{49}The first cell thus made is stored with pollen and honey, and an egg is deposited. Then a barrier has to be constructed to close this chamber; the material used for the barrier is the pith of the stem, and the Insect cuts the material required for the purpose from the walls of the second chamber; the excavation of the second chamber is, in fact, made to furnish the material for closing up the first cell. In this way a chain of cells is constructed, their number being sometimes as many as fifteen. The mode in which the bees, when the transformations of the larvae and pupae have been completed, escape from the chain of cells, has been the subject of much discussion, and errors have arisen from inference being allowed to take the place of observation. Thus Dufour, who noted this same mode of construction and arrangement in another Hymenopteron (_Odynerus nidulator_), perceived that there was only one orifice of exit, and also that the Insect that was placed at the greatest distance from this was the one that, being the oldest of the series, might be expected to be the first ready to emerge; and as the other cocoons would necessarily be in the way of its getting out, he concluded that the egg that was last laid produced the first Insect ready for emergence. Fabre tested this by some ingenious experiments, and found that this was not the case, but that the Insects became ready to leave their place of imprisonment without any reference to the order in which the eggs were laid, and he further noticed some very curious facts with reference to the mode of emergence of _Osmia tridentata_. Each Insect, when it desires to leave the bramble stem, tears open the cocoon in which it is enclosed, and also bites through the barrier placed by the mother between it and the Insect that is next it, and that separates it from the orifice of exit. Of course, if it happen to be the outside one of the series it can then escape at once; but if it should be one farther down in the Indian file it will not touch the cocoon beyond, but waits patiently, possibly for days; if it then still find itself confined it endeavours to escape by squeezing past the cocoon that intervenes between it and liberty, and by biting away the material at the sides so as to enlarge the passage; it may succeed in doing this, and so get out, but if it fail to make a side passage it will not touch the cocoons that are in its way. In the ordinary course of events, supposing all to go well with the family, all the cocoons produce their inmates in a state for emergence within {50}a week or two, and so all get out. Frequently, however, the emergence is prevented by something having gone wrong with one of the outer Insects, in which case all beyond it perish unless they are strong enough to bite a hole through the sides of the bramble-stem. Thus it appears that whether a particular _Osmia_ shall be able to emerge or not depends on two things—(1) whether