172 PRINCIPLES AND CONCLUSIONS. 



plants in the wire cage produced only a very small amount of seed, viz, 

 0.004 mg. per flower, although they had been artificially pollinated. The 

 215 decorollate flowers yielded 10.77 gm. of seed or 0.05 gm. per fruit, while 

 215 normal ones produced 25.23 gm. or 0.117 per fruit. The 28 decorollate 

 flowers that had been artificially fertilized gave 0.115 gm. of seed per fruit. 



The criticisms of Plateau as to the methods and conclusions of Giltay 

 have already been given (p. 159). These brought in return a critique of 

 Plateau's studies, in addition to a new series of experiments with mutilated 

 poppies, in which especial attention was devoted to the insect visitors 

 (1904:368). The first emphasized the discrepancy between Plateau's 

 earlier and later conclusions as to the importance of color, as Perez had 

 already done, and summarized the experiments of Forel with masked Dahlia 

 heads, which gave results very different from those obtained by Plateau (p. 

 145). Papaver rhoeas was again used by Giltay for his studies, owing to its 

 many advantages, such as complete sterility to its own pollen, the ease of 

 cultivation, the abundance of flowers and of insect visitors, and the readiness 

 with which the corolla may be removed. To insure better manipulation 

 the plants to be used were transferred to pots and kept in a screened cage 

 until needed. Checks were employed in connection with the removal of the 

 calyx and corolla to show that this in no wise affected the visits of insects 

 and that Plateau's fears were consequently groundless. In the first series 

 of 1902 two potted plants were placed on a greensward 2 meters apart, 

 and the corollas removed from the flowers of one of them. For an equal 

 number of flowers the normal ones received 96 visits and the mutilated 9 

 visits, while the second series with the plants near each other and with but 

 one or two flowers each gave 38 and 1 visits respectively. In the third 

 series the intact flowers obtained 34 visits, those with the calyx removed 

 and the corolla still folded 14, ordinary buds 1, and young fruits 2. When 

 flowers were so masked by pots that they could not be seen but any odor 

 present was free to escape, they received no visits, but bees came readily as 

 soon as the pot was removed. When the pot was so placed that the flower 

 was visible, it received frequent visits, thus showing that the odor of the 

 pot was not a factor in the preceding case. The relative importance of 

 color and odor was further demonstrated by placing petals in a dish and 

 covering them with a pot so that they were invisible but the odor could 

 escape. These were never visited, but when the pot was removed, bees 

 alighted on them, sometimes immediately. 



Several additional series of experiments were carried out in 1903, in which 

 the response to the two kinds of flowers was shown to be greatly influenced 

 by habit. At first when one or more pots of mutilated flowers were placed 

 among normal ones, the visitors gave the former practically no attention. 

 In the case of two flowers of each sort the intact received 15 visits, the muti- 

 lated, none, while on the next day a honey-bee was seen that went to muti- 

 lated flowers as often as to normal ones. Three days later an observation 

 on 4 each of intact and decorollate flowers yielded 12 visitors to the former 

 and none to the latter, and 1 flower of each on the same plant gave respec- 

 tively 8 and 4. However, 1 marked honey-bee went to 18 intact and 13 

 mutilated flowers. When all the flowers in the experimental plot were picked 

 and 2 pots with equal numbers of the two sorts of flowers were placed at 



