1887.] VALUE OF COLOUR AND MARKINGS IN INSECTS. 217 



exceedingly imperfect, while very perfect protection is gained in 

 other ways. In fact I have suggested above that the species is 

 not now "mimetic" at all, but retains two points of an ancestral 

 condition in which it actually derived shelter from the reputation 

 possessed by many Hymenoptera, these being (1) a structural 

 point — its transparent wings, and (2) a habit — its diurnal flight. 

 If this view be correct tlie species should be moved into the table 

 given below of the results of experiments upon insects protected 

 by evading their enemies, and there it would add to the instances 

 which support Wallace's converse suggestion, that such insects 

 will be found to be eaten just as the class we are discussing 

 are generally refused. I have placed the species in the fourth 

 table in deference to Wallace's opinion on the kindred species 

 S. hombyliformis, which it closely resembles ; but I believe that 

 the place of both species should be below. I made several 

 attempts to obtain the imagos of Sphecia apiformis, in order to offer 

 these to my Lizards ; for this species certainly does suggest a 

 stinging Hymenopterous insect in the most remarkable way ; but 

 the attempt was unsuccessful. >S'. bembeciformis is equally well 

 suited for experiment, but even more difficult to obtain ; I hope, 

 however, to be able to experiment with one or both species in the 

 course of next summer. 



It is noteworthy that (excluding S.fuciformis) there is probably 

 no instance in this list which proves a difficulty in the way of 

 Wallace's suggestion at all equal to that raised by the larva of 

 D. euphorbicE. For the pupa of A. grossulariata was not swallowed, 

 but only well bitten by a very hungry lizard, and although the 

 imagos of S. lubricipeda and P. auri/lua were eaten (also by very 

 hungry lizards in the former case), it cannot be asserted that they 

 were eaten with relish ; and, further, the experiment should be 

 repeated with Birds and other species of Lizards. In all other cases 

 the insects were refused by at least one of the animals to which they 

 were ofiFered. The comparison between the treatment received by 

 Bees (described below) and Wasps, together with a comparison 

 between their appearances, is strongly in favour of Wallace's sug- 

 gestion. It now remains to analyze the list in the same manner as 

 was adopted before, showing the transition of conduct observed (see 

 Table, page 224). 



Comparing the three analyses of all experiments ever conducted 

 upon presumably nauseous or dangerous insects, omitting S.fuci- 

 formis and including C elpenor and C. vinula, we find that out of a 

 total of 44 cases which have been tested 13 were entirely disregarded ; 

 but this number includes P. machaon, which I believe ought to be 

 placed below under forms chiefly defended by protective resemblances, 

 and also S. menthastri (larvae, which were tasted in the youngest 

 stages). Furthermore four species (including the two just men- 

 tioned) belong to Table III., containing larvae which can hardly 

 be called conspicuous. Hence, omitting this table, the number is 

 reduced to 9 cases out of a total of 37 ; and of the 9, not a single 

 species has been offered to more than one out of the three groups of 



