ng resides in the scales, the places occupied by the roots of
the latter being marked by rows of dots.

Hitherto we have been looking at these scales as the component parts of a
picture, like the _tesserae_ of mosaic work; but they are no less
interesting as individual objects, when viewed microscopically. To do this,
delicately rub off a little of the dust or scales with the finger; then
take a slip of glass, and pressing the finger with the adhering dust upon
it, the latter will come off and remain on the glass, which is then to be
placed under the microscope. These scales may be treated either as opaque
or transparent objects, and in both conditions display exceeding beauty,
some of these single atoms showing, by aid of the microscope, as {24} much
complexity of structure as the whole wing does to the unassisted vision.

A few of the highly varied forms they present are shown on Plate II. Figs.
23 to 38 are selected from among the commoner forms, as seen by a
comparatively low power. The small stalk-like appendage is the part by
which the scale is affixed to the wing: it may be called the root. Figs.
17, 18, 19, 20, 21, show some very remarkable forms, which are, so far as
has been ascertained, peculiar to butterflies of the _male_ sex, though the
use or reason of this masculine badge, only visible to highly magnifying
optics, is neither known nor probably to be known at present; but
singularly beautiful and curious they are to look at. The little balls at
the end of threads are the root portion, and fit into cup-like sockets,
placed here and there among the ordinary scales. The surface of these
scales is beautifully ribbed and cross-ribbed, and at the upper end is a
plume-like tuft of delicate filaments. The curious scale aptly called, from
its shape, the Battledore scale, and shown at fig. 22, also belongs to the
male of various butterflies, especially those pretty little ones known as
the "Blues." Its surface is most curiously ornamented with rows of
bead-like prominences.

Probably one would imagine that in such wee specks as are these scales, one
single layer of substance would suffice for their whole thickness (if we
can talk of _thick_ness, with objects almost immeasurable in their
_thin_ness). But such is not the case, for when scales have {25} been
injured by rubbing we now and then find a part with the sculptured surfaces
torn off on each side, showing a plain central layer, so that at least
three layers--