8 



of the northeastern States, and three have adopted burrowing habits 

 (elapsoides, virginiana, and rliomhomaculata) , They may all be 

 classed as truly land snakes; so far as known none seek their food 

 in the water. They have not the stout body of the snakes that 

 depend for safety upon inconspicuousness or poison, nor the attenu- 

 ated body and tail of the tree snakes and racers, but rather a well- 

 proportioned body and short tail, adapted to life on the ground and 

 more or less in the open. They seem in general to be fearless and 

 self-reliant. Their immunity to snake venom and their powerful 

 constricting ability j^ender them truly kings among the reptiles of 

 North America. Much remains to be learned about the food prefer- 

 ences of the various forms, but in general it may be said that they 

 are the enemies of all small rodents and of all snakes and lizards; 

 fledgling birds are undoubtedly eaten on opportunity. 



But little is knov/n about hibernation and reproduction. The 

 fullest accounts will be found under the descriptions of triangulum 

 and getulus. Probably all are oviparous. 



Variation. — Since so little careful study has been made of variation 

 in scutellation and color pattern in snakes, it is unsafe to generalize 

 on the systematic value of any of these points without a thorough 

 knowledge of the particular situation. Ruthven has shown, in his 

 excellent review of the garter snakes (1908), the extent and meaning 

 of variations in the genus TJiamnopMs , but a comparison with the 

 present review will shov/ that not the same value can be placed in 

 all instances upon variation in any particular characteristic. Thus 

 the forms of TJiamnopliis exhibit great similarity in color pattern, 

 while in both series of labials there is much variation of specific value. 

 Lampropeltis, by comparison, exhibits the widest variation in pattern, 

 while there is remarkably little variation in the labials. This is but 

 one of numerous possible illustrations that color pattern may be less 

 variable than scutellation. It seems usually to be more so, but, as 

 said above, it is not safe to generalize upon the value of characters, 

 for what is true in one case may be quite otherwise in another. The 

 relative value of characters must be determined separately for each 

 group of genetically related forms. By a careful study of the extent 

 and meaning of variations in different characters in different genera 

 we may learn something more definite of the genetic relations between 

 genera, as we are now able to learn these relations between species 

 and subspecies. It is a lack of knowledge of the meaning of differences 

 between genera that results in errors and uncertainties as to their 

 relations. For instance, it is not now known what meaning to place 

 upon differences in method of change in number of scale rows. The 

 present review can not, of course, determine this, but it serves to 

 confirm the results obtained by Ruthven in the garter snakes to the 

 extent of showing the method of change in number of scale rows to 



