182 DB. R. BROOM ON THE STEIJCTURE AND [June 4, 



whether this single arch is the homologue of the upper or o£ the 

 lower arch in the typical reptiles. As, however, the single arch in 

 the Dicynodonts differs in structure from either of the normal rep- 

 tilian arches, and as it is moderately certain that the reptiles with 

 the two arches have been derived from the ancestral forms which 

 had the temporal region completely roofed, by quite a different 

 line from that by which the Dicynodonts have arisen, the single 

 arch in the latter cannot be regarded as homologous with either 

 of the arches in the more typical reptile. In the branch which 

 gave rise to the majority of Eeptilian orders the temporal roof 

 became transformed on either side into a couple of arches by 

 an upper fenestra formed between the parietal, squamosal, 

 postorbital, and postfrontal, and a lower fenestra between the 

 jugal, quadrato-jugal, supratemporaP, and postorbital. In the 

 other great Eeptilian branch which gave rise to the Anomodonts, 

 Theriodonts, and Mammals, a single fenestra only has been formed 

 in the temporal roof, corresponding to the upper fenestra of the 

 branch from which the Ehynchocephalians and allied forms have 

 sprung. After the formation of the fenestra, or possibly before, 

 the skull in the anomodont line became much simplified by the 

 reduction and loss of a number of elements and the corresponding 

 increase in size of others. The squamosal increased apparently 

 early iu relative size, and with its increase the supratemporal and 

 the quadrato-jugal became first reduced and then lost. In a 

 number of different lines of descent we find a similar process 

 has gone on. Thus, in Sphenodon the squamosal by its increase 

 has led to the complete loss of the supratemporal and the great 

 reduction of the quadrato-jugal ; in Aetosaurus the increased de- 

 velopment of the squamosal has led to the loss of both the supra- 

 temporal and the quadrato-jugal ; and in the Plesiosauria, which 

 are possibly an offshoot from the Anomodont stem, we find a very 

 large squamosal with complete loss of the supratemporal and 

 quadrato-jugal. The temporal region of Udenodon only differs 

 from that in the Plesiosaur in that, whereas the latter has both a 

 postfrontal and a postorbital, one of those elements is lost in 

 the former ; in my opinion it is the postorbital which is lost in 

 the Anomodont. 



The only essential difference between the structure of the 

 postero-lateral region of the skull in the Mammals and that in 

 the Anomodonts and Theriodonts is that the mammalian skull 

 has become further simplified by the loss of the postfrontal and 

 the almost complete loss of the quadrate. A number of years ago 

 I advanced (10) the view that the reptilian quadrate had its 



^ Concerning the two bones found in the upper and posterior temporal 

 region, opinions have differed as to which is to be regarded as the squamosal 

 and which supratemporal. The opinion to which I have been led is that it is 

 invariably the upper and inner which is the squamosal — an opinion which 

 agrees with that of Baur, but differs from that of most British authorities, 

 including Parker. When only the one element is found it appears to be 

 always the squamosal. 



