10 TROCEEDINGS OF THE ACADEMY OF [Juu , 



A REVIEW OF THE GENERA AND SPECIES OF AMERICAN SNAKES, 

 NORTH OF MEXICO. 



BY ARTHUR ERWIN BROWN. 



In recent years, investigation of the lower groups in classifica- 

 tion has largely taken the form of observing and noting the most 

 minute variations, occurring in however small numbers. Among 

 snakes, this method has lieen carried to such an extreme that 

 Prof. Cope's " Characters and Variations of North American 

 Snakes " ^ contains the names of twenty-three species and subspe- 

 cies which were founded upon one, or at most two specimens each. 



Two propositions, both fundamentally correct, have contributed 

 to this result: first, that a knowledge of the laws under which new 

 forms are developed is to be ])est gained by a study of variations ; 

 and second, that subspecies are an essential part of classification. 

 As a general truth the first proposition is unassailable, but there 

 appear to be good reasons why limits should be placed upon its 

 application to the present group and why a cautious valuation 

 should here be made of minor variations. This should be true if 

 it can be shown that unmeaning departures from type are especi- 

 ally common among its members. 



It is a law of organisms that a high degree of instability is 

 associated with degenerative processes. That the serpents, as a 

 whole, are a degenerate group is probable, and while some lines 

 among them have become much specialized, there are large num- 

 bers of small and degraded forms, always highly variable, which 

 can be connected with higher types. 



It is, furthermore, a moi'phological fact that where repetition of 

 parts is the rule, variability, in number at least, is to be looked 

 for.'^ Among snakes, generic and specific characters are chiefly 



' Proc. U. S. National Museum, 1892, pp. 589-684. 



* A very simple summary of a long series of observed facts is contained iu 

 Bateson's Materials for the Study of Variation, p. 571 (London, 1894): ''It 

 is perhaps true that, on the whole, series containing large numbers of undif- 

 ferentiated parts more often show Meristic Variation than series made up of 

 a few parts much differentiated, l)i]t throughout the evidence a good many 

 of the latter class are nevertheless to be seen." 



