Economy of Fuel. 331 
taken off when the temperature of the entering gas was increasing, 
when it was diminishing, and when stationary. It was likewise im- 
_ portant to determine whether the proportion withdrawn when the 
excess of the temperature in the entering gas above that of the roont 
was great, were the same as when it was less; or whether, on the 
contrary, the difference in the velocity of movement, of the gas due 
to the difference of density, would sensibly increase the proportion 
of that excess which would be withdrawn at the lower tem ures. 
The area of the convex vertical sides of the drum was 3091.4 
Do. of the two ends of do. 2 - - 
Do. of the pipe ¢ from the drum to the thermometer ¢”, 1234.6 
er Total area, 5249.6 
Equal to about 37.8 square feet. 
- Having made this arrangement, I found the temperature of the apart- 
ment which had, in former years, required a separate grate, or stove, 
to keep it in a comfortable condition, entirely freed from that neces- 
ity, and during the whole season, which will long be remembered 
as one of uncommon severity, not a single hour is known to have 
found it untenantable from cold. At night, the fire in the kitchen 
was prepared for a slow operation, by adding a fresh supply of coal 
covered closely by a layer of the finer kind called “ chestnut coal,” 
or, what is still better, the coarser parts of the sifted cinders from 
which the earthy and vitrified portions were always carefully rejected. 
In this state of the fire, the temperature of the entering gas was con- 
siderably reduced, but in no instance was it found lower than 100°. 
It is probable that the mass of brick work, constituting the chimney 
having become hot during the day, contributed to keep up the tem- 
perature inthe drum during the night,—a contribution which would - 
haye been utterly wasted on the “ upper air” according to the usual 
method of arranging both wood fires and open grates. == 
- In the following table are given the results of numerous observa- 
tions made, some at irregular intervals during the winter, others at 
regular periods of five or ten minutes apart, coutinued for several 
hours in immediate succession. They are arranged according to 
the temperature of the entering gas which, it will be observed, was 
never higher at the time of any observation, than Tic. “ie chan- 
ges were, in general, so gradual as to allow a perfect facility in noting 
