BEHAVIOR AND COLOR CHANGES OF TREE FROGS 59 



the later accounts of experiments similar diversity was evinced 

 under identical tests. 



But not only the anomalous color features call for explana- 

 tion ; there is also the hardly less anomalous peculiarity of 

 habitat already referred to, which likewise is a matter of beha- 

 vior the meaning of which is not at all obvious. I do not forget 

 that occasionally these creatures may be found on old fences 

 or walls. Still it is exceedingly rare to find them in such places. 

 In all my observations, covering many years in a general way, 

 only once have I actually found a specimen in such habitat, 

 and then it was on an old rail fence not far from the wood. In 

 this case it was in mid-summer, and the creature seemed quite 

 at home in the devious labyrinths of light and shade, etc. This, 

 however, is quite a different story from that here under review, 

 and still leaves doubtful the why and wherefore. 



Of course, the contiguity of the wood-side might suggest the 

 possibility of descent from the trees looking toward hiberna- 

 tion. But when first observed it was in early September while 

 the weather was still very warm, the trees in full foliage, and 

 food quite abundant; too early apparently for the hibernating 

 instinct to manifest itself. Nor could it be prompted by quest 

 after food ; for of all places this would seem the most unpromis- 

 ing. Even the elaborate webs of autumnal spiders were con- 

 spicuously absent on this fence. Nor on the other hand, could 

 it have been prompted by unusual satiety of appetite, for when 

 placed in the cages they were quite ready and eager to take an 

 insect. I merely submit the facts without further comment, 

 and leave the problem open to further light. It may be noted 

 in passing that during one fall not a single specimen was found 

 in this locality, even after diligent search. As hinted above, I 

 am still disposed to regard the matter as in some way, not very 

 clear, related to the approach of the hibernating season, the 

 first expression of which on the part of the frogs would prob- 

 ably be descent from the trees. But even at best this is merely 

 a guess. 



In an account of color reaction in which an attempt is made 

 to measure or determine with some approximation the amount 

 of color change which takes place at a given time, or under a given 

 test, it is very desirable that one have some basis of comparison. 

 In the case of Hyla this is somew T hat difficult. It has seemed, 



