ests of the Chalicodoma of the Sheds. I therefore thought
out a scheme for making the Three-horned Osmia accept my study as her
settlement and build her nest in glass tubes, through which I could
easily watch the progress. To these crystal galleries, which might well
inspire a certain distrust, were to be added more natural retreats:
reeds of every length and thickness and disused Chalicodoma-nests taken
from among the biggest and the smallest. A scheme like this sounds mad.
I admit it, while mentioning that perhaps none ever succeeded so well
with me. We shall see as much presently.

My method is extremely simple. All I ask is that the birth of my
insects, that is to say, their first seeing the light, their emerging
from the cocoon, should take place on the spot where I propose to make
them settle. Here there must be retreats of no matter what nature, but
of a shape similar to that in which the Osmia delights. The first
impressions of sight, which are the most long-lived of any, shall bring
back my insects to the place of their birth. And not only will the
Osmiae return, through the always open windows, but they will also
nidify on the natal spot, if they find something like the necessary
conditions.

And so, all through the winter, I collect Osmia-cocoons picked up in
the nests of the Mason-bee of the Sheds; I go to Carpentras to glean a
more plentiful supply in the nests of the Anthophora. I spread out my
stock in a large open box on a table which receives a bright diffused
light but not the direct rays of the sun. The table stands between two
windows facing south and overlooking the garden. When the moment of
hatching comes, those two windows will always remain open to give the
swarm entire liberty to go in and out as it pleases. The glass tubes
and reed-stumps are laid here and there, in fine disorder, close to the
heaps of cocoons and all in a horizontal position, for the Osmia will
have nothing to do with upright reeds. Although such a precaution is
not indispensable, I take care to place some cocoons in each cylinder.
The hatching of some of the Osmiae will therefore take place under
cover of the galleries destined to be the building-yard later; and the
site will be all the more deeply impressed on their memory. When I have
made these comprehensive arrangements, there is nothing more to be
done; and I wait patiently for the building-season to open.

My Osmiae leave their cocoons in the second half of April. Un