134 COMPETITION AND CONSTANCY. 



RESUME. 

 Experimental results in competition. — In the case of adult insects, 

 especially bees, habit is the controlling factor in competition. This is 

 shown by the fact that the standard plant, which is the one that the insect 

 was in the habit of visiting, was favored in 36 cases, while the bouquet, 

 consisting of flowers out of their natural position, had the advantage in 

 but 11 cases. Moreover, the difference between the number of visitors in 

 each experiment was regularly much greater when the plant was preferred 

 than when the bouquet was. The reversals of choice were caused by flowers 

 with a large amount of nectar, mostly members of the rose family, such as 

 Rubus strigosus, Prunus demissa, and Opulaster opulifolius, or by regular 

 flowers well-supplied with nectar and pollen, such as Chamaenerium, in 

 competition with specialized ones such as Monarda. In a few cases, the 

 abundance of pollen was the deciding factor, as when Rubus deliciosus or 

 Rosa acicularis proved more attractive, doubtless owing to a shift in the 

 nutrition requirements at the time. Flowers in vials obtained very few 

 visitors, even in comparison with bouquets, probably because of the strange- 

 ness of the vial itself. The effect of the vial is well exhibited in the experi- 

 ment where the two species of Geranium were employed reciprocally as 

 standard plants and in vials. With G. richardsoni as the standard the re- 

 spective figures were 88 and 9; withG. caespitosum as the standard, 83 and 

 10. When standard roses were in competition with roses and with Cha- 

 maenerium in vials, the figures were 85, 49, and 7. The vials alone reduced 

 visits nearly a half, and a strange flower in addition further decreased them 

 seven times. In the case of natural competition, or competing bouquets, 

 habit is partly eliminated, and the relative attractive power of the species 

 is the chief factor. This varies, however, with respect to the composition 

 of the visiting population, and a decisive test is possible only with young 

 insects in which habits have not yet been fixed. 



Apis mellifica gave the preference to the standard plant in 24 instances 

 and to the bouquet in 4; in 15 experiments it went to the plant alone. 

 The total number of visitors to the standard was 972, to the bouquet, 153. 

 Bombus bifarius preferred the standard in all of the 9 experiments where 

 it was an important visitor, and went to it alone in 6 cases, the total number 

 of visitors to standard and bouquet being 206 and 20. B. proximus chose 

 the plant in 7 experiments out of 8, visited it alone in 6 instances, and 222 

 bees went to it in contrast to 48 to the bouquets. B. juxtus was far less 

 faithful to the standard, giving it the preference in 13 cases out of 19, 

 visiting it exclusively but twice, and yielding 564 visitors to the bouquets 

 for 920 to the standards. The preference as to species naturally shifted 

 with the progress of the season, but during the respective flowering periods 

 the marked preferences were as follows: Apis for Rubus strigosus, Bom- 

 bus for Rubus deliciosus, Rosa, Geranium, and Chamaenerium, Clisodon 

 for Monarda, Andrena for Opulaster and Prunus, and Osmia for Pentstemon. 



Effects of competition. — The final outcome of competition for pollin- 

 ators in terms of the setting of fruit and the production of seeds is difficult 

 to determine in nature, owing to the wide range of preferences among the 

 many species of insects usually present. The only direct study of this so 



