The Theromorpha comprise a great number of extraordinary, extinct reptiles, which as a group had a wide range in space and time. The earliest known occur in the Lower Red Sandstone of Thuringia and Bohemia, and in the middle Permian strata of Russia. The majority have been found in strata transitional between the Permian and the Triassic age, notably in the Karroo sandstone of South Africa and in corresponding levels of North America. Closely allied to them are those of the Triassic sandstone of Elgin in Scotland, and of India. They seem to have died out with the Muschelkalk or Middle Trias.
The various genera exhibit such a diversity of structure, shape, and size, and many are still so imperfectly known, that any diagnosis is liable to be faulty, even assuming that they are a homogeneous group. To avoid confusion, we characterise the Theromorpha as Reptiles with a firmly fixed quadrate, a single temporal arch, an interparietal foramen, and a pelvis in which the pubes and ischia form one stout, ventral symphysis.
The dentition is most abnormal, and permits the division of the Theromorpha into two or three main groups. In the Pareiasauri the teeth of the upper and lower jaws form rather even series of nearly equal size; smaller teeth are carried by the palatal bones. In the Theriodontia the teeth are differentiated in a truly Mammalian fashion into incisors, prominent canines, and multicuspid or tubercular molars. Each tooth, and this applies to all Theromorpha, is implanted in a separate alveolus; Tritylodon only seems to have double-rooted molars. The lower canines cross in front of the upper, just as in Mammals. In Placodus, which probably belongs to this assembly, the teeth are few in numbers, very broad and flat, especially those of the palate. In Dicynodon and Gordonia the teeth are restricted to a pair of conical, sometimes very large, tusk-like upper canines, and in Oudenodon the whole mouth is toothless.
The configuration of the skull shows two main types. In the Pareiasauri it is completely roofed in by dermal bones, the only holes on the surface being the nostrils, orbits, and the interparietal foramen.
The most striking feature of the second type of skull is the tendency to form an almost Mammalian zygomatic arch by the junction of the much elongated squamosal with the jugal bone, both abutting against a downward process of the postfrontal bone. The skull shows a pair of wide supratemporal foramina bordered by the parietals, squamosals, and postfrontals. The composition of the temporal arch varies considerably in detail, and in Cynognathus crateronotus at least there is a small hole within the arch, between the squamosal and jugal, probably the last remnant of the otherwise absent infratemporal foramen. Except in the roofed-in skulls of Pareiasaurus and Elginia there is no separate quadrato-jugal element. The quadrate is firmly fixed by the overlapping squamosal, and the whole pedicle for the support of the mandible is rather elongated, and either stands vertically or slants forwards. The mandible itself is compound. The pterygoids extend backwards so as to approach or reach the distal portion of the quadrate; separate ectopterygoids do not seem to be developed. The shoulder-girdle consists on either side of a large scapula, which is mostly directed obliquely backwards, and is fused with the coracoid; a precoracoid is present or at least indicated by a notch or foramen; it is usually fused with the other bones. At least some genera possess a T-shaped interclavicle and clavicles; Pareiasaurus possesses also a pair of cleithra.
The pelvis is in every respect constructed upon the Mammalian plan. The three constituent parts meet at the acetabulum, and the ventral bones, pubes and ischia, form one broad symphysis, leaving two, sometimes very small, obturator-foramina. The ilium is attached to one to five sacral vertebrae, and since the whole pelvis slants obliquely downwards and backwards, this sacral attachment is distinctly pre-acetabular, perhaps most markedly so in Dicynodon. The limbs are mostly stout, humerus and femur with strong crests; the feet are thoroughly plantigrade, with five fingers and toes. The details of the carpus and tarsus are not well enough known to permit of generalisation, but there is a tendency to form a heel, and to develop the cruro-tarsal joint into the chief joint of the hind feet. The vertebrae are amphicoelous, sometimes with rather thin-walled centra, so that in these cases the chorda was continuous. Intercentral wedges, or basiventral elements, are frequent in the cervical and caudal regions. Most of the ribs, especially those of the neck, have a tuberculum attached to the neural arch, and a distinct capitulum which articulates either with the centrum or with the intercentrum, or lastly, if the latter is absent, between two centra. The axis and atlas vertebrae are united.
The occipital condyle exhibits every stage between the single median knob (Pareiasaurus) formed almost entirely by the basioccipital bone, a triple condyle (Dicynodon) to which both lateral and the basioccipital bones contribute, and a kidney-shaped or double condyle (Cynognathus) from which the middle or basioccipital portion is more or less withdrawn.
Dermal bony armour reached an extraordinary development on the head of Pareisaurus and Elginia; whether other parts of the body were protected is doubtful, but the flattened tops of the neural spines of Pareiasaurus suggest that they carried bony scutes. Abdominal protective ossifications are unknown. Many of the Theromorpha[[126]] reached a considerable size, massive skulls of one foot in length being not uncommon. The tail was comparatively short.
The many resemblances of these strange creatures to Mammals have naturally suggested that the Mammalia have sprung from some such Theromorpha or "beast-shaped" animals. The resemblances are chiefly the dentition, the zygomatic arch, the pelvis, the cruro-tarsal joint, the scapula which is sometimes possessed of a spine, and the occasionally double occipital condyle. The general shape of the skull of Cynognathus is indeed strikingly like that of a Carnivorous Mammal, and the shape of the whole body suggests rather a Mammal than a reptile; and when we have to deal with the fragmentary skulls of Tritylodon (cf. p. [309]) it is, indeed, difficult to decide to which of the two classes such a creature belongs. But the Theromorpha possess a number of important characters by which they reveal themselves at once as reptiles: (1) the large and fixed quadrate bone, which is still the sole support of the lower jaw; (2) the compound mandible, which is composed of at least an articular, dentary, angular, supra-angular, and splenial element; (3) the interparietal foramen; (4) the possession of prefrontal and postfrontal bones, sometimes also postorbital, supratemporal, and quadrato-jugal bones. Of course, any of these ancestral bones may be lost, and the interparietal hole may be closed as in tortoises and crocodiles. We can also imagine that the quadrate may be relieved of its jaw-bearing function and become loosened, but this is not easy, considering the strong development of the squamoso-quadrate pedicle. Those Theromorpha in which the quadrate itself is small, whilst the squamosal reaches down, or at least approaches the mandible, as in Dicynodon and Gordonia, are so hopelessly pledged, or specialised in other directions, that it is impossible to connect them ancestrally with Mammals.
However, it is beyond reasonable question that the Mammals have sprung from some reptilian stock (the attempts to derive them from Amphibia, without the intervention of Reptiles, are as gratuitous as they have proved futile), and the Theromorpha undoubtedly comprise creatures which of all animals approach nearest to Mammals, and coincide with them in most important features. But we have not yet found a single Theromorph which can claim to be a direct ancestor of Mammals. Since the latter occur already in the Trias, we have to look for their reptilian forefathers at least in the Lower Permian, and this naturally excludes all the known forms. The filling up of this gap is but a question of time.