196 
mites in such a manner as to give constantly 400° of light, 
and it burnt 12.7 grammes of oil; that is, in the ratio of 361 
parts of oil per hour, for 400° of light; or for 100 parts of 
oil, 112° of light. 
This is the first experiment where a given quantity of oil 
has furnished more light than an equal given quantity of 
wax; but without stopping to make any observations on this 
fact, I pass on to more important results ; and to save time, 
I shall exhibit in a tabular form, the results of the four ex¬ 
periments, of which I have given some account, and shall 
add to them the results of five others, which, as well as the 
four preceding, were all made the same day, with the same 
lamp, the same candle, and the same care. 
Order of Ex¬ 
periments. 
Intensity of the light 
furnished by the lamp 
during 30 minutes. 
Quantity of oil 
consumed per 
hour. 
Light furnished 
per hour,from the 
consumption of 
100 parts of oil. 
I 
100° 
228 parts 
48° 
2 
200 
271 
74 
! 3 
300 
305 
93 
4 
400 
361 
112 
5 
500 
405 
121 
6 
600 
441 
138 
7 
700 
470 
149 
8 
800 
515 
155 
9 
900 
560 
160 
On comparing the results of these nine experiments, we 
find that the quantities of light furnished, were far from 
being in a constant proportion to the quantities of oil con¬ 
sumed, as should have been the case, if light was one of 
the chemical products of combustion. The quantity of 
light afforded by 100 parts of oil, was almost four times as 
much in the ninth experiment as in the first, though the 
flame was equally clear in both, as indeed it was in all the 
other experiments, and had neither manifested smoke or 
odour. 
As it appeared very probable to me, that a small flame 
of a particular form, always yielded less light in proportion 
