e in such a manner that other persons using instruments
similarly constructed would be able to obtain the same results.

Perhaps the most important tests needed are in regard to the
sensitiveness of the plates. Most plate makers use the wet plates as
their standard, giving the sensitiveness of the dry plates at from two to
sixty times greater; but as wet plates vary quite as much as dry ones,
depending on the collodion, condition of the bath, etc., this system is
very unsatisfactory. Another method, employed largely in England, depends
on the use of the Warnerke sensitometer. In this instrument the light
from a tablet coated with luminous paint just after being exposed to a
magnesium light is permitted to shine through a colored transparent film
of graduated density upon the plate to be tested. Each degree on the film
has a number, and, after a given exposure, the last number photographed
on the plate represents the sensitiveness on an empirical scale. There
are two or three objections to this instrument. In the first place, the
light-giving power of the luminous tablet is liable to variations, and,
if left in a warm, moist place, it rapidly deteriorates. Again, it has
been shown by Captain Abney that plates sensitized by iodides, bromides,
and chlorides, which may be equally sensitive to white light, are not
equally affected by the light emitted by the paint; the bromides being
the most rapidly darkened, the chlorides next, and the iodides least of
all. The instrument is therefore applicable only to testing plates
sensitized with the same salts.

In this investigation it was first shown that the plates most sensitive
for one colored light were not necessarily the most so for light of
another color. Therefore it was evident that the sun must be used as the
ultimate source of light, and it was concluded to employ the light
reflected from the sky near the zenith as the direct source. But as this
would vary in brilliancy from day to day, it was necessary to use some
method which would avoid the employment of an absolute standard of light.
It is evident that we may escape the use of this troublesome standard, if
we can obtain some material which has a perfectly uniform sensitiveness;
for we may then state the sensitiveness of our plates in terms of this
substance, regardless of the brilliancy of our source. The first material
tried was white filter paper, salted and sensitized in a standard
solution of silver nitrate. This