crumbles to pieces. It is slow in coming away; but it comes away
all the same, atom by atom. At long intervals, the method
changes. With its crown of awls driven into the pith, the animal
frets and fidgets, sways on the pivot of its anal armor. The work
of the auger follows that of the pickaxe. Then the blows
recommence, interspersed with periods of rest to recover from the
fatigue. At last, the hole is made. The pupa slips into it, but
does not pass through entirely: the head and thorax appear
outside; the abdomen remains held in the gallery.

The glass cell, with its lack of supports at the side, has
certainly perplexed my subject, which does not seem to have made
use of all its methods. The hole through the sorghum is wide and
irregular; it is a clumsy breach and not a gallery. When made
through the mason bee's walls, it is cylindrical, fairly neat and
exactly of the animal's diameter. So I hope that, under natural
conditions, the pupa does not give quite so many blows with the
pickaxe and prefers to work with the drill.

Narrowness and evenness in the exit tunnel are necessary to it.
It always remains half caught in it and even pretty securely fixed
by the graters on its back. Only the head and thorax emerge into
the outer air. This is a last precaution for the final
deliverance. A fixed support is, in fact, indispensable to the
Anthrax for issuing from her horny sheath, unfurling her great
wings and extricating her slender legs from their scabbards. All
this very delicate work would be endangered by any lack of
steadiness.

The pupa, therefore, remains fixed by the graters of its back in