1887.] VALUE OF COLOUR AND MARKINGS IN INSECTS. 2G1 



fallino; therefore under the influence of natural selection. Knowhig; 

 that increasing efficiency in protective measures is couuterbiilanced 

 by increasing keenness and cunning on the part of enemies, it is easy 

 to see how, as a response to an advance l)y the latter, a species might 

 take advantage of such an incidental quality to adopt an entirely 

 new line of defence. The concealment of the larvse we are considering 

 is evidently very successful, but if it were seen through far more 

 frequently than at present, and yet the larvae were always rejected 

 with disgust, there would be more and more opportunity and 

 necessity for the enemies to remember the experience ; and the 

 further the species varied away from the beaten path of protective 

 colouring, the greater aid would it afford to memory, which, although 

 that of another animal, is in this respect of far less importance for 

 the possessor than for the larva itself. I need hardly point out 

 that in speaking of an advance in the keenness of Vertebrate insect- 

 eaters, I mean an advance in the power of detecting all such larvae, 

 so that there would always remain a large proportion of palatable 

 species ; while the new line of defence would only be open to such 

 few of them as possess the quality of distastefulness in a marked 

 degree. I am quite aware that there is another possible explanation 

 of the unpleasant qualities in M. tijpica ; i. e. that they are the 

 remnant of a former defence by such means accompanied by 

 corresponding coloration, &c. ; but while this may exjjlain similar 

 facts in the case of certain other species, I do not think that it is 

 likely to hold in the instance of M. tijpica, for the protective habits 

 and appearance are correlated in so perfect a manner that we are 

 compelled to assume that a very long period of time must have been 

 covered in the attainment of so unusual and specialized a result. 



It now remains to consider the other exceptions which are of less 

 theoretical importance although of extreme interest. As the same 

 species have occurred before under other tables, it will be well to 

 shortly tabulate the results of all the instances among Lepidoptera 

 in which experiments have been made upon more than one stage 

 (see Table, p. 262). 



I much hope that future experiments will enable us to extend this 

 Table, but short as it is, it appears to point to several interesting 

 conclusions. In the first place there is no known instance of 

 distasteful qualities in stages later than the larva when the latter is 

 itself palatable. This statement will doubtless be true of the great 

 majority of species however complete be theexperimental investigation, 

 and it points to the conclusion that this method of defence arose 

 first in the larval stage. Such a relation is to be expected ; for the 

 species is exposed to more danger and is more helpless at this 

 period than at either of the subsequent stages. Tlie unpleasant 

 taste appeals to non-parasitic enemies which devour insects ; but the 

 almost complete limitation of the attacks of insect-parasites to the 

 larval stage must bear in an important way upon the other modes of 

 protection in this stage, tending to produce tiiat extraordinary 

 S|)ecialization in defensive methods which are well known to occur. 

 The imago can escape by flight, and the pupa, if exposed, may render 



