of any metal having an atomic weight differing by more than 5 from that of sodium could be present, supposing it to be a chloride. On converting the chloride into a sulphate an equally accordant result was got. This convinced me that I had a perfectly pure substance.

§3. Preparation of the Brine.-The clearest lumps of the rock-salt were washed in common and then in distilled water. With these a large beaker was filled. Distilled water was poured on so as to fill the beaker about two thirds. If part of the salt is thus always above the surface of the brine, the latter appears to be saturated in twenty-four hours. After forty-eight hours the brine was poured into another beaker and allowed to stand until crystals began to be formed. It was then transferred to a stoppered bottle and was ready for use. Although Regnault and others have shown that for temperatures above 0° C. the solubility of chloride of sodium in water is constant, yet the precaution was taken of suspending a muslin bag containing some crystals of salt from a string passing through the cork of the bottle so that the crystals touched the surface of the brine. By this means the perfect saturation of the brine was assured. At 18° C. this solution contained 26-2724 per cent. of salt. The solution was weighed in a long-necked flask, and was evaporated to dryness and heated to about 300° C.

The following numbers show the results obtained by some other experimenters :

per cent. Sat. at 25 C. contains 26.32

spec. grav.

18.75

26.72

1.2046

The above specific gravities are referred to water at the same temperature as the brine, as far as my own experiments are concerned.

4. Cold produced on mixing saturated Brine with Water.-To examine this point, a series of covered beaker-glasses containing known weights of water, and a similar series containing saturated brine, were placed side by side in a tin tray containing water and resting on flannel. The whole arrangement was covered with flannel and allowed to stand over night. In the morning the temperature of eight or ten was found to be exactly 21° C. The liquids were prepared by measuring out from a burette in proportions that, taking 1·1997 as the specific gravity of the brine, the ratios by weight were as in columns 3 and 4 (Table I.). The experiments were performed by taking out a water-beaker, rapidly drying its outside, pouring in the brine, stirring with the thermometer, and at once reading.

TABLE I.

Cold produced on mixing saturated Brine with various proportions of Water by weight. Initial temperature 21° C.

The greatest depression of temperature accordingly in the range of ratio between 84 and 90 of water and 16 and 10 of salt, and the greatest depression of temperature to be got by mixing saturated brine with water, is under no circumstance quite 1° C.

§ 5. If now anhydrous salt on dissolving in water absorbs heat only by reason of its conversion into a liquid, and if no hydrate of salt exists at the ordinary temperature above 0° €., no such absorption of heat as we have seen to occur could ensue on mixing brine and water. Two series of experiments are therefore at once demanded the first to see what proportion must exist between anhydrous salt and water to absorb the most heat; the second to see if at any ratio between salt and water an abnormal change of density takes place.

§ 6. Cold produced on mixing NaCl with Water.-A quantity of rock-salt having been finely powdered, was passed through muslin so as to be quite mealy. Various quantities of distilled water were weighed into beakers, and various quantities of the rock-salt were thrown in under constant stirring. The abasement of temperature being noted gave the results of Table II. The temperature of the salt was about 2° C. above that of the water.

And it is seen that the greatest abasement of temperature occurs when there is about 40-60 of salt to 160-140 of water (say 25 per cent. of salt). This result is what we should expect, remembering that the saturated solution contains 26-27 per cent. of salt.

TABLE II.-Maximum Cold produced on mixing Salt and Water in different proportions by weight.

§7. Specific Gravity of Brines of various strengths.—The question as to the source of cold when brines are diluted can be conveniently approached from an examination of the density of brines of different strengths, and the comparison between the observed specific gravities and the theoretical specific gravities, supposing no change of volume to take place. The brine was found to have a specific gravity of 1.2011 at 26° C., the temperature at which the determinations were made. In Table III. the columns 1 and 2 give the weights in grams of the quantities of water and saturated brine. Columns 3 and 4 give the percentage of the water and NaCl in the resulting brines. Column 5 gives the specific gravity found, and column 6 gives the specific gravity calculated under the hypothesis that no alteration of volume ensues.

TABLE III.-Specific Gravity of Brines of various strengths.

There is accordingly an increase of volume when a solution of salt is diluted; and this is of course connected with the absorption of heat examined in § 4, Table I. When the original temperatures are restored, a mixture of strong brine and water may have a volume two hundredths greater than the sum of the volumes of its constituents. And when the brine contains as little as 4 per cent. of salt, its specific gravity is sensibly smaller than if such a proportion had been the result of the mixture without contraction of the strongest brine and water.

§ 8. Refraction of Light by Brines of different strengths.—It seemed, from a consideration of the numbers in Table III., that at or near the point of saturation a definite hydrate of salt exists; and that this is not merely diluted, but also actually decomposed on the addition of water, so that expansion takes place. This I imagined might be tested by the change in the refractive index of the brine. Accordingly brines of various strengths were placed in a hollow prism of 60°. The refraction was measured on a goniometer (Babinet's) table provided with telescope and collimator. An alcohol-flame containing sodium was employed as a source of light shining through a very fine slit. The minimum refraction being obtained, the D line was split by the spiderthread. In Table IV. the empty prism is considered as giving a displacement of 0°. The angular displacements are alone here given, as they perfectly suffice for the detection of singular values.

These numbers show a singular value at about the 10- to 11

per-cent. solutions.

§ 9. Another method of attacking the question is offered by the examination of the boiling-points of various brines. Accordingly I took solutions made by mixing saturated brine with water; and waiting for a day on which the barometer stood nearly at the mean, I determined the boiling-points of the brines. The vessels were tall copper cylinders. In the Table V. the temperature of the steam, as well as that of the brine itself, is given. The boiling was, of course, only continued a short time, to avoid the error of strengthening the brine.

TABLE V.

Boiling-points of Brines of various strengths in copper cylinder.

We have here again a singular value about the 10- to 11-percent. solutions.

§ 10. Freezing-points of Brines of various strengths.-The molecular separation of water from salt when a brine is boiling has to some extent a counterpart in the separation of ice from brine when the latter is subjected to cold. If a weak brine such as 9 of water to 1 of saturated brine (that is, 1 of salt to 24-2 of water) is subjected to cold, pure ice begins to be formed in this case at -1°5. The temperature gradually sinks; but, as has been shown, the solid part consists of ice, which may be completely freed from salt by mere pressure. This formation of ice continues, and the temperature sinks until the inferior limit attainable by an ice-salt freezing-mixture is reached (-22° C.). The brine, of course, becomes richer and richer in salt. In other words, brines richer in salt yield up ice at lower temperatures. Thus, in Table VI. The same brines are examined as were examined in the preceding Tables. The various brines were examined in succession, being contained in small beakerglasses in an ice-salt freezing-mixture. It is noteworthy that all