le smaller, it is
scarcely noticeable, whereas, in the Horned Osmia and the Three-horned
Osmia, the male is only half or a third the size of the female, as we
have seen from the respective bulk of their cocoons. In the Mason-bee
of the Walls there is also a difference in size, though less
pronounced.

The Three-pronged Osmia has not therefore to trouble about adjusting
the dimensions of the dwelling and the quantity of the food to the sex
of the egg which she is about to lay; the measure is the same from one
end of the series to the other. It does not matter if the sexes
alternate without order: one and all will find what they need, whatever
their position in the row. The two other Osmiae, with their great
disparity in size between the two sexes, have to be careful about the
twofold consideration of board and lodging.

The more I thought about this curious question, the more probable it
appeared to me that the irregular series of the Three-pronged Osmia and
the regular series of the other Osmiae and of the Bees in general were
all traceable to a common law. It seemed to me that the arrangement in
a succession first of females and then of males did not account for
everything. There must be something more. And I was right: that
arrangement in series is only a tiny fraction of the reality, which is
remarkable in a very different way. This is what I am going to prove by
experiment.

The succession first of females and then of males is not, in fact,
invariable. Thus, the Chalicodoma, whose nests serve for two or three
generations, ALWAYS lays male eggs in the old male cells, which can be
recognized by their lesser capacity, and female eggs in the old female
cells of more spacious dimensions.

This presence of both sexes at a time, even when there are but two
cells free, one spacious and the other small, proves in the plainest
fashion that the regular distribution observed in the complete nests of
recent production is here replaced by an irregular distribution,
harmonizing with the number and holding-capacity of the chambers to be
stocked. The Mason-bee has before her, let me suppose, only five vacant
cells: two larger and three smaller. The total space at her disposal
would do for about a third of the laying. Well, in the two large cells,
she puts females; in the three small cells she puts males.

As we find the same sort of thing in all the old nests, we must needs
admit that the mother knows the sex of the eggs which