DORSAL SCALE- ROWS IX BRITISH SNAKES. 361 



invariably coincides with the termination of series IV. ; and as the 

 corresponding points vary on the two sides of the snake, a variety 

 of combinations result, all of which have the effect of reducing the 

 number of scale-rows from 21 to 17 in aii abrupt and often highly 

 irregular manner. One such combination of the ending series 

 was as follows:— 21-20, lYth left side: 20-19, Xth right side; 

 19-18, IXth left side ; 18-17, IVth and llird united, right side 

 (see text-fig. 3). In some other specimens the penultimate series 

 ended over an inch anterior to the lYth, the regression of the 

 one pair thus overlapping the progression of the other. This 

 overlap is always present in specimens having 23 rows, as in 

 these the amount of regression remains the same, whilst the 

 progression of series V. and TV. has advanced reciprocall3\ 

 17-16. 

 Just before the commencement of the tail there is yet another 

 reduction caused by the vertebral row, which drops out usually 

 within about four ventrals of the end of the body. Occasionally 

 the loss is less advanced, and constitutes the first reduction on the 

 tail-proper; conversely regression may be more advanced, and in 

 one extreme instance this series ended 22 ventrals before the end 

 of the body. 



Anatomical reasons for the halting -points in the rechictiooi 

 and ])roduction of scale-rows. 



It is clear, since the size of the dorsal scale is more or less fixed, 

 that the number of rows is correlated with the calibre of the body, 

 and that this number must decrease or increase, if smaller or 

 larger forms are to be produced. Also, the series must vary in 

 number according to the taper of the body. The posterior reduc- 

 tions are amply accounted for by this, but the general taper of 

 the body seems quite insufficient to account for the constant 

 halting-point in reduction or production of two, in the case of 

 V. herus four row^'^, about two-thirds or three-fifths down the 

 body. 



Now the number of rows must be regulated by the normal 

 girth of the body ; or by the girth attained by necessary expan- 

 sion at certain points. It also stands to I'eason, if scale-roAvs are 

 to be dropped, in the case of reduction, that the abridgment Avill 

 commence where the loss in girth will be least missed, and pro- 

 gress in this manner, so that the portion of body where these 

 rows finally persist will be that part Avhere the skin has the 

 greatest strain imposed upon it. Conversely this will be the 

 point where the development of new rows will commence *. 



In T. natrix this portion is from the neck to a point just 

 beyond the middle of the body. In the Garter-snakes wdien 

 fvu'ther reduction is necessar}^ after the abridged lows have 



* In Naia, for instance, the move developed the hood, the more numerous the 

 scales. 



