GOLD.— Quantitative Estimation or Ores and Allots. 111

This substance, in the moist state, is dark purple-red, and after drying brown. Its true nature andcomposition have given rise to much discussion.According to some authorities it contains metallicgold with hydrated stannic oxide; according toothers the gold is oxidized. Berzelius concludedfrom the researches of M. Figuier, that it consistsof a hydrated double stannate of gold and tin, asexpressed by the formula—

SriAn^Si]oO :1 4 II 2 0.

It seems certain that whatever the composition bebefore ignition, after this has taken place the powderconsists of a mixture of stannic oxide (Sn0 2 ) andmetallic gold; but the fact of the powder in itsunignited state being soluble in ammonia, and notyielding any gold when treated with mercury, ap-pears to be convincing proof that free gold does notthen exist.

The composition assigned by Berzelius , whichhas been confirmed by M. Barual, accounts for theprincipal reactions of the purple of Cassius. Thus,heat decomposes it, and expelling water leaves be-hind a mixture of gold and stannic oxide.

Purple of Cassius is decomposed by the acids, butis not changed by the action of light. Hence it issomewhat extensively employed in the arts, chieflyfor colouring glass and porcelain. When mixed witha little borax or some fusible glass, and applied tothe surface of china, it imparts to it a beautiful roseor a rich purple colour. Its different applicationsas a colouring material have already been explainedin the articles Enamels and Glass. It is remarkablethat the oxide of tin does not appear to be essential,since even finely-divided metallic gold alone willgive the same purple hue.

Spectrum of Gold. — IIuggins has observed no lessthan twenty-three lines in the spectral band pro-duced by gold, but of these only six are of import-ance, the others being so faint as to be barely visible.The two brightest (equal in intensity to the well-known D lines of sodium) occur one in the yellow -green near to D, and the other in the green betweene and l) of Fraunhofer’s lines. Of the four otherimportant lines, three occur in the yellow-orangebetween c and d, and one, possessing a high re-frangibility, in the blue between F and G.

Recently Lockyer has suggested the use of thespectrum of gold in connection with that of copperas a means of determining the amount of gold con-tained in a given alloy, thus superseding the ordinarymethod of assay. A number of experiments havebeen made by him, and in the Royal Mint, on thesubject, and a paper was read before the RoyalSociety by Lockyer and Roberts ; but as yet nopractical application has been made of the discovery.

Quantitative Estimation of Gold Ores andAlloys. —In proceeding to determine the exactamount of gold present in an ore or alloy, it is ob-viously necessary to exercise the greatest caution inthe sampling. Excellent advice on this point isgiven by Percy in his admirable lecture on themetallurgical treatment and assaying of gold ores,

delivered at the Museum of Practical Geology.Careless sampling, he remarks, can only mislead ;assays of individual specimens may be accurate, butthey are worse than useless if the assayer has notoperated upon an average sample of the ore. lietherefore advises the capitalists, to whom prospec-tuses of gold-mining schemes may be submitted, notto be allured with glittering specimens of gold ore,with assays yielding a high produce, and with theglowing statements of sanguine promoters or en-thusiastic adventurers, without having ascertainedon good evidence that the samples which are pre-sented to his notice are really average samples of theore, and that something like a continuous supplymay reasonably be expected. If such specimens, headds, do not represent an average, they becomewhat the Cornish miner calls slocking-stones, whichare at all times enticing and dangerous to the inex-perienced and unwary, and never more so than inthe case of auriferous ores. The sampling gener-ally devolves upon the miner, but the assayer andmetallurgist should likewise understand the business.Assayers of great experience and high integrity mayoccasionally, he adds, commit unintentional mistakes.Thus, a few years ago, two small pigs of lead fromSouth America , very rich in silver, were offered forsale. Portions had been taken from the top andbottom of each pig, with a view to obtain a fairaverage. Percy had occasion to attempt to verifythe report of the assayers. Portions were takenfrom the same parts of each pig as in the first in-stance ; but the results did not agree with the report,nor did Percy’s assays agree with each other ontaking fresh portions. It was therefore certain thatthe composition was not uniform, and that the por-tions taken for the purpose of assaying in neither caserepresented an average. The pigs were accordinglysent again to the same assayers. Each pig wasmelted separately, and while melted a sample wastaken. A second report was given, which differedfrom the former to the extent of 1000 ozs. to theton! In the sampling of gold ores most especialcare should be taken, as the precious metal existsirregularly diffused through the mass, in particles ofvery different size, and as minute errors in samplingwill necessarily be greatly multiplied when the.quantity of gold per ton is calculated from the assayof, it may be, 500 or 1000 grains of ore.

The quantitative estimation may be made, moreor less correctly, by one of four methods:—1. Bydetermining the specific gravity of the ore or alloy ;2. By the touchstone ; 3. By chemical analysis, orthe wet method of assaying; and 4. By the metal-lurgical, or dry method, in which fire is the principalagent. The first two afford only approximations,and the second is applicable only to alloys of gold ;but both may be useful when circumstances renderimpossible or inconvenient the performance of anexact assay by the wet or dry method.

1. Quantitative Estimation by Specific Gravity. —Itis evident that the high specific gravity of gold notonly affords facilities for separating it, by means ofwashing, from other matters, but also for estimating