like the Leptus, with its mandibulated mouth and jointed legs,
seems at first well nigh impossible. We have the faintest possible
indication in the structure of some mites, and of the Tardigrades and
Pentastoma, where there is a striking recurrence, as we may term it, to
a worm-like form, readily noticed by every observer, whatever his
opinion may be on the developmental theory. In the Demodex we see a
tendency of the mite to assume under peculiar circumstances an
elongated, worm-like form. The mouth-parts are aborted (though from what
we know of the embryology of other mites, they probably are indicated
early in embryonic life), while the eight legs are not jointed, and form
simple tubercles. In the Tardigrades, a long step lower, we have
unjointed fleshy legs armed with from two to four claws, but the
mouth-parts are essentially mite in character. A decided worm feature is
the fact that they are hermaphrodites, each individual having ovaries
and spermaries, as is the case with many worms.

[Illustration: 191. Nauplius.]

When we come to the singular creatures of which Pentastoma and
Linguatula are the type, we have the most striking approximation to the
worms in external form, but these are induced evidently by their
parasitic mode of life. They lose the rudimentary jointed limbs which
some (Linguatula especially) have well marked in the embryo, and from
being oval, rudely mite-like in form, they elongate, and only the claws
or simple curved hooks, like those of young tape worms, remain to
indicate the original presence of true jointed legs.

In seeking for the ancestry of our hypothetical Leptus among the worms,
we are at best groping in the dark. We know of no ancestral form among
the true Annelides, nor is it probable that it was derived from the
intestinal worms. The only worm below the true Annelides that suggests
any remote analogy to the insects is the singular and rare Peripatus,
which lives on land in warm climates. Its body, not divided into rings,
is provided with about thirty pairs of fleshy tubercles, each ending in
two strong claws, and the head is adorned with a pair of fleshy
tubercles. It is remotely possible that some Silurian land worm, if any
such existed, allied to our living Peripatus, may have been the ancestor
of a series of types now lost which resulted in an animal resembling the
Leptus.

[Illustration: 192. Platygaster error.]

We may, however, as bearing upon this difficult question, c