1901.] NATUKAL SCIENCES OF PHILADELPHIA. 11 



found in the teeth, in the plates upon the head, and iu the numljer 

 and form of the rows of body scales; all of these are numerous, 

 and variability under the above law is common. 



Color is largely used in specific, and almost wholly in subspecific 

 determination, and this, too, we should expect to find inconstant 

 in a group whose structure is such that the whole exterior is 

 brought into close contact, surface or subterranean, with earth, 

 sometimes swamp and sometimes desert sand, and whose slow 

 metabolism brings such physiological activities as temperature, 

 nutrition and epidermal repair into close dependence upon exter- 

 nal conditions. 



There is, again, a class of anomalies not uncommon in this 

 group, such as are shown^ at times in genera like Coluber or 

 Zamenis, in which the young of some species are spotted or cross- 

 banded, becoming striped when adult. Here, occasionally, more 

 or less of the juvenile pattern is retained, showing through, as it 

 were, the later stage. Examples of such are Coluber guttatus 

 ■sellatus Cope and C. rosaceus Cope. This class of variations is 

 purely physiological, and when occurring in isolated cases, has no 

 more zoological significance than the occasional retenlion to matu- 

 rity of the youthful livery of spots in lions. 



Aside from anomalies, there are characters which are too variable, 

 normally, to be of use except in broad definitions. Form and 

 proportions, both of the whole and of parts, vary considerably; 

 among those which change with growth are the relative length of 

 the tail (which also varies with sex), and the reciprocal propor- 

 tions of some head plates; breadth of head and stoutness of body 

 change to an extraordinary degree with nutritive conditions, a 

 fact which can be best learned by observation of snakes kept in 

 captivity.* 



The system of trinomials has added greatly to the facilities for 

 expressing the relationship of transiti(mal forms, but while its 

 value is fully conceded, so also must be the existence of the danger 

 which has attended and not infrequently overtaken it — that the 

 very ease of its methods may lead the systematist to overvalue the 

 importance of individual and insignificant variations. 



^ A suggestion as to the possible origin of occasional specimens presenting 

 mixed characters, is that among snakes which breed iu captivity there seems 

 to be little orno aversion to cnss-breeding. This is especially true of 

 Eutmnia and Tropklonotus, both of which produce young free from the 

 egg, and breed not infrequently. 



