What stimulus does the insect obey when it employs the reserve powers that slumber in its race? Of what use are its industrial variations? The Osmia will yield us her secret with no great difficulty. Let us examine her work in a cylindrical habitation. I have described in full detail, in the foregoing pages, the structure of her nests when the dwelling adopted is a reed-stump or any other cylinder; and I will content myself here with recapitulating the essential features of that nest-building. We must first distinguish three classes of reeds according to their diameter: the small, the medium-sized and the large. I call small those whose narrow width just allows the Osmia to go about her household duties without discomfort. She must be able to turn where she stands in order to brush her abdomen and rub off its load of pollen, after disgorging the honey in the centre of the heap of flour already collected. If the width of the tube does not admit of this operation, if the insect is obliged to go out and then come in again backwards in order to place itself in a favourable posture for the discharge of the pollen, then the reed is too narrow and the Osmia is rather reluctant to accept it. The middle-sized reeds and a fortiori the large ones leave the victualler entire liberty of action; but the former do not exceed the width of a cell, a width agreeing with the bulk of the future cocoon, whereas the latter, with their excessive diameter, require more than one chamber on the same floor. When free to choose, the Osmia settles by preference in the small reeds. Here, the work of building is reduced to its simplest expression and consists in dividing the tube by means of earthen partitions into a straight row of cells. Against the partition forming the back wall of the preceding cell the mother places first a heap of honey and pollen; next, when the portion is seen to be enough, she lays an egg in the centre of it. Then and then only she resumes her plasterer's work and marks out the length of the new cell with a mud partition. This partition in its turn serves as the rear-wall of another chamber, which is first victualled and then closed; and so on until the cylinder is sufficiently colonized and receives a thick terminal stopper at its orifice. In a word, the chief characteristic of this method of nest-building, the roughest of all, is that the partition in front is not undertaken so long as the victualling is still incomplete, or, in other words, that the provisions and the egg are deposited before the Bee sets to work on the partition.