FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 65 



neighborhood, but this is purely hypothetical. It is most reasonable 

 to conclude that all these animals did live together, at the same time, 

 witliin the same area. 



It is therefore to be expected that genera present in this quarry 

 will either have only one species each or will have species not inter- 

 grading at all and reflecting structurally their pertinence to different 

 ecologic stations in the area. With this in mind, the assumption was 

 made in dealing with each genus that it did include (in this quarry 

 fauna) just one species unless the contrary could be proved beyond 

 reasonable doubt. 



Matthew, in the paper cited above, and most other writers on the 

 question of species making in paleontology have insisted on making 

 due allowance for variation, or using for taxonomy only non variable 

 characters, but they have adduced no real, objective criterion as to 

 what "due allowance" should be, and they sometimes seem to over- 

 look the fact that there is no such thing as a truly and completely 

 "nonvariable" character. Not merely as mechanical, mathematical 

 procedures but as a general system of logic and a grouping m^ethod 

 useful both explicitly and as an implicit background for dealing with 

 both numerical characters and attributes, the methods of statistics 

 provide the desired means of measuring variation accurately and the 

 necessary criterion as to whether this variation is or is not of the sort 

 normal Avithin a species. These tests and this logical background 

 have been the basis for taxonomy in tliis study. If the specimens 

 pertaining to one genus could not indubitably be separated into differ- 

 ent groups, the conclusion has been that the fundamental hypothesis 

 of one species to each genus was correct. If they necessarily had to 

 be separated into different groups, and these groups could not be 

 interpreted as based on nontaxonomic differences (such as age or 

 sex), then and only then has the liA'^pothesis been discarded. 



Since tliis largely objective testing has intervened, it is not arguing 

 in a circle to start the study with the assumption that Cabrera's 

 Law applies, and then to consider the results as a test of the validity 

 and an example of the operation of that law. (See figs. 3 and 4.) 



The actual results are as follows: Tliirty genera are represented in 

 the Gidley Quarry by one species each. Since this includes the greater 

 part of the quarry fauna, in general it is true here that the related 

 (congeneric) animals living at that time in this area were of the same 

 species. The apparent exceptions belong to six genera, each of which 

 is worthy of brief special consideration. 



The multituberculates, with four species tentatively referred to 

 Ptilodus and three to Edypodus, are the most striking apparent 

 exceptions. In the fu'st place, however, the generic designations are 

 very dubious. It is not at all certain that the species referred to 



