500 PROF. HOTVES AND MR. A. M DAVIES ON THE [DeC. 4, 



teristic of the syndesmosis. Apart, however, from this important 

 structural diflFerence between the two things, there are certain other 

 considerations which point to the conclusion that the syudesmodial 

 pad is not, as might appear, a late infolding of the investing 

 membrane. In many instances, especially in its phalangeal condition, 

 it enters into that w'hich is, at best, a loose connection with the same 

 at all points {p.s./i., Hyla arborea, fig. 1), and there are developed 

 true synovial surfaces at its points of contact with the penultimate 

 and terminal phalanges. All this being so, the question next in need 

 of solution is that of the primary origin of the syudesmodial pad — 

 Is it, as its histological structure in the larvae of Hyla and Rana 

 would suggest, related to the phalanges ? — or is it, as its condition 

 in Proteus would seem to indicate, a derivative of the investing 

 tunic 1 



The base of the terminal phalanx is, in some Annra, greatly en- 

 larged and swollen, giving rise (ex. Hylodes) to an immense tubercle 

 which underlies the distal extremity of the penultimate one. In 

 certain others the ojiposite ends of the phalanges are sever- 

 ally expanded, in a somewhat similar fashion; and, when first our 

 attention was drawn to the subject, it appeared not unlikely that the 

 supernumerarv phalanx might represent the dismembered tubercle 

 of either the penultimate or terminal one — more probably that of 

 the latter. The coexistence of the two things in Nototrema and other 

 Hylidsh, in itself, fatal to this supposition ; while it shows the struc- 

 ture in question to be of independent origin. Choice seemed then to 

 lie between two alternatives ; it appeared to be either an intercalary 

 structure of comparatively late origin, or a sesamoid. We have al- 

 ready given reasons (supra) for doubting the former hypothesis. Our 

 chief reason for supposing it to be a sesamoid lay in the assumption of 

 a similarity of relationship with the proximal end of the terminal pha- 

 lanx to that so common among the higher Vertebrata. This, however, 

 is inconstant, for in many instances (ex. Rhacophorus eques, fig. 2) 

 the structure is disposed lineally with the penultimate phalanx. 

 Appeal to the details of muscular insertion proved fruitless, as the 

 tendons become, for the most part, merged into the investing tissues 

 proximally to the structure under consideration. 



The youngest stages examined by us were those of Rana tempor- 

 ariu and Hyla arborea. In the former, at a period at which the tail 

 is almost absorbed, the condition of the joints recalls that of Proteus ; 

 this is especially the case with the most distal joint (sy., sy.', fig. 10), 

 which is strictly similar to that of the Urodele, except for the pos- 

 session of a triangular projection from one side of the outer mem- 

 brane. A similar condition is met with in Hyla arborea. 

 Examination of our youngest specimen of this species shows, however, 

 that, at a stage at which the parts in general are first becoming 

 differentiated (fig. 9), the syudesmodial pad is centrally histologically 

 identical, to minute detail, with the adjacent incipient phalanges ; 

 while, peripherally, it graduates off into the surrounding indifferent 

 tissue. Thus it is seen that the phalanges and syndesmoses are, 

 together with their investing sheatli, differentiations of a continuous 



