centimetre of the decimal solution, capable of precipitating 0.001 gr. of pure silver, is introduced. If any silver remain in solution, the liquor becomes cloudy, and after being again shaken another centimetre of the decimal solution is added. If the liquor again becomes turbid, it is again well shaken, allowed to become clear, and a third centimetre of the decimal solution poured in, and so on, until no further cloudiness is produced by the addition of the decimal solution. Suppose that 5 of the cubic centimetres of the decimal solution have been introduced in succession, and have produced a precipitate in the liquor, while the addition of the sixth has not affected its transparency, it may be concluded that after the precipitation of 1 gramme of pure silver by the decilitre of the normal solution, the liquor still contained at least Tooths of a gramme of silver. But as the fifth cubic centimetre of the decimal solution produced a cloudiness and the sixth did not, it is evident that the liquor did not contain more than ths of a gramme of silver, and thus in adding 45ths, the exact result is certainly attained within one half-thousandth of the truth. The standard of the alloy in question will thus be 896+4)=900 thousandths. If, on the contrary, the decimal solution produce no further precipitate in the solution of silver which has already received the decilitre of the normal liquid, the standard of the alloy is evidently below ths. Its exact composition is ascertained by means of a standard solution of silver in nitric acid, so adjusted that 1 litre of the liquor contains exactly 1 gramme of pure silver: this is called the decimal solution of silver. In using it, a cubic centimetre thereof is dropped from a pipette into the bottle containing the assay, and it occasions a precipitate of salt exactly corresponding to the same volume of the decimal solution of common salt which was added for the purpose of ascertaining if all the silver were precipitated. The liquor being made clear by agitation, another cubic centimetre of the silver solution is added. If cloudiness be produced, the bottle is again shaken, and a third measure of the solution introduced, and so on until the silver solution ceases to produce a precipitate. Suppose the first 5 cubic centimetres of the silver solution produced a precipitate, and that the sixth did not, it is probable that the whole of the fifth was not entirely decomposed, and it is usually stated that 44 cubic centimetres of the silver solution have decomposed the excess of the chloride of sodium left in the liquor after the addition of the decilitre of the normal solution. Hence we must subtract 4 thousandths from the presumed title of the alloy, the correct standard of which will be expressed by 896-448914 thousandths.

In places where numbers of assays of silver and copper alloys have to be made, the apparatus is so arranged as greatly to assist the operation. The normal solution is kept in a large vessel of sheet-copper, tinned on the inside, and supported on a shelf near the ceiling of the laboratory, as shown at v, Fig. 1989; and to prevent evaporation it is furnished with an immovable cover, through which passes a tube for the

pipette P contains exactly a decilitre of the saline solution, and it is connected with the tube t by another tube containing a thermometer. The metal connecting piece, by which the tube t' is fastened to the pipette, is furnished with 2 stop-cocks s and s'. In an assay the operator closes the extremity of the pipette with the forefinger of the left-hand, and with the right opens the stop-cocks & s': the stop-cock s', shown separately in Fig. 1989, is so constructed as to allow the air to escape in proportion as the solu tion enters the pipette. When this is filled a little above the mark m, the cocks ss are closed, and the pipette remains charged with the solution after the

Pig. 1990.

finger is removed. Below this apparatus is a sliding support T, in which is a case of copper b for supporting the bottle which contains the solution of the alloy in nitric acid, and close to it is a small stand containing a sponges covered with linen and arranged at the exact height of the beak of the pipette. The operator slides the plate p in the grooves of the stand T, so that the sponge may be in contact with the end of the pipette, and by carefully admitting air through the stop-cock s', the solution is allowed to descend until it reaches the mark m scratched on the glass. The sponge removes the last drop of the solution which would otherwise be attached to the beak, and as the sponge becomes saturated the solution passes down the hollow stem of s into the cylindrical vessel beneath. The slide is now drawn forward until it is stopped by a peg, when the neck of the bottle will be exactly under the beak of the pipette: the stop-cocks is now opened, and the solution flows into the bottle. When a number of alloys are to be assayed at the same time, they are contained in a number of bottles arranged in a metal frame something like a cruet stand, Fig. 1990, and marked with numbers; and after the acid has been added, the stand is placed in hot water, in order that the heat may promote the solution. By the side of each bottle is a small cup for holding the stopple. When the solution is completed the nitrous fumes are removed from the bottles by blowing into them with a glass tube, and a decilitre of the normal solution of salt is added to each, as already noticed. The bottles are afterwards placed in another metal frame F, Fig. 1991, suspended from the extremity of a steel springs, and steadied below by an elastic spring IR of vulcanized india-rubber. The bottles being

which, when thrown up, is supported by a rest r, are well shaken by an assistant, who takes hold of the handles hh, and moves the whole apparatus briskly up and down for a few minutes. When the liquors have become clear the bottles are removed from this frame to a black table furnished with divisions numbered to correspond with the numbers on the bottles. The decimal solution, which is contained in a bottle with a pipette passing through its stopple, is now employed for determining the exact standard of the various assays. The pipette has a line drawn on its surface, graduated so as to allow the operator to measure out 1 cubic centimetre of the solution. For this purpose the point of the forefinger closes over the upper extremity of the tube, which is then removed from the bottle, and by the careful admission of air at the top, the liquid is allowed to drop until it has fallen to the level of the line marked upon the glass. The top of the tube is again closely stopped by the finger, and the cubic centimetre of the solution is transferred to the first bottle of the series, into which it is allowed to flow by removing the finger. Each of the other assays receives the same quantity of the solution. The bottles are then examined, and a chalk mark is made on the black table before those bottles which exhibit precipitates. These are briskly shaken at the shaking apparatus; returned to the black table, and another dose of the decimal solution is added to each bottle, in which a precipitate was obtained by the last operation. In this way the several bottles which give no precipitate are discovered, and the number of chalkmarks before each shows the number of cubic centimetres of the decimal solution added to each assay. Half a centimetre is deducted from each to allow for the loss on the last addition, a portion only of which has probably been decomposed.

The normal solution of common salt is prepared at the temperature of 15° Cent. But as the density of the solution varies with the temperature of the air, it is necessary to apply certain corrections when the assays are made at temperature above or below 15° C. For this purpose the thermometer contained in the tube t, Fig. 1989, should be consulted, and the corrections read off from tables prepared for the purpose; but in order to prevent error, the assayer usually makes an experiment every morning on the saline solution with 1 gramme of pure silver, and the indications thus afforded enable him to ascertain the exact constitution of the solution, and to correct it accordingly.

The standard solution of chloride of sodium is made by dissolving 500 grammes of common salt in 4 litres of water, and filtering. The additional quantity of water required for the normal solution, supposing the salt to be pure, is now added, and the solution is carefully adjusted to the proper strength by testing it with the solution of pure silver in nitric acid. The decimal solution is prepared by pouring a cubic decilitre of the normal solution into a bottie of the exact capacity of a litre, and filling it up with pure distilled water. The decimal solution of silver

is preparel by dissolving 1 gramme of pure silver in nitric acid, and to this distilled water is added to make up an exact litre of the liquid.

If the alloy operated on contain mercury or lead, the results of the humid assay are not exact, since these metals are precipitated with the silver and decompose a portion of the normal solution. The presence of mercury is detected by the difficulty of obtaining a transparent solution by agitation. Such being the case the assay is not to be relied on. The assay of alloys containing mercury may, however, be conducted by the humid process by adding a solution of acetate of soda to the nitric acid liquor containing the silver previously to the introduction of the normal solution: the acetate prevents the formation of chloride of mercury.

cess.

Professor Graham, of University College, London, and Professor Miller, of King's College, conduct the silver assays for the Royal Mint by the humid proProfessor Graham adopts the French system of weights and measures above described. Professor Miller has modified it so as to suit English weights and measures. The following is a brief outline of the process as conducted by that gentleman, and kindly communicated by him: -10 grains of the alloy to be assayed are accurately weighed, and dissolved in 120 grains of pure nitric acid. The standard solution of salt is so regulated that 1000 grains thereof will exactly precipitate 1 grain of pure silver. Now, taking the standard of English silver at 925 silver and 75 copper, care is taken to keep an excess of silver in the solution, for if salt be in excess, however small, no amount of shaking will get it clear. Accordingly the pipette P, Fig. 1989, is charged with the standard solution to the mark m, which is equivalent to 923 grains, a quantity which is capable of precipitating 923 grains of pure silver: the liquor is agitated for about a minute, and then allowed to settle. 10 grains of the decimal solution, capable of precipitatingth of a grain of silver, are now added: if a cloud be produced a chalk mark is made against the bottle. The liquor is agitated for a minute: a second dose of the decimal solution is added, and if no precipitate is produced, it is thus shown that of the 10 grains of the alloy 9-234 grains consist of pure silver, the rest being alloy; thus showing a result a little below the standard. The standard solution of silver is formed by dissolving 10 grains of pure silver in pure nitric acid, and diluting it so that 1000 grains thereof shall contain 1 grain of silver.

The alloy of silver and copper used for coin and for the manufacture of silver-plate, is adjusted by the legislature of the country. In this country the same alloy is used for both purposes: the standard silver of England consists of 111 parts of silver and 9 of copper; or in 1,000 parts, 925 of silver, and 75 of copper, as already noticed. In order to prevent fraud, all vessels of silver are required to be stamped by the Goldsmith's Company, who are authorized to search the shops of silversmiths and seize articles (1) We have also to thank Professor Miller for allowing us to copy his apparatus, as shown in Figs. 1989 to 1992.

which do not bear the hall-mark of the company. The company makes a charge of 1s. 6d. per ounce on the weight of the object for the assay thereof and the impression of the stamp. The larger portion of this sum is paid over to Government as a tax, a small deduction being made for the assay. In France there are three different standards. The alloy used for the silver coinage is composed of 9 parts silver and 1 of copper; for plate, 9 parts silver to a part of copper; and for small articles of silver, such as those used for ornaments, the alloy consists of 8 parts silver to 2 of copper. The silver coins of the ancients and many Oriental silver coins are nearly pure: they contain only traces of copper and of gold.

The addition of a small proportion of copper increases the hardness of silver in a remarkable degree, without greatly diminishing its whiteness. An alloy of 7 parts silver and 1 of copper has a decidedly white colour, although less pure than that of virgin silver. Even with equal weights of the 2 metals the alloy is white. The maximum of hardness is obtained when the copper amounts to th of the silver. Articles formed of alloyed silver are subjected to a process called whitening, which has the effect of removing the baser metal from the surface. The article to be whitened is heated nearly to redness, and plunged, while still hot, into water acidulated with nitric or sulphuric acid, by which means the oxide of copper formed by heating the surface in contact with air is immediately removed. The matted appearance of the surface, called blanched or dead silver, due to the isolation of the particles of silver, is removed by burnishing. The blanks for coin undergo the process of whitening, whence the blanched appearance of new coin, and the darker appearance of the projecting portions, occasioned by wear in consequence of the alloy appearing beneath the pure surface. Articles of plate are also deadened or matted by boiling them in bisulphate of potash, which acts in the same way as sulphuric acid.

Silver solder consists of 667 parts of silver, 233 of copper, and 100 of zinc. Silver is largely used for plating the surfaces of articles made of inferior metals, for which purpose some of the methods described under PLATING and ELECTRO-METALLURGY may be adopted, or the silver may be applied to the surface of the object in the form of an amalgam, the mercury being driven off by heat. By a process of this kind buttons are gilt. [See BUTTON.] For silvering brass, a mixture of chloride of silver, chalk and pearlash is used: the metal is made chemically clean, and the mixture moistened with water is rubbed on the surface.

SIMPLE BODIES. A list of elementary bodies with their atomic weights and symbols is given under ATOMIC THEORY.

SINGEING. See BLEACHING, Fig. 138.

SIPHON (σipov, a tube), a bent tube, one leg or branch of which is longer than the other, used in raising fluids or emptying casks, &c. In Fig. 1993, the siphon ABC is represented with its shorter leg A B, immersed in the liquid which is to be transferred