state in which those of some of the soldiers of Napoleon's old guard,
that had been committed unstripped to the earth, may be dug up in the
future on the fatal field of Borodino, or along the banks of the Dwina
or the Wap. The cuirass lies still attached to the helmet, but we find
only the naked skeleton attached to the cuirass. The Pterichthys to its
strong helmet and cuirass added a posterior armature of comparatively
feeble scales, as if, while its upper parts were shielded with plate
armor, a lighter covering of ring or scale armor sufficed for the less
vital parts beneath. In the Asterolepis the arrangement was somewhat
similar, save that the plated cuirass was wanting: it was a strongly
helmed warrior in slight scale armor; for the disproportion between the
strength of the plated head-piece and that of the scaly coat was still
greater than in the Pterichthys. The occipital star-covered plates are,
in some of the larger specimens, fully three-quarters of an inch in
thickness, whereas the thickness of the delicately-fretted scales rarely
exceeds a line.

Why this disproportion between the strength of the armature in different
parts of the same fish should have obtained, as in Pterichthys and
Asterolepis, or why, while one portion of the animal was strongly armed,
another portion should have been left, as in Coccosteus, wholly exposed,
cannot of course be determined by the mere geologist. His rocks present
him with but the fact of the disproportion, without accounting for it.
But the natural history of existing fish, in which, as in the Pimelodi,
there may be detected a similar peculiarity of armature, may perhaps
throw some light on the mystery. In Hamilton's "Fishes of the Ganges" I
find but little reference made to the instincts and habits of the
animals described: their deep-river haunts lie, in many cases, beyond
the reach of observation; and of the observations actually made, the
descriptive naturalist, intent often on mere peculiarities of structure,
is not unfrequently too careless. Hamilton describes the habitats of the
various Indian species of Pimelodi, whether brackish estuaries, ponds,
or rivers, but not their characteristic instincts. Of the Silurus,
however, a genus of the same great family, I read elsewhere that some of
the species, such as the _Silurus glanis_, being unwieldy in their
motions, do not pursue their prey, which consists of small fishes, but
lie concealed among the mud, and seize on the