MAIN RESEARCHES OF PLATEAU. 153 



(2) Insects cease their visits when, without disturbing the bright-colored parts, one 



removes the nectar-bearing part of the flower, and they begin them again if 

 the missing nectar is replaced by honey. 



(3) It is sufficient to put honey in or upon inconspicuous green or brownish anemoph- 



ilous flowers rarely visited, to attract numerous insects. 



(4) The r61e of the sense of smell is also demonstrated by leaf flowers filled with honey. 



Role of vexillary organs. — The attractive value of colored bracts 

 and show-flowers furnished the theme of the first investigation in a new 

 series of studies on the relations between insects and flowers, begun by 

 Plateau in 1897 (1898:339). The vexillary role of the terminal group of 

 vivid bracts of Salvia horminum was first studied, making use of a bed of this 

 species more than a square meter in extent. This resembled a rose carpet 

 as seen from above, while the flowers themselves were hardly visible, except 

 at the edge. The numerous bees present, chiefly Apis and Anthidium, 

 went directly to the flowers upon their arrival, noticing the colored bracts 

 only as they flew upward from flower to flower or when the folded bracts 

 simulated the flowers. The butterflies were more easily deceived, 1 

 Pieris and 3 Rhodocera landing on the bracts and trying to probe them with 

 the proboscis. Six honey-bees were observed to make but one visit and one 

 inspection of the bracts to 388 visits to the flowers. In order to eliminate 

 the possibility that this behavior was due to habit arising from the fact 

 that Salvia had been grown for several years in the same spot, plants were 

 transferred to another garden more than 2 km. distant, where they were 

 placed in a circle within another of Dianthus barbatus. Thus, the visiting 

 insects were offered three choices, between the flowers of Dianthus, those of 

 Salvia, and the bracts of the latter, the last two being presumably new to 

 them. The insects that hovered without landing on either Dianthus or 

 Salvia were 106 in number, all butterflies or moths, with the exception of 3 

 individuals of Bombus. During the period, 26 individuals of bees, 49 

 butterflies, 48 flies, and 16 beetles visited Dianthus alone, without paying 

 any attention to Salvia. Of 21 individuals that visited Dianthus and were 

 also confused by Salvia, all were Lepidoptera, with the exception of one 

 Andrena. Three out of 6 individuals of Macroglossa hovered over the 

 colored bracts of Salvia, one making this mistake 3 times. In the case of 

 the insects visiting Salvia alone, the bees observed numbered 250, but only 

 24 of these made even a slight mistake, while there was on an average one 

 error for each butterfly and each fly. 



With reference to the attraction exerted by show-flowers, Plateau em- 

 ployed two species of Hydrangea, the one native and with a clear distinction 

 between the small fertile flowers of the center of the umbel and the large 

 sterile ones of the margin, the other cultivated and exhibiting umbels with 

 large sterile flowers alone. The number of visitors to the native species 

 was small, a few bees and syrphids coming for pollen alone. The behavior 

 of these was characteristic, for, instead of alighting first on the large mar- 

 ginal flowers as demanded by the theory of their attractive rdle, they flew 

 by or above them as though they did not exist, in order to land directly on 

 the fertile central ones. The number of mistakes made was but 1 for the 

 bees out of 79 visits, while for the less intelligent flies it was 1 for 18 umbels 

 visited. In the case of the cultivated species neither the brilliance of the 



