the Coloration of Eggs. 539 



had a Roller in my aviary. I frequently went in and fed it 

 by hand. On a few of these occasions I offered that large, 

 gaudy, evil-smelling locust, Plnjmateus. It was each time 

 readily eaten. I also found it in the stomachs of one or two 

 wild Rollers. I concluded, very naturally (and I believe I 

 was rash enough to publish the conclusion) that Rollers are 

 probably liighly indiscriminuting. A few years later I began 

 definitely to experiment on Rollers. Then I found that, 

 while they readily ate Phymateus up to a certain point in 

 the satisfaction of their hunger, that point was by no means 

 a very advanced one, and that from that point up to repletion- 

 point they refused it with symptoms of dislike. Again, my 

 mongoose sometimes left the eggs that formed part of his 

 ordinary diet uneaten for quite a time. Had this occurred 

 in the days of my first Roller I would have perceived no 

 significance in the fact. The other fact, that they were 

 always sooner or later devoured, would have satisfied me 

 that they were all much liked, and the delay, if I thought of it, 

 would have been put down to repletion. Had I experimented, 

 I would sometimes have found (as I sometimes did find) 

 that the mongoose was far from repletion and eager for 

 food — but not as yet for that particular egg. 



When told of acceptances, we should ask : how hungry 

 was the animal? What went just before — was there, per- 

 haps, special stimulation ? Could there have been a special 

 craving through the insufficient presence of that class of 

 food in the everyday diet ? Were not the eggs, perhaps, in 

 anv case all of a high-grade nature? Unless the first three 

 questions can be satisfactorily answered, the acceptances, 

 whether by wild animals or tame, are completely valueless 

 as evidence of indiscriminateness. No number of unchecked 

 records of eggs eaten can ever show that those eggs are not 

 sometimes, perhaps often, refused. 



Similarly, acceptances that are not seen to be accompanied 

 by neglect or rejection of other species — or their relative 

 immunity in the same locality — are quite useless as evidence 

 of preference, and this is why such evidence must always be 

 hard to procure in the field. It can only be obtained in 



