AMPHIBIA - EVOLUTION The First Land Vertebrates.—The most primitive of the fossil Amphibia are unquestionably the Embolomeri. Already in Carboniferous times the group contained, as shown by Watson, "primitively aquatic animals which show no signs of ever having possessed terrestrial ancestors," as well as others which were ob viously terrestrial, and still others which had secondarily re turned to life in the water. "Despite their diverse habits, the fundamental morphology of the skeleton is strikingly uniform throughout the group." It is clear from the researches of Watson that the ancestral fish changed to a tetrapod before the latter be came permanently adapted to land life.
The first tetrapods, therefore, swam in the waters with their fish ancestors. Eogyrinus, the best known of these most primi tive tetrapods, had a long body and probably a flattened tail. It very probably lived on fishes in the pools of a rather arid and quickly drying country. As the pools dried up Eogyrinus would move, or at least was equipped to move, in the manner of an eel overland to other pools. Its skull agreed closely with that of its osteolepid ancestors differing chiefly in having a long tract of the basis cranii unossified, permitting a certain movement in the skull. The spiracular notch was very likely covered by a tympanic membrane for, as shown in certain other Embolomeri, the hyomandibular had already been converted into a stapes, even though no opening for its insertion into the otic capsule had yet appeared. The pectoral girdle was attached to the skull by post temporals as in many fish, and this girdle differed from the primi tive bony fish apparatus only in the addition of the interclavicle, a new dermal element, to its ventral surface in the mid-line. The pelvis was not firmly attached to the vertebral column as in later vertebrates, but the ilia rested on long sacral ribs and were held by muscles very much in the same way that the scapula is secured in the thorax. The limbs were small and capable of only a limited motion in any direction.
The Embolomeri are the most primitive labyrinthodonts and gave rise directly to the Rachitomi, and the latter in turn to the Stereospondyli. The Phyllospondyls and Lepospondyls, already present in the Carboniferous, evolved either from specialized Embolomeri or primitive Rachitomi. There is no good evidence of these having arisen independently from fishes. The urodeles and frogs have come from Phyllospondyls, while the coecilians seem to have arisen from Lepospondyls. Fossil intermediates between any modern Amphibia and these ancient orders are lack ing and our conclusions are based on a comparison of their skele tal characters.
Skull.—Thefirst tetrapods inherited a skull very similar to that of their fish ancestors. It consisted of a cartilaginous box or chondocranium, covered completely above by dermal bones and internally strengthened by bones which replaced the cartilage in part. This skull, the prototype of all tetrapod skulls, differed re markably from that of modern Amphibia by the far greater number of skull elements it contained, by the greater ossification of the chondocranium and by its reptile-like form.
The skull of the Embolomeri agreed in most details with that of the osteolepid fishes. The skull roofs were similar (figs. 2 and 3), except for a cluster of small bones in the nasal region in osteo lepids. The distribution of the lateral line canals in both skulls facilitates the identification of lacrymal, prefrontal, jugal, postor bital, postfrontal, supra and intertemporals. The spiracular notch of Ostelepis has become the otic notch and the elements bordering this and the orbit may be readily homologized (fig. 2) . The palates agree remarkably in the small size of the interptery goid vacuities and the almost identical position of the palatal and jaw bones (fig. 3). The large pterygoids are bordered externally by the prevomers, palatines and ectopterygoids. The distribu tion and replacement of the large teeth are identical in fish and tetrapod. An internal nares is present in all osteolepids and "is identical in position, size, and borders with that of certain Em bolomeri." The osteolepids very probably breathed air in the manner of the Embolomeri.
The lower jaw of the first tetrapods differed greatly from that of modern Amphibia in the far greater number of elements it con tained. It was, however, identical with that of the osteolepid jaw.
A. Eusthenopteron B. Baphetes The brain case of osteolepids and Embolomeri were very similar but not identical. In both the basioccipital formed a circular condyle perforated by the notochord. The exoccipitals extended upwards to a well-developed supraoccipital. The basisphenoids of both had definite basipterygoid processes with which epiptery goids articulated. The great difference is seen in the palatal view. The osteolepids had a large part of their basis cranii unossified. In this they differed from all other vertebrates, save their de scendants, the coelacanths. This character would seem to exclude the osteolepids from the immediate ancestorship to the tetrapods, but the numerous resemblances of brain case, skull roof and jaws show that the osteolepids must have been closely allied to the ancestor fish.
Once this reptile-like, highly ossified and complex skull of the ancestral fish became established in land forms, it underwent an enormous specialization. The successive stages in its specializa tion may be traced in the Rachitomi and Stereospondyli and to a lesser extent in the other orders of Amphibia. Very soon there began a progressive fenestration, the bones tending to segregate along lines of greatest stress. The solid dome-like skull roof progressively lost various dermal elements until in the modern frogs only the premaxillary, maxillary, nasal, quadratojugal, squamosal, frontal and parietal are left. The more primitive urodeles still retain a lachrymal and prefrontal in addition to the frog's equipment. The quadratojugal, although appearing as a separate element during ontogeny in the urodeles, later fuses with the quadrate. This loss of elements was perhaps largely due to a shortening of the skull. The Embolomeri had 12 cranial nerves; and in the evolution of this group and the Rachitomi the hypoglossal foramen in the exoccipital is carried farther back ward until in the most specialized labyrinthodonts, the Stereo spondylia of the Upper Trias, it no longer perforates the skull wall, the twelfth nerve passing out behind the skull as in modern Amphibia.
The second great change which developed as the tetrapod series evolved was the gradual flattening of skull. It changed from tropibasic to platybasic, not the reverse as some embryo logists have maintained. The trabeculae were no longer squeezed together, and the interorbital septum became hollowed out into a trough.
The third fundamental change was a progressive weakening of the bony organization of the skull. The brain case became pro gressively less well ossified. Basioccipital, basisphenoid and supraoccipital became reduced and disappeared. Hence, the original tripartite condyle became converted into a pair of well separated exoccipital condyles. The tripartite condyle is retained in many reptiles, but in the promammals a withdrawal of the basioccipital leads to a condition exactly parallel to that of mod ern Amphibia.
On the palate the interpterygoid vacuity became progressively wider (fig. 3). The pterygoid lost its connection with the basi pterygoid process and higher labyrinthodonts depended entirely on the parasphenoid for its support. In frogs and certain primitive urodeles (some hynobiids, etc.) the pterygoid cartilage extends forward and fuses with the nasal capsule. The bony pterygoid which forms around it has much the form of the pterygoid of the Rachitomi but posteriorly it makes different contacts. The ectopterygoid is lost in all urodeles and frogs but still retained in a few coecilians. The palatine is variable in occurrence. Within a single genus of frogs some species may possess and others lack an ossified palatine. The most remarkable changes occur in the prevomer of modern forms. In urodeles the palatine undergoes a radical change at metamorphosis, part becoming lost and part apparently fusing with the prevomers. This combined prevomer and palatine tends to invade the parasphenoid region in the vari ous families of urodeles, sending back long dentigerous processes on either side of the parasphenoid region (salamandrids) or ven tral to it (plethodontids). In most plethodontids the posterior processes break off as one or two separate dentigerous areas over lying the parasphenoid as "parasphenoid teeth." A few frogs (brevicipitids) have succeeded in similarly dividing the pre vomers, but the posterior elements overlie the palatine, not the parasphenoid region. A few others have the prevomers fused into a single element (Xenopus). In all urodeles except certain primi tive salamandrids and hynobiids and a few higher forms, the pterygoid fails to reach the maxilla, and in the Plethodontidae it never ossifies at all, although it may extend forward to the maxillae. It is the reduced pterygoid and maxilla which make the urodele palate seem so unlike the anuran, although they are both fundamentally alike and referable to the Rachitomi plan. If the palate of such a primitive salamandrid as Tylototriton is compared with that of a frog or of Eryops, the fundamental resemblance becomes obvious.
The most reduced part of the modern amphibian skull is the lower jaw. The Embolomeri inherited a complex mandible of ten pieces. In the Rachitomi and Stereospondyli there occurs a gradual reduction of the mandible in width, a loss or fusion of elements, until in primitive urodeles there was left only a dentary, a prearticular and articular and an angular. Coronoids are present in the larvae of most urodeles. The angular is fused with the prearticular in all urodeles above the Cryptobranchoidea. In coecilians a very early fusion of the jaw elements occurs in most genera, but the resulting element probably contains a coronoid as well as the dentary, articular and prearticular, for two rows of teeth are present in many forms. In the Salientia the reduction is carried to an extreme, for only the dentary and prearticular are recognizable. The articular ossifies in very few Salientia and is then fused to the prearticular. The anterior end of Meckel's cartilage ossifies as a pair of distinct elements in some Ranidae, Hylidae and Bu f onidae.
While the changes in the skull and jaws discussed above pro ceeded regularly in the evolution of the Amphibia and may be classified as trends of evolution or orthogenetic changes without a primary adaptive significance, there were other modifications tak ing place which seem more definitely correlated with the environ ment. Perhaps the most important of these were those associated with the ear. The first tetrapods as represented by Eogyrinus possessed a stapes and very probably a tympanic membrane, but no fenestra ovalis. The sound waves were thus transmitted from the stapes through the otic capsule. This condition was scarcely different from that of fishes where no stapes was present. An other piscine character was the absence of a cranial wall between the labyrinth and the brain. The latter character was handed on to certain Rachitomi but later Amphibia walled off the auditory apparatus from the brain, leaving only a foramen for the auditory nerve. With the assumption of land life the fenestra ovalis was completed in the lateral wall of the otic capsule. The Salientia inherited this auditory apparatus in a slightly modified form. All Caudata have suffered a loss of the tympanum and middle ear. The stapes is connected in the larva with the quadrate, and sound waves are transmitted through the lower jaw resting on the sub stratum to the quadrate and thence via the stapes to the inner ear. In the metamorphosed individual the head is raised from the ground and the stapes is usually connected by a muscle with the shoulder girdle. Vibrations are thus apparently sent through the f ore-legs of the adult urodele to the ear.
The osteolepid ancestors of the Amphibia had five branchial arches, while the modern Amphibia have at most f our. The latter possess, however, a pair of laryngeal cartilages which may or may not represent the fifth branchial arch. The gill arches in the adult urodele or frog are modified by reduction and fusion.
Vertebrae.—Themajor groups of the Amphibia are classified to a large extent on the form and composition of their vertebrae. To evolve the first tetrapod vertebrae from those of their fish ancestors did not require any radical alterations. The immediate ancestors of the Embolomeri are not known to have embolomer ous vertebrae, but they probably had vertebrae which could be easily resolved into the embolomerous type. Most fish and Amphibia have the vertebrae more or less ossified and the carti laginous blocks of the embryonic vertebrae after certain unequal growths fused. The basidorsals almost invariably grow up and around the neural tube tending to enclose it as a neural arch. The basiventral may either grow up around the notochord or it may remain rudimentary, while the basidorsal grows down toward it. The interdorsal and interventral may fuse and ossify or remain separate either ossifying or not. In the tail region of the fish Amia the interdorsals sometimes fuse with the interventrals but not with the basiventral. Thus, each neural arch may have two centra, an embolomerous condition. In the body region of the same fish interdorsals, interventrals and basiventrals may fuse as a single centrum.
Such a variability of vertebral composition becomes greatly restricted in the Amphibia. The first tetrapods retained the em bolomerous or double centrum type of vertebrae for their entire column. The Rachitomi, being more cartilaginous, have the f our pairs of embryonic blocks ossified, but separate, producing the rachitomous or "backbone-cut-into-pieces" type. In the most specialized labyrinthodonts a decided one-sided growth of the blocks has occurred, for the centrum is composed entirely of basiventral, the interdorsal and interventral either being greatly reduced or remaining cartilaginous.
The Lepospondylia and Phyllospondylia have sprung from either the Rachitomi or from a labyrinthodont stock slightly more specialized than the known Embolomeri. Their vertebrae are more embryonic than that of the Embolomeri in that the notochord is less restricted by the surrounding centrum. The vertebrae are, on the other hand, more specialized in that the interdorsal and inter ventral remain cartilaginous (i.e., are not present in the fossils), the definitive centrum being formed by the basidorsal and basiven tral which may remain separate in some branchiosaurs, but al ways fuse into a single ring in the microsaurs. In both these orders the vertebral column seems to consist of a series of rings, one for each somite, the septum, as shown by the position of the rib being, therefore, between the successive vertebrae unless by a secondary shifting, such as occurs in modern Amphibia, the rib arises from the side of the vertebra. All Phyllospondylia have their ribs arising in this secondary position while most Lepospon dylia retain the more primitive position.
Primitive urodeles and frogs (Liopelmidae) have vertebrae which are very similar to those of Branchiosaurs, except that they are more ossified. Further, in the trunk region of the recent forms, the basiventrals remain very small, and the main part of the centrum is formed by the basidorsals. Interdorsals and ven trals remain cartilaginous in the primitive types of modern Amphibia and frequently, as in Ascaphus, no joints are formed between the successive vertebrae. In the higher forms, however, the intervertebral cartilage (interdorsal plus interventral fused) splits in such a way as to form a ball-like surface on the anterior (opisthocoelus) or on the posterior end (procoelous) of the verte brae. Usually in Salientia and frequently in Caudata the articular ball calcifies or ossifies. In a few frogs (Megalophrys, etc.) the ball may remain quite free of either the vertebrae before or the one behind it. This ball has been called a centrum by various authors. As it contains almost all of the interdorsal and inter ventral tissue it is homologous to the centrum of reptiles and mammals, but not to that of the Phyllospondyls, Lepospondyls or Stereospondyli. In the Salientia the form of the vertebrae as determined by the manner of fusing of the intervertebral ball affords one of the primary characters of classification.
The Salientia are in various features of their skeleton more primitive than the urodeles, but the development of their verte brae is far less diagrammatic. In the most extreme types, some terrestrial (Pelobatidae), others aquatic (Pipidae, Discoglossidae) the basidorsals and ventrals extend only a short distance towards each other before metamorphosis. At this time the notochord is greatly reduced and a much flattened vertebra results, formed, except for the articular ball, almost entirely by the basidorsal. The explanation of these coenogenic features may rest in the large notochord and very active tadpoles found in these forms. The developmental history if taken alone would throw very little light on the past history of these vertebrae.
Ribs.—Theribs of primitive Amphibians and reptiles articu late with all vertebrae, at least as far back as the middle of the tail. In phylogeny the sacral ribs become fused to the vertebrae first, then the caudals, next the lumbars, the cervical and finally the thoracic. In elongate urodeles the ribs have become reduced in number. Siren and Amphiuma have only the anterior thoracic vertebrae still retaining the ribs. In frogs the reductions have reached an extreme. Discoglossids retain three, liopelmids only two ribs in the adult. The pipids have two ossified ribs in the larva, but these fuse later to the diapophyses. The bits of carti lage on the ends of the diapophyses of the higher frogs have been considered homologues of the ribs. This is perhaps a matter of definition. No ossified ribs appear as distinct elements in the development of any Salientia above the Pipidae.
The ribs are long in the primitive labyrinthodonts and surround the body as a series of hoops. They never meet in the mid-line and never connect with the sternum. In some higher Rachitomi (Cacops) the ribs are short and directed outward. The micro saurs have long curved ribs, the branchiosaurs and modern Amphibia very short and rod-like ones. The ribs are shortest in the frogs and the long-bodied perennibranchs ; longest in the primitive salamandrids. In Tylototriton, and its close relative Triturus waltl, the tips of the ribs may be long, pointed and actually protruding through the skin. This apparently serves as a special mode of protection, for when the animal is quickly seized the needle-like ribs may give a sharp prick.
The ventral side of the body of many Labyrinthodontia, Lepospondylia and Phyllospondylia were sheathed with a coat of closely set rods or plates. These apparently gave rise to the ab dominal ribs of reptiles. In modern Amphibia only Liopelma has large myoseptal cartilages of apparently the same form as the abdominal ribs of lizards, but a few urodeles have splints of cartilage in a few of the myosepta which may possibly be similarly interpreted.
Aquatic labyrinthodonts secondarily broadened their clavicles and interclavicles. A similar broadening of the ventral girdle elements is found in many modern Amphibia which are wholly aquatic, The dermal elements, however, continued to dwindle in the higher Amphibia. This seems to be directly correlated with a greater development of the leg and shoulder musculature brought into play by the increased use of the forelimbs.
The branchiosaurs possessed a pectoral girdle very similar to that of the Rachitomi but more cartilaginous. The entire glenoid and coracoid regions remained unossified. A somewhat similar difference distinguishes the pectoral girdle of certain aquatic urodeles from that of their more terrestrial relatives.
The Salientia have retained the pectoral girdle of the more terrestrial Rachitomi in almost its entirety. The interclavicle has been lost and a new structure, the omosternum, has arisen by a forward elongation and usually a division of the coracoid carti lages. The broadened ventral ends of scapulo-coracoid bars have become fenestrated. The posterior rim ossifies as a separate centre, the coracoid, while the mesial and anterior part usually remains cartilaginous. The anterior rim is called the procoracoid, while the mesial part the coracoid cartilage. In spite of these dif ferent names the fundamental resemblance of the pectoral girdle of frog and labyrinthodont is obvious.
In higher Salientia many modifications of the pectoral girdle occur. In different families the coracoid cartilages may fuse in the mid-line producing a firmisternal type of girdle. In the Brevicipitidae procoracoid and clavicle may become lost entirely. Omosternum and sternum are subject to either ossification or loss. In the Ranidae and the Polypedatidae the former may be come widely forked posteriorly. The sternum was possibly carti laginous in the labyrinthodonts and branchiosaurs, although as it is never found fossil in these groups its presence has been denied. The sternum in the most primitive Salientia resembles greatly the abdominal ribs and may, in fact, have arisen from this series of elements.
The urodeles are specialized in the complete loss of the dermal elements. As if in compensation for this loss the coracoid carti lages are usually broadly dilated. The deltoid region has become extended into a long, so-called procoracoid in most forms, but the primitive genera of several families retain this in a less extended condition. The urodeles are primitive in that the scapulo-coracoid usually ossifies as a single piece, the ossification often extending far down into the coracoid region producing a structure very much as in Eryops; in Siren and Pseudobranchus the posterior ventral margin of the cartilage ossifies as a separate coracoid.
The ilium soon gained a firmer support to the sacral ribs, and even within the Embolomeri the girdle underwent some modifi cation. Diplovertebron, according to Watson, failed to ossify its pubis, and this condition has been handed down to most modern Amphibia. The terrestrial labyrinthodonts retained a more ossi fied girdle which was almost indistinguishable from that of the contemporary cotylosaur reptiles.
Modern Amphibia have inherited the same plate-like pelvis of the primitive labyrinthodonts. The obturator foramen so charac teristic of modern reptiles is not present and the pubis is usually unossified, as in the branchiosaurs and higher labyrinthodonts. In a few Salientia the pubis region is either calcified or ossified, but it rarely forms a separate bone. The pubis cartilage is of variable extent in the different families of urodeles, it being more extensive in aquatic than in terrestrial forms. A part of the pubis is unossified in even the most terrestrial Amphibia.
Many urodeles exhibit at the anterior end of the pubis a Y shaped cartilage called an epipubis or an ypsiloid apparatus ; as shown by Whipple, this structure serves to assist in controlling the form of the lungs in those urodeles which use this structure as a hydrostatic organ. It seems to be a neomorph in the urodeles, although certain liopelmids and pipids have a cartilaginous plate immediately anterior to the pubis. The primitive Ascaphus has developed a pair of rod-like cartilages which lie over the ventral surface of the pubis and function in directing forward the "tail" or copulatory apparatus of the male.
Limbs.—The Amphibia arose from some generalized cross opterygian. The skeletons of the paddles of only two of these primitive types of fishes are known. These fin bones are leg-like in that they consist of a proximal element or humerus subtending two distal elements, a radius and ulna. The outer row of elements, however, are more numerous than the supporting elements in the hands or feet of tetrapods. Undoubtedly a series of fusions and losses must have occurred in the change from fin to limb. The modifications are confined to the distal segment of the appendage. The fundamental ground plan of single-proximal, double-median and multiple-distal elements was already well established in the fish ancestors of the tetrapods.
The chief problem in the origin of the chiropterygium concerns the distal segment of the limb. How many fingers and toes were present in the first land vertebrates and how many carpal and tarsal elements formed their support? Watson has shown that Diplovertebron had five well-developed fingers and toes, as in the case of the oldest reptiles. He finds no evidence of a sixth or seventh digital ray, such as has been described in the rachitom ous Eryops. This makes it highly probable that the prepollex and postminimus, if present at all, were cartilaginous, as in certain recent reptiles and mammals.
All Amphibia above the Embolomeri had only f our digits in the hand, but many, to judge from their tracks, must have had a well-developed prepollex. Further, Eryops apparently had a stout prepollex, a rudiment of the fifth finger and a cartilaginous block representing a sixth. The prepollex was originally a sup porting ray on the inner side of the hand. In modern Salientia it is almost universally present and often hypertrophied in the male to ensure a better grip in amplexus. In urodeles, even in forms which are not known to practise an amplexus, the prepollex is sometimes present and even bony. The modern Amphibia in herited a prepollex, four digits and a rudiment of a fifth digit in the hand.
In the foot of modern Amphibia there are five digits present, as in Diplovertebron, but also in most Salientia a prehallux, and in some primitive urodeles, both a cartilaginous prehallux and a postminimus. In burrowing Salientia the prehallux is greatly en larged to form the core of a "spade." This gives the impression that the prehallux is a recent adaptive structure which has arisen in connection with the burrowing habit. Its wide occurrence in non-burrowing forms is, however, against such a hypothesis. It seems highly probable that the prehallux is a fundamental element and that the original tetrapod foot, like the hand, was seven rayed.
The oldest known carpus and tarsus are both from rachitomous types. In both there is found the maximum number of elements found in any Amphibian. The proximal row consists of f our ele ments. Distal to this proximal row of elements there is a series of three elements called medialia by Schmalhausen, because of their position. Immediately distal to the medialia are the car palia, i–S in the hand, and tarsalia, 1-5 in the f oot.
Numerous fusions or, perhaps better, failures of the blastema to differentiate distinct elements have occurred in the carpus and tarsus of modern forms. The Hynobiidae, Ambystomidae and Cryptobrancliidae approach most closely to the primitive condi tion. Some specimens of Ranodon, according to Schmalhausen, differ in their tarsus from Trematops only in that tarsalia 1-2 are fused. Urodeles with a reduced number of digits have suffered the greatest number of fusions. The Salientia have also diverged considerably from the primitive type. In the carpus fusions have occurred in all three rows. The discoglossids and pelobatids ex hibit fewer fusions than the higher families. The tarsus of the Salientia is peculiar in the great elongation of the tibiale and fibiale, and the loss or fusion of all the other elements save three or four of the tarsalia. This elongation of the proximal series gives the frog's hind leg the appearance of having three long segments instead of the usual two. Such a specialization is un doubtedly an adaptation to jumping, although various species living to-day have given up that habit.