Birch in Relation to their Prey. 37 



myself offered the species to have shown the very greatest 

 repugnance to it. 



My offering of butterflies to these birds will be criticised 

 by those who are only superficially acquainted with their 

 well-known bark-searching habits and with the habits of the 

 butterflies that inhabit the same wooded pastures. That 

 cockroaches, wood-lice, and bark-hauuting beetles and larvse 

 would have been letter may be at once admitted. But many 

 butterflies rest on bark, some, as Crenis and some Lycsenids, 

 habitually, whether temporarily or for their prolonged rest. 

 Again, this Irrisor, as I have occasionally seen, will sometimes 

 attack insects passing close on the wing ; and Danaida, 

 Acrcea acara, and various smaller Acrseinae, YpMliima, various 

 species of Precis and of some other Nymphaline genera, and 

 Terias and BeJenois amongst Pierines, are not only the most 

 conspicuous of such passing insects, but they are so 

 abundant in Irrisor s, hunting-ground that they are really 

 quite likely to be the fairly frequent mark of such attacks, 

 experimental or for food, as I have referred to. My own 

 experience of the bird by no means leads me to endorse the 

 view that it necessarily feeds chiefly on cockroaches, and the 

 stomachs of six specimens collected for me by Odendaal in 

 1907—8 also suggests a very varied diet. The contents are 

 spread before me as I write this, and a glance over them shows 

 that beetles come easily first in importance — beetles of various 

 families, and not all of them of bark-haunting species. 

 Next come larvre, including two moth-larvpe that feed on 

 the foliage of Brachjstegia. Grasshoppers, spiders, green 

 Pentatomid bugs that frequent foliage, cockroaches — one or 

 two of each, — and a Tipulid fly (I think) are also present. 

 No butterfly-remains are visible, but I have placed a small 

 sample of debris from each stomach under the microscope, 

 and in one (Odendaal's No. 309) have found the scales and 

 socketed membrane- scraps of some Lepidopteron — whether 

 moth or butterfly, I cannot say. 



Expt. .501. Jan. 10. — It became fairly clear in the course 

 of the following experiment that A, with long bill and tail 

 and the usual adult green gloss on the head, was a parent 



