a fine sponge and equally distributed over its
surface. The previously ground and fined salt surface (this work is
done the same as in glass working) is now placed upon the polisher and
motion instantly set up in diametral strokes. I usually walk around
the polisher while working a surface. It is well to note that motion
must be constant, for a moment's rest is fatal to good results, for
the reason that the surface is quickly eaten away, and irregularly so,
owing to the holes that are in the pitch bed. Now comes the most
important part of this method. After a few minutes' work the moisture
will begin to evaporate quite rapidly. No new application of water is
to be made, but a careful watch must be kept upon the pitch bed, and
as the last vestige of moisture disappears the prism is to be slipped
off the polisher in a perfectly horizontal direction, and if the work
has been well done, a clean, bright, and dry surface is the result.
The surface is now tested by the well-known method of interference
from a perfect glass test plate (see Fig. 178).

"If an error of concavity presents itself the process of polishing is
gone over again, using short diametral strokes. If the error is one
of convexity, the polishing strokes are to be made along the chords,
extending over the edge of the polisher. The one essential feature of
this method is the fact that the surface is wiped dry in the final
strokes, thus getting rid of the one great difficulty of pitch
polishing, a method undoubtedly far superior to that of polishing on
broadcloth. If in the final strokes the surface is not quite cleaned
I usually breathe upon the pitch bed, and thus by condensation place
enough moisture upon it to give a few more strokes, finishing just the
same as before. In ten minutes I have polished prisms of rock salt in
this manner that have not only shown the D line double, but Professor
Langley has informed me that his assistant, Mr. Keeler (J. E.), has
seen the nickel line clearly between the D lines. This speaks for the
superiority of the surfaces over those polished on broadcloth.

"In polishing prisms I prefer to work them on top of the polisher, as
they can be easily held, but as it is difficult to hold lenses or
planes in this way without injuring the surfaces, I usually support
them in a block of soft wood, turned so as to touch only at their
edges, and work the polisher over them. Though it takes considerable
practice to succeed at