35, 48, 65, 80, 100, 150, and 200. If desired the -200 product may be further separated into granular and colloidal matter by panning, though such a division made by this means is arbitrary, and results obtained by different operators are not comparable. For the screen analysis it is often more satisfactory first to separate out the colloidal matter by panning and then dry the granular part before screen sizing. Another plan is to remove the colloids and the finest of the sand by washing them into a bucket through a No. 200 sieve under the water tap. The two products are then dried separately, and the oversize graded in the usual way. In this case there will probably be some -200 sand left in the original oversize which may be screened out dry and added to that which was washed through the sieve with water. Some metallurgists prefer to do the whole of the screensizing in water, and this has the advantage of avoiding any possible error due to decrepitation of the larger particles in drying, but such a risk is small when the drying is conducted at a low temperature; moreover, water screening takes longer to attain the ultimate limit of separation between each grade, and there is a danger of stopping the operation before it is complete.

In dry screening the common practice of putting a washer on the screen or rubbing the pulp on it with a block of wood is not to be recommended, because it forces through the meshes particles which would not otherwise pass, giving a higher reading for the minus product than the correct one, and incidentally it displaces the wires making the mesh irregular and soon impairing the standard quality of the screens besides materially shortening their life. The main incentive to this practice is the difficulty often experienced through blinding of the meshes by the clayey matter in the ore. For this reason it is recommended that the colloids be first removed by water washing, but if it is found necessary or advisable to dry-screen a sample entire, the meshes may be kept free from the fine slime by continually tapping sharply on the top edge of the sieve with the handle of a spatula or other similar instrument, and in this way good time may be made even on very colloidal material.

In the case of ground pulp that has been wet and subsequently dried, however, this method is not satisfactory as the colloidal part in drying forms hard pellets or granules which remain on the screens, and are recorded as sand. In such cases in order to obtain reliable results the dry pulp should be thoroughly disintegrated in water and the colloids floated off before the screen test is made. If a known quantity of the dry pulp be first weighed up the sample after disintegration in water may be washed on the 200-mesh screen under the water tap, the part that passes being allowed to run to waste, and the amount calculated by difference. In this case, as already stated, the portion remaining on 200 should after drying be re-screened on a 200 sieve as it will be found possible to make a further elimination of -200-mesh material in this way.

A list of screens and their apertures comparing the Tyler Standard, and the Institution of Mining and Metallurgy (London) standard will be found among the tables at the end of this book.

The assays of the various sizings will give a good indication of the degree to which comminution should be carried in order to liberate the precious metals, and expose them to the action of the cyanide. It is well to check the result by multiplying each assay by its proportional percentage, and dividing the sum of all the products by 100. The result should check closely the assay of the sample taken out before screen sizing.

The test may afterward be repeated using 0.1% cyanide strength, and only 2 or 3 days agitation, in order to get an idea as to the maximum size of particle that will give a good extraction under the usual conditions of slime treatment.

It is often instructive to make a screen analysis of the pulp before cyaniding, so as to see how the assay values are distributed, and what percentage is extracted from each grade.

"All-sliming" Tests.-Having obtained an idea of the effect of cyanide treatment on each grade of material, an "all-sliming" test may be made, with the object of finding out the probable maximum extraction by simple cyanidation.

The ore is ground so that the whole will pass a No. 200 sieve, and 250 grams weighed up for a charge. Lime is added in accordance with the experience gained in the previous test, and the dilution made up to 3:1 with water to which the necessary amount of cyanide has been added to give a strength of, say, 0.3% KCN. The bottle is placed in the agitating wheel and treated for 3 or 4 days, the solution being titrated each day for cyanide and alkali, and the strength maintained at the original figure.

Grinding Samples to a Given Screen Size. It is impossible in the laboratory to imitate exactly the degree of grinding in a working scale mill by using a certain screen size as a standard. For instance, if it is decided that a -100-mesh product is desired an operator may produce several entirely different degrees of grinding though using the same mesh screen, according to the frequency of the screening. If screening is done after long intervals of grinding, on a bucking table or with closely set discs if using a disc grinder, the resulting product though nominally -100 mesh will have a far finer character as a whole than if the screening be frequent during bucking or if the discs of the grinder be set apart and gradually closed up after each screening. And probably none of the products obtained by these different procedures will correspond to the pulp of a mill where -100-mesh grinding is aimed at. In one case known to the writer the product obtained from a bucking table, putting everything through a 100-mesh sieve, only contained 50% of — 200 material whereas the finished mill pulp on the same ore contained 5% of +100mesh material and 75% of -200. It is thus apparent that deductions drawn from laboratory extraction results on pulps all of which pass a given screen size must be received with discrimination.

Computing Cyanide Consumption. This may be done as in the following example: Ore taken, 250 grams. Ratio of solution to ore, 3:1 750 cc: 250 grams. Cyanide strength, 0.3% KCN.

It should be noted that cyanide consumption as figured above represents chemical consumption only. In practice there will also be a mechanical loss of cyanide in residues, and also a loss to some extent in precipitation, if the zinc process be used. The mechanical loss in residues will depend on the strength of solution used, and also on whether the slime residue is dewatered by a filter or by decantation,

In the case of silver ores the chemical consumption indicated in a bottle test usually corresponds very closely with the actual total consumption to be expected in practice, and for this reason; an important factor in the cyanide consumption is the amount that combines with the silver: on precipitation this silver is replaced by zinc, and this zinc solution on coming in contact with the new lime added in the mill is partly dissociated and some free or available cyanide liberated. The amount of such cyanide thus recovered (a phenomenon which is not apparent in the ordinary laboratory test) comes near to offsetting the mechanical loss of cyanide in the filter residues.

Tests made as described will be found in most cases to approach very closely to working conditions, and to give results concordant with large-scale figures. A variety of combinations will suggest themselves to the operator. Assuming the ore to be "allslimed" he will need to ascertain, among other things, the best ratio of solution to ore, the best and most economical cyanide strength, and the time of treatment necessary; and a series of bottle charges may be made up and carried through to elucidate each point separately.

The figure for the extraction obtained on such tests is usually based on the difference between the assays of the head and the residue, and this seems to be a perfectly proper procedure, since both head and residue sampling is completely under control. Some experimenters make a point of checking the extraction, calculated as above, by an assay of the solution, but on these small-scale bottle tests the fact of a portion of the solution being withdrawn each day for titration complicates the operation and renders the result less reliable than the figures based on the ore and residue assays.

Leaching Tests.-If the preliminary test seemed to indicate that excessively fine grinding was not necessary, it is probable that it will be more profitable to separate the pulp into sand and slime, percolating the former and agitating the latter.

As already stated, a good idea of the extraction to be expected by leaching may be obtained by agitating the sample in a bottle.