1884.] Batrachia of the Permian Period of North America. 27 
The characters displayed by the three divisions in question, 
indicate their relationships to be as follows: The orders of 
division I present in their cranial structure a greater resemblance 
fo the limi>-finned or crossopterygian fishes, than do either of the 
_ others. The third division is the most divergent from that type, 
and is ia various respects the most specialized. This specializa- 
tion cčasfsts not only in a departure from the primitive Batrachia, 
but also from all other forms of Vertebrata. Its specialization is 
seen in the loss and codssification of various parts of the skele- 
ton. The Urodela display characters intermediate between the 
extremes of the class. Near them the Trachystomata (Sirenidz) 
are inferior by loss of parts of the skull, and of the pelvic arch, 
the result, probably, of a process of degeneration. The same is 
probably true of the Proteida, which have lost the maxillary arch 
of the Stegocephali, but retain their os intercalare.’ 
As regards the extinct orders, the primitive type is evidently 
the Rhachitomi, whose vertebral column displays an arrest of 
characters which are transitional in the higher Vertebrata. From 
this group the orders Embolomeri and Stegocephali have evi- 
dently been derived. We may then present the following gene- 
alogical table of the class Batrachia: 
. — Anura 
Gymnophiona 
Urodela Trachystomata 
Proteida 
Embolomeri Stegocephali 
‘ we 
Rhachitomi 
As regards the connection of the class, as a whole, with other 
- classes of Vertebrata, it is very probable that the extinct orders, 
as the Rhachitomi, were derived from some extinct form of Dip- 
noan fishes more or less related to the group of which the genus 
Ceratodus is a representative. In this type we have a persistent 
chorda dorsalis; fins which present the type from which ambulatory 
Of Cuvier; Epiotic of Iuxiey, according to Vrolik. 
