RESUME. 135 



far made seems to be that of Dodel-Port (1882:294), though Lovell (1912, 

 1914) mentions a number of observations by bee-keepers, which belong in 

 this general category. Dodel-Port found that two beds of the scarlet- 

 runner bean, though blooming profusely, failed to set fruits with the excep- 

 tion of a single legume, in spite of the fact that elsewhere in Zurich it fruited 

 abundantly. During this period repeated observation failed to reveal a 

 single pollinating insect on these flowers, though honey-bees, bumble-bees, 

 butterflies, wasps, and flies were numerous on the adjacent flowers of 

 Cerinthe, Calendula, Centaurea, Bidens, Cichorium, etc., which exercised a 

 greater selective power upon them. Just as long as these flowers that were 

 richer in honey continued to bloom, the bean was at a complete disadvan- 

 tage in the competition, but after the middle of August they began to 

 disappear and the brilliance of the beans then brought the visitors to them 

 and fruits began to develop frequently. A comparison with the plants of 

 this species in other gardens led to the inevitable conclusion that it formed 

 no fruits in the Botanical Garden as a direct consequence of the diversion 

 of the bumble-bees through the other more attractive flowers. 



Constancy as shown by pollen loads. — The composition of pollen 

 loads can only reveal the behavior of bees in the collection of pollen. Too 

 little is known of the detailed habits of the various species of bees with 

 respect to the gathering of pollen and nectar, and more exact knowledge 

 of constancy must await upon this among other matters. The analysis of 

 pollen loads showed that 86 individuals carried mixed loads in comparison 

 with 121 carrying pure pollen. With respect to species, none represented 

 by 5 or more individuals was characterized by pure loads, the latter being 

 found only on 4 bees belonging to Halictus pulzenus, on 3 of Monumetha 

 albifrons, and 2 of Osmia melanotricha. In Apis the relation of pure loads 

 to mixed was 28:3, in Andrena crataegi 15:3, in Bombus juxtus 19:6, and in 

 bifarius 7:5. Variation in behavior within a genus is shown by the fact 

 that for B. occidentalis the ratio was 9 : 17, Andrena madronitens 1 : 3, and 

 A. vicina 2:3, while for Halictus medionitens it was 2:2. It is clear that 

 bees in general are no more constant in the collection of pollen than in the 

 gathering of nectar. 



Determination and designation of constancy. — Our present know- 

 ledge of the habits of bees is inadequate to a comprehensive and accurate 

 treatment of constancy. Nearly all those interested in the subject have 

 defined constancy in different terms and the studies that bear directly on 

 this relation have been so few and fragmentary that further advance must 

 await special investigations which are at once extensive and intensive. 

 These must take account of behavior with respect to both pollen and 

 nectar, the differences between individuals, sexes, ages, species, etc., 

 single flights and those made on the same day, and on different days 

 and at different parts of the flowering period of preferred species. Further- 

 more, they must employ experiment to vary conditions in a definite manner 

 in order to test habit under new circumstances and to furnish quantitative 

 results, which will permit the expression of the degree of constancy for 

 individual or species in numerical terms. 



