s, etc. Now, if we construct three cells (or any
other number), so faced that the three, collectively, will be acted upon
by all the colors, and arrange them around the light to be tested, at
equal distances therefrom, each cell will produce a current corresponding
to the colored rays suited to it, and all together will produce a current
corresponding to all the rays emitted by the light, no matter what the
proportions of the different colors may be. The three currents may act
upon the same index, but each should have its own coil, not only for the
sake of being able to join or to isolate their influences upon the index,
but also to avoid the resistances of the other cells. If a solid
transparent conductor of electricity could be found which could be thick
enough for practical use and yet would transmit all the rays perfectly,
i.e., transmit white light unchanged, that would be still better. I have
not yet found a satisfactory conductor of that kind, but I think the plan
stated will answer the same purpose. This portion of my system I have not
practically tested, but it appears to me to give good promise of removing
the color stumbling-block, which has so long defied all efforts to remove
it, and I therefore offer it for your consideration.

_Photo-electric regulator_.--My regulator consists of a
current-generating cell arranged in front of a light, say an electric
lamp, whose light represents the varying strength of the current which
supports it. The current produced in the cell by this light flows through
an electro-magnetic apparatus by means of which mechanical movement is
produced, and this motion is utilized for changing resistances, actuating
a valve, rotating brushes, moving switches, levers, or other devices.
This has been constructed on a small scale, and operates well, and I
think it is destined to be largely used, as a most sensitive, simple, and
perfect regulator for currents, lights, dynamos, motors, etc., etc.,
whether large or small.

In conclusion, I would say that the investigation of the physical
properties of selenium still offers a rare opportunity for making very
important discoveries. But candor compels me to add that whoever
undertakes the work will find it neither an easy nor a short one. My own
experience would enable me to describe to you scores of curious
experiments and still more curious and suggestive results, but lack of
time prevents my giving more than this very incomplete outline of my