140 PRINCIPLES AND CONCLUSIONS. 



The least attractive of all the colors of true bee-flowers is glaring yellow, 

 the ratio between this and yellow-white being 35:68, with pink-red 27:74, 

 with purple 42:62, with indigo-blue 28:56, and with violet 24:78, while 

 chrome-yellow and cobalt-blue gave 11:40. White and yellowish- white 

 were visited about as readily as many shades of purple or even more so, 

 but less readily than blue or violet; for example, white against dark- 

 purple gave a ratio of 7:30, and against sky-blue 36:64, while yellowish- 

 white and purple gave 52:49, with blue 18:25, and with violet 11:24. 

 Blue is either preferred to red, or the two are equally attractive, in accord- 

 ance with the particular shades employed; thus, violet-blue in competition 

 with dull-purple gave 61:31, sky-blue and bright purple 45:36, sky-blue 

 and rose 57:54, cornflower blue and purple 52:52, and impure violet-blue 

 and dark-purple 57:57. A pure deep blue excels violet in attraction in 

 the proportions, 50:35 and 81:67. In its attraction for the honey-bee 

 violet excels all other flower colors except blue. Among the brilliant flower 

 colors, bright yellow is the most attractive to bees, giving with scarlet 

 the ratio 50:29, and with bright orange 42:31. The green of leaves is less 

 pleasing to honey-bees than the colors of bee-flowers, giving in competition 

 with rose 75:33, but it is more attractive than scarlet, 45:40 or orange 

 57:46. As a result of field observations, Kranichfeld (1915:40) has come 

 to the conclusion that the color preference demonstrated by Mueller's 

 experiments does not obtain in the normal visits of bees to flowers (see 

 p. 133). 



Color preference of nocturnal moths. — Gratacap (1883:791) made 

 cylinders of variously colored tissue-papers and drew them over common 

 kerosene lamps with gas chimneys in order to test the color perceptions 

 of night-flying insects. The colored lights were first placed in a row at long 

 distances from each other, but this arrangement seemed defective by reason 

 of the fact that the brilliancy of the light first visited interfered with the 

 visitor's freedom of choice between that color and another that reached 

 it but dimly. The lights were next arranged in a square, at first in such 

 manner that the circles of light touched each other, but later so that they 

 overlapped. Since this did not permit an exactly equal choice, the final 

 plan was to make use of but two lamps at a time. The need of causing the 

 insect to choose instantly between the colors before it reached either arose 

 from the infatuation produced by the lights, preventing the insect from 

 freeing itself except in an accidental manner. The results indicated the 

 absence of marked preference for certain colors over others, and demon- 

 strated the almost invariably greater charm of the white lantern, which 

 on account of its translucency appeared more brilliant than the colored ones. 



Response to color without antennae. — Forel (1886:24) considered 

 that the ingenious and patient studies of Lubbock had demonstrated the 

 ability of bees and wasps to distinguish colors and at the same time a feeble 

 sense of smell in the former. While this seemed to be contradicted by 

 the experiments of Plateau with artificial flowers, such negative results 

 were regarded as of less value, especially since other factors rendered the 

 investigation incomplete. Above all, it is possible and even probable 

 that an imitation exact to our eyes may affect those of an insect so that 



