measuring the comparative Intensities of Light . 95 
and the light made to pass through them both, the loss of 
light, in four different experiments, was ,3089 ; ,3259 ; ,3209 ; 
and ,3180 ; mean ,3184. 
With another pane of glass of the same kind, but a little 
thinner, the mean loss of light, in four experiments, was ,1813. 
With a very thin clean, pane of clear, white, or colourless 
window-glass, not ground, the loss of light, in four experi- 
ments, was ,1324; ,1218; ,1213; and ,1297; mean >1263. 
When the experiment was made with this same pane of 
glass, a very little dirty, the loss of light was more than 
doubled. 
Might not this apparatus be very usefully employed by the 
optician, to determine the degree of transparency of the glass 
he employs, and direct his choice in the provision of that 
important article in his trade ? 
In making these experiments, a great deal of the trouble may 
well be spared, for there is no use whatever in bringing the two 
lamps A and B to burn with the same degree of brilliancy ; all 
that is necessary being to bring the shadows to be of the same 
density, with the glass, and without it, noting the distance of 
the lamp B in each case (the lamp A remaining immoveable 
in its place) ; for the relative quantity of light lost will ever be 
accurately shewn by the ratio of the squares of those distances, 
whatever be the relative brilliancy with which the two lamps 
burn. The experiment is more striking, and the consequences 
drawn from it rather more obvious, when the lamps are made 
to burn with equal flames ; otherwise that equality is of no 
real advantage. 
