It is well known that bees and many other flower-visiting insects are accustomed to visit a certain kind of flower at certain times and at those times to neglect other kinds. This habit places before us a problem which Darwin has raised. "Bumblebees and the honey-bees are good botanists, " wrote the illustrious observer, "because they know that varieties can present great differences in the color of the flower wliile belonging to the same species. I have frequently seen," he adds, "bumblebees fly directly to a plant of common Dictamnus fraxi-nella; and all went to a white variety, to a purple variety of Viola tricolor, and to another of golden yellow." Gaston Bonnier made similar observations upon the varieties of hollyhocks, and Plateau .upon a certain number of plants frequented by different insects, both concluding from their observations that honey-visiting species are indifferent to colors.

But we know that this is not so, and that is why we are going to try to solve the problem proposed by Darwin. If anthophilous insects are "good botanists" which distinguish the corollas of a given species whatever its color may be, it is evident that they have as guides the two elements of the flower which vary only slightly, the form and the perfume. What part in this attraction the one or the other takes, we cannot say, but we must acknowledge with Bonnier that floral perfumes are of two very different kinds, the one produced by the essential oils which have their principal seat in the whole extent of the corolla, and the other by the nectar, which has a particular aroma. Insects come to the flowers for nectar, but the experiments of Lubbock (1875) have shown that they strongly sense the essential oils, and we must have to agree that the odor of these oils must play a part in the complex sensations which permits insects to be "good botanists."

There are flowers which exhale a fetid odor and which are distasteful from that fact to most insects (flowers with the nauseous, idiopathic odor of Delpino, the indoloid perfumes of Kerner). Aristolochia sipho smells like urine, Aristolochia labiosa like spoiled fish, Arum dracunculus like a decomposed cadaver. But these flowers are visited by insects, and some even have the habit of visiting excrement and putrefying flesh. Paul Knuth reports that upon five flowers of Arum dracunculus he found 463 Coleoptera of which 377 belonged to a species (Saprinus nitidulus), which feeds on cadavers, and that Walker at Gibraltar saw in the same flowers CallipJiora vomitoria, Scatophagus, Bermestes, and three species of Saprinus. We know that these sarcophagic and stercophagic insects perceive odors from afar,, and for them it is undoubtedly odor which is the principal attraction. On the other hand, flowers with agreeable perfumes (the sympathetic odors of Delpino, the benzoloid and terpinoid perfumes of Kerner) may attract from afar certain very sensitive anthophilous species. Has not Kerner seen the convolvulous sphinx proceed directly to the invisible flowers of honeysuckle situated a hundred meters away?

It is impossible to apply a general rule to anthophilous insects. Without doubt they depend on both sight and smell to direct them toward flowers, but with some it is sight which plays the principal part and with others it is the olfactory sense. In general, we may say with Perez that "for distances at which sight cannot be any help, the bees, and without doubt most insects, are guided in their search for flowers by odors transported by the currents of air. . . . When the distance is short enough for the sense of sight to intervene, the insect flies straight to the flower whose appearance attracts him, but this appearance is enough in the absence of any probable perception of perfume when the accidents of flight bring it near a flower. It alights then if the perfume is one which pleases it ; if not, it leaves.