516 



BULLETIN 58, UNITED STATES NATIONAL MUSEUM. 



practically from the same locality whence came Brandt's t^^e of 

 Trionyx schlegelii. It follows that if Brandt's T. maacJcii from the 

 Amur region is a distinct form, it is different from our Chinese 

 specimen also. 



The absence of specimens from the drainage of the Yangtse River, 

 which may be supposed to represent the true Amyda sinensis, is 

 particularly to be regretted, as it prevents me from arriving at any 

 but the most inconclusive and preliminary results. 



That my specimens represent tln-ee separable forms, however, 1 

 have but little doubt. 



In the first place, in the Formosan specimens, wliich I call A. sinen- 

 sis, the shell is evenly arched, without any depression on either side 

 of the neural plates and without any median raised keel. The north 

 China specimen, as well as all my Japanese specimens, agree, on the 

 other hand, in having a well-marked depression on each side of the 

 neural plates of the upper shell, so that a broad and bhmt but very 

 distinct keel is formed along the median line. In addition, the For- 

 mosan specimens seem to have the dermal ridges on the back smooth 

 and continuous, while in the north China and Japan specimens the 

 great majority of these ridges are dissolved into lines of tubercles. 



To demonstrate the difference between the north Cliina specimen, 

 which I call A. scMegelii, and the Japanese specimens (A. japonica) 

 the following table is presented : 



Comparative measurements of Japanese and Chinese soft-shelled turtles. 



