position of the frontal, n, to the parietal, 7, and of this to the superoccipital, 3, which is overlapped by the parietal, just as itself overlaps the flattened spine of the atlas, we gain a conviction which cannot be shaken by any difference in their mode of ossification, by their median bipartition, or by their extreme expansion in other animals, that the above-named single, median, imbricated bones, each completing its neural arch, and permanently distinct from the piers of such arch, must repeat the same element in those successive arches --in other words, must be ' homotypes,' or serially homologous. In like manner the 138 ANATOMY OF VERTEBRATES. serial homology of those piers, called e neurapophyses,' viz., the lamina; of the atlas, the exoccipitals, the alisphenoids, and the orbitosphenoids, is equally unmistakable. Nor can we shut out of view the same serial relationship of the paroccipitals, fig. 95, 4, as coalesced diapophyses of the occipital vertebra, with the mastoids, ib. 8, and the postfrontals, 12, as par- or di-apophyses of their respective vertebra?. All stand out from the sides of the cranium, as tranverse processes for muscular attachment ; all are alike autogenous in the Turtles ; and all of them, in Fishes, offer articular surfaces for the ribs of their respective vertebrae ; and these characters are retained in the postfrontals as well as in the mastoids of the Crocodiles. The frontal diapophysis, figs. 93, 95, 12, is wedged between the back part of the spine, 11, and the neurapophysis, 10; its out- wardly projecting process extends backward, and joins that of the succeeding diapophysis, 8 ; but, notwithstanding the retro - gradation of the inferior arch, it still articulates with part of its own pleurapophysial element, 28, which forms the proximal element of that arch. There finally remain in the cranium of the Crocodile, after the successive detachment of the foregoing arches, the bones termi- nating the fore part of the skull, 1ST 4, fig. 93 ; but, notwithstand- ing the extreme degree of modification to which their extreme position subjects them, we can still trace in their arrangement a correspondence with the vertebrate type. A long and slender symmetrical grooved bone, fig. 93, 13, is con- tinued forward from the coalesced pterygoids, 24, and stands in the relation of a centrum to the vertical plates of bone, 14, which expand as they rise into a broad, thick, triangular plate, with an exposed horizontal superior surface. These bones, the prefrontals 14, stand in the relation of neurapophyses to the rhinencephalic prolonga- tions of the brain commonly called ' olfactory nerves ; ' and they form the piers or haunches of a neural arch, which is completed above by a pair of symmetrical bones, is, called ' nasals,' which I regard as a divided or bifid neural spine ; the independent basal ossification, answering to the vomer in Fishes, figs. 81, 84, 13, and Chelonians fig. 98 B, n, is in advance of its proper segment, and divided in the middle line as in Ganoid Fishes and Batrachia. In some Alligators (All. niger) the divided vomer extends far for- ward, expands anteriorly, and appears upon the bony palate. Almost all the other bones of the head of the Crocodile are adjusted so as to constitute four inverted arches. These are the haemal arches of the four segments or vertebras, of which the neural arches have been just described. But they have been the ANATOMY OF VERTEBRATES. 139 seat of much greater modifications, by which they are made sub- servient to a variety of functions unknown in the haemal arches of the rest of the body. Thus the two anterior hremal arches of the head perform the office of seizing and bruising the food ; are armed for that purpose with teeth : and, whilst one arch is firmly fixed, the other works upon it like the hammer upon the anvil. The elements of the fixed arch, called f maxillary arch,' fig. 93, H, iv, have accordino-lv undergone the greatest amount of change, in o *> o o o order to adapt that arch to its share in mastication, as well as for forming part of the passage for the respiratory medium which traverses it. Almost the whole of the upper surface of the max- illary arch is firmly united to contiguous parts of the skull by rough or sutural surfaces, and its strength is increased by bony 94 Section of cranium, Crocodile appendages, which diverge from it to abut against other parts of the skull. Comparative Anatomy teaches that, of the numerous places of attachment, the one which connects the maxillary arch by its element, 20, with the centrum, is, and with the descending plates of the neurapophyses, u, of the nasal segment, is the normal or the most constant point of its suspension ; the bone, 20, being the pleurapophysial element of the maxillary arch : it is called the ' palatine,' because the under surface forms a portion of the bony roof of the mouth, called the ' palate,' as in fig. 98 c, 20. It is articulated at its fore part with the bone, 21, which is the haema- pophysial element of the arch. This bone is called the ' maxil- lary,' and is greatly developed both in length and breadth, fig. 95, 21 : it is connected with 20, figs. 94, 98 C, behind and with 22 in front, which are parts of the same arch, and with the diverging appendages of the arch, viz., fig. 95, 26, the malar bone, and fig. 98 c, 25, the ectopterygoid : the maxillary is also united with the nasals, is, and the lacrymal, 73, as well as with its fellow of the opposite side. The smooth, expanded 140 ANATOMY OF VERTEBRATES. horizontal plate, which effects the latter junction, is called the palatal plate of the maxillary, fig. 98 c, 21 ; the thickened external border, where this plate meets the external rough surface of the bone, and which is perforated for the lodgment of the teeth, is the ( alveolar border.' The haemal spine or key-bone of the arch, 22, is bifid, and the arch is completed by the symphysial junction of the two symmetrical halves ; these halves are called f premaxillary bones : ' these bones, like the maxillaries, have a rough facial plate, fig. 95, 22, and a smooth palatal plate, fig. 98 c, 22, with the connecting alveolar border. The median symphysis is perforated vertically through both plates ; the outer or upper hole being the external nostril, fig. 95, 22, the under or palatal one being the premaxillary aperture, fig. 98 C, p. Both the palatine and the maxillary bones send outward and backward parts or processes which diverge from the line of the haemal arch, and give attachment to distinct bones, which form the ' diverging appendages ' of the arch, and serve to attach it, as do the diverging appendages of the thoracic haemal arches in the bird, to the succeeding arch. The appendage, 24, called ' pterygoid,' effects a more extensive attachment, and is peculiarly developed in the Crocodilia, As it extends backward it expands, unites with its fellow both below and above the nasal canal, encompassing it so as to form the hinder or palatal nostril, fig. 98 C, n ; the coalesced pterygoids articulate anteriorly with the divided vomer, the palatines, and the basi- pre-sphenoid : posteriorly each broad wing, extending outward, gives attachment to a second bone, ib. 25, called l ectopterygoid,' "which is firmly connected with the maxillary, 21, the malar, 26, and the postfrontal, 12. The second diverging ray of the maxil- lary arch is of great strength ; it extends from the maxillary, fig. 95, 21, to the tympanic, 28, and is divided into two pieces, the malar, 26, and the squamosal, 27 ; both of which begin to assume more lengthened and slender proportions than in the Turtle (compare fig. 95 with 91). Such are the chief Crocodilian modifications of the hamial arch and appendages of the anterior or nasal vertebra of the skull. The hamial arch of the frontal vertebra, fig. 92, H, iii, is someAvhat less metamorphosed, and has no diverging appendage. It is slightly displaced backward, and is articulated by only a small proportion of its pleurapophysis, 28, to the parapophysis, 12, of its own segment ; the major part of that short and strong rib articulating with the parapophysis, 8, of the succeeding segment. The bone, fig. 95, 2$, called 'tympanic,' because it serves to support ANATOMY OF VERTEBRATES. 141 the c drum of the ear ' in air-breathing vertebrates, is short, strong, and immovably wedged, in the Crocodilia, between the paroccipital, 4, mastoid, 8, postfrontal, 12, and squamosal, 27 ; and the conditions of this fixation of the pleurapophysis are exemplified in the great developement of the hasmapophysis (mandible), which is here unusually long, supports numerous teeth, and requires, therefore, a firm point of suspension, in the violent actions to which the jaws are put in retaining and overcoming the struggles of a powerful living prey. The movable articulation between the tympanic, 28, and the rest of the haemal arch is analogous to that which we find between the thoracic pleurapophysis and haamapophysis in birds. But the haamapophysis of the mandibular arch in the Crocodiles is subdivided into several pieces, in order to combine the greatest elasticity and strength with a not excessive weight of bone. The different pieces of this adaptively subdivided element have received definite names. That numbered 29, fig. 93, which offers the articular concavity to the convex condyle of the tympanic, 28, is called the 6 articular ' piece ; that beneath it, so, which developes the angle of the jaw, when this projects, is the 'angular' piece; the piece above, 29, and e, fig. 95, is the ' surangular ;' the thin, broad, flat piece, 31, fig. 93, applied, like a splint, to the inner side of the other parts of the mandible, is the ( splenial ; ' the small accessory ossicle, 31', is the f coronoid,' because it developes the process, so called, in lizards ; the anterior piece, 32, which supports the teeth, is called the ( dentary.' The purport of this subdivision of the lower jaw-bone has been well explained by Conybeare 1 and Buckland, 2 by the analogy of its structure to that adopted in binding together several parallel plates of elastic wood or steel to make a crossbow, and also in setting together thin plates of steel in the springs of carriages. Dr. Buckland adds ' Those who have witnessed the shock given to the head of a crocodile by the act of snapping together its thin long jaws, must have seen how liable to fracture the lower jaw would be, were it composed of one bone only on each side.' The same reasoning applies to the composite structure of the long tympanic pedicle in fishes. In each case the splicing and bracing together of thin flat bones of unequal length and of varying thickness, affords compensation for the weakness and risk of fracture that would otherwise have attended the elongation of the parts. In the abdomen of the crocodile the analogous subdivision of the haamapophyses, there called abdo- minal ribs, allows of a slight change of their length, in the expansion and contraction of the walls of that cavity ; and since 1 'Geol. Trans.' 1821, p. 565. - 'Bridgewater Treatise,' 1836, vol. i. p. 176. 142 ANATOMY OF VERTEBRATES. amphibious reptiles, when on land, rest the whole weight of the abdomen directly upon the ground, the necessity of the modifi- cation diminishing liability to fracture further appears. These analogies are important, as demonstrating that the general homo- logy of the elements of a natural segment of the skeleton is not affected or obscured by their subdivision for a special end. The purposive modification of the hrcmapophyses of the frontal vertebra is but a repetition of that which affects the same elements in the abdominal vertebrae. Passing next to the haemal arch of the parietal vertebra, fig. 93, ir, ii, we are first struck by its small relative size. Its restricted functions have not required it to grow in proportion with the other arches, and it consequently retains much of its embryonal dimensions. It consists of a ligamentous ' stylohyal,' retaining the same primitive histological condition which obstructs the ordinary recognition of the pleural element of the lumbar haamal arches ; of a cartilaginous ( epihyal,' 39, intervening between this and the ossified hamiapophysis, or ceratohyal, 40 ; and of the haemal spine, 4J, which retains its cartilaginous state, like its homotypes, in the abdomen : there they get the special name of l abdominal sternum,' here of f basihyal.' The basihyal has, however, coalesced with the thyrohyals to form a broad cartilaginous plate, the anterior border rising like a valve to close the fauces, and the posterior angles extending beyond and sustaining the thyroid and other parts of the larynx. The long bony ( ceratohyal ' and the com- monly cartilaginous ' epihyal ' are suspended by the ligamentous f stylohyal ' to the back part of the tympanic at its junction with the paroccipital process ; the whole arch having, like the man- dibular one, retrograded from the connection it presents in Fishes. This retrogradation is still more considerable in the succeeding haemal arch, fig 92, H i ; fig. 57, 51. In comparing the occipital segment of the Crocodile's skeleton with that of the Fish, fig. 81, the chief modification that distinguishes that segment in the Cro- codile is the apparent absence of its haemal arch. We recognise, however, the special homologues of the constituents of that arch of the Fish's skeleton, fig. 34, in the bones 51 and 52 of the Cro- codile's skeleton, fig. 57 ; but the upper or suprascapular piece, 50, fig. 92, retains, in connection with the loss of its proximal or cranial articulations, its cartilaginous state : the scapula, 51, is ossified, as is likewise the coracoid, 52, the lower end of which is separated from its fellow by the interposition of a median, symmetrical, partially ossified piece called ' episternum.' The power of recog- nising the special homologies of 50, 51, and 52 in the Crocodile, ANATOMY OF VERTEBRATES. 143 with, the similarly numbered constituents of the same arch in Fishes though masked, not only by modifications of form and proportion, but even of very substance, as in the case of 50 depends upon the circumstance of these bones constituting the same essential element of the archetypal skeleton, viz. the fourth haemal arch, numbered pi, 52, in fig. 17. For although in the pre- sent instance there is superadded to the adaptive modifications above cited the rarer one of altered connections, Cuvier does not hesitate to give the same names, f suprascapulaire' to 50, and f scapulaire' to 51, in both Fish and Crocodile ; but he did not per- ceive or admit that the narrower relations of special homology were a result of, and necessarily included in, the wider law of general homology. According to the latter law, we discern in fig. 93, 50 and 51, a compound ' pleurapophysis,' in 52 a ( haemapophy sis,' and in hs, the ( haemal spine,' completing the haemal arch. 1 The scapulo-coracoid arch, both elements, 51, 5-2, of which retain the form of strong and thick vertebral and sternal ribs in the Crocodile, is applied in the skeleton of that animal over the anterior thoracic haemal arches. Viewed as a more robust hamial arch, it is obviously out of place in reference to the rest of its vertebral segment. If we seek to determine that segment by the mode in which we restore to their centrums the less displaced neural arches of the antecedent vertebras of the cranium or in the sacrum of the bird, 2 we proceed to examine the vertebra before and behind the displaced arch, with the view to discover the one which needs it, in order to be made typically complete. Finding no centrum and neural arch without its pleurapophyses from the 1 The author of No. CLXXI, in criticising this conclusion, omits consideration of the cartilaginous element, fig. 93, so : as it exists and required due attention, I was led to regard it as the homologue of the ossified element, figs. 81, 85, 50, in Fishes, and as being part, one might say, half, of the pleurapophysis. No anatomist has impugned such determination of the special homology of the ' lame cartilagineuse du bord spinal de Pomoplate ' of the Crocodile, with the ' partie spinal del'omoplate ' of the Frog, and with the ' os surscapulaire ' of the Fish. Now the latter is the homotype of the proximal half of the compound pleurapophysis of the pelvic arch, of which the part called ' ilium ' answers to the part called ' scapula.' There remains, therefore, for Dr. Humphrey's consideration, the serial and general homologies of the * suprascapula ; ' in the omission of which lurks the fallacy of his criticism. CLXXI, pp. 27, 28. The alleged difference of developement, at most one of direction of growth, is futile. A ' haemal arch ' having been defined as including the ' pleurapophysis ' as well as * hremapophysis,' by altering the meaning of the term and restricting the ' haemal parts of the vertebra ' to the ' ha?mapophyses and hasmal spine,' Dr. Humphrey makes ground for pronouncing the part of the hcemal arch, 50 and 51, in figs. 81 and 92, as being the hasm- not the pleur-apophysis. 2 See 'On the Archetype and Homologies of the Vertebrate Skeleton,' pp. 117 and 159. 144 ANATOMY OF VERTEBRATES. scapula to the pelvis, we give up our search in that direction ; and in the opposite direction we find no vertebra without its ribs, until we reach the occiput ; there we have centrum and neural arch,, with connate parapophyses, but without the haemal arch, which arch can only be supplied by a restoration of the bones 50-52 to the place which they naturally occupy in the skeleton of the fish. And since the bones 50-52 in the Crocodile, fig. 57, are specially homologous with those so numbered in the Fish, fig. 34, we must conclude that they are likewise homologous in a higher sense ; that in the Fish the scapula-coracoid arch is in its natural or typical position, whereas in the Crocodile it has been displaced for a special purpose. Thus, agreeably with a general principle, we perceive that, as the lower vertebrate animal illus- trates the closer adhesion to the archetype 1 by the natural articu- lation of the scapulo-coracoid arch to the occiput, so the higher vertebrate manifests the superior influence of the antagonizing power of adaptive modification by the removal of that arch from its proper segment. The anthropotomist, by his mode of counting and defining the dorsal vertebrae and ribs, admits, unconsciously perhaps, the important principle in general homology which is here exemplified ; and which, pursued to its legitimate consequences, and further applied, demonstrates that the suprascapula and scapula are the modified rib of that centrum and neural arch, which he calls the f occipital bone ; ' and that the change of place which chiefly masks that relation (for a very elementary acquaintance with Compara- tive Anatomy shows how little mere form and proportion affect the homological characters of bones), differs only in extent, and not in kind, from the modification which makes a minor amount of comparative observation requisite, in order to determine the relation of the shifted dorsal rib to its proper centrum in the human skeleton. With reference, therefore, to the occipital vertebra of the Cro- codile, if the comparatively well-developed and permanently distinct ribs of all the cervical vertebrae prove the scapular arch to belong to none of those segments, 2 and if that haamal arch be required to complete the occipital segment, which it actually does complete in fishes, then the same conclusion must apply to the same arch in other animals, up to man himself. 1 The term ' simple primary form ' appears to Dr. Humphrey, CLXXI, p. 34, to be more correct than the word ' archetype.' 2 Close the eyes to the fact of the suprascapular element in the Crocodile, and you then may, with Dr. Humphrey, see its representative in one of the cervical pleur- apophyses. Comp. ib. p. 28, and note, p. 144, of the present work. ANATOMY OF VERTEBRATES. 145 The locomotive extremity, fig. 92, 53-57, is the diverging ap- pendage of the arch, under one of its numerous modes and grades of developement. Coadjusted as the above-defined vertebral elements are in the skull of the Crocodile, they compose such a whole as is represented in fig. 95. Each temporal fossa is circumscribed externally by two horizontal bony arches ; the upper one formed by the post- frontal, 12, and mastoid, 8 ; the lower one by the malar, 26, and squamosal, 27 : the tympanic, 28, and mastoid, 8, bound the fossa behind : the coarticulated processes from the postfrontal and malar form a partial division between the fossa and the orbit in 95 Skull of Crocodile front. The orbit is circumscribed by these bones, with the frontal, 11, prefrontal, u, and lacrymal, b. A superorbital or palpebral derm-ossicle strengthens the upper eyelid. The external nostril, single and advanced in Crocodilia, is surrounded sometimes, as in Gavials, by the premaxillaries, 22 ; sometimes, as in fig. 95, admit- ting also the points of the nasals, is. The internal nostril opens far back, beneath the occiput, fig. 98 c, n, and is exclusively surrounded by the pterygoids, 24 : its plane is horizontal in Gavials and some Alligators ; but is more or less oblique, looking backward, in Crocodiles. Behind and above it are the median and lateral Eustachian bony outlets, from which the membranous continua- tions of the tubes converge and unite in the single valvular aper- ture on the soft palate. L The vast extent of the bony roof of the 1 CLXXII., pi. xii. fig. 5. This paper may be referred to for other cranial foramina, and for the details of the complex bony structure of the median and lateral VOL. I. L 146 ANATOMY OF VERTEBRATES. mouth is interrupted by the large ' ptery go-maxillary ' vacuities, ib. //, bounded externally by the maxillaries and ectopterygoids : at the fore part is the small ( prepalatine ' opening, ib. p. In the Gavials each pterygoid expands at its outer and fore border into a large oval bulla. The palatines and maxillaries are excavated by sinuses communicating with the nasal passages. The form of the maxillo-premaxillary palatine suture helps by its variation to the distinction of species. 1 The anterior expanded parts of the divided vomer appear upon the bony palate in some Alligators. 2 The otic capsule remains in great part cartilaginous : towards the cranial cavity it is defended by the thin otocranial plates of the alisphenoid, superoccipital and paroccipital, with occasionally a small scale, representing a rudimental petrosal. The eye-capsule is not defended by bony plates, as in Chelonia. The turbinals remain cartilaginous. o The cranial cavity is miserably small in these huge cold-blooded Carnivora; its main part, shown in section, fig. 94, 2, o, 10, may be filled by a man's thumb in a skull of three feet in length. The proper brain-chamber is, however, continued along the groove beneath the interorbital platform to the second slight expansion between the prefrontals, 14, where the rhinencephalic (olfactory) lobes send forward the true olfactory nerves. If the foregoing statement of the grounds for determining the homologies, general and special, of the skull-bones of the Crocodilia may have seemed tedious or unnecessary, I excuse myself by the importance attached to the subject by Cuvier, who, in the last lecture which he delivered, stated : l If we were agreed as to the Crocodile's head, we should be so as to that of other animals ; be- cause the Crocodile is intermediate between mammals, birds, and fishes.' Admitting, with some latitude, the reason, a sense of the importance of a determination of the bones answerable to those previously defined in Chelonia and Fishes, has influenced me in the foregoing description of the skull of the Crocodilia. 33. Skull of Opliidia.- -The skull in Lacertians and Ophi- dians departs from the vertebral pattern by a greater degree of confluence and a minor extent of neurapophysial ossification, than in Crocodilia : and that of Serpents manifests more strongly the principle of adaptive developement. Eustachian canals in Crocodilia. See also the preparations, XLIV., Nos. 706, 727, 728, 750, pp. 154164. 1 Ib. XLIV., p. 1 63, where that characteristic of Crocodilus rhombifer is specified. 2 Ib. No. 764, p. 166. ANATOMY OF VERTEBRATES. 147 In the Python, figs. 96 and 97, the basioccipital, i, is subhex- agonal, broadest anteriorly, smooth and concave above, suturally rough on each side, with a recurved pointed hypapophysis : the hinder facet forms the lower half of the occipital condyle, on each 96 Section of the Skitil of a Python side of which is a small sharp process. The basioccipital unites above with the exoccipital, 2, and alispheiioid, 6 ; and in front with the basisphenoid, 5. The exoccipitals (2, 2) are each pro- duced backward into a peduncular process supporting a moiety of the upper half of the occipital condyle : at the outer side of the base of the peduncle is an obtuse process, forming the upper part of the ridge continued upon the basioccipital. The outer and fore part of the exoccipital expands, and is perforated by a slit for the eighth pair of nerves, articulates below with the basioccipital, is excavated in front to lodge the petrosal cartilage where it articulates with the alisphenoid, and unites above with the superoccipital, 3. This is of a subrhomboidal form, sends a spine from its upper and hinder surface, expands laterally into oblong processes, is notched anteriorly and sends down two thin plates from its under surface, bounding on the mesial side the surface for the cerebellum, and by the outer side forming the inner and upper parts of the acoustic cavities. The superoccipital articulates below with the exoccipitals and alisphenoids, and in front with the parietal, by which it is over- lapped in its whole extent. The occipital vertebra is as if it were sheathed in the expanded posterior outlet of the parietal one, the centrum resting on the oblique surface of that in front, and the anterior base of the neural spine entering a cavity in and being overlapped by that of the preceding neural spine : the analogy of this kind of ' emboitement ' of the occipital in the parietal vertebra with the firm interlocking of the ordinary vertebra? of the trunk is L 2 148 ANATOMY OF VERTEBRATES. very interesting : the end gained seems to be, in grovelling reptiles liable to have the head bruised, an extra protection of the epencephalon- -the most important segment to life of all the primary divisions of the cerebrospinal axis. The thickness of its immediately protecting walls (formed by the basi-, ex-, and super-occipitals) is equal to that of the same vertebral elements in the human skull ; but they are moreover composed of very firm and dense tissue throughout, having no diploe : the epen- cephalon also derives a further and equally thick bony covering from the basisphenoid and the parietals, the latter being partly overlapped by the mastoids, fig. 97, 8, which form here a third layer of the cranial Avail. The basisphenoid, fig. 96, 5, and presphenoid, 9, form a single 97 2-2 Skull of a Python bone, and the chief keel of the cranial superstructure. The posterior articular surface looks obliquely upward and backward, and supports that of the vertebral centrum behind, as the posterior ball of the ordinary vertebrae supports the oblique cup of the succeeding one : here, however, all motion is abrogated between the two vertebrae, and the co-adapted surfaces are rough and sutural. The basisphenoid presents a smooth cerebral channel above for the mesencephalon, in front of which a deep depression (sella) sinks abruptly into the expanded part of the bone, and there bifurcates, each fork forming a short cul-de-sac in the sub- ~ stance of the bone. The transverse processes from the under and lateral surfaces are well marked, strong, but short, much thicker in the Python than in the Boa. The alisphenoids, 6, form the anterior half of the fenestra ovalis, which is completed by the exoccipitals ; and in their two large perforations for the posterior divisions of the fifth pair of nerves, as well as in their ANATOMY OF VERTEBRATES. 149 relative size and position, the alisphenoids agree with those of the Frog. Each alisphenoid is a thick suboval piece, with a tuber- cular process on its under and lateral part : it rests upon the basisphenoid and basioccipital, supports the posterior part of the parietal and a portion of the mastoid, 8, and unites anteriorly with the descending lateral plate of the parietal bone. The parietal, 7, is a large and long, symmetrical roof-shaped bone, with a median longitudinal crest along its upper surface, where the two originally distinct moieties have coalesced. It is narrowest posteriorly, where it overlaps the superoccipital, and is itself overlapped by the mastoid : it is convex at its middle part on each side the sagittal spine, and is continued downward and in- ward to rest immediately upon the basisphenoid, 5. This part of the parietal seems to be formed by an extension of ossification along a membranous space, like that which permanently remains so in the Frog, between the alisphenoid and orbitosphenoid : the mesencephalon and the chief part of the cerebral lobes are protected by this unusually developed spine of the mesencephalic vertebra. The optic foramina are conjugational ones, between the anterior border of the lateral plate of the parietal and the posterior border of the corresponding plate of the frontal. The frontals, n, rest by descending lateral plates, representing connate orbitosphenoids, upon the presphenoidal prolongation of the basisphenoid : the upper surface of each frontal is flat, sub- quadrate, broader than long in the Boa, and the reverse in the Python, where the roof of the orbit is continued outward by a detached superorbital bone : there is a distinct, oval, articular sur- face near the anterior median angle of each frontal to which the pre- frontal, 14, is attached : the angle itself is slightly produced to form the articular process for the nasal bones. The smooth orbitosphe- noidal plate of the frontal joins the outer margin of the upper surface of the frontal at an acute ano-le ; the inner side of each frontal is o * deeply excavated for the prolongation of the cerebral lobes, and the cavity is converted into a canal by a median vertical plate of bone at the inner and anterior end of the frontal. The frontals join the parietals and postfrontals behind, and, by the connate orbito- sphenoid plates, the presphenoid below, the prefrontals and nasals before, and the superorbitals at their lateral margins. The orbito- sphenoidal plates have their bases extended inward, and meet below the prosencephalon and above the presphenoid, as the neurapo- physes of the atlas meet each other above the centrum. The anterior third part of such inwardly produced base is met by a downward production of the mesial margin of the frontal, forming a septum 150 ANATOMY OF VERTEBRATES. between the olfactory prolongations of the brain, but is not con- fluent with the frontal septum : the outer portion of the orbitosphe- noidal plate is smooth externally, and deeply notched posteriorly for the optic foramen. The post-frontal, fig. 97, 4, is a moderately long trihedral bone, articulated by its expanded cranial end to the frontal and parietal, and bent down to rest upon the outer and fore angle of the ecto- pterygoid, 25. It does not reach that bone in the Boa, nor in poisonous Serpents. In both the Boa and Python it receives the anterior sharp angle of the parietal in a notch. The natural segment which terminates the cranium anteriorly, and is formed by the vomerine, prefrontal and nasal bones, is very distinct in the Ophidians. The vomer is divided, as in some ganoid Fishes and Batrachians, but is edentulous : each half is a long, narrow plate, smooth and convex below, concave above, with the inner margin slightly raised : pointed anteriorly, and with two processes and an inter- vening notch above the base of the pointed end. The prefrontals, 14, are connate with the lacrymals. The two bones which inter- vene between the vomerine and nasal bones are the turbinals, fig. 96, d, they are bent longitudinally outwards in the form of a semicylinder about the termination of the olfactory nerves. The spine of the nasal vertebra is divided symmetrically as in the Frog, forming the nasal bones, fig. 97, 15 ; they are elongated, bent plates, with the shorter upper part arching outward and downward, completing the olfactory canal above ; and with a longer median plate forming a vertical wall, applied closely to its fellow, except in front, where the nasal process of the premaxillary is received in the interspace of the nasals. The acoustic capsule remains in great part cartilaginous : there is no detached centre of ossification in it : to whatever extent this capsule is ossified, it is by a continuous extension from the alisphe- noid. The long stapes, fig. 97, 16, extends from the ' fenestra vesti- buli ' to the subcutaneous ear-drum attached to the tympanic bone, 28. The sclerotic capsule of the eye is chiefly fibrous, with a thin inner layer of cartilage ; the olfactory capsule is in a great measure ossified, as above described. Maxillary arch.- -The palatine, fig. 96, 20, or first piece of this arch is a strong, oblong bone, having the inner side of its obtuse anterior end applied to the sides of the prefrontals and turbinals, and, near its posterior end, sending a short, thick process upward and inward for ligamentous attachment to the lacrymal, and a second similar process outward as the point of suspension of the ANATOMY OF VERTEBRATES. 151 maxillary bone : between these processes the palatine is perforated, and behind them it terminates in a point. The chief part of the maxillary bone, 21, is continued forward from its point of suspen- sion, increasing in depth, and terminating obtusely : a shorter process is also, as usual, continued backward. The point of suspension of the maxillary forms a short, narrow, palatine process : the dental branch of the supramaxillary nerve penetrates the upper and fore part of this process, and its chief division escapes by a foramen on the outer and fore part of the maxillary. A space occupied by elastic ligament intervenes between the maxillary and the premaxillary, 22, which is single and symmetrical, and firmly wedged into the nasal interspace : the anterior expanded part of this small triangular bone supports two teeth. Thus the bony maxillary arch is interrupted by two ligamentous intervals at the sides of the premaxillary key-bone, in functional relation to the peculiar independent movements of the maxillary and palatine bones required by Serpents during the act of engulfing their usually large prey. Two bones extend backward as appendages to the maxillary arch ; one is the ( pterygoid,' 24, from the palatine, the other the ectopterygoid, 25, from the maxillary. The pterygoid is continued from the posterior extremity of the palatine to abut against the end of the tympanic pedicle : the under part of its anterior half is beset with teeth, fig. 96, 24. The ectopterygoid, 25, overlaps the posterior end of the maxillary, and is articulated by its posterior obliquely cut end to the outer surface of the middle expanded part of the pterygoid. Mandibular arch.- -The tympanic bone, 28, is a strong, trihedral pedicle, articulated by an oblique upper surface to the end of the mastoid, 8, and expanded transversely below to form the antero- posteriorly convex, transversely concave, condyle for the lower jaw. This consists chiefly of an articular 31, and a dentary 32, with a small coronoid and splenial piece. The articular piece, 31, including the angular and suranffular elements of the Crocodile, ends ob- o ~ tusely, immediately behind the condyle : it is a little contracted in front of it, and gradually expands to its middle part, sends up two short processes, then suddenly contracts and terminates in a point wedged into the posterior and outer notch of the dentary piece. The articular is deeply grooved above, and produced into a ridge below. The coronoid is a short compressed plate : the splenial is a longer plate applied to the inner side of the articular and dentary. The outer side of the dentary has a single perforation near its anterior end : this is united to that of the opposite ramus by elastic ligament. 152 ANATOMY OF VERTEBRATES. The skull of the Boa Constrictor differs from that of the Python, not only in the greater breadth of the frontals, but in that of the nasals ; in the absence of the superorbital, in the more slender and cylindrical form of the ectopterygoid, and in the larger and higher internal border of the coronoid. But the mechanism of the jaws is the same. By the elastic matter join- ing together the extremities of the maxillary and mandibnlar bones, those on the right side can be drawn apart from those on the left, and the mouth can be opened not only vertically, as in other vertebrate animals, but also transversely, as in insects. Viewing the bones of the mouth that support teeth in the great constricting serpents, they offer the appearance of six jaws --four above and two below ; the inner pair of jaws above are formed by the palatine and pterygoid bones, fig. 96, 20-24, the outer pair by the maxillaries, ib. 21, the under pair by the mandibles, or ' rami,' as they are termed, of the lower jaw, fig. 97, 31-32. Each of these six jaws, moreover, besides the movements ver- tically and laterally, can be protruded and retracted, independently of the other : by these movements the Boa is enabled to retain and slowly engulf its prey, which may be much larger than its own body. At the first seizure the head of the prey is held firmly by the long and sharp recurved teeth of all the jaws, whilst the body is crushed by the overlapping coils of the serpent ; the death-struggles having ceased, the Constrictor slowly uncoils, and the head of the prey is bedewed w T ith an abundant slimy mucus : one jaw is then unfixed, and its teeth withdrawn by being pushed forward, when they are again infixed, further back upon the prey ; the next jaw is then unfixed, protruded, and reattached ; and so with the rest in succession- -this movement of protraction being almost the only one of which they are susceptible whilst stretched apart to the utmost by the bulk of the animal encompassed by them : thus, by their successive movements, the prey is slowly and spirally introduced into the wide gullet. In comparing the skull of a poisonous with that of a constrict- ing Serpent, the differential characters consist, in the Rattlesnake (^Crotalus) e.g., chiefly in the modification of form and attach- ments of the maxillary, which is movably articulated to the palatine, ectopterygoid, and lacrymal bones ; but chiefly supported by the latter, which presents the form of a short, strong, three- sided pedicle, extending from the anterior external angle of the frontal to the anterior and upper part of the maxillary. The articular surface of the maxillary is slightly concave, of an oval shape : the surface articulating with the ectopterygoid on the poste- ANATOMY OF VERTEBRATES. 153 rior and upper part of the maxillary is smaller and convex. The maxillary bone is pushed forward and rotated upon the lacrymal joint by the advance of the ectopterygoid, which is associated with the movements of the tympanic pedicle of the lower jaw by means of the true pterygoid bone. The premaxillary is edentulous. A long, perforated poison-fang is anchylosed to the maxillary. The palatine bone has four or five, and the pterygoid from eight to ten, small, imperforate, pointed, and recurved teeth. The frontal bones are broader than they are long : there are no superorbitals. A strong ridge is developed from the under surface of the basisphenoid, and a long and strong recurved spine from that of the basioccipital ; these give insertion to the powerful e longi colli ' muscles, by which the downward stroke of the head is performed in the infliction of the wound by the poison-fangs. The skull of the typical Ophidian reptiles most resembles that of Lizards, but lacks the outer diverging appendage, formed by the malar and squamosal, of the maxillary arch. It differs from that of Batrachians in the distinct basi- and superoccipitals ; in the superoccipital forming part of the ear-chamber ; in the basioc- cipital combining with the exoccipital to form a single articular condyle for the atlas ; in the ossification of the membranous space between the elongated parietals and the sphenoid ; in the constant coalescence of the parietals with one another ; in the connation of the orbitosphenoids with the frontals, and in the meeting of the orbitosphenoids below the prosencephalon upon the upper sur- face of the presphenoid ; in the presence of distinct postfrontals, and the attachment thereto of the ectopterygoids, whereby they form an anterior point of suspension of the lower jaw, through the medium of the pterygoid and tympanic bones ; in the connation of the prefrontals and lacrymals. In the AmphisbcEna fulifjinosa coalescence still further simplifies the cranial structure : the parts of the epencephalic arch con- stitute a single occipital bone ; the superoccipital crest extends forward into a sagittal one ; a small foramen marks the boundary : the premaxillary is single, and, with the rest of the upper jaw, is fixed ; the tympanic is short, compared with that of true Serpents, and extends almost horizontally forward, in a line with the lower jaw which it supports ; the coronoid is more developed. The nostrils, divided by the premaxillary, are terminal ; or even, as in Lepidosternon, may open behind the fore end of the skull : in this A mphisb amian the maxillaries overlap the nasals to join the premaxillary. l 1 CLXXIII., pi. 15, figS. 8, 11. 154 ANATOMY OF VERTEBRATES. 34. Skull of Lacertilia.- -Lizards, like Serpents, have the cra- nial bones, especially those of the haemal arches and appendages, more elongated, slender, and liberated than in Crocodiles and Che- lonians; the temporal vacuities and orbits are large, and the external nostrils are apart. Lizards retain the malo-squamosal bar connect- ing the maxillary with the tympanic ; and some of them develope, as in the Crocodile, the upper zygomatic arch formed by the post- frontal and mastoid. The neurapophysial walls of the parietal and frontal segments retain much of their fibro-cartilairinous tissue ; and O O the cranial roof is there sustained by a bony pillar on each side ( f columella ' of Cuvier), which has its base implanted in a fossa of the pterygoid, and underprops the parietal near its outer border. The homologies of the cranial bones of the Python, figs. 96 and 97, with those of the Crocodile, figs. 93, 94, and 95, being recog- nised, those of any Lizard will be readily understood. In a New Zealand Gecko (liliynclioceplialus ! ) the occipital con- dyle is unusually elongated transversely, and presents the form of a crescentic, convex bar, bent upward. The basisphenoid sends down two short processes to abut against the pterygoids. The parietal bone is perforated by a small median fontanelle close to the sagittal suture : its upper surface presents two strong curved and approximated temporal crests, divided by a median, angular, longitudinal furrow : the crests are continued outward upon the posterior bifurcated part of the parietal to be continuous with that forming the upper border of the mastoid : the frontal is divided by a median suture, as is the parietal in the common Gecko. The posterior frontal supports a strong, obtuse ridge forming the back part of the frame of the orbit, and unites below \vith the malar and behind with the mastoid. The premaxillary bones are divided by a median suture, and their dentigerous border projects below the level of that of the maxillary bones. The vomer is likewise divided by a median suture. The palatal apertures of the nostrils are bounded behind by the vomer and palatal plate of the maxillary : this plate is of unusual breadth, as compared with the Lizards generally, and presents the unusual peculiarity of a dentigerous ridge parallel with the posterior half of the alveolar border. It is situated close to the inner side of this border, leav- ing only space sufficient for the reception of the teeth of the under jaw. The teeth are confluent with the summits of the proper and accessory alveolar ridges. The palatine bones are united together along the anterior halves. The rami of the lower jaw are not anchylosed at the symphysis. The alveolar border is 1 CLVIII. ANATOMY OF VERTEBRATES. 155 serrated by a single row of anchylosed teeth. The coronoid piece is triangular, rises into a point, and presents a smooth articular surface on its inner side, adapted to the anterior lateral projection of the pterygoid. In the skull of the black Scink ( Cyclodus nic/cr-), the frontal and parietal bones are thick and expanded ; the parietal is bifurcated behind, and articulated with the mastoids and paroccipitals. The postfrontals are separated from the malars by the squamosals, which extend between the malars and the mastoids to form the strong lateral bony arch resting anteriorly upon the malar and the maxillary, and posteriorly on the parietal and tympanic. Con- comitantly with the strong osseous roof of the cranium, there is an arrest of osseous developement in the fibro-membranous neurapophysial walls of the cranium : two lateral processes extend downward into these walls from the parietal and for- ward from the exoccipitals ; but the protective office of the alisphenoids is solely performed by the columnar { columellas,' which extend from the interspaces of the processes above mentioned, to rest upon the upper groove of the pterygoids. The orbitosphenoids are represented by still more slender bony styles, which circumscribe the outlets for the optic nerves, and form the anterior boundary of the prosencephalic division of the cranium. The lacrymal bones are large and divided on each side, as in most Lizards. The premaxillaries are confluent, and their nasal process separates the external nostrils from each other. Each pterygoid presents a rough surface towards the palate, but does not support teeth. There is a small ossicle between the pterygoid processes of the sphenoid and the true pterygoid bones. The columelliform stapes is extremely long and slender. In the Iguana the parietal supports a single median crest : the posterior margin of the frontal is notched by the fronto-parietal fontanelle : both lacrymal and postfrontal are subdivided into two pieces ; the lacrymal foramen is a e conjugational ' one between the two pieces. The upper portion of the lacrymal represents the facial part of the prefrontal ; it does not send down a neurapo- physial plate to join the vomer or palatines, nor forms any part of the lateral walls of the rhinencephalic cavity, or of the foramen for the transmission of the olfactory nerves. The palatine nostrils, fig. 98, D, n, are very long, and notch the large palatines, 20; the pterygoids, 24, each support a row of small teeth. In the skull of the Monitor Lizard ( Varanus niloticus) the basioccipital sends down a pair of short, obtuse hypapophyses : those of the basisphenoid are larger and abut against the ptery- loG ANATOMY OF VERTEBRATES. goicls : these bones are applied to the back part of the tympanic, and the slender e columella' rests upon the middle of their upper surface. The parietal is perforated near its anterior border. The postfrontal has a descending postorbital process. The pre- frontal developes a partial post-lateral wall for the rhinencephalic chamber ; externally it supports an antorbital dermal bone : the small perforated lacrymal is a distinct bone. The nasal and pre- maxillary are both single bones, as in most Lacertians. The malar, wedged anteriorly between the maxillary, palatine, lacrymal and ectopterygoid, curves backward as a slender style terminating in a point, leaving the orbit uncircled by bone behind : the squamosal, wedged behind between the mastoid and tympanic, curves forward to a point beneath the postfrontal. In the American Monitor ( Tejus nicjropunctatus) the nasals are divided: the malar articulates behind with the postorbital- -a dis- memberment of the postfrontal, which continues the zygomatic arch with the squamosal : there is no ' foramen parietale.' In the Chameleon the teeth are short, and so confluent with the jaws that these appear to have simply a serrated margin. The external nostrils perforate the maxillary bone ; a long, compressed, serrated crest arches upward and backward from the superoccipital and parietal bones, and joins the processes of bone continued from the mastoids. In the Chameleo bifurcus the anterior fork-like productions are formed by the maxillary and prefrontal bones. The premaxillary at the bottom of the cleft is very small. In Draco volans there is merely the rudiment of a spine or ridge from the superoccipital ; an arched transverse ridge separates the occipital from the parietal region of the skull. The post- frontal, mastoid, and paroccipital project successively from their respective cranial segments, and well manifest their character as the transverse processes of these. The vacuities in the bony palate are many, and show much variety in the cold-blooded, especially the reptilian, series, in regard to their number, kind, and relative size. The most con- stant are those which are more or less circumscribed by the max- illary and pterygoid, and constitute a pair. They are present in Polypterus and most Ganoids, bounded outwardly by the maxil- lary, medially by the palatine, and behind by the pterygoid. In the Menopome the vomer, fig. 73, /, forms the median and the pterygoid,/, the posterior boundary. In the Frog, fig. 98, A, the pterygo-maxillary vacuities, ?/, are divided from each other by the basispheiioid ; whilst the palatine forms the front boundary and separates them from the nasal apertures, n. In ANATOMY OF VERTEBRATES. 157 Lizards, ib. D, the palatine 20, and pterygoid 24, form the median boundary, the maxillary, 21, and ectopterygoid the outer one oft/. In the Crocodiles, ib. C, the palatine 20 forms the median, the ectopterygoid 25 the outer, the maxillary 21 the fore, and the ptery- goid 24 with the ectopterygoid the hind, boundary. In the Che- Ionia there is no ectopterygoid to divide the pterygo-maxillary vacuity from the lower opening of the temporal fossa. The next openings in point of constancy are the palatal, or posterior, or internal nostrils ( palatonares ; ' but they are variously formed and situated. In the Menopome, fig. 73, there is no palatine bone to divide them from the pterygo-maxillary vacuity ; in fig. 98 A, the Frog, the transverse palatine forms the posterior boundary of the palatonares, n, the vomer the inner, and the maxillary the outer, boundary; they are similarly encompassed in the Lizards, 98 B Frog. Tortoise. Crocodile. Palatal apertures, Eeptilia. Iguana. ib. D, n. In the Crocodiles, the palatonares, ib. C, n, form a single aperture surrounded by the pterygoids, and situated far back. There is also a single premaxillary foramen, ib. C, p, at the fore part of the bony palate. This is sometimes divided into two by the premaxillary, like the external nostrils, as in the Iguana, ib. D, p. In most Lizards there is a more or less elongate ( interpterygoid' vacuity, ib. D, s, bounded behind by the hypapophyses of the basisphenoid, laterally by the pterygoids, and usually extending some way between the palatines. In the Mosasaurus the inter- pterygoid fissure does not extend far back between the pterygoids, but is bounded in a greater proportion by the palatines. Some- times there is a distinct small ( interpalatine ' vacuity, ib. m, in 158 ANATOMY OF VERTEBRATES. advance of the interpterygoid ; and more rarely there occurs an ( intervomerine ' vacuity still more in advance. Thus there are definable and iiameable, in the bony palate of reptiles, the f pterygo-maxillary,' ( palatonarial,' f premaxillary,' 6 interpterygoid,' 'interpalatal,' and ' intervomerine ' vacuities or foramina- - more or less valuable as characters of recent and extinct species. 35. Skull of Ichthyopterygia. Amongst the illustrations of extreme varieties in the reptilian skull which Palaeontology has brought to light, may be cited the Ichthyosaurus, the Dicynodon, and the Pterodactylus. That of the first combines in a peculiar manner some piscine with reptilian characters. It differs from all existing Reptilia in the great size of the premaxillary, fig. 105, 22, and small size of the maxillary, 21 ; in the lateral aspects and antorbital position of the nostrils ; in the immense size of the orbits, and in the large and numerous sclerotic plates, which latter structures give to the skull of the Ichthyosaurus its most striking features. The two supplemental bones of the skull, which have no homo- logues in existing Crocodilians, are the postorbital and super- squamosal ; both, however, are developed in Archecjosaurus and the Labyrinthodonts. The postorbital is the homologue of the inferior division of the postfrontal in those Lacertians e. g., Iguana, Tejus, Ophisaurus, Anguis, in which that bone is said to be divided ; but in Ichthyosaurus it more resembles a dismember- ment of the malar, 26. Its thin obtuse scale-like lower end over- laps and joins by a squamous suture the hind end of the malar : the postorbital expands as it ascends to the middle of the back of the orbit, then gradually contracts to a point as it curves upward and forward, articulating with the supersquamosal and post- frontal, 12. The supersquamosal may be in like manner regarded as a dismemberment of the squamosal, 27 ; were it confluent there- with, the resemblance which the bone would present to the zygo- rnatic and squamosal parts of the mammalian temporal bone would be very close ; save that the squamous part w^ould be removed from the inner to the outer wall of the temporal fossa. The nostril is bounded by the lacrymal, 73, nasal, 15, maxillary, 21, and pre- maxillary, 22, bones. It is distant from the orbit about half its own long diameter. Like the orbit, the plane of its outlet is vertical. The pterygo-maxillary vacuities are very long and narrow, broadest behind, where they are bounded, as in Lizards, by the anterior concavities of the basisphenoid, and gradually narrowing to a point close to the palatine nostrils. These are smaller than ANATOMY OF VERTEBRATES. 159 in most Lizards, and are circumscribed by the palatines, ecto- pterygoid, maxillary, and premaxillary. The pterygomalar fis- sures are the lower outlets of the temporal fossa? ; their sudden posterior breadth, due to the emargination of the pterygoid, relates to the passage of the muscles for attachment to the lower jaw. The parietal foramen is bounded by both parietals and frontals, 11 ; its presence is a mark of labyrinthodont and lacertian affini- ties ; its formation is like that in Iguana and Rhynchocephalus. The occipitoparietal vacuities are larger than in Crocodilia, smaller than in Lacertilia ; they are bounded internally by the basi-, ex-, and super-occipitals, externally by the parietal and mastoid. The auditory apertures are bounded by the tympanic and squa- mosal : the tympanic, 28, takes a greater share in the formation of the ' meatus auditorius' in Lizards ; in Crocodiles the bone 28 is restricted to that which it takes in Ichthyosaurus. 1 In comparing the jaws of the Ichthyosaurus tenuirostris with those of the gangetic Gharrial, an equal degree of strength and of alveolar border for teeth result from two very different propor- tions in which the maxillary and premaxillary bones are combined V / together to form the upper jaw. The prolongation of the snout is the same : the difference of structure relates to the collective tendency of the affinities of the Ichthyosaurus to an antecedent hrcmatocryal type of structure still partly shown by Lizards. The backward or antorbital position of the nostrils, like that in whales, is related to the marine existence of the Ichthyosaurs. But in the Labyrinthodonts, in which the nostrils are nearer the fore part of the head, their anterior boundaries are formed by the premaxillaries, as in modern Lizards : it appears, therefore, to be in conformity with these affinities that the premaxillaries of the Ichthyosaur sjiould enter into the same relation with the nostrils, although this involves an extent of anterior develope- ment proportionate to the length of the jaws, the forward pro- duction of which sharp-toothed instruments fitted the Ichthyosaur, like the modern Dolphin for the prehension of agile fishes. 36. Skull of Dicynodontia. The skull of the Dicynodon^ fig. 99, is articulated with the atlas by a single condyle, formed by the basi- and ex-occipitals in equal proportions : the latter have coalesced, as in the Crocodiles, w T ith the paroccipitals. The parie- tals form one bone, perforated by a small ( foramen parietale ' close to the coronal suture. The frontals, n, contribute a share to the superorbital border ; their median suture is distinct, as is that 1 CXLVII. p. 388. 1GO ANATOMY OF VERTEBRATES. between the nasals, 15. The prefrontal, 14, extends to the nostril, n. The lacrymal, is, forms the rest of the fore part of the orbit, ex- tending forward upon the face. The sides of the premaxillary, 22, bend abruptly down in front of the nostrils, to join the maxillary, 20, 21 ; this forms the lower boundary of the nostril, n, and joins above and behind with the prefrontal, lacrymal, and nasal bones : the maxillary projects below the orbit, like a forward continuation of the zygoma, becomes more prominent as it advances, and soon forms the outer angle of the three-sided socket of the canine tusk, c. There is a single but strong zygomatic arch formed by the malar, 26, and squamosal, 27, abutting against the upper end of the tym- 99 Skull of Dicynodon panic pedicle, 28. The rami of the lower jaw augment in depth from the angle to the symphysis, where they are confluent. The angle projects a very little way beyond the articulation. The articular surface is moderately concave, and looks obliquely up- ward and backward. The elements of the posterior half of the ramus answer to the articular, angular, 26, and surangular, 25. A thin vertical splenial plate, on the inner side of the ramus, begins about an inch in advance of the angle, and extends forward to the symphysis, at the back part of which it appears to become con- fluent with its fellow. The part answering to the angular diverges from the surangular, and forms the hind boundary of an oblong vacuity at the middle of the side of the ramus, the fore part of which vacuity is formed by a bifurcation of the dentary element, 23. This is thickened and strengthened by a ridge, subsiding at the vertical channel upon the side of the symphysis, receiving the ANATOMY OF VERTEBEATES. 1<>1 tusk, s, when the mouth is closed. The symphysis of the man- dible is peculiarly massive broad, high, and thick. Anteriorly it is convex in every direction ; it is bent or produced upward, terminating in a broad trenchant margin, like the fore part of the lower mandible of a macaw. The modification of the back part of the cranium, especially the great expansion due exclusively to the developement of ridges for augmenting the surface of attach- ment of muscles (for the brain of the cold-blooded reptile would need but a small spot of the centre of the occipital plates for its protection), indicates the power that was brought to bear upon the head as the framework in which were strongly fixed the two large tusks. The strength or resistance of the cavities receiving O O * ' the deeply implanted bases of the tusks was increased by the ridges developed from the outer part of their bony wall. Only the Crocodiles now show a like extent of ossification of the occiput, and only the Chelonians the trenchant toothless mandible ; but in both the outer nostril is single and median : the Lizards repeat the divided apertures for respiring air : in Mammals alone do we find a developement of canine tusks like that in the Dicynodonts. 37. Skull of Pterosauria. The skull of the Pferodactyle, fig. Ill, was as remarkable for its light and delicate structure as that of the Dicynodont for its compact massiveness. It had a single occipital condyle : a post-fronto-mastoid arch and a malo- squamosal arch on each side ; the latter abutting against the end of the tympanic pedicle. The orbit was large, and the eyeball defended by sclerotic plates. The external nostrils were divided, and placed about midway between the orbits and the muzzle. There was a large vacuity between the orbits, o, and nostrils, n. The jaws varied much in length in different species. 38. Scapular arch and appendage. Parts which project from the body to act on the surrounding medium commence as a bud or fold of skin, within which is formed the framework, in texture and structure according to the work to be done. The reaction of the medium, whether air, water, or earth, calls for the due resistance usually afforded by junction of the projecting part with a segment of the endoskeleton. Thus, in Fishes, the frame of the opercular flap articulates with the tympano-mandibular arch : that of the branchiostegal (gill-covering) flap with the hremal arch of the parietal vertebra : that of the pectoral flap or limb with the same arch of the occipital vertebra. The frame of the caudal flap or fin is attached to the terminal vertebrae of the body : those of the dorsal and anal fins are less firmly inter- VOL. I. M 162 ANATOMY OF VERTEBRATES. locked with the neural and haemal spines of more advanced vertebra?. All these various supports of flaps, fins, or limbs belong to the same natural genetic group of skeletal parts : their peripheral rays are not f dermal bones ; ' they are developed between folds, not in the substance, of the integument ; although in some instances they press away the skin and become coated by a ganoid conver- sion or calcification of its outer layer. The most simple condition of the parial (pectoral and pelvic) limbs is manifested by the Lepidosiren, fig. 100. A filamentary appendage is sustained by a single many-jointed cartilaginous ray, fig. 101 A, a. In one species there are attached at right angles to the pectoral ray fine filaments sustaining the narrow fold of membrane continued from its posterior side. A similar series of finer rays supports the membrane continued from the dorsal detached dermoneurals of Polyptcrus. Protoptcrus (Lepidosiren} annectens. xxxm. The arch sustaining the pectoral limbs of Lepidosiren is also simple, departing least from its archetypal condition. A long straight cylindrical bone, fig. 101, A, 51, pi, is attached by a short ligamentous mass to the epencephalic arch, ib. n, of which it is the rib, or ' pleurapophysis,' assuming in ulterior developements the special name of ' scapula.' With each scapula is articulated a larger and more flattened bone, ib. 52 : the two converge and meet at their lower ends, .completing, as hremapophyses, a widely expanded hremal arch. The entire segment, A, conforms to the thoracic modification of the Archetype vertebra, fig. 19 ; and, simi- larly, is expanded in order to encompass and protect the heart : but it is simplified by the absence of the hamial spine in Protopterus, as the neural spine is sometimes wanting in a neural arch. The hremapophysis, h, in ascending the vertebrate scale, assumes special forms, signified by the term f coracoid,' with the number 52. In ANATOMY OF VERTEBRATES. 163 JProtopteri, as in more piscine Hcematocrya, the coracoid ex- clusively supports the appendage or limb. From the condition exemplified in fig. 101, A, the developement 101 Elementary limbs, A, c, Lepidosiren ; B, D, Ampliiuma, CXL. of the pectoral member diverges in two directions : one by multi- plication of many-jointed rays, the other by simplification as to number of rays and joints, with special modification and differen- tiation of the latter. 39. Pectoral limb of Fishes. The first series of modifications is now confined to Fishes : but, before describing the appendage, 1 f x* J? xl 1 . a briei notice ot the arch is requisite. In most Osseous Fishes the pleurapophysis of the occipital, like that of the two antecedent cranial vertebras, is in more than one piece ; but the divisions do not exceed two. The upper piece (suprascapula) is commonly bifurcate, as in the Cod, figs. 34, 75, 81, 50, the lower prong answering to the ( head,' the upper one to the ' tubercle ' of the thoracic rib in the Crocodile : the latter articulates with the transverse process (jparoccipitaT). The lower piece (scapulci), ib. 51, is a slender straight bone, pointed below, and mortised into a groove of the coracoid, ib. 52. The two parts of the scapula are confluent in the Siluroid Fishes. In the Murasnoids the suprascapula is ligamentous, and loosely appends the scapular arch to the skull. In the Playiostomi the arch is detached from its vertebra, and has receded in position, to allow, as it seems, for the great expanse of the appended fin. The hrcmapophysis, or 'coracoid,' figs. 34, 38, 39, 75, 85, 52, is longer and usually broader than the scapula. In the Cod-tribe, M 2 164 ANATOMY OF VERTEBRATES. its pointed upper extremity projects behind that bone and almost touches the suprascapula ; a broad angular plate of the coracoid projects backward and gives attachment to the radiated appendage, below which it bends inward and forward, gradually decreasing to a point, which is connected by ligament to its fellow, and to the urohyal bone, fig. 43. The inner side of the coracoid is ex- cavated, and its anterior margin folded inward and backward, lodging the origin of the great lateral muscle of the trunk. In most fishes the lower end of the arch is completed, as in the Cod, by the ligamentous symphysis of the coracoids ; but in the Siluri and Platycephali the coracoids expand below, and are firmly joined together by a dentated suture. In all Fishes they support and defend the heart, and form the frame, or sill, against which the opercular and branchiostegal doors shut in closing the great branchial cavity ; they also give attachment to the aponeurotic diaphragm, dividing the pericardial from the abdominal cavity. To the inner side of the upper end of the coracoid there is attached, in the Cod and Carp, a bony appendage in the form of a single styliform rib ; but in other Fishes this is more frequently composed of two pieces, as in the Perch. This single or double bone, figs. 34, 38, 85, 58, is slightly expanded at its upper end in the Cod-tribe, where it is attached by ligament to the inner side of the angular process of the coracoid : its slender pointed portion extends downward and backward, and terminates freely in the lateral mass of muscles. In the Batrachus its upper extremity rises above the coracoid, and is directly attached to the spinous process of the atlas. In some Fishes, as the Snipe-fish ( Centriscus Scolopax), the Cock-fish (Aryyreiosus Vomer), the Lancet-fish (Sf'yanus), it is joined by the lower end to the corresponding bone of the opposite side, thus completing an independent in- verted arch, behind the scapular one. There is some reason, therefore, for viewing the bone 58 as representing the ha3inal arch of the atlas, or its hasmapophysial portion. The usually free lower extremities of these ha3inapophyses, to- gether with their taking no share in the direct support of the pec- toral fins, and their inconstant existence, oppose the view of their special homology with the coracoids of higher Vertebrates. To that with the ( clavicles' of higher classes it has been objected that these bones are always situated in those classes in advance of the coracoids ; but this inverted position may be a consequence of the backward displacement of the scapula and coracoid in the air-breathing Vertebrates. The appendage of the scapular arch, in most Osseous Fishes, ANATOMY OF VERTEBRATES. 165 is composed of three segments : the first, of two, rarely of three, bones immediately articulated with the coracoid ; the next, of a series of from two to six smaller bones ; which, lastly, support a series of spines or jointed rays. These rays serially repeat the branchiostegal rays in the hyoidean appendage, and the opercular rays in the tympanic appendage. The vegetative repetition of digits and joints, and the vegetative sameness of form in those multiplied peripheral parts of the fins of Fishes, accord with the characters of all other organs on their first introduction into the animal series. The single row of fewer ossicles, figs. 34 and 81, 56, supporting the rays, 57, obviously represents the double carpal series in Mammals ; and the bones of the brachium and antibrachium seem in like manner to be reduced to a single series, 54, 55. In the ventral fin, fig. 34, v, no segment is developed between the arch, 63, and the digital rays, 70 : it is in this respect like the branchiostegal fin, 40, 44. The pectoral fin is directed backward, and being applied, prone, to the lateral surface of the trunk, the ray or digit answer- ing to the thumb is toward the ventral surface. The lowest of o the bones supporting the carpus should, therefore, be regarded as the radius (figs. 34 and 81, 54), holding the position which that bone unquestionably does in the similarly disposed pectoral fin of the Plesiosaur, fig. 45, 54, and Cetacea. The upper bone, which commonly affords support to a smaller proportion of the carpal row, may be compared to the ulna (ib. 55). As a third small bone is articulated to the coracoid, in some Osseous Fishes, at least in their immature state, the name of humerus may be confined to that bone : but in these it is generally above and on the inner side of the ulna, and seems to be rather a dismember- ment of it. In the Salmonida, it is more distinctly developed; it is articulated in the Bull-trout (S. eriox} 1 to the middle of the back part of the coracoid by a transversely elongated extremity ; and is expanded at its distal end, where it articulates by cartilage with the radius and ulna. In the Cod, Haddock, and most other Fishes there is no separate representative of the humerus : in these the ulna is a short and broad plate of bone, deeply emargi- nate anteriorly, attached by suture to the coracoid, and by the opposite expanded end to the radius, and to one or two of the carpal ossicles, and directly to the upper or ulnar ray of the fin. In the Bull-head and Sea-scorpion (Cottus), the radius and ulna are widely separated, and two of the large square carpal 1 XLIV. p. 18, No. 46. 166 ANATOMY OF VERTEBRATES. plates in their interspace articulate directly with the coracoid. A similar arrangement obtains in the Gurnards and the Wolf-fish ; but the carpals in the interspace of the radius and ulna are sepa- rated from the coracoids by a space occupied by clear cartilage ; and in the Wolf-fish the intermediate carpals are almost divided by two opposite notches. The ulna is perforated in all these fishes. The radius is of enormous size in the Opah (Lampris), the Cock-fish, fig. 38, and the Flying -fish ; it is anchylosed with the coracoid in the Silurus, to give firmer support to the strong serrated pectoral spine. Both radius and ulna are connate with the coracoid in the Angler (Lophius, fig. 102, 54, 55). The ossicles called carpals are usually four or five in number, 102 Coracoid aiid bones of pectoral fin, Angler (Lophiits) as in the Cod tribe, fig. 81, 56; they progressively increase in length from the ulnar to the radial side of the carpus, especially in the Parrot-fish (Scarus^) and the Mullets (MugiT). They are three in number and elongated in the Polypterus, fig. 103, 56, but are reduced to two in number, and more elongated in the Lophius, fig. 102, se) ; thus they retain in this species and in the Sharks, fig. 104, their primitive form of ( rays ;' but change to broad flat bones in the Wolf-fish, just as the rays of the opercular fin exchange that form in the Plagiostomes for broad and flat plates in ordinary Osseous Fishes. The rays representing the metacarpal and phalangial bones are, in the Cod, twenty in number, and all soft, jointed, and sometimes bifurcate at the distal end. Their proximal ends are slightly expanded and overlap each other, but are so articulated as to permit an oblique divarication of the rays to the extent permitted by the uniting fin-membrane, the combined effect being a move- ment of the fin, like that called the ' feathering of an oar.' Each ANATOMY OF VERTEBRATES. 167 soft and jointed ray splits easily into two halves as far as its base, and appears to be essentially a conjoined pair. In the series of Osseous Fishes the rays of the pectoral and ventral fins offer the same modifications as those of the median fins, on which have been founded the division into ' Malacoptery- 6 gians ' and ( Acanthopterygians : ' in the former, the last or ulnar fin-ray, is usually thicker than the rest ; in the latter it is always a hard, unjointed spine : in some Fishes it forms a strong pointed or serrated weapon (Silurus). In the Gurnards, fig. 82, the three lowest rays are detached and free, like true fingers ; and are soft, multi-articulated, and larger than the rest ; they are supplied by special nerves, which come from the peculiar ganglionic enlarge- ments of the spinal chord, and are organs of exploration and of subaqueous reptation. 1 In all the Gurnards the natatory part of the pectoral member is of large size ; but in one species (Dactylo- pterus) it presents an unusual expanse, and is able by its stroke to raise and sustain for a brief period the body of the fish in the air. The pectoral fins present a still greater developement in the true Flying-fish (Exoccetus). In some Malacopteri and Ganoidei a segment analogous to a metacarpus may be distinguished by modification of structure from the phalangeal portion of the fin rays : in the Polypterus there are seventeen simple cylindrical metacarpal bones, fig. 103, 57, the middle ones being the longest : they sustain thirty -five digital rays, and are supported by 103 carpal bones, ib. 103, 56, of which two are almost as remarkable for _< 41 their length as in the Lophius ; the third, shorter and broader, is wedged into the interspace of the two longer ones, but does not directly join the Bones of P ectoral metacarpus. The carpus is supported by a small radius, 55, and ulna, 54, which articulate directly with the coracoid. A further approach to the higher conditions of the pectoral member is made by the same Fish in the carpal portion projecting freely from the side of the body, as in the Lophioid Fishes. In the Salmon, where eleven such metacarpals support thirteen or fourteen fin- rays, the carpus is short and consists of four bones. In the Plagiostomes the scapular arch is detached from the oc- ciput, the conditions of its displacement being the more varied and vigorous use, or the enormous expanse, of the pectoral fin ; per- 1 CLIX. p. 46. 168 ANATOMY OF VERTEBRATES. haps, also, the more posterior position of the heart in these Fishes. In the Sharks and Chimaerae the arch is loosely suspended by ligaments from the vertebral column : in the Rays the point of re- sistance of their enormous pectoral fins has a firmer, but somewhat anomalous attachment, by the medium of the coalesced upper ends of the suprascapular pieces to the summits of the spines of the confluent anterior portion of the thoracic abdominal vertebra?. In the Sharks the scapular arch consists chiefly of the coracoid por- tions, fig. 104, 52, which are confluent together beneath the peri- cardium which they support and defend ; the scapular ends of the arch, connected to the coracoids by ligament, project freely upward, backward, and outward. To a posterior prominence of the cora- coid cartilage corresponding with the anchylosed radius and ulna, ib. 54, 55, in the Lophius, there are attached, in the Dog-fish and most other Sharks, three sub- compressed,, sub-elongated carpal 104 Cartilages of the pectoral fin and arch of the Dog-fish (Spinax acanthias) cartilages, the uppermost, ib. 56, the smallest, and styliform ; it supports the upper or outer phalangeal ray. The next bone, ib. se', is the largest and triangular, attached by its apex to the arch, and supporting by its base the majority of the phalanges. The third carpal, ib. 56", is a smaller but triangular cartilage, and supports six of the lower or radial phalanges. Three joints (metacarpal and digital) complete each cartilaginous ray or representative of the finger, ib. 57 ; and into the outer surface of the last are inserted the fine horny rays or filaments, ib. 57 /x , the homologues of the claws and nails of higher Vertebrata, but which on their first appearance, in the present highly organised class of Fishes, mani- ANATOMY OF VERTEBRATES. 169 fest, like other newly introduced organs, the principle of vegetative repetition, there being three or four horny filaments to each carti- laginous ungual phalanx. On the fore part of the coracoid arch, near to the prominence supporting the fin, there are developed a vertical series of small bony cylindrical nuclei in the substance of the cartilage in most Sharks. In the Rays the coraco-scapular arch forms an entire circle or girdle attached to the dorsal spines : it consists of one continuous cartilage in the Rhinobates, but in other Rays is divided into coracoid, scapular, and suprascapular portions, the latter united together by ligament. The scapula and coracoid expand at their outer ends, where they join each other by three points, to each of which a cartilage is articulated homologous to the three above described in the Shark, and which immediately sustain the fin-rays. The posterior cartilage answering to the upper one in the Shark curves backward and reaches the ventral fin : the an- terior cartilage curves forward, and its extremity is joined by the antorbital process as it proceeds to be attached to the end of the rostral cartilage ; the middle proximal cartilage is comparatively short and crescentic, and sustains about a sixth part of the fin-rays, which are the longest, the rest being supported by the anterior and posterior carpals, and gradually diminishing in length as they approach the ends of those cartilages. Developement by irrelative repetition of parts reaches a maximum in the present plagiostomous group. In the common Ray, fig. 64, there are upwards of a hundred many-jointed fingers in each pectoral limb : but all are bound up in a common function of the simplest kind. 40. Pectoral limb of Reptiles. The other route of develope- ment from the prototypal condition exemplified in fig. 101, A and C, leads to a differentiation of the several divisions and parts of the limb, and their adaptation to particular functions or parts of com- bined and varied mechanical actions. The first step, as manifested in the Amphimne, ib. B, c, is the formation of a Ions; inflexible segment, as a lever of greater resist- o o o ance, 53 and 65 ; this is followed by a pair of similar, but shorter cylindrical bones, each sustaining a ray of few joints. The proximal bone assumes through ulterior developements the special name f humerus,' or arm-bone, with the symbol 53, in the fore limb; and of 'femur,' or thigh-bone, with the symbol 65, in the hind limb. The two bones of the next segment become, in the fore limb, f radius,' 54, and ' ulna,' 55 collectively, antibrachium or 6 fore-arm;' in the hind limb, tibia, 66, fibula, 67 collectively, 170 ANATOMY OF VERTEBRATES. 105 the cnemion or leg. The mass of fibro-cartilage, in which more or fewer ossicles are subsequently developed, interposed between the antibrachium and terminal rays, is the 6 carpus,' 56 : the corre- sponding mass in the hind limb is the tarsus, 68. The terminal rays are the digits, called f hand,' and ( fingers,' 69, in the fore limb ; ( foot ' and ' toes ' in the hind limb. The proximal joints of these rays, being bound together in a sheath of integument, are differentiated as c metacarpals ' in the hand, and ( meta- tarsals ' in the foot. The other joints are the ( phalanges,' ultimately distin- guished as ( proximal,' ( middle,' ( distal ' or ' ungual,' as usually supporting a claw or nail. In the extinct Ganocephala, and in the few surviving ichthyomorphous or per- ennibranchiate Batrachia, the simple type of limb, as in fig. 101, B, is re- tained; only that the digital rays in- crease in number from the tf two ' in Amphiuma, to ( three ' in Proteus, and to f four ' in Menopoma, fig. 43, 57, and Axolotes. In the extinct Ichthyopterygia the digits may be seven, eight, or nine in number, and consist of numerous short joints a significant mark of piscine affinity : they are bound together, but converge towards a point : the joints are of a flat- tened angular form, and interlock with a ' those of the contiguous digit, the whole forming a continuous, broad, slightly flexible basis of support to the fin. The essential distinction from the fin of the fish is shown by the well developed 6 humerus,' 53, and by the complex sca- pular arch. The two antibrachial bones retain the piscine shortness and breadth ; skeleton of ichthyosaurus, with and the metacarpal series is less distinctly cast of spiral intestine. CLXIII. defined than ill SOlllC fishes. ANATOMY OF VERTEBRATES. 171 The scapula, 51, is short and straight, displaced backward from the occiput, and contributing to form the shoulder-joint, as in the Batrachia and higher air-breathers : but it shows a certain breadth and flatness. The coracoid, 52, is still broader, not cartilaginous as in most perennibranchs, but well ossified, and united below with its fellow, and with a small ' episternum ' of a triradiate form, one ray of which is wedged into the fore part of the intercoracoid fissure. There is also a pair of bones, so, long and slender, articulated with the fore border of the scapula and the transverse rays of the episternum : they are the clavicles. A supplementary flattened bone, the ' epicoracoid,' is wedged between the scapula, clavicle, and coracoid. The above complex and powerful scapular arch would enable the fore-paddles to act upon the land with sufficient power to effect a shuffling forward move- ment of the body, as in the Turtle ( Chelone) and Seal tribe : but the main office of the fore-limb in the Ichthyosaur was that of a pectoral fin. In the Plesiosaurus, fig. 45, the limbs acquired a developement more closely accordant with that in Chelone. The scapula, 51, developes an acromial process representing the clavicle. The coracoid, 52, is unusually extended in the trunk's axis, and is united with its fellow by a long symphysis interposed between the an- terior abdominal rib and the episternum ; it articulates at its fore part with the episternum and clavicular process, and, further back, with the lower end of the scapula to form the humeral joint. The humerus is proportionally longer than in Ichthyosaurus ; the radius is better developed, and slightly expanded at both ends ; the ulna retains a flattened reniform shape. The carpal series is distinct, in a double row of ossicles, the largest at the radial side of the wrist, the opposite side retaining more unossified material. The digits are five in number, with the proximal and more elongated joints representing a metacarpus. The phalanges are shorter, and decrease in size to the tips of the digits, which converge. The first or radial digit has generally 3 phalanges, the second from 5 to 7, the third 8 or 9, the fourth 8, the fifth 5 or 6 : all are flattened and included in a common sheath of integument like those of the Turtle ; but the paddle had no claws. The scapular arch retains the same essential simplicity in the Chelonian as in the Sauropterygian order, only the acromial or clavicular process is relatively longer, more like a collar bone ; it extends from near the articular part of the scapula toward the median line, in advance of the coracoid, fig. 51, O, with the medial end ligamentously attached to the episternal. In the 172 ANATOMY OF VERTEBRATES. 106 Tortoise (Testudo) it is shorter, in Chelys, fig. 106, b, it is longer than the scapula, a. This bone in all Chelonians is a strong, straight columnar one, with the upper end connected by ligament with the inner surface of the first costal plate, fig. 51, N; it descends almost vertically to the shoulder-joint, of which it forms, in common with the coracoid, the 'glenoid' cavity, fig. 106, //. The coracoid, suturally united at that end with the scapula, passes inward and backward, fig. 51, o, expanding and becoming flattened at its median end, which does not meet its fellow nor articulate with the sternum. The coracoid is broad and short in the Tortoise ; long and slender in Chelone and Emys, fig. 51, o, of intermediate proportions in Trionyx and Chelys, fig. 106, c. The scapular arch and proximal part of the limb being included in the thoracic abdominal box, the humerus is peculiarly bent and twisted in the terres- trial species in order to emerge from the front fissure, and plant the foot on the ground, fig. 51, p. In the Tortoise the ordinary position of the fore-limb is that of extreme pronation, with the olecranon forward and outward, and the radial side of the hand downward. The capsule of the shoulder-joint includes a consider- able part of the neck of the humerus. The hemispheroid 'head projects unusually from the back part of the bone, which looks upward : the tuberosities are large and bent toward the palmar aspect : that which is internal in most animals is here ' postero-superior ; ' that called ' external ' is 6 postero-internal ' in position ; from the former is continued the ' deltoid crest.' The distal end is expanded and rather flattened from before backward. In the Turtle the humeral shaft and its lower end is compressed laterally : and the bone is almost straight in those marine species ; in all Chelonia it is solid throughout. The ulna is shorter, and in the Turtle, fig. 107, b, the olecranon is less developed than in the Tortoise, fig. 108, by 55. The contrast between the marchers and the swimmers is most striking in the proportions of the toes. In the Turtle, fig. 107, the pollex, /, is short and has two phalanges after the Scapular arch, Chehjs. CLI. ANATOMY OF VERTEBRATES. 173 107 metacarpal : the last phalanx supporting a claw. The three middle digits, it, Hi, iv, have each three long phalanges, the last being flattened and without a claw ; the fifth has two pha- langes. All these are connected together by a web. In the Tortoise, fig. 108, all the toes are very short and subequal ; and each has one metacarpal and two phalanges, the last supporting a claw ; the few species in which the fifth has but one phalanx and no claw form the genus Homopus, Dum. and Bib. In Emys europ&a, fig. 51, T, u, the first and fifth digits have each a metacarpal and two phalanges ; the others have three phalanges ; the last bears a claw in each digit. In the Soft or Mud-turtles. C5 the pollex has two phalanges, the second with a claw ; the three middle digits have each three phalanges, but only the index and meclius have the claw; the fifth digit has two o phalanges and no claw, whence the generic name Triomjx, proposed for these frequenters of the muddy estuary. In the Crocodilia the scapular arch consists of a simple scapula, fig. 57, 51, and coracoid, ib. 52, and fig. 54, 8 : these compressed, narrow, mode- rately long plates of bone, are thickest where they are united to- gether to form the glenoid cavity for the humerus. In each, the bone contracts beyond the articular ex- pansion, becomes sub-cylindrical, but soon again flattens and expands to its opposite end ; that of the sca- pula is free, that of the coracoid joins the lateral border of the ster- num. There is no trace of clavicle, no acromial projection from the sca- pula. The humerus, fig. 51, 53, presents two curves : the articular head is a transversely elongated, sub-oval convexity ; it is continued upon the short, obtuse, angular prominence, answering to the inner or ulnar tuberosity. The radial crest begins to project from the shaft at some distance from the head of the bone. There is a longitudinal ridge on the anconal surface close to the radial border. Bones of fore-arm and paddle, Chelone. CLI. 174 ANATOMY OF VERTEBRATES. 108 I II III IV V Laud Tortoise. CLI. The distal end is transversely extended and divided anteriorly into two condyles. The shaft has a medullary cavity smaller than in land lizards. The radius, fig. 57) t, fig. 109, Z>, 54, has an oval head, an almost cylindrical and straight shaft, with an oblong and subcompressed distal end. The ulna, fig. 57, s, fig. 109, #, 55, articulates with the outer condyle of the humerus by an oval facet, the thick convex border of which swells out be- hind like the beginning of an ' olecranon ; ' the O O shaft of the ulna is compressed transversely and curves slightly outward ; the distal end is less than the proximal one, and articulates with the second and third bones of the carpus. The first metacarpal supports two phalanges, I, the second three, n, the third and fourth, each four, the fifth, V, three phalanges which 109 are very slender ; but the proportions are shown in the cut ; only the toes, I, n, and in, have the claw. All are basally united by a short web, but the fore-foot is chiefly used in movements upon land. In the Monitor ( Varanus niloticus) the supra- scapula is a broad semiossified plate : the scapula is short and broad, and appears to have coalesced with the coracoid. This bone is much expanded, and has two deep notches anteriorly, and a perforation near the humeral articulation. In some Lizards it sends for- ward an acromial process. The coracoid is shorter and broader than in the Crocodile, abuts against the upper margin of the rhorn- boidal sternum, and sends off two processes from its anterior border, the one next the sca- pula abutting against the transverse branch of the episternum ; the other against the sub- ossified epicoracoid : this element overlaps that of the opposite side. In the Monitor, as in most Lizards, there are distinct clavicles : usu- ally long and slender bones, with more or less expanded extremities, extending from the body of the episternum and accompanying the trans- verse branch to abut against the scapula ; and sometimes also reaching the outer process of the coracoid. In Lacerta, Cuv. Bonos of fore-arm and foot, Crocodile. CLI. ANATOMY OF VERTEBRATE S. 175 110 and Scincus, the clavicle expands at its medial half, which has a large vacuity or perforation occupied by membrane. In the Chameleon the scapular arch is as simple as in the Crocodile, but the coracoid is shorter and broader. The humerus in Lacertians is usually larger and straighter, fig. 50, Draco volans, than in the Crocodiles, with a more compact wall and wider medullary cavity. The radius, ib. and fig. 110, b, 54, is almost straight, and slender, with an oval proximal articular concavity, and a distal surface partly convex, partly concave. The ulna, fig. 110, a, 55, shows the olecranon better developed than in the Crocodile : its dis- tal articular surface is convex. The dibits are five in number, o * the phalanges are 2, 3, 4, 5 and 3, counting from the metacarpal of the first to that of the fifth digit : each has a claw supported on a moderately long, compressed, curved, and pointed phalanx. The Chameleon offers an exception to the numerical rule, the phalanges being 2, 3, 4, 4, 3 ; and the direction of the digits modified for the scansorial function in these arboreal Lacertians : I, n, and in, enveloped by the skin as far as the claws, are directed forward ; iv, and v, similarly sheathed, are directed backward : and the joints are shorter and broader than in Land-lizards, The fore limbs in Draco volans accord with the usual lacertian type, and take no share in the support of the parachute. But in the extinct order of truly volant Reptiles (Pterosauria) they were modified for the exclusive support and service of the wings. The scapula, fig. Ill, 51, long, narrow, flattened, and slightly expanded, lay more parallel with the spine than in land and sea Reptiles. The coracoid, strong and straight, and combining, as usual, with the scapula to form the glenoid cavity, articulated at the opposite end with a groove at the fore-part of a discoid sternum, which part is produced and keeled. The humerus, ib. Bonds of fore-arm and foot, Chameleon. CLI. 176 ANATOMY OF VERTEBRATES. 53, is more expanded at its proximal end than in the Crocodile or Lizard ; the inner (ulnar) tuberosity is more prominent, the radial crest much more developed : with a base coextensive with one fifth of the shaft of the bone, it extends in a greater proportion from the shaft, affording a powerful lever to the muscles inserted into it. The articular head is reniform. The shaft is cylindrical : ill Skeleton of Pterodactylus crassirostris. A. Restoration of Pterodactyle. CLXXX. the walls thin and compact, the cavity large, and was filled with air as in birds of flio-ht. 1 o The e pneumatic foramen,' or that by which the air passed from a contiguous air-cell into the bone, is situated on the fore (palmar) side, a little below the radial end of the head of the bone. The radius, ib. 54, and ulna, ib. 55, are very long, straight, and closely connected together. The digits show the lacertian number of 1 CXLIX. p. 16. CLXVI. p. 451. This discovery breaks down the following distinction : ' Au rcste, on distingue toujours 1'humerus d'un lezard do celui d'un oiseau, parceque le premier n'est pas creux ni percc de trous pour Fentree de Fair dans son intc-ricur.' CLI. v. pt. 2, p. 296. ANATOMY OF VERTEBRATES. 177 phalanges from the first to the fourth, and slightly increase in length: each terminated by a deep compressed,, curved and pointed ungual phalanx. The modification converting the limb into a wing is confined to and concentrated upon the fifth digit, ib. 5 : its metacarpal presents almost the thickness of an antibrachial bone: the proximal phalanx, of equal thickness, has more than twice the length, and at the proximal joint shows a process like an olecranon. This is usually followed, as in Pterodactylus crassirostris, by three similarly elongated phalanges, of which the last gradually tapers to a point. The fore limb thus exceeds in length the whole body, and is presumed to have supported a membranous wing, as in the sketch A, fig. 111. Such are the chief modifications by which the fore-limb, in the Reptilian series of cold-blooded air-breathers is or has been adapted for aquatic, amphibious, terrestrial, arboreal, and aerial life. Before, however, quitting this subject, it may facilitate the comprehension of the homologies of the carpal series of ossicles, by concluding with a separate and serial review of them in the Reptilian group. In the Toad (Bufo) the carpus includes eight bones : the two principal are the ' lunare,' fig. 44, c, /, and ( cuneiforme,' ib. c, respectively articulating with the radial and ulnar divisions of the antibrachial bone, ib. 54, 55 ; the scaphoid, c, s, presents its ( inter- medial' position between the lunare and the four ossicles on the radial side of the distal series : these consist of the trapezium t, trapezoides tr, magnum m, and the divisions of the unciform u for the fourth and fifth digits; that for the fifth being the largest of the five bones. The thumb, I, is represented by its metacarpal only ; the index, fig. 44, A, n, and medius, in, have each a metacarpal with two phalanges ; the digits IV and v have each three phalanges. In the Tortoise ( Testudo, fig. 108), the antibrachium articulates with three carpals forming the proximal row ; the first or radial bone, ib. , answers to the ( scaphoid ' with the ' intermedium ' e ; the second ib. c, to the lunare ; the third, ib. d, to the cunei- forme ; the lunare being interposed between the ends of the radius and ulna. In the Emys, fig. 5 1 , s, the carpus has a similar struc- ture ; but in some species there is a distinct pisiforme. In the Turtle ( Chelone), the scaphoid is reduced in size, and represents only the intermedium, fig. 107,