UPPER CRETACEOUS MAMMALS 99 



the lack of association absolutely prevents the definition of useful species which are 

 even distantly comparable in scope to the species of other and better known mammalian 

 faunas. 



The condition is not much changed today. Knowledge of later, related mammals 

 has increased greatly and given greater scope for the employment of analogy, but it is 

 still impossible to reconstitute the entire dentition of even one genus. Thus the same 

 primary problems which faced Marsh and Osborn face the present attempt. It is clear 

 that a revision, in the strictest sense, is impossible. One can only bring the earlier con- 

 ceptions more into line with the advanced knowledge and methods of today; one can- 

 not establish a detailed taxonomy comparable to those lately established for many of 

 the Tertiary faunas. In attempting even this much one must again make a choice be- 

 tween analysis and synthesis. Should one revive all of Marsh's names and propose the 

 very numerous new names which would be necessary for a thorough analysis of tooth 

 tj-pes, or should one follow Osborn in essentials and use generic names for agglomera- 

 tions of different forms which can be separated, in various parts of the dentition, from 

 other similar agglomerations but which must be much more comprehensive than the 

 genera of any other fauna ? 



After due consideration, the latter course has been adopted, with such exceptions 

 as are forced by proper respect for the rules of nomenclature. Many of Marsh's generic 

 names and perhaps the majority of his specific names are thus neither recognized as 

 valid and used nor definitely reduced to synonymy. They are simply included in other 

 genera or species which are used in a very broad sense. In the case of genera based on 

 isolated teeth, no serious attempt is made to distinguish species beyond the necessary 

 recognition of one species as a genotype. Separation of such genera into species is im- 

 possible and unnecessary at present. 



The main result of this work is hence not an actual systematic revision of the 

 upper Cretaceous mammals, but a broader treatment which nevertheless permits one to 

 gain an adequate and important conception of the general character of the fauna and 

 of the range of morphological variation exhibited by its different groups. 



Three orders occur: Multituberculata, Marsupialia, and Insectivora — the most 

 ancient of all mammalian orders and the most primitive orders of the two great groups 

 of modern mammals."* The several faunas are essentially one for morphological pur- 

 poses as there are no clear differences due to age or facies, and they are treated together. 

 Separate faunal lists are given in a later section of this work. The fragments from the 

 probable Cretaceous of Patagonia are also included. 



The following treatment, especially of the multituberculates, follows Osborn's 

 revision of 1893 in many of the broader essentials. There are, of course, numerous dif- 

 ferences both in morphological conceptions and, particularly, in views as to affinities, 

 but the credit for the first fundamental approach toward the solution of the broader 

 problems (many of them, however, even now not completely solved) is Osborn's. 



^* Aside, of course, from the monotremes, the pre-Pleistocene history of which is at present totally 

 unknown. 



