violet. When reporting these results, we said (Science, LV, May 12, 1922):

The question may still be raised that these phototropic reactions of Drosophila are due to fluorescence of eye media, similar to that experienced by the human eye when exposed to ultraviolet light. All that can be said at present in this connection is that the intensity was so low that we did not experience the visual sensation characteristic of such fluorescence, but the files reacted promptly and definitely.

We were unable to follow up this work in 1922, but during that year there appeared in Die Naturwissenschaften, XXXVII, pp. 738-740, a short article by Kiihn and Pohl, it apparently being a preliminary announcement of a more detailed report. An actual spectrum was projected upon a horizontal plate in a semi-dark place. Bees (probably ordinary honeybees, but this is not clearly stated) were trained to find food at given places on this spectrum, and thereafter the bees hunted for food there, even though none was present. They were successfully trained to come to .365 (ultraviolet) as well as to other parts of the spectrum, such as .578, ;546, .436, etc. It is not clear that in the case of the ultraviolet the authors have ruled out the possibility that the bees remembered its location with respect to the "visible" light. They could move the spectrum around from place to place, but the ultraviolet would always be on the other side of the blue from the yellow. The higher Hymenoptera have such a wonderful and to us still mysterious ability to note and to remember locations that this ability must on no account be forgotten. Still, I believe that honeybees could be trained to distinguish between colors which are indistinguishable to us but which differ in their reflection of ultraviolet; for example, between a magnesium oxide surface and one painted with Chinese white.

Kiihn and Pohl explain Frisch's finding that the bees did not distinguish between his Green No. 10 paper and gray by the fact that this paper reflects from below .365 to .492. Therefore the green paper must have appeared to the bees strongly veiled in white (''cine starke Weissverhiillungzustande"). Whether it would be white or some other color is a matter of terms; whether its ultraviolet reflection entered into the problem at all depends upon the extent to which the double transmission of light through glass cut out these wave-leilgths (isee table 24 and p. 265).

During the past season I was able to carry out experiments on a fair variety of flower-visiting insects. I used the method of putting the insect to be tested in a box having a window at each end. These windows consisted of one or another of the light-filters listed on page 238 and used also in determining by photography the reflection of certain colors from flowers. The top of the box was glass surrounded by a light-tight hood within which I could put my head, drawing the free end of the hood around my neck so that, although I could see clearly what was going on in the box, no stray light was admitted. This box was taken into the field where insects freshly captured in the act of visiting flowers could be immediately tested, ruling out as far as possible any ill effects that might come from confinement. Also, the flower which the insect was visiting could be noted and certain of its colors determined by means of the same filters as were used in testing the vision of the insect. The portability of these two pieces of apparatus was a great convenience.

A word more should be said about the filters (see p. 238) that were used. The ultraviolet (G586A) was so dense that under no normal conditions of light could print be read through it. The blue (49C) was also dense, but print could be read through it without much difficulty in a fairly good light. The green, red, yellow and "noviol" were very transparent, the transparency increasing in the order given, and print could easily be read through them.

Many tests were made, chiefly on various species of bees and wasps, but also on butterflies, moths, and flies, especially the Syrphidse, so abundant at flowers. There seems to be no point in copying my entire note book into this paper. I was not engaged primarily in testing the relative visibility to insects of the colors we can see; neither was I attempting to determine whether the insects see colors, including ultraviolet, as colors or merely as different intensities of illumination. My object was to get an idea as to the extent to which at least flower-visiting insects see ultraviolet. The results may be summarized as follows: Certain individuals, especially Lepidoptera, stubbornly refused to react to any conditions of illumination, even clear sunlight, when they were in the box. Certain others reacted only when urged to do so by jarring the box, but then their reactions were definite. Most of the flower-visiting insects, especially the bees and wasps, reacted to all of the tests without artificial urging. In every case in which an insect reacted to any illumination, either clear or colored, it definitely reacted to ultraviolet, going to the apparently opaque G586A as though it were clear glass and buzzing against it in an attempt to get through. They did this, even though the other end of the box contained the, to us, very transparent red filter. The objection that the ultraviolet filter passed small but appreciable-to-us amounts of other colors falls before experiments in which insects were given their choice between this filter and one passing large amounts of the other colors.

The following samples show the character of the tests made. When an insect was at one end of the box either the filter at the other end was changed or light was cut off from the filter at which the insect was by an opaque slide placed outside of it. Thus, unless otherwise stated, the insect traveled the entire length of the box from one filter to another at each test. Under each species the tests are recorded in the order in which they were made for that individual.

Apis mellifica (the honeybee) worker.—Went from dark to clear glass; from dark to ultraviolet. From dark to blue; from blue to ultraviolet. From dark to green; from green to ultraviolet. From dark to red; from red to ultraviolet. When at the ultraviolet filter, would not leave it to go to several thicknesses of clear glass.

Hombus bimaculatus (a common bumblebee) worker.—Went from dark to clear glass; from dark to ultraviolet. From dark to blue; from blue to ultraviolet, but not quickly. From dark to red; from red to ultraviolet, but not quickly. When at ultraviolet, would not leave it for either blue or red or several thicknesses of clear glass.

Polistes annularis (a common w-asp, making an open paper nest) worker.—Went from dark to clear glass; from dark to ultraviolet. Went from red to ultraviolet. No reaction, even when jarred, in dark vs. green and in dark vs. blue, but then went without stimulation from dark to clear glass, from dark to ultraviolet, and from dark to "noviol." Went from midway between ultraviolet and "noviol'' to ultraviolet. Then from dark to "noviol." Would not leave "noviol" for ultraviolet. Then from dark to ultraviolet. Would not leave ultraviolet for "noviol." Went from dark to yellow; from yellow to ultraviolet; and then from dark to blue, although it did not show this reaction in the trial noted above. In four trials of ultraviolet vs. blue, each starting from the middle of the box, it went to ultraviolet three times and to blue once. Went from dark to green and from dark to ultraviolet, but when at either ultraviolet or green would not go to the other. Showed a slight preference for green over red. Went from dark to red. Showed a marked preference for blue over red.

Xenoglossa pruinosa (a solitary bee visiting especially squash or pumpkin flowers) female.—Went from dark to clear glass; from dark to ''noviol"; from dark to ultraviolet. Went from dark to red, but only slowly; from red to ultraviolet at once. Went from dark to yellow rather slowly; from yellow to ultraviolet. From dark to green; from green to ultraviolet. From dark to blue; from blue to ultraviolet. From dark to "noviol"; from "noviol" to ultraviolet.

Melissodes bimaculaia (a solitary bee frequently found visiting especially flowers like the morning glories) male.—This sample is given here chiefly because the individual was a male. Went from dark to clear; from dark to ultraviolet. From dark to red; from red to ultraviolet. From dark to yellow; from yellow to ultraviolet. From dark to green; from green to ultraviolet. From dark to blue; from blue to ultraviolet. From dark to "noviol"; from "noviol" to ultraviolet.

A Psammocharid (Pompilid) Wasp, species not determined.—These are spider-hunters, but they are not rarely found visiting flowers. Went from dark to clear; from dark to ultraviolet. From dark to red; from red to ultraviolet. From dark to yellow; from yellow to ultraviolet. From dark to green; from green to ultraviolet. Went from dark to "noviol," and when at "noviol" wandered slightly away from it toward ultraviolet, but always returned to the ''noviol." Went from dark to ultraviolet, and Avhen at ultraviolet showed no inclination to wander in the direction of "noviol."'

A Syrphid Fly, species not determined.—Went from dark to clear: from dark to ultraviolet. Went from dark to red rather slowly; went quickly from red to ultraviolet. From dark to yellow; from yellow to ultraviolet. Went, but only after stimulation, from dark to green; went without stimulation from green to ultraviolet. Went from dark to blue, but only slowly and after stimulation; went from blue to ultraviolet quickly and without stimulation. Went, after stimulation, from dark to "noviol '; went from "noviol" to ultraviolet without stimulation.

E//argyreus iifysus (one of the "skipper" butterflies frequently found about flowers).— Fluttered from dark to clear. When the windows were changed to dark vs. ultraviolet the butterfly "went to sleep" and refused to move even when the box was jarred, but when the insect was touched it flew at once to ultraviolet and tried to get through the filter. In dark vs. red it could not be persuaded to react even by touching, but in red vs. ultraviolet it rlew to the ultraviolet. In dark vs. yellow much stimulation evoked only a slight reaction in favor of the yellow; in yellow r.s. ultraviolet no reaction could be evoked. Then dark vs. ultraviolet was tried again and a slight jarring of the box sent it at once to the ultraviolet. Fpon jarring the box it went from dark to blue and then from blue to ultraviolet. Keyetitions of the last two tests gave the same results.

The last two paragraphs arc quoted from my notes as examples of individuals (perhaps species) that required stimulation to make them react under these experimental conditions. Xeedless to say, care was taken not to give the stimulation in such a way as to drive the insect toward a selected end of the box, and it will be seen that the reactions, when made, were quite in harmony with the spontaneous reactions of the other insects.

Two things seem to be quite certain in these results. The first is that, whether insects perceive red, green, yellow, blue, and ultraviolet as colors or merely as illumination, they are affected by and react to each of these, including red and ultraviolet. The second is that the insects see44 ultraviolet as a strong light; in other words, they see it well—even better than other light. Not only do positively phototropic insects leave a brilliant red or relatively bright blue window that they are trying to get through to fly to an opaque-to-us ultraviolet window, but at least some of the flower visitors leave a "noviol" window for an ultraviolet one. Since

"An extremely mechanistic psychologist may object to the word "see," but If a creature having eyes is "affected by and reacts to" light by going toward it In a very definite manner, it Is pure logomachy to say that It does not see that light.

"noviol" is supposed to transmit practically the entire visible-to-us spectrum with little reduction of intensity of illumination, while G586A cuts out practically all of this and transmits only a part of the ultraviolet of sunlight (see figure 24), this last reaction is quite striking and it was repeatedly made by various species and in an apparently definite, often quick manner, not as though by accident.

If, then, on the one hand, insects, though able to see the light we call red, do not see it as a color distinct from some shade of gray and, on the other hand, they can see (whether as a color or not) the ultraviolet light which we can not see, what kind of vision shall we say they have and how do the flowers look to them? I have spent considerable time searching the literature of human optics (and friends, including the Research Information Service of the National Research Council, have probably spent even more time searching in my behalf).for some record of a test on a color-blind man as to his ability to see ultraviolet. Apparently no record of such a test has been published. Although T know no theory of human optics that would fit such a case, it does not seem impossible that colorblind men do see ultraviolet. Perhaps they have just as great a range of color-vision as normal men, but both limits of the range have been shifted toward shorter wave-lengths, possibly a reversion to the vision of our invertebrate ancestors. Such a man, if such there be, might describe for us the world of flowers from an insect's point of view.

Discussion

It has been shown that, in addition to the colors which man can see, some flowers are ultraviolet, while others are not at all ultraviolet, and still others have an ultraviolet pattern. It has also been shown that flower-visiting insects can see ultraviolet as well as or even better than they can see the rays perceived as light by man. The effect of this work is to add one more color—and apparently an important one—to the list of those which plants use in facilitating the visits of insects, if plants do use colors for such a purpose. It would be easy and it is rather tempting to develop the subject along this line. Xewbigin was both heterodox and cynical when he said:

Thus, according to the school which is usually known as the Darwinian, colour, wherever seen, is due to the favouring influence of Natural Selection, and is in some way useful to the species. In the view of the popularisere of the subject, it therefore becomes the main object of the naturalist to invent as ingenious an explanation as possible of the way In which it is useful. If the naturalist's powers of invention fall, though this happens but rarely, then the colour Is non-significant, or better still, the animal has recently changed its habitat, and is no longer perfectly adapted to its environment. The theory

« PreviousContinue »