Get access

The Amphibian Fauna from the South Joggins. Nova Scotia


*Communicated by Prof.D.M.S. Watson, F.R.S, F.Z.S


The morphology of the Labyrinthodontia is now well known, so that the classification of the order into the three grades Embolomeri, Rachitomi, and Stereospondyli rests on an accepted basis. These clearly-defined grades predate one another in time, and are ancestral to one another in the structure of the vertebral column and the skull, intermediate forms being rare. The interest of the Nova Scotia Labyrinthodonts lies therefore in Dendrerpeton acadianum, which provides, not only in the structure of the skull but also of the vertebral column, a form intermediate between the Embolomeri and Rachitomi, combining structural features of both grades.

The Embolomerous vertebra, known in Eogyrinus, Pholidogaster pisclformis, and Cricotus, as well as by numerous isolated neural arches and centra, consists of a neural arch which rests equally on two centra, a pleura- and interceontrum. Both pleura- and intercentrum are complete circum-notochordal rings, of equal size, well ossified, showing in section a double cone ossification, with a small central perforation for the notochord. In the Rachitomous vertebra the intercentrum is represented by a u- or wedge-shaped element, lying vontral to the notochord, the pleuracentrum by a more posterior pair of elements, dorsal to the notochord. The vertebral column of Dendrerpeton acadianum, known only by one connected series of vertebrae, by detached neural arches, and pleura- and intercentra, and by sections of these vertebrae, is, in its structure, intermediate between these two types; the vertebrae resemble these of the Embolomerids in the mode of contact of the neural arch and centra (the neural arch articulating equally with pleura- and intercentrum), the equal size of the pleura- and intercentrum, and the fact that the pleuracentra are complete circum-notochordal rings, showing no decrease mid-ventrally in antero-posterior length. They differ from the Embolomerous vertebrz, and resemble those of the Rachitomous in that the intercentra are hemicyclinders, situated ventral to the notochord, and only partially surrounding the notochordal space–that is, the pleuracentra are Embolomerous, the intercentra Rachitomous in structure, the complete vertebra, in the non-reduction of the pleuracentra, approaching more nearly to an Embolomerous than to a Rachitomous type.

It is clear that the vertebrae of Dendrerpeton acadianum, with the vertebræ of such Labyrinthodonts as the caudal vertebræ of Pholidogaster pisciformis and the type-specimen of Diplovertebron punctatum, together with such a stage as is present in Trimerorachis and the caudal vertebrae of Eryops, in all of which the pleuracentra are complete circum-notochordal rings and the intercentra u-shaped cyclinders, can only be interpreted as a morphological stage in the evolution of the Rachitomous from the Embolomerous type of vertebra.

The skull of Dendrerpeton acadianum has the high type of skull characteristic of all the early Embolomeri. In the skull:

  • 1Flanges from the tabulars and dermo-supraoccipitals extend on to the occiput, a condition which occurs in no known Embolomerous form.
  • 2The pterygoids do not meet in the mid-line of the palate, a condition characteristic of all Rachitomi but occurring in no Embolomerous form.
  • 3The pterygoids articulate movably with the basipterygoid process of the parasphenoid, as in the Embolomerids Eogyrinus, Palæogyrinus, and the Rachitomous forms Trimerorachis and Archegosaurus, but in no later Rachitomous form.

The neural cranium, belonging to the third largest of ten skulls, is characterized by:

  • 4A single nearly circular concave occipital condyle, the greater part of which is formed from basioccipital, as in all Embolomeri and early Rachitomi.
  • 5An unossified supraoccipital. An ossified supraoccipital occurs in all known Embolomeri and Eryops, but is lacking in Trimerorachis and the Stereospondyli.
  • 6The paroccipital is sufficiently well ossified to preserve sharp impressions of that part of the horizontal and posterior vertical semicircular canals which lie in the paroccipital bone, in contrast to Eogyrinus and Trimerorachis, in which the labyrinth lies as an impression on the inner surface of the cranial wall, and to Eryops, where the canals are enclosed in spongy bone.
  • 7The foramen for the tenth nerve lies between the exoccipital and paroccipital, as in Palæogyrinus and Eryops, and unlike later Rachitomi.
  • 8There is a twelfth cranial nerve.

This skull is Embolomerous in the structure of the condyle, the relation of the exoccipital to the paroccipital, and the movable articulation of the pterygoid with the parasphenoid. It is Rachitomous in possessing occipital flanges from the dermo-supraoccipitals and tabulars and the reduction of the pterygoids, so that they do not meet in the mid-line of the palate, two structural features which occur in no Embolomerid skull. In its skull-structure Dendrer-peton acadianum is intermediate between the Embolomerous and Rachitomous Labyrinthodonts, a position consistent with what is known of the structure of the vertebral column and in agreement with the geological age of the fauna which is Upper Middle Coal Measures in age.

The remaining Labyrinthodonts of this fauna are known only by skull-tables, asSociated with certain of the limb- and girdle-bones and isolated vertebrae. The skull-tables of Dendryazousa dikella and Calligenethlon watsoni are interesting in possessing slender tabular horns–that is, there is still a relic of the connection between the shoulder-girdle and the skull, although there is no trace of a functional connection such as is present in the Anthracosaurs and Loxommids. In these skulls no occipital flanges from the dermo-supraoccipitals and tabulars occur. The vertebræ of Calligenethlon watsoni are unknown in section, but those of Dendryazousa dikella agree with the vertebræ of Dendrerpeton acadianum in showing reduction of the intercentrum.

Two other features of these Labyrinthodonts are of general interest. The pelvis of Dendrerpeton acadianum, Dendryazousa dikella, and the unidentified Labyrinthodont pelvis agree in possessing an ossified pubes; this is characteristic of all known Embolomeri (except Diplovertebron) and the early Rachitomi. The ilium in all three specimens shows a well-developed post-acetabular process, and, where the anterior margin is complete, a small anterior dorsal process. In all the Embolomeri where the pelvis is known, e. g., Diplo-vertebron, the Embolomerous pelvis from Pictou (Watson, 1926, fig. 27) and Pholidogaster pisciformis (anterior region of the pelvis not known), both these processes are present. An anterior dorsal process occurs in no Rachitomous pelvis. The Labyrinthodont pelvis, which is the only one in which the inner surface is preserved (PL II. fig. 3), shows no trace of a sacral articulation, as in Pholidogaster and Gricotus; it is interesting that this pelvis has a comparatively short ventral symphysis. An entepicondylar foramen in the humerus of Amphibia, with the exception of Diplocaulus, has not been recorded; it is, however, present in Dendrerpeton acadianum and the Lepospondyls Hylonomus lyelli and Fritschia curtidentata.

The four species included in the Lepospondyli are very incompletely known. Thoy agree in possessing vertebræ of a normal Lepospondylous type, i. e., in which the neural arch and centrum are co-ossified, no neurocentrum suture being apparent.

One anomalous feature of Hylonomous lyelli is the presence of caudal hæmal arches articulating freely between successive centra of the caudal vertebræ. The occurrence of free intercentra in certain of the species attributed to the group Microsauria (of which Hylonomus lyelli is the type) led Goodrich to regard the Microsaurs as belonging to the Reptiles rather than the Amphibia. The morphology of the species included in the Microsauria, the occurrence or absence of intercentra, and the structure of the complete vertebral column in the group is, however, so inadequately known that it is impossible to discuss the relations of the Microsaurs. Hylonomus lyelli is regarded as an Amphibian because no presacral intercentra are present. The majority of primitive Reptiles (e. g., all Seymauriamorpha) have a complete series of small intorcentra throughout the vertebral column. Further, no neurocentrum suture is ever apparent, even at young stages, in the development of the Lepo-spondyl vertebra. A neurocentrum suture occurs in the young stages of all Reptiles, persisting often into the adult.

I can express only very inadequately my thanks to Professor D. M. S. Watson, F.R.S., for his constant help and advice. I would also thank Dr. W. E. Swinton, of the British Museum (Natural History), for the trouble he has taken in providing me with material. I am deeply indebted to the Trustees of the British Museum of Natural History, London, for the extensive use they have allowed me to make of their collection of fossil Amphibia; and especially to Professor Clark, of the Peter Redpath Museum, McGill University, Montreal, for sending to me here Dawson's entire collection of fossil Amphibia from Nova Scotia.

Get access to the full text of this article