Tlie Bionomics of South African Insects. 325 



giving one or two only edible butterflies, and confining 

 the remainder to a single species of Acr/ea or Amauris 

 each. But at present I find it not only difficult to get 

 hold of a Mantis, but it is even quite a job to catch sufficient 

 Acrseas to continue the experiments. I have not seen a 

 single specimen of A. pctrxa for over six weeks, though 

 normally it should be swarming at this time of year. 



[Mr. Marshall subsequently carried out a part of the 

 programme which he liere suggests. See Experiments 

 IX., X., XI. on MantidR'?\ 



When the experiments on spiders are compared with 

 those on Mantiscs the conclusion is suggested tliat Aciw- 

 inie were distasteful to both, and only eaten under the 

 stress of hunger, while Danainm were far less distasteful 

 to the Mantises than to the spiders. To the latter they 

 appeared to be at least as distasteful as the AcrivinK. 

 Such differences in the susceptibilities of insect-eaters lielp 

 us to understand the puzzling case of Tcrias, and the 

 Ethiopian Lyca^nid genera which appear undoubtedly to 

 mimic it, and permit us still to look on Fapilio clcmodocus as 

 a possible model. We see that the various insectivorous 

 groups have different tastes, and within each group we 

 must expect to find individual species adapted to feed 

 largely on insects which are as a rule rejected by the 

 other members of the same group. 



In one respect spiders are extremely satisfactory for the 

 purpose of these experiments. They remain throughout 

 wild animals with their natural sources of food still avail- 

 able. The same may be the case with Mantides, as in 

 the Gongylas watched day after day by Col. Yerbury at 

 Trinkomali (see p. olG). 



The late Thomas Belt (" Naturalist in Nicaragua," Lon- 

 don, 1888, p. 317) states that a " spider that frequented 

 flowers seemed to be fond of" the HcJiconidcii (including 

 Itlwmiinc'B), althougli a large species of A^ejjhila used to 

 drop them out of its web when he put them into it. 



Dr. A. G. Butler (Trans. Ent. Soc. Lond., 18G9, p. 27) 

 long ago showed that the larvas of Abraxas grossularictta and 

 Halia %vavaria were not eaten by the spiders he employed 

 — Epcira cUadema (the name given in the paper is Erciha 

 dicalcma) and Lycosa species (?). In the former case they 

 were cut out of the web, and in the latter seized and 

 carried down the "dark silken funnel," but then relin- 

 quished apparently uninjured. Professor Plateau (Mem. 



