128 BULLETIN 169, UNITED STATES NATIONAL MUSEUM 



their family characters, such as the ectoloph construction, are ap- 

 proached only as decidedly aberrant generic characters by one or a few 

 primate genera. 



If we grant Matthew's second opinion that the astragalus of Mixo- 

 dectes does not have diagnostic ordinal rodent characters, the evidence 

 for rodent relationships is little more than the presence of enlarged in- 

 cisors and (in some but not all genera) of more or less molariform pre- 

 molars. Both these characters also appear independently in insecti- 

 vores, numerous different lines of ungulates, primates, marsupials, 

 and other orders. Nor are they really rodentlike in precise detail in 

 this group. Indeed, there now seems to be no actual evidence that 

 the mixodectids are related to rodents. 



Granting the usage of Insectivora not only to include the recent 

 groups but also numerous extinct forms that necessitate definition on 

 primitive characters only, we may well caU the mixodectids insecti- 

 vores. Negatively, it may be said that no other defined order could 

 receive them, and positively that their dentition is insectivorelike at 

 least in habitus, that the astragalus is more like that distinctive of 

 insectivores than like any other group except rodents (which are 

 excluded by the dentition), and that in general they have the primitive 

 features by wliich the Insectivora sensu lato are defined. 



It is probable that the mixodectids include a related group of phyla 

 that diverged from the primitive placental stock, and apparently from 

 the Insectivora in a more limited sense, at a very early date. Had 

 such a sideHne evolved more rapidly, or had it run a longer span and 

 occupied a more important place in mammalian history, it would be 

 more conveniently defined as an order, as, for instance, are the tillo- 

 donts, which probably had a very similar history but developed more 

 striking specializations. Since, in fact, the mixodectids were a short- 

 lived and relatively unimportant group, it is most convenient simply 

 to classify them in the order Insectivora, from which they probably 

 arose. 



Within the Mixodectidae there have been included two apparently 

 distinct groups: Mixodedes and Indrodon of the Torrejon, on one 

 hand, and Cynodontomys and Alicrosyops of the Lower and Middle 

 Eocene, on the other. The Torrejon forms are certainly closely re- 

 lated, indeed the distinction between them is not clear, and the Eocene 

 genera are also closely allied and rather difficult to distinguish. 

 Matthew (1915c) defined the two groups as subfamilies, Mixodectinae 

 and Microsyopinae, and he repeatedly expressed doubts as to their 

 really being related to each other, rather than merely convergent. 

 In the lower jaw (the upper being uncertainly known in this respect) 

 the "Mixodectinae" retain a canine, and the enlarged tooth is an 

 incisor, while in the "Microsyopinae" there is only the enlarged tooth 



