legs, that when, during the
following summer, another officer described to me a great spider which
was very common on Long Island, where he was stationed, I knew it was
the same, and told him what I had done the year before, adding that I
was sure something would come of it in time.

With leisure and many spiders at his command, this officer improved upon
my suggestion, by substituting for my quill turned in the fingers a
wooden cylinder worked by a crank, and by securing, at a proper
distance, (between pins, I think,) one or more spiders, whose threads
were guided between pins upon the cylinder. He thus produced more of the
silk, winding it upon rings of hard rubber so as to make very pretty
ornaments. With this simple machine I wound the silk in two grooves cut
on a ring of hard rubber and parallel except at one point, where they
crossed so as to form a kind of signet. Another officer now suggested
and put in operation still another improvement, in the shape of the
"gear-drill-stock" of our armorer's chest. This, being a machine for
drilling iron, was rough in its construction and uneven in its action,
but, having cog-wheels, a rapid and nearly steady motion could be given
to its shaft. To this shaft he attached a little cross of rubber, and
covered it with silk, which was of a silver-white color instead of
golden-yellow, as in other cases. The difference in color was then
supposed to depend upon individual peculiarities, but the true
explanation will be given farther on. With this gear-drill-stock, upon a
larger ring, one inch in diameter and three eighths of an inch in width,
in a groove upon its periphery one fourth of an inch in width, and
across the sides of the ring in two directions, I wound _three thousand
four hundred and eighty-four yards_, or _nearly two miles, of silk_. The
length was estimated by accurately determining the different dimensions
of the ring where wound upon, and multiplying by this the number of
revolutions of the cylinder per minute (170), and this product again by
the number of minutes of actual winding (285), deducting from the gross
time of winding (about nine hours) each moment of stoppage for any
cause.

This was late in the fall of 1864, and, our specimens being sent home,
further experiments, and even thoughts upon the subject, were prevented
by the expedition against the Charleston and Savannah Railroad, and the
many changes of station that followed the disastrous battle of Honey