MAIN RESEARCHES OF PLATEAU. 147 



(1) Insects visit actively heads in which form and color are masked by green leaves. 



(2) Neither the form nor bright colors of flower heads seem to possess attractive power. 



(3) The colored ray-flowers of simple dahlias, and in consequence those of other 



radiate composites, do not have the vexillary, or signal, role attributed 

 to them. 



(4) Since form and color appear to have no attractive role, insects are evidently guided 



to composite heads by another sense than sight. This is probably the 

 sense of smell. 



Removal of corolla. — Plateau removed the corolla or ray-flowers in 

 several species in order to determine the role in attraction (1896:505) 

 The first experiment was made with Lobelia erinus, with which Darwin had 

 already made a test of this response (1876:420). The latter cut off all the 

 petals of certain flowers and only the lower striated lip of others and found 

 that not a single honey-bee visited them, although bees were abundant on the 

 normal flowers. The removal of the small upper lip had no effect. Plateau 

 used two pots of this species, one with 30 and the other with 40 flowers. 

 These were placed near groups of Dahlia, Petunia, and Tagetes in order to 

 afford a wider choice to visitors. The corollas of one pot were cut back to the 

 tube, and the behavior recorded for three different periods of 1 to 2 hours. 

 The total number of visitors to the whole flowers was 33 against 25 to the 

 mutilated ones, Eristalis tenax giving much the largest number, while 

 there were 29 inspections to the former and 16 to the latter. Thus, while 

 there was a larger number of visitors to the normal flowers, it was pointed 

 out that the suppression of the corolla did not prevent them, as in Darwin's 

 results. In the case of Oenothera biennis, one honey-bee was seen to take 

 nectar from 14 mutilated flowers in succession, and other individuals made 

 respectively 10, 3, and 15 visits to them, the last one returning to flowers 

 already visited. When the corolla of Ipomoea purpurea was cut back to 

 the sepals, Bombus muscorum visited successively 5 mutilated flowers, and 

 B. terrestris in like manner 2, 3, and 5 respectively. 



Kurr (1833:135) removed the floral envelopes of Delphinium ajacis and 

 D. consolida and found that the flowers produced seeds, a fact that aston- 

 ished Darwin (1. c). Plateau made a similar test with D. ajacis, removing 

 all the petaloid parts except the nectary, but the only visit was made by a 

 bumble-bee. In the next test only part of the flowers of a spike were 

 mutilated; B. terrestris visited the intact flowers and two of the mutilated 

 ones. The mutilation experiment made by Bonnier (1878:61) with Digi- 

 talis purpurea was also repeated, 5 out of 19 clusters having the corolla 

 cut back to a tube 1 cm. long, and the plants being intermingled. Anthid- 

 ium and Bombus made 12 and 1 visits and 11 and 5 inspections respectively 

 of the mutilated flowers. With 3 mutilated spikes out of 5, 4 individuals of 

 B. terrestris merely inspected the cut flowers, while 2 visited 5 and 6 of them 

 respectively. The next day the same setting was visited by 4 bumble-bees 

 of the same species, the first going only to intact flowers, the second merely 

 inspecting the mutilated ones, the third visiting 6 of the latter in succession 

 after normal ones, and the last, 9 in similar manner. Antirrhinum majus 

 was mutilated by cutting the corolla back to 1 cm. in 8 spikes out of 25. 

 Of 5 bumble-bees observed, 3 merely whirled about the cut clusters, while 

 the other two flew near them and then left for the entire flowers. The 



