MAIN RESEARCHES OF PLATEAU. 151 



them without the slightest notice. Similar results were obtained with the 

 apple, except that the flowers with a drop of honey were visited by one 

 wasp and many flies. No visitors came to the artificial flowers of Myosotis 

 alpestris, even with the addition of honey, while with those of Saxifraga 

 umbrosa 2 individuals of Apis and 1 of Andrena gave them a hurried in- 

 spection. 



A more comprehensive study was made of Digitalis purpurea, in which 

 the habit of the inflorescence, the form of the corolla, its color, and markings 

 were copied from nature with great care. Among 6 natural spikes were dis- 

 posed 3 artificial ones, on one of which honey was placed in the flowers. On 

 its two visits Megachile paid no attention to the false flowers, while 4 of 10 

 individuals of Bombus inspected them, 1 repeating this three times. In the 

 second experiment, 3 natural spikes were interspersed with 3 false ones, 1 

 of which was provided with honey. Anthidium ignored the artificial flowers 

 on 38 visits, but inspected them briefly during 11 others, while one Oxybelus 

 actually entered a false corolla containing honey, as did one Odynerus also. 

 In the next test, made when the natural flowers were few, Megachile again 

 ignored the false flowers, Anthidium inspected them on 9 visits out of 31, and 

 two individuals of Musca alighted, but without entering the corolla. To 

 eliminate the competition of the natural flowers of Digitalis, 3 clusters of 

 artificial ones were placed at a distance, but among Dianthus and Tagetes. 

 Of the many visitors to the latter, but 3 individuals inspected the false 

 flowers for an instant. To check this behavior an old trunk of a conifer, 

 a few feet in height, was placed in the spot occupied before by the imitation 

 clusters, with the result that 3 different insects were observed to inspect 

 it on two different occasions, indicating the precautions that must be taken 

 in interpreting mere inspections. 



Three clusters of false flowers were suspended in a mass of normal ones 

 of Lathyrus latifolius, the mimicry being so close that it would not have been 

 suspected by one not forewarned. One individual of Stelis and one of Bombus 

 gave an instant's attention to an artificial cluster, while Megachile inspected 

 it in 10 of 17 visits. To test the significance of this, 3 pieces of pole were 

 placed where artificial clusters had been, when it was found that Megachile 

 made 9 inspections of them in the course of 36 visits. When the imitation 

 clusters were placed 4 dm. in front of the main group, Bombus ignored them 

 and Odynerus merely alighted for a moment on one of the artificial leaves. 

 In 50 visits Megachile hovered above the false flowers in but 8 cases, paying 

 no attention whatever to them in the others. 



Artificial flowers of green leaves. — To avoid the objection arising from 

 the use of something other than vegetable tissue and to further test the 

 response to green objects furnished with honey, leaves of Ribes rubrum and 

 Acer pseudoplatanus were rolled into cups, the free edges trimmed, and 

 honey placed in the center. Six flowers of this type were attached to several 

 branches of Symphoricarpus racemosus, where they were almost invisible 

 in the mass of foliage. In a half-hour the leaf flowers had received 19 visits, 

 chiefly from flies; the bees were represented by two visits of Apis. In the 

 next test Apis made 8 visits, 1 more than by the other insects, and in the 

 third, Bombus furnished 2 of the 15 visits, again chiefly due to flies. The 



