The evolutionary story of Dicynodon, as set out in this paper, may be summarized as follows:‐ The less advanced Pelycosaurs have an elongated skull with a short temporal region within which lay the relatively feeble temporal muscles, probably already sub‐divided into musculus temporalis and musculus pterygoideus posterior whose thickening was allowed to take place by the presence of a laterally placed temporal fenestra. These muscles arose from the skull roof and probably not at all from the neural cranium. The articulation between the palate and the basis cranii lay between the transverse flanges of the pterygoids. The quadrate condyle is cylindroid and the lower jaw opened as on a hinge; there is a musculus depressor mandibuli. In the advanced Sphenacodonts the face is greatly deepened, not apparently to accommodate the much enlarged anterior maxillary teeth, but to increase the girder strength of the structure, and to fit a temporal region sufficiently high to hold masticatory muscles long enough to allow the mouth to be widely opened. The pterygoid, in the region of the transverse flanges, becomes a most massive bone and the quadrate ramus is so deep that it is certain that the temporal muscles arose only from the visceral surface of the wide skull roof. Thus the face with the pterygoid, zygoma and quadrate make a mechanically adequate, self‐contained structure, housing the masticatory muscles without any assistance from the neural cranium. The brain is small compared with the skull, even allowing for the great absolute size of large Dimetrodons. Mechanically the neural cranium is important as the connecting link between the face and the “neck” and body. It is connected with the rest of the skull by overlap of the occipital plate by the interparietal and tabulars, and by the attachment of the backwardly turned lateral extremities of the paroccipital processes to the tabulars and squamosals. The basi‐cranial articulation is with a facet on the epi‐pterygoid, high up in the skull and much behind the transverse flanges of the pterygoids, a position unknown to any other reptile except the Therapsids, and feeble mechanically. In Sphènacodonts the quadrate‐articular joint is remarkable for its narrowness from side to side, in that the outer condyle lies high above the inner, and in the long radius of curvature of the articular surfaces, a condition which implies that the lower jaw slid backward and forward when the mouth was opened and shut. Why such a condition should be valuable is very difficult to conceive, but it is present in Captorhinus and in Edaphosaurs and even in Seymouria. An antero‐posterior movement of the jaw of this kind implies a muscle attached anteriorly in the skull and posteriorly in the jaw, and this muscle is the musculus pterygoideus anterior, already in existence in Captorhinus. To lengthen this muscle in Sphenacodonts to allow sufficient range in motion its attachment moves up on to the outer side of the hinder end of the angular, and the muscle cuts a notch bounded by a reflected lamina in that bone. The next known stage of mammal‐like reptiles is represented by the primitive Deinocephalia from the earliest reptile containing zone of the Russian Permian (of which a full description by Professor Efremov will shortly be published), many of the primitive features of whose skulls are retained in the Deinocephalia of the later zone II, which have been more fully described. The skulls of Brithopus and Deuterosaurus are full of resemblances to Sphenacodons, but show changes which are fully realized only in the later Titanophoneus and Ulemosaurus. The significant changes in Tapinocephaloids are that the premaxillac become deep below the nostril and their anterior surface slopes back to give a wedge‐shaped profile to the anterior part of the skull, the temporal fossae enlarge upward so that they reduce the width of the upper surface of the skull, and the free border of the squamosal behind them is enlarged and produced backward. This results in a new origin for the temporal museles, which arise from the inner and posterior borders of the temporal fossa instead of from the under surface of the skull roof, and as the need for expansion during contraction is otherwise met the lateral part of the temporal fenestra, still large in Deuterosaurus (Nopcsa, 1928), becomes reduced in Ulemosaurvs and vanishes in the South African forms, the zygoma there becoming very deep. The great extension of the squamosal and quadrate below the floor of the brain cavity is retained but is buttressed by exceedingly powerful, down‐turned paroccipital processes. The exaggeration of these structures is associated with a general increase in size and massiveness of the whole occipital part of the brain‐case which is perhaps in part, though certainly not to a large extent, associated with an increase in brain size. But the essential improvement is in the much more secure attachment of the jaws to the neck which results from this enlargement. The increased size of the brain‐case pushes down the basi‐pterygoid processes until, though still in contact with the epi‐pterygoids, they come to rest, secondarily, on the upper surfaces of the pterygoids, far behind the transverse flanges. These events lead to a reduction in depth of the quadrate rami of the pterygoids, the mechanical needs met by those structures in Dimetrodon being now assumed by the strong paroeeipital‐quadrate articulation. In later Tapinocephaloids an additional change, the bowing outward of the zygoma, makes space for an enlargement of the masseter muscle, which seems to have existed in Dimetrodon. A small musculus depressor mandibuli is still present, so placed that there can have been no tympanic membrane. The Tapinocephaloids, from their beginning in Deuterosaurus, possess enlarged grinding incisors, the large canine undergoing reduction. Associated with this arrangement is a reduction of the size of the pterygoid flanges, accurate closure of the mouth being no longer desirable. The Titanosuchids are reptiles which represent the same evolutionary grade as the Tapinocephaloids. Their later members resemble Tapinocephaloids in their great size, in the pachyostosis which exists in the cranial roof, and (according to Broom) in the presence of crushing incisors of similar type, and in the sharp distinction of cranium and face. On this basis the two groups are commonly included in the order Deinocephalia. The early Russian forms of Titanosuchids, which in the case of Titanophoneus seem to be clearly related to the later South African forms, differ considerably from the Tapinocephaloids and suggest that the later resemblances are secondary. In the Russian Titanosuchids the incisors are not unduly enlarged, the canine is powerful, and the cheek teeth small. The transverse flange of the pterygoid remains functional. The palate bears patches of small teeth but otherwise resembles in its general structure, in the mode in which it is attached to the basis cranii, and in the shallow quadrate rami of the pterygoids, the Tapino‐cephaloids. But the articular surface of the quadrate is no longer carried down below the basis cranii, the lower jaws lack the rounded upper border posteriorly which is carried on from Sphenacodonts to the Tapinocephaloids, and the temporal fossa is necessarily elongated. The surface of the skull between the temporal fossae remains wide, but the free upper and jiosterior border of the squamosal is pressed backward to give length to the temporal muscles, and, as the temporal fenestra is large and to a considerable extent laterally placed, the zygoma becomes shallow and outwardly bowed. In later Titanosuchids the quadrate condyle moves forward and a little downward from its original position in the plane of the basi‐occipital condyle. These differences are of such a nature as to suggest that the two groups are not very closely allied, the features in which the early members resemble one another being those which distinguish a Therapsid from a Pelycosaur. With the earliest Deinoeephalia in Russia occurs Venjukovia. As Efremov has shown, this animal is related to the Tapinoeephaloids which it recalls in the enlarged grinding incisor teeth, in the structure of the face, and in general of the lower jaw. But it possesses an articular bone whose articular surface recalls that of a Pelycosaur in the large radius of curvature, so that the jaw movement is a sliding rather than a rotational one. This suggests an origin from the very base of a Therapsid stock, probably not long after the separation of the Tapinocephaloid and Titanosuchid stocks. The most remarkable feature of Venjukovia is the presence in the lower jaw of an expansion of the dentary laterally to the cheek teeth, covered with horn, against which the upper canine and post canine maxillary teeth bit, and of a growth forward and inward of the anterior part of the palatine between the maxilla and the palatal nostril, which opposes the mandibular cheek teeth and was certainly horn covered. The quadrate lay far ventrally, the quadrate ramus of the pterygoid is slender and the sutural attachment of the upper surface of the pterygoids (far behind the vestigial transverse flanges) to the basi‐sphenoid is exactly as in Tapinocephaloids. From Venjukovia, or from a closely allied form lacking such specialized features as the dew‐lap on the dentary, the Anomodonts seem to have arisen. Venjukovia occurs in the Deuterosaurus and Titanophoneus zones of the U.S.S.R., the typical Anomodonts are found first at the base of the Tapinoeephalus zone in South Africa in the form Dicynodon jouberti. A little later is Brachyprosopus (Olson, 1937) from higher in the Tapinocephalus zone, still retaining a series of cheek teeth in the maxilla, and perhaps the premaxilla, which pass mesially to the canine. In the succeeding lower Endothiodon zone are many representatives of the true Endothiodonts, in which the canine is lost in both sexes, whilst the maxillary cheek teeth are increased in number, and may be arranged in more than one row. They have tall slender crowns with a serrated posterior edge, and are rapidly replaced; similar teeth lie at the extreme lingual side of the dentary. Laterally to these teeth the dentary is grooved and was horn covered to bite against the maxillary teeth, the condition being derived from that in Venjukovia. But this horny cover extends so far forward that it functionally replaced the incisors, and led to the disappearance of the canine. In the upper jaw a corresponding horncovered outgrowth of the maxilla, which is quite unrepresented in Venjukovia, makes its appearance and the horny cover continues forward to cover the premaxillae and replace the upper incisors. The antero‐lateral processes of the palatines still, as in Veniukovia, bite against the mandibular teeth and their horny coats were probably confluent with that over the premaxilla. The Endothiodonts have a remarkable articulation between the lower jaw and the skull, in which the quadrate has two cylindroid condyles, the outer lying far dorsal to the inner. These articulate with convex surfaces on the articular, the inner lying at a lower level than the outer, connected with it by a vertical articular surface. This arrangement can only have arisen for a condition recalling that in Dimetrodon. The Endothiodonts have already developed a typical Anomodont squamosal in which the bone consists of a broad sheet extending the occipital plate laterally with a backwardly turned upper border from which arose temporal and masseter muscles, separated by the root of the zygoma and plunging down into a very large temporal fossa. The Endothiodon zone, throughout its whole extent, yields smaller Anomodonts which retain large canine tusks, at least in males, and have a very variable number of small teeth in the maxilla and dentary and, very rarely, also in the premaxilla. These animals seem to represent a line of descent differing from that in the true Endothiodonts, the persistence of the canine in them, even although it is no longer used in feeding and owes its enlargement perhaps to its general restriction to males, is difficult to understand if they arose from a form in which, as in Venjukovia, the upper canine is a crushing tooth opposed by a horny plate in the lower jaw. The conditions in Synostocephalus, where the horny sheet lateral to the maxillary cheek teeth is short, narrow and had in all probability a marginal ridge which bit against the outer surface of the mandibular beak, suggest that this sheet did not meet that over the premaxillae, being separated from it by the canine. In such later forms as Tropidostoma a reduction in the number and size of the cheek teeth is associated with a great expansion and fusion of the palatal horny plates, which now cover the whole region except for the teeth. In the lower jaw the horny pad which opposed the post canine maxillary teeth spreads forward on to the lateral grooves, which make their appearance on the upper surface of the symphysis, and teeth are inserted into the dentary on its inner margin. In Tropidostoma the articular retains a relic of the Endothiodon condition in that the inner condylar surface is still on a lower level than the outer, and is connected with it by a rather low but still nearly vertical articular face, but the articular faces have now a much greater anteroposterior extent, the jaw having the power of sliding backward and downward, forward and upward over a considerable range. Endothiodon has a reflected lamina of the angular which passes out abruptly from the outer surface so as to enclose a capacious pocket, which cannot be accounted for by any condition of the attachment of the anterior pterygoid muscle, but must have housed an independent organ, which Broom has suggested may have been a salivary gland. In small “Endothiodonts”‐such as Synostocephalus‐this pocket still exists, not so wide from side to side but sometimes covered by a very extensive lamina. In Tropidostoma it is obviously present (cf. fig. 15 A, Ref. Lam.), but is very greatly lessoned both in thickness and area. In Dicynodon it is lost, the reflected lamina fitting closely to the rest of the bone so as to leave space only for the ligamentous sheet which attached the muscle to the jaw. In fact, Dicynodon secondarily reverts to the condition found in the Deinocephalia. Thus Venjukovia, Synostocephalus and Tropidostoma form a series of structural stages intermediate between a primitive Tapinocephaloid and Dicynodon. The resemblance of Tropidostoma to Dicynodon is so great that a very small continuation of change in the direction shown in the series would lead to the latter form, and it is indeed difficult to believe that the descendants of Tropidostoma did not become Dicynodons. But the genus had long been in existence when Tropidostoma lived, and it seems extremely probable that Dicynodon is polyphyletic, perhaps highly so. In these circumstances it is evident that only a study of a large series of completely prepared skulls of Dicynodon would make it possible to disentangle the lines of descent amongst the Anomodonts and allow the great group of species of Dicynodon to be divided amongst a soundly based series of genera. The most interesting result of the present study lies in the confirmation it gives to Romer's view that the Therapsids are of Sphenacodont descent. It lends additional support to the view that the Deinocephalia are the most primitive group of Therapsids, and to Professor Efremov's conclusion that Venjvkovia is of Tapinocephaloid origin. The structure of Venjukovia explains the origin of the Anomodontia in that it shows a very early stage in the development of their horny beak. And the “Endothiodonts” provide a series of forms which elucidate further stages of this process. An analysis of the Dicynodon skull shows that it is a structure singularly well fitted to house a powerful jaw moving musculature and to stand the stresses set up by its action. In sum this study serves to show that all the features which distinguish a skull of Dicynodon from that of a Tapinocephaloid, or indeed of a Pelyeosaur, serve a mechanical need and are hence in essence adaptive. The only exception is the retention, even the enlargement of the upper canine tooth, and in most cases its restriction to males. I am indebted to Professor A. S. Romer for the opportunity of examining much material of Pelycosaurs, to Professors Orlov and Efremov, of the Academy of Sciences of the U.S.S.R., for showing to me materials of the Russian Permian reptiles and explaining to me the results of their unpublished researches. And to Miss Joyce Townend who has made the drawings which illustrate this paper. This work is offered to Robert Broom as a testimony of respect in his 81st year.

EXPLOR_STEP=$WICRI_ROOT/Wicri/Santé/explor/EdenteV2/Data/Istex/Curation

HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 006639 | SxmlIndent | more

HfdSelect -h $EXPLOR_AREA/Data/Istex/Curation/biblio.hfd -nk 006639 | SxmlIndent | more

{{Explor lien
   |wiki=    Wicri/Santé
   |area=    EdenteV2
   |flux=    Istex
   |étape=   Curation
   |type=    RBID
   |clé=     ISTEX:CDFD99E7F8AE5989EEAABF8C7783FD64438025B7
   |texte=   The Insectivorous Genera Microgale and Nesogale (Madagascar)
}}