J. C. Merriam — Triassic Ichthyosauria. 29 



of secondary origin would seem almost a violent assumption. 

 That splitting and intercalation of digits have occurred in some 

 of the broad-paddled Ichthyosaurs is beyond question, but evi- 

 dence of the character which we find in these forms is lacking 

 in the paddle of Mixosau?"us. The digits are of equal size 

 and their relations to the intermedium are such as one might 

 expect to find in a fairly primitive limb. 



Though there is a tendency for the short first digit to disap- 

 pear in the evolution of a natatory limb, it has not always 

 done so, as for example in the Plesiosaurs. Farther, in the 

 history of the Ichthyosauria two quite distinct types of paddles 

 have appeared ; the broad form, illustrated in the Latipinnati 

 and in Mixosaurus / and the narrow form, represented in the 

 Longipinnati and in the Calif ornian genera. In all probability 

 the course of evolution has in all cases been fairly direct. That 

 is, the broad paddles have tended toward greater width and 

 the narrow ones toward slenderness. It is not easy to imagine 

 that after limbs had been reduced to a narrow type they would 

 again increase in w T idth. There is therefore good reason to 

 believe that Mixosaurus and the Latipinnati have retained the 

 first digit. 



If the first digit in the limb of Mixosaurus represents digit 

 I of a primitive pendactyle form, this type may be consid- 

 ered as the most generalized known in the Ichthyosauria. If, 

 on the other hand, the first digit represents number two of the 

 primitive form, the limb can hardly be considered as less 

 specialized than the tridactyle form seen in Merriamia, one 

 form having lost two digits, the other having lost one and 

 gained one. Supporting the first suggestion we have the fact 

 that Mixosaurus is the oldest described form in which the 

 limb structure is known. The beds in which it occurs are con- 

 sidered by Fraas as the equivalent of some portion of the 

 Middle Triassic, while the Californian genera belong to the 

 Upper Triassic. Evidently Mixosaurus is the only described 

 genus which could be considered as ancestral to the Jurassic 

 forms. In the other genera the reduction of the digits has 

 gone farther than in the Jurassic Longipinnati. 



It is not impossible that other forms with wider paddles will 

 be found in the American Trias, but up to the present time 

 only the leptochirous or narrow-paddled group seems to be rep- 

 resented. These forms may be closely related to the Longipin- 

 nati or may represent a branch of the order which diverged and 

 specialized early. The Longipinnati and Latipinnati may 

 have developed from a persisting primitive stock after the 

 American Triassic forms had become well separated from the 

 rest of the order. 



It should be borne in mind that while the evidence furnished 

 by limb structure is some of the most valuable material that we 



J 



