318 Mr. G. A. K. Marshall on 



from the behaviour of MantidiV in the wild state, and 

 such as Avo do possess entirely confirms the conclusions to 

 be drawn from Mr. Marshall's experiments. In the first 

 place we have the followint;' observation of his own, made 

 in the Karkloof, Natal, in February 1897 : — 



" Saw a Mantis catch a male horta on a flower in the 

 veldt. It began eating at tliebase of the abdomen, which 

 it consumed entirely, and then started on the thorax, of 

 which it only ate a very little, and then thi'ew it away." 



This observation corresponds almost precisely with 

 many made upon the captive insects. Mr. Roland Trimen 

 also says that he never found the wings of Danais or Aci\va 

 among the fragments of butterflies wliich sprinkle the 

 ground below the feeding-place of a large Mantis, although 

 he is careful to add that he could not be sure that these 

 butterflies visited the exudations of Acacia sap, round 

 which the predaceous insects secure a plentiful supply of 

 food (Linn. Soc. Trans., vol. xxvi, 1870, p. 500). It has 

 already been pointed out that Colonel Yerbury's and 

 Colonel Bingham's observations upon Mantid;v in the wild 

 state are entirely confirmatory of Mr. Marshall's observa- 

 tions of them in captivity, as regards the food which 

 appears to be freely provided by certain Pierine genera 

 refused or disliked b}^ other insect-eating animals. 



Another question of deep interest raised by Mr. 

 Marshall's experiments on Maniidrc. is the inquiiy how 

 far the species which they reject or eat only sparingly is 

 unwholesome or even poisonous to them. There is 

 strong a 'priori probability for the view that the 

 preferential appetite of such a form as a Mantis is merely 

 the strong instinctive tendency to eat the food which best 

 suits its organization and reject that which suits it least. 

 We should expect therefore that such marked disinclina- 

 tion to eat Aeneas as we observe in Mantidn} indicates, 

 not distaste or unpalatability in an anthropomorphic sense, 

 but merely that Acrseas are unwholesome to Mant.idie. 

 The evidence requires to be sifted in detail. 



In Experiment III. the signs of weakness seem to be a 

 too-excessive result of the single Acr/ea, and portion of 

 another, which were eaten. At the same time generic and 

 specific differences are almost certainly of great import- 

 ance, and it must be remembered that III., IV., and VII. 

 belonged to probably the same species, and all exhibited 

 weakness after an Ao^^a diet, resulting in the death of 



