118 Dr. C. W. Andrews on some 



The grooves marking the boundaries of the epidermal 

 shields are fairly clearly marked both on the carapace and 

 plastron. 



The upper surface of the hinder region of the plastron bears 

 the bases of the ischia and pubes, which are closely united 

 with it, in the manner characteristic of the Pleurodira. 

 When found tlie pelvic girdle was m situ, but was broken 

 away in removal. As far as can be seen, the pelvis differs in 

 liO essential respect from that of Podocnemis. 



Systematic Position. — The presence of raesoplastra shows 

 that this tortoise is referable to the Pelomedusid?e, and the 

 small size and lateral position of those elements further show 

 a close relationship with Podocnemis and Pelomedusa. Tiie 

 large size of the plasti'on and the width of the bridge between 

 it and the carapace indicate a closer relationship with the 

 former of these two genera than with the latter : indeed^ at 

 first it seemed possible to refer this species to Podocnemis, 

 but further consideration, aided by the advice of Mr. G. A. 

 Boulenger, leads to the conclusion tl;at its generic separation 

 is justifiable. The chief points in which it difters from 

 Podocnemis are : — 



(1) The narrowness of the anterior vertebral shield, Avhich 



is narrower than the underlying nuchal. 



(2) Tlie large size of the intergular and the separation of 



the gulars and humerals by it. 



(3) The separation of the series of neural bones both from 



the nuchal in front and the suprapygal behind. 



(4) The comparatively slight degree to which the buttresses 



are developed. 



(5) Another point that may be of some significance is that 



in the fossil the anterior border of the plastron is not 



convex but concave, though to a slight extent only, 



and behind the antero-lateral angles the bone is 



considerably thickened, so that the upper surface of 



the plastron is gently concave from side to side in 



the middle line, the shallow depression being bounded 



by the thickened ridges just mentioned. 



All these characters except the first (which, owing to the 



imperfection of the specimens, cannot be observed) occur in 



the Chelonians from the Middle Eocene which are described 



below and referred to Stereogenys Cromeri, because they are 



found both on the same horizon and in the same locality as the 



skulls on which that species was founded, and with them are 



the commonest of the Chelonian remains. The present species, 



also, will therefore be referred provisionally to the same genus 



and its specific name will be Stereogenys libyca ; it differs from 



