Morphological Variations in Vipera berns. 3 



" (a) The Syncipital Shields. — In British specimens they 

 are neariy always well developed. In one specimen only 

 (from Hampshire) I find the frontal much reduced and the 

 parietals broken up into scales, the specimen in this respect 

 agreeing with V, asjois . . . Any such extreme speci- 

 mens . . . are worth recording and preserving. 



" (d) The number of labial shields, not being always the 

 same, has to be counted on both sides of the animal. In 

 35 out of 74 cases I find 8 shields, in 30 cases 9, in 4 cases 

 10, in 4 cases 7, in 1 case 6. Continental specimens have 

 much more frequently 9 than 8 shields. F. aspis has from 

 9 to 11. 



" (e) The number of scales round the eye (prae-, sub-, and 

 post-oculars) varies between 6 and 11 in British specimens, 

 the two extreme numbers occurring only once in my lists, 

 8 or 9 being the usual number. 



" (6) The Scaling of the Body. — The scales number 21 across 

 the middle of the body. But there are exceptions. In one 

 specimen from Petersfield I counted 23 scales, and in another 

 from Scotland only 19. . . . Such exceptional specimens 

 should be recorded and preserved. 



" (6) The Ventral Shields. — In 37 British specimens I have 

 counted 137 to 146 ventral shields (exclusive of the anal) in 

 males, 139 to 154 in females. . . . The limit of variation 

 on record (sexes not discriminated) is from 124 to 159. 



" (7) The Suh-caudal Shields. . . The number (count- 



ing each pair as one, and not reckoning the terminal, conical, 

 or spine-like shield) is 35 to 40 in males, 28 to 35 in females. 

 In one male I find as few as 33. The number of sub-caudals 

 is stated to vary between 25 in males and 48 in females." 

 (Quoted from the Zoologist, March 1892.) 



The following records of scaling variations in my own 

 specimens will show how far they agree with and differ from 

 those hitherto recorded. The actual work of counting these 

 scales is most laborious and tedious, in fact, it is quite an 

 hour's work to count and verify by recounting the scaling 

 of a single specimen, as the slightest distraction of the 

 attention is apt to lead one to make a mistake. For greater 

 convenience of reference, I have arranged my results in 



