14 BULLETIN 17 5, UNITED STATES NATIONAL MUSEUM 



number throughout the length of the body. The maximum number 

 for any given individual occurs slightly anterior to the middle of the 

 body, and the number decreases both anteriorly and posteriorly, the 

 minimum number of rows occurring just anterior to the vent. Thus, 

 to represent the scale rows of any individual a formula of three 

 numbers is used, e. g., 29-33-23. In such a formula the first number 

 represents the minimum number of rows in the anterior part of the 

 body, generally found a short distance posterior to the neck; the 

 second number represents the maximum number for the individual; 

 and the last, the minimum number in the posterior part of the body. 

 It is obvious that the reduction in the number of the scale rows must 

 indicate the loss of some of the rows. If on a typical specimen the 

 scale rows are counted at frequent intervals throughout the body 

 length, it will be found that generally two lateral rows (one on each 

 side) are dropped simultaneously, and if the sides of the body are 

 carefully examined the actual point where the loss occurs in each 

 case can be discovered, and the rows involved can be determined. 

 Thus, the formula of a typical specimen as given above, in order to 

 represent the true situation, should read 29-31-33-31-29-27-25-23. 

 For convenience, this formula is shortened to include only the maxi- 

 mum and anterior and posterior minimum numbers, the others being 

 understood. It will be noted that in occasional specimens one or 

 more numbers in the formula are even, rather than odd, as normally. 

 This peculiarity is due to an abnormality in the method of reduction, 

 by which either a single vertebral row is lost somewhere on the body, 

 or a single lateral row is lost in one reduction. For a given specimen 

 the maximum number of rows lost anteriorly is 6, posteriorly 12, 

 while the minimum number is none anteriorly and 5 posteriorly. 

 The amount of variation in the number of scale rows is very great, 

 not only within the genus as a whole but in many of the separate 

 forms as well. Thus, the range of variation in sayi, one of the most 

 variable forms, of from 25-28-23 to 35-37-27, is nearly as great as 

 the range of from 25-27-19 to 35-37-27 for the entire genus. 



Ruthven (1908, p. 17) found that in Thamnophis the reduction in 

 the number of scale rows was due to the loss of definite rows through- 

 out the genus. His results may be expressed in the following diagram, 

 in which the upper row of figures represents the numbers of scale rows, 

 and the lower series the scale rows concerned in each change, the num- 

 ber being counted dorsaUy from the ventral series of scutes: 



23 21 19 17 15 



5 5 4 4 



In Lampropeltis (Blanchard, 1921a, p. 9), Coronella, Vipera, Tropi- 

 donotus (Proctor, 1920, p. 357), Coluber, Masiicophis (Ortenburger, 

 1928, p. 8), Elaphe, Natrix, and other genera, it has been demonstrated 



