496 PROF. HOWES AND MR. A. M. DAVIES ON THE [DeC. 4, 



observation of Peters's (1 6), as to the existence of " a small additional 

 phalanx between the ultimate and what is normally the penultimate " 

 one in his PolypedatincB. Peters's description was very meagre, and 

 it is to Boulenger that credit is due for having placed the matter on a 

 sound footing. The conclusions arrived at by the two authors differ 

 ill important respects, and these we shall duly consider. Both deal 

 exclusively with the Ranidce. The presence of a structure in the 

 HylidcB which at once suggests that in question had been previously 

 noted by one of us, while engaged upon an allied investigation (7), 

 and as we had, in pursuit of the same, accumulated material sufficient 

 for a fuller elucidation of the problem, we decided to work it out. 



We give in Section IV. a complete list of the animals examined. 

 Our sincere thanks are due to Dr. Giinther, F.R.S., for certain of 

 these, and to Mr. Boulenger for others and for ever generous 

 assistance and advice. 



Method. — The digits were, if necessary, first decalcified in | p.c. 

 chromic-acid solution, with a few drops of nitric acid. When soft 

 they were transferred to an aqueous solution of NaHCOg until the 

 acid was for the most part removed. They were then dehydrated 

 with alcoholic solutions of increasing strengths and stained in borax- 

 carmine, the excess of stain being subsequently removed by 75 p. c. 

 alcohol, to which a few drops of HNOg were added. They were then 

 finally dehydrated in absolute alcohol and transferred, via turpen- 

 tine, to paraffin, and cut with the Rocking Microtome. 



Staining was, in many instances, resorted to without dehydration. 

 We found that a couple of hours sufficed in most cases for decalci- 

 fication in chromic and nitric acids. The preparations were re- 

 moved from this as soon as possible, for it was found that if allowed 

 a prolonged immersion therein the chances of successful staining 

 became reduced. 



In deciding upon the above method we found Whitman's 

 • Methods of Research ' of great service, and we gladly acknow- 

 ledge the help afforded us by that work. 



II. GENERAL AND ANATOMICAL. 



The supernumerary phalanx was, as seen by both Peters and 

 Boulenger, ossified, and in such a condition that, had it been found 

 fossil, would have left no doubt of its value as such, except perliaps 

 for its small size. It will be shown later on that the structure is 

 subject to a wide range of modification anatomically and histo- 

 logically ; and its fundamental relationships may best be studied in 

 detail in such a form as the tree-frog {llyla arborea, Plate XXIV. 

 fig. 1). Neither Peters nor Boulenger were aware of its existence in 

 this family ; and we find that our own discovery of the same was 

 anticipated by Leydig in 1876. He writes (14. p. 166), "von 

 Andern und mir wuide bei IJyla arborea ein sich zwischen die 

 letzte und vorletzte Phalanx einschiebender Knorpel erwahnt." 

 This author, then, has a long priority of claim to the first discovery. 



The supernumerary phalanx is, in the adults of most forms in which 



