Variation of the Edible Frog, 245 



this is by no means always so, and specimens are to be found 

 in which the snout is much more rounded than in some 

 jR. ridibunda. I have selected tliree specimens, of which 

 outline figures are here given, to show that the above 

 definition o£ the three forms cannot be relied upon. 



Figr. 3. 



Upper views of heads of typical form, c^, St. Male (a) ; var. ridibunda, 

 $ , Capljina, Herzegovina (b) ; and var. chinensis, § , Broughton 

 Bay, Corea (c). f nat. size. 



The width of the head varies between 1 and 1^ times its 

 length in the typical form (28 : 32 in ? from Havre), be- 

 tween 1 and 1^ times in tlie var. ridibunda (=in some 

 specimens from Herzegovina, France, Portugal, Algeria, 

 Asia Minor, Persia, 36 : 43 in $ from Kiev), between 1 and 

 1^^ in the var. chinensis. Tiie width of the interorbital space 

 is ^ to ^ tiiat of the upj)er eyelid in specimens of the typical 

 form from St. Malo and Paris, ^ to | in otiiers from Poitiers. 

 In the var. ridibunda, taking only specimens from Germany 

 and Austria-Hungary into consideration, it is between ^ and ^, 

 but it may be exceptionally f ( (^ from Laaerberg near 

 Vienna) ; | (in a large ? from Damascus) is another excep- 

 tioji. in the var. chinensis it varies between ? and |. 



Tlie head varies much in shape, and exceptionally may 

 even be not unlike that of a typical R. temporaria ( ? , var. 

 ridibunda, from Crete). The canthus rostralis is always 

 very obtuse ; I have never seen a specimen in which it may 

 be said to be " strongly marked/' as stated by Bolkay in his 

 description of R. chinensis. 



The Hind Limb. 



That there are very considerable differences in the pro- 

 portions of the hind limb, I was the first to point out, and 

 I have proposed to make use of those for defining varieties, 

 with the necessary restrictions in the diagnoses imposed by 



