17 
irrespective of pressure. Experiments 4 and 5 show an in- 
creased consumption of 75 per cent. under an increased 
pressure of three atmospheres. 
The chief point of interest brought to view, is the close 
correspondence between the relative consumption and the 
square roots of the pressures. The series of tests was not 
really extensive enough to establish a definite law, but the 
results thus far are of much interest, as pointing to this 
principle. 
Much may depend, however, on the capillary attraction of 
the wick, and therefore upon its structure. It is probably 
due to this cause that, in one set of experiments, conducted 
by Prof. Chas. A. Seely, standard candles, with small wicks, 
increased only 5 per cent. in consumption under a pressure of 
46 Ibs. On the other hand, alcohol, far lighter than the 
melted hydrocarbon fats, and supplied through a very open 
wick, gave an increase of consumption at the same pressure, 
of 200 per cent., or directly in proportion to the pressure. 
The alcohol flame under such circumstances, changes its 
aspect completely, becoming white and luminous, so as to 
give three-fourths the light of a “coach” candle. 
Gas from the street-pipes was finally imtroduced, and 
employed for lighting the caisson. Here again, increased 
energy of combustion was noticed speedily, and much trouble 
was also found from smoking of the gas under advancing 
pressures. Six-foot “fish-tail” burners had to be abandoned 
for two-foot “bat-wing” ones, at 20 lbs. pressure; and at 
28 lbs. these latter had to be replaced by one-foot burners of 
the same pattern. These were used until the close of the 
work, with fair results. 
The paper concludes with a discussion of the theoretical 
conditions of this increased combustion. Conceiving a flame 
as the oxidation of a distilled hydrocarbon, its volume will 
be determined as follows :—the vapor or gas will expand 
until combustion goes on over its surface at precisely the 
same rate as the supply goes on from below. In the case of 
a candle, the supply depends on the heat of the flame, i. e. 
ge oe oer of combustion. Compressed air allows a 
