1906.] MR. R. SHELFORD O^ " FLYING " SNAKES. 229 



the latter by virtue of a pronounced concave surface is buoyed up to 

 a certain extent, and very frequently its fall terminates in a slight 

 upward swoop, so that it reaches the ground with but little violence 

 of impact. The same holds good, as I believe, for those snakes 

 that can convert their cylindrical shape into the semblance of a 

 split bamboo. A specimen of Chrysopelea ornata was taken to a 

 height of fifteen to twenty feet and allowed to fall several times ; 

 after one or two false starts the snake was felt to glide from the 

 experimenter's hands, straightening itself out and hollowing in 

 the ventral surface as it moved, and it fell not in a direct line to 

 the ground, but at an angle, the body being kept rigid the whole 

 time. The height from which the snake fell was not great 

 enough for it to be possible to determine with any accuracy 

 whether it fell more slowly than when it fell in iri-egular coils, but 

 it certainly appeared to be so. If the snake was thrown up into 

 the air, it seemed unable to straighten itself out ; it had to be 

 laimched, so to speak, from the hands in order to induce it to 

 assume the rigid position ; and no doubt in its natural haunts the 

 snake prepares itself for a parachute flight by gliding with some 

 force from off a branch, and does not fall in the casual manner of 

 such a species as Tropidonotus inaculatus, 



Text-fio-. 57. 



Diagrammatic transverse sections of tlie body of Chrysopelea ornata. 

 A, in the normal condition ; B, driring " flight." 



It was not until 1904 that another Dyak collector brought me 

 a specimen of Dendi'ophis pictus, with the assertion that he had 

 witnessed its "flight" from a tree; the story of this quite inde- 

 pendent witness was to the effect that he had seen the snake 

 shoot out from a tree and fall at an oblique angle to the ground, 

 its body being held straight during the fall. This species also has 

 the hinged ventral scales characteristic of the genus Chrysopelea, 

 but whereas Chrysopelea belongs to the Opisthoglyphous group of 

 Oolubrines, DendropMs is one of the Agiypha ; it is larger than 

 either of the Chrysopelea}. Experiments that were carried out 

 with this species did not prove so conclusive as those with 

 C. ornata, but it was observed that if the snake was held up by 

 the tip of the tail the ventral surface of the body became concave 



16* 



