Omalus auratus (parasite of various bramble-dwellers).
     Cryptus bimaculatus, GRAV. (parasite of Osmia detrita).
     Cryptus gyrator, DUF. (parasite of Tripoxylon figulus).
     Ephialtes divinator, ROSSI (parasite of Cemonus unicolor).
     Ephialtes mediator, GRAV. (parasite of Psen atratus).
     Foenus pyrenaicus, GUERIN.
     Euritoma rubicola, J. GIRAUD (parasite of Osmia detrita).

     4. COLEOPTERA.
     Zonitis mutica, FAB. (parasite of Osmia tridentata).

Most of these insects have been submitted to a learned expert, Professor Jean Perez, of Bordeaux. I take this opportunity of renewing my thanks for his kindness in identifying them for me.—Author's Note.)

They include members of very diverse corporations. Some, more industrious and equipped with better tools, remove the pith from the dry stem and thus obtain a vertical cylindrical gallery, the length of which may be nearly a cubit. This sheath is next divided, by partitions, into more or less numerous storeys, each of which forms the cell of a larva. Others, less well-endowed with strength and implements, avail themselves of the old galleries of other insects, galleries that have been abandoned after serving as a home for their builder's family. Their only work is to make some slight repairs in the ruined tenement, to clear the channel of its lumber, such as the remains of cocoons and the litter of shattered ceilings, and lastly to build new partitions, either with a plaster made of clay or with a concrete formed of pith-scrapings cemented with a drop of saliva.

You can tell these borrowed dwellings by the unequal size of the storeys. When the worker has herself bored the channel, she economizes her space: she knows how costly it is. The cells, in that case, are all alike, the proper size for the tenant, neither too large nor too small. In this box, which has cost weeks of labour, the insect has to house the largest possible number of larvae, while allotting the necessary amount of room to each. Method in the superposition of the floors and economy of space are here the absolute rule.

But there is evidence of waste when the insect makes use of a bramble hollowed by another. This is the case with Tripoxylon figulus. To obtain the store-rooms wherein to deposit her scanty stock of Spiders, she divides her borrowed cylinder into very unequal cells, by means of slender clay partitions. Some are a centimetre (.39 inch.—Translator's Note.) deep, the proper size for the insect; others are as much as two inches. These spacious rooms, out of all proportion to the occupier, reveal the reckless extravagance of a casual proprietress whose title-deeds have cost her nothing.

But, whether they be the original builders or labourers touching up the work of others, they all alike have their parasites, who constitute the third class of bramble-dwellers. These have neither galleries to excavate nor victuals to provide; they lay their egg in a strange cell; and their grub feeds either on the provisions of the lawful owner's larva or on that larva itself.

At the head of this population, as regards both the finish and the magnitude of the structure, stands the Three-pronged Osmia (Osmia tridentata, DUF. and PER.), to whom this chapter shall be specially devoted. Her gallery, which has the diameter of a lead pencil, sometimes descends to a depth of twenty inches. It is at first almost exactly cylindrical; but, in the course of the victualling, changes occur which modify it slightly at geometrically determined distances. The work of boring possesses no great interest. In the month of July, we see the insect, perched on a bramble-stump, attack the pith and dig itself a well. When this is deep enough, the Osmia goes down, tears off a few particles of pith and comes up again to fling her load outside. This monotonous labour continues until the Bee deems the gallery long enough, or until, as often happens, she finds herself stopped by an impassable knot.

Next comes the ration of honey, the laying of the egg and the partitioning, the last a delicate operation to which the insect proceeds by degrees from the base to the top. At the bottom of the gallery, a pile of honey is placed and an egg laid upon the pile; then a partition is built to separate this cell from the next, for each larva must have its special chamber, about a centimetre and a half (.58 inch.—Translator's Note.) long, having no communication with the chambers adjoining. The materials employed for this partition are bramble-sawdust, glued into a paste with the insects' saliva. Whence are these materials obtained? Does the Osmia go outside, to gather on the ground the rubbish which she flung out when boring the cylinder? On the contrary, she is frugal of her time and has better things to do than to pick up the scattered particles from the soil. The channel, as I said, is at first uniform in size, almost cylindrical; its sides still retain a thin coating of pith, forming the reserves which the Osmia, as a provident builder, has economized wherewith to construct the partitions. So she scrapes away with her mandibles, keeping within a certain radius, a radius that corresponds with the dimensions of the cell which she is going to build next; moreover, she conducts her work in such a way as to hollow out more in the middle and leave the two ends contracted. In this manner, the cylindrical channel of the start is succeeded, in the worked portion, by an ovoid cavity flattened at both ends, a space resembling a little barrel. This space will form the second cell.

As for the rubbish, it is utilized on the spot for the lid or cover that serves as a ceiling for one cell and a floor for the next. Our own master-builders could not contrive more successfully to make the best use of their labourers' time. On the floor thus obtained, a second ration of honey is placed; and an egg is laid on the surface of the paste. Lastly, at the upper end of the little barrel, a partition is built with the scrapings obtained in the course of the final work on the third cell, which itself is shaped like a flattened ovoid. And so the work goes on, cell upon cell, each supplying the materials for the partition separating it from the one below. On reaching the end of the cylinder, the Osmia closes up the case with a thick layer of the same mortar. Then that bramble-stump is done with; the Bee will not return to it. If her ovaries are not yet exhausted, other dry stems will be exploited in the same fashion.

The number of cells varies greatly, according to the qualities of the stalk. If the bramble-stump be long, regular and smooth, we may count as many as fifteen: that, at least, is the highest figure which my observations have supplied. To obtain a good idea of the internal distribution, we must split the stalk lengthwise, in the winter, when the provisions have long been consumed and when the larvae are wrapped in their cocoons. We then see that, at regular intervals, the case becomes slightly narrower; and in each of the necks thus formed a circular disk is fixed, a partition one or two millimetres thick. (.039 to.079 inch.—Translator's Note.) The rooms separated by these partitions form so many little barrels or kegs, each compactly filled with a reddish, transparent cocoon, through which the larva shows, bent into a fish-hook. The whole suggests a string of rough, oval amber beads, touching at their amputated ends.

In this string of cocoons, which is the oldest, which the youngest? The oldest is obviously the bottom one, the one whose cell was the first built; the youngest is the one at the top of the row, the one in the cell last built. The oldest of the larvae starts the pile, down at the bottom of the gallery; the latest arrival ends it at the top; and those in between follow upon one another, according to age, from base to apex.

Let us next observe that there is no room in the shaft for two Osmiae at a time on the same level, for each cocoon fills up the storey, the keg that belongs to it, without leaving any