150 0. N. Rood on Photometric Experiments. 
two branches of the supply abe in point of fact, on giving 
them equal lengths on c 
tions slowly, so that the tien Seren, to L was not me 
shaken or set in oscillation, this source of error a cam 
so. 
greatly reduced as in no way to interfere with the obeorvea 
Mode. of measuring the amount of light reflected from a mir- 
ror.—To adjust the apparatus, it is first necessary that that = 
ameter of the circle joming the 0° and 180° points, should 
made parallel to the axis of the instrument, which is effected 
RY the use of the vernier at X, fig. 2, the circle is then clamped. 
ext, the vernier attached to the mirror is placed at 90°, and 
the mirror itself is made to assume a position at right angles 
to the axis of the instrument, by the aid of its three screws, 
and a small gas flame placed on a support which rests on the 
parallel bars at a distance of two or three feet from the mirror, 
it being so contrived that the center of this small flame shall 
just lie in the axis of the instrument. The mirror is adjuae 
till it reflects the light of this small flame back to the e 
through a flame itself, securing thus the collimation with the 
desired ac aes The arm A is then set at ei? desired angle 
ing n aed behind the flame so as to be out of the 
way ad to provid light from reaching the mirror. 
Mode of registering the observations.—These were always reg: 
istered with a sharp point on a angel fillet of paper, attached 
to — long wooden rod_R T, fig. 9, used for moving the mirror. 
int was one end of a small lever placed at N, fig. 1 
In consequence of this, at the end of a set of experiments tw° 
groups of dots were found on the paper, admitting of the most 
exact measurement on the following day. Before removing 
the pa p aper from the rod two dots were always made on it in the 
borhood of these groups, the corresponding positions i 
mill 
detached fle kee and the observations recorded in the 
k. In determinations given below, the observa 
tions on the direct and reflected light w were made alternate, 50 
= 
ng of the compensation point 
