424 
CAPTAIN ABNEY AND MAJOR-GENERAL FESTING 
(‘Paris, Comptes Rendus/ vol. 90 (1880), p. 252) has applied this plan to a spectrum 
photometer, and gave results which do not quite agree with ours. A great many 
measures of the solar spectrum were made by this plan, but it was felt to be 
not as uniformly accurate as could be desired, and to be very fatiguing to the 
observer. In fact, though feasible in making experiments with our own eyes, it was 
a plan which we could hardly expect others, whose eyes we might wish to test, to 
try, seeing that their interest in the results might hardly he even personal and not 
general. One cause of inaccuracy in this method is the liability of the eyes to 
astigmatism—a difficulty which one of us had to encounter, and found very hard to 
overcome, the slightest difference in the inclination of the axis of the eyes rendering 
readings discordant. By this method it would be necessary, therefore, that any 
casual observer who might he called in should first have undergone an examina¬ 
tion for this defect before any idea could be formed of the value of his readings. 
Another drawback to it is the fact that from the blue to the violet the accom¬ 
modation of the eyes is insufficient to permit of the lines being seen sharply, the 
black lines invariably appearing to both of us hazy and covered with a blue mist. 
This latter difficulty was, however, partially avoided by the use of a small observing 
telescope, the focus being altered to suit each ray. But the trouble and uncertainty 
of this plan caused us to seek for some other method which would be more generally 
useful, and more easily applicable to others as well as to ourselves. 
Our next plan was to cut an aperture in a mask of black paper fastened to ground 
glass, which was made of such a size that the image of it at a fixed distance when 
thrown on the screen should be of the same size and shape as the coloured patch of 
light. Behind this ground glass was placed a candle or incandescence lamp, which 
illuminated it, and thus a patch of yellowish light was thrown on the screen in juxta¬ 
position to the patch whose intensity was to he determined. By moving the light 
towards or away from the translucent screen the patch of light became brighter or 
dimmer, and when the two patches appeared of equal luminosity the distance of the 
candle from the ground glass was noted, the inverse square of this distance being a 
measure of the brightness of the patch. This method gave fairly concordant results; 
but it occurred to us that the close juxtaposition of small patches lighted from different 
sources, as in the Rumford system of photometry, would be a more convenient method, 
and this was accordingly tried. 
The light was focussed on the slit, S, of the collimator, C, fig. 1, by a condenser filling 
the collimating lens, and after passing through two flint prisms, P, of medium density, 
each of an angle of 62°, fell on the lens, L, of a, camera, B, which brought the spectrum 
to a focus on the ground-glass screen. As before described in the paper quoted, a 
sliding slit, S', formed in a card, A, was substituted for the dark slide, and the patch 
of light was formed by being collected by L on a white card, D, placed about 4 feet 
from the sliding slit.* In front of the card was placed an upright rod, E, about half 
an inch in diameter, to intercept the light from the two sources. A slide, II, carrying 
