lining of the hind-gut, for the pupa to be
supported while it jerks its cremaster out of the larval cuticle and
works it into the meshes of the silken pad. The moult is thus completed
and the pupa hangs securely all the time. In the numerous cases where
the pupa is enclosed in a cocoon, the cremaster serves to fix the pupa
to the surrounding silk. Chapman (1893) has drawn attention to the fact
that among the more highly organised moths the pupa remains in the
cocoon, the emergence being entirely left to the imago, while the pupae
of the more primitive moths work their way partly out of the cocoon
before the final moult begins. In the latter case, the cremaster is
anchored by a strand of silk which allows a certain degree of emergence,
and the pupa has rows of spines on its abdominal segments, of which a
greater number retain the power of mutual motion than in those pupae
which do not come out of their cocoons.

[Illustration: Fig. 23. Pupa of White Butterfly (_Pieris_), side view;
_f_, feeler; _w_, wing; _sp_, spiracle; _p_, anal pro-leg; _cr_,
cremaster. Magnified 8 times. In part after Hatchett-Jackson, _Trans.
Linn. Soc._ 1900, and Tutt's _British Butterflies_.]

While the pupa on the whole resembles the imago that is to emerge from
it, there are not a few cases in which a special structure necessary for
some contingency in pupal life is retained or adopted in this stage. A
butterfly pupa, like the imago, has no mandibles, but in the case of the
Caddis-flies (Trichoptera) and two families of small moths, the most
primitive of all Lepidoptera, the pupa, like the larva, has
well-developed mandibles. These enable the caddis pupa to bite its way
out of the shortened larval case in which it has pupated, and then to
swim upwards through the water ready for the caddis-fly's emergence into
the air. Pupae that are submerged require special breathing-organs. In
the previous chapter (p. 77) mention was made of the gnat's aquatic
larva with its tail-spiracles adapted for procuring atmospheric air
through the surface-film. The pupa of the gnat[10] also has 'respiratory
trumpets' serving the same purpose, but these are a pair of processes on
the prothorax, so that the pupa, which is fairly active, hangs from the
surface-film with its abdomen pointing downwards through the water. This
change of position is correlated with the necessity for the imago to
emerge into the air; were the pupa to hang head downwards as the larva
does, t