1018 REPORT OF NATIONAL MUSEUM, 1898. 



The following table represents the geographical range of the species : 



The above table gives but a very general view of the distribution of 

 the species, siuce some of them are restricted to districts of the regions 

 only, while a few species are known from so few examples that their 

 range is unknown. Of the latter class are U. butlerii, E. hr achy stoma,, 

 E. rutiloris, E. angustirostris, and E. nigrilatus. The E. sackenii is 

 restricted to Florida, and the E. radix to that part of the central 

 region that lies east of the Rocky Mountains, entering the eastern 

 region. The widely distributed species, as E. sirtalis and E. elegans, 

 are represented in special districts by peculiar subspecies, which look 

 very different from each other. The E. proxima has a rauge which 

 does not coincide with any zoological district, inhabiting eastern 

 Mexico, Texas, and the Mississippi Valley. 



The study of the several hundred si^ecimens of species of this genus 

 which are contained in the TJ. S. National Museum and my private collec- 

 tion shows that in most of the species the number of rows of scales and 

 the number of the labial plates are quite constant. In only one species, 

 the E. leptocephala, is the number of scale rows varied by the i^res- 

 ence or absence of a single row on each side, and in none is the 

 number of labial plates frequently variable. The position of the lateral 

 stripe is, as stated by Baird and Girard, very constant. The relative 

 length of the tail is constant within certain limits and in certain 

 species. In some of the species it varies a good deal. The coloration 

 varies within limits in each species, and often characterizes subspecies 

 with considerable precision, transitional forms in some such cases being 

 rare and in others more frequent. The species of the Pacific coast 

 present the greatest difficulties to the systematist. Here the eastern 

 E. sirtalis comes into contact with the western E. elegans, and some 

 close parallels in coloration occur. Thus the E. s. parietalis resembles 

 very much the E. elegans ordinoidcs, and the E. s. sirtalis resembles 

 considerably the E. c. lineolata. Tlie E. infernalis intervenes between 



