GENERAL RESUME. 241 



primarily, and perhaps slightly as to surface texture and odor. Their 

 greater naturalness is attested by the fact that they received about five 

 times as many visitors as the paper composites employed. With respect 

 to total response they were half as attractive as normal blossoms, though 

 in the case of particular species they were equally or even more attractive. 

 Thus, painted flowers of Rubus deliciosus were visited as readily as normal 

 ones, and in the installations of Aconitum and Chamaenerium they received 

 a larger number of visits absolutely. The greatest number of visits to such 

 flowers was made by Bombus juxtus, but in two experiments the honey-bee 

 made twice as many visits to them. The behavior of the latter was espe- 

 cially significant, since it ignored paper flowers altogether, but visited about 

 half as many painted as normal ones. On the whole, it seems clear that 

 changing the color of flowers as well as giving a wider range in color dis- 

 turbed the response of bees, but to a much smaller degree than artificial 

 flowers proper. The results secured by painting the white flowers of Rubus 

 suggest that making these more conspicuous overcomes the disturbance 

 arising from the addition of the water-colors, and indicate that white flowers 

 should be preferred for experiments of this kind. 



Inclosing flowers in glass. — Glass containers, such as beakers, tubes, and 

 globes, have been employed by several investigators to eliminate odor and 

 determine the response to color alone. Andreae placed flowers under bell- 

 glasses or beneath inverted beakers and found that they received many visits, 

 in spite of the absence of odor. Wery also found out that cut flowers in a closed 

 globe exerted nearly normal attraction. The conical disk of three heads 

 of Rudbeckia was covered with a glass tube by Detto, with the consequence 

 that numerous bees flew against the glass. Allard made use of boxes with 

 a glass face and found that the inspections of the cotton blossoms were 

 reduced about half. These experiments show conclusively that color by 

 itself serves to attract insects in quantity, and to a degree equaling or 

 exceeding that of odor. This is confirmed by the experiments in masking 

 and covering discussed in the next section, the effect being to eliminate 

 color and to render odor the sole attractive force. 



Green flowers and showy nectarless flowers. — Plateau contended 

 that green or dull-colored flowers which are visited by insects demonstrate 

 that color is unnecessary and hence ineffective in attraction. Andreae 

 showed that the color of inconspicuous flowers was often more attractive 

 than either odor or honey, and Lovell has proved that green leaves with 

 honey are much less attractive than yellow immortelles. When a head of 

 golden glow and a green spike of Amarantus were placed in competition, 

 the former obtained all told 51 visits to 16 for the latter. Plateau likewise 

 insisted that the failure of showy nectarless flowers to attract insects proved 

 that color was unimportant, and that it was necessary to supply them with 

 a sweet fragrant liquid to induce bees to come. Perez observed that bees 

 did go to showy flowers without nectar, but that they soon learned the futil- 

 ity of this and ignored such blossoms. They came readily to geranium 

 flowers supplied with honey, and not only returned to them after the honey 

 was exhausted, but also went to others that had received no honey. Giltay 

 found that bees trained to come to geranium flowers with honey went read- 



