NATURAL SELECTION 245 



results. Heikertinger (1929, 1929a) has recently considered the 

 case of the Hymenoptera, many of which are protected by 

 stings, a device whose protective value can be assumed with 

 greater safety ; yet in this group Heikertinger has endeavoured 

 to show that the stinging forms are more, rather than less, 

 attacked than other groups ; his evidence is considered in a 

 later paragraph (p. 255). 



Fisher (1930, pp. 158-62) appears to be one of the few 

 authors who have considered the difficulties involved in the 

 development of an unpalatable character, of the sort requiring 

 for its demonstration the actual tasting of the animal. It 

 would be expected that though the unpleasant taste would 

 disgust the eater, yet the victim could not survive, and no 

 selection in the direction of increase of unpalatability could 

 result. It has been maintained that some of the most con- 

 spicuous and probably unpalatable butterflies have an integu- 

 ment so hard or so flexible and leathery (Swynnerton, 1926, 

 p. 504 ; Eltringham, 1910, p. 109) that the insect can survive 

 experimental tasting, so that selection in the required direction 

 may well occur. Some of the Cantharid (Telephorid) beetles 

 which have conspicuous colours and appear to be distasteful 

 to birds have also an extremely flexible integument : in trying 

 to box these beetles in a tin they may be clipped between the 

 lid and the bottom to a degree which would cut any other 

 beetle in half, but which in this case only flattens the flexible 

 abdomen. But these are extreme cases which are not very 

 helpful in explaining the early stages of the development of 

 such a character. The difficulties are typical of those encoun- 

 tered by any explanation of the evolution of complex struc- 

 tures (see p. 306). In any palatable insect with a normal 

 integument, changes in palatability or in hardness or flexi- 

 bility occurring alone would appear to be of little survival 

 value, and we have no reason to assume that the appropriate 

 variations would occur simultaneously. We are faced with 

 the usual dilemma that if certain characteristics could develop 

 to a certain point ' on their own ' (e.g. if a certain degree of 

 either unpalatability or flexibility were developed), then selec- 

 tion could evolve the necessary complementary features ; but 

 that ' development on their own ' requires an evolutionary 

 process independent of selection. 



Fisher (I.e.) has advanced the alternative hypothesis that 



