EARLY EXPERIMENTS OF PLATEAU AND OTHERS. 139 



had sipped for a quarter of a minute, the slide with honey was moved so 

 that she flew to another slide. This was next taken away, and so on, until 

 she was induced to visit all the drops before returning to the hive. While 

 she was absent, all the upper glasses with honey were transposed and the 

 colored slips also moved, so that position could not influence selection 

 by the bee. The record was made by noting the order in which the bee 

 went to the different colors. The experiment was repeated a hundred 

 times with the use of two different hives at different places and the obser- 

 vations extended over some time in order to work with different bees and 

 under varied conditions. The first day's results gave a decided preference 

 for blue; white, yellow, and green were nearly equal, followed by orange 

 and red also about equal, with the plain glass last. In the next series, 

 bees had been trained for three weeks to come to a particular spot on the 

 lawn by placing honey on a piece of plain glass. In spite of the advantage 

 thus given the latter, blue was again first, followed at some distance by 

 white, and this by yellow, red, green, orange, and plain glass. 



Bonnier's results to the contrary were regarded as inconclusive, since 

 the colors were largely covered up by the bees and since the presence of 

 so many would attract their companions. Moreover, he omitted blue and 

 his squares were all colored, the absence of colorless checks being especially 

 serious. 



Response to detached petals.— Mueller (1883:273) regarded Lubbock's 

 results as inconclusive, since he did not use the natural colors of the flower 

 and did not give the bee a distinct choice between two plates differing in 

 color alone. To obviate these difficulties Mueller made use of detached 

 petals placed between glass slides which were cemented at the edges, 

 so that the odor could not be effective. A drop of honey was placed on 

 each test-object thus prepared. As a rule, petals of two distinct colors 

 were submitted to the choice of bees that had been marked and accustomed 

 to coming for honey on glass slides. In their color preference different 

 honey-bees exhibited marked individual response, one showing equal 

 liking for both colors, another a preference for one, and a third for the other. 

 Thus, while the total number of visits made by a group of 6 bees to the purple 

 of a rose and the blue of a cornflower was the same for each, only three 

 bees were equally sympathetic to them, the other three visiting them 

 in the following ratios: 10:14, 5:3, and 3:1. In the case of the fire-red 

 of nasturtium and the violet of the pansy, seven of eight bees preferred the 

 latter, most of them very decisively, and but one the former, also very 

 distinctly, viz, 10:3. Individual bees also differed in the constancy of 

 their response to related colors, two preferring honey-yellow to brilliant 

 yellow throughout the period of experiment, and the other two choosing 

 the one in the first half and the other in the second. Bright colors, bril- 

 liant yellow and orange, fire-red, scarlet, are less pleasing to the honey- 

 bee than the softer colors with which the bee-flowers are adorned, viz, 

 white, honey-yellow, rose, pink-red, violet, and blue. The preference 

 for bright yellow and honey-yellow was 48:78, for the former and white 

 21:48, for brilliant orange and rose 22:71, and 26:77, for fire-red and 

 violet 29:80, for scarlet and rose 9:55, for scarlet and pink 34:69, and for 

 scarlet and blue 13:78. 



