the Coloration of Egys. 537 



prey tliat is being- particularly experimented with to avoid 

 an undue craving for it. Naturally the experimenter must 

 also know the principles on which an animal feeds : uiy own 

 early experiments, before I gradually learned these and 

 discovered the various complicating factors that must bo 

 eliminated or watched for, were worth very little. He must 

 realize that a hungry animal will eat almost anything that 

 comes within the category of its natural prey, and not jump 

 to the conclusion^ when it does so, that it is necessarily 

 indiscriminating at all times. He must know something of 

 the animal's state of hunger at the outset and recognize the 

 symptoms of growing repletion, must be able to follow the 

 twists and turns of an appetite increased a little through ten 

 minutes' fasting, decreased through a further morsel of food, 

 or, it may be, stimulated suddenly by a specially savoury 

 ofiering or depressed by an unsavoury one. And so on. 

 It is insufficiency of knowledge on such points that has 

 vitiated some food-experiments in the past and caused the 

 results obtained sometimes to seem contradictory and 

 unreliable. That the method may, carefully used, yield 

 perfectly reliable results is indicated, I think, by the fact 

 that in my own experiments wild insectivorous birds corro- 

 borated in the most ample manner the results I had obtained 

 from captive birds, while totally unrelated carnivoroxis 

 mammals (one of them unconfined) confirmed generally 

 each other's verdicts on bird and mammal prey. That 

 my egg-experiments only (and these the last to be carried 

 out, with the precautions suggested by their predecessors) 

 should, for some reason, be unreliable, is, indeed, perfectly 

 possible, but not very likely, though they certainly require to 

 be corroborated, as my insect-experiments were, in the field. 

 (2) " Non-corroboration from field-observation.'' In 

 insects this objection has been brought against two views : 

 (a) that birds eat butterflies ; [b) that insectivorous birds 

 have preferences. In each case the abundant results of 

 specially-undertaken observation at once showed that the 

 reason for the paucity of evidence had been a previous lack 

 of special observation. Yet entomologists had had their 



