154 PRINCIPLES AND CONCLUSIONS. 



umbels nor their large number served to attract honey-bees, bumble-bees, 

 or butterflies, except for rare visits or inspections. The sole visitors were 

 Syrphus and Melanostoma, which went almost exclusively to the flowers 

 with stamens. The total number of errors for the bees was 5 in 397, for the 

 flies 5 in 93. The author dismissed the possibility that this was due to 

 habit, partly because of the short life of the individuals and partly because 

 of the way he assumed this would affect the numbers found on the umbels. 

 He regarded these results as proving conclusively that the vexillary organs 

 studied actually attracted the majority of insects so little and the higher 

 bees to such a minimum degree that the fertilization of these species would 

 in no wise suffer by their absence, and that they could no longer be consid- 

 ered as organs of attraction. 



Choice of colors by insects. — The first half of this second paper of the 

 series (1899:336) is devoted to a critical resume" of experiments by others 

 made by means of colored objects and by means of natural intact flowers, 

 together with an account of desirable and undesirable methods to be em- 

 ployed in such studies. The first experiments were made with the two 

 varieties of Salvia horminum, one of which had pale-rose flowers and bright- 

 rose bracts, while the other possessed violet-blue flowers and deep-blue 

 bracts. These grew in contiguous beds and were trimmed back to give the 

 same dimensions to the two groups. The observations were made during 

 an hour each day for 12 days, a representative record giving 125 visits to 

 rose and 131 to blue, with 5 abrupt changes from rose to blue and 7 from 

 blue to rose. The addition of the numbers series by series gave the predom- 

 inance alternately to rose and blue, the final totals being 1,085 rose and 

 847 blue, and this was regarded as explaining the preference obtained by 

 Lubbock for blue and Bennett for rose. The equality of the two colors 

 was further shown by the fact that Anthidium and Megachile flew suddenly 

 from rose to blue in 64 cases and from blue to rose in 75, while one individual 

 of the former visited first 5 blue, then 2 rose, 4 blue, 4 rose, 1 blue, and 4 rose. 



For a number of insects and flowers the percentages of each color and 

 of the visits were computed and the results expressed in table 90. 



In further experiments it was found that the honey-bee went to scarlet 

 without the least repugnance. Pieris brassicae eagerly visited the red 

 and yellow heads of Zinnia, while the related Goniopteryx seemed to neglect 

 these colors. One individual of Pieris visited indifferently the red, rose, and 

 yellow heads of Zinnia, but another went only to rose heads of Scabiosa. 



In his conclusions, Plateau emphasized the fact that he had at no time 

 said that insects did not see the colors of flowers. He affirmed, however, 

 that we have no practical means of assuring ourselves that they perceive 

 colors and that this perception is the same as with man. The results of all 

 experiments made with colored papers, cloths, or glass, or with isolated petals 

 or normal flowers can be explained either by differences in the amount of 

 light reflected, or by differences in the refrangibility of the rays transmitted 

 or reflected. Whatever may be the visual perception of insects, the 

 question whether the insects that visit flowers are guided in their choice by 

 the colors that flowers present to the human eye must be answered in the 

 negative. In studying the response of insects to varieties of the same 



