1892.] NATURAL SCIENCES OF PHILADELPHIA. 293 



Oxi/ccaa, Protopsalis and Hrjcenodon. (4) Amhhjctonidce, with 

 Amb/yetonus aud Pakeonidis. (5) Meson ycliidoi, with Mesoiiyx, 

 Dissacus, Scij-cothraustes, Patriofelis and Thereutherium. Later in 

 the same work (1889) he modities this arrangement by removing 

 the Miaeidce from the Carnivora and erecting them into a special 

 group, "Creodonta Adaptiva." I regard Schlosser's sclieme as, on 

 the whole, a very natural one and have adopted it with some modi- 

 fication. 



For the reasons already explained, the following classification of 

 the Creodonta must be regarded as merely tentative, a provisional 

 attempt to arrange the genera by jiliyla or lines of divergence, 

 which can be but imperfectly expressed by rigid definitions, even if 

 these lines were much more clearly and completely understood than 

 is the case at present. It may seem that I have unduly increased 

 the number of genera, and that many of these are but vaguely 

 defined. But it should be remembered that in the Puerco, and to 

 a somewhat less degree in the Wasatch, the pattern of the teeth is 

 similar throughout many difi'erent groups, and minute differences of 

 tootli-structure must be employed when we have reason to believe 

 that they indicate important differences in the general character of 

 the animal. This is illustrated by Panto lestes, an artiodactyl, which 

 was regarded, and with good reason, as a creodont. With our 

 present materials a rigid system is impracticable, because it leads to 

 the unnatural association of forms really very different. But at 

 the same time, it must be admitted that such vagueness is most 

 undesirable and may lead us quite astray. Hence, the emphasis 

 laid upon the provisional character of the scheme. 



In the following table of families and in the enumeration of the 

 genera under those families, the problematical and doubtful forms 

 are omitted from consideratiim and will be discussed at the close of 

 the paper. It must not, however, be inferred from this, that all 

 those genera which are enumerated belong unmistakably to the 

 creodonts ; merely that they very probal)ly do so. 



I. Fourth upper {)remolar not forming a well-developed secto- 

 rial ;' sectorials present, if at all, in more than one pair. 



1. Superior molars tritubercular, not trenchant; cusps erect and 

 acute; inferior molars luberculosectorial, with trigonid moderately 

 elevated above the talon and not forming a shearing blade ; pre- 



^ Oxyana and Palaonictis form a partial exception to this statement. 



