Drs Christison and Turner on the Construction of 
tion, it was surrounded by a glass-tube graduated to lOths of an inch. As a 
three-inch jet of good coal-gas burnt in this instrument about two hours and a 
third, and one of oil-gas about four hours and a quarter, we could easily make, 
with one charge of the gasometer, from six to twelve comparative experiments ; 
and we could also compare with each other those made on different days, pro- 
vided the gas was tolerably uniform in specific gravity. 
Another gasometer, in every respect the same as that now described, was 
used for supplying gas to the moveable light, whose power was to be ascer- 
tained. 
In order to measure the consumption of the gas, a graduated glass tube 
was attached to the side of each gasometer, communicating above and below 
with its cavity. The degrees were 50ths of a cubic foot, measured with the 
greatest care ; and in our observations the time was generally counted for 4, 
6, or 10 degrees. The measurements were so exact, even when the smallest 
of these quantities was observed, that, on afterwards determining the expen- 
diture of various burners on the great scale, we had not to alter any of our 
former results. 
It is unnecessary to add any thing farther to the proofs formerly given of 
the extreme accuracy ensured by this method of operating — Some may sug- 
gest, however, that what has just been said of the inadequacy of candles to 
form a standard of comparison, is at variance with one of the facts regarding 
their light formerly and now again appealed to in proof of the accuracy of this 
mode of experimenting. In answer, we have to observe, that, independently 
of the exact agreement between the two observations with the candle being per- 
haps to a certain degree accidental, the mode in which they were made is quite 
incompatible with the idea of a standard. The candle, after burning some time, 
was carefully snuffed ; its light in relation to that of a gas-jet was repeatedly 
compared, at intervals of a minute, till it began to fade ; and the distance was 
taken at the maximum. And to show how impossible it is to make use of a 
candle for the standard, we need only add, that, however carefully the snuf- 
fing was performed, the light, in the experiment now mentioned, increased 
after it was apparently at its brightest, in the ratio of five to six. To con- 
clude ; the accuracy of our method has been shown to our satisfaction, not only 
by the frequency of such coincidences as we have noticed, but likewise in a 
less fallacious manner, by verifying by direct observation (as in trigonometri- 
cal surveys), results deduced through a connected series of mixed observation 
and calculation. 
II. 
We shall now proceed, in the second place, to detail our experiments rela- 
tive to the circumstances which affect the degree of light emitted by the gases 
during their combustion. The consideration of these circumstances will lead 
to the discovery of the principles on which burners ought to be constructed. 
The object held in view, so far as we know, universally in burning the 
gases for the purpose of illumination, has been to render the combustion as 
vivid as possible. This principle has been followed most probably, because it 
secures the complete combustion of the gas, and because, by increasing the vi- 
