138 Mr. G. A. K. Marshall on Diaposematism, with reference 



The value of experiment as evidence of palatdbility. 



The divergence of opinion which exists on the subject of 

 palatability between the more extreme and more moderate 

 supporters of Mliller's theory practically resolves itself 

 into a question of the relative value and reliability of the 

 various lines of evidence available. The following are the 

 principal sources from which information may be derived, 

 given in the order of their importance. 



(1) Exact observation as to the insects actually eaten and 



avoided by wild insectivorous animals. 



(2) Careful experiments upon wild or captive animals. 



(3) Field observations on the flight and general habits of 



insects in relation to their environment. 



(4) The comparative study of colour patterns and 



structure. 



The first class of evidence, which is obviously most 

 essential, is unfortunately the most difficult to procure, 

 more especially in the case of butterflies ; for very many 

 birds cut off the wings of these insects before eating them, 

 and they cannot therefore be identified by an examination 

 of the stomach contents, as in the case of insects having 

 easily recognisable chitinous structures. Thus we must 

 rely to a very great extent upon the three remaining 

 sources of information. 



Now in almost all those cases where the existence of 

 unpalatability is in dispute it will be found that the 

 arguments in its favour are based exclusively upon the 

 fourth line of evidence. But I believe that most ento- 

 mologists will agree that deductions of this kind are 

 peculiarly liable to error unless they be supported by 

 evidence from some other line of investigation. For my 

 part, where I find that conclusions derived from (2) and 

 (3) point in one direction, and those derived from (4) in 

 another, I have little hesitation in accepting the former 

 and rejecting the latter. 



As against this, it will be contended that habits cannot 

 be taken as an absolute proof of edibility or otherwise. 

 For although all the butterflies that have nauseous 

 qualities announce their presence by a characteristic de- 

 meanour, yet we must not suppose that these qualities 

 have suddenly appeared as we now see them ; on the 



