238 PRINCIPLES AND CONCLUSIONS. 



GENERAL RESUME. 



TREATMENT. 



Plateau focussed attention so exclusively upon attraction that it has 

 remained the chief theme of experimental pollination for 45 years. While the 

 present series of studies has endeavored to deal with the whole field in a 

 more comprehensive fashion, attraction is necessarily the central theme of 

 a resume that attempts to summarize all the experimental results in brief. 

 This is considered with respect to color, odor, form and size, distant and near 

 attraction, and competition. The first is discussed under the captions of 

 mutilation, artificial flowers, painted flowers, inclosing in glass, green and 

 showy nectarless flowers, color preference, and the second under those 

 of masking and concealing flowers, honey, perfumes and scents, and relative 

 values. The remaining major divisions are learning and habit, memory 

 and intelligence. For the treatment of these there is much less definite 

 and accurate experimental material, and an adequate exposition of them 

 must await extensive experimentation under the quantitative conditions 

 afforded by control and measurement. 



ATTRACTION. 



COLOR. 



Mutilation. — Here it is necessary to take into account only those ex- 

 periments in which the corolla, perianth, ray-flowers, or bracts were removed 

 or reduced in expanse by shortening or otherwise. Covering the attractive 

 part produced essentially the same effect, but such cases are considered 

 under masking or artificial flowers in accordance with the device employed. 

 In Plateau's first experiments with decorollate flowers (1896), his results 

 in general failed to confirm his conclusion as to the insignificant role of the 

 corolla in attraction. In the case of several species of flowers and insects, 

 the mutilated flowers were visited less than the normal, for example, 

 24:44, 4: 10, 250:330, and 16 inspections to 29, and in others no comparison 

 was possible because the visits to normal flowers were not noted. More- 

 over, no account was taken of habit and the effect of removing the corolla 

 on the diffusion of the odor, especially of the nectar. In the studies of 

 decorollate poppies (1902 3 ), he obtained an average of 4.5 visits to the mu- 

 tilated flowers to 3.4 for the normal. These were wholly exceptional and 

 are negatived by the results of all other workers. They serve to illustrate 

 the variations in behavior that may be expected, but in this case they are 

 doubtless to be explained on the basis of habit and accessibility. 



In working with decorollate poppies and geraniums (1904), Giltay 

 regularly obtained such ratios between normal and mutilated flowers as 

 96:9, and 38:1, while in the later studies of 1906, the intact flowers were 

 always much more visited likewise. On the second day of an installation, 

 the ability of the bees to learn changed such ratios as 15:0 and 12:0 to 

 18:13, but they were rarely if ever reversed. Similarly, Schroder (1901) 

 secured ratios of 37:0 and 21:4 with Syritta pipiens on the first day, but by 

 the second day experience increased the visits to the mutilated to almost 

 half, viz, 19 to 46. Andreae (1903) found decorollate flowers to be several 

 times less attractive than normal ones (3 : 24), and Errera (1904), about half as 



