RECENT INVESTIGATIONS. 207 



tween the color sense of the honey-bee and that of the hawk-moth, as is 

 equally true for the latter and Bombylius. 



Since attraction from a distance is a matter of vision, the hawk-moth 

 will regularly visit only those flowers that belong to its particular color 

 groups. This is supported by the records of visits, which show that the 

 flowers commonly sought are blue or purple in color, though in southern 

 Europe the yellow flowers of Linaria vulgaris may be exclusively visited. 

 This difference in behavior in various regions may be regarded as a passing 

 stage of the natural association, and therefore as a consequence of the insect's 

 experience. To effect pollination it is desirable that the hawk-moth visit 

 the flowers of one species for as long a period as possible. Therefore, all 

 the features of the flower are of significance that produce and fix the associ- 

 ation of the moth with this particular species. Of paramount importance 

 in this are those optical features which guide the moth to adequate supplies 

 of nectar. In connection with such optical associations the nectar guides 

 may play an important role. 



The above results confirm in large measure the views of the earlier 

 flower biologists as to the significance of the optical features of flowers for 

 the visits of the hawk-moth. The importance of nectar guides was some- 

 what overemphasized by them, though the essential facts of the older con- 

 cept can be maintained. However, no support has been found for the 

 view that the odor of flowers plays the important part in the attraction 

 of the hawk-moth heretofore assigned to it. 



Response of bees to spectral bands. — The studies of Kuhn and Pohl 

 (1921:738) mark a distinct advance in the technique employed for the 

 analysis of color vision in insects. They have utilized the spectral bands 

 obtained by means of the mercury spectrum in order to extend their experi- 

 ments into the region of the ultra-violet and have consequently had the 

 advantage of working with pure colors. The bees were habituated to fly 

 in a room in which a Hg spectrum was thrown on a horizontal table-top, 

 upon which it could be turned and shifted at will. To human eyes the 

 lines at 578 mm (yellow), 546 (jl/jl (green), and 436 nn (blue) appeared brilliant; 

 405 ^n (violet) was distinctly and 492 nn (blue-green) still noticeably to be 

 seen in the daylight of the room. The line 365 nn (ultra-violet) could be 

 recognized in weaker light as a sensible bluish glow on the white paper of 

 the table. In training the bees all the lines were eliminated except one, 

 which was directed upon a small elongated combustion dish filled with 

 sugar solution. In doing this the position of the line was frequently shifted 

 to avoid any association with a particular portion of the table. In the 

 tests without food a new sheet of paper was spread and the particular wave- 

 length presented in proper sequence with the remaining lines of the spectrum. 

 After a training period of an hour the bees always gave a constant response. 



After training to 578 ixn (yellow) the bees concentrated regularly upon 

 this band. When this was covered, they went in almost equal number to 

 546 uti (green). Lines of shorter wave-length were not noticed, and varia- 

 tions in the brightness within wide limits had no effect. After feeding upon 

 436 ij-h (blue) a dense mass of bees formed on this; violet (405 nn) was 

 equally strong in effect after uncovering from 436 n/x. After eliminating 

 436 nn and 405 nn, there was a marked movement to the ultra-violet line 



