ART. 25 CHINESE AMPHIBIANS AND EEPTILES STEJNEGER 105 



Variability. — To illustrate this it is only necessary to recite the 

 fact that Pere Heude, who once attempted the study of these turtles, 

 felt constrained to propose 8 specific names (each with a different 

 generic name) for the form inhabiting the lower reaches of the 

 Yangtsekiang. 



Breaking doivn of the natural harriers. — Two factors are here of 

 importance, human agency in carrying these valuable food animals 

 from place to place in order to market them or transplant them; 

 physical changes in the environment, some of which may be due 

 directly to man's activity in building canals, thus opening up direct 

 water communication between different river systems, or the rivers 

 themselves changing their course. Thus the Hwangho in 1852 broke 

 through in a northeasterly direction debouching into the Gulf of 

 Chili instead of 4 degrees of latitude further south. AVe are in- 

 formed that because these turtles are considered a delicacy and 

 fetch higher prices in Japan, the}'^ are shipped in great quantities 

 to the latter country and elsewhere, so that one can not be sure 

 that the specimens obtained in a locality actually is a native of that 

 place. The history of these animals goes back to the tertiary epoch, 

 and we know now how different the drainage of those times may 

 have been from that of the present time. While one might be 

 tempted to approach the problem of these forms on the hypothesis 

 that each of the great river systems, such as the Amur, the Hwangho, 

 tlie Yangtsekiang and the West River might have favored the 

 differentiation of its own peculiar form, experience from elsewhere 

 shows that specific or even subspecific differences in these turtles 

 may be older than the present river drainages. A glance at the 

 map suggests that the great northern loop of the Hwangho, en- 

 circling Ordos and northern Shensi, may in part at least have 

 belonged to an entirely different river system at some earlier period. 



The Hwangho may therefore easily share two different forms of 

 closely related turtles Avith other rivers, as does the Tennessee River, 

 and the explanation may be similar.^* 



The material received by the National Museum, since the pub- 

 lication of the Herpetology of Japan, is not of sufficient quantity or 

 quality to affect the preliminary views there expressed. 



No specimens from the Amur river drainage representing A. 

 "iiiaackii are in the museum, and none has been received since the 

 publication of the Herpetology of Japan, which can be referred to 

 -4. schlegelii. 



The specimens which have been added I am now listing under the 

 name of A. sinensis Avith some doubt. Only one is supposed to be 

 from near the type of locality (No. 46488) having apparently been 



"^ See Stejneger, Proc. U. S. Nat. Mus., vol. 62, art. 6, Feb. 10, 1923, pp. 1-3. 



