90 I'ROCEEDINGS OF THE 



were diplasioccielous. Tlie remaining specimens, 7?. ma.rhmis. 

 It. mudagascariensis, 11. ncJilfr/elii, and II. reinwardtU [2], \\ere uni- 

 formly procoelous. Ot" the five specimens but one {11. reinwnrdtii) 

 permitted of direct observation of the vertebral faces, but 

 Mr. Boulenger, to whom I referred the question, agreed that the 

 vertebral articulation in these specimens is as stated. 



Such a condition is apparently inexplicable in view of the fact 

 that the lihucopliori are generally accepted as true Eanidte. It is, 

 however, of peculiar interest in view of the fact that the Bliaco- 

 pliorl have not alw ays been regarded as Eanidse. Originally they 

 \\ere placed with the Hylidae (which are of course procoelous), to 

 winch they bear a most remarkable resemblance which is, at the 

 ])resent time, attributed merely to convergence. How close is 

 this resemblance may be judged from a fact recently brought to 

 light by Stejneger ('07). This author has pointed out that the 

 specimen originally figured by Schlegel and regarded as the type 

 specimen of Polypedates {lihacoj)1iorus) schhgelii is actually a mere 

 vai'iety of Hyla arhorea (H. arhorea japon7ca)\ This view, he 

 remarks ('07, p. 77), has been confirmed by an examination of 

 the original specimen in the Leiden Museum. He figures this 

 variety of Hi/la as possessing the tongue typical of the Hylidse 

 but as having a foot which, so Mr. Boulenger informs ine, is 

 absolutely characteristic of ItJiacojyhoms I 



The skeletons of the several species of Rliacopliorus in the col- 

 lection of the British Museum show the coracoids united but 

 somewhat widely separated from the weak clavicles by conspicuous 

 and well-calcified precoracoids. 



Apart from these apparent exceptions, the whole of the 

 remaining forms, at present grouped in the I'irmisternia and 

 which I have been able to examine, are diplasiocojlous. More- 

 over, this character, as already pointed out, is one which, in this 

 group as in the Procoela, is exceptionally free from individual 



variation. 



******* 



It may, nevertheless, be objected that the occurrence of excep- 

 tions renders the character of the vertebral articulation of little 

 value for systematic purposes. The condition of the sternal 

 apparatus is, however, open equally to such objection. Forms 

 with sternal apparatus as diverse as those of liana and of 

 llemmis or of Brtviceps and of Cacopus are all grouped together 

 as Firmisternia. I have shown, too, in a recent paper ('15), that 

 there is normally in Rana tir/rina a very marked overlap of the 

 coracoids rivalling that of many Arciferous forms. 



It is obvious, moreover, that whereas the precise manner in 

 which the intervertebral spheres unite with adjoining centra has 

 no pliysiological importance, the consolidation of the sternal 

 apparatus may well be of considerable ])hysio]ogical value and 

 therefore be a character w liich might well arise independently in 

 widely separated forms*. 



* Cf. the existence of a consolulated sternal apparatus in tlie Aglossa. 



