520 SODLMES TRORASHEGLD EINES: 
bre were greatly flattened in the antero-posterior direction and 
expanded laterally; this feature, with the short and stout 
limbs, led Cope to suggest that they were possibly fossorial in 
habit. 
The three remaining forms mentioned have been placed ina 
separate family from the preceding, the Pariotichide; they are 
all smaller than Diadectes, never exceeding a length of two or 
three feet,and are further ‘distinguished by the fact thatthe 
teeth are not equal in size in all parts of the jaw, the teeth of 
the middle part of the maxillary series being larger than the 
others. Certain of the less well known of the Texas forms 
show broad bony plates that extend outward from the middle 
of the back and cover the ribs to a large extent; they corre- 
spond to the ribs in number and position. This is a very close 
approximation to the condition in the turtles; in Osocwlus the 
plates are especially well developed and the lateral edges of 
adjacent plates meet. 
The taxonomy of this group and its related forms is very 
imperfectly understood and there has resulted little but confusion 
from the numerous schemes of classification that have been pro- 
posed. When the first of the African fossils were discovered 
and described the name Anomodontia was proposed by Owen for 
the whole series of African forms, as he supposed that they were 
all closely related, but he soon recognized that this was not true 
and separated a group, the Zheriodontia, to include the forms 
with a more carnivorous dentition. Later writers have found it 
necessary to depart very widely from this scheme and have 
founded new orders and in some cases done away with the 
original ones. Out of this tangle one thing seems very clear, the 
order FPareiasauria is a distinct order separate from all the rest 
of.the Permian reptiles and is the primitive form of all the 
reptiles. Other forms that are closely related to this order, but 
whose structure is too incompletely known to make the deter- 
mination of their position definite, had best here be retained in the 
original order Anomodontia as suggested by Seeley and Lydekker. 
The classification of the forms will then stand as follows : 
