SNAKES 81 



along its length, convex on one side, concave on the 

 opposite side. Now if a cylinder of bamboo and a 

 split length of bamboo of equal weight be dropped 

 from a height, the former will reach the ground before 

 the latter, provided that the latter be dropped with 

 the concave surface directed downwards, for by virtue 

 of its concave surface it will be buoyed up to a certain 

 extent ; now, it seemed to me that the same might be 

 the case with the " flying " snake. So I took the snake 

 up to the verandah of the Museum and threw it into 

 the air, but I was disappointed to see it fall in writh- 

 ing coils to the ground, which it hit with a distinct 

 thud. Then I allowed the snake merely to fall from 

 my hands to the ground, and after one or two false 

 starts eventually I felt it glide rapidly through my 

 hands, straightening itself out, and hollowing-in its 

 ventral surface as it moved ; this time it fell not in a 

 direct line to the ground but at an angle, the body 

 being kept rigid all the time. The height from which 

 the snake descended was not great enough for it to be 

 possible to determine with any accuracy whether the 

 snake fell more slowly than when it tumbled in 

 irregular coils, but this certainly appeared to be the 

 case, and there can be little doubt that the hinged 

 ventral scales of these snakes, enabling them to draw 

 the belly inwards, are a modification of structure render- 

 ing a parachute flight possible. Some years later I 

 gained an indirect confirmation of the "flight" of 

 snakes, for another Dayak hunter brought me a speci- 

 men of Dendrophis pictus, and asserted that he had seen 

 it "fly" down from a tree. On examining the snake 

 I found that this species also possessed hinged ventral 

 scales, and that it too had the power of hollowing-in 



7 



