bramble a transparent wall which will enable me to follow the hatching day by day, at any moment which I think opportune.

The male Osmia splits his cocoon at the end of June and the female at the beginning of July. When this time comes, we must redouble our watch and inspect the tubes several times a day if we would obtain exact statistics of the births. Well, during the six years that I have studied this question, I have seen and seen again, ad nauseam; and I am in a position to declare that there is no order governing the sequence of hatchings, absolutely none. The first cocoon to burst may be the one at the bottom of the tube, the one at the top, the one in the middle or in any other part, indifferently. The second to be split may adjoin the first or it may be removed from it by a number of spaces, either above or below. Sometimes several hatchings occur on the same day, within the same hour, some farther back in the row of cells, some farther forward; and this without any apparent reason for the simultaneity. In short, the hatchings follow upon one another, I will not say haphazard—for each of them has its appointed place in time, determined by impenetrable causes—but at any rate contrary to our calculations, based on this or the other consideration.

Had we not been deceived by our too shallow logic, we might have foreseen this result. The eggs are laid in their respective cells at intervals of a few days, of a few hours. How can this slight difference in age affect the total evolution, which lasts a year? Mathematical accuracy has nothing to do with the case. Each germ, each grub has its individual energy, determined we know not how and varying in each germ or grub. This excess of vitality belongs to the egg before it leaves the ovary. Might it not, at the moment of hatching, be the cause why this or that larva takes precedence of its elders or its juniors, chronology being altogether a secondary consideration? When the hen sits upon her eggs, is the oldest always the first to hatch? In the same way, the oldest larva, lodged in the bottom storey, need not necessarily reach the perfect state first.

A second argument, had we reflected more deeply on the matter, would have shaken our faith in any strict mathematical sequence. The same brood forming the string of cocoons in a bramble-stem contains both males and females; and the two sexes are divided in the series indiscriminately. Now it is the rule among the Bees for the males to issue from the cocoon a little earlier than the females. In the case of the Three-pronged Osmia, the male has about a week's start. Consequently, in a populous gallery, there is always a certain number of males, who are hatched seven or eight days before the females and who are distributed here and there over the series. This would be enough to make any regular hatching-sequence impossible in either direction.

These surmises accord with the facts: the chronological sequence of the cells tells us nothing about the chronological sequence of the hatchings, which take place without any definite order. There is, therefore, no surrender of rights of primogeniture, as Leon Dufour thought: each insect, regardless of the others, bursts its cocoon when its time comes; and this time is determined by causes which escape our notice and which, no doubt, depend upon the potentialities of the egg itself. It is the case with the other bramble-dwellers which I have subjected to the same test (Osmia detrita, Anthidium scapulare, Solenius vagus, etc.); and it must also be the case with Odynerus rubicola: so the most striking analogies inform us. Therefore the singular exception which made such an impression on Dufour's mind is a sheer logical illusion.

An error removed is tantamount to a truth gained; and yet, if it were to end here, the result of my experiment would possess but slight value. After destruction, let us turn to construction; and perhaps we shall find the wherewithal to compensate us for an illusion lost. Let us begin by watching the exit.

The first Osmia to leave her cocoon, no matter what place she occupies in the series, forthwith attacks the ceiling separating her from the floor above. She cuts a fairly clean hole in it, shaped like a truncate cone, having its larger base on the side where the Bee is and its smaller base opposite. This conformation of the exit-door is a characteristic of the work. When the insect tries to attack the diaphragm, it first digs more or less at random; then, as the boring progresses, the action is concentrated upon an area which narrows until it presents no more than just the necessary passage. Nor is the cone-shaped aperture special to the Osmia: I have seen it made by the other bramble-dwellers through my thick disks of sorghum-pith. Under natural conditions, the partitions, which, for that matter, are very thin, are destroyed absolutely, for the contraction of the cell at the top leaves barely the width which the insect needs. The truncate, cone-shaped breach has often been of great use to me. Its wide base made it possible for me, without being present at the work, to judge which of the two neighbouring Osmiae had pierced the partition; it told me the direction of a nocturnal migration which I had been unable to witness.

The first-hatched Osmia, wherever she may be, has made a hole in her ceiling. She is now in the presence of the next cocoon, with her head at the opening of the hole. In front of her sister's cradle, she usually stops, consumed with shyness; she draws back into her cell, flounders among the shreds of the cocoon and the wreckage of the ruined ceiling; she waits a day, two days, three days, more if necessary. Should impatience gain the upper hand, she tries to slip between the wall of the tunnel and the cocoon that blocks the way. She even undertakes the laborious work of gnawing at the wall, so as to widen the interval, if possible. We find these attempts, in the shaft of a bramble, at places where the pith is removed down to the very wood, where the wood itself is gnawed to some depth. I need hardly say that, although these lateral inroads are perceptible after the event, they escape the eye at the moment when they are being made.

If we would witness them, we must slightly modify the glass apparatus. I line the inside of the tube with a thick piece of whity-brown packing-paper, but only over one half of the circumference; the other half is left bare, so that I may watch the Osmia's attempts. Well, the captive insect fiercely attacks this lining, which to its eyes represents the pithy layer of its usual abode; it tears it away by tiny particles and strives to cut itself a road between the cocoon and the glass wall. The males, who are a little smaller, have a better chance of success than the females. Flattening themselves, making themselves thin, slightly spoiling the shape of the cocoon, which, however, thanks to its elasticity, soon recovers its first condition, they slip through the narrow passage and reach the next cell. The females, when in a hurry to get out, do as much, if they find the tube at all amenable to the process. But no sooner is the first partition passed than a second presents itself. This is pierced in its turn. In the same way will the third be pierced and others after that, if the insect can manage them, as long as its strength holds out. Too weak for these repeated borings, the males do not go far through my thick plugs. If they contrive to cut through the first, it is as much as they can do; and, even so, they are far from always succeeding. But, in the conditions presented by the native stalk, they have only feeble tissues to overcome; and then, slipping, as I have said, between the cocoon and the wall, which is slightly worn owing to the circumstances described, they are able to pass through the remaining occupied chambers and to reach the outside first, whatever their original place in the stack of cells. It is just possible that their early eclosion forces this method of exit upon them, a method which, though often attempted, does not always succeed. The females, furnished