EXPERIMENTS ON PROTECTIVE COLORATION 497 



former the bird approached from the side of the chosen combina- 

 tion (showing greater contrast) in four cases (15 per cent), 

 from the opposite side in eighteen cases (69 per cent), and took 

 a middle course in four cases (15 per cent).'*-' In the negative 

 experiments on the other hand, the approach was from the same 

 side in 15 cases (68 per cent), from the opposite side in 5 cases 

 (23 per cent), while a middle course was taken in 2 cases (9 per 

 cent), which is nearly an exact reversal of the former -pro-portions. 

 This comparison shows perhaps even more clearly than do the 

 results of the more definite series (III, VI, VIII, and X) the pro- 

 tective effect of color in the animals used in these experiments; 

 for it shows that in spite of a deliberate approach of the birds 

 to the backgrounds, it was possible to so far deceive them that 

 in eighteen out of forty-eight cases they passed by the incon- 

 spicuous combination. On the other hand, in only five cases 

 out of the forty-eight, did a bird pass over a conspicuous combina- 

 tion in favor of an inconspicuous one.'^'^ Further in Series III, 

 the martins usually approached the backgrounds on foot, rather 

 than on the wing, giving them also time to inspect the back- 

 grounds more or less closely before feeding. Yet in spite of that 

 fact, the results of this series show clearly the protective value 

 of resemblance to the background in the case of motionless insects. 



It is further to be borne in mind that the resemblance between 

 the prey and the background in these experiments, was in many 

 cases not very striking. Had animals with more striking re- 

 semblance to their surroundings been available for experiment, 

 I am confident that even in experiments with crows, grackles, 

 chickens, and birds of similar feeding habits still more marked 

 results would have been obtained. 



There is another factor also which possibly modified the results 

 in a few cases. Namely, choice of the bird for one background 

 over another, regardless of the relative contrasts presented by 

 the combination of background and prey. This factor, I be- 



^1 Percentages given to nearest unit only. 



35 In this connection, see Experiments 3 and 5, Series I, 3, II, 4, 6, 8, 15, 18, 

 21, 22, 23, 28, and 30, YII, and 2, 3, 4, 7, and 9, IX. 



