Summary of the most important new observations 
(1) The present material from the Permian of Brasil represents 
anew type.of Proganosaurian, very’ similar in general character: to 
Stereosternum tumidum, Cope, but differing from it in having a 
relatively larger skull, and in having in’ the pubisa notch instead 
of a- foramen. 
(2) In these points of divergence from Stereosternum the form 
approximates the genus Mesosaurus, hitherto known only from the 
Karoo beds of Southe Africa —is in fact indistinguishable from it, — 
and I therefore propose for the new form the name Mesosaurus 
brasiliensis. 
(3) In the skull] the following new observations have been made. 
The cranial region proper is short, the orbits rather large, and the 
external nares placed far back, not far from the eyes, and separeted by 
the elongate nasal bones. About one-half of the pre-orbital part of the 
skull (rostrum) is premaxilary.. There is a single pre-nariai row of 
teeth on each vomer (pre-vomer). The occipital condyle is entirely 
basi-occipital. It is practically certain that the skull is of the diapsid 
type as regards the temporal arcades. Teeth are thecodont, about 48 in 
the mandible, and fewer, about 38,.in the upper jaw. 
(4) The cervical vertebrae number 11, including the atlas, and 
there are 18 thoraco-lumbars, making ‘29 pre-sacrals. 
(5) The material elucidates completely the structure of the 
shoulder-girdle, which had previously been misinterpreted by several 
morphologists. There is a low fan-shaped scapula, coossified with a 
large coracoid, the latter pierced by, a foramen, and there is no sepa- 
rate precoracoid, the coracoid arising from a single ossific center. 
There is a slender elongate interclavicle, expanded anteriorly, where 
it articulates with well-developed clavicles. 
(6) The pubis is notched instead of perforated by a foramen as in 
Stereosternum. 
(7) The humerus shows a radio — or ectepicondylar groove in 
addition to the ulno-condylar (entepicondylar) foramen. 
(8) From a limited series of meausurements, to limited to be 
absolutely trustworthy, it appears that the ratio of length of tibia 
and femur is about three-fifths, while in Stereosternum it is about 
