THE BATRACHIA OP NORTH AMERICA. 245 



identity between the two types it would only be necessary to elongate 

 the ilia of the latter. The developed sternal apparatus and shoulder 

 girdle of the Salientia is only found among Batrachian orders in the 

 Khachitomi and Stegocephali. Thns in Eryops of the former there are 

 clavicles, coracoids, and episternum (the last reduced as in Urodela), 

 and in Actinodon there is also an epiclavicle (Graudry). The posterior 

 direction of the suspensorium of the lower jaw of the Salientia is also 

 only found in the extinct orders named, pointing again to this origin. 

 In other recent orders these bones are directed forwards. 



The modifications effected in the Ehachitomous skeleton to produce 

 the Salientian, have been partly the same as those which have produced 

 the other existing orders. Thus the true vertebral centra have been re- 

 placed by complete intercentra, and several posterior cranial bones have 

 been lost. The ilia have been greatly elongated, and in so doing have 

 embraced vertebrae successively more and more anteriorly, so that the 

 number between th^:^ ilia and the cranium has been greatly reduced, and 

 the vertebrae posterior to the point of attachment become atrophied in 

 part and concrescent in part. This process has been carried to the 

 greatest degree in the extinct family of the Palteobatrachida?. Here 

 the ilia extend to two vertebrae in advance of the ninth or usual sacral, 

 thus inclosing three vertebrae in the sacrum, and leaving onl^^ six for 

 the remainder of the column. The coracoid is iDrobably that of the 

 Stegocephalous order, as it is better developed than in the Ehachitomi. 

 The second row of the tarsus has also become reduced from these primi- 

 tive types by atrophy, while the first row lias been reduced to two bones, 

 as in the Mammalia, which have been greatly elongated. A parallel 

 case occurs in the Mammalia in sorau lemurs, particularly in tlie Tar- 

 siidae. 



I have discovered that the Ganocephala (Trimerorhachis), and the 

 Khachitomi (Tatrachys) possessed an elongate columella auris, which 

 is directed outwards, backwards, and upwards to a. possible mem- 

 hranum tynipanij which may have occupied the notch external to the 

 OS inter calare.* (Plate 50, figs, 4-7.) 



The subdivision of this rod may have given origin to three of the 

 four distinct elements exhibited by the Salientia. (Plates 49, 50.) The 

 homologies of these with the three ])rinci[)al ossicula auditUH is possi- 

 ble. The history of these parts shows that the lack of auditory ossicles 

 displayed by some Salientia and by all Urodela (Plates 48, 49) is the 

 result of degenerac3^ 



The cause of some of the modifications of the skeleton can be traced 

 to use. Thus the constant muscular stress on the ilia in humping the 

 back previous to leaping must have had a tendency to draw it for- 

 wards not only on itself, bnt on its vertebral attachments, which are 

 cartilaginous and yielding. The elongation of the first row of tarsal 



^American Naturalist, 1888, p. 465; American Journal of Morphology, Vol. ii, Pt. 

 11, 1888. 



