ated by the following formula.

the weight of water in the vessel.

.....

Putting w

g

n = that of the vessel itself.

that of the glass in the immersed part of the thermometer.

m = that of the liquid in the same.

x=

y=

z=

the specific heat of container.

do. do.

of glass.

of the thermometric liquid.

v = the weight of vapour condensed.

T= the temperature gained by the water.

t = the distance of the final temperature of the water below that at which

the steam enters it,

and = the latent heat of vapour at the boiling point.

Results of Experiments on Latent Heat.

The accompanying table presents the determination, in the manner already described, of the latent heat of the vapour of water. The trials were made in four different cylindrical vessels, one of copper, two of glass, and one of sheet iron.

The quantities of water varied from about 13000 to upwards of 39000 grains.

The equivalents of the thermometers were either approximately estimated by knowing the size and thickness of the bulbs, or were actually determined by weighing before and after filling, and in every instance the calculation for the thermometrical equivalent, was made only on the part of the instru ment actually immersed.

It will be perceived that three out of the four vessels, give mean results which differ from each other by not more than 3 degrees. The third set, or that made in thin glass, and which differs widely from all the rest, ought probably to be rejected. If this be done, the other three sets give a mean result equal to 1037 degrees; which is 3.8 degrees less than that obtained by Count Rumford. Including the third set, the mean result will be 1026.83. As the steam rising up in the case L, necessarily came in contact with the hot iron S, it became, to a certain extent, surcharged with heat; but as the thermometer indicated its temperature at the moment of escape, an allowance is easily made for the surcharge. The rapidity of flow being duly regulated by the stop-cock k, the steam was prevented from carrying over any water in an unvaporized state. As the amount of surcharge seldom exceeded 3 degrees, it was not considered necessary to calculate for the difference between the specific heat of vapour and that of water. By the experiments of Delaroche and Bérard, the specific heat of vapour, compared with that of water, is .847 to 1.000. Admitting this to be true, the result must, in any case which has occurred to the committee, be but little affected by allowing for the difference.*

=

Then the heating effect is represented by Tx (w + nx + gy+mz), and the cooling effect by v × (1+t); whence v × (1+1) · Tx (w+nx+gy+mz) and consequently =. T (w+ nx + gy+mz)

• The experiments hitherto published, had left some doubt as to the true latent heat of vapour. Black first obtained the number 810°; Watt afterwards gave it 950°; Southern produced 945; Lavoisier made it rather more than 1000°; Rumford 1010.8; Despretz 955.8; Ure 1000; Thompson "more than 1000°." Watt and Clement have both established the position that the latent heat of steam, added to the sensible heat above 32°, is nearly a constant quantity. As, however, the point 32° is entirely arbitrary, and as no temperature is now known, at which vapour does not rise from water or ice, there is reason to suppose that in strictness, the constant-if there be one-is different from that which these experimenters have derived. If not, the latent heat of vapour must diminish below 32° as the temperature diminishes.

TABLE XII.

Experiments to determine the latent heat of Steam, employing a given weight of water in a vessel of known weight and specific

Sheat, to receive and condense the vapour as it passed from the mouth of the pipe.

Specific Heat by Vaporization.

Having determined the latent heat of vapour, it is not difficult to verify our preceding determinations of the specific heat, by operating in precisely the same manner as we do to obtain the temperature of a body,-except, that the temperature of the bath of melted metal is now first ascertained by the mercurial thermometer; and the actual temperature of the standard piece being then known, is compared with the weight of vapour which it produces, by cooling in boiling water from its initial temperature down to 212°. These experiments were made both before and after the screw beam and counterpoise were changed. The weight of the standard piece is, in both cases, taken in degrees of the pyrometer scale as existing at the time.

It will be seen, that, assuming as correct the determination of latent heat, made by the committee (1037°), the experiments given in the accompanying table (No. XIII.) afford results for the specific heat of iron as follows:1. Taking the mean of 29 experiments, it is

.11325

.11340

have

2. Taking only those made before the screw beam was changed, (9 experiments,) we obtain

3. Taking together the last 20 experiments of the table, we

.11324

.11336

.11356

4. Experiments Nos. 7 and 8 with the first screw and coun-
terpoise, differing only in the 5th place of decimals, give
5. Six out of the last 20, differing only in the fourth place,
give

6. The mean result of these 5 comparisons, is

.11336

As a mean of the nine sets of experiments in different vessels, the specific heat below 212°, determined by heating water, as above detailed, was found=.113374. As the calculations just detailed are carried only to the 5th place, the two results may be considered as differing from each other only by 7th part of the total value.

Of those experiments which differ considerably from the general result, about the same number was found above, as below the mean, showing that if these discrepancies be due to errors of observation, they are, as we ought to expect, liable to be either in excess or defect; and that they counterbalance each other.

The eight experiments, of which the results differ only in the 4th place of decimals, were made at temperatures varying from 392 to 595, without indicating any decided difference in the specific heat of the metal within those limits.

Of the extreme results in the table, the highest was obtained at 480° and the lowest at 488°;-the next to the highest, at 500°, and the next to the lowest, at 292°.

From the exact conformity of the general results of the method of evaporation, and that of heating water, in trials below 212°, it appears that if the specific heat of the standard piece be determined by the latter method, and its weight be duly regulated to conform to the length of the threads of the screw beam, and to the weight of the revolving counterpoise, its indications of temperature will be such as to connect themselves immediately with those of the mercurial thermometer.