92.-According to the analysis of Dr. John Murray 10,000 parts sea water of the specific gravity 1·029+ contain

Or 1 part of sea water contains '030309 parts of salts = of its weight. 93.-Now as the salts do not rise with the steam, the water in a boiler supplied with sea water becomes gradually more saturated, and after a certain time begins to deposit salt, if the means that have been invented for that purpose be not employed to prevent it. (See Sect. III.) And even then a certain degree of a saturation must be allowed to take place. The following table, with the constant numbers for different degrees of saturation will serve to illustrate this matter. The boiling point of water appears to be increased one degree by each addition of 26 parts to the proportion of common salt in 100 parts of water; at least so nearly that this regular law does not materially differ from the mean results of my experiments, which were made with a considerable degree of care. But it is difficult to make them on account of the degree of saturation constantly varying during the experiment.

* Quarterly Journal of Science, Vol. XVIII. p. 90.

+ Thomson's Chemistry, Vol. II. p. 14.

+ Phil. Mag.

94.-The next point is to compare the formula with experiment, and we will commence with Mr. Watt's experiments on salt water. The water was nearly saturated with salt. It was more free from air than common water, but it parted with difficulty from that it contained. The results compared with the formula for saturated salt are shewn in the following table.

In these as in all the early experiments on the force of steam, the force is less than it ought to be at low temperatures.

Mr. Watt's experiments on pure water afford a like discrepancy, as will be found by comparing the following table of results taken at random out of his series.*

The explanation offered by Mr. Watt himself is not sufficient to account for the difference except in the lower temperatures.

He supposes the stationary barometer must have had its scale placed 2 of an inch too low; and if so, the same addition would be required to the forces in the preceding table on salt water. These tables, however, are not selected for minute accuracy, but to shew the important fact that the force of the steam of water depends on the temperature of the liquid which produces it, or which is in contact with it. For this they are sufficiently correct, and it is a circumstance which affects its elastic force both in the boiler and in the condenser ; and is peculiarly interesting to those concerned in steam vessel engines. The temperatures not being the same the comparison is not so easy, but at 180° the force of salt water is 10.85; that of pure water 14-73 inches: at 212° salt water has a force of 22.74; pure water 29.56.

95.-The experiments made by Professor Robison were tried in a similar manner; and

Robison's Mechan. Phil. Vol. II. p. 32–34.

FIG. 10.

as a method the same in effect was used by Bettancourt, whose results agree extremely well with Robison's, the description of it may be useful. Professor Robison's apparatus for determining the force of steam.-This apparatus, in the first trials, consisted of a small digester of copper, A B C D, in the figure; the top of which had a thermometer inserted through the centre, and a loaded valve at V; with a third hole for inserting a barometer tube S G F, to ascertain the force at lower temperatures than 212°. The force at the higher temperatures was measured by the steelyard on the valve, and a plug was inserted in the place of the tube S G F, but the results with the valve were irregular and not satisfactory. Hence, the glass tube M NK, having a cistern L for mercury, was adapted to the hole in the digester, and instead of measuring the force by the valve, it was measured by the ascent of the mercury in the tube M N. The digester was heated by a lamp.

M

To determine the pressure at temperatures below 212°, the tube SGF was inserted as in the figure, and a basin of mercury provided at F. The lamp being applied, the water in the digester produced steam till it issued at both the valve and the pipe F, so as to expel the air; the lamp being removed, and both the valve and tube being closed, the latter by immersing it in the mercury, the mercury rose in the tube F G as the apparatus cooled, and the heights corresponding to different temperatures were noted; like observations were made as it reheated.

To determine the pressure at higher temperatures with the apparatus, the end K of the tube M N K was inserted at E, and as the temperature increased, the pressure of the steam in the cistern L caused the mercury to ascend, and consequently afforded a means of measuring the force of the steam.

The objection to this mode of trial is that the temperature of the mercury must be continually changing during the trial, and steam must either be condensing or generating on its surface during the time of observation. At each observation the temperature of the whole of the apparatus ought to be the same, and then the column exhibiting the pressure ought to be reduced to its equivalent at the mean temperature. The only observation where these circumstances would have place was that which appears to have been made when the thermometer was at 42°; then the column in the syphon was 29-7, and the barometer stood at

29-84: the difference is the force of steam at 42°, and is 0.14 inches. By cooling down to 32o the force was not perceptibly different, and we know from later trials that this is nearly correct. Professor Robison, however, seems to have thought it was necessary to have the force 0 at 32°.*

If the elastic force 14 from which Robison began to register had been added to all the experiments below 212°, as it ought to have been, they would have agreed extremely near with the results of later experiments. The experiments made by Achard seldom vary more than a degree or two from those in the above table.

96.—Mr. Dalton's inquiries were conducted by a different method. He took a barometer tube, made perfectly dry, and filled it with mercury just boiled, marking the place where it was stationary; then graduated the tube into inches and tenths by means

Mechan. Phil. Vol. II. p. 36.

K