Amphibia

AMPHIBIA

LINNAeUS originally employed this term to denote a class of the Animal Kingdom comprising crocodiles, lizards and salamanders, snakes and Coecilioe, tortoises and turtles, and frogs; to which, in the later editions of the Systema, Naturoe, he added some groups of fishes. In the Tableau Elémentaire, published in 1795, Cuvier adopts Linnaeus’s term in its earlier sense, but uses the French word “Reptiles,” already brought into use by Brisson, as the equivalent of Amphibia. In addition, Cuvier accepts the Linnaean subdivisions of Amphibia-Reptilia for the tortoises, lizards (including crocodiles), salamanders, and frogs; and Amphibia-Serpentes for the snakes, apodal lizards, and Coecilioe.

In 1799 Brongniart pointed out the wide differences which separate the frogs and salamanders (which he terms Batrachia) from the other reptiles; and in 1804, Latreille, rightly estimating the value of these differences, though he was not an original worker in the field of vertebrate zoology, proposed to separate Brongniart’s Batrachia from the class of Reptilia proper, as a group of equal value, for which he retained the Linnaean name of Amphibia.

Cuvier went no further than Brongniart, and, in the Régne Animal, he dropped the term Amphibia, and substituted Reptilia for it. Meckel, on the other hand, while equally accepting Brongniart’s classification, retained the term Amphibia in its earlier Linnaean sense; and his example has been generally followed by German writers; as, for instance, by Stannius, in that remarkable monument of accurate and extensive research, the Handbuch der Zootomie (Zweite Auflage, 1856).

In 1816, De Blainville, adopting Latreille’s view, divided the Linnaean Amphibia into Squamiféres and Nudipelliféres, or Amphibiens; though he offered an alternative arrangement, in which the class Reptiles is preserved and divided into two sub-classes, the Ornithoides and the Ickthyoides. The latter are Brongniart’s Batrachia, plus the Coecilioe, whose true affinities had, in the meanwhile, been shown by Duméril; and, in this arrangement, the name Amphibiens is restricted to Proteus and Siren.

Merrem’s Pholidota and Batrachia (1820), Leuckart’s Monopmoa and Dipnoa (1821), Miiller’s Squamata and Nuda (1832), are merely new names for De Blainville’s Ornithoides and Ichthyoides, though Müller gave far better anatomical characters of the two groups than had previously been put forward. Moreover, following the indications already given by Von Bär in 1828, Müller calls the attention of naturalists to the important fact, that while all the Squamata possess an amnion and an allantois, these structures are absent in the embryos of all the Nuda.

Müller makes an appeal for observations on the development of the Coecilioe, and of those Amphibia which retain gills or gill-clefts throughout life, which has unfortunately yielded no fruits from that time to this.

In 1825, Latreille published a new classification of the Vertebrata ,6 which are primarily divided into Hoematherma, containing the three classes of Mammifera, Monotremata, and Aves ; and Hoemacryma, also containing three classes — Reptilia, Amphibia, and Pisces. This division of the Vertebrata into hot and cold blooded is a curiously retrograde step, only intelligible when we reflect that the excellent entomologist had no real comprehension of vertebrate morphology; but he makes some atonement for the blunder by steadily upholding the class distinctness of the Amphibia. In this he was followed by Dr J. E. Gray; but Duméril and Bibron in their great work, and Dr Günther in his Catalogue, in substance, adopt Brongniart’s arrangement, the Batrachia being simply one of the four orders of the class Reptilia. Professor Huxley has adopted Latreille’s view of the distinctness of the Amphibia, as a class of the Vertebrata, co-ordinate with the Mammalia, Aves, Reptilia, and Pisces; and the same arrangement is accepted by Gegenbaur and Haeckel. In the Hunterian lectures delivered at the Royal College of Surgeons in 1863, Professor Huxley divided the Vertebrata into Mammals, Sauroids, and Ichthyoids, the latter division containing the Amphibia and Pisces. Subsequently he proposed the names of Sauropsida and Ichthy-opsida tor the Sauroids and Ichthyoids respectively. It is proper to mention, finally, that Professor Owen, in his work on The Anatomy of Vertebrates, follows Latreille in dividing the Vertebrata into Hoematotherma and Hoemar tocrya, and adopts Leuckart’s term of Dipnoa for the Amphibia.

The Amphibia are distinguished from the Sauropsida and Mammalia by very important and sharply-defined characters. The visceral arches of the embryo develop gills, which temporarily, or permanently, perform the respiratory function. There is no trace of an amnion, and it is still a question whether the urinary bladder, which all Amphibia possess, answers to the allantois of the higher Vertebrata or not. At any rate, it plays no part in the respiration of the embryo, nor is it an organ by which nutriment is obtained from the parent. There are two occipital condyles, and the basi-occipital region of the skull is either very incompletely, or not at all, ossified. There is no basi-sphenoidal ossification. When young, the Amphibia are provided with, at fewest, three, and usually four, cartilaginous, or more or less ossified, branchial arches. From Pisces, on the other hand, they are distinguishable only by the characters of their locomotive apparatus. When they possess median fins and limbs, these never present fin-rays; and the limbs exhibit, in full development, the type of structure which obtains among the Sauropsida and Mammalia, and differ very widely from the fins of any fish at present known. This difference obtains even among the long extinct Amphibia of the Carboniferous epoch. In other respects, the lower Amphibia approach the Chimoeroe, the Ganoidei, and the Dipnoi very closely; while, in their development, they present curious approximations to the Marsipobranchii.

With respect to the primary subdivisions, or orders, of the class Amphibia, no one can doubt the propriety of the separation of the recent forms into what may be broadly termed Newts (Urodela) ; Frogs and Toads (Anura); and Concilia (Peromela) effected by Duméril; while all that is known of the organisation of the extinct Amphibia of the newer Palaeozoic, and older Mesozoic, formations tends to show that they form a fourth natural assemblage of equal value to each of the others.

The names of Urodela and Anura, given to the first two of these divisions, are undoubtedly open to criticism; but if well-understood terms, which have acquired a definite scientific connotation, are to be changed whenever advancing knowledge renders them etymologically inappropriate, the nomenclature of taxonomy will before long become hopelessly burdened; and, to set a good example, the names of Urodela, Anura, Peromela, and Labyrintho-donta are adopted here for the four orders of the Amphibia, even although it be true that the Labyrinthodonta do not all possess the dental structure on which the name was founded; though there is reason to believe that some Labyrinthodonts were devoid of limbs, or peromelous; that the Anura are not more tail-less than are the Peromela; and that the tails of the Urodela are not more conspicuous than were those of the Labyrinthodonts.

The Urodela, are Amphibia with elongated bodies and relatively short limbs, devoid of scales or pectoral plates, with numerous prae-caudal vertebrae, and with amphicoelous, or opisthocoelous, vertebral centra. The hyoidean arch remains connected with the suspensorium throughout life, and its cornua are large in proportion to its body. The mandible is dentigerous. There are one or two pairs of limbs, the pectoral arch and limbs being always present. The manus never possesses more than four digits. The bones of the antebrachium and of the crus remain distinct, and the tarsus is not elongated. So far as the spermatozoa are known, they are elongated filaments with a vibratile fringe. The larva develops external gills only; and, except Siren, none are known to possess, at any time, a homy masticatory apparatus.

The Anura have relatively short and broad bodies, and both pairs of limbs are constantly present, the hinder being the longer and stronger. There are no scales, nor pectoral plates, but ossification sometimes occurs in the dorsal integument. The vertebrae vary in character, but are usually procoelous. The prae-sacral vertebrae never exceed nine in number, and the caudal portion of the vertebral column is represented by a peculiar styliform coccyx. The hyoidean arch detaches itself from the suspensorium, and almost always becomes connected with the pro-otic region of the skull. The cornua are usually slender, as compared with the broad body of the hyoid. The mandible is almost always devoid of teeth. The bones of the antebrachium and of the crus early ankylose, and the astragalus and calcaneum are much elongated. The manus has a rudimentary fifth digit. Except in Bombinator, the spermatozoa have, flagelliform appendages, like those of ordinary Vertebrata. The larvae develop first external, and afterwards internal, gills, and, so far as is known at present, are provided with deciduous horny masticatory plates. The gill apertures are closed by the growing over them of an opercular membrane.

The Peromela have snake-like bodies, totally devoid of limbs and limb arches. In most, the integument is provided with transverse rows of imbedded cycloid scales, but there are no pectoral plates. The vertebrae of the trunk are very numerous, and are amphicoelous; those of the caudal region are very few, and are free. The hyoidean arch is attached neither to the suspensorium, nor to the skull; its cornua are very slender, and no distinct body is developed; it is followed by several slender, hooplike, branchial arches. The mandible is dentigerous. Nothing is known of the early stages of development; but Müller discovered branchial clefts, with rudimentary branchial filaments, in young Coedlioe.

The Labyrinthodonta for the most part resembled the Urodela in the proportions of the tail and limbs to the body, but some (as Ophiderpeton) were serpentiform, and apparently apodal; no raniform Labyrinthodonts have yet been discovered. The vertebrae are amphicoelous. The mandible is dentigerous. The bones of the antebrachium and crus remain distinct, and the tarsus is not elongated. The manus and pes appear to have been pentadactyle. Three sculptured pectoral plates and a peculiar dermal armour of small scales, confined to the ventral face of the body, are present in many genera. Nothing is known of the early stages of development, but the γG ∖ mgArchegosauria appear to have possessed ossified branchial arches.

In giving a sketch of the organisation of the Amphibia, it will be necessary to enter much more fully into the characters of the skeleton than into those of the other systems of organs. ’

The Vertebral Column, Bibs, and Sternum.— Leaving the extinct Archegosauria aside for the present, all the Amphibia possess well-ossified vertebrae, the arches of which, in the adult condition (except, perhaps, in some Labyrinthodonts), are not separated by a neuro-central suture from the centra. The latter may be amphicoelous, as in the lower Urodela, the Peromela, and the Labyrinthodonta; or opisthocoelous, as in the higher Urodela and some Anura (e.g., Pipa and Bombinator) ; or procoelous, as in the majority of the Anura (with the exception of the eighth vertebra, which is usually amphicoelous; and of the ninth, which commonly has one convexity in front and two behind). In all the recent forms which have been examined, the centra and intervertebral masses contain more or less distinct remains of the notochord. The arches of the trunk vertebrae are connected by zygapophyses; the spinous processes are usually low, but attain a great relative length in the caudal region of some of the Labyrinthodonts (e.g. Urocordylus). Transverse processes are present in all the trunk vertebrae, except the atlas; they are shortest in the Peromela, longest in the Anura. In most Urodela, these transverse processes, at any rate in the anterior trunk vertebrae, are divided into two portions, a dorsal and a ventral, which diverge towards their free ends; or, more strictly speaking, these processes are made up of two subequal transverse processes, a dorsal “tubercular” process, and a ventral “capitular” process. Sometimes this division prevails throughout the whole length of the trunk, but, more commonly, the two transverse processes become fused into one, posteriorly.

In the long-bodied Urodela (Siren, Proteus, Amphiuma), only a small number of the vertebrae which succeed the atlas present traces of double transverse processes; further back, the coalesced transverse processes form trihedral projections, their dorsal and ventral contours converging instead of diverging, and giving a very characteristic aspect to these vertebrae.

In some Labyrinthodonts, the capitular and tubercular processes, divergent and subequal in some (probably the anterior) vertebrae, coalesce into one in other vertebrae; and the capitular division being shorter than the tubercular, transverse processes, like those of the middle thoracic region of the crocodiles, are thus produced.

In the Peromela, there is a short capitular process, but the tubercular process is represented by a mere facet placed below the prezygapophysis.

In the Anura, finally, the vertebrae have only a single transverse process (possibly representing the coalesced double transverse processes of the posterior trunk vertebrae of such Urodela as Menopoma), which, in some of the vertebrae, may attain a great length.

Bibs are present in a few Anura, in all Urodela, Peromela, and Labyrinthodonta; and, in the last-named and some Urodela, they attain as great relative dimensions as in other Vertebrata. But they are always vertebral ribs, no Amphibian being known to possess more than rudiments of sternal ribs. The atlas is never provided with ribs. In the Peromela, ribs are borne by all the other vertebrae, except the very hindmost. In the long-bodied Urodela, on the other hand, they are restricted to a few of the anterior vertebrae. In the other Urodela they are usually confined to the prse-sacral and sacral vertebrae; but, in some cases, one or two of the anterior caudal vertebrae have free ribs. The form of the proximal end of the rib corresponds with that of the transverse processes or process. Where this is double, the rib presents a fork, formed by the capitulum and tuberculum; and when the capitular and tubercular transverse processes are of equal length, the capitula and tubercula of the ribs are equal; but when either of the former is shorter than the other, the corresponding part of the rib is longer. The Peromela have no sternum, and that of the Labyrinthodonta (if they possessed any) is unknown. In the Urodela, the sternum never ossifies, and there is no trace of even a cartilaginous sternum in Proteus. In Menobranchus, there is a very small cartilaginous sternal plate, which sends lateral prolongations into two of the intermuscular ligaments, representing rudimentary sternal ribs.

In the Newts the sternum becomes a broad and stout plate of cartilage, with a median, posterior, cristate, xiphoid process, and with articular surfaces on its antero-lateral margins for the reception of the coracoids. The sternum attains its highest development in the raniform Anura, the xiphoid process becoming elongated and dilated at its extremity, and more or less converted into bone, while calcification of the body of the sternum itself may also occur.

In the Urodela, the first vertebra always presents two slightly concave articular facets, the faces of which look outwards and forwards to adapt themselves to the occipital condyles. Between these facets, the dorsal moiety of the anterior face of the centrum gives rise to a process, which is little more than a ridge in Proteus, but in other genera becomes very prominent, and has a curious resemblance to the odontoid process of the axis vertebra of a bird. This “intercondyloid” process of the atlas sometimes {Amphiuma, e.g.) exhibits on each side, near its termination, an articular facet, which plays on a corresponding facet of the adjacent face of the occipital condyle.

Mayer (Analecten, p. 10) was misled by the form of this process into the supposition that the vertebra to which it belongs is not the atlas, but the odontoid, vertebra. But there is a similar process of the first vertebra in the Rays, and the relations of the vertebra to the nerves show that it is certainly not the homologue of the axis vertebra of other Vertebrates. The first spinal nerve, which haa the distribution of the hypo-glossal of the higher Vertebrata, passes out of the spinal canal, either between the first and second vertebrae, or through a foramen in the arch of the first, in the Amphibia, which have no proper suboccipital nerve. This is a very curious circumstance, and requires elucidation by the study of development.

In the Anura, the atlas has the same general form, but the median process is either inconspicuous, as in Rana esculenta, or may be absent.

Among the Labyrinthodonta, the atlas of Mastodonsaurus only is known. It presents two concave facets anteriorly, separated for about half their length by a notch, which probably lodged a ligament.

The atlas of the Peromela has the two characteristic facets for the occipital condyles, but the intercondyloid process is absent, and the anterior margin of the arch of the vertebra projects forwards towards the corresponding margin of the occipital foramen (Epicrium).

In those Urodela which possess posterior limbs (except Proteus and Amphiuma), one vertebra, or sometimes two (Meno-pomai), are distinguished from the rest as “sacral” by having stouter ribs, the outer ends of which abut against, and are united by ligaments with, the ilia.

The Anura always possess a sacral vertebra (sometimes ankylosed with its predecessor or successor), the transverse processes of which are often enlarged, and sometimes greatly expanded, at their iliac ends. The characters of the sacrum of the Labyrinthodonta are not known.

In the Urodela, the anterior caudal vertebrae, except the first, have inferior arches, which, like the neural arches, are continuously ossified with the centra; and the same condition obtained in the caudal vertebrae of the Labyrinthodonts.

In the Anura, the caudal vertebrae are replaced by a long coccyx, consisting of an osseous style, to the dorsal aspect of the anterior end of which two neural arches are ankylosed. The anterior face of the style usually presents two concavities (one in Bombinator and some other genera), which articulate with the corresponding convexities on the posterior face of the centrum of the sacral vertebra. The number of the vertebrae in the spinal column of the Urodela and Peromela varies very much. In the long-bodied Urodela and Peromela they may be very numerous.

According to Cuvier, Caecilia has 230; Siren, 99; Amphi-uma, 75; in Menobranchus there are 18 prae-sacral and 25 caudal; in Salamandra, 15 and 26; and a similar variation appears to have obtained in the Labyrinthodonts. On the other hand, in the Anura the number of vertebrae (excluding the coccyx) is very constantly nine; though this number undergoes an apparent reduction, in some cases, by the ankylosis of the first and second vertebrae (Cera-tophrys dorsata, Pipa, Lactylethra, Breviceps), and in others by that of the sacral vertebra with the coccyx (Pipa, Dactylethra, Breviceps, Pelobates).

In the carboniferous Labyrinthodont, Archegosaurus, the notochord appears to have persisted throughout life, and the ossification of the centra of the vertebrae to have gone no further than the development of bony rings, such as those with which the ossification of the centra of the vertebrae of a tadpole commences.

The Cranium,.— The skull is always very depressed, and is usually broad in proportion to its length, though, in this respect, there is considerable variation, the skulls of Proteus, Menobranchus, and Amphiuma being narrow, when compared with those of Siredon, Menopoma, and the Anura. The occipital foramen is situated in the middle of the posterior face of the cranium, and there are always two occipital condyles. The long axis of the suspensorium, or pedicle by which the mandible is connected with the sidewalls of the brain-case, varies much in its direction— passing obliquely downwards and forwards in the lower Amphibia and in the larval condition of all, but swinging back until it stands out at right angles to the axis of the skull, or becomes directed downwards and backwards, in the higher Amphibia. The suspensorium is almost immovable upon the skull, being clamped thereto by the squamosal bone, besides being, as a general rule, united with some part of the wall of the skull by synchondrosis. The “primordial skull,” or chondro-cranium, usually remains, to a great extent, unossified, even in the adult. In the Urodela, the hyoidean arch· is always connected by strong ligaments with the suspensorium; but, in the Anura and in the Peromela, it becomes completely detached from the suspensorium, and may be free (Peromela), or acquire a new attachment to the periotic region of the skull in front of the fenestra ovalis (Anura).

The bones which are always present in the Amphibian skull are the exoccipital, pro-otic, parasphenoid, vomer, parietal, frontal, squamosal, premaxillary, palatine, quadrate, dentary, splenial, and angular. The basi-occipital and the basisphenoid are always absent, or are represented by mere partial calcifications of the chondro-cranium. There is always a fenestra ovalis closed by a stapes. The branchial arches do not exceed four pairs in number, and, in the perennibranchiate Amphibia, there are never fewer than three pairs.

The skull of the Frog (Figs. 4-7), as the most accessible member of the group, and that, the development of which has been most carefully studied, may be taken as the starting-point from whence to follow the various modifications of the Amphibian skull. At the sides of the occipital foramen, it presents two large exoccipi-tal ossifications (E. O.), which bear the prominent occipital con dyles, and, in old specimens, may meet in the middle ventral line.

Dorsally, however, they remain separated by a narrow tract of cartilage, which may become more or less calcified. External to the condyles, are the foramina,by which the vagus and glosso-pharyn-geal nerves emerge from the cranial cavity; and, beyond these, the bones expand outwards and forwards, so as to embrace the posterior half of the fenestra ovalis, while above, they enclose the greater part of the posterior vertical semicircular canal. The cartilage which incloses the summit of the arch of that canal, however, appears always to remain unossified, and its place is occupied by a groove in the dry skull (Fig.

4, Bp.) These ossifications, therefore, answer primarily to the exoccipitals, but, in addition, represent the opisthotic and epiotic elements. Above the fenestra ovalis, the Avail of the otic capsule is produced outwards into a stout shelf, which forms the roof of the tympanic cavity, and corresponds with the tegmen tym? pani in man (ί.ί.) This is largely ossified in continuity with the exoccipital 1 posteriorly, and the pro-otic (to which in all probability it properly belongs) in front. The outer extremity of the tegmen, however, remains cartilaginous, and, in front, it passes into a curved band of cartilage, which, as it is continued backwards into the suspensorium, may be termed the “dorsal crus” of the suspensorium (Fig. 9, c.d.) The T-shaped squamosal bone (*S⅛∙) sends a broad, flat process inwards, which rests upon the tegmen tympani, while its long descending process lies external to the cartilaginous suspensorium, and the posterior half of its cross-piece, or proper squamoso-zygomatic part, has the same relation to the dorsal crus of the suspensorium. The suspensorium has a second attachment to the skull, by a “ventral crus” (Fig. 9, e.v.), which diverges from the dorsal crus at the anterior extremity of the suspensorium, and is continued into two branches. One of these, passing outwards and forwards, becomes the pterygoid cartilage. The other (Fig. 9, pd.), directed backwards and inwards, may be termed the “pedicle of the suspensorium;” it becomes thickened at its inner extremity, and articulates with a facet in front of the fenestra ovalis, and close to the attachment of the hyoidean cornu. A thin fibrous band extends from this inferior crus to the side walls of the skull, passing between the first division of the fifth nerve in front, and the second and third divisions behind. The space between the dorsal crus of the suspensorium and the pedicle is filled, in the fresh state, with fibrous tissue, which constitutes the anterior boundary of the tympanum. It is traversed (as Dugès long since pointed out) by the posterior division of the seventh nerve, which therefore lies above the pedicle. The pro-otic ossification (Pr. 0.) not only walls in the anterior part of the otic •capsule, but extends for a short distance forwards in the side walls of the skull. Hence, the foramen of exit for the trigeminal and portio dura (V.) is pierced in this portion of the pro-otic; and the foramen for the sixth nerve is seen at its lower margin. In front of the pro-otic, the lateral walls of the skull remain cartilaginous for some distance, and are perforated by the large optic foramen (II.) Anteriorly to the exit of the optic nerves, the side walls of the skull are formed by elongated plates of bone, which are parts of an extensive ossification of the anterior moiety of the brain-case and the posterior part of the nasal capsules, constituting the complex structure termed by Cuvier “os en ceinture,” and by Dugès “ethmoide.” As it takes the place of the ethmoid, presphenoid, and orbito-sphenoids, it may be termed the sphen-ethmoid (S.e.) It may be compared to a dice-box, one-half of which is divided by a longitudinal partition. This half is anterior, the longitudinal partition being represented by the ossified mesethmoid; while the posterior, undivided, half lodges the anterior portion of the cerebral hemispheres and the olfactory lobes. The front wall of this posterior cavity is perforated by the olfactory foramina; while the outer and posterior wall of each anterior, or nasal, chamber presents, where it forms the anterior and inner boundary of the orbit, a small aperture (V) through which the orbito-nasal nerve passes. The exoccipitals, pro-otics, and sphen-ethmoid are ossifications which involve the chondrocranium, though they largely consist of secondary bone. The supra- occipital is represented, if at all, by a mere calcification of the cartilage, and the like is true of the quadrate, which is an ossification of the distal end of the suspensorium. The quadrate, however, very early becomes continuous with a slender style of membrane bone, the proper jugal, which applies itself to the inner face of the posterior end of the maxilla, and thus gives rise to the quadrato-jugal (Q.J.) Ligamentous fibres also connect the anterior end of the zygomatic process of the squamosal directly with the pterygoid, and indirectly with the maxilla and jugal, and pass from the same process to the fronto-parietal bone, forming a fascia over the levators of the mandible, and encircling the orbit. A strong band is continued forwards, over the ascending process of the maxilla, to the alinasal cartilage of the chondrocranium.

The short premaxillae (Pmx.) are united suturally in the middle line, and have stout ascending processes, which become closely connected with the “rhinal processes * by means of oval nodules of cartilage adherent to their poste rior surfaces. The long maxillae unite with the premaxillae in front, and with the jugals behind: each sends up a short anterior ascending process towards the alinasal cartilage, with which it is united by ligament, and further back, gives off a longer ascending process which becomes connected with the nasal bones. The palatine bones (2V.) are straight, slender, and flattened. They lie transversely to the axis of the skull, behind the posterior nares, closely applied to the ventral surface of the sphen-ethmoid and of the antorbital processes of the chondrocranium. Externally they come into contact with the pterygoids and maxillae; internally and anteriorly with the vomers. Each pterygoid (Pt.) is a triradiate bone, with an anterior, an inner, and a posterior, or outer, ray. The first, or anterior, process of the pterygoid (Pt v ) is the longest, and lies, for a considerable distance, in contact with the maxilla, reaching forward to the outer extremity of the palatine. The second (F⅛ 2 ) underlies the pedicle of the suspensorium, and comes into contact with the end of the transversely elongated, sub-auditory, portion of the parasphenoid. The third, or posterior, process of the pterygoid (Pt 3 ) is prolonged, in correspondence with the backward elongation of the suspensorium, along the inner side of which it lies, as far as the articulation of the mandible.

The vomers (Vo.) are broad triangular plates of bone, with irregularly-notched outer edges, which are closely applied to the ventral surface of the sphen-ethmoid. Their inner edges are separated by a narrow interval, and each bears numerous teeth, set along a line which is not quite transverse to the axis of the skull. On the dorsal aspect of the skull two elongated flattened bones, united in a median suture, represent the coalesced frontal and parietals (Fr., Pa.), which are separate in the young frog. In front of these, also meeting in the middle line, are two triangular bones, the apices of which extend outwards to the ascending processes of the maxillae, and which roof over the nasal capsules (Act. ) These correspond in position and relations with the nasal and so-called “prefrontal” bones of Saurop-sida, and perhaps with the lachrymals: for brevity’s sake, they may be termed nasals.

The parasphenoid has the form of a dagger with a very wide guard and short handle. The latter lies beneath the ventral junction of the exoccipitals, while the blade extends forwards, and its point underlies the posterior moiety of the sphen-ethmoid, but does not reach the vomers. The “guard” passes outwards on each side beneath the auditory capsules, and ends by an abruptly-truncated extremity, its anterior and external angle coming into relation with the inner process of the pterygoid.

The slender, permanently cartilaginous, hyoidean cornu (Hy.) passes into the cartilage of the auditory capsule on the ventral side, between the fenestra ovalis and the articular surface for the inferior crus of the suspensorium. The fenestra ovalis lies in a cartilaginous interspace between the exoccipital and the pro-otic, and is filled bv the oval cartilaginous stapes (St.) The anterior face of this presents a concave facet, for articulation with a corresponding surface occupying the posterior half of the inner end of the columella auris (C. a.), the anterior half of which fits into a fossa in the pro-otic bone. The columella auris itself consists of three portions—a middle elongated osseous rod, an inner swollen and enlarged cartilaginous part, which articulates partly with the pro-otic and partly with the stapes, and an outer portion, which is elongated at right angles to the rest, fixed into the tympanic membrane, and attached by its dorsal end to the tegmen tympani.

The mandible presents one cartilaginous and three osseous constituents on each side. Of the latter, one, the “Mento-Meckelian” (Parker), is a short curved rod of bone, which unites with its fellow in the symphysis, and is, in fact, the ossified symphysial end of Meckel’s cartilage, which extends thence through the length of the ramus, becoming thicker posteriorly, and furnishing the articular surface for the quadrate. The second, and largest, bony constituent of the mandible is a long membrane bone, which ensheaths the inner and under region of the outer surface of Meckel’s cartilage, rising at one part into a low coronoid process. It obviously represents the angular, coronary, and splenial elements, and may be termed the angulo-splenial (An.) A small dentary element, which bears no teeth, lies over the outer face of the anterior half of Meckel’s cartilage.

The hyoidean apparatus of the adult frog (Fig. 8) presents a body and two slender cornua. The body consists of a broad and thin squarish plate of cartilage, produced on each side into three processes, which may be called anterior, lateral, and posterior. The anterior process (a)-is slender, curves outwards, and very soon divides into two processes, one short, anterior, forming a loop by its ligamentary connection with the second, or posterior, branch, which passes into the long and slender cornu of the hyoid. The lateral process (δ) passes outwards and slightly dorsad—expanding into a broad, hatchet-shaped free extremity. The posterior process (c) is a mere prolongation of the postero-lateral angles of the body of the hyoid. Finally, from the middle of the posterior margin of the body of the hyoid there project two strong bony rods, wider at the ends than in the middle, which embrace the larynx, and have been termed the thyro-hyals (d).

The parieto-frontals, nasals, premaxillae, maxillae, squa-mosals, palatines, pterygoids, and parasphenoid, the dentary and angulo-opercular bones, may be removed from the frog’s skull without injury to the chondrocranium, the structure of which then becomes apparent (Fig. 9).

It furnishes a floor, side walls, and roof to the brain case, interrupted only by a large space (fontanelle), covered in by membrane, which lies in the interorbital region under the parieto-frontals; and by the foramina for the exit of the cranial nerves. It consists entirely of cartilage, except where the exoccipitals, the pro-otics, and the sphen-eth-moidjnvade its substance. In front of the septum of the anterior cavity of the sphen ethmoid, it is continued forward, between the two nasal sacs, as the cartilaginous septum narium, from which are given off, dorsally and ventrally, transverse alae of cartilage, which furnish a roof and a floor, respectively, to the nasal chamber. Of these, the floor is the wider. The dorsal and ventral alae pass into one another where the chondrocranium ends anteriorly, and give rise to a truncated terminal face, which is wide from side to side, narrow from above downwards, and convex in the latter direction. The lateral angles of this truncated face are produced outwards and forwards into two flattened prae-nasal processes (p. nl.)∙, these widen externally, and end by free edges, which support the adjacent portions of the premaxillae and maxillae. From the ventral face, just behind the truncated anterior end of the chondrocranium, spring two slender cartilages (r.p., r.p.), which do not seem to have been noticed hitherto.

Each of these inclines towards the middle line, and ends against the middle of the posterior face of the ascending process of the premaxilla by a vertically elongated extremity. These may be termed the rhinal processes. An oval nodule of cartilage is attached to the posterior face of the dorsal end of the ascending process of the premaxilla, and serves to connect it with the rhinal process. On the dorsal face of the chondrocranium, just above the point of attachment of the rhinal processes, the external nasal apertures are situated, and the outer and posterior margins of each of these apertures is surrounded and supported by a curious curved process of the cartilaginous ala — the alinasal process (a.n.) Where the sphenoidal and the ethmoidal portions of the sphen-ethmoid meet, a stout, transverse, partly osseous and partly cartilaginous, bar (A.O.) is given off, which is perforated, at its origin, by the canal for the orbito-nasal nerve. It then narrows, but becoming flattened from above downwards, rapidly widens again, and its axe-head-like extremity abuts against the inner face of the maxilla. The anterior angle of the axe-head is free; the posterior angle is continued back into a slender cartilaginous rod, whichbifurcates posteriorly; the outer division passes into the ventral crus (c.v.) of the suspensorium. The inner (pd.) is the pedicle of the suspensorium already described.

Meckel’s cartilage, articulated to the free end of the suspensorium, is unossified throughout the greater part of its extent, no osseous articulare being developed; but, at its symphysial end, each cartilage becomes ossified, and forma the “mento-Meckelian” element of the mandible.

The slender cornu of the hyoid passes directly into the periotic cartilage immediately in front of, though below, the fenestra ovalis. It is unossified throughout its whole length.

With many variations in detail, the skulls of the Anura in general are readily reducible to the type of that of the frog. In the Λglossa, which differ so widely in many respects from the other Anura, the cranium presents some notable peculiarities.

In Dactylethra the skull is similar to that of the ordinary Anura in general form, but the nasal region is small in proportion to the orbito-temporal space. The fronto-parietals are ankylosed together, and extend forwards as far as a line drawn along the anterior edge of the antorbital processes. Here they overlap a very singular bone, consisting of two broad alae, which lie between the anterior edge of the frontal and the external nares, and of a median portion which is continued forwards, as a narrow, flat, curved process, between the nasal chambers, being received into a sort of groove of the chondrocranium. The bone is readily raised up from the subjacent chondrocranium, of which it appears to be quite independent. At the outer end of each of its alae, and between the antorbital process and the nasal capsule, is a small, transversely elongated, slender bone, loosely connected by fibrous tissue with the foregoing. The ethmoid is completely cartilaginous. The parasphenoid has the ordinary sword shape, except that the “guard” is extremely short; but its point extends along the base of the skull, passing between the nasal sacs, underlying their septum, and terminating close to the premaxillae. The vomers are represented by a transversely elongated rhomboidal osseous plate, devoid of teeth, which lies between the two posterior nasal apertures, and therefore much behind the anterior end of the parasphenoid. The side walls of the cranial cavity are ossified from the antorbital process to the anterior boundary of the foramen for the fifth nerve, just in front of which they are pierced by the optic foramen. There is no palatine bone. The pterygoid, in the main, resembles that of the ordinary frogs; but, in consequence of the shortness and little backward extension of the suspensorium, the outer process passes almost directly outwards, with hardly any backward inclination. A bony plate, which extends backwards from the posterior edge of the inner and outer branches of the pterygoid, underlies the tympanic cavity and the auditory capsule, and forms the floor of the Eustachian canal. The squamosal is a short broad bone, with a long anterior process, which becomes connected, by direct articulation, with the pterygoid, and by ligament with the maxilla. The premaxillae are small, and the maxillae are connected merely by ligaments with the suspensorium, there being no jugal. The columella auris is remarkably strong, and is bent in the middle, so that its two halves form an obtuse angle; the anterior half lies against the inner face of the tympanic membrane. The posterior half runs parallel with the posterior edge of the tegmen tympani, towards the fenestra ovalis. Ligamentous fibres fix the columella firmly, though movably, to the superior margin of the tympanic cavity, where it is bounded by the squamosal. The stout cartilaginous hyoidean cornua are attached just beneath the anterior and inferior part of the margin of the fenestra ovalis. The body of the hyoid is very small, but the two “thyro-hyals” are extremely long and broad cartilages. There is no ossified “mento-Meckelian” element.

In Pipa, the skull is extraordinarily depressed and broad. The nasal bones are wide, flat, triangular, and quite distinct from one another, a forward prolongation of the coalesced fronto-parietals extending between the two. The parasphenoid, very broad in the greater part of its extent, and having the guard rudimentary, sends a narrow median process forwards underneath the nasal septum, as in Dactylethra. No trace of a vomer, or palatine bone, was to be found in the specimen examined. The pterygoid is very like that of Dactylethra, but its inner branch is greatly prolonged, and the floor sent under the Eustachian tube unites much more closely with the produced exoccipital. The squamosal is very small, and the place of its zygomatic process is taken by ligament. This ligament, however, unites with the pterygoid in the same way as the bony process which answers to it in Dactylethra. The columella is less massive than in Dactylethra, and the end which abuts against the tympanic membrane is imbedded in a disk of cartilage. The occipital condyles look outwards and backwards, instead of inwards and backwards, as in Dactylethra.

The hyoidean cornua are wanting, the thyro-hyals being large, but not so large proportionally as in Dactylethra.

The skulls of the Urodela present a very interesting series of modifications, leading from a condition in which the cranium retains, throughout life, a strongly-marked embryonic character, up to a structure which closely approximates that found in the Anura.

In Menobranchus, for example, the chondrocranium of the adult is in nearly the same state as that in which it exists in a young tadpole or larval salamander (Figs. 10, 11, 12).

Instead of there being a well-developed cartilaginous brain-case, interrupted only by a dorsal fontanelle, as in the frog, both the floor and roof of the cranial cavity are formed by merely fibrous tissue, which underlies the frontal and parietal bones, and overlies the parasphenoid; and only its sides and its anterior end are bounded by cartilage.

The occipital region remains membranous in the middle line, both dorsally and ventrally, and exhibits, in the latter aspect, the remains of the notochord. The cartilaginous rods (7 7 r.), which bound the cranial cavity laterally, and represent the trdbeculoe of the embryonic vertebrate skull, are separated by a wide oval space, which occupies the whole length of the floor of the cranial chamber. Anteriorly, they converge, and, just before they do so, give attachment to the slender antorbital processes (A. O.} which lie behind the posterior nares. They then unite, and, becoming applied together, coalesce into a flattened narrow mesethmoid, to the anterior extremity of which the premaxillary bones are applied. They give off neither alinasal nor subnasal processes, and therefore furnish neither roof n®r floor to the nasal chamber. Posteriorly, they become flattened from above downwards, and coalesce with the auditory capsules, and with the cartilage which extends beneath these, and gives rise to the occipital condyles (A. 0.)

Each auditory capsule has a generally oval form, but is produced posteriorly, so as to give rise to a conical epiotic process (Ep. 0.), which projects beyond the level of the occipital condyle. Fitted into the outer wall of each is the relatively large, conical, stapes, whence strong ligamentous fibres proceed to the posterior face of the suspensorium.

Immediately in front of the auditory capsule, the suspensorium passes by a strong pedicle (shown, but not lettered in Fig. 12) into the trabecula, and then, directed outwards, downwards, and forwards, ends in the quadrate process (<?), with which the dorsal end of Meckel’s cartilage articulates. A large process (o) ascends from the posterior face of the suspensorium, and applies itself to the outer and anterior face of the auditory capsule. A small and hardly perceptible elevation (p) is seen near the quadrate process of the suspensorium. Finally, a flat process (a, Fig. 11) ascends above the pedicle, and applies itself to the dorsal face of the trabecula.

On comparing this with the suspensorium of the frog, it is clear that the rudimentary process (p) answers to the pterygoid cartilage; and that the process o (the otic process) answers to the dorsal crus of the suspensorium. In fact, the posterior, or hyo-mandibular, branch of the seventh nerve passes back beneath this, and above the stapedial ligament, to its distribution.

The pedicle answers to the part so named (including the ventral crus of the suspensorium) in the frog, though it retains the embryonic relations to the trabecula, such as exist in the tadpole. The ascending process (a) lies between the orbito-nasal and the other branches of the trigeminal, the orbito-nasal passing between it and the trabecula. A similar process is very generally found in the Urodela (being particularly large, for example, in Menopoma), but appears to be represented only by fibrous tissue in the Anura.

Meckel’s cartilage (3fc⅛., Fig. 12) is thick and deep at its articular end, but, after furnishing a surface of attachment for the elevator muscles of the jaw, it rapidly narrows, and ends in a point, at some distance from the symphysis of the dentary bones.

The hyoidean apparatus (Fig. 13) is represented, on each side, by a cartilaginous rod, subdivided into a short hypo-lιyal (IL Λ.) and long cerato-hyal (C. A.) A strong ligament extends from the front face of the latter, below its free summit, to the suspensorium, reaching this at the same place as the stapedial ligament, into which it is continued. From a point a little above the attachment of this ligament, another ligamentous band arises, and, crossing the former, on the inner side of which it passes, becomes attached to the angle of the mandible. It answers to the interoperculum of a fish, and has nearly the same relations as the stylo-maxillary ligament of the higher Vertebrata.

In the ventral median line, the hypo-hyals are connected only by fibrous tissue. Firmly united with this, however, there is a median first basibranchial cartilage (72δ 1 ), succeeded by a second basibranchial (Eb 2 ), which is ossified, and is the only bony constituent of the hyo-branchial apparatus. There are only three branchial arches. The first consists of a stout and long cerato-branchial (Cb v ), bearing an equally well - developed epibranchial (Epb r ). The second is represented by a mere nodule of cartilage (Cδ 2 ), with which the expanded end of the second epibranchial (Ep. b 2 ) articulates. The third and smallest epibranchial (Ep. δ 3 ) is articulated with the step-like broad end of the second. There is no trace of any fourth branchial arch, such as exists in tadpoles and in young salamanders; and in Siredon, Siren, Amphi uma, and Menopoma.

The most curious feature in the bony skull, oτosteo-cranium, of Menobranchus is the presence of the prominent conical ossifications which lie external to the exoccipitals,and occupy the place of the epiotic and opisthotic bones. In possessing these elements of the skull, in so large and distinct a form, Menobranchus differs from all Amphibia, save Proteus and the extinct Labyrinthodonts. The parietal bones are separate from the frontals, and send, as is usual in the Urodela, a long process forwards on each side of the latter. In Menobranchus, this process is extremely long, reaching the olfactory foramen, the posterior margin of which it bounds. There are no nasal nor prefrontal bones, nor any distinct ali-sphenoidal, orbito-sphenoidal, or ethmoidal ossifications; of the maxilla, nothing but a rudiment appears, and this is sometimes absent. There are no jugal or quadrato-jugal ossifications. A palato-pterygoid plate, bearing teeth on its expanded palatine portion, extends from the antorbital process to the inner face of the suspensorium, which, as stated above, has a mere tubercle in the place of a pterygoid process. The vomers are long, bear teeth along their outer edges, and diverge backwards so as to leave an interspace between their inner edges. Posteriorly, they articulate with the anterior ends of the palato-pterygoids.

The squamosal is a long, slender, curved bone, devoid of any zygomatic process, which extends from the articular end of the suspensorium, along its outer edge, to the outer side of the epiotic. From its posterior margin it sends down a short process over the stapedial ligament.

The parasphenoid is a broad thin plate of bone, which extends from near the junction of the vomers to the occipital foramen. The distal end of the suspensorium is incompletely ossified, as a quadrate bone; and a dentary and a splenial element, both dentigerous, lie, the former external to and below, the latter internal to, Meckel’s cartilage.

The skull of Proteus is, in its general characters, similar to that of Menobranchus, but is more extensively ossified.

In Siren, the skull, which has the same elongated form and forwardly directed suspensoria as in ΛIendbranchus and Proteus, possesses, in the epiotic region, two strong crests, which project backwards beyond the level of the occipital condyles, but are otherwise very different from the epiotics of the latter genera. The “exoccipitals” and prootics are completely fused together, even in half-grown specimens, a mere rim of cartilage being left around the fenestra ovalis. The lateral walls of the skull present ossifications extending from the exits of the orbito-nasal nerves, forwards to the mesethmoid, or internasal portion of the trabeculae, and completely encircling the olfactory foramina. But these ossifications remain distinct for a considerable time, if they ever coalesce. Each, therefore, represents half of the sphen-ethmoid of the Frog.

The flat and wide parasphenoid extends forwards to the space left by the divergence of the vomers. On the roof of the skull, each broad parietal sends forward a prolongation along the outer edge of the frontal, which reaches the antorbital process. The premaxillae have very long ascending processes, which lie upon the mesethmoid, and are received between the ends of the frontal bones. Between these ascending processes one or two elongated ossifications are situated. These were termed “nasals” by Cuvier, but their position does not accord with this determination. The horizontal, or oral, portions of the premaxillae, on the other hand, are very short and edentulous, each being coated by a thin plate of horny substance. The maxillae are rudimentary or absent. The vomers are two, flat, oval, bony plates, the ventral aspect of which is beset with parallel rows of teeth set in obliquely-disposed curved lines, the convexities of the curves being turned inwards and backwards. The vomers are in contact anteriorly, but diverge posteriorly. The posterior extremity of each abuts upon a plate of similar form, but much smaller, and bearing fewer series of teeth, which lies on the ventral side of the origin of the antorbital process, and represents the palatine bone. The aperture of the posterior nares lies just opposite, and external to, the junction of these two bones.

The suspensorium is short, thick, and completely cartilaginous. Dorsally and internally, it is attached by a stout pedicle to the trabecula in front of the auditory capsule, while its dorsal and posterior face lies against the truncated anterior face of the pro-otic. The angle formed where this face joins the outer face represents the otic process of Menobranchus. There is no pterygoid process, nor any trace of a pterygoid bone. Just above the suspensorium, and seemingly connected with it, there proceeds from the anterior face of the pro-otic region of the skull, a strong, triangular, forwardly-directed cartilaginous process. From the free anterior end of this, a band of fibrous tissue passes, and, encircling the eye, is attached to the antorbital process. The squamosal is a slender curved bone, extending from the epiotic ridge to the articular end of the suspensorium, where it is widest. It exhibits only a rudiment of the well-marked process which extends towards the stapes in Menobranchus and Proteus. The mandible presents a dentary, an angular, and a dentigerous sρlenial element; and the proximal end of Meckel’s cartilage is ossified, giving rise to a dense nodular articulare. The dentary is toothless, and supports the inferior horny beak.

The cornu of the hyoid is very stout, and its ventral moiety is ossified. The much thicker dorsal moiety is cartilaginous, and its recurved dorsal end extends beyond the extremity of the skull. At a considerable distance below its apex, a strong short ligament proceeds from its anterior face to the stapes. A broad sheet of ligamentous fibres further unites the hyoid with the lateral walls of the ear capsule, and with the posterior face of the suspensorium (hyo-suspensorial ligament); and a slender ligament (mandibulo-hyoid) proceeds from near the insertion of the hyo-stapedial ligament to the angle of the mandible. There are two basibranchials, both ossified, the posterior ending in short radiating processes; two cerato-branchials and four epibranchials.

In Amphiuma, the suspensoria are very little inclined forwards, and their long axes make nearly a right angle with that of the skull. The portion of the skull which lies behind a line joining the articular ends of the ossa quadrata is very much shorter than the region in front of it. Moreover, although the epiotic processes are prominent, the occipital condyles project far beyond them. One bone represents the exoccipital, epiotic, and opis∙∙ thotic on each side. The pro-otic is large, and gives rise to the anterior moiety of a strongly-marked temporal ridge. Its exposed surface presents two fossae, divided by a nearly vertical linear elevation. The large pariétale form the posterior portion of the temporal ridge, and diverge anteriorly, to be continued forwards, on each side of the frontale, to the sphen-ethmoid. The frontals, in like manner, diverge in front to receive a median ossification, which is continuous with the coalesced median processes of the premaxillae. The anterior half of each frontal is rugose, as are the exposed surfaces of the sphen-ethmoid and of the nasal bones, and the integument is firmly adherent to these rugosities. The nasals are broad and triangular. The truncated base of each lies over the nasal aperture; the inner edge articulates with the ascending process of the premaxilla; the outer edge joins first the maxilla, and then the sphen-ethmoid. The premaxillae are so thoroughly ankylosed that no trace of their primitive distinctness is to be seen. The large maxillae extend back for half the length of the skull, are firmly united with the adjacent bones, and are connected by dense ligament with the extremity of the quadrate bone. The greatly elongated vomers diverge but little; nevertheless, they come in contact only by their anterior extremities. In the rest of their extent they are separated, in front, by a median ossification representing the anterior part of the sphen-ethmoid, and, behind, by the anterior forked prolongation of the parasphenoid which embraces this ossification. The rest of the parasphenoid is broad and flat; it widens a little, in front of the auditory capsules, so as to form a rudiment of the “guard” in the frog’s skull.

The osseous pterygoid is a curved plate of bone, convex inwards and concave outwards, which articulates posteriorly with the quadrate, and, in front, stops short, at little more than half the distance from its posterior end to the internal nostril. The cartilaginous pterygoid process of the suspensorium extends some way beyond it, and, widening, is attached by ligament to the maxilla. Posteriorly, the cartilaginous pterygoid is traceable, as a comparatively narrow band, on the inner side of the bony pterygoid, to the pedicle of the suspensorium, which is attached in front of the fenestra ovalis, and above the rudimentary “guard” of the parasphenoid. An ascending process passes from it between the orbito-nasal and the other divisions of the trigeminal. The otic process of the suspensorium, which is articulated with the outer face of the auditory capsule, is cartilaginous; but the rest of the suspensorium is ossified as a quadrate bone. This is, as usual, clamped to the skull by the squamosal, which is broad and expanded above, and narrow below. Behind, the suspensorium is directly articulated with, the styliform projection of the centre of the stapes. A very strong hyo-suspensorial ligament passes from near the distal end of the suspensorium to the cornu of the hyoid. The mandibulo-hyoid ligament is much weaker. The hyoidean apparatus presents a median basi-hyal, connected by a rounded hypo-hyal on each side, with a long and curved cerato-hyal, which is almost completely ossified. The first basibranchial is elongated and cartilaginous—the second is absent. The first branchial arch is a single elongated bone, representing the similarly coalesced cerato-branchial and epibranchial in Menopoma. The second cerato-branchial is small and cartilaginous. The three posterior epibranchials are simple curved cartilages; and the single branchial cleft is placed between the third and fourth epibranchials.

The skulls of the four genera, AIenobranchus, Proteus, Siren, and Amphiuma, now described, resemble one another, and differ from those of other Amphibia, in their elongated form; and, especially, in the relative narrowness of the facial region in front of the orbits, which, as the case of Amphiuma shows, arises, not from any want of development of the maxillary bones, when they exist, but from their taking a direction which but slightly diverges from parallelism with the axis of the skull. Moreover, they all possess well-marked epiotic prominences. Amphiuma differs widely from the other three, in the great size of its maxillary bones, in the absence of palatine bones, in the projection of the occipital condyles beyond the epiotic processes, in the ankylosis of the premaxilløe, in the presence of well-developed nasal bones, in the coalescence of the first cerato-branchial with the first epibranchial, and in the transverse direction of the suspensorium.

In most of those respects, in which Amphiuma differs from Aienobranchus, Proteus, and Siren, it approaches the Salamanders; especially if we take such forms as Anaides into account. On the other hand, in the entire absence of a palatine bone, and in the fusion of the first cerato-branchial with its epibranchial, it agrees with Afenopoma and Cryptobranchus.

In Al·enopoma, the skull has a broadly-rounded snout, and its posterior contour slopes forwards and outwards (without being interrupted by conspicuous epiotic prominences), in the manner characteristic of the higher Urodela. The small pro-otics are separated from the exoccipitals (which also represent the epiotics and episthotics), by a wide cartilaginous interspace, in which the fenestra ovalis is situated. The parietal sends a process forwards, along the outer edge of the frontal, between it and the orbito-sphenoid. This meets a curved flat bone, which bounds the orbit anteriorly and internally, and articulates with an ascending process of the maxillary bone. It may therefore be regarded as a prefronto-lachrymal. The frontale unite in a long median suture, and then, diverging, embrace the nasal bones, and articulate externally with the fore part of the ascending process of the maxillary bone, which is thus received between the frontal and the prefronto-lachrymal. The very broad parasphenoid extends from the exoccipitals to the vomers, with which it unites by a denticulated squamous suture. The wide vomers are united by a median suture, and expand in front, ending in arched edges, close behind which the teeth are set. The premaxillae are separate and small, articulate with the arched edges of the vomers, and send up strong ascending processes to the dorsal face of the skull, where they firmly unite with the nasals. The squamosal is a flattened prismatic bone, as broad at one end as at the other, which articulates with the parietal externally, and with the quadrate internally. Like the suspensorium, which it covers, it stands out at right angles with the axis of the skull. There is no palatine bone. The pterygoid is broader and more square than in any other Amphibian, in consequence of the great expansion of its internal process, which articulates by its whole length with the parasphenoid. The anterior process ends in a free pointed cartilage, directed outwards and forwards, and united with the maxilla by ligament, as in the higher Urodela. The external process extends to the articular end of the quadrate, as usual, and is continued thence along the cartilaginous suspensorium to its attached end.

The chondrocranium forms a complete ring of cartilage round the occipital foramen, continuous at the sides with the auditory capsules. From these the trabeculae are continued forwards, as in Menobranchus, leaving a very wide ventral fontaneUe. At the anterior end of this they unite and form the mesethmoid, from which roof and floor plates of the nasal capsules are continued. The suspensorium is connected by a pedicle with the trabecula, in front of the auditory capsule, and gives off a broad ascending process, which becomes ossified continuously with the pterygoid, over the orbito-nasal nerve. A stout otic process is articulated with a facet on the antero-external region of the periotic capsule, and is further connected with it by ligamentous fibres. The quadrate ossification involves a small portion of the articular end of the suspensorium; it thence extends upwards, on the dorsal aspect of the suspensorium, gradually becoming more slender, and nearly reaches the point at which the otic process of the suspensorium articulates with the periotic cartilage.

The osseous skull of Cryptobranchus is extremely like that of Afenopoma.

In Aienopoma the hyo-branchial apparatus presents the same general structure as that of Siredon, except that the second basibranchial seems to be wanting, while the first is very broad and rounded; at the same time, the epibranchial and the cerato-branchial of the first arch aro represented by only one continuous cartilage.

In Cryptobranchus, however, a considerable reduction has taken place, the two posterior pairs of branchial arches present in Aienopoma having disappeared. The second arch still presents a division into cerato-branchial and epibranchial, but the dorsal end of the latter is closely united with that of the preceding arch. It is interesting to observe, however, that the modification thus effected is quite different from that which occurs in the Salamanders, in which, in the adult state, the first branchial arch retains its ’ two segments; while the second, reduced to its cerato-branchial, is applied against the first, at the junction of the cerato-and epi-branchial; and the second basibranchial persists as the ossiculum thyroideum of Von Siebold.

Aienopoma and Cryptobranchus further differ from the proper Salamanders in having the vomerine teeth disposed along the anterior edges of the expanded vomers. Unfortunately nothing is known of the larvae of these forms, but it would seem as if, in them, the primitive vomers enlarge by extension of ossification behind, and not in front of, the originally existent teeth.

Tn the remaining Urodela, the Salamanders proper, the skull has the broadly-arched snout and the shelving posterior contour of Menopoma, but the vomers and pterygoids are very different.

The structure of the skull in these animals will be best understood by commencing with that of Siredon, which, though perenno-branchiate under ordinary circumstances, is totally unlike the other so-called Perennibranclιiata in cranial structure, and is, in fact, to all intents and purposes a larval Salamander.

An ossification on each side of the occipital foramen represents the exoccipitals, epiotics, and episthotics. In front of each of these is a pro-otic, and an orbito-sphenoid. The latter is sometimes united with its fellow of the opposite side. 1 The skull is roofed in by pairs of parietal, frontal, prefronto-laehrymal, and nasal bones. The maxillae are short, and are united with the ends of the suspensorium only by fibrous tissue. There is a very broad and flat parasphenoid, which extends from the inferior margin of the occipital foramen, and ends, anteriorly, by a wide, irregularly convex edge, which does not reach the vomers. The latter bones are elongated and curved, and their long axes diverge posteriorly, as in Meno-branchus; but their anterior ends are far apart, and they lie, separated by the whole breadth of the parasphenoid, and between that bone and the premaxilla and maxilla, adherent to the ventral face of the subnasal process of the chondrocranium.

The short palatine bones are situated immediately behind, and on the inner side of, the posterior nares, but their somewhat tapering, external and posterior, ends do not articulate directly with the pterygoids. The latter are triradiate bones, with an inner process which passes towards the base of the skull; an outer, which runs down the suspensorium; and a long anterior process, which gradually diminishes in breadth forwards, and is connected only by ligament with the palatine. Three ossifications embrace Meckel’s cartilage. The dentary covers its outer face throughout its whole length. The angular lies on the inner face of its posterior two-thirds, and the small dentigerous splenial is also applied to its inner face between the angular and the dentary. The chondrocranium is in much the same condition as that of Menopoma. There is a broad basicranial cartilage situated between the auditory capsules, and passing, at the sides and above, into a complete occipital arch. But, in front, the trabeculae, though they have increased in vertical height, remain united by fibrous tissue only, both in the floor and in the roof of the skull, which thus presents two great “fontanelles” when the parasphenoid, parietals, and frontals are removed. In front, they coalesce, each giving off, as it does so, a flat antorbital process, which is expanded at its outer end, where it supports the maxilla. Below, this process gives attachment to the pala tine. By their coalescence, the trabeculae give rise to a broad internasal septum (or mesethmoid cartilage), and they expand, on each side, below, into subnasal plates, which are separated, anteriorly, by a wide notch in the middle line. The curved outer edges of these plates give attachment to the premaxillae and maxillae, and they answer to the prae-nasal processes of the chondrocranium of the frog. Between the posterior edge of each of these and the anterior edge of the corresponding antorbital process, the posterior nostril is situated; and the inferior surface of the subnasal plate gives attachment to the vomer. Superiorly, the mesethmoid cartilage expands into a very thin (alinasal) plate, which roofs in each nasal chamber, and supports the prefronto-laehrymal and nasal bones.

The suspensorium is connected, above and internally, with the trabecula of its side by a pedicle; and it has an ascending process which lies over the orbito-nasal nerve (which is therefore included between the pedicle and the ascending pri i >ess) immediately after its exit from the skull. Posteriorly and superiorly, it gives off an otic process, which is articulated with the outer and front part of the auditory capsule; while, inferiorly and externally, it furnishes an articular surface to the mandible. The pterygoid process has the form of a style tapering forwards, and nearly reaching the antorbital process, with which it is connected, however, only by ligamentous fibres. The posterior moiety of Meckel’s cartilage is very stout as far as the coronoid process, and then tapers rapidly to its free, pointed, symphysial extremity.

The hyoidean and branchial apparatus is entirely fibrous and cartilaginous, none of its parts having undergone ossification. Each cornu of the hyoid is connected with the upper and posterior face of the suspensorium, and with the angle of the mandible, by ligamentous fibres—the hyosuspensorial and mandibulo-suspensorial ligaments. The cornua are not subdivided, and are united in the median ventral line by ligament. A triangular first basibranchial extends back from their junction, and is succeeded by a second, as in Menobranchus; but this second basibranchial is not ossified. Two cerato-branchials are attached to the posterior extremity of the first basibranchial on each side, and the anterior is, as usual, followed by a long and strong epibranchial, which supports the anterior gill. The posterior cerato-branchial supports the second epibranchial directly, and the third and fourth epibranchials indirectly.

The interesting observations of Professor A Duméril have shown that, under certain conditions, the ordinarily perenni-branchiate Siredon passes into the caduci-branchiate Amblystoma; and this metamorphosis is accompanied by some very interesting modifications in the structure of the cranium, especially in the vomerine, palatine, and pterygoid regions. Ossification extends forwards from the vomers beneath-the prae-nasal processes, so that the series of teeth, which originally lay along the anterior margins of these bones, come to be situated at their posterior edges. At the same time they take up a direction at right angles to the axis of the skull, instead of being greatly inclined to that axis, as they are in Siredon, and the two sets of vomerine teeth thus form a single transverse row. Moreover, the anterior process of the pterygoid moves outwards until it comes into contact with the inner face of the maxilla. The one end of the palatine remaining attached to the vomer, the other swings outwards, in correspondence with the change of position of the pterygoid, and thus becomes directed transversely to the axis of the skull, immediately behind the posterior nostril, its teeth continuing the transverse line of the teeth of the vomers. Salamanders with the teeth thus disposed have been termed “lechrio-dont.” The maxillary bones are larger than in Siredon, but the jugal arch remains ligamentous. The dorsal ends of the cornua of the hyoid retain their ligamentous connection with the suspensorium, and the ventral ends with the anterior basibranchial. The first cerato-branchial and epibranchial persist, and retain their articulation with one another. The second cerato-branchial remains, but its dorsal or outer end becomes attached to the preceding, and all the three posterior epibranchials disappear. The second basibranchial becomes detached as a Y-shaped piece, which lies in the middle line, in front of the larynx.

In all the other Salamanders, the vomers, in the adult, present the same enlargement of the part in front of the teeth, at the expense of the region behind them, as in Amblystoma. But the teeth rarely offer the same disposition. More commonly they form two series, inclined to one. another at a more or less acute angle, open forwards, and supported upon bony plates, which appear like prolongations of the vomers, extending backwards on the ventral face of the parasphenoid. This “mc-codont” arrangement is strikingly exemplified by Salamandra maculosa, and still more remarkably by Plcthodon and Anaides, where these longitudinal series of teeth beneath the parasphenoid are commonly termed “sphenoidal” teeth. Dugès 3 and other observers, however, have shown that, in larval Salamandrae and Tritones, the vomerine and palato-pterygoid apparatus have, at first, the same disposition as in Siredon; and Dugès has described the process by which the palatine bones, becoming detached from the pterygoids, which rotate outwards, ankylose with the vomers, taking up a position beneath the parasphenoid, and more or less parallel with the axis of the skull; and it can hardly be doubted that the so-called “sphenoidal” dentigerous plates of other genera of Salamandrida have the same origin. If this conclusion be correct, it indicates a very curious morphological difference between the “mecodont” and “lechriodont” Salaman∙ drida.

In all the Salamandrida the parietal bones send long processes forw ards on each side of the frontals. The parasphenoid is a broad flat .plate. V ery often the premaxillae are ankylosed into one bone, and the bones of the periotic capsule coalesce. In some cases there are epiotic processes or ridges. Maxillae are always present, and the snout is usually broadly arched. Nasal bones, distinct from the prefronto-lachrymals, are usually present.

In the genus Anaides the skull is comparatively long and narrow, and the muzzle is less arched than usual. The single premaxilla, and the two well-developed maxillae, follow the semicir∙ cular curve of the broad subnasal plates, to the edges of which they are attached. The hinder free extremities of the maxillae are curved upwards, and the jugal arch is represented only by ligament. Thus far the skull is salamandrine; as it is also in the presence of distinct nasal and prefronto-laehrymal bones, in the disposition of the vomerine and so-called “sphenoidal” teeth, in the absence of an apparent palatine bone, and in the manner in which the pterygoid is produced into a long process, which becomes connected with the inner face of the maxilla. But, in the well-marked downward and forward inclination of the suspensorium, and in the strong crests into which the epiotic processes are developed, the skull oi Anaides is very like that of Siren. .

In the skull of Epicrium glutinosum (Fig. 14), which may be selected as an example of the Peromela, the strong occipital condyles are continued into two ossifications, which rise on to the roof of the skull, where they unite in a short suture, and, spreading out so as to embrace the parietals, are continued over the auditory apparatus, as far as the squamosal and the quadrate bones. Ventrally, no indication of any suture between these bones and the broad parasphenoid is visible; laterally, they pass forward into a continuous ossification, which constitutes the side walls of the auditory capsule, and, in front of this, is perforated by the wide foramen for the trigeminal nerve, and enters largely into the lateral wall of the cranial cavity. The parietal bone rests on the dorsal edge of this lateral ossification, which terminates, anteriorly, by an irregularly 1 See Friedrich and Gegenbaur—“Der Schädel des Axolotl {Siredon pisciformis)" in the Berichte der Königlichen Zootomischen Anstalt tu Würzburg, 1849. Iliis memoir contains an excellent account of the chondrocranium of the Axolotl.

a Recherches, pp. 172, 173, pl. xiv. fig. 89. Rusconi, Observations Anatomiques sur la Sirbne, pl. vi. figs. 3 and 10.

excavated border, between which and the posterior margin of the βphen-ethmoid the cranial wall is unossified. Throughout its whole extent, this large ossification, which represents the exoccipitals, the elements of the periotic capsule, and the alisphenoids, is firmly ankylosed with the parasphenoid. There is a well-developed sphenethmoid, similar in its general characters to that of the frog. It is very closely united, if not ankylosed, with the vomers and premaxillae. The roof of the skull is completed by two parietals, two frontals, and two large nasals, which unite in a long suture, except in front, where, for a short distance, they are separated by the ascending processes of the premaxillae. The dentigerous oral processes of these bones are short, and unite by suture with the maxillae. These send up broad plates which lie in front of and below the orbit, on the sides of the face. The canal for the suborbital tentacle perforates the maxilla in front of the orbit. Posteriorly, the maxilla unites with the squamosal, which is a broad plate firmly fixed to the quadrate, but somewhat loosely united with the frontal and parietal and with the · complex occipito-otic bone. A small crescentic post-orbital bone (denoted by 1 in Fig. 14) articulates with the maxillary and squamosal, and with another bone (2), which answers very nearly to the prefrontal of a reptile. Between the nasal bone and the premaxilla, above and below, and the maxilla behind, a small bone (3) is fitted. The quadrate bone is represented by the ossification of the distal end of the suspensorium, which is inclined a little backwards. The stapes is large and well ossified. Two distinct ossifications, an angulo-articular and a dentary, are discernible in the mandible; and the second short row of teeth, inside those of the dentary, seems to indicate the existence of a splenial element.

In the Labyrinthodonta the skull presents the extremes of form which are met with among the Amphibia, from the elongation observable in Archegosaurus, to the short and broad form of Metopias and Brachyops. The chief characters by which the laby-rinthodont cranium differs from that of its existing allies are the following :—

The occipital condyles in some genera remained long, if not permanently, cartilaginous; and one or two supra-occipital ossifications (probably membrane bones) very generally occur. The epiotics are prominent, and appear to remain permanently separate from the adjacent bones. In front of them, and articulated with the outer edges of the parietal and frontal, are two bones, which are commonly identified with the “squamosal” and “post-frontal” of the higher Vertebrata. The “post-frontal” articulates anteriorly with a large “prefrontal” bone, which bounds the dorsal and anterior contour of the orbit. The outer edges of the “squamosal” and “post-frontal” articulate with two bones, termed the “postorbital” and the “supra-temporal.” The post-orbital lies in front of the other, and contributes to the posterior margin of the orbit,— the rest of the contour of which, between the post-orbital and the prefrontal, is usually completed by a large jugal. Articulated, dorsally and internally, with the “supra-temporal,” and, anteriorly, with the jugal, is a “quadrato-jugal,” which sometimes extends into the articular surface for the lower jaw, and in some cases, at any rate, overlaid a quadrate ossification. There are long paired nasals, between the anterior ends of which the ascending processes of the premaxillae are received; and between these bones, the unusually long maxillae, and the prefrontals, distinct “lachrymal” ossifications occur.

The vomers are large, meet in a long median suture, and bear teeth. The palatine bones, also dentigerous, bound the posterior nares in front, and are elongated antero-posteriorly.

The mandible presents a dentary (probably including a splenial) element, an angular and an articular ossification.

Those surfaces of the cranial bones which were covered by the skin are usually rugose, and sculptured much in the same way as those of the crocodiles, and they frequently present symmetrically-disposed grooves, the so-called “mucous canals,” which, very probably, lodged sensory apparatuses resembling the similarly-named structures in fishes—the homologues of which are found in existing Urodela.

The hyoid is unknown, and what appears to be traces of a branchial apparatus have been observed only in young specimens of Archegosaurus. Hence it is probable, not only that no known Labyrinth-odonts were perennibranchiate, but that the air-breathing condition supervened early in the course of their development.

The Labyrinthodonts doubtless possessed a well-developed chondrocranium, but such a structure would necessarily perish in the course of fossilisation. The singular resemblance of the labyrinth-odont skull to that of the Peromela, in the arrangement of the bones which bound the cavity of the mouth, and the disposition of the teeth upon them, suggests a comparison of the other cranial bones in the two groups. Starting from the nasals of Epicrium, which may be safely identified with those of the Labyrinthodonts, the bone marked (2) in Epicrium corresponds very closely with the labyrinthodont “prefrontal;” and that numbered (1), with the “post-orbital.” No. 3 in Epicrium, in some respects answers to the so-called “lachrymal” of the Labyrinthodonts; while the maxilla of the Caecilian may be taken to represent both maxilla and jugal of the Labyrinthodont. But if this be so, the squamosal of Epicrium corresponds with the supra-temporal of the Labyrinthodont; and a question arises as to the true nature of the “squamosal” and “post-frontal” of the latter. The Limbs.— The pectoral arch in the Amphibia is distinguishable into a scapular, a coracoidal, and a praecoracoidal region, although the extent to which these parts of the primitive cartilaginous arch become separately ossified varies very much in the different members of the group.

In Proteus, Menobranchus, Cryptobranchus, and Menopoma, ossification occurs only in the scapular region. In Siren and Aιnphiuma an additional broad coracoidal ossification occurs, but it does not meet the scapular ossification in the glenoidal cavity. The junction, however, takes place in Siredon and the Salamanders. In none of the Urodela does any ossification appear in or upon the praecoracoidal or supra-scapular cartilage.

A supra-scapular ossification exists in all known Anura. All but Microps and Hylcedactylus have a praecoracoid, which acquires a sheath of bony matter. The glenoidal cavity is bounded by tho praecoracoid, coracoid, and scapula; and in some cases (e.g., Dacty-lethra) the ossified ends of the three unite and give rise to a triradiate suture in the glenoidal cavity, just as the pubis, ischium, and ilium of most Vertebrata unite in the acetabulum. In Systoma gibbosum, contrary to the usual rule, the praecoracoid is far broader than the coracoid (Parker).

In the higher Anura, a median piece, of very variable size, form, and consistency, extends forwards from the junction of the prae-coracoids. Mr Parker considers it to be an outgrowth from these, and terms it the omosternum.

The long bones, both in the fore and hind limbs, consist of an axis of cartilage, sheathed in, and more or less replaced by, a diaphysis of membrane bone. The extremities of the cartilages frequently undergo calcification, and are thus converted into epiphyses. A strong crest characterises the humerus in many male Anura. In the latter, the radius and ulna coalesce into one bone, while in all other Amphibia they remain distinct.

In Siredon, Cryptobranchus, and Menopoma, the carpus contains eight separate cartilages, of which three—the radiale, intermedium, and ulnare — form a proximal row; and four distalia, a distal row. Between these two series lies a single centrale.

In Menobranchus, there are only six carpal cartilages—the ulnare and intermedium, and the radiale and radial distale, respectively, having apparently, as Gegenbaur suggests, coalesced.

In Amphiuma didactylum, the number of carpal cartilages is reduced to four, and in Proteus to three. In both these cases the two largest cartilages form a proximal row.

The Salamandrida usually have seven carpal elements. In the proximal row there are two—a radiale and a coalesced intermedium and ulnare. There is a single centrale and four distalia. These are variously ossified until, in Triton cristatus and alpestris, all are ossified.

No urodele amphibian has more than four digits in the manus, and the number may be reduced to three, or even two (Amphiuma didactylum). When four digits are present the number of the phalanges is usually 2, 2, 3, 2.

Among the Anura, Dugès and Gegenbaur have shown that Bom-binator and Pelobates have eight distinct carpal bones—two in the proximal row (radiale, intermedium-ulnare), five in the distal, and one between these two rows. This last, which is the centrale, lies on the radial side of the manus, and articulates with the three radial distalia, much as the navicular hone articulates with the three cuneiformia in the mammalian tarsus. In Rana esculenta, there are also two bones in the proximal row, and the centrale lies on the radial side of the carpus. But there are only three bone3 in the distal row; one large, on the ulnar side, which bears the third, fourth, and fifth metacarpals, and two small ossicles on the radial side, which articulate with the first and second metacarpals.

There are five digits in the manus of the Anura; but the pollex is rudimentary, being represented only by a cartilaginous or more or less ossified style. The second and third digits usually have two phalanges each, and the fourth and fifth, three (2, 2, 3, 3).

The pectoral arch of the Labyrinthodonts is best known in Archegosaurus, where it presents three ossified elements, which probably answer to the coracoid, praecoracoid, and scapula. The bones of the fore-limb in the Labyrinthodonts are always weak relatively to the size of the body. There appear to have been five digits, the carpus remaining unossified.

In Proteus, Menobranchus, and Amphiuma, the pelvic arch is not connected with any distinctly modified sacral vertebra, and the ilium is very small. The pubes and ischia are represented by broad cartilaginous plates, which unite, and may become fused together in the middle line.

In Menobranchus, the pubic portion of the pelvis is continued forwards into a broad triangular median process. In Siredon, Menopoma, Cryptobranchus, and the Salamanders, there is a similar median process, reminding one of the omosternum in the pectoral arch of the Anura. It becomes bifurcated anteriorly. The ilium is always ossified; and there are ischial ossifications in all but Proteτιs. On the other hand, the pubic region always remains cartilaginous in the Urodela.

Hyrtl has shown that Cryptobranchus has no proper knee-joint, the femur being united with the tibia and fibula by a solid fibrous mass; and that, in Menopoma, the cavity of the knee-joint is very small.

The tibia and fibula in the Urodela arc always separate, and the proximal elements.' of the tarsus are not elongated. The greatest number of tarsal elements is found in Cryptobranchus and Menopoma, which, according to Hyrtl, have three cartilages in the proximal, and five in the distal, row, while two are central. In Siredon, the tarsus completely resembles the carpus, but there is one more distal cartilage. The tarsus therefore consists of three proximal cartilages (tibiale, intermedium, fibulare), one central (centrale), and five distal (distalia). In the Salamanders, there is usually the same number and disposition of the tarsal cartilages; but more or fewer are ossified, and it is interesting to remark that the two fibular distalia sometimes become united into a “cuboid.”

Menobranchus has two (or three) proximal, one central, and three distal tarsal cartilages; Amphiuma, three proximal and two distal; Proteus, two cartilages on the fibular, and one on the tibial side.

Siredon, Cryptobranchus, Menopoma, and most Salamanders have five digits in the pes; Menobranchus, four; Amphiuma, three; and Proteus, two. The number of the phalanges in the pentadactyle foot is usually 2, 2, 3, 3, 2. In Siredon, Hyrtl found 1, 2, 3, 4, 2.

In the Anura, the ilium is greatly elongated, and the pubes and ischia are flattened, discoidal, and applied together by their inner surfaces. The ilium and the ischium, alone, become completely ossified, and there is no praepubic process.

The tibia and fibula coalesce into one bone. Two elongated bones form a proximal row in the tarsus, and are commonly united by their epiphysial ends (e.g., Rana; they remain separate in Bom-binator). In Rana esculenta, the distal confluent ends of these bones (which possibly answer to the astragalus and calcaneum) present a transversely elongated articular surface, which is convex from the dorsal to the plantar side. Between this and the proximal end of the second and third metatarsals lies a discoidal, more or less calcified, cartilage. The convex distal face of this cartilage articulates with these two metatarsals. From its fibular side a strong ligamentous band passes to the proximal end of the fifth metatarsal, and a fibrous plate to the fibular and plantar edge of the fourth metatarsal, so that the band and plate are interposed between these metatarsals and the coalesced astragalus and calcaneum. On the tibial side of the discoidal cartilage lies another, which is elongated from the dorsal to the plantar side, and concave proximally, to articulate with the tibial side of the distal end of the coalesced astragalus and calcaneum The inner or tibial face of this cartilage articulates with the proximal end of the elongated first joint of the calcar. Its distal end is connected by a strong band of ligamentous fibres, within which a nodule of cartilage may be enclosed, with the proximal ends of the first and second metatarsals. The second joint of the calcar has the form of an ungual phalanx.

In Bufo and Bombinator, according to Gegenbaur, the calcar consists of only a single piece.

The pelvic arch of the Labyrinthodonts appears to have contained a well-ossified pubic element, in which respect it differs from that of all other Amphibia. The hind-limb, like the fore-limb, was relatively weak. The tibia and fibula are distinct. In the few cases in which the pes is preserved it is pentadactyle, with a short cartilaginous tarsus.

The Integumentary Organs.— In all recent Amphibia, the integument is remarkable for the great abundance of simple follicular glands which are distributed through it, and are sometimes all of one kind (e-.g., Proteus), though in other cases two sorts of such glands can be distinguished (Rana). In many Anura and Urodela, these glandular structures attain a greater complication of structure, especially near the angle of the jaw, and constitute what are termed the “parotoid” glands. In some cases, the secretion of these glands is extremely acrid and irritating. In some Urodela (Proteus and Siredon), and in the tadpole, the epidermis becomes, modified in relation with the termination of sensory nerves, in the head and along the body, in the region of the nerve of the lateral line, and gives rise to sensory organs of the same nature as those which are found in the lateral line and the so-called mucous sacs and canals of fishes.

In a few Anura, ossification takes place in the dorsal integument, and this process may go so far as to give rise to bony plates, which may become closely connected with the spines of the subjacent vertebrae (Brachycephalus, Ceratophrys). In the majority of the Peromela, oval, cycloid scales are imbedded in the transverse folds of the integument, and constitute another point of resemblance between the members of this group and the Labyrinthodonts. But the rows of scales are not confined to the ventral surface, and the scales themselves differ in structure from those of the Labyrinthodonts.

In the Urodela and Anura, the epidermis is periodically exuviated.

The Alimentary Organs. —The teeth of the recent Amphibia vary a good deal in form. In the Urodela, they are usually conical and pointed; frequently more or less curved; sometimes, as in Anaides, lancet-shaped. Siren has the surfaces of the vomers and palatines covered with parallel series of small dents en brosse. In Ceratophrys, the bases of the teeth are slightly grooved longitudinally. In Archegosaurus, similar grooves are more marked, and give rise to folds of the wall of the tooth. These, extending inwards and ramifying, give rise to the complicated or “labyrinthic” structure exhibited by transverse sections of the teeth of the typical Labyrinthodonts. Very generally, the teeth become ankylosed with the subjacent bones, and are replaced by others developed at their bases. In the Labyrinthodonts, some of the anterior teeth frequently become much larger than the rest. The Anura are remarkable for the total absence of teeth in the mandibles, in all but one or two genera, while many have no premaxillary or maxillary teeth. The Toads have no teeth in the upper jaw. Pipa is altogether edentulous. Siren alone presents plates of horn upon the gingivel surfaces of the premaxillae and of the dentary elements of the mandible.

Teeth may be developed upon the premaxillae and maxillae, the palatines, and the dentary and the splenial elements of the mandible; but they do not occur elsewhere,—the so-called sphenoidal teeth of some Salamanders being really borne, as has been seen, on the peculiarly modified palatines.

The buccal cavity is usually spacious, and the widely-separated posterior nares open into the anterior part of it. In the lower Urodela, the branchial clefts lie at the sides of the pharynx, and the median aperture of the glottis is situated far back. In the Urodela, and some Anura, there are no Eustachian passages; but, in most Anura, these passages have the form of wide recesses leading out of the pharynx. In Pipa and Dactyleihra alone, the “recesses” are converted into Eustachian “tubes,” which open by a common median aperrare; tms is relatively wider in Dadylethra. Two grooves in the mucous membrane of the roof of the mouth pass from the Eustachian to the posterior nasal apertures, and enclose a lyrate space, in these genera.

The tongue is rudimentary in .the lower Urodela; but, in the Salamanders, it may be free, fleshy, and even mushroom-shaped. In Pipa and Dactylethra, no trace of a tongue is to be observed. In Rana, as in most Anura, the anterior end of the tongue is comparatively small and little elevated above the mucous membrane of the floor of the mouth, but the posterior end is produced into a free fleshy mass, bifurcated at its extremity. It is this free end which is thrown forward, in the act of prehension, the tongue turning on its anterior end as on a hinge. Rhinophrynus is the only Anuran in which the anterior end of the tongue alone is free.

In the males of many Anura the mucous membrane of the mouth is produced outwards, on each side, between the mandible and the hyoid, into a sac, which becomes filled with air, and gives rise to a conspicuous projection of the integument of the throat. In some cases these two sacs coalesce into one.

Salivary glands have not been discovered in any Amphibia.

Except in the Peromela, the gullet is short. It passes into an elongated stomach, the long axis of which coincides with that of the body in the Urodela and Peromela, but becomes oblique, or transverse, in the Amira. The intestine is never very long, and, consequently, its convolutions are few and simple. There is always a marked distinction between the small and the large intestine. The latter opens into a cloaca, which receives the ducts of the urinary and genital apparatus. The stomach and intestine are enclosed in peritoneum, and suspended to the roof of the abdominal cavity by a mesenteric fold. The liver is always provided with a gall-bladder. It is distinctly bilobed in most Anura; and, in Pipa and Dactylethra, the two lobes are completely separate, the gall-bladder being attached to the right lobe. In the Peromela, the liver has an exceptional form, being divided into a great number of small lobes, arranged in a longitudinal series so as to overlap one another.

A pancreas is always present; but sometimes, as in Rana, it is small, and its glandular substance surrounds the hepatic duct. The spleen, enclosed in the mesentery, is elongated in the Urodela and Peromela, rounded in the Anura. The. Organs of Circulation.— The heart is contained within a pericardium, the walls of which generally exhibit numerous scat- z. s .v.e. tered pigment z ∙4ς *tK7∙ - ∖

cells, and though whole apparatus 2⅛, >7 t ⅛

enclosed within Fl015 _ The heart of s, re <ι on m erιcanus. Lateral view of th© pericardium, the heart contained within the pericardium, the left βvΛi>nt th ο von<n κa * 1 of the sinu8 venosus and of the auricles being except tne ven<*i removed; <S. sinus venosus; I. v. c. inferior vena cava; Cavae) presents a £· *∙ v ∙ c ∙ superior vena cava; L.A. left auricle; . ' , r . 2 , R. A. right auricle; Spt. septum auriculorum; V.

Senes 01 live ventricie; T. a. truncus arteriosus; 1, 2, 3, 4, the søcrmpnfq to aortic arches. The arrow traverses the sinu-auricular ιu aperture. The auriculo-ventricular aperture lies to which, enumera-the right of the arch formed by the free edge of the ting them from scptum '

behind forwards, the following terms may be applied:—1, The sinus venosus ; 2, the atrium; 3, the ventriculus; 4, the pylangium (from 7rυλωv, a gateway, and αγγεtov, a vessel); and 5, the synangium. Atrium here denotes the auricular division of the heart, comprising the right and left auricles. Pylangium and synangium, together, are the equivalents of that portion of the heart which lies between the ventricle and the anterior wall of the pericardium, and which has been variously named bulbus, cavus, and truncus, arteriosus.

These five segments of the heart are so arranged, that the sinus and atrium lie on the dorsal and posterior aspect of the organ, while the others occupy its ventral and anterior region. Viewed sideways, in fact, the heart has the shape of a of winch the sinus and atrium occupy the upper, and the other segments the lower half. But it also always presents, more or less, a lateral flexure, between its anterior and posterior points of adherence to the middle line of the pericardium; so that, viewed from above, it approximates the form of an N, of which the right half is represented by the synangium, pylangium, and ventricle, and the left half by the atrium and sinus. The pylangium, in fact, always arises from the right side of the ventricle, while a large part of the atrium and of the sinus very often lies to the left of the ventricle, the auriculo - ventricular aperture of the ventricle looking to the left side and forwards.

There is an interesting difference to be observed in the relative position of these segments of the heart in the lower and the higher Amphibia. In Siredon, for example, the greater part of the sinus lies completely behind the ventricle, and the sinu-auricular aperture is situated on the posterior face of the atrium, on a level with the posterior part of the ventricle; but, in the Frogs, the sinus lies altogether above the ventricle, without sensibly projecting behind it, and the sinu-auricular opening lies in the dorsal face of the atrium, in front of the level of the auriculo-ventricular aperture. In other words, the segments of the heart have a less marked vertical flexure in the lower, than in the higher Amphibia, and more nearly approach the condition of the embryonic heart. In correspondence with this, the superior cavae traverse the pericardium to enter the sinus near its posterior end in Siredon, but about its middle in the Frog.

The sinus venosus is a thin walled sac, which is relatively largest in the lower Amphibia and smallest in the Frogs. Anteriorly, it usually receives, on each side, one of the two superior venae cavae; posteriorly, the single vena cava inferior opens into it. But, in some cases (as in Meno-branchus and Pipa, according to Meyer) the inferior vena cava divides into two branches, each of which coalesces with the superior vena cava of its sidebefore opening into the sinus. The superior cavae may open into the sinus immedi- 1 ately after they have traversed the pericardium, as in Meno-branchus; or, they may be short trunks, as in the Frogs; or, as in Siredon and the Salamanders, the right cava may be long and the left short.

The sinu-auricular aperture, by which the sinus and the right auricle communicate, is small, relatively to the size of these two cavities, and has an oval form. Its lips may be slightly prolonged towards the cavity of the auricle, but do not give rise to very definite sinu - auricular valves.

The auricular segment of the heart, or atrium, is always more or less bi-lobed, the truncus arteriosus being embraced by the two lobes, one of which projects on its right side and the other on the left. The right and left lobes are equal in Proteus; both lobes are large and sacculated in Siren; in Menobranchus, the left lobe seems to be large, in Epicrium, the right; but many of these differences are probably accidental. In the Bullfrog (Rana pipievs) the two lobes of the atrial segment of the heart completely envelop the truncus arteriosus, and become united together by fibrous tissue, which connects their walls on the ventral side of the truncus. The atrium is usually divided into two cavities, of which the left is smaller than the right, by a septum, which extends from the left wall of the atrium towards the auriculo-ventricular aperture. The cavity of the auricular segment thuβ becomes divided into a smaller, left, auricle, which lies behind and to the left of the septum, and a larger, right, auricle, to the right and in front of the septum. In the Frogs, the septum auricularum is a complete partition, containing muscular fibres, and the septal branches, with their ganglia, of the cardiac nerves of tho pneumogastric.

It divides the auriculo - ventricu - lar aperture, passing from one auriculo - ventricular valve to the other, and ending between them by a free edge, which might almost be said to lie in the cavity of the ventricle (Fig. 23).

In Lissotriton punctatus, and in Siredon, the septum, still complete, ends in the cavity of the auricular segment by a free edge, which arches over the auriculo-ventricular aperture. In Menobranchus, the septum is reduced to little more than a wide-meshed network of branched muscular bands, and, in Proteus, the existence of a septum is doubtful.

The auriculo-ventricular aperture is always situated at the left side of the posterior end of the auricular segment, where the latter joins the ventricle. In Rana esculenta and pipiens it possesses distinct, though short, membraneous valves, the free edges of which, directed towards the ventricular cavity, are kept down by fine tendinous filaments.

The common trunk, formed by the union of the two pulmonary veins, runs over the dorsal wall of the sinus venosus, passes between the two superior cavae, and, usually dilating, opens into the cavity of the left auricle, close to the sinu-auricular aperture; and, in fact, separated from it only by the septum, which continues the direction of the right wall of the pulmonary vein.

The ventricular segment always has thick walls and a comparatively small cavity, which lies in the anterior half or base of the ventricle, and takes a direction from left to right, or from the auriculo-ventricular aperture to that of the truncus arteriosus. In consequence of the loose and spongy texture of the greater part of the thickness of the ventricular wall, it must be recollected that its apparent cavity by no means represents its capacity.

The truncus arteriosus of Menobranchus is subcylindrical, in that half which is nearest the ventricle, but, in the other half, has a dilated and ovoid form. The latter, in reality, consists of the origins of the aortic arches, closely united together (synangium), while the former subdivision is the gateway between the ventricle and the great vessels, or the pylangium. It presents two transverse rows of semilunar valves, three in each row; the lower or posterior row being close to the opening of communication between the pylan-gium and the ventricle, while the other row is near the anterior end of the pylangium.

In Siredon (Figs. 15 and 16) there is the same division into a pylangium proper and a large oval bulb-like synangium, formed by the united aortic arches. Three valves in a transverse row are situated at each end of the pylangium. An oblique ridge projects from the dorsal wall of the pylangium, beginning low on the left side, and gradually increasing in size, until it passes into the dorsal valve of the anterior row. There is a small space in front of the anterior row of valves, into which projects the posterior free end of an oblique, but nearly horizontal septum, which divides the cavity of the synangium. From this thick partition thinner septa radiate to the walls of the synangium, which they thus divide into five longitudinal canals, of which that which lies to the right is twice as large as any of the others. In fact, it also becomes subdivided, further forwards, by a longitudinal septum, and then there are six canals answering to the six aortic arches which spring from the synangium, where it reaches the anterior end of the pericardium. According to Hyrtl’s account, the pylan-gium of Cryptobranchus has a very similar structure; but the synangium is completely split into two trunks, each of which contains three canals.

This leads to the structure of the truncus arteriosus observed in the Frogs, which consists almost wholly of the pylangium. Three thick semilunar valves are placed at the ventricular end of this region, and three others, also of unequal dimensions, at its synangial end. A longitudinal ridge, with a rounded, free, ventral edge, projects from the dorsal wall of the pylangium. It is thicker anteriorly than posteriorly, and is directed obliquely, so that its anterior end passes into the right anterior valve, while its posterior extremity is close to the left posterior valve. The anterior valves of the pylangium (v 2 ) are much larger than the posterior valves; and, of the three anterior valves, that which lies on the dorsal side is the smallest. Immediately beyond it is situated the aperture (p), which leads into the pulmonary trunks. In front of the pulmonary aperture is a wide cavity, whence the two great aortic trunks (Ao, Ao v ) spring. A tongue-like projection springs from the dorsal wall, and divides the cavity imperfectly. On the ventral side of the base of this tongue are the two openings (Ca) which lead into the carotid trunks. The three trunks—carotid, aortic, and pulmonary—pass out of the pericardium together, so closely united that they appear one. It is only at some distance beyond the pericardium that they separate,—the anterior ending in the rete mirabile, which has received the name of the “carotid gland;” the middle becoming the arch of the aorta; the posterior, the pulmo-cutaneous artery.

In the Peromela (e.g. Epier turn), the heart presents many singular peculiarities (Fig. 24). In the first place, it is moved back to a distance which is relatively far greater than in any other Amphibia and in most lizards. Next, it is extremely elongated, and the truncus arteriosus is relatively more prolonged than any other part of the heart. But the relative proportions of the pylangium and Synangium are the reverse of those which obtain in the Anura. Tho two transverse rows of valves which mark the boundaries of the pylangium are situated close to one another, near the origin of the truncus, all the rest of which is made up of the synangium. A longitudinal partition, at first, divides the cavity of the synangium into two unequal passages; but, towards its anterior end, it contains four equal canals. Having reached the anterior extremity of the pericardium, the synangium divides, and the two pairs of canals become independent, but closely united, trunks, which run, on each side of the trachea, to about the level of the glottis. Here the two trunks join, and pass into the single arch of the aorta, which turns sharply back beneath the vertebral column. The carotid artery is given off from the junction of the two trunks with the single dorsal aortic arch. Shortly before the two trunks join, that upon the dorsal side gives off the pulmonary artery. A single pulmonary vein opens into the left auricle; and it is worthy of notice, that the auricles and sinus are situated as far forwards on the dorsal aspect of the heart as in the Frogs.

As regards the number and destination of the great vessels which arise from the synangium, great differences obtain in the different groups of the Amphibia.

In the perennibranchiate Urodela, each of the three, or four, branchial arches has its appropriate aortic trunk, which springs mediately, or immediately, from, the synangium. The three anterior aortic trunks supply the gills, but are not wholly distributed to them; so that the trunks which unite to form the dorsal aorta are derived partly from the gills and partly come directly from the ventral aorta. The anterior aortic arch gives off, on its ventral side, a hyomandibular artery to the walls of the oral cavity, which appears to represent the remains of the hyoidean and mandibular aortic arches, while, dorsally, it supplies the internal carotid. The pulmonary artery is given off from the fourth aortic arch, or from the common trunk, which is formed by the union of this with those which precede it.

In Cryptobranchus, according to Hyrtl, three trunks are given off on each side from the synangium. The most anterior corresponds with the hyomandibular artery of the perennibranchiate forms. The second belongs to the first branchial arch. It gives off no branch, but unites with the third and largest vessel to form a common trunk, which unites with its fellow beneath the vertebral column, and gives rise to the dorsal aorta. The posterior aortic arch gives off the pulmonary artery (which supplies a branch to the alimentary canal). From the common trunk a maxillary and an internal carotid artery are supplied; while a third branch passes to the ventral side of the atlas, and, turning backwards, passes between the transverse process of the second and succeeding vertebrae as a collateral vertebral artery.

In Salamamdra, there are four aortic arches. The most anterior of these belongs to the first branchial arch. It gives off a hyomandibular branch, then breaks up into a rete mirabile, whence the internal carotid artery is continued, and is connected by a mere ductus Botalli with the second arch. The second and third arches give off no branches; but, along with the slender ductus Botalli of the fourth arch, coalesce into the trunk which unites with its fellow to form the dorsal aorta. The fourth arch gives off the pulmonary artery, and a smaller dorsal cutaneous branch (Hyrtl). The pulmonary artery gives twigs to the stomach.

It is clear that the posterior trunk of Cryptobranchus represents the second, third, and fourth aortic arches of Salamandra; and that the first aortic arch of Salamandra answers to the first and second trunks which spring from the synangium of Cryptobranchus.

In the Anura there are apparently only two aortic arches; but, as has already been observed, each of them is divided into three canals. The anterior canal ends in a rete mirabile, whence the internal carotid artery proceeds, and it gives off the hyomandibular or lingual artery. It therefore answers to the first arch of the Salamanders. The second or middle canal is the largest, and passes into a trunk which runs along the sides of the gullet; and curving backwards, unites with that of the opposite side in the dorsal aorta. The third canal ends in a trunk which divides into the pulmonary and the great cutaneous arteries, which latter is distributed to the dorsal integument. It answers to the third and fourth arches in the Salamanders.

In the Peromela {Epicrium) the two aortic trunks which spring from the truncus arteriosus would seem to correspond with the second and third of the frog, the first having become absorbed into the second. This is a point which can be cleared up satisfactorily only by the study of development; but it is obvious that the heart and its arches have undergone greater changes in this group than in any of the others.

With respect to the venous system, it is worthy of notice that the blood returning, from the hinder part of the body and the posterior extremities is, in part, carried to the kidneys, and in part poured into a vein which runs in the anterior wall of the abdominal cavity,—the anterior abdominal vein. Of the branches in which this vein terminates anteriorly, one communicates with the portal vein, and one is distributed to the liver directly. In the Anura, venous radicles in the integument covering the back of the head and shoulders, unite to form a great cutaneous vein, which passes backwards, perforates the external oblique muscle, and then turning abruptly forwards, ends in the subclavian vein. This vein carries away a large part of the blood of the cutaneous artery, which accompanies it in a great part of its course.

The lymphatic system has been most carefully studied in the frog, where it consists of ( 1. ) widely-distributed lymphatic capillaries, and sinuses which ensheath the blood-vessels; (2.) subcutaneous lymph sacs; (3.) a large subvertebral cisterna, enclosed between the diverging lamellae of the mesentery, and placed in communication with the peritoneal cavity by minute openings or stomata; (4.) four lymph hearts, two situated close to the transverse process of the third vertebra and two at the sides of the coccygeal style. These hearts pump the lymph into the adjacent veins. As the two pairs of lymph hearts have been discovered in Triton and Salamandra as well as in Rana, it is probable that they are present in the TJrodela generally. No Amphibia possess lymphatic glands.

The Thymus gland in the Urodela lies behind the angle of the mandible (2VΛon, Salamandra), or close to the dorsal ends of the branchial arches {Proteus, Menobranchus, Siredon, Amphiuma, Menopoma'). In the Peromela it has the same position as in the abranchiate Urodela. In the tadpole the thymus occupies a place similar to that which it possesses in the branchiate Urodela. In the adult frog it is to be found just behind the suspensorium. The Thyroid gland, usually double, but single (according to Leydig) in Proteus, always lies in the immediate vicinity of tho lingual vessels. The Respiratory Organs.— The glottis in the Amphibia is situated in the middle line of the floor of the pharynx. In the perennibranchiate Urodela, it is a very small longitudinal slit leading into a narrow passage, which widens into a chamber into which the elongated pulmonary saca open. The Urodela and the Peromela present mere cartilaginous rudiments of a larynx; but, in the Anura, this structure attains a great development, and becomes the instrument of the powerful voice with which many of these animals are provided. The larynx is lodged in the angle between the two thyro-hyals, with which it is closely connected. The chief part of the larynx is an annular cricoid cartilage, with which two arytaenoid cartilages are articulated. Membranous folds, or freely projecting cartilaginous processes of the arytaenoid cartilages {Pipd), play the part of vocal ligaments. In Pipa the larynx is extensively ossified. In Proteus, the lungs are long tubes, dilated at their posterior blind ends, and fixed to the dorsal walls of the abdominal cavity by folds of the peritoneum. In Triton they are somewhat wider sacs, but, in both, the inner surfaces of the pulmonary sacs are smooth. In Siren and Salamandra, the walls of the sacs become cellular, and in Amphiuma, Menopoma, Cryptobranchus, and the Anura, the cellulation acquires a considerable development.

In Amphiuma, Menopoma, Cryptobranchus, and in the Peromela, there is a distinct trachea, which is of great length in the Peromela. In Pipa and Dactylethra there is no trachea, but each lung is connected with the laryngeal cavity by a bronchus.

The R,enal Organs.— The kidney is a more or less elongated organ—longer in the Urodela and Peromela, shorter in the Anura— which lies on each side of the vertebral column, its posterior end being close to, or even extending back on the dorsal side of, the cloaca.

In the female the efferent ducts of each kidney unite into a longer or shorter common trunk, which appears always to open into the cloaca by an aperture distinct from that of the oviduct, though the contrary statement is very generally received. In Rana esculenta, there can be no doubt as to the distinctness of the minute urinary apertures from the large and conspicuous oviducal openings, close to which they are situated. Hyrtl says of Crypto~ branchus —“Ureter .... super latera cloacae descendens in collum allantoidis exoneratur” {op. cit., p. 84).

In the male Amphibia, on the other hand, there is a longer or shorter duct common to both the renal and the genital products, which opens into the cloaca. In the Urodela, the duct is continued forwards along the outer side of the kidney to the anterior end of the ab-dominai cavity, and clearly represents the Wolffian duct of the embryo. Both the urinary tubuli and the vasa efferentia of the testis open into this duct. In Cryptobranchus the kidney is divided by a constriction into two portions—a slender, anterior, and a much thicker and longer, posterior, division. From the latter the efferent urinary canals proceed, and, curving outwards and backwards, join the posterior part of the Wolffian duct. The former is traversed by the vasa efferentia of the testes, which pass from its outer edge to the anterior portion of the Wolffian duct, so that it resembles an epididymis.

In Proteus, according to Leydig, the anterior end of the Wolffian duct is infundibuliform and open; the vasa efferentia of the testes open into the anterior moiety of the duct, the renal ducts, into its posterior moiety. The numerous arcuated renal ducts of the Salamanders and Tritons unite together, and open into the Wolffian duct near its cloacal end. The Wolffian duct persists in Bombinatσr igneus and Disco-glossus pictus ; but, in most Anura, it becomes obliterated for the greater part of its extent, and the same canals serve to convey both the urinary and the spermatic fluids to the persistent cloacal end of the Wolffian duct, which ordinarily receives the name of ureter. The urinary bladder is always large, and is often bifurcated anteriorly.

The Nervous System.— The amphibian brain is remarkable for the rudimentary condition of the cerebellum, which has the form of a mere band arching over the anterior part of the fourth ventricle. The mesencephalon is divided above, more or less distinctly, into two optic lobes. The cerebral hemispheres are always relatively large, subcylin-drical in the Urodela, but wider behind than in front in the Anura, and they are generally closely united together by their inner faces.

Ten pairs of cranial nerves are always found-·—viz., 1, The olfactory; 2, optic; 3, oculomotor; 4, pathetic; 5, trigeminal; 6, abducens; 7, portio dura; 8, auditory; 9, glossopharyngeal; 10, pneumogastric. The hypoglossal is always an extra-cranial nerve.

1. The olfactory is usually a rounded cord, not dilated at its anterior end. Fischer has observed it to arise by two roots in Pipa.

2. The optic nerves are attached, as usual, to the floor of the thalamencephalon. Fischer found no chiasma in Siredon or Menobranchus. Dr Humphrey found none in Cryptobranchus; but sections of the brain are needful before the actual absence of the chiasma can be considered to be satisfactorily proved.

3. The oculomotor nerve remains distinct from the trigeminal in most Amphibia, but its branch to the superior rectus muscle appears to coalesce with the orbito-nasal division of the fifth in Salamandra terrestris (Fischer).

4. The pathetic nerve remains distinct in Siredon and Cryptobranchus, and in the Anura; but in Salamandra terrestris, Fischer found that the superior oblique muscle was supplied by a branch from the orbito-nasal, with which, therefore, the pathetic had probably coalesced.

5. The trigeminal gives rise, as usual, to a Gasserian ganglion; and this ganglion remains distinct from that of the seventh nerve in all the Urodela, though united with it by a commissural band, which appears to answer to the nervus petrosus superficialis minor of the higher Vertebrata. In the Anura, on the contrary, the two ganglia are closely approximated {Pelobates, Bombinator) , or confounded together {Rana, Hyla, Bufo) in the adult, though they are distinct in the tadpole. The orbito-nasal, or first division of the trigeminal, is always separated from the second and third divisions by the ascending process of the suspensorium, when this structure is present. It supplies the tentacles of the Peromela. In the tadpole, and in some Urodela, a cutaneous branch to the dorsum of the head is given off from the fifth.

6. The abducens is distinct from the trigeminal in Salamandra and Bufo, but coalesces with the Gasserian ganglion in Rana, Pipa, and most Anura.

7 and 8. The portio dura and portio mollis arise by a common trunk, from which the portio dura soon separates, and either forms a distinct ganglion, as in the Urodela and Peromela, or fuses with the trigeminal·

9. The ganglion of the glossopharyngeal nerve appears to coalesce with that of the vagus, and the roots of the two nerves pass out of the same foramen in all the Amphibia except Siren, where, according to Fischer {op. cΛ.,p.l47),the nerve leaves the skull by a distinct aperture, close in front of that for the pneumogastric, and forms a ganglion of its own.

10. The vagus or pneumogastric, in the perennibranchiafe Amphibia, supplies the second and third branchia, and the cucullaris muscle; gives off cutaneous, laryngeal, cardiac, pulmonic, and gastric branches, and sometimes as many as three cutaneous branches, one of which runs along the junction of the dorsal and ventral muscles to the binder part of the body. These lateral nerves of the pneumogastric exist also in Menopoma, Amphiuma, and Triton, and in tadpoles; but appear to be absent in Salamandra terrestris and in the adult Anura (Fischer, Z.c.) These, however, possess a cutaneous branch of the vagus, which accompanies the cutaneous branch of the pulmo-cutaneous artery, and is distributed more or less widely to the dorsal integument of the head and trunk.

Fischer considers that a fine nerve, arising lower down than the vagus, and distributed to the abductors of tho head in Pipa, is to be regarded as an accessorius. But, seeing that, in the Amphibia generally, the motor nerves of the larynx, and, where a cucullaris exists, the nerves of that muscle also, are supplied from the pneumogastric, tho question of the presence or absence of an accessorius seems to reduce itself to this: Does the pneumogastric receive nerve fibres arising from the sides of the medulla oblongata and spinal cord between the roots of the spinal nerves? And, as it certainly does not, the accessorius, as it exists in the higher Vertebrata, must be admitted to be absent in Amphibia.

In most Amphibia, the first cervical nerve has the distribution of the hypoglossal; in Jfenobranchus, however, the corresponding nervous supply is furnished by the second and third cervical nerves,—the first spinal nerve, in this genus, perforating the sides of the body of the atlas, and being distributed to a muscle which passes from th⅛ vertebra to the occiput (Fischer, l.c., p. 158). In Pipa the hypoglossal is furnished by the second cervical nerve; in Salamandra, by the first and second. There is no trace of any suboccipital nerve in the Amphibia; and as, in the absence of this nerve, the first spinal would appear to answer to the second cervical of the higher Vertebrata, the fact that it takes the place of the hypoglossal becomes very perplexing.

In the Anura (Rana) the sympathetic is represented by a double chain of ganglia, situated at the sides of the aorta, and receiving branches from the anterior divisions of the spinal nerves. It appears to be continued in the skull by commissural cords which pass forwards on the inner side of the auditory capsule, and connect the ganglion of the vagus with that of the trigeminal.

The Organs of the Higher Senses.— The nasal sacs are elongated in Proteus, Menobranchus, and Siren, and not covered by nasal bones or alinasal cartilages. In the other Amphibia they are broader, and enclosed by cranial cartilages and ossifications. The olfactory mucous membrane is variously folded; and, in Rana, some of these folds are supported by ingrowths of the anterior cartilaginous wall of the nasal chamber.

In Proteus the eye is completely hidden by the continuation of the unaltered integument over it, and the organ of vision is almost as much obscured in the Peromela. In the other perennibranchiate TJrodela, and in Pipa, the integument covering the eye forms a transparent cornea, but there are no eyelids. The abranchiate Urodela have an upper and a lower lid; and, in the higher Anura, the lower lid becomes transparent, and is usually regarded as a membrana nictitans, as it is provided with a peculiar motor apparatus. In the Anura, the eye possesses not only the ordinary four recti muscles and the two obliqui, but there is a retractor bulbi. The Frogs and probably other Anura, possess a Harderian gland; but no lachrymal gland has been observed. The sclerotic may be chondrified, but it is not ossified. There is no pecten.

With regard to the organ of hearing, the membranous labyrinth is enclosed between the pro-otic bone, in front, and the representatives of the opisthotic and epiotic (usually confounded with the exoccipital), behind. The fenestra ovalis always occupies a space in the line of junction of the pro-otic with the posterior ossification, whether it be occupied by a broad unossified space, as in Menopoma, or the two bones be ankylosed together, as in Siren, Triton, and old Frogs. The stapes is more or less ossified, and its outer face is frequently provided with a styliform appendage, in the Urodela. In the Urodela (which have no tympanic cavity), a ligament passes from the stapes to the suspensorium, and there is no columella auris. The like absence of columella auris and of a tympanum obtains in several Anura. Dugès states that the columella is wanting in Bombinator and Pelobates (Recherches, p. 41), and the absence of the columella auris, as of the tympanum and Eustachian tubes, has since been noticed in Telmatobius, Phryniscus, Atelopus varius, and Brachycephalus ephippium (Stannius, op. cit., p. 61). In the higher Anura, there is a complete tympanum, with Eustachian tubes, and a columella auris, which extends from the stapes to the membrana tympani. The tympanic membrane is either quite similar to the rest of the integument or markedly different from it. In some genera (e.g. Rana), the tympanic membrane is set in a frame of cartilage. In Pipa and Dactyleihra, the Eustachian tubes, as already remarked, have a common opening, and the columella is very peculiar. Recent investigations make it probable that the Amphibia possess a rudimentary cochlea. Whether the opening, which in the Frogs has been described as a fenestra rotunda, is really of that nature, is doubtful.

The Reproductive Organs.— The ovana and testes are attached to the dorsal wall of the abdominal cavity, in the immediate vicinity of the kidneys, by the mesoarial and mesorchial folds of the peritoneum, which invest them.

The ovaria, when fully developed, become hollow, and in the Anura their internal cavities may be divided by septa.

The oviducts are long, usually more or less convoluted, tubes, which open posteriorly into the cloaca; while, anteriorly, their funnel-shaped apertures lie in the anterior part of the abdomen, sometimes, as in the Frogs, as far forward as the root of the lung. Their walls are glandular, and secrete a viscid substance which invests the ova in their passage down the oviduct.

In the male Urodela, the persistent Wolffian duct, already mentioned, occupies the position of the oviduct in the female, and the vasa efferentia, after traversing the kidney, open into it. This duct persists in Bombinator igneus and Discoglossus pictus ; but in the male Anura, in general, the greater part of it is obliterated, only so much remaining as plays the part of ureter and vas deferens. In the Urodela accessory glands open into the cloaca, and in Triton there is a rudimentary copulatory papilla. Some female Urodela are provided with receptacula seminis. In the terrestrial Salamanders and in the anurous Rhinoderma Gayi the young are developed within the dilated uterine terminations of the oviducts. In Pipa the eggs are deposited on the b⅛ck of the female, and the integument grows up round each, and encloses it in a cell, in which it undergoes its development. In some tree-frogs (Nototrema and Opis-thodelphys) the eggs are received into a sort of marsupial pouch formed by an up-growth of the margins of the dorsal integument, which, when complete, has a small posterior aperture. On the other hand, it is the male Alytes obstetricans which twists the strings of eggs laid by the female round his hind-legs, and, thus cross-gartered, retires into seclusion until the young are ready to be hatched, when he resorts to the water in which the tadpoles are to perform their further metamorphoses.

Development of the Amphibia.— The yelk of the ovum undergoes complete division, in which respect the Amphibia agree with the Pharyngobτanchii, Marsipol·ranchii, and Mammalia, and differ from other Vertebrata.; though it must be remembered that the process of yelk division in the Ganoidei and Dipnoi is not yet known.

Except in some viviparous species, the embryo, when hatched, is pisciform and apodal; and three pairs of external gills, which becomo more or less branched, are developed from the first, second, and third branchial arches. In the larval Triton a very singular elongated appendage makes its appearance on each side of the head, in front of the branchiae; and in the tadpole two eminences of the ventral integument, with glandular terminal faces, are developed—one on each side of the middle line, behind the mouth. The larvae of Dactylethra have two long tentacles attached near the an"le of the mouth. An opercular fold of the integument grows back from each hyoidean arch, and the two are connected by a free fold of the sub-gular integument. The limbs make their appearance as buds from the sides of the body, the anterior pair appearing first. The anterior limbs attain a considerable size before the posterior pair are developed in Triton; but, in the Frogs, the posterior limbs grow much faster than the anterior, which long remain inconspicuous and hidden.

In the Urodele larvae, teeth are very early developed in the premaxillary, maxillary, vomerine, palatine, splenial, and dentary regions; and, indeed, in Triton and Siredon, the teeth precede the corresponding bones, which arise by the ossification of the mucous membrane about the bases of the teeth; and there are no labial cartilages, and no horny labial papillae, or beak-like armature of the jaw. The abdomen is slender, in accordance with the brevity of the intestine, and the little animal is altogether carnivorous.

In the Anura, on the other hand, teeth are not developed until a later stage. A pair of cartilages appear in the roof of the mouth in front of the ends of the trabeculae (“rostraux supérieurs,' Dugès; “upper labials,” Parker), and another pair opposite them (“rostraux inférieurs," Dugès; “lower labials,” Parker); and the epithelium of the mucous membrane covering them becomes converted into an upper and a lower brown horny toothed plate, having seme resemblance to the beaks of a Chelonian. The curtain-like lips, which surround the oral aperture, are also beset with horny papillae, which call to mind the corneous teeth of the Marsipo-branchii. The abdomen is swollen and almost globular, and lodges a long and spirally-coiled intestine. The animal is herbivorous, though it does not despise animal food, even in the shape of the weaker members of its own family.

The space allotted to this article does not allow the details of the development of the Amphibia to be even sketched; but attention may be directed to one or two of the more important points.

The skull presents some singular differences in the course of its development in the Urodela (Triton, Siredon) and the Anura (Rana, Aly tes) respectively. In the former, the mandibular and trabecular arches become connected only at their dorsal ends, by the pedicle of the mandibular arch; the pterygoid arch is developed late; and the mandibular arch appears to give rise to no orbital process. In the latter, the mandibular and trabecular arches not only unite at their dorsal ends by the pedicle, but, at a very early period, the mandibular arch is united with the antorbital process of the trabecula; and the pterygoid grows pari passu with the subsequent divergence of the mandibular and the trabecular arches. A large orbital process is developed from the mandibular arch.

In the Urodela, the hyoidean and branchial apparatus consists, at first, of elongated cartilaginous hyoidean cornua, united with a median chondrification, which represents the basihyal and basi-branchial pieces, to which last two cerato-branchials are attached. The first cerato-branchial is continued dorsally into the first epibranchial, while the second cerato-branchial supports the other three epibranchials. As the development of the Triton proceeds, the hyoidean arch becomes connected with the suspensorium, and with the stapes, by ligament. The second basi-branchial ossifies, detaches itself from the first, and lies as a forked bone in front of the larynx; and only the two cerato-branchials, with the first epibranchial, remain—the rest of the branchial apparatus disappearing.

In the Anura, the hyoidean arches are, at an early period, very thick, and relatively short, and are articulated with the suspensoria. A relatively broad and short cartilage represents the basihyal and basibranchial, and at the sides of this are two very broad cartilages, which correspond with the two cerato-branchials, inasmuch as their dorsal edges bear the four epibranchial cartilages. As the tadpole grows older, the hyo-branchial apparatus becomes more like that of the Urodele larva, the hyoid arch elongating into a slender rod, and the two cerato-branchials becoming distinct. The basibranchial region of the median cartilage, which unites the cerato-branchials ventrally, becomes forked, and the processes which form the fork ossify and become the thyro-hyals, which therefore would seem to correspond with the os ypsiloides of the Urodela. Finally, the extreme dorsal end of the hyoidean arch detaches itself from the suspensorium, and enters into close union with the periotic capsule, from the outer wall of which the columella auris is developed. The Distribution of the Amphibia.— Darwin has pointed out (Origin of Species, p. 350) that Amphibia are met with on no islands but New Zealand, New Caledonia, the Andaman Islands, and perhaps the Solomon Islands and the Seychelles. “This general absence of frogs, toads, and newts in so many true oceanic islands cannot be accounted for by their physical conditions; indeed, it seems that these are peculiarly fitted for those animals, for frogs have been introduced into Madeira, the Azores, and Mauritius, and have multiplied so as to become a nuisance. But as these animals and their spawn are immediately killed (with the exception, so far as is known, of one Indian species) by sea-water, there would be great difficulty in their transportai across the sea, and therefore we can see why they do not exist in strictly oceanic islands.”

Leaving the oceanic islands aside, the distribution of the Am phibia is world-wide, but the different groups are very remarkably localised.

The Urodela, for example, are limited not only to the arctogaeal province, but to the temperate parts of that province; and, in curious correspondence with the Ganoid fishes, their headquarters are in North America. Siren, Menobranchus, Amphiuma, Menopoma, Dicamptodon, Heredia, Anaides, Desmognathus, Batrachoseps, Hemidactylus, and Plethodon are exclusively North American; and the majority of species of Amblystoma and Spelerpes appertain to that region,— Amblystoma being represented in North Asia, and Spelerpes in the circum-Mediterranean area. Triton alone is spread over the whole temperate arctogaeal area. Salamandra, Pleurodeles, Bradybates, Chioglossa, and Salamandrina are confined to Europe and North Africa. The singular Salamandra atra is limited to the Swiss and Austrian Alps, Proteus to Camiola and Carinthia. Four genera— Ellipsoglossa, Isodactyliurm, Onychodactylus, and Ranodon— are confined to North Asia; and Cryptobranchus, if it be a distinct genus, is limited to Japan.

If the distribution of the Urodela calls to mind that of the Ganoid fishes, that of the Peromela is rather comparable to the distribution of the Tapirs. Of the four genera, Siphonops and Rhinatrema are exclusively inhabitants of the hotter part of the Austro-Columbian province—as are the great number of the species of Caecilia; but the remaining species of that genus are East Indian, and Epicrium is confined to Java and Ceylon.

In strong contrast with the foregoing, the Anura are of worldwide distribution, being abundantly represented in all the great provinces. A great preponderance of the genera and species, however, are Austro-Columbian, the Anura having their headquarters in South America, as markedly as the Urodela have theirs in the northern division of that continent. North America, in fact, is poor in Anura, having only three peculiar genera, viz., Scaphiopus t Acris, and Pseudacris; while the rest of northern Arctogaea has five, viz., Pelodytes, Discoglossus, Alytes, Pelobates, and Bombinator.

The genus Rana itself, however, is characteristically arctogaeal, having only a single species in the Mexican border-land of Austro-Columbia, and none in Australia. Rana esculenta extends from France to China and Japan, and from North Europe to Tunis. Rana temporaria covers even a larger area, as it occurs in the British Islands and in North America, as well as in North Asia and Japan.

The Austro-Columbian region not only presents the greatest number of species, but among them are some of the most singular forms, such as Pseudis, Ceratophrys, Braxhycephalus, Rhinoderma, Engystoma, Otilophus, Nototrema, Opisthodelphys, Rhinophrynus, and Pipa; in which respect the South American Anura run parallel with the birds of the same region. And, as is seen in other cases, the nearest allies of many of these singular forms are to be found in Ultra-Saharal (Aethiopic) Africa, e.g., Hemisus (Brachycephalies), Breviceps (Engystoma), Dactylethra (Pipa). It is remarkable that Pseudophryne, which is closely allied with the Aethiopic Hemisus and the Austro-Columbian Brachycephalus; and Chelydobatrachus, which is similarly related to the Aethiopic Breviceps and the Austro-Columbian Engystoma, are Australian.

The Australian region is remarkable for the absence of the genera Rana and Bufo, which occur everywhere else; and for the occurrence of Cystignathus, which is an Austro-Columbian, North American, and Aethiopian form, but does not occur in India. If it were not for its tree-frogs, Australia would be poorer in Anura than Europe is.

These Ailùra, modified for arboreal life, or “tree-frogs,” are represented in all the distributional provinces—the genus Hyla having its chief seat in Austro-Columbia, and extending thence over North America, Europe, North Africa, Western and Eastern Asia, and Australia, but not into India or Ultra-Saharal Africa, in which other forms of the same group are met with.

The British Islands possess the following species of Amphibia; —Rana temporaria, Bufo vulgaris, B. calamita, Triton cristatus, T. Bibronii, IÄssotriton punctatus, L. palmipes. Geological Distribution.— No fossil Peromela are known. Anura occur in the Miocene deposits of France and Germany. The best preserved forms belong to the genera Palaeobatrachus and Latonia, and occur in the schists of (Eningen along with their tadpoles. They possess maxillary teeth, and present no important differences from existing Anura, except that, in Palaeobatrachus, the sacral vertebra has coalesced with the two preceding vertebrae, while, in existing forms, only one of the prae-sacral vertebrae is known to become confluent with the sacral. Urodela also occur in the same Miocene deposits. Of these the famous Andrias Scheuchzeri is very closely allied to Menopoma and Cryptobranchus, while other forms appear to be generically identical with Triton and Salamandra. The singular genus Orthophysis presents a good deal of resemblance to Proteus, but appears to have possessed no limbs.

The older Cainozoic and the uppeι∙ and middle Mesozoic formations have yielded no Amphibia. A doubtful form, Rhinosaurus, occurs in the Lias of Simbirsk. In the earliest Mesozoic deposits—the Trias,—and in the later Palaeozoic—the Permian and the Carboniferous formations, Amphibia occur, sometimes in great abundance. In the Trias, they have been found in greatest numbers in Germany, while the Carboniferous formations have furnished the largest supply in the British Islands, Germany, and North America. It is interesting to observe that the last-named region has recently yielded elongated apodal forms, allied to the Ophiderpeton of the Kilkenny coal measures.

Aetiology of the Amphibia.— In taking a general survey of the relations of the different great divisions of the Amphibia, the most striking fact is their singular distinctness and isolation from one another. None of the Peromela present the slightest indication of an approximation towards the Anura or the Urodela.

It may be suggested that the incompleteness of the jugal arch in Breυiceps, Pipa, and Dactylethra; the absence or rudimentary condition of the palatine bones in Breviceps, Bombinator, and Aly tes; the rudimentary condition of the tympanum, and the absence, or reduction to a rudiment, of the columella auris, in so many forms; the presence of rudimentary ribs attached to some of the anterior vertebra of Bombinator and Alytes; the presence of mandibular teeth in Hemiphractus and Grypiscus; and the peculiar spermatozoa of Bombinator, are so many indications of an approach towards the type of structure observed in the higher Urodela.

But, without underestimating the force of these considerations, it must be admitted that they count for very little, λvhen we take into consideration the fixity of the number of the vertebrae, and of the characters of the pelvis and of the limbs, in the Anura.

It is to be regretted that nothing is known of the development of any of the Peromela; of any of the Urodela, except Salamandra, Triton, and Siredon (Amblystoma) ; and of more than a few of the Anura. Among the lower forms of this division, the development of Alytes and Pelobates has been studied thoroughly by Vogt and Van Bambeke; and the more advanced conditions of the tadpole of Dactylethra are known. So far as these observations go, however, they tend to show that thelarvae of all the Anura possess the horny beak, which distinguishes them from those of the Urodela.

If we assume, as the fιmdamental similarities between the different divisions of the Amphibia lead us to do, that they have resulted from the modification of some one primitive form, the problem, at present seemingly insoluble, presents itself, whether these differences in structure and habit of the larvae of the Urodela and Anura indicate that the caudate ancestor of the Anura was already different from the ancestor of the Urodela, or whether they result from modifications which have taken place in the larvae of the Anura, since that group came into existence.

In view of this problem, Siren possesses a particular interest. Its horny jaw-sheaths might be compared to those of the Anuran tadpole, and it might be regarded as showing the way by which the Anuran became differentiated from the caudate original stock. But the homy sheaths in Siren rest directly upon the premaxillae and the dentaries, and not on labial cartilages; and as to its habits of life, Siren appears to be eminently carnivorous (Duméril et Bibron, Erpétologie Générale, i. 196). As has been already stated, no fossil remains of Peromela are known, but Urodela and Anura occur in some abundance, and, in certain cases, in an excellent state of preservation, as far back as the middle of the Tertiary epoch. Now, these fossils show that the Anurous and Urodelous types of organisation were, at that time, thoroughly differentiated from one another. Palaeobairachus, with its three vertebrae ankyl03ed into a sacrum, is, in fact, a singularly modified frog; while among the Urodela, the Salamandrida, the Mcnopomida, and very possibly the Proteida, are severally represented. The young of the Miocene Anura were tadpoles so similar in form to those of the existing frogs and toads, that there is no reason to doubt their resembling them in other respects.

There can be little question, then, that the Anurous and the Urodelous types must have been represented before the Tertiary epoch; but here their history breaks off, no amphibian belonging to any living groups having been discovered in Mesozoic or older strata, as far as the Lias.

From the Trias to the Carboniferous formations, inclusively, the fresh-water deposits abound in Amphibia. But all these, so far as we have any positive knowledge, are referable to the Labyrintho-dont type. No Labyrinthodont presents the slightest approximation towards the Anura; but elongated and apodal, as well as sala-mandroid forms occur; and in their cranial structure, no less than in the presence of scale-like dermal ossifications, they approach the Peromela. In regard to their possible relations with the Urodela, it is interesting to observe that in some Labyrinthodonts, at any rate, the manus has the five digits, one of which, at least, is lost in all the Urodela, and the pelvis appears to have had a distinct and completely ossified pubic element, which has also disappeared in all existing Amphibia (Miall, Report, l.c.)

The Labyrinthodonts present a few characters—such as the paired supra-occipital ossifications and the complications of the folds of their teeth—by which they approach the Ganoid fishes more than any other Vertebrata ; and it is worthy of notice that the lowest Labyrinthodonts, such as Archcgosaurus, present no approximation to the cranial characters of the lower Urodela, and show no evidence of the largely-developed branchial apparatus which is so characteristic of the latter.

Thus, if upon such slender evidence as exists, it is justifiable to speculate at all concerning the “phylogeny” of the Amphibia, the most probable conclusion appears to be that the Labyrinthodonta, the Urodela, and the Anura diverged from one another at a very early period of geological history; while, possibly, the Peromela are the last remnants of the peromelous modification of the Labyrinthodont type.

With respect to the origin of the amphibian stock itself, the following considerations appear to be of fundamental importance :— 1. The early stages of development of the Amphibia do not resemble those of any known Ganoid, Telostean, or Elasmobranch fish, and are similar to the corresponding stages of the hiarsipobranchii. 2. The skull of the lowest Urodela has, in some respects, advanced but little beyond the Marsipobranch stage. In the higher Urodela there are numerous points of resemblance with the Ganoids. The skull of the tadpole, on the other hand, has much in common with that of Chimaera (as Müller has pointed out), and with that of the Dipnoi, while the chondrocranium of the adult frog has many singular affinities with that of the Elasmobranchii, and particularly of the Rays. 3. The only Vertebrata, besides the Amphibia, which have transitory external gills are the Elasmobranchii, the Dipnoi, and perhaps some Ganoids. 4. The only fishes in which the cerebellum is rudimentary are the Marsipobranchii and Ganoidei. 5. The only fishes in which the amphibian and embryonic connection between the male reproductive organs and the renal efferent ducts is observed are the Ganoids. 6. The only fishes which have a “pylangium,” with valves disposed as in the Amphibia, are the Ganoids, Elasmobranchs, and Dipnoi. 7. The only fishes which possess morphological (Polypterus) or functional (Dipnoi) lungs are the Ganoids and Dipnoi. The conclusions suggested by these facts appear to be that the Amphibia took their origin from some primordial form common to them, the Elasmobranchii, the Ganoidei, and the Dipnoi; and that the main distinction by which their earliest forms were marked off from those of the other groups, was the development of that pentadactyle type of limb, which is common to all the higher Vertebrata. And seeing that the Elasmobranch, Ganoid, and Dipnous types were fully differentiated from one another in the Devonian epoch, it is reasonable to believe that the existence of the Amphibia, as a group, dates back at least as far as that remote period of the earth’s history.

Taxonomic Synopsis op the Amphibia.

I. The Urodela.

A. Branchiae persistent throughout life. (Pcrennibranchiata.)

1. Trachystomata.— Skull elongated; praemaxillae and dentary piece of the mandible provided with horny plates; premaxillae not ankylosed; no nasal bones, but ossifications between the ascending processes of the premaxillae; maxillae rudimentary or absent; palatines small, oval, and beset with “dents en brosse;" pterygoid absent; four persistent branchial arches; pelvic arch and limbs absent.

Siren.

2. Proteida. —Skull elongated; premaxillae and dentaries dentigerous; maxillae rudimentary or absent; premaxillae not ankylosed; no nasal bones; palatines bearing a single row of teeth, and coalescent with the pterygoids; three persistent branchial arches; both the pectoral and the pelvic arches and limbs developed.

Proteus, Menobranchus.

B. Branchiae caducous; gill-clefts persistent. (Derotremata. )

3. Amphiumida.- Skull elongated; premaxillae and dentaries dentigerous; maxillae large; premaxillae ankylosed; large nasal bones; palatines absent; pterygoid present, elongated; a basinyal cartilage; four persistent branchial arches; both the pectoral and the pelvic limbs developed, though very small.

Amphiuma.

4. Menopomida. —Skull broad; premaxillae and dentarios dentigerous; maxillae large; premaxillae not ankylosed; large nasal bones; palatines absent; pterygoid present and very broad; a basihyal cartilage; persistent branchial arches may be reduced to the first and. second; the cerato-hyal and epibranchial are confluent in the first, distinct in the second branchial arch; both pectoral and pelvic limbs well developed.

Menopoma, Cryptobranchτcs.

C. Branchiae cadu∞us, and gill-clefts closed in the adult condition. (Myctodera.)

5. Salamandrida.— Skull broad; premaxillae and dentaries dentigerous; maxillae large; premaxillae separate or ankylosed; nasal bones present, and usually large; palatines present in the young state, and situated as in the Trachystomata awiProteidea, but changing their relations in the adult; pterygoids present; the first and second branchial arches persistent,—the first two-jointed, the second a single piece.

[The latest writer on the classification of the Urodela, Professor Strauch (“Revision der Salamandrinen-gattungen,” Mêm. de l'Acad. Imp. des Sciences de St Petersbourg, se. vii. tome xvi. ), divides the Salamandrida into two tribes,— Mecodonta and Lechriodonta; the first comprising all those species, the vomero-palatine teeth of which are disposed along the inner edges of two backwardly diverging processes of the bones, and therefore form two longitudinal series divergent posteriorly; and the second, those which have the teeth disposed along the posterior edges of the vomero-palatine bones, which are sometimes truncated posteriorly, sometimes produced into a longer or shorter median process, and on which, therefore, the teeth are either directed transversely, or form two oblique series, more or less rapidly converging backwards.]

a. Mecodonta— Salaanandra, Pleurodeles, Bradybates, Triton, Chioglossa, Salamandrina. i. Lechriodonta— Ellipsoglossa, Isodactylium, Onychodactylus, Am-blystoma, Ranodon, Bicamptodon, Plethodon, Bcsmognathus, Anaides, Hemidactylium, Heredia, Spclerpes, Batrachoscps.

II. The Anura.

[For the classification of the Anura , consult Dr Giinther’s valuable Catalogue of the Batrachia salionbia; Mr Cope’s papers in the Natural History Review, 1865, and in the Journal of the Academy of Natural Science of Philadelphia, N.S., vol. vi.; and Mr Mivart’s essay “On the Classification of the Anurous Batrachians,” in the Proceedings of the Zoological Society, 1869. Far more minute investigation of the structure of the Anura than has yet been carried out seems to be requisite before their classification can be placed upon more than a provisional footing. The phases through which the Frog passes in the course of its development, show that those Anura which are devoid of a tympanic cavity are of a more embryonic character than those which possess one. The arboreal habit is so evidently adaptive, that it can hardly be regarded as a safe basis for classification. Even Rana temporaria, at a year old, will climb up the vertical side of a glass vessel, flattening out the ends of its toes, and applying its belly against the surface of the glass, like a Tree-frog.]

A. The tympanic cavity, with its Eustachian passage of communication with the mouth, may be present or absent. When present, the oral apertures of the Eustachian tubes are separate, and the pterygoid bones do not furnish a floor to them.

a. No teeth in the premaxillae or maxillae; tongue free, either in front or behind, but usually behind.

<χ. No tympanic cavity. Eustachian recesses sometimes present.

Rhinophrynus(X), Phryniscus, Pseudophryne, Brachycephalus (2), Hcmisus (3), Micro-hyla. β. A tympanic cavity and Eustachian tubes.

Hylaplesia, Kalophrynus, Bufo, Otilophus, Peltaphryne, Pseudobufo, Schismaderma, Xenorhina (4), Engystoma (4), Biplo∙ pelma (4), Cacopus (Systoma) (4), Gly-phoglossus (4), Callula (4), Brachy-merus (4), Adenomera (4), Pachyba-trachus(V), Breviceps, Chclyddbatrachus, Hypopachus, Rhinoderma, Atelopus, Copca, Paludicola.

∕ (.1) Tongue free !n front. (2) Dorsal dermal ossifications. (3) Tongue retractile. (4) No prsecoracoids.

b. Teeth in the premaxillae and maxillae; the tonguø may be fixed by its whole circumference, but is usually free behind.

a. No tympanic cavity; Eustachian recesses sometimes present.

Bombinator (1), Pelobates, Bidocus, Alsodes, Tclmatobius, Cacotus, Liopelma. β. A tympanic cavity and Eustachian tubes.

Plβctromantis, Aly tes (1), Scaphiopus, Hy-perolius, Helioporus, Nattereria, Phyllomedusa, Pelodryas, Chirodryas, Hyla, Hylclla, Otolygon, Pseudacris, Pohlia, Litoria, Triprion, Opisthodelphys, Trachycephalus, Nototrema, Ixalus, Megalixalus, Hylarana, Leptomantis, Hylambates, Platymantis, Comufer, Hemimantis, Rhacophorus, Chiromantie, Polypedates, Theloderma, Rappia, Acris, Leiyla, Elosia, Epirhexis, Phyllobates, Hylodes, Crossodactylus, Strabomantis, Calostethus, Rana, Odontophrynus, Dicroglossus, Oxyglossus, Phrynoba∙ trachus, Hoplobatrachus, Phrynoglossus, Clinotarsus, Pseudis, Pithecopsis, Mixo∙ phyes, Pyxicephalus, Ceratophrys (2), Zachoenus, Platyplectrum, Neobatrachus, Cyclorhamphus, L·imnodynastes, Crinia, Eusophleus, Plewrodema, Leiuperus, Hylorhina, IAmnoeharis, Cystignathus, ∏cmiphractus(3), Chiroleptes, Calypto-cephalus, Cryptotis (4), Asterophrys (4), Xenophrys(f), Megalophrys(4), Nanno-phrys(f), Pelodytes, Leρtobrachium, Bis-coglossus (1), Zaphrissa (1), Latonia (f), Paloeobatrachus, Arthroleptis, Grypis- eus (3).

(1) Opisthocoelian verteorae; rudimentary ribs attached to the anterior vertebrae. (2) Dorsal dermal ossifications, (3) Mandibular teeth. (4) Opisthocoelian vertebras.

B. The Eustachian tubes of the well-developed tympanic cavity have a common median aperture in the roof of the mouth, and the pterygoid bones extend beneath and form a floor to them. The tongue is wanting. The lungs are attached to bronchial tubes; and the vertebra are opistho-coelous.

a. No teeth.

Pipa. v b. Teeth in the premaxillae and maxillae.

Dactylethra.

III. The Peromela. a. With a tentaculiferous fossa on the fore part of the face.

Ccedlia, Siphonops (1), Epicrium. b. With no tentaculiferous fossa.

Rhinatrema.

(1) Siphonops annulatus has no scalea

IV. The Labyrinthodonta. Amphibamus, Anthracosaurus,* Apateon, Archegosaurus,⅛Baphetes,* Batrachider-peton, Bothriceps,Υ Brachydβctes, Brachy∙ ops,* Capitosaurus,* Chalcosaurus,* Cocy tinus, Colosteus, ∖ Easy ceps,* Ben-drerpeton,* Bictyocephalus, Bolichosoma, Erpetocephalus,* Eupelor,* Eurythorax,* Gonioglyptus,* Hylerpeton, Hylonomus, Ichthyerpeton,* Ichthyocampsa, Kerater ρeton,* Labyrinthodon,* Lepidotosaurus, Lepterpeton,* Leptophractus, Loxomma,* Mastodonsaurus,* Melosaurus,* Metc-pias,* Micropholis,* Molgophis, ∖ Oesto- cephalus (?)+ Ophiderpetonfr Osteo∙ phorus,* Pachygonia,* Pariostegus, Pholi- derpeton,* Phlegethontior,^ Pteroplax, Ptyonius,↑∙ Raniceps,∙ ∖ ∙ Rhinosaurus, Sauropleura,* Trematosaurus, j r Tudi∙ tames, f Urocordylus, ∖ Xestorrhytias,* Zygosaurus.*

[A satisfactory grouping of these genera has not yet been effected; and it is possible that some of the forms here enumerated may not be true Labyrinthodonts. To those about the truly Labyrinthodont character of which there seems no doubt a * or a † is attached— the † denoting the serpentiform genera. Batrachiderpeton, Pariostegus, and Pteroplax are remarkable for the incompleteness of the jugal arch, and some other characters by which they appear to represent the Proteidea. The true position of Hylerpeton and Hylonomus is still doubtful.]