This is a safe and very common method of proving a stone, since its specific gravity does not vary more than a point or so in different specimens of the same stone.

There are several ways of arriving at this, such as by weighing in balances in the usual manner, by displacement, and by immersion in liquids the specific gravity of which are known.

Cork is of less specific gravity than water, therefore it floats on the surface of that liquid, whereas iron, being heavier, sinks. So that by changing the liquid to one lighter than cork, the cork will sink in it as does iron in water; in the second instance, if we change the liquid to one heavier than iron, the iron will float on it as does cork on water, and exactly as an ordinary flat-iron will float on quicksilver, bobbing up and down like a cork in a tumbler of water.

If, therefore, solutions of known but varying densities are compounded, it is possible to tell almost to exactitude the specific gravity of any stone dropped into them, by the position they assume.

Thus, if we take a solution of pure methylene iodide, which has a specific gravity of 3.2981, and into this drop a few stones selected indiscriminately, the effect will be curious: first, some will sink plump to the bottom like lead; second, some will fall so far quickly, then remain for a considerable time fairly stationary; third, some will sink very slowly; fourth, some will be partially immersed, that is, a portion of their substance being above the surface of the liquid and a portion covered by it; fifth, some will float on the surface without any apparent immersion. In the first case, the stones will be much heavier than 3.2981; in the second, the stones will be about 3.50; in the third and fourth instances, the stones will be about the same specific gravity as the liquid, whilst in the fifth, they will be much lighter, and thus a rough but tolerably accurate isolation may be made.

On certain stones being extracted and placed in other liquids of lighter or denser specific gravity, as the case may be, their proper classification may easily be arrived at, and if the results are checked by actual weight, in a specific gravity balance, they will be found to be fairly accurate.

The solution commonly used for the heaviest stones is a mixture of nitrate of thallium and nitrate of silver. This double nitrate has a specific gravity of 4.7963, therefore such a stone as zircon, which is the heaviest known, will float in it. For use, the mixture should be slightly warmed till it runs thin and clear; this is necessary, because at 60° (taking this as ordinary atmospheric temperature) it is a stiff mass. A lighter liquid is a mixture of iodide of mercury in iodide of potassium, but this is such an extremely corrosive and dangerous mixture, that the more common solution is one in which methylene iodide is saturated with a mixture of iodoform until it shows a specific gravity of 3.601; and by using the methylene iodide alone, in its pure state, it having a specific gravity of 3.2981, the stones to that weight can be isolated, and by diluting this with benzole, its weight can be brought down to that of the benzole itself, as in the case of Sonstadt’s solution.

This solution, in full standard strength, has a specific gravity of 3.1789, but may be weakened by the addition of distilled water in varying proportions till the weight becomes almost that of water.

Knowing the specific gravity of all stones, and dividing them into six groups, by taking a series of standard solutions selected from one or other of the above, and of known specific gravity, we can judge with accuracy if any stone is what it is supposed to be, and classify it correctly by its mere floating or sinking when placed in these liquids.

Phenakite 2.965
Turquoise 2.800
Beryl 2.709 and occasionally to 2.81
Aquamarine 2.701 ” 2.80
Labradorite 2.700
Emerald 2.690
Quartz 2.670
Chrysoprase 2.670
Jasper 2.668
Amethyst 2.661
Hornstone 2.658
Citrine 2.658
Cordierite 2.641
Agate 2.610
Chalcedony 2.598 and occasionally to 2.610
Adularia 2.567
Rock-crystal 2.521 and occasionally to 2.795