1912.] NATURAL SCIENCES OF PHILADELPHIA. 357 



against natural enemies are unavoidable: (1) The experiments 

 are very inconsistent : (2) They have been misinterpreted, and (3) 

 They are not trustworthy guides to behavior under natural con- 

 ditions. 



The Experiments are very Inconsistent. — Inconsistency in the details 

 of various series of experiments have been set forth in previous 

 pages (see pp. 298, 300, 313, 316 and 319). Inconsistency 

 in the results of entire series is plainly shown by the strongly 

 contradictory conclusions different experimenters have drawn. 

 Thus Weir, Poulton, Marshall, Pocock, and Finn, for instance, thought 

 their experiments supported the selectionist theories concerning 

 protective adaptations, while Butler, Manders, Punnett, Plateau, 

 Reighard, and Pritchett, among others, drew just the opposite 

 conclusion. Beddard's opinion was that distastefulness was not 

 more definitely associated with conspicuous colors, than with plain 

 ones. The characteristic inconsistency of experimental results are 

 described by him in the following language: 92 "None of these 

 experiments are thoroughly satisfactory; it is so difficult to interpret 

 them, and they are often contradictory, for a bird will eat one day 

 what it has refused before. The experiments that have been 

 made are like most other statistics — they may be made to prove 

 anything." 



The Experiments have been Misinterpreted. — This charge weighs 

 not so much against the experiments themselves as against their 

 makers, but it throws doubt upon the desirability of such tests, 

 since the personal equation is so large a factor in the interpretation 

 of results. 



Definite instances of misinterpretation have been cited in 

 previous pages (295, 303, 305-316, 325 and 328-330). A 

 chronic case is well illustrated by the following quotations from 

 Prof. E. B. Poulton (Trans. Ent. Soc. Lond., 1902) : 



" A mantis is probably less affected in this respect [food preferences] 

 by confinement than a vertebrate animal; but the same general 

 criticism will probably hold in both cases — that while the rejection 

 of an insect by a not over-fed insectivorous animal in captivity is 

 evidence of unpalatability or dislike, its acceptance is not sufficient 

 evidence of appreciation or that it constitutes an element of the 

 normal diet. An insect may be eaten readily in captivity which 

 would be rejected or only eaten under the stress of hunger in the 

 wild state" (p. 317). 



32 Animal Coloration, 1892, p. 166. 



