THE MOST PRIMITIVE KNOWN REPTILE. 



277 



There is a large hemicylindrical intercentrum whose width 

 agrees with that of the basioccipital part of the condyle. The 

 lower surface is smooth, short blunt-ended backwardly directed 

 processes forming its latei-al margin. The upper surface of this 

 bone has laterally two large flat triangular areas for articulation 

 with the neural elements; between these areas the anterior end 

 of the bone is excavated into a conical pit, only half of which is 

 present, but which resembles exactly that in the end of an 

 ordinary centrum. 



Text-figure 7. 



IN 



C OakC^ C^ WC^ C 



8eymouria-hayloriensis Broili.— Three anterior vertebrae, X 2. 

 Right lateral aspect. Left surface. 



The atlas is composed of drawings of the elements in my skeleton replaced in what 

 appears to be their natural relations, the centrum of the axis and the inter- 

 centrum before it are hypothetical. 



C2, &, centra of 2nd & 3rd vertebrae; In.C.i>2, 3^ intercentra 1, 2, & 3; 

 Od., odontoid ; N' & N^, neural arches of 1st & 2nd vertebrae. 



One pair of other elements is preserved. Each of these has 

 a high laterally compressed but antero-posteriorly elongated 

 neural spine, whose posterior margin descends directly to the 

 hinder end of a well-formed posterior zygapophysis. The 

 anterior margin of the spine descends vertically about to the 

 level of this articular face, and then turns forward so as to 

 lie horizontally and form the upper edge of a squarish area of 

 bone whose front edge is rounded and articular and of consider- 

 able width. The lower part of this region ends in a short blunt 

 point. It is obvious that the neural spines of the pair of elements 

 lay in contact with one another and that the zygapophyses were 

 supported by those of the succeeding vertebra. The anterior end 

 could have articulated with the exoccipital part of the condyle, 

 the intercentrum supporting the basioccipital part. 



Proc. Zool. See— 1918, No. XX. 20 



