240 BULLETIN 114, UNITED STATES NATIONAL MUSEUM. 



gaster group of the present genus, the genus Elaflie, in particular, and 

 numerous other genera. This, hov/ever, is an assumption that must 

 be proved. SimpHcity and speciaHzation in color pattern can not be 

 decided on a 'priori grounds. It is not safe to say that because a 

 certain type of pattern is primitive or specialized in one or several 

 genera that it is so in any given case. Close examination may bring 

 out reasons why the rule docs not hold in a particular instance. 

 Thus we have a spotted pattern in Brymobius margaritijerus and in 

 the three forms of the getulus group, floridana, JioTbroolci, and splerv- 

 dida. In Brymobius there is good reason for considering this pat- 

 tern specialized, namel}^-, as belonging to an end form. As these 

 are all closely allied and all different, they are probably not all 

 specialized, or, if so, some are less specialized than others. In the 

 case oi floridana geographic probabilities alone are enough to place 

 its pattern as recent, and as derived from something else, even if we 

 failed to note that it still bears the vestiges of an earlier pattern; 

 and careful study of geographic probabilities and structural and pat- 

 tern differences have shown that the pattern of TiolhrooM is a deriva- 

 tive of that of splendida, and that the pattern of the latter is the only 

 one from which all the others in the group may be derived. Thus 

 we would have been in error had we said in the first place that because 

 the pattern of splendida looked similar to that of Brymobius that it 

 was specialized and not a suitable starting point for the derivation 

 of other patterns. A type of pattern therefore that is specialized in 

 one form may prove, for reasons quite evident upon examination and 

 comparison, to be primitive in another. 



In the case of triangulum there are numerous real difficulties in 

 the way of considering this form as primitive. It should be noticed 

 that (1) the pattern is best and most perfectly developed only in a 

 region that can not have been a center of preservation of reptilian 

 life, the glaciated portion of the northeast; (2) in the southeast, from 

 West Virginia and Maryland to Florida, it is erratic and not typical; 

 (3) it is much less stable and widespread than the ringed type of pat- 

 tern, and in the central Atlantic States it is undergoing degeneration 

 in company with structural reduction; (4) no other form of the 

 triangulum group presents any approach to this pattern except syspila, 

 the nearest relative of triangulum; (5) if triangulum be considered the 

 stem, form of the group the explanation of elapsoides becomes ex- 

 ceedingly difficult. It is well known that the pattern and color of 

 some snakes change dm^ing development from young to adidt, and 

 the 3^oung stages in some cases are supposed to indicate ancestral 

 conditions. This is the case with the racer. Coluber constrictor, 

 which changes from a spotted to a plain black or bluish snake 

 during growth from young to adult. In the young of conjuncta the 

 body pattern can not be told from that of boylii, but that of the 



