VARIATION IN EPIDERMAL COLOR OF NAJADES. 253 



smaller numbers. Against such treatment however was the check 

 of a separate color comparison for each shell. By means of this 

 method, it was possible to determine for each species shades of 

 epidermal color peculiar to the locality where the shells had been 

 collected, and this being done, charts were prepared showing the 

 sequence of color changes passing down stream, or the distribution 

 in different parts of Ihe same body of water. This data, in con- 

 solidated form is presented in PI. III. Even by this process of 

 condensation, a very large number of colors was obtained for each 

 species, making it imperative to simplify further in order that the 

 evidence for the relative prevalence of different epidermal colors 

 in those species where more than one was described might be 

 rendered more intelligible for report. Just as the systematist for 

 rough descriptive purposes has picked out a number of the more 

 prominent epidermal colors of each species, the writer, following 

 largely Simpsons Descriptive Catalogue of the Naiades, (17), 

 chose from the previously prepared charts the ten to thirty leading 

 colors, (primary or secondary), in the epidermal colors of each 

 species to which the large majority of the rest could be assigned. 

 Percentages of these leading colors were then calculated for each 

 species in the bodies of water, drainages, groups of drainages in 

 which they were found, as best seemed to throw light on the prob- 

 lems to be attacked. While all recorded shades in a large number 

 of cases would not conform to this treatment, they represent 

 percentages in the extreme minority, and may be inferred to exist 

 in those species where the tables of Distribution of Colors as a 

 Whole does not add up to 100%. Strictly speaking, even this 

 comparatively large number of "leading colors" could have been 

 condensed to a smaller number, but the larger number was necessary 

 in order that certain close distinctions in the colors of the epidermis 

 for the purposes of the investigation might be made, for exami.le, 

 between the colors of shells from a river and those from its tribu- 

 taries. In the discussion of any particular body of water, however, 

 the leading colors given, represent my reduction to lowest terms 

 of the colors represented in it. 



Data on the sex of the animal, prevalence of rays, etc., were taken 

 at the time color comparisons were made. So far as observations 

 on epidermal color as associated with the sex of the animal are 

 concerned, the small number of shells on which they are based is 

 explained t)y the f^ct that the specimens were collected before Dr. 



