432 L. D. BURLING CAMBRIAN AND ORDOVICIAN BRACHIOPODA 



the percentage of shale localities is 27, but the non-accordance of the 

 other percentages would place this as a mere coincidence. Perhaps the 

 most noticeable feature is the uniformity in the figures for each of tiie 

 types of sediment; in fact the figures were so general that they were be- 

 lieved to indicate little or nothing until the study of the number of si)e- 

 cies per locality was completed. This number varied from 1 to 15 in 

 both sandstone and limestone and from 1 to 13 in shale, and the averages 

 for the groups betrayed so little evidence of system that the ratio between 

 them was computed. Thus the ratio between the number of species per 

 locality in shale and limestone for Group C is .892 (^) ; for Group B 

 it is .896 (f^). The ratio between the iiuml)er of species per locality in 

 shale and limestone for Group A was .765 (ij^), a number differing so 

 greatly from the ratios for Groups B and C that the ratios between the 

 number of species per locality in sandstone and limestone was computed 

 for Groups A and B, giving .985 (|^) in the former and .982 (};^) in 

 the latter case. For Group C this computation was of course impossible, 

 owing to the absence of collections from sandstone. In view of the mul- 

 tiplicity of the problems involved and the entire lack of int^rrelatiou be- 

 tween the groups into which the localities were divided, the ratios afford 

 fairly clear evidence of the accuracy of all of the figures except 2.1."), ilic 

 number of species per locality in shale for Group A, and would seem to 

 indicate that the number of species per locality is smaller in shale than 

 in sandstone and greatest in limestone. 



The accordance of the results is all the more remarkable when we re- 

 member that the following factors, if they Were not more or less com- 

 pensatory, might easily have vitiated the results: (1) Errors in the iden- 

 tification of the sediment due to the presence in the collections of gra- 

 dational types whose classification was of necessity somewhat arbitrary; 

 beds neither predominantly arenaceous or calcareous, for example, which 

 might on second examination be referred differently; (2) errors in the 

 determination of the true nature of the inclosing sediment ])ecause of 

 local variations in the matrix immediately surrounding the fossil ; and 

 (3) errors in the identification of the species themselves. Most of the 

 determinations of sediment were made by examination of the hand speci- 

 mens on which the species are preserved, and in no case have these origi- 

 nal results been verified or changed. For this reason the one sandstone 

 7'epresentative of the genus Tluendla may either be incorretcly assigned to 

 that genus or may occur in a sandstone calcareous enough to justify its 

 transfer to the limestone column. On the other hand, both of these identir 

 fications may be correct and the genus may occur in both types of sediment. 



