234 KNOWLTON 



less marked between the mammals of the "Ceratops beds" and those 

 of the Puerco. His statement is as follows: 



Estimating the geological intervals by dental evolution and fauna] 

 succession, there is first the great gap between the Trias of Micro- 

 lestes and Dromotherium and the Jurassic of the Stonesfield state; 

 there is a relatively shorter interval, but still a considerable one, 

 between this and the Purbeck or Atlantosaurus beds. Then fol- 

 lows another long and very important interval between the Atlanto- 

 saurus beds and the Laramie (Upper Cretaceous) . The gap between 

 the Laramie and Puerco was relatively short as indicated by the 

 comparatively limited evolution both of the Plagiaulacidae and 

 Trituberculates. 



Mr. J. W. Gidley, of the United National States Museum, who 

 has had opportunity of studying much of the material seen by Marsh, 

 together with that upon which Osborn based his conclusions as well 

 as some new material, takes even stronger ground than Osborn as 

 to the afl5nities between the Ceratops and Puerco and Torrejon 

 mammals. He says: 



I do not hesitate to say, judging from the known forms, that the 

 differences between the mammals of the Jurassic and those of the 

 Ceratops beds are at least ten times as great as they are between 

 the mammals of the Ceratops beds and those of the Puerco. This 

 conclusion is based on morphological ground rather than actual 

 relationships. Leaving the multituberculates out of consideration 

 at present the pattern of the tooth-crowns, especially of the upper 

 molars, in the Jurassic mammals is fundamentally different from 

 that of the known mammals of any later horizons, while most of the 

 forms of the Ceratops beds have attained the typical tritubercu ate 

 pattern, which, with very slight variations, has been repeated over 

 and over again in every higher horizon where mammals are found, 

 and is the dominant form in most of living orders. The natural 

 inference follows, therefore, that the time interval between the Cera- 

 tops beds and the lowest acknowledged Tertiary above is very slight 

 as compared with that between the Ceratops beds and the Morrison 

 beds below. 



The multituberculates show almost as wide a di (Terence in time 

 between the Morrison and Purbeck, and Ceratops beds, while the 

 Ceratops beds and Torrejon probably contain at least three genera 

 of mammals {Ptilodus, Cimolomys and Meniscoessus) in common, 

 and there are no genera in common with the Puerco. It is thus 



