198

REPLY TO OBJECTIONS.

by the color. So far, then, from weakening my conclusions, the fact, so far as it goes, tends to strengthen them, because it shows that notwithstanding this tendency the blue was preferred, and the honey on colorless glass neglected. The legitimate conclusion to be drawn seems, I confess, to me, not that my mode of observation was faulty, but rather that the preference of the bees for particular colors is even somewhat greater than the numbers would indicate.

Next, Dr. Müller objects that when disturbed from one drop of honey, the bees naturally would, and that in his experiments they actually did, fly to the next. As a matter of fact, however, this did not happen in mine, because, to avoid this source of error, when I removed the color I gave the bee a good shake, and so made her take a flight before settling down again.

According to my experience, bees differ considerably in character, or, I should rather perhaps say, in humour. Some are much shyer and more restless than others. When disturbed from the first drop of honey, some are much longer before they settle on the next than others. Much also, of course, depends on how long the bee has been experimented on. Bees, like men, settle down to their work. Moreover, it is no doubt true that, cæteris paribus, a bee in search of honey will go to the nearest

source.

But, as a matter of fact, in my hundred experiments I had but very few cases like those quoted above from Dr. Müller. This arose partly from the fact that my bees were frequently changed, and partly because, as already mentioned, I took care, in removing the color, to startle the bee enough to make her take a little flight before alighting again. Dr. Müller says that in

PREFERENCES OF BEES.

199

his experiments, when the bee did not go to the next honey, it was when he shook her off too vigorously. I should rather say that in his observations he did not shake the bee off vigorously enough. The whole objection, however, is open to the same remark as the last. The bee would have a tendency, of course, like any one else, to go to its goal by the nearest route. Hence I never supposed that the figures exactly indicate the degree of preference. The very fact, however, that there would naturally be a tendency on the part of the bees to save themselves labour by going to the nearest honey, makes the contrast shown by my observations all the more striking.

I have never alleged that it was possible, in the case of bees (or, for that matter, of men either), to get any absolute and exact measure of preference for one color over another. It would be easy to suggest many considerations which would prevent this. For instance, something would probably depend on the kind of flower the bee had been in the habit of visiting. A bee which had been sucking daisies might probably behave very differently from one which had been frequenting a blue flower.

So far, however, as the conclusions which I ventured to draw are concerned, I cannot see that they are in any way invalidated by the objections which Dr. Müller has urged, which, on the other hand, as it seems to me, rather tend to strengthen my argument.

I may perhaps be asked, If blue is the favourite color of bees, and then pink, and if bees have had so much to do with the origin of flowers, how is it there are so few blue and pink ones?

The explanation I believe to be that all blue flowers

200

THE COLORS OF FLOWERS.

have descended from ancestors in which the flowers were red, these from others in which they were yellow, while originally they were all green-or, to speak more precisely, in which the leaves immediately surrounding the stamens and pistil were green; that they have passed through stages of yellow, and generally if not always red, before becoming blue.

It is, of course, easy to see that the possession of color is an advantage to flowers in rendering them more conspicuous, more easily seen, and less readily overlooked, by the insects which fertilize them; but it is not quite so clear why, apart from brilliancy and visibility at a distance, one color should be more advantageous than another. These experiments however, which show that insects have their preference, throw some light on the subject.

Where insects are beguiled into visits, as is the case especially with flies, they are obviously more likely to be deceived if the flowers not only, as is often the case, smell like decaying animal substance, but almost resemble them in appearance. Hence many fly flowers not only emit a most offensive smell, but also are dingy yellow or red, often mottled, and very closely resemble in color decaying meat.

There remains another case in which allied flowers, and species, moreover, which are fertilized by very much the same insects, are yet characterized by distinct colors. We have, for instance, three nearly allied species of dead nettle-one white (Lamium album), one red (Lamium maculatum), and one yellow (Lamium galeobdolon or luteum).

Now, if we imagine the existence in a single genus of three separate species, similar in general habit and

THE COLORS OF FLOWERS.

201

appearance, and yet mutually infertile, it is easy to see that it would be an advantage to them to have their flowers differently colored. The three species of Lamium above mentioned may be growing together, and yet the bees, without difficulty or loss of time, can distinguish the species from one another, and collect pollen and honey without confusing them together. On the other hand, if they were similarly colored, the bees could only distinguish them with comparative difficulty, involving some loss of time and probably many mistakes.

I have not yet alluded especially to white flowers. They seem to stand in a somewhat special position. The general sequence, as I have suggested, is from green, through yellow and red, to blue. Flowers normally yellow seldom sport into red or blue; those normally red often sport into yellow, but seldom into blue. On the other hand, flowers of almost any color may sport into white. White is produced by the absence of color, may therefore appear at any stage, and will be stereotyped if for any reason it should prove to be an advantage.*

* The genesis of the color is a large and interesting question. It may be due to various causes, and is by no means always owing to the presence of a different coloring matter. For instance, as Professor Foster has observed to me, many species of Iris occur in blue and yellow forms. The yellow is largely, or wholly, produced by chromatoplacts, the purple or blue to cell-sap, and if the latter is absent the yellow becomes apparent.

( 202 )

CHAPTER X.

ON THE LIMITS OF VISION OF ANIMALS.

ANTS AND COLORS.

I HAVE elsewhere recorded a series of experiments on ants with light of different wave-lengths, in order, if possible, to determine whether ants have the power of distinguishing colors. For this purpose I utilized the dislike which ants, when in their nest, have for light. Not unnaturally, if a nest is uncovered, they think they are being attacked, and hasten to carry their young away to a darker and, as they suppose, a safer place. I satisfied myself, by hundreds of experiments, that if I exposed to light the greater part of a nest, but left any of it covered over, the young would certainly be conveyed to the dark part. In this manner I satisfied myself that the various rays of the spectrum act on them in a different manner from that in which they affect us; for instance, that ants are specially sensitive to the violet rays.

But I was anxious to go beyond this, and to attempt to determine whether, as M. Paul Bert supposed, their limits of vision are the same as ours. We all know that "Ants, Bees, and Wasps."