E. L. Troxell — Ccenopus, the Ancestral Rhinoceros. 51 



ungulate evolution extending throughout the Oligocene. 

 Clearly separated from the Hyracodontidse and Amyno- 

 dontidse, this gave rise to all later forms, including the 

 modern genera of rhinoceroses. 



Four new subspecies are proposed: Ccenopus alius is 

 noted for its strong protoloph and weak metal oph; C. 

 nanolophus is so named from the dwarfed or incomplete 

 metaloph on the premolars and the large posterior fossa. 

 C, metalophus reverses the conditions of the others men- 

 tioned and has a metaloph as long as the protoloph and 

 the two are parallel. This is probably one of the progeni- 

 tors of C. tridactylus. C. avus, although a valid sub- 

 species of C. tridactylus, yet shows a very definite simi- 

 larity to Diceratherium Marsh and is ancestral to it. It 

 is notable for the parallel cross lophs, complicated enamel 

 bands, and its late Oligocene age. 



SCIENTIFIC INTELLIGENCE. 

 I. Chemistry and Physics. 



1. A Recalculation of the Atomic Weights. — Dr. Frank 

 WiGGLESWORTH Clarke, cliemist of the U. S. Geological Survey, 

 has now published the fourth edition, revised and enlarged, of 

 this important work in the form of a volume of 418 quarto pages. 

 The data of atomic weight determinations are presented very 

 fully and the various ratios, with the exception of some that 

 appear to be undoubtedly inaccurate, are calculated in con- 

 nection with their mathematical probable errors. Then the 

 available ratios are combined, with weightings according to 

 their probable errors, in order to find the atomic weights. The 

 author admits certain deficiencies in this mathematical method 

 of calculation, due to the effect of constant errors in the deter- 

 minations, such as the effect of impurities in the substances 

 weighed, but it appears that his method is the best one available. 

 A final table of atomic weights is presented in which the greater 

 part of these constants are carried out to five figures. The table 

 varies in this last respect from that of the International Com- 

 mittee, and there are also other appreciable variations in the two 

 tables. For instance, Clarke gives 27.039 for aluminium, while 

 the International Tables give 27.1, and it is interesting to observe 

 that Richards has found, just recently and too late for Clarke's 

 use in the work under consideration, that this atomic weight is 

 probably about 26.96. 



