EXPLANATION OF FORM AND COLOUEINQ. 219 



captives, so accurate and instinctive does the judgment by taste even among 

 young birds become after a very little practice. 



Mistaken refusals too. Thus it seems quite probable that, given an equally 

 perfect view, a wild bird would, as the result of greater practice, distinguish 

 between the commoner models and mimics somewhat more readily on the 

 average than did my more long-caged birds. Yet, as I will show, even wild 

 birds are deceived — and they seldom have so good a view as I used to go 

 out of my way to provide for my captive birds. 



3. Through surfeit in a particular direction. I have had cases — very few 

 indeed that mattered — in which probable over-eagerness for a particular 

 order has been displayed. To these cases I will draw special attention as 

 I come to them. 



The above are three difficulties that are likely to confront even a careful 

 experimenter on captive animals. The first two can be partly remedied and 

 entirely allowed for, and the third can be remedied : none of them need be 

 allowed to vitiate one^s deductions in any appreciable degree. 



A fourth difficulty is worth stating here, though it has certainly had no 

 especial connection with captivity. The animal experimented on would 

 sometimes appear suddenly to awaken to a realization of greater appetite ; it 

 would fall back (sometimes with an actual " brisking up " of manner, which 

 at once gave the clue) to a point already left behind and once more begin to 

 take things that were already becoming unacceptable to it. Even amongst 

 my insectivorous birds this was probably in most cases equivalent to that 

 marked increase in appetite that I later found it possible — and useful — to 

 provoke at will in my carnivorous animals by the offering of something 

 sufficiently pleasant to stimulate, it would seem, a flow of the digestive secre- 

 tions. Conversely, repletion — or the symptoms of it — might be produced 

 prematurely by the eating (in, at any rate, one of my mammals merely the 

 smelling) of some unduly indigestible species of prey. 



To this phenomenon — the sudden stimulation or inhibition of the gastric 

 secretion — may be attributed, I believe, nearly all the cases of apparent incon- 

 sistency that have occurred in the course of these experiments. Its occurrence 

 demands increased caution in the deduction of preferences generally, but, as 

 I have said, there seems no reason to suppose that it is a phenomenon peculiar 

 to or aggravated by the captive state. 



In conclusion, it must in all fairness be acknowledged that the only absolute 

 proof that a captive animaPs preferences are the same as those of that animal 

 in the wild state would be that afforded by a testing of the preferences of a 

 single individual, (1) in the wild state, (2) after capture ; and I have often 

 regretted that it did not occur to me to attempt this in the case of my wild 

 bush-shrike (Dryoscopus guttatus). But even the evidence I have obtained 

 (and referred to above), namely a rough general agreement between unrelated 

 species in captivity and, yet more important, the rough confirmation of these 



LINN. JOUEN. — ZOOLOGY, VOL. XXXIII, 17 



