496 R. T. YOUNG 



Now if birds have such keen sight as is frequently attributed 

 to them, keen enough to discern protectively colored animals 

 on their backgrounds, it seems unlikely that they should be 

 unable to distinguish a man in crouching posture until within 

 gunshot of him, or to distinguish between real and decoy birds. 



In the case of insectivorous birds those which swoop down 

 upon their prey from a distance, like the Tyrannidae, the Hiru- 

 dinidae, etc., prey almost exclusively on moving insects, in 

 which case color (apart from mimicry and warning color) can 

 obviously have little or no protective value; while others, like 

 the Icteridae, Sturnellidae, etc., approach their prey so closely 

 before seizing it, that even if it were motionless, it would need 

 to bear an exceedingly close resemblance to its background in 

 order to escape their scrutiny. 



Many of the experiments described in this paper, especially 

 those of Series I, II, VII, and IX, are inconclusive for various 

 reasons, some of which have been already stated in connection 

 with the individual experiments. The chief reason however is, 

 as already stated by me in a preliminary communication be- 

 fore the American Society of Zoologists (Young, '15), that in 

 many of the experiments the approach of the birds to the back- 

 grounds was deliberate and gave them ample time to inspect 

 the latter carefully before the prey was seized. In this con- 

 nection, see Experiments 6, Series V, 28, VII, 1, IX, etc. Under 

 these circumstances it is probable that in many experiments 

 chance determined the bird's choice, that combination being 

 chosen, which was nearest to the point of the bird's approach. 

 In this connection an analysis of the results of Series I, III, VII, 

 and IX is of interest. My notes unfortunately do not indicate 

 in every case the point of the bird's approach. The number 

 of experiments in which this is indicated however is sufficient 

 to enable a fairly definite conclusion to be drawn from them. 

 Of the forty-eight experiments, the combination presenting 

 the greater contrast was chosen in twenty-six instances, and the 

 one showing less contrast in twenty-two. The former may be 

 called positive and the latter negative experiments. In the 



