282 PROCEEDINGS OF THE ACADEMY OF [1895. 



rather than retrograde. In very many, perhaps the majority of 

 cases, the skull is modernized earlier than the feet. 



A very much more difficult problem is to determine what 

 relation, if any, exists between Protoptychus and the Heteromyidce. 

 In most schemes of classification this family and the Geomyidce are 

 referred to the Myomorpha, but several authorities deny the pro- 

 priety of such a reference. Winge/in particular, rejects the division 

 of the simplicidentate rodents into three sections, and founds his 

 arrangement of the order principally upon the characters of the mas- 

 ticatory muscles and upon the modifications of the skull which 

 accompany these characters. On account of the anterior position of 

 the infraorbital foramen and the imperforate zygomatic process of the 

 maxillary, the "Sacomyidce" (which include both families) are 

 brought into relation with the beavers, marmots and squirrels. The 

 obvious objection to this method of classification is that it ignores the 

 possibility of the independent acquisition of similiar structures in 

 widely separated series, through convergent or parallel development, 

 processes, the reality and frequency of which are not open to question. 

 All available evidence goes to show that in the primitive rodent type 

 the zygomatic process of the maxillary was perforated by a large 

 opening, and that the jngal extended up along the anterior edge of 

 the orbit to the lachrymal. There is nothing impossible in the sug- 

 gestion that the condition of these parts found in the Geomyidce and 

 Heteromyidce may have been derived from some such arrangement 

 as that which occurs in Protoptychus. In the latter the dentition and 

 the structure of the cranium also suggest a relationship with the 

 Heteromyidce. It is, however, not at all probable that the Uinta 

 genus is directly ancestral to any of the latter family, because the 

 condition of the incisive foramina is so totally different. Until the 

 connecting links are recovered, it will not be possible to decide 

 whether the suggestive resemblances between Protoptychus and the 

 pocket-mice are due to any actual relationship, or are merely 

 instances of convergence. 



As a contribution to this question, it may be of use to consider the 

 relationship between the Geomyidce and the Heteromyidce. In con- 

 trasting these two families, Cones 2 has enumerated the following points 



1 Winge, If. Rongeurs fossiles et vivants tie Lagoa Santa, Miuas Geraes, 

 Bresil. E Museo Lundii. I. Copenhagen, 1888. 



2 Cones, /'.. and J. A. Allen, Monographs of North American Kodenls. 

 U. S. Geolog. Survey of the Territories, Vol. XI, pp. 492-3. 



