182 PRINCIPLES AND CONCLUSIONS. 



can be determined by their direct and rapid flight to them, the flight by the 

 lower types is quite different. The latter change in direction and each 

 time toward the side from which the fragrance comes. Such insects also 

 perceive colors, but only when they are near at hand. Similar differences 

 also occur among Diptera; Eristalis responds differently to a color than a 

 gnat does, and Bombylius and Volucella, two highly developed flies, show 

 but very slight response to odor, as Forel has already shown. 



The conclusion that color attracts the biologically higher insects from a 

 distance and odor when near at hand is supported not only by the experi- 

 ments recorded, but also by the following facts. A brightly colored object 

 does not change its position, and, if forced out of it, becomes more or 

 less conspicuous only as a result of changes in lighting, while odor behaves 

 very differently. In the first place, its intensity varies from time to time, 

 in the second its conveyance depends upon a variable wind direction, and in 

 the third its perception is determined by the amount of humidity in the 

 air, since the antennae are much more sensitive in moist than in dry air. 

 There are constantly many odors in the atmosphere which pervade and 

 replace each other, and are constantly changing their position in response 

 to the slightest breeze. Thus, the view of Sprengel and of Darwin that 

 color attracts insects from a distance is rehabilitated. 



Andreae's criticisms of Plateau's work. — In connection with Plateau's 

 statement that we have no means of knowing that the perception of color 

 by insects is the same as with man, Andreae pointed out that this was 

 equally true of odor, declaring that the whole question was an idle one and 

 that the only thing of importance was to determine whether what we call 

 color exerts an attractive effect upon insects or not. He also emphasized the 

 fact that many of Plateau's experiments were excellent in demonstrating 

 choice among colors, but they had no value for the question of the relative 

 merits of color and odor. The chief error in his investigations lay in the 

 fact that he failed at the outset to organize his queries properly. In the 

 first place all possible factors that bring about attraction must be considered, 

 and in the second as many of these as possible must be tested at the 

 same time. In this way alone can comparative results be obtained that 

 will lead to an objective conclusion as to the importance of form, luster, 

 brightness, and color, honey-odor, or fragrance. Many of Plateau's results 

 were to be ascribed to the effect of habit, and the others were not organized to 

 }deld a definite decision as to color and odor. However, Andreae's state- 

 ment that the greater proportion of visits to the decorollate poppies did 

 not change the essential facts with respect to the attraction of the corolla 

 does not seem valid. Finally, it was stated that Plateau had not sufficiently 

 distinguished the behavior of various insects in regard to color and odor, 

 since the lower apids gave results directly opposite to those obtained with 

 the higher ones. 



Plateau's criticisms of Andreae's work. — Plateau stated (1906:13) 

 that it was probable that the artificial flowers of Andreae and of Reeker 

 would have exhibited the same sources of error as those employed by Wery 

 (p. 185). With respect to the experiments with Crocus in which the flowers 

 were covered, it was objected that the bees upon returning to the habitual 



