The test carriage shown with this brick-bottomed test is peculiar in being provided with a simple hand-wheel arrangement at one side of the front for lowering the front of the test ring as the cupellation proceeds, so as to regulate the flow of litharge without cutting down the litharge notch, and also to permit of pouring the contents of the test if required.

With all the above test rings the litharge very soon eats away the material forming the test, and, therefore, many devices have been adopted for cooling by means of water pipes coiled round the test ring, &c., combined often with a separate water breast, so as to diminish the corrosion. One of these devices in common use in America is the completely water-jacketed test of Steitz, shown in Figs. 155 and 156. The water jacket, a, of boiler iron is rectangular and complete, except in front, where a smaller cast-iron jacket, b, bearing the litharge notch, c, is bolted on. The only wear of the jacket is in this notch, and when it is worn through, the small breast jacket is taken out and replaced by a new one. This form of test ring for concentrating rich lead up

to 70 per cent. silver has the advantages of trebling the life of the hearth material, while saving trouble with the litharge notch, which remains at a constant level and suffers no corrosion. This constant level of the notch, however, coupled with the enormous cooling effect, renders it impossible to bring up the enriched alloy to fine silver on such a test; consequently, Steitz test is used for concentrating alone.

the

Not only, however, is the cooling effect with these completely water-jacketed tests so great that more fuel has to be employed to avoid retardation of the process, but the tests themselves are greatly racked by the varying degrees of intense heat to which they are subjected, and soon begin to leak at the seams. For these reasons metallurgists outside of the United States generally prefer the plain oval flanged cast-iron test ring, with a 1-inch pipe all round, fitting tightly against the ring and imbedded in the material of the test. At the front is a small cast-iron tapping breast carrying the litharge notch or lip. The breast is easily replaceable when it begins to leak, though it usually lasts out from five to six tests. Before using one of these breasts over again it is tested for leaks by hydraulic pressure in order to minimise the danger of springing a leak while the cupel is running.

When the system of concentrating and finishing on a separate hearth is adopted it is convenient to be able to tilt the test so as to pour the molten alloy into moulds. This can be done with the hand-wheel carriage shown in Fig. 151; but the Lynch test support, shown in Fig. 157,* is also very convenient. It consists

of two bars on the lower side of the test ring, by means of which the front end of the test is supported upon two turnbuckles, a, which in turn, by means of an iron triangle, hang from a differential chain block, b. The back of the test rests upon jack-screws. c, working in a fixed bar. With the block the front of the test can be easily raised or lowered during cupellation as required, and, with the turnbuckles, either side can be raised or lowered so as to regulate the corrosion of the hearth.

* Blake, Trans. A.I.M.E., vol. x., p. 220.

This support is never required for water-jacketed tests with a constant discharge level, which are, therefore, commonly mounted on one or other of the forms of carriage already figured; whereas the Lynch support is sometimes seen in America on finishing tests which are not water-jacketed, and when it is thought more convenient to pour the silver than to ladle it. The total quantity to be handled, however, on the finishing test is so small that, even in the United States, it is more common to use the fixed jack-screw test carriage and ladle out the fused silver into moulds, instead of attempting to pour it.

Hearth Material.-Formerly the only material used for the tests of English furnaces was bone-ash, which is still commonly used in England itself for that purpose. At Eureka, Nev., bone-ash has been practically superseded by mixtures of siliceous limestone and clay (in proportions of from 2 to 5 per cent. of the former to 1 of the latter), ground to pass a 12-mesh, and Portland cement, which has given excellent results at many American works; a mixture of cement and ground firebrick has also given good results.

Bone-ash must not be beaten down too hard, as that would make the test crack on firing. The limestone-clay tests, however (and, according to Hofman,* those of cement also), must be tamped as hard as possible, and dried, and warmed very gradually and thoroughly. In using cement it is essential to do the tamping as rapidly as possible before it commences to set.

The bottom of the test shown in Figs. 150 and 151 is made of firebrick, and experiments have been made with magnesia brick as a lining for sides as well as bottom,† but no results have been published. A brick hearth is specially used for refining argentiferous matte into bottoms (v. Chapter xvii.).

Whatever its material, the hearth should be dried out for at least a month before commencing to warm up. A plain (not water-cooled) cement test used right through for concentrating as well as finishing may last seven days, while one used four or five hours daily for finishing only will last for months, and even a limestone-clay test, under these circumstances, lasts a month; a water-jacketed concentrating test of limestone-clay will last two months.

At Port Pirie, § all the tests are alike, except that those used for finishing have no water-cooled lip or litharge breast. The mixture employed is fireclay, siliceous limestone, and Portland cement in certain proportions, which are mixed dry, damped to the required consistency, and quickly pressed into the test, which lies bottom upwards upon a cast-iron mould the shape of the required cavity Contrary to what is stated by Hofman, * E. and M. J., May 4, 1893. + Hofman, Metallurgy of Lead, p. 391. All these figures are on the authority of Hofman, op. cit., p. 392. § Private notes.

it is found that beating down a cement test too hard injures its durability. The gentle ramming is done as quickly as possible before the cement commences to set, just sufficient pressure being exercised to get out all the air in the interstices, and then the test is immediately turned up and filled with water, which is allowed to be absorbed and to evaporate naturally. The tests are then piled in a warm part of the refinery, where they dry out as gradually as possible during four to six months before being used. The average life of tests prepared in this way is, for concentrating up to 50 per cent. concentrated bullion, three weeks, at an average of 30 bars per shift, or, say, a total of 72 tons lead passing through the test, at the end of which its capacity has become enlarged from 1 to 2 tons. On the finishing side of the cupellation house each test lasts, on an average, ten days, going constantly day and night, or while an average of 400,000 ozs. silver is cupelled upon it.

Mode of Working.-After gradually heating up the test to a red heat, lead is charged through the front and back openings, and, when melted down so as to fill the cavity, the blast is turned on and cupellation started. The litharge runs off at the litharge notch, and the rate of running off is regulated, partly by the rate at which pigs of lead are fed forward at the tuyere door, partly (in tests unprovided with water breasts) by the depth of the litharge notch. In any case, it is usual to keep about half the lead covered with litharge to reduce the loss by volatilisation. When the test has become filled with rich alloy (sometimes called "bullion concentrates "), the blast is stopped, the contents of the test removed into moulds by ladling or pouring, and a new concentration begun, the furnace working continuously as long as the test lasts.

The operation of "finishing" is conducted similarly, but at a higher temperature, in an uncooled test, until sufficient crude silver is obtained, when the addition of pigs of concentrated alloy is stopped, and the last litharge removed. The phenomenon brightening" is seldom or never seen on the English

of "

hearth.

Refining of the crude molten silver is sometimes conducted in a separate hearth, but generally in the finishing hearth itself, and by the blast alone. A little bone-ash, however, is sprinkled on to absorb base metal oxides, and, generally, also a few shovelfuls of soda nitre, one at a time; the slag so produced is usually not removed until just before casting. The refining is concluded when the metal has a clean, smooth, brilliant surface, like a mirror, which does not become tarnished on stirring, and when a small sample bar shows a finely granular silky fracture and great malleability. It is usual, however, to make careful assays before casting, in order to be sure that the bars will be above the standard for fine silver (viz., 997 + ).

Casting is occasionally done with the help of the Lynch test support already described, the silver being poured direct into the moulds, but more usually the metal is ladled out into the warmed moulds. The sample is best taken from each mould as poured, and just before "setting," by means of a long iron spoon. Examples of English Cupellation.-At Eureka (Nev.)* bone-ash tests, holding 1 ton each, are filled with rich lead, containing 550 ozs. silver, which is melted down and fresh lead added as fast as it oxidises. Once in each twenty-four hours the rich alloy is tapped, yielding six to eight bars of 60 to 70 per cent. silver. In the "finishing" operation sixty such bars are cupelled in sixteen hours, giving 16,000 ozs. of doré silver, 965 fine in silver, and 30 in gold. From the concentration some litharge is obtained, running only 1 oz. Ag per ton, which is reduced to market lead in a reverberatory with charcoal. The litharge from the finishing test all averages 75 ozs. per ton, and goes to the ore furnace.

At Port Pirie (S.A.) there are six concentrating hearths and two finishing hearths, all being alike in general form and dimensions. In the former, the bars of rich bullion ("retort bullion"), averaging 4500 ozs. Ag per ton, are melted down with a 12-oz. blast at the rate of thirty bars (say, 22 cwts.) per 8-hour shift, the capacity of the hearth being from 1 to 2 tons, according to the amount of corrosion it has undergone. Concentration is stopped when the lead reaches 15,000 to 16,000 ozs. to the ton, and the contents of the test are then ladled out into moulds. The litharge from the concentration averages only 30 ozs. per ton, and is smelted in the refinery blast furnace together with siliceous ores.

About ton of the bars of concentrated bullion are melted down in the finishing hearth, and cupellation is begun, more being added at the back, as usual, to supply the place of that oxidised. The rate of cupellation on these finishing tests is about 14,000 ozs. of silver per 8-hour shift, and the quantity of silver produced from one filling is about 23,000 ozs., which is ladled out by hand into moulds, sprouting and effervescing violently as it solidifies. The average fineness of the finished silver is about 997, the chief impurity present being copper, which obstinately remains after all the lead has been removed. The litharge produced on the finishing hearth contains about 200 ozs. Ag per ton, and goes to the softening furnaces to assist in the oxidation of antimony, as explained in Chapter xv.

The silver is "dried," or refined, upon a perfectly new finishing test by exposure for six hours to the action of the blast, which oxidises the last traces of base metals. No fluxes are employed, but a little bone-ash or hearth mixture is sprinkled on to the bath towards the end of the operation in order to collect the * Curtis, Mon. U.S. Geol. Survey, vol. vii.. 1884.