212 MR. C. t: M. SWYNISTERTON ON TfiE 



stomach; yet a captive hee-eater refused to touch them, emphatically and 

 repeatedly, though she readily accepted and ate certain other insects. 



This observation by itself would be most puzzling, and might well help to 

 discredit iitterly experimentation on captive birds— as might any one oH 

 various other observations o£ exactly the same kind that I have in mind. 

 Yet my bee-eater only thus refused bees when a certain point in the satis- 

 faction of her appetite had been reached. Up to that point I found that she 

 ate them readily, and careful and repeated observations on numerous wild 

 bee-eaters of the same species (Alerops apiaster) showed that they too appa- 

 rently only ate hive-bees up to a certain point, thereafter ignoring all they 

 were seen to meet, but attacking and eating various other species of insects. 



(b) I have myself accused rollers of being/, probably, indiscriminate feeders 

 on the strength of such strongly-smelling insects as the grasshopper, Phymateus, 

 having been foxmd in their stomachs and because a roller in my aviary 

 (Coracias olivaceiceps) accepted and ate them on several occasions ivith 

 apparently the greatest relish. 



I had no knowledge of the state of repletion of the latter roller, nor could 

 I possibly know the exact state of hunger or otherwise of the wild roller 

 whose stomach was examined, at the moment each ate his Pliymateus. And 

 I now realize that, without that knowledge, I was utterly and completely 

 unjustified in coming to any conclusion at all, either as to the insect being 

 highly pleasant or the roller indiscriminate. Offered to C. garridus (B), 

 the Phymateus was eaten, it is true, but only up to a certain point in the 

 satisfaction of the bird's hunger, and not beyond it. The finding of a par- 

 ticular insect in the stomach of a wild bird, even frequently and in large 

 quantities (as in the bee-eater instance just given), at most proves that that 

 species of bird will eat it when hungry enough. It in no way invalidates 

 such conclusions as may be suggested by the rejection of the same insect by 

 the same species of bird, captive or otherwise, when it is not hungry enough 

 for it, 3^et is hungr}^ enough for certain other species of insects. It can be 

 seen, too, from what I have said, how futile any attempt must be to deduce 

 the status of, say, the Acrasinas used in any series of experiments by merely 

 counting up their total rejections against their total acceptances. If they 

 were offered mainly to hungry birds the acceptances probably would be more 

 numerous than the rejections, but this would not indicate that the Acrseinas 

 are as digestible as things eaten to nearer repletion-point, or that they would 

 ever be eaten at all under conditions other than those under which they were 

 offered — a nearly emj^ty stomach and a rousing appetite. 



Stomach and pellet investigation must, I think, remain our greatest and 

 most valuable source of information with regard to the general preferences of 

 a species, the orders it usually feeds on, and the insects it mostly " fills up " 

 on, as a bee-eater does on bees. But facts like the above seem to indicate 

 that our knowledge of its detailed preferences must come in the main from 



