No. 527] COLOR SENSE OF THE HONEY-BEE 689 



But if the species are closely allied in form and color, 

 as among the buttercups, spiraeas and golden-rods, the 

 bees do not carefully discriminate between them. Yet 

 even in these genera the honey-bee often exhibits a re- 

 markable power of distinguishing between allied species, 

 even when they are of the same color. I have described 

 in the American Naturalist how in an upland pasture 

 honey-bees showed a * marked preference for the flat- 

 topped corymbs of So!i</<if/<> hn/ccohi fa ( Euthumia t/nuni- 

 nifolia) to the panicled inflorescence of 8. rugosa. 



They were repeatedly seen to leave S. lanceolata, and after flying 



they had left only a few moments before. In another instance a bee 

 was seen to wind its wa\ among the plants of the latter species until it 

 found an isolated plant of S. lanceolata. A plant of each of the above 



as a single cluster; a honey-bee rested on 8. lanceolata and it seemed 



such was not the case, for presently it flew away to another plant of 

 the former. 20 



The bees must, therefore, have perceived differences 

 between these two species of Solidago, though they occa- 

 sionally ignored them. 



When a plant species displays variously colored 

 flowers, it is obvious that they are alike in shape, odor 

 and nectar, and differ in color alone. Under these cir- 

 cumstances it is for the advantage of bees to pass from 

 one color to another, and this they speedily learn to do. 

 In an earlier paper I have pointed out that form is a more 

 important factor than color in determining the visits of 

 the long-tongued bees and butterflies. 21 This conclusion 

 is confirmed by Dr. Graenicher in a very important con- 

 tribution on the pollination of the Composite. After a 

 careful comparison of the effect of tube length, color, 

 and odor on the limitation of visitors, he says : 



It may be stated that according to the results obtained from a study 



