ness, from the presence of impurities. Many substances are recommended and used for the toughening and brightening of jewellers' gold alloys, consisting, for the most part, of substances containing chlorine, such as common salt, bichloride of mercury, and sal-ammoniac. Common salt is not advisable, as it produces a very liquid slag, which is liable to run into the ingot-mould along with the metal, making the ingot irregular and full of small holes. The same remarks apply to borax. Bichloride of mercury is very useful when the brittleness of the gold arises from the presence of lead or tin; the fractured surface of the bar then presents a close grain, of a pale yellow colour. In most cases sal-ammoniac is the best agent for producing tough gold. The sal-ammoniac, like the mercury chloride, is a volatile substance, and partly vaporises, while another portion is decomposed into ammonia and hydrochloric acid. The latter probably acts on the base metals, yielding up to them its contained chlorine and forming volatile chlorides. Thus, no slag is left to run into the moulds and interfere with the ingot of purified metal. In most cases a very small amount of this flux is required, and, indeed, a great excess will be injurious. In all cases, where gold alloys containing copper are melted for toughening, a little charcoal should be added along with the chloride employed.

In making gold alloys, the ingredients of which do not contain sufficient impurities to interfere with the toughness of the gold, it is not advisable to use anything else except a little powdered charcoal, which forms a protective coating and prevents the oxidation of the copper.

In melting precious metals the quality of the coke employed is not such a matter of indifference as some manufacturers suppose. The following observations by Mr. C. Tookey, formerly assistant assayer at the Japanese Mint, bear on this subject.

1 In the Imperial Mint at Osaka, Japan, bars of gold from San Francisco were frequently converted into standard 1 Percy's Metallurgy, Gold and Silver, Part I. p. 489.

1000

307

metal by melting them with the requisite proportion of Japanese copper to produce an alloy containing 900 of gold. In melting one particular consignment there was an unusual loss. No good reason could be assigned in the Melting Department, but on investigation Mr. Tookey found that the plumbago muffles and stirrers used in the operation were coated with minute particles of the gold alloy. It appeared as though the particles had been projected from the surface of the molten metal by the escape of a gas. Mr. Tookey had often watched a similar phenomenon when pots containing a silver-copper alloy were removed from the furnaces previous to pouring. While one of these pots, holding 2500 ounces, was cooling on the floor of the melting room to the proper temperature, a strong effervescence took place at the surface of the metal, projecting it in most minute particles, which were deposited on the floor of the room. During the effervescence there was a powerful smell of sulphurous acid, which had no doubt been absorbed or occluded by the metal, while the pot containing it had been exposed, at a high temperature, to the products of combustion of coke containing much sulphur.

The size of the cast bars vary according to the branch of the jewellery trade for which they are required. Locket-makers cast the gold in broad and thin plates. Chain-makers cast their metal in long and tolerably thick strips, which, when rolled to about Nos. 10, 11, or 12 of the Birmingham wire gauge, are annealed and cut into strips in the slitting mill, when they are drawn into wire.

An ordinary jeweller's melting furnace is a wind furnace about 9 inches square and 18 inches deep, lined with fire-brick. The draught-hole is 6 by 3 inches, and the ash-pit should have a capacity at least equal to that of the furnace. The bars are 10 inches long, 14 inches wide at the top, and gradually taper towards the bottom. The furnace must be connected with a chimney at least 40 feet high, in order to obtain an active and strong draught. mouth is closed by two fire-bricks, each of which is clamped

The furnace

by a piece of flat bar-iron, well wedged on. The draught is regulated by a suitable damper fixed in the flue. Coke is the fuel generally employed. This should be of good quality and practically free from sulphur.

The furnaces used for melting gold for coinage are of larger dimensions than the one just described, and correspond in capacity and arrangement to those used for ordinary brass melting (see Figs. 7-12) described under the head of "brass." The gold is sent from the Bank of England to the Mint in ingots of 400 ounces each. The crucibles are made of plumbago and are capable of melting 1200 ounces of standard gold. The pots are heated previous to the introduction of the charge, to prevent them cracking or flying when containing the precious metal. To ascertain whether a crucible is sound a cold bar of iron is put to the bottom, when, if any cracks exist, they will become visible. When the mixture of gold and copper is melted it is thoroughly stirred with an iron rod, then poured into ingot-moulds.

The plates of gold are then assayed, two assays being made by two different men for each plate, and the correct standard thus determined. If found correct, the plates are then weighed and rolled to the desired size for cutting out the blanks.

§ 143. Preparation of Pure Gold.-The following method has been adopted for the manufacture of the pure gold TrialPlate now in the custody of the Warden of the Standards. Fine gold is dissolved in aqua-regia, the excess of acid driven off, and alcohol and potassium chloride added to precipitate traces of platinum. The chloride of gold is then diluted with distilled water in the proportion of half an ounce to the gallon, when the solution is allowed to stand three weeks. The solution is then carefully syphoned off, and oxalic acid in crystals added from time to time until the solution is colourless, the precipitation of the gold towards the end being aided by a gentle heat. The spongy gold so obtained is washed repeatedly with hydrochloric acid, distilled water,

IX

REFINING GOLD

309

ammonia water, and lastly with distilled water. It is then melted in a Picardy crucible with a little pure bisulphate of potash and borax, and poured into a stone mould. The Trial-Plate prepared in this way weighed 70 ounces and was of the average purity of 999.96 parts of gold per 1000.

$144. Refining of Gold. The gold employed for coinage, when it does not contain more than 10 per cent of silver, is usually purified by Miller's process with chlorine gas. This process consists of melting the impure gold in a clay crucible, which has been glazed inside with borax, and passing chlorine through the molten metal by means of a clay pipe. The chlorine combines with the silver to form silver chloride, which rises to the surface of the molten metal, whilst the chlorides of base metals which may be present, such as zinc, bismuth, antimony, arsenic, etc., are volatilised. A layer of borax is placed on the top to prevent the silver chloride from being volatilised. The gold thus refined varies in purity from 991 to 997 in 1000 parts, which is purer than ordinary fine gold.

Dr. J. C. Booth, of the United States Mint, has discovered a general method of toughening gold and silver which he described to the American Chemical Society a few years back. Some time ago Mr. Booth found that a quantity of brittle gold accidentally melted with some tough gold in a crucible had rendered the whole mass very brittle, crystalline in fracture, and therefore useless for coining. The whole was toughened by him in one and a half days at a trifling cost by the new process. The 75,000 ounces of gold were divided into 14 "melts" of 5400 ounces each, and each melt separately toughened. The ingots, which could be broken into pieces by striking them on the edge of a wooden box, were put into the crucible with soda ash and anhydrous fused borax, in the ratio of one or two ounces to a melt, until the crucible was nearly full. After melting it appeared as a quiet mass of metal covered with a viscid slag, disposed to swell and puff.

A few crystals of saltpetre, say one or two ounces, were then dropped into the centre of the metallic surface, and as they melted, their spreading out over the whole surface was aided by the concentric motion of the bottom of a small crucible. The moment the visible oxidising action began to slacken, the melter skimmed off, by a small blacklead dipping crucible, the fluxed matter as rapidly as was consistent with the care necessary to avoid taking up metal. The remainder in the melting pot was the toughened metal.

§ 145. Gold Plating. This consists of joining a bar of gold and gilding or other metal together by sweating or soldering. A bar of gold of a desired quality and a bar of base metal are first made perfectly flat under a stamp or press; then the surfaces to be joined are filed or scraped clean; borax is next prepared and well rubbed over the surfaces. The two bars of metal are firmly secured together by iron wire, placed in a muffle, and the temperature raised nearly to the point of fusion when the metals unite into one compact bar. This is termed joining by "sweating."

Another method, which is more generally adopted, is to join the two bars by soldering. The two bars are prepared as in the former case. The metal-bar, being larger than the gold-bar, supports the pieces of solder, which are placed along one side, and half-way along each end. The whole is then strongly heated in a muffle until the solder melts, and joins them together. These compound bars may be rolled, stamped, spun, or otherwise manipulated, as though they had been melted to form one homogeneous mass. The gold follows the reduction of the base metal during the processes of rolling, etc., and retains the relative proportion of thickness between them.

HARD SOLDERS FOR GOLD WORK

§ 146. With respect to hard solders, it may be taken as a general guide that the more nearly the solder approximates in composition and properties to that of the metal to