OF THE SCAPULA IN THE BATItACHIA. 213 



from end of coracoid by cartilage, longer than pars acromialis ; 

 dividing cleft pierced horizontally, only visible when viewed 

 obliquely. 



H. pieties Ptrs. — Scapula as in II. albopunctattis. 



Example. 



Length of praecoracoid 10 units. 



,, scapula 25 ,, 



,, „ shaft 17*5 „ 



Breadth of ,, (distal end) 8 - 8 ,, 



,, „ (middle) 2 „ 



,, pars acromialis + pars glenoidalis 8*8 „ 



Length of supra-scapula 17*7 ,, 



Text-figure 10. 



Pectoral girdle of Heleioporus pictus, X2 ; from within. 



The abnormal development of the scapula in this genus is 

 similar to that in Phyllomedusa as regards length. In Phyllo- 

 medusa, however, the pars glenoidalis is much smaller than the 

 2?ars acromialis, which is separated from the head of the praecora- 

 coid by acromion cartilage ; also the dividing cleft is long and 

 broad, forming a very large central foramen, pierced vertically. 



Summary and Conclusion. 



These data are sufficient to show that the form of the scapula 

 is remarkably stable within a genus, and that in many instances 

 it is so highly characteristic that a glance at this bone alone 

 suffices to recognize the genus to which the skeleton belongs. 

 This is the case throughout the Pipidae and Discoglossidae, and, to 

 a great extent, in other families. In the most recent families, 

 however, there are many genera conforming to one pattern of 

 scapula, which makes the character, in these cases, worthless for 

 diagnostic purposes. 



Of the genera which I have studied, the greatest specific varia- 

 tion occurs in Bufo ; in Hyla, which is a veiy large genus, there 

 is practically none. 



Although in many ways the Pelobatid genera resemble Palwo- 

 batratchus, the Aglossa must be regarded as having the most 

 primitive type of scapula, as in Pipa, Hymenochirus, and Xenopus 

 the proximal end of the bone is entire. In all other genera* 



* Genera of the Arcifera. 



