The cover of the pot consists solely of round pieces, very nearly alike and more or less numerous. Sometimes I find only two, sometimes I count as many as ten, closely stacked. At times, the diameter of these pieces is of an almost mathematical precision, so much so that the edges of the disk rest upon the ledge. No better result would be obtained had they been cut out with the aid of compasses. At times, again, the piece projects slightly beyond the mouth, so that, to enter, it has to be pressed down and curved cupwise. There is no variation in the diameter of the first pieces placed in position, those nearest to the honey. They are all of the same size and thus form a flat cover which does not encroach on the cell and will not afterwards interfere with the larva, as a convex ceiling would. The subsequent disks, when the pile is numerous, are a little larger; they only fit the mouth by yielding to pressure and becoming concave. The Bee seems to make a point of this concavity, for it serves as a mould to receive the curved bottom of the next cell.

When the row of cells is finished, the task still remains of blocking up the entrance to the gallery with a safety-stopper similar to the earthen plug with which the Osmia closes her reeds. The Bee then returns to the free and easy use of the scissors which we noticed at the beginning when she was fencing off the back part of the Earth-worm's too deep burrow; she cuts out of the foliage irregular pieces of different shapes and sizes and often retaining their original deeply-indented margins; and with all these pieces, very few of which fit at all closely the orifice to be blocked, she succeeds in making an inviolable door, thanks to the huge number of layers.

Let us leave the Leaf-cutter to finish depositing her eggs in other galleries, which will be colonized in the same manner, and consider for a moment her skill as a cutter. Her edifices consist of a multitude of fragments belonging to three categories: oval pieces for the sides of the cells; round pieces for the lids; and irregular pieces for the barricades at the front and back. The last present no difficulty: the Bee obtains them by removing from the leaf some projecting portion, as it stands, a serrate lobe which, owing to its notches, shortens the insect's task and lends itself better to scissor-work. So far, there is nothing to deserve attention: it is unskilled labour, in which an inexperienced apprentice might excel.

With the oval pieces, it becomes another matter. What model has the Megachile when cutting her neat ellipses out of the delicate material for her wallets, the robinia-leaves? What mental pattern guides her scissors? What system of measurement tells her the dimensions? One would like to picture the insect as a living pair of compasses, capable of tracing an elliptic curve by a certain natural inflexion of its body, even as our arm traces a circle by swinging from the shoulder. A blind mechanism, the mere outcome of its organization, would alone be responsible for its geometry. This explanation would tempt me if the large oval pieces were not accompanied by much smaller ones, also oval, which are used to fill the empty spaces. A pair of compasses which changes its radius of its own accord and alters the curve according to the plan before it appears to me an instrument somewhat difficult to believe in. There must be something better than that. The circular pieces of the lid suggest it to us.

If, by the mere flexion inherent in her structure, the Leaf-cutter succeeds in cutting out ovals, how does she succeed in cutting out rounds? Can we admit the presence of other wheels in the machinery for the new pattern, so different in shape and size? Besides, the real point of the difficulty does not lie there. These rounds, for the most part, fit the mouth of the jar with almost exact precision. When the cell is finished, the Bee flies hundreds of yards away to make the lid. She arrives at the leaf from which the disk is to be cut. What picture, what recollection has she of the pot to be covered? Why, none at all: she has never seen it; she does her work underground, in utter darkness! At the utmost, she can have the indications of touch: not actual indications, of course, for the pot is not there, but past indications, useless in a work of precision. And yet the disk to be cut out must have a fixed diameter: if it were too large, it would not go in; if too small, it would close badly, it would slip down on the honey and suffocate the egg. How shall it be given its correct dimensions without a pattern? The Bee does not hesitate for a moment. She cuts out her disk with the same celerity which she would display in detaching any shapeless lobe that might do for a stopper; and that disk, without further measurement, is of the right size to fit the pot. Let whoso will explain this geometry, which in my opinion is inexplicable, even when we allow for memory begotten of touch and sight.

One winter evening, as we were sitting round the fire, whose cheerful blaze unloosed our tongues, I put the problem of the Leaf-cutter to my family:

'Among your kitchen-utensils,' I said, 'you have a pot in daily use; but it has lost its lid, which was knocked over and broken by the Tomcat playing among the shelves. To-morrow is market-day and one of you will be going to Orange to buy the week's provisions. Would she undertake, without a measure of any kind, with the sole aid of memory, which we would allow her to refresh before starting by a careful examination of the object, to bring back exactly what the pot wants, a lid neither too large nor too small, in short the same size as the top?'

It was admitted with one accord that nobody would accept such a commission without taking a measure with her, or at least a bit of string giving the width. Our memory for sizes is not accurate enough. She would come back from the town with something that 'might do'; and it would be the merest chance if this turned out to be the right size.

Well, the Leaf-cutter is even less well-off than ourselves. She has no mental picture of her pot, because she has never seen it; she is not able to pick and choose in the crockery-dealer's heap, which acts as something of a guide to our memory by comparison; she must, without hesitation, far away from her home, cut out a disk that fits the top of her jar. What is impossible to us is child's-play to her. Where we could not do without a measure of some kind, a bit of string, a pattern or a scrap of paper with figures upon it, the little Bee needs nothing at all. In housekeeping matters she is cleverer than we are.

One objection was raised. Was it not possible that the Bee, when at work on the shrub, should first cut a round piece of an approximate diameter, larger than that of the neck of the jar, and that afterwards, on returning home, she should gnaw away the superfluous part until the lid exactly fitted the pot? These alterations made with the model in front of her would explain everything.