HERPETOLOGY OF JAPAN. 517 



In the first four columns I have given tlie exact measurements of 

 the membranous sliell, and in the last four columns I have reduced 

 the same dimensions to per cent of the length of the membranous cara- 

 pace. Although the series is small, the figures probably deserve some 

 confidence because of their remarkable uniformity and harmony. If 

 we examine the table, the first observation to impress itself upon us 

 is that the males and females from the same locality differ from each 

 other but very little in their proportions, except in the depth of the 

 shell — i. e., the greatest distance between A^entral and dorsal surfaces 

 on the middle line measured with curved calipers, the difference 

 amounting to 5 and 7 per cent, respectively. At the same time it 

 should be noted that the difference between the various individuals 

 of the same sex does not reach 2h per cent. 



If we now compare the north China specimens with the Japanese 

 specimens, we observe that in the males the depth of the shell of the 

 former is 9 per cent greater than in the latter. Unfortunately we have 

 no female specimen from northern Cliina; but the difference between 

 the island specimens and the mainland male is so great that the latter 

 even exceed the females of the former by 4 per cent. The difference 

 may be very strikingly formulated b}" stating that in the males of A. 

 japonica before me the depth of the shell is one-fourth the length of 

 the carapace, while in the male A. scJiIegelii it is one-third the length 

 of the carapace. 



The Formosan specimens appear to be slightly narrower and also 

 somewhat deeper than the Japanese without even approaching the 

 north China specimen in this respect, inasmuch as the latter, a male, 

 presents exactly the same proportions as the female of the more south- 

 ern form. 



The Formosan and north China specimens agree in the lateral out- 

 line of the shell, for while in the Japanese specimens the greatest width 

 of the body is situated more nearly at the middle, in the former the 

 greatest width is farther back. 



The material is so scant that I shall pass over another point wherein 

 the Formosan specimens and the north Cliina one seem to differ from 

 the Japanese, viz, in the somewhat shorter plastron of the latter. 

 Compared with the width of the body, the difference amounts to 

 nearly 10 per cent; but this measurement is less satisfactory than the 

 the others, and, as stated, the series is too small to give positive 

 results. 



It would then appear that the Japanese soft-shelled turtles are 

 neither identical with the north Chinese nor with the Formosan forms. 

 The question therefore naturally arises, whether it may not agree 

 with the south Chinese species. As said before, I have no specimens 

 for comparison, and the literature offers very little of a definite nature. 

 Both Gray and Boulenger, examining Formosan as well as Chinese 



