GENERAL RESUME. 239 



attractive (24:46). Wery (1904) regularly obtained more visits to the normal 

 flowers ; the difference was often small with a heterogeneous group of visitors, 

 such as 19 : 13, 27 : 24, 43 : 27, but when the honey-bees alone were counted the 

 ratios were 1 1 : 6, 29 : 10, and 20 : 6. The totals for one series was 138 normal 

 and 46 mutilated for all visitors, and 72 to 28 for the bees. Detto (1905) 

 observed that the removal of the corolla stopped visits completely, and 

 that replacing the petals caused them to be resumed. Lovell (1909) secured 

 decisive results as a rule by the removal of the corolla, as shown by the 

 following ratios for normal and mutilated, 15:0, 12:0, 7:0, and 12:1. 

 In Allard's experiments (1911) the loss of the corolla caused the inspections 

 to drop from 81 to 4, while a single petal received 16 inspections to 26 for 

 the normal. Placing a single petal on a decorollate flower yielded 8 inspec- 

 tions to 9 for such a petal pinned to a stem and 27 for the normal; replacing 

 the petals regularly brought the inspections back to the usual number. 



In the present studies, the removal of the corolla has rarely eliminated 

 visitors entirely, but it has usually decreased them from three to ten times, 

 depending upon the flower and the habit of the visitors. Reducing the 

 size of the corolla was somewhat less effective as a rule, the decrease ranging 

 from a half to a tenth. With both lips removed in Monarda, visits were 

 reduced to a fortieth of the total for all mutilations and with the lower 

 lip gone, to a tenth. The most striking exception occurred in Chamae- 

 nerium, where the excision of all the flower parts but the ovary and nectar 

 ring resulted in nearly 5 times as many visitors and visits as to the normal. 

 This was apparently due to the marked effect of habit in responses to this 

 flower, as well as to the exposure of the nectar. This probably also explains 

 the discrepancy involved in Detto's results, in which the reduction of the 

 corolla was without effect. Experiments with newly hatched insects will 

 be required to evaluate the effect of habit, but practically all the existing 

 experimental results are in accord in demonstrating that the loss or re- 

 duction of the corolla has a significant and usually a decisive effect in de- 

 creasing visitors and visits. This applies with greatest force to the bees, 

 and is somewhat modified by the type of flower and especially its fragrance. 



Artificial flowers. — Plateau devoted four papers to the attraction exerted 

 by artificial flowers, in which he insisted upon the perfection of his imitations 

 and ascribed the abundant visits obtained by other investigators to the 

 presence of materials attractive by their odor or to imperfections in the 

 method. Neither the presence of attractive substances nor the effect of 

 habit can explain the frequent initial visits obtained by the great majority 

 of workers, and the proper explanation seems to be that of the variation 

 to be expected in consequence of differences in time, place or installation. 

 In his preliminary studies (1877), Plateau obtained few visits to artificial 

 flowers and these were made chiefly by butterflies, but visits and inspections 

 were much more numerous in the experiments of 1897, though as a rule 

 far below those to normal flowers. Bombus gave 4 inspections out of 10 

 to imitations, Anthidium 11 out of 38, and Megachile 10 out of 17; in some 

 cases of those reported in the series of 1906, artificial flowers received as many 

 as 8 visits and 7 inspections to 15 visits to normal ones. Forel agreed with 

 Plateau's conclusions as to the ineffectiveness of artificial flowers, but 

 his results likewise were not at all in harmony with each other. Bees dis- 



