Dedicated to the memory of
Prof. Lev Lvovich Tove.

Study of Gold-Bearing Ores in Metallurgical Laboratory 1 of the Tomsk Technological Institute

V. Ya. Mostovich and V. A. Pazukhin.

I.

Introduction

For the extraction of gold from gold-bearing ores, pyro- and hydrometallurgical methods are used at the present time. The pyrometallurgical method consists in smelting ores and is applicable in cases where, together with the noble metals, the ores contain copper or lead in quantities sufficient for the complete extraction of the noble metals. During smelting, copper and lead act as so-called collectors of gold and silver, and the latter are concentrated in copper matte or work-lead, the further processing of which is carried out by well-known methods established in metallurgical practice.

The economic feasibility of smelting is determined by the composition of the ores and by a number of local conditions, such as, for example, the kind and cost of fuel, the cost of mechanical power, labor, and so forth.

Where smelting is applicable, laboratory investigations are limited merely to determining the qualitative and quantitative composition of the materials subject to smelting.

Hydrometallurgical methods include amalgamation and cyanidation, based on the solubility of gold in mercury and potassium cyanide. In applying these methods, each ore requires individual industrial treatment, since the quantity of gold extracted depends on the composition and character of the ore and on the conditions of its treatment. The flowsheet and treatment conditions for a given ore that are most rational from the technical and economic standpoints can be established by laboratory investigation of a small quantity of material, provided only that this latter represents a true average sample corresponding in its composition to the entire mass of material subject to treatment. Since

the individuality of the material—for example, of the ore—is fully revealed in the average sample; and since, in laboratory investigation, the entire treatment process can be divided into separate stages, this greatly facilitates the investigator’s work and makes it possible to note a number of facts that escape the investigator’s attention when working with large masses. Thus preliminary laboratory investigations of ores that are to be subjected to amalgamation or cyanidation are not only expedient, but also necessary for establishing the scheme and the appropriate conditions for treating the given ore.

Experience has shown that systematic, scientifically organized tests carried out on a laboratory scale with small samples make it possible, with a degree of probability sufficient for practice, to predict the results of bulk treatment of the material on a plant scale. Metallurgical practice knows many cases in which large sums were spent on the construction of plants whose operation proved to be economically quite unprofitable because a process had been adopted whose suitability had not been clarified beforehand by laboratory investigations. These investigations, simple and inexpensive to carry out, and not very numerous in general, could have saved enterprises from mistakes and from considerable losses of money, labor, and time. Therefore, in countries with a developed mining industry, not a single cyanide plant, for example, or concentrating mill is built without the ores to be treated first being comprehensively investigated on a laboratory scale. The procedure for testing ores is, for example, as follows. One begins with a sample of several kilograms and, when the characteristic features of the ore have been studied, experiments are set up with 50–100 kg of ore; and finally, especially with silver ores or with ores showing a refractory character, the next stage of testing is a pilot plant for treating several tons of ore. The results obtained in such tests serve as the basis for establishing a rational scheme for treating the given ore.

The necessity of preliminary ore tests has not yet been sufficiently realized among broad circles of our industrialists and even among persons directing the technical side of enterprises. Evidence of this is the fact that plants for ore treatment are often built in our country according to ill-considered, inexpedient plans—either too complicated and expensive, or primitively simple. In this, the basis of the scheme is not data from a preliminary investigation of the ores, but imitation of existing models and templates, owing either to the presence of some similarity of the given ore with other ores that are treated according to the borrowed scheme, or to a striving for cheapness of treatment. In most cases—

... in a number of cases all this leads to opposite results or is associated with predatory exploitation and the squandering of the country’s natural ore wealth.

Another reason for the limited prevalence among us of preliminary laboratory testing of ores is unfamiliarity with the methods of testing and the absence of corresponding literature in the Russian language. To be sure, in almost every English manual on the metallurgy of gold or on special processes—for example, cyanidation—there is information on methods for the preliminary testing of ores; but this information is either brief and of too general a character, or is presented by the authors without proper criticism—in other words, it is simply copied from other similar manuals without experimental verification.

The description of the methods of laboratory investigation of gold-bearing ores employed in Metallurgical Laboratory I of the Tomsk Technological Institute constitutes the subject of the present article. It is necessary, however, to point out that at present it is not yet possible in the laboratory to carry out complete and systematic investigations of ores, since the equipping of the laboratory for this special purpose, begun only in 1913 with very insufficient funds, has not yet been completed; moreover, in 1914 the war began, which greatly hampered the matter of equipping the laboratory, owing to the impossibility of obtaining the necessary instruments and machines ordered from specialized foreign firms.

The following principle lay at the basis of the adopted plan for equipping the laboratory. Modern practice in the industrial treatment and laboratory investigation of ores is based on taking a correct average sample, which, by systematic reduction, can be brought down to a comparatively small quantity—one hundred or more kilograms. Further tests can be successfully carried out with this small quantity of ore on apparatus of laboratory type and scale. Therefore, the equipping of the laboratory in advance with large installations on a factory scale was excluded, since this is cumbersome, expensive, and has no advantages whatsoever.*)

II.

Amalgamation.

General remarks. The extraction of gold by means of amalgamation is based on the property of metallic gold to give fairly firm solutions or alloys with mercury, called amalgam. The latter

*) This view is also shared below by M. F. Ortin. See Zh. R. M. O., 1915, 613.

is mechanically separated from the gangue by washing. The success of amalgamation depends on the completeness and duration of contact between the particles of gold and the mercury. Contact may be effected in two ways, namely: by grinding the ore in the presence of water and mercury, or by passing streams of ore slurry, the so-called pulp, over amalgamated copper sheets, called sluices. Accordingly, a distinction is made between internal and external amalgamation. In internal amalgamation, mercury is poured into the grinding apparatus, for example into a stamp-battery mortar; in external amalgamation, the ground ore is carried by water out of the grinding apparatus and, in the form of a thin stream, is passed over sluices. In our country, in most cases both of the indicated methods are applied simultaneously, although the question of whether to use internal or external amalgamation must be established experimentally for each ore. If internal amalgamation increases the extraction of gold, then it is necessary; otherwise it is superfluous.

Amalgamation is a very important preliminary process in the treatment of ores containing free, metallic gold not associated with sulfides. In such a case, the greater part of the gold can be extracted directly by amalgamation, and for the so-called freely amalgamating ores the extraction of gold reaches 60–80%. To extract the gold remaining in the tailings from amalgamation, subsequent treatment by cyanidation is necessary.

The questions of the necessity of amalgamation in general, and of the percentage extraction of gold at which one should aim in treating a given ore by amalgamation, are purely economic and in practice receive different answers.

It is known that, with sufficiently fine grinding of freely amalgamating ores containing even coarse gold, the ore can be brought into such a state that by cyanidation without preliminary amalgamation it is possible to extract almost all the gold contained in the ore, and preliminary amalgamation proves superfluous. Sometimes such treatment is advisable, but in the majority of cases it proves economically less advantageous than amalgamation followed by cyanidation, owing to the following circumstances. First, such treatment is associated with the high cost of the preliminary fine grinding of the ore; second, the enterprise loses interest on the gold, which is realized in cyanidation much more slowly than in amalgamation; and third, amalgamation is still the simplest, most easily performed, and cheapest method of extracting gold. In view of the great value of the latter, in the practice of gold mining a rule has become established according to which

although even a partial recovery of the gold should, insofar as possible, be pursued at the very earliest stage of the ore-treatment process.

This is why amalgamation, although in a technical respect a process of little perfection, has to this day enjoyed the widest use among the hydrometallurgical methods of treating gold-bearing ores.

III.

Laboratory investigations relating to the process of amalgamation.

Laboratory investigations have as their aim the determination of the percentage recovery of gold from a given ore and the most suitable conditions for treating the latter. Laboratory investigations must provide answers to the following questions:

1. what is the gold content in the ore;

2. what is the percentage recovery of gold attainable in the amalgamation of the ore;

3. whether internal or external amalgamation is necessary, or both processes together;

4. what is the gold content in the tailings from amalgamation; on this content depends the decision of the question of further treatment of the tailings;

5. what is the quantity of sulfides (concentrates) in the ore and the distribution of gold between the pyritic and non-pyritic parts of the ore. The quantity of concentrates and their relative richness determine the necessity both of separating the concentrates before amalgamation and of treating them independently.

The percentage recovery of gold attainable in the amalgamation of a given ore depends on the following circumstances:

1. on the composition and character of the ore,

2. on the kind and form of the gold in the ore,

3. on the size of the ore grains, or the degree of its comminution,

4. on the size of the surface and the inclination of the sluices in the case only of external amalgamation,

5. on the density of the pulp, or the ratio of the weight of water to the weight of ore during amalgamation.

The last two conditions, as well as certain other details, can be finally established only by experiments on a works scale.

The composition and character of the ore, as well as the kind and form of the gold contained in the ore, can be determined by chemical and mineralogical analyses of an average sample of the crushed ore. From the average sample of ore arriving at the laboratory for testing, a test sample is taken by successive crushing and reduction of the sample.

to reduction to an amount of 500–1000 g. The sample is ground in a ball or disk mill to such a degree that all of it passes through an I. M. M. 150-series sieve. Separate portions of the ground sample are subjected to careful sampling, qualitative, quantitative, and mineralogical analysis. In quantitative analysis one determines not only those components that are present in the ore in predominant amounts, but also those whose quantity, though very small, nevertheless strongly affects the results of amalgamation. Such are, for example, copper, sulfur, arsenic, antimony, and sulfuric acid in free and combined form. As is known, decomposed sulfides and water-soluble sulfates cause “sickening” of the mercury, its coating, the greening of amalgamated plates, and all this reduces the extraction of gold and increases losses, or consumption, of mercury, a rather expensive reagent.

Microscopic study of the material, in connection with its microchemical analysis, is a very essential method for clarifying the mineralogical composition of the ore, the character of the grains, and the form and appearance of the gold. From the data of microscopic analysis one can, to a certain extent, foresee the results of amalgamation.

The grain size of the ore is the most important factor determining the results of amalgamation. Amalgamation is a process of dissolution; the rate and completeness of dissolution of gold in mercury depend on the degree of completeness and the duration of contact between particles of metallic gold and mercury. The completeness of contact is determined by the degree of grinding of the ore. The purpose of grinding the ore is to free the particles of gold from their mechanical bond with the minerals enclosing them and to expose the metallic particles of gold. The grinding of the ore should be carried out only to that maximum grain size at which, under the given conditions, a sufficiently high extraction of gold is obtained. The most advantageous degree of grinding is determined solely by economic conditions, chiefly by the cost of grinding and by the method of further treatment of the tailings from amalgamation. In practice the degree of grinding of ores for amalgamation varies within wide limits, from 9 to 60 openings per 1 linear inch.

Laboratory investigations can indicate at what grain size a given ore gives maximum extraction and how the gold remaining in the tailings is distributed among grains of different sizes. For this purpose a sieve analysis of the ground ore is carried out; that is, it is screened by grain size into classes, and the gold content in each class is determined. This operation is called fixing the gold by classes. By fixing the gold by classes before and after amalgamation, from the difference in gold content we obtain the necessary data for calculating extraction by classes. Obviously, the most advantageous grain size, in terms of percentage ...

extraction will be the one that shows the maximum extraction of gold.

Sieve analyses occupy a very important place in all methods of laboratory investigation of ores, and therefore it will not be superfluous to give indications as to how sieve analyses are carried out.

IV.

Standard Sieves.

The size of the grains of crushed ore is denoted by the number of openings in 1 linear inch of the sieve through which the grains pass. But the term “number of openings in one linear inch” does not yet determine the actual size of the grain. The latter depends on the length of the free space between the wires of the sieve, or on the size of the sieve opening. The latter, in turn, depends on the diameter of the wire. Thus, for example, commercial meshes with 20 openings in one linear inch are made from wires of 16 different diameters, from 0.0095 to 0.032 dm. A sieve with 20 openings, or a “20” sieve, made from wire 0.0095 dm in diameter, has openings of 0.0405 dm, whereas the same sieve made from wire 0.032 dm in diameter has openings of 0.018 dm. Thus, in a “20” sieve, with 17 different wire diameters from 0.0095 to 0.032 dm, the size of the openings varies from 0.0405 to 0.018 dm, and we can obtain, using these sieves, which have exactly 20 openings per linear inch, 16 different grain sizes—“20,” the largest grain being 2¼ times greater than the smallest.

To prevent such diversity in the construction of sieves, standards have in practice been established and adopted for the manufacture of sieves used in assaying for sieve analyses.

Assay sieves are made of iron, brass, bronze, or copper wire. All openings in each sieve must be perfectly identical and have the exact shape of a square. The wire must not change its position or slip during the operation of the sieve. The construction of a series or set of sieves must be based on some definite relation between the sizes of the openings in two successive sieve numbers. This definite relation must remain constant throughout the entire series of sieves and is called the basis, or modulus, of the given series. Thus, a series of standard sieves represents a scale of precisely measured openings, increasing or decreasing according to a definite law throughout the entire series.

At the present time two series of sieves are considered generally accepted: the Tyler series (Tyler Standard Screen Scale) and the series of sieves of the Institute of Mining and Metallurgy in London, abbreviated as the I.M.M. series (Institute of Mining and Metallurgy).

Tyler sieves are manufactured by the firm W. S. Tyler (Cleveland, Ohio, United States of North America) and, as meeting scientific requirements, are especially recommended by Professor R. Richards. Tyler sieves have been adopted in the United States of North America and are used there for screen analyses in concentrating mills and in assay laboratories.*) The modulus of the series is taken as \(\sqrt{2}=1.414\), recommended in its time by Rittinger in his work on ore concentration. The starting point is a sieve with 200 openings in 1 linear inch; the length and, consequently, the width of the opening of this sieve is equal to 0.0029 in., or 0.074 mm. The entire series consists of 18 sieves with openings from 0.0029 in. to 1.05 in. In this series the diameters of the openings in two successive sieves are in the ratio \(1.414\); thus, for example, the length of an opening in sieve “8” is 1.414 times greater than in sieve “10” and 1.414 times smaller than in sieve “6.”

Table I gives the series of standard Tyler sieves.

TABLE I. Standard Tyler sieves.

Opening size (modulus \(\sqrt{2}=1.414\)), in.	Opening size (modulus \(\sqrt{2}=1.414\)), mm	Number of openings in one linear inch	Wire diameter, in.
1.050	26.67	—	0.149
0.742	18.85		0.135
0.525	13.33		0.105
0.371	9.423		0.092
0.263	6.680	3	0.070
0.185	4.699	4	0.065
0.131	3.327	6	0.046
0.093	2.362	8	0.032
0.065	1.651	10	0.035
0.046	1.168	14	0.025
0.0328	0.833	20	0.0172
0.0232	0.589	28	0.0125
0.0164	0.417	35	0.0122
0.0116	0.295	48	0.0092
0.0082	0.208	65	0.0072
0.0058	0.147	100	0.0042
0.0041	0.104	150	0.0026
0.0029	0.074	200	0.0021

*) Proceedings Amer. Soc. Testing Materials, Vol. XIII, (1913). G. A. Dishro, “Screen—Scale Sieves Made to a Fixed Ratio.”

Eighteen sieves of Tyler’s series are sufficient for ordinary laboratory practice. For more complete sieve analyses, three more intermediate sieves with 80, 115, and 170 openings are included in the series between the “65” and “200” sieves. These sieves form an additional series of sieves with openings from 0.0029 in. (the “200” sieve) to 0.0082 in. (the “65” sieve); the modulus of this series is \(\sqrt[4]{2} = 1.189\).

Table II gives the additional series of Tyler sieves.

TABLE II. Additional Series of Tyler Sieves

Size of openings, modulus \(\sqrt[4]{2} = 1.189\), in.	Size of openings, modulus \(\sqrt[4]{2} = 1.189\), mm.	Number of openings per linear inch	Diameter of wire, in.
0.0082	0.208	65	0.0072
0.0069	0.175	80	0.0056
0.0058	0.147	100	0.0042
0.0049	0.124	115	0.0038
0.0041	0.104	150	0.0026
0.0035	0.088	170	0.0024
0.0029	0.074	200	0.0021

In England and on the continent, in particular in Russia, the I.M.M. series of sieves is accepted.*)

As was indicated above, the term “number of openings in one linear inch” does not yet determine the true size of the openings in the sieve, and consequently the size of the grain. To determine the latter, it is necessary to specify the size of the opening, or the number of openings per linear inch of the sieve and the diameter of the wire. I.M.M., foreseeing the inconvenience of such a designation of grain size, adopted the following standards, which formed the basis for the construction of a series of normal laboratory sieves:

1. the openings in the sieves are square;

2. in each sieve the diameter of the wire is equal to the size of the opening;

3. the sum of the areas of all openings in each sieve, or the live section of the sieve, amounts to 25% of the total area of the sieve;

*) Trans. Institution Min. Metallurgy Vol. XX, (1910), 486. Metallurg. Chemic. Eng. Vol. XI, 1913, 168.

4, the product of the number of openings in a linear inch in each sieve by the length of an opening is a constant value and is equal to 0.5. Consequently, with \(n\) openings in 1 linear inch, the size of the opening in any sieve of the series is \(= \frac{0.5}{n}\) in. The modulus of the I. M. M. sieves is not constant and varies between 1.258 and 1.33. In the entire series there are seventeen sieves, with the number of openings from 5 to 200 in one linear inch.

Table III gives the series of standard I. M. M. sieves.

TABLE III. Series of standard laboratory sieves of the I. M. M.

Number of openings per linear inch	Opening size, in.	Opening size, mm	Wire diameter, in.	Wire diameter, mm	Clear area of sieve, %
5	0,10	2,54	0,10	2,54	25,0
8	0,062	1,574	0,063	1,600	24,60
10	0,050	1,270	0,050	1,270	25,0
12	0,0416	1,056	0,0417	1,059	24,92
16	0,0312	0,792	0,0313	0,795	24,92
20	0,025	0,635	0,025	0,635	25,0
30	0,0166	0,421	0,0167	0,424	24,80
40	0,0125	0,317	0,0125	0,317	25,0
50	0,010	0,254	0,10	0,254	25,0
60	0,0083	0,211	0,0083	0,211	24,80
70	0,0071	0,180	0,0071	0,180	24,70
80	0,0062	0,157	0,0063	0,160	24,60
90	0,0055	0,139	0,0055	0,139	24,50
100	0,005	0,127	0,005	0,127	25,0
120	0,0042	0,107	0,0041	0,104	25,4
150	0,0033	0,084	0,0083	0,084	24,5
200	0,0025	0,063	0,0025	0,063	25,0

Although the I. M. M. sieves have found wide application in practice, they have certain substantial shortcomings.

Since, for laboratory sieves, the sieve diameter, for practical reasons, is made no more than 8–10 dm, it is desirable for the sieve to have as large a working, or clear, area as possible, in order to speed up the sieving operation. In the I.M.M. sieves, however, this clear area amounts to only 25% of the total area of the sieve, and the remaining 75% are a non-working, or dead, area. The adoption of so small an open area in the I.M.M. sieves is explained by the desire to increase the stability of the wires in the weave and to prevent their displacement or slipping; for this purpose, in constructing the sieves, a thicker wire was chosen, which naturally reduced the open area of the sieves. Another drawback of the I.M.M. sieves must be recognized as the absence in them of a constant modulus.

Very recently, in the United States of America, it has been proposed to introduce a “metric” series of sieves.* In April 1916, a convention of representatives of American technical and engineering societies, convened on the initiative of the Bureau of Standards, adopted a normal series of sieves. The scale in this series is metric. The base is a sieve with an aperture size of 1 mm. For sieves with apertures from 1 to 8 mm, the modulus adopted is \(\sqrt{2} = 1.4142\); for sieves with apertures smaller than 1 mm, \(\sqrt[4]{2} = 1.1892\). The sieves are designated by the aperture size, expressed in mm. In the entire series there are 25 sieves, with aperture sizes from 8 to 0.044 mm and with the number of apertures from 2.5 to 330 in one linear inch.

Let us now turn to sieve analysis.

V.

Sieve analysis.

Sieve analysis is carried out by the dry or the wet method, manually or with the aid of automatically operating apparatus.

Analysis by the dry method consists in taking a definite sample of crushed ore, usually 100 g, sieving it successively through a series of sieves, and weighing the material remaining on each sieve and that which has passed through the last sieve. The product remaining on a sieve is designated by the corresponding sieve No. with the sign \(+\); the undersize is designated by the sieve No. with the sign \(-\).

The material is poured onto the coarsest sieve of the selected series of sieves; the sieve is inserted into the corresponding pan, covered with a lid, and shaken. Through each sieve the material is sieved to refusal. In doing this, an attempt is made to keep the sieve in a horizontal position, since when the sieve is inclined, through the apertures of the latter there may—

* Bulletin Amer. Inst Min. Eng. june, 1917, 945. A. Standard Screen Scale for Tesing Sieves.

can pass prismatic or cylindrical particles whose cross-section is smaller than or equal to the opening of the sieve, but whose length may considerably exceed the opening of the latter. To speed the work and break up lumps during sieving, weights (washers), in the form of wooden or iron plates with smooth surfaces, are sometimes placed in the sieve. These weights, exerting pressure on the particles, facilitate their passage through the sieve. But the use of such weights is not recommended, since they promote the passage through the sieve of prismatic or cylindrical particles—for example, long threads of gold, silver, or copper—before the latter have been sufficiently comminuted. In addition, the weights deform the sieve. All this introduces errors into the results of assaying the products of sieve analysis.

The factors that cause the grains of material to pass through the sieve are the intensity, duration, and number of shakes per unit time. Manual sieving provides only a constant duration of shaking and, of course, not with complete precision, since an exact accounting of all interruptions in the work would greatly lengthen the entire sieving operation. Usually manual sieving is carried out through one sieve at a time, and not through a set of several sieves.

Mechanical sieving makes it possible to make the influence of the above-mentioned factors the same for all classes and to make the results of sieve analysis independent of the operator’s will, and therefore more correct. For mechanical sieving in the laboratory, an apparatus manufactured by the firm W. S. Tyler in Cleveland is used. The apparatus consists of a platform suspended on springs and set into strong vibration by a special electric motor. A set of sieves, inserted one into another telescopically, is fixed in the apparatus, and the whole system is set into vibration for a definite period of time, regulated by an automatically acting current switch. The personal factor is thus completely eliminated, and at the same time 400–500 g of material are sieved on 6–8 sieves, which greatly shortens the duration of the sieve analysis.

Microscopic study of the products of dry sieving shows that a considerable part of the coarse grains is encrusted with the finest particles of the material, the so-called slime, and thus the separation of particles from one another according to grain size is not sufficiently complete.

To eliminate this phenomenon and obtain clean products, a weighed portion of the material taken for sieve analysis is sometimes stirred with an excess of water, and the turbid water with suspended particles of slime is decanted. The operation is repeated until the slime is completely removed, until the decan-

the decantation water becomes sufficiently clean and transparent. The decantation waters are allowed to settle; the clear water is drawn off with a siphon, the sediment is transferred to a filter, allowed to drain, and dried on an iron or cast-iron baking sheet. As a filter, the laboratory uses a square wooden frame fixed on four legs. A canvas filter is stretched over the frame, on which filter paper is placed. The sediments are dried together with the paper, which is easily separated from the dry silt. The washed and dried materials (sand and silt) are weighed and subjected separately to sieve analysis.

The method of preliminary washing of the material for the purpose of removing “amorphous” silt and obtaining clean size classes is also recommended by the U. S. Bureau of Mines); however, this method nevertheless does not fully attain its purpose, since microscopic examination of the products of the sieve analysis shows that much fine material still remains on the coarse particles, and, most importantly, as experiments in our laboratory have also shown, preliminary washing before sieve analysis often completely changes the results of assaying individual classes; evidence of this will be given below. This is why this method cannot be recommended.*)

Sieve analysis by the wet method is carried out as follows. A sieve containing a weighed portion of ore is lowered to half its height into a vessel with water. By carefully raising and lowering the sieve, a current of water is created, by which the particles of the material are drawn through the sieve. If there is a large amount of silt in the material, the water in the vessel becomes turbid, but the turbidity does not prevent the same water from being used for further work. The product remaining on the sieve is washed in a new portion of water until a clear wash water is obtained; this water is used for washing the products remaining on the subsequent sieves. For materials containing much silt, it is more advantageous to wash out the silt at once through the finest sieve; the method is carried out as described on p. 12, but the turbidity is filtered through sieve No. 200. The product that has passed through the first sieve is passed through the next one, and the operation is repeated with the entire series of sieves. This method gives a completely clean material, but requires much time for its execution.

In addition, it must be noted that, as laboratory investigations have shown, dry and wet screening give non-coincident results.

Determination of gold by classes. To determine gold by classes, a preliminary sieve analysis of 100 g of ore is performed.

) Min. Sci. Press. Vol. 115, (1917), 236.
*) Min. Sci. Press. Vol. 115, (1917), 344.

From the data obtained they determine how much material must be taken so that in each class there will be not less than 100 g—a quantity sufficient for two crucible assays for gold. Having weighed all the classes, they determine in each of them the gold content and carry out a microscopic analysis of each class. This gives very useful indications both as to the character of the grains in each class, their form and composition, and as to the form of the visible gold.

The results of sieve analyses are arranged in the form of the following approximate table (see Table IV).

TABLE IV. Sieve analysis and recording of gold by classes.

WEIGHED PORTION ........ g,

Average gold content ........ gold, in 100 poods.

Classes	Quantity, g	Quantity, % \(a\)	Gold content in 100 poods, zolotn. \(b\)	Total quantity of gold in class, \(\dfrac{a \cdot b}{100}\) zolotn.	Distribution of gold by classes, %

VI.

Determination of the quantity of sulfides or concentrates in the ore.

For this determination, in the laboratory they use the method of washing in a bowl or in a scoop. The operation requires some dexterity and considerable experience and is carried out as follows.

A wide vessel, for example an enameled bucket or a glazed pot, is filled three-quarters with water. The height of the vessel must be sufficient for the free immersion of the bowl or scoop. The vessel is placed under a water tap. An ore sample of 500–1000 g, in small portions of 50–70 g, is stirred up with water in the scoop; the latter is held over the water in the wide vessel. By shaking and rotating the scoop, and by frequently immersing it in the water, they try to wash off the lighter particles of silt and sand. Gradually, as the latter are removed, the inclination of the scoop is increased. The remaining concentrates are washed into a separate vessel and the operation is repeated with the following portions of ore until the entire sample has been washed. The collected concentrates are finally washed, dried, weighed, and sampled. In this way the quantity of sul-

sulfides and the distribution of gold between the sulfides and the non-pyritic part of the ore, which is very important for deciding the question of the necessity of concentrating the ore and of separately treating the concentrates.

For the separation of sulfides one might use heavy solutions employed in mineralogical analysis, such as, for example, methylene iodide with sp. gr. 3.33, cadmium borotungstate with sp. gr. 3.28, etc. But the great cost of these reagents excludes the possibility of using them for the concentration of large quantities of material, for example, 1–2 kg, which is often necessary in view of the small content of sulfides in the material.

VII.

Determination of the amount of slimes in crushed ore.

It is often necessary to establish the amount of slimes, colloidal or crystalline, in ore crushed in a given apparatus. For this purpose laboratory concentrators are used, which, according to statements in the literature, give good results; but in our laboratory there are as yet no such concentrators. The simplest method of determining the amount of slimes, requiring no special apparatus, is the following. A weighed portion of the average ore sample is stirred with 4–5 times its quantity of water in a cylindrical vessel. After settling for 30–60 seconds, the turbid water with suspended slime is decanted into a special vessel. The operation of stirring and decantation is repeated with fresh portions of water until the decantate is free from suspended slime. The material washed free of slime is dried and weighed. After settling, the decantation waters are poured off; the coagulated slime is filtered off, dried, and weighed. The assaying of the sands and the slime will indicate the distribution of gold between these products.

VIII.

Determination of the maximum extraction in amalgamation.

The determination of the maximum extraction, i.e., the extraction in internal and external amalgamation, is carried out in the laboratory in a porcelain mortar 20–25 cm in diameter, or in a cast-iron ball mill.

The ore is crushed to 40–60 mesh, and from it an average sample is taken by quartering for the determination of the gold content. The total acidity of the crushed ore is preliminarily determined, i.e., the amount of alkali, usually CaO, that is necessary for the neutralization of sulfuric acid, both free and combined with metallic oxides in the form of sulfates. The sulfates of certain

metals, such as, for example, iron, copper, zinc, manganese, and others, present in oxidized or weathered ores, impair amalgamation by promoting the fouling of the mercury and its “pumicing,” or by causing, during amalgamation on plates, the formation of a deposit or film on the active mirror-like surface of the amalgamated plates. This substantially lowers the extraction of gold from the ores, as shown by the following experiments of F. Thomson and R. Keffer*.

For the experiments a quartz ore was used, crushed to 40 mesh and containing 2.49 ounces of free gold per ton (12.7 zolotniks per 100 poods). The pulp, with a liquid-to-solid ratio of 8:1, was passed over a copper plate amalgamated with pure mercury, repeatedly, up to three times, until the content in the amalgamation tailings no longer decreased. To the pulp were successively added \(\mathrm{FeSO_4}\), \(\mathrm{MnSO_4}\), and \(\mathrm{ZnSO_4}\) in an amount of 5% by weight of the dry ore. The mixture was subjected to amalgamation in the same way as the pure ore, and the following results were obtained:

Composition of the pulp	Gold extraction, %
Pure ore	76.9
Pure ore + 5% \(\mathrm{ZnSO_4}\)	55.2
Pure ore + 5% \(\mathrm{MnSO_4}\)	43.0
Pure ore + 5% \(\mathrm{FeSO_4}\)	43.4

These figures show how harmful the presence of soluble sulfates in the ore is.

The determination of total acidity is described below, in the chapter on cyanidation.

Into a mortar are poured 500–1000 g of the average sample of the crushed ore; \(\mathrm{CaO}\), calculated in advance, is added, and then, gradually from a graduated cylinder, enough water is added to obtain a thick pulp. The required amount of water depends on the character of the ore and is determined experimentally. The density of the pulp is of great importance for proper amalgamation. If the pulp is too thick and not very mobile, the mercury cannot be distributed uniformly throughout the whole mass, and it is impossible to achieve sufficient contact between the droplets of mercury and the grains of ore. If the pulp is liquid, the mercury quickly passes through the layer of ore and collects in separate places at the bottom of the bowl. By grinding the pulp it is not possible to disperse mercury that is not held by the ore.

The pulp is mixed with a pestle to a homogeneous consistency, and mercury is added to it in an amount of 5–10% by weight of the ore. The pulp is ground for two hours. After this, an excess of water is added to the mortar, and the liquid pulp is stirred so that the mercury can collect into large drops at the bottom of the bowl.

* Metallurg Chemic. Eng. Vol XIII, 1915, p. 367.

Then the ore is washed free of mercury on a scoop or on Richards’ round pan (see Fig. 1), trying, as far as possible, to wash out the finest drops of mercury. The latter is collected in a porcelain cup, washed with a stream of water, the water is carefully separated off by means of filter paper, and the mercury is dried at 60–70° and weighed. The loss in the weight of the mercury shows how carefully the washing has been carried out. The washed ore or tailings are separated from the water by decantation, filtration, and drying. After weighing, the tailings are mixed, screened through a 20–30 sieve, while all lumps are rubbed down. An average sample is taken from the tailings, ground to 120 or 150 apertures, and assayed. The gold content in the tailings is used to determine the extraction. The latter can also be determined from the quantity of gold that has passed into the amalgam from the given charge of ore. For this it is necessary to distill off the mercury or to dissolve it in nitric acid. These operations require much time, special apparatus, and are associated with inevitable and very serious losses of gold, if one takes into account that the concentration of the latter in the amalgam is very small.

Amalgamation in a ball mill. For amalgamation in the laboratory, the ball mill shown in Fig. 2 is used.

The procedure is as follows. Into the mill are charged 2–3 kg of ground ore, the calculated required quantity of CaO, and enough water to obtain a thick pulp in which drops of mercury could be retained. The quantity of water depends on the composition and character of the ore and varies from 20 to 40% of the weight of the ore. The quantity of water is best established experimentally; for this purpose 100 g of ore are placed in a mortar and, while stirring, water is added from a graduated cylinder until a pulp of the required consistency is obtained. The consumption of water is determined by noting the remaining quantity of it in the graduated cylinder.

Mercury is taken in an amount of 8 to 10% by weight of the ore. The mercury is poured from a small bottle, the neck of which is covered with canvas. Passing through the canvas filter, the mercury breaks up into the finest droplets and is uniformly distributed over the surface of the pulp. To mix the latter, two cast-iron balls are placed in the mill, and after closing the charging hatch the mill is set in rotation at a speed of 30–40 revolutions per minute for two hours. Then, through opening a, several liters of water are poured into the mill to dilute the pulp, and rotation is continued for approximately 5–10 minutes. In this way the greater part of the mercury collects and can be drained off by opening the plug.

Then the pulp is discharged into an enameled bucket or another suitable vessel and the mercury is washed on Richards’ pan.

If the weight of the washed mercury differs considerably from the weight charged, then the amalgamation tailings are washed again.

The tailings are separated from the water and subjected to assay, as described for amalgamation in a mortar. From the gold content in the tailings, the percentage extraction is calculated.

During amalgamation the mercury is often “sickened,” i.e., it is broken up into the finest droplets, having lost mobility and not coalescing into a single mass; this is explained by the formation, on the surface of the mercury droplets, of films of foreign substances.

Severe sickening of mercury is observed when sulfides of heavy metals are present in the ore. To revive such mercury, a little strong solution of KCN, NaOH is added to it and it is heated slightly. Sometimes the addition of nitric acid helps.

Such are the methods for determining the maximum extraction in the amalgamation of ore of a given degree of comminution.

Amalgamation on a sluice. To determine the extraction of gold in external amalgamation, the laboratory uses a copper amalgamated sluice, shown in Fig. 3. The whole process is an imitation of amalgamation on sluices.

Into vessel A, which serves as a pulp-former, water is poured and the pulverized ore and lime are charged in the required quantities. The pulp is stirred by a four-bladed agitator B, set in motion by a motor and making 600–800 revolutions per minute. The stirring does not cease during the course of amalgamation.

The pulp, through the neck of the vessel and a tube, passes onto the amalgamator. The velocity of the stream is regulated by a clamp; and for uniform distribution of the pulp over the plate, at the head of the latter an adjustable sill is arranged, by lowering and raising which the thickness of the pulp layer can be changed. The amalgamator consists of a copper plate 100 cm long, 36 cm wide, and 2 mm thick.

The preparation of the amalgamator is carried out as follows.*) The plate is annealed; then its edges along the long sides are bent at an angle of 90° so as to form rims 15–20 mm high. After removing the scale with emery paper, the plate is rubbed with fine sand until its surface becomes clean. Then the plate is rubbed with a moist mixture of 10 parts sand and 1 part ammonium chloride, washed with water, and this is repeated two or three times. After this, mercury is sprayed into the mixture of sand with ammonium chloride, and with this mixture the plate is rubbed until a mirror-like surface is obtained, which is washed with water. Then the amalgamated surface is wetted with a 2% KCN solution and the plate is left under the action of the latter for 12 hours; afterward it is washed with water and again rubbed with mercury until the surface

) R. Richards. Ore dressing*, vol. II, 753.

will become shiny and mirror-like. If, after standing for several hours, green spots begin to appear on the mirror surface of the plate, then the washing operation with subsequent rubbing with mercury is repeated, or the plate is rubbed with sodium amalgam.

The amalgamated plate is mounted on a frame made of two supports parallel to the length of the plate. The inclination of the frame, and consequently of the plate, can be varied as desired. Since this inclination depends on the density of the pulp, the size of the grains, and the composition and character of the ore, the suitable inclination of the plate for a given case can be established only experimentally. For a given pulp, the inclination must be such that the ore slime flows off in an even, thin layer, does not flood the plate, and does not allow accumulations of sulphides or other heavy particles to appear on it. The rate at which the pulp is fed to the amalgamator may vary from 1 to 2 liters per minute.

Procedure. Having set the ratio of liquid to solid, which in practice amounts to from 6 to 10 parts by weight of water to 1 part by weight of ore, the measured quantity of water is poured into the pulp-former or agitator, the stirrer is set in motion, and the ore is gradually charged. When the pulp has been prepared, it is allowed to run down the amalgamator, the stream being regulated by means of a clamp. The pulp that has run off is collected in a bucket, transferred to the agitator, and again passed over the plate. The operation is repeated from 5 to 10 times. After a certain number of passes, all tailings from the amalgamation from the given operation are collected, separated from the water, and the gold content in them is determined. In this way it is established after how many repeated passes over the plate the gold content in the tailings ceases to decrease. Then the amalgamation is repeated according to the data from the preceding experiments. From the gold content in the ore and in the tailings, the extraction attainable by external amalgamation under the given conditions is determined; these conditions, of course, can easily be varied. Thus, for example, one may increase the density of the pulp, the degree of grinding of the ore, add lime, etc.

To determine the most suitable degree of grinding, screen analyses are carried out and the gold is recorded by classes in the ore and in the tailings. From the results obtained it is possible to establish which of the classes gives the greatest percentage of gold and, consequently, which grain size is the most suitable for obtaining maximum extraction.

The duration of experimental amalgamation depends on the amount of pulp subjected to amalgamation in a single operation and on the velocity of the stream. Thus, in the laboratory, with an agitator capacity of 10 liters and a pulp-stream velocity of 1–2 liters per minute, a single passage of all the pulp over the plate takes 10–12 minutes. Owing to the small quantity of material and to the slow and calm flow,

of the process during amalgamation on a plate it is possible to make very valuable observations characterizing the behavior of the ore in this process.

Below are given examples of amalgamation illustrating the application of the laboratory methods of ore investigation described above.

IX

1. Amalgamation of ore from the Bogodarovannyi mine of the Asiatic Gold-Mining Company (Uryankhai Territory).

The ore consisted of pure white quartz without visible foreign inclusions. The laboratory was set the task of determining the maximum extraction of gold from this ore by amalgamation.

The entire sample of ore delivered to the laboratory, in the amount of 8580 g, was crushed to pass an 80-mesh I. M. M. sieve. During screening, forged particles containing visible gold remained on the sieve. The material that did not pass through the sieve was carefully collected, weighed, and assayed. In this material, whose total weight was 11.5 g, 143.4 mg of gold was found; referred to the original weight of the sample (8580 g), this gives 1.7 mg of gold per 100 g of ore, or 6.53 zolotniks per 100 poods. The gold content in the screened ore proved to be 26.88 zolotniks per 100 poods. Consequently, in 100 poods of ore there was 6.53 zolotniks of coarse gold, 26.88 zolotniks of fine gold, in all 33.41 zolotniks.

Amalgamation. Into a porcelain mortar were placed 1400 g of the average ore sample; 400 cubic cm of water (29% by weight of the ore) and 168 g of mercury (12% by weight of the ore) were added. The pulp was stirred for two hours; then it was diluted with water and washed in a launder. The amount of washed tailings was 1360 g; consequently, 40 g of ore, or 2.9% of it, was lost. The content in the tailings proved to be 0.58 zolotniks per 100 poods.

Thus, the maximum extraction of gold by amalgamation was

$$ \frac{26.88 - 0.58}{26.88}\cdot 100 = 98\%. $$

The extraction is referred to the gold content in the screened ore, without taking into account the amount of coarse gold that remained on the sieve; in amalgamation on an industrial scale, it is clear, this gold too would be extracted. A content of 0.58 zolotniks of gold per 100 poods of tailings would not justify their further treatment under local conditions.

The ore investigated represents an example of a freely amalgamating ore, from which almost all the gold can be extracted by means of a single amalgamation process.

2. Amalgamation of ore from the Varvarinsky mine in the South Yenisei mining district.

From the Varvarinsky vein, 200 poods of ore were sampled on site. By gradual crushing and reduction, the sample was brought down to 1.5 poods, or 25 kg, with a maximum piece size of up to 12 mm.

The ore consisted of solid quartz of a dark-gray color, with rare inclusions of pyrite, galena, and sphalerite. In the laboratory the sample was crushed in a Blake crusher to 5 mm and reduced on a Jones divider to 7.5 kg, which were ground to −40 mesh, since it was decided to carry out amalgamation at this degree of grinding. Further, the sample was again reduced to 1980 g, and then 200 g of material were taken for the determination of gold. These 200 g were ground to −150, thoroughly mixed, and from a charge of 50 g of ore the gold was determined; its content proved to be 11.26 zolotniks per 100 poods of ore.

Sieve analysis and fixing of gold by classes. For the sieve analysis the second half of the last reduced sample was taken, after deducting the 200 g used for fire assays, i.e., 1780 g.

For the analysis, sieves 60, 80, 100, and 120 were used. Each class was examined under the microscope, then ground to −150 and assayed by crucible test.

The results of the sieve analysis and of the fixing of gold by classes are given in Table V.

Table V. Sieve analysis of the ore and fixing of gold by classes.

CHARGE 1780 g.

Average gold content: 11.26 zolotniks per 100 poods.

Classes	Quantity, g	Quantity, %	Gold content in 100 poods, zolotniks	Total amount of gold in the class, zolotniks	Distribution of gold by classes, %
+60	520	29.8	20.08	6.20	55.10
+80	240	13.8	9.79	1.35	11.99
+100	155	8.9	7.49	0.66	5.84
+120	130	7.5	8.06	0.60	5.33
−120	700	40	6.14	2.45	21.74
Total	1745	100%	11.26	11.26	100%

The amount of gold which each class contributes in 100 poods of ore is determined by multiplying the gold content in 100 poods, co-

of the corresponding class by the percentage of this class in the sample, i.e., alternately multiplying the figures of columns 3 and 4 and dividing the product by 100.

Microscopic analysis. Each class was examined under the microscope separately, at a magnification of 60–100.

Class \(+60\). Crystalline grains of quartz. Sulfides are not separate, but have intergrown with the quartz particles. Gold is clearly visible in the form of fine dust on the quartz grains. When the material is moistened with hydrochloric acid, \(CO_2\) is liberated from the dark-colored limestone grains.

Class \(+80\). Crystalline grains of quartz. Few sulfides. Individual quartz grains are covered with specks of gold.

Classes \(+100\) and \(+120\). The picture is the same as in classes \(-60\) and \(+80\).

Class \(-120\). The picture is unclear, owing to the fact that almost all grains are covered with dust of iron oxides. The material is crystalline.

The predominance of crystalline quartz grains without noticeable inclusions, the presence of free gold, and the very small quantity of sulfides—all this indicates that extraction by amalgamation promises to be quite satisfactory.

Qualitative chemical analysis revealed the presence of \(Pb\), \(Fe\), \(As\), \(Mg\), \(Ca\), and \(S\) in very small quantities. Assuming that all the sulfur in the ore is combined with iron in the form of pyrite, which, according to the indications of the microscopic analysis, predominates among the sulfides, it was determined from the amount of sulfur in the ore that the latter contains \(1\%\) \(FeS_2\), in view of which separation of the sulfides appeared unnecessary.

Determination of the maximum extraction in amalgamation. This determination was carried out in a ball mill. The total acidity of the ore was \(0.5\%\) \(CaO\). Into the ball mill were charged 2700 g of ore, crushed to \(-40\); 20 g \(CaO\); 680 cubic cm of water, which, according to preliminary determinations, gave a pulp of suitable consistency; and 220 g of mercury. Two balls were placed in the mill, and amalgamation lasted 2 hours. After dilution with water, the pulp was discharged, and the ore was twice washed free of mercury on a Richards blanket. The amalgamation tailings were dried, screened through a No. 20 sieve, weighed, reduced, and assayed. From the tailings, 1500 g were taken for sieve analysis and for determining the gold by classes.

Results of amalgamation:

Item	Quantity
ore charged	2700 g
tailings obtained	2630 g
ore lost	70 g \(= 3.4\%\) of the total charge
mercury added	220 g

washed off as amalgam . . . . . 215 gr.
loss of mercury . . . . . . . . 5 “ = 2.3% of the amount charged

amount of gold in 100 poods of ore . . . 11.26 zolotn.
“ “ tailings . . . 3.07 “
gold extracted from 100 poods . . . 8.19 “
% extraction . . . . . . . . . 72.7%.

Table VI gives the results of sieve analysis of the tailings and the accounting of gold in them by classes.

TABLE VI. Sieve analysis of tailings and accounting of gold by classes.

Charge—1500 gr., gold content—3.07 zolotn. in 100 poods.

Classes	Quantity, gr.	Quantity, %	Gold content in 100 poods, zolotn.	Total amount of gold in class, zolotn.	Distribution of gold by classes, %
+ 60	327.5	23.3	2.84	0.66	24.3
+ 80	250.0	17.8	3.34	0.59	21.7
+ 100	154.5	11.0	2.69	0.29	10.7
+ 120	121.0	8.6	2.78	0.24	8.9
− 120	554.0	39.3	2.38	0.93	34.3
Sum . . .	1470	100%	3.07	2.71	100%

Table VII compares the sieve analyses and the distribution of gold by classes in the ore and in the tailings from amalgamation.

TABLE VII. Sieve analyses and extraction of gold by classes during amalgamation in a ball mill.

Classes	Quantity, %, before amalgamation	Quantity, %, after amalgamation	Gold content in 100 poods of each class, before amalgamation	Gold content in 100 poods of each class, after amalgamation	Gold extracted from each class, zolotn. per 100 poods	Gold extracted from each class, %
+ 60	29.8	23.3	20.08	2.84	17.24	85.1
+ 80	13.8	17.8	9.79	3.31	6.45	65.9
+ 100	8.9	11.0	7.49	2.69	4.80	64.0
+ 120	7.5	8.6	8.06	2.79	5.27	65.4
− 120	40.0	39.3	6.14	2.38	3.76	61.3
Sum . . .	100	100	11.16	3.07	8.19	72.7

From the data of Table VII we see that the most suitable grain size for amalgamation must be regarded as \(-40+60\), since the \(+60\) class showed the greatest extraction of gold. Finer grinding of the ore does not increase the extraction and, consequently, is superfluous. A certain increase of the \(+80\) and \(+100\) products in the amalgamation tailings must be attributed to the additional grinding in the ball mill of the coarser classes.

Analysis of the data of Table V indicates that the gold in the ore is coarse, for the greater part of it is associated with the \(+60\) and \(+80\) classes, amounting in both classes to 67% of the total gold content in the ore.

External amalgamation on a plate. 4.5 kg of the average sample, ground to \(-40\) apertures, were divided on a Jones splitter into four approximately equal parts.

Water was poured into the pulp-former in such quantity that the ratio of liquid to solid in the pulp was equal to 10:1; ore was charged and calcium oxide was added in the quantities shown in Table VIII. The pulp from the first quarter of the sample was passed over the plate five times, from the second quarter—seven, from the third—eight, and from the last—nine times. The inclination of the plate was 1 dm per foot of plate length, or \(1/12\); at this inclination the plate did not flood, and the ore slime flowed in a calm stream.

The number of revolutions of the agitator was 700–800 per minute. In all experiments the working conditions remained identical, with the exception of the number of passages of the pulp over the plate. The tailings from each operation, comprising the above-mentioned number of passages over the plate, were collected, dried, and assayed.

The results of amalgamation on the plate are given in Table VIII.

TABLE VIII. Results of amalgamation on a plate.

Number of passages over the plate	Duration of one complete amalgamation, min.	Per minute passing over the plate, liters of pulp	Weight of ore, g	Quantity of water, cm³	CaO, g	CaO, % by weight of ore	Weight of tailings, g	Ore lost, g	Ore lost, %	Gold content in 100 poods of tailings, zolotniks	Gold extraction: from 100 poods of ore, zolotniks	Gold extraction: in %, of the gold content in the ore
5	48	1.26	1130	1100	10	0.88	1070	50	4.4	3.53	7.73	68.70
7	43	1.46	1060	1060	8	0.75	1020	40	3.4	2.84	8.42	74.80
8	44	1.41	1050	1050	8	0.76	1010	40	3.5	2.47	8.79	78.10
9	50	1.70	1170	1200	8	0.68	1125	45	3.8	2.46	8.80	78.15

Experiments in amalgamation on a plate show that after the pulp has been passed over the plate eight times, the extraction does not increase, and the maximum extraction with external amalgamation is 78%. The gold content in the tailings is thereby reduced to 2.47 zolotniks per 100 poods, which indicates the necessity of further treatment of the tailings by cyanidation.

The extraction of gold in external amalgamation proved to be 5.4% (78.1–72.7) higher than in amalgamation in a ball mill. This may be explained by the fact that, when a thin stream of ore pulp moves over the plate, more complete contact is achieved between the particles of gold and mercury than in amalgamation in a ball mill, where the conditions of contact are less favorable for the dissolution of gold in mercury. However small the drops of mercury dispersed in the pulp may be, the density and low mobility of the latter prevent each particle of ore from coming into contact with the mercury, as takes place on the plate. Moreover, in the ball mill, owing to the grinding action of the balls and ore, the mercury becomes coated, and in this way the active surface of the latter is greatly reduced. On the plate, the surface of the mercury remains at all times bright and active.

The results of the investigation of the ore and of its amalgamation make it possible to draw the following conclusions.

1. The ore belongs to the category of freely amalgamating ores.
2. The gold in the ore is coarse.
3. Concentration of the ore before or after amalgamation is unnecessary, since the quantity of sulphides is very small.
4. The most suitable conditions of amalgamation for this ore are the following:
   a) internal amalgamation is not needed;
   b) a sufficient degree of pulp density will be 10:1;
   c) the reaction of the pulp during amalgamation should be slightly alkaline, corresponding approximately to 0.5% CaO by weight of ore.
5. The maximum extraction of gold under the above conditions in external amalgamation is 8.8 zolotniks per 100 poods, or 78% of the gold content in the ore (11.26 zolotniks per 100 poods).
6. The tailings from amalgamation, as still containing 2.5 zolotniks per 100 poods, are subject to further treatment with the aim of extracting part of the gold contained in them.

X.

Cyanidation.

General remarks. In the vast majority of cases, cyanidation of gold-bearing ores is preceded by amalgamation, because

that ores are very rarely encountered after amalgamation of which there would not still remain in the tailings enough gold for them to be profitably treated at cyanide plants; much more often the metallurgist has to deal with ore which it would be unprofitable to amalgamate, either owing to the insignificant extraction of gold, or because of the excessive difficulties with which amalgamation is associated (high specific gravity of the ore, loss of mercury, etc.). It should also be noted that the cyanide method of extracting gold is universal. The modern technology of cyanidation stands so high that, without risk of falling into error, one may assert that by means of the cyanide process and its modifications, gold can be extracted from any ore, which cannot be said of amalgamation. If the latter is nevertheless so widely used, the reason for this must be seen not in the impossibility of extracting amalgamable gold by cyanidation, but in the easy capture of a more or less considerable part of it (chiefly coarse gold) and in its more rapid realization, as was stated above.

As is known, theoretically the cyanide process appears very simple and is entirely contained in only two chemical equations:

1) dissolution of gold in potassium cyanide:

\[ 2Au + 4KCN + O + H_2O = 2KAu(CN)_2 + 2KOH \]

and

2) precipitation of gold from solution by means of metallic zinc:

\[ 2KAu(CN)_2 + Zn = K_2Zn(CN)_4 + 2Au. \]

In the practical realization of the process, however, a number of difficulties arise: first, because of the comparatively low stability of the reagent itself, which changes with time in aqueous solutions with the formation of a whole series of more or less complex organic compounds, such as, for example, potassium formate,

\[ KCN + 2H_2O = HCOOK + NH_3, \]

potassium cyanate,

\[ 2KCNO + O_2 = 2KCNO, \]

which decomposes according to the equation:

\[ 2KCNO + 3H_2O = K_2CO_3 + CO_2 + 2NH_3 \]

and many others; secondly, owing to the properties of hydrocyanic acid—an exceedingly weak acid, the salts of which hydrolytically decompose in aqueous solutions with the formation of caustic alkali and free acid:

\[ KCN + H_2O = KOH + HCN \]

and which is readily displaced from salts both by the carbonic acid of the air and by acids formed during the weathering of ores (\(H_2SO_4\)); in aqueous solutions, hydrocyanic acid rapidly passes into other

organic compounds, polymerizing and becoming resinified;*) thirdly, because of the tendency of potassium cyanide to form complex compounds with most heavy metals and their salts, such as Au, Ag, Cu, Fe, Mn, Ni, Co, Zn, Hg, Cd, for example:

\[ \mathrm{CuSO_4 + 4KCN = K_2Cu(CN)_4 + K_2SO_4} \]

\[ \mathrm{ZnSO_4 + 4KCN = K_2Zn(CN)_4 + K_2SO_4,} \]

and so forth; fourthly, as a consequence of the great diversity both in the size and form of the gold particles and in the character of the minerals with which it is associated in ores. The factors listed above, which affect the results of the cyanide process and its course, being numerous in themselves, give an infinite number of varied combinations, the probability of whose recurrence is very small. Hence it becomes clear that every ore entering treatment is the bearer of a combination of properties unique in its kind, which exert an influence on the course of the process.

The individual features of an ore naturally require an individualization of the methods of its treatment, which is possible owing to the great flexibility of the cyanide process, allowing it to be modified according to the metallurgist’s wishes. For this reason, no ready-made schemes borrowed from the rich experience accumulated in the literature on gold metallurgy are applicable in their entirety when deciding the question of a method for treating any new material. In the treatment scheme on the basis of which a plant is designed, all the individual features of the material, ascertained through a comprehensive study of its properties, must be taken into account. Neglect of this elementary rule—unfortunately very widespread among domestic entrepreneurs—proves each time to be very costly, many times more costly than a preliminary investigation would have cost.

The study of the properties of gold-bearing material, the determination of the conditions under which the cyanide process, at the least expense, gives the most complete extraction of gold—all these are tasks whose solution is most readily approached by means of laboratory methods of investigation. Thus, work on an industrial scale must be preceded by the study of the material in the laboratory.

Of course, the question is appropriate: is it possible, on the basis of data obtained in the laboratory, to conclude that the operation of a plant under the same conditions will give the same results? This question may be answered in the affirmative, since the cyanide process, with all its modifications, is easily reproduced in the laboratory, and,

) The frequently encountered statement in the literature that hydrocyanic acid is rapidly volatilized from dilute aqueous solutions does not correspond to reality, as shown by E. Kühn’s experiments described in the dissertation “Die chemischen Vorgänge bei der Cyanlaugung von Silbererzen,”* Halle a. S., 1912.

consequently, to obtain results close to those which, under similar conditions, would be obtained at the plant. Of course, there are questions whose complete elucidation in the laboratory is either difficult or altogether impossible. The most correct solution of such questions could be found at an experimental plant of small capacity, the construction of which precedes the creation of production on a large scale, as is now practiced by most newly arising large enterprises in the United States of North America and in other industrial countries. For the cyanide process as such, however, these questions are of secondary importance, and the impossibility of resolving them in the laboratory in no way diminishes the importance of the preliminary investigations, to the exposition of whose methods we now turn.

Tasks of the laboratory investigation. Laboratory investigations of gold-bearing ores or materials (tailings from amalgamation, sands, slimes, concentrates) subject to cyanidation have as their aim the establishment of the following:

1. Gold content.
2. Chemical and mineralogical composition of the material.
3. Maximum extraction of gold.
4. The most advantageous degree of comminution.
5. The most advantageous strength of the solution.
6. The most advantageous ratio of liquid to solid.
7. The most advantageous duration of treatment.
8. The most advantageous alkalinity of the solution.
9. Applicability of percolation.
   Percolation rate.
   Porosity of the material.
   Amount of moisture retained in the tailings.
   Extraction of gold.
   Consumption of potassium cyanide.
10. Extraction of gold by means of an agitation test.
11. Necessity of concentration.
12. Acidity of the material, or its alkalinity.
13. Physical constants:
    specific gravity,
    bulk density.
14. Necessity of auxiliary operations:
    washing with water,
    oxidation.

XI.

Determination of the Gold Content.

In determining the gold content in the material under investigation,

usually always precedes cyanidation; to determine the maximum extraction of gold, the material is taken as it is. In the second case, it is necessary to settle upon some definite degree of comminution which a priori might be considered sufficient. As we shall see below, in our laboratory material passing through a 150 sieve is regarded as sufficiently comminuted. The questions, however—whether the comminution is sufficient in the first case and what is the most advantageous maximum particle size in the second—are resolved by means of separate tests, described below.

In carrying out the actual procedure for determining gold extraction, the so-called bottle test, there is far from that uniformity which would be desirable.

Turning to the literature in the Russian language, we find the only indication that the determination of “the amount of gold extracted from tailings by means of potassium cyanide” consists in the following: “a charge of not less than 500 grams of an average sample of tailings is shaken for 2–3 hours in an ordinary bottle, stoppered with a cork, with a solution of potassium cyanide of strength from 0.15 to 0.5%, taken in a number of cubic centimeters equal to half the gram weight of the charge,” after which the gold content is determined in the separated solution, and from it the extraction is judged.*)

The method described suffers from substantial shortcomings: first, the application of the method is limited to tailings, since no indication is given as to how ores are to be investigated, to what grain size they must first be comminuted; second, it may be doubted that within the indicated interval of time, 2–3 hours, the gold soluble in potassium cyanide will in all cases have time to dissolve; third, the duration of the treatment, a most essential factor, is not precisely fixed.

I. E. Clennell, in his book**), proposes, as a standard preliminary test serving as the basis for further investigations, the following method: from material previously comminuted to 40 mesh a 100 g sample is taken, which is then treated, with continuous agitation for 16 hours, with 100 cubic centimeters of a 0.5% solution of potassium cyanide with the addition of calcium oxide. The extraction of gold is determined both from its content in the tailings and by assay of the solution.

It is obvious that the method proposed by Clennell does not make it possible to determine what amount of gold can be extracted under the most advantageous conditions, since the degree of comminution recommended by him cannot be considered sufficient. This is confirmed

) M. Rybalkin, “The Cyanide Method of Gold Extraction.” Vestnik Zolotopromyshlennosti*, 1904, vol. XIII, p. 248.

) I. E. Clennell, The Cyanide Handbook, 1910, p. 490.

usually always precedes cyanidation; in order to determine the maximum extraction of gold, the material is taken as it is. In the second case one has to stop at some definite degree of comminution, which a priori might be considered sufficient. As we shall see below, in our laboratory material passing through a 150 sieve is considered sufficiently ground. The questions as to whether the grinding in the first case is sufficient, and what is the most advantageous maximum size of the particles in the second, are decided by means of the separate tests described below.

In carrying out the very procedure for determining gold extraction, the so-called bottle test, there is far from that uniformity which would be desirable.

Turning to the literature in Russian, we find the sole indication that the determination of “the quantity of gold extracted from tailings by means of potassium cyanide” consists in the following: “a charge of not less than 500 grams of an average sample of tailings is shaken for 2–3 hours in an ordinary bottle, stoppered with a cork, with a solution of potassium cyanide of strength from 0.15 to 0.50%, taken in a number of cubic centimeters equal to half the gram weight of the charge,” after which the gold content is determined in the separated solution, and from this the extraction is judged*).

The method described suffers from essential shortcomings: first, the application of the method is limited to tailings, since no indication is given as to how to conduct the investigation of ores, or to what grain size they must first be ground; secondly, it may be doubted that, within the indicated interval of time, 2–3 hours, gold soluble in potassium cyanide will in all cases have time to dissolve; thirdly, the duration of the treatment, a most important factor, is not precisely fixed.

I. E. Clennell, in his book**), proposes the following method as a standard preliminary test serving as the basis for further investigations: from material previously ground to 40 mesh, a sample of 100 g is taken, which is then treated, with continuous shaking for 16 hours, with 100 cubic centimeters of a 0.5% solution of potassium cyanide, with the addition of calcium oxide. The extraction of gold is determined both by its content in the tailings and by assaying the solution.

It is obvious that the method proposed by Clennell does not make it possible to determine what quantity of gold can be extracted under the most advantageous conditions, since the degree of grinding recommended by him cannot be considered sufficient. What has been said is confirmed

*) M. Rybalkin. The cyanide method of extracting gold. Vest. Zolotopr., 1904, vol. XIII, p. 248.
*) I. E. Clennell. The Cyanide Handbook*, 1910, p. 490.

practice at Rand, where it has been established that the greater the proportion of all the material entering the percolation vats that passes through a 90 sieve, the higher the percentage of gold extraction. Therefore, at Rand it is considered that normally 66–70% of all the material should pass through a 90 sieve. Special investigations carried out there showed that 43.5% of the gold remaining in the tailings after percolation falls in the +90 class.) Thus, practice indicates that for the percolation process, which generally gives an extraction lower than that attainable under more favorable conditions (for example, with the slime process), grinding is required that is considerably finer than that adopted by Clennell for the samples he recommends.

At Rand several methods have also been developed for solving the question that interests us.

Since Johnston**) gives no indication of what the maximum size of the particles of the material under investigation should be. This omission is explained by the fact that at Rand all ores are preliminarily amalgamated, and the investigator has to deal exclusively with tailings, which he takes as such, without concern for further comminution.

The most rapid results are obtained by treating 250 grams of tailings with 0.25 g CaO and 500 cu. cm of KCN solution of 0.25% strength, in a bomb of 2-liter capacity under a pressure of about 7 atmospheres (100 lb per sq. inch) of air or oxygen. The bomb is shaken continuously for 6 hours, which is sufficient for extracting all the gold that can dissolve.***) This method requires special apparatus, not usually at hand, and therefore cannot claim wide adoption.

Instead of treating the material under the action of compressed air or oxygen, at Rand barium peroxide is added to the potassium cyanide solution. By applying this oxidizer the same goal is achieved—the acceleration of gold dissolution—as was pursued by increasing the pressure of oxygen.

The determination of the maximum extraction of gold may in this case be carried out by two methods.****)

1. 250 g of material, 0.25 g of barium peroxide \((\mathrm{BaO}_2 \cdot 8\mathrm{H}_2\mathrm{O})\), 0.25 g of calcium oxide, and 500 cu. cm of potassium cyanide solution of 0.25% strength are left to stand in an open vessel, shaking from time to time, for 24 hours. Then the clear solution is de—

*) Rand Practice, Vol. I, p. 169.
*) Author of the chapter “On testing,” Rand Practice, vol. I, p. 350–351.
*) It is interesting to note that, essentially, the treatment of ore at increased pressure is the basis of a process patented by “The Koering Cyaniding Process” Company and already used in practice at one time.
***) Rand Practice, vol. I, p. 350–351.

decanted, a fresh portion of KCN solution (500 cubic cm), with the same amount of BaO·2H₂O, is added, and it is left to stand for another 24 hours. The extraction is determined from the tailings, after first washing them free of dissolved gold.

1. A sample prepared in the same way as in the first method is agitated continuously for 24 hours, the stopper being opened from time to time. With continuous agitation, the dissolution of the gold proceeds, of course, more rapidly, and therefore the duration of treatment is reduced by half in comparison with the first method.

In the literature there are still a number of directions concerning the performance of the bottle test, differing from one another just as much as those given above. We shall not dwell on them, since what has already been set forth sufficiently confirms that practice has not yet developed strict standards for determinations of this kind.

In our laboratory we more than once had to encounter difficulties due to the circumstance indicated; it was therefore necessary to establish these standards for ourselves and to be guided by them in all determinations of this kind. In order that these standards should not be arbitrary, and that determinations of gold extraction carried out according to them should give figures truly close to the maximum attainable under the most favorable laboratory conditions, a series of corresponding experiments was set up to establish them.

The following materials were used for the experiments:

1. Slimes from the Ioannovsky mine, Yenisei Province.
2. Slimes from the Kochkarsky mine, Orenburg Province.
3. Tailings from laboratory amalgamation of ore No. 4.
4. Slimes from the Udaloi mine, Semipalatinsk Region.
5. Tailings from laboratory amalgamation of ore from the Varvarinsky claim, Yenisei Province.
6. Slimes from the Berikulsky mine, Tomsk Province.

A series of samples of 400 g, taken from the first three materials, was treated with 600 cubic cm of KCN solution (0.2%) in the presence of 0.4 g CaO for 1, 3, 6, 9, and 12 hours. The extraction of gold was determined from the solutions. The results of the experiments are brought together in Table IX.

TABLE IX. Effect of treatment duration on extraction.

Slimes	Ioannovsky	No. 4	Kochkarsky
Gold content in 100 poods	4.17	2.21	1.47
Agitation 1 hour, extracted %	45.87	17.36	18.13
Agitation 3 hours, extracted %	57.34	20.83	35.20
Agitation 6 hours, extracted %	63.78	20.83	40.00
Agitation 9 hours, extracted %	70.18	26.04	44.00
Agitation 12 hours, extracted %	76.15	26.04	45.33
Maximum extraction %	78.34	34.72	49.34
Extraction in the percolator %	83.21	50.63	—

ore. The extraction of gold from the Elizavetinskaya ore, in the form in which it was carried out by M. Rybalkin (see p. 31), is wholly unsuitable for determining the maximum extraction. In all three cases the curves showing the relation between treatment time and extraction rise rather sharply between the 1st and 3rd hours of treatment; then the rate of dissolution falls, and the curve rises more slowly, not reaching the maximum possible extraction even with 12-hour treatment. It is evident that the figures obtained for gold extraction when the material was treated for 2–3 hours, as recommended by M. Rybalkin, are entirely fortuitous.

The maximum extraction, also plotted on diagram I, was determined by a method similar to that adopted at Rand (see p. 32). Samples of 200 g (for the poor Kochkar tailings, 400 g) were agitated for 9 hours with 400 cc (for Kochkar tailings, 600 cc) of a 0.25% KCN solution in the presence of 0.2 g CaO and 0.2 g BaO₂·8H₂O. The samples were then left at rest for 14 hours, after which agitation was continued for another 3 hours. For each of the three materials studied, two samples were taken. The extraction was determined from the solutions.

Whether, under such conditions, the maximum extraction is actually attained was a question to which direct experiments with a laboratory percolator—whose results are given in Table IX and in diagram I—gave a negative answer. The extraction in the percolator proved to be greater than that given by the “maximum” extraction. The determination of extraction by means of the laboratory percolator is described below.

Unfortunately, because the materials used for the experiments described had been consumed, it was not possible with them to clarify definitively the question that had been posed.

The method of determining the maximum extraction of gold by means of successive double treatment of the sample with KCN solution in the presence of barium peroxide and calcium oxide, proposed at Rand and described on p. 32, likewise proved not entirely satisfactory, as is shown by experiments with tailings from the Udaly mine, the results of which are presented in Table X on p. 34.

From the figures given it is clear that 22-hour continuous agitation gives a greater extraction than treatment of the material for 48 hours with standing and shaking only from time to time. Thus, continuous agitation of the samples appears to be the most advisable procedure for attaining the possible maximum dissolution of gold.

There arises, however, the question: what should be the duration of agitation that might be adopted as a standard more or less suitable for all ores?

TABLE X. [[unclear: title]]
(gold content: 1,500 fractions per 100 poods of ore)

No.	Treatment	Hours	Tailings, g	KCN solution, cu. cm	KCN content, %	CaO, g	BaO₂H₂O, g	Extraction, %, by tailings	Extraction, %, by solution
1	Agitation	22	300	300	0.49	2.5	—	66	52.7
[[unclear]]	[[unclear]]	24	[[unclear]]	400	0.25	2	1	[[unclear]]	[[unclear]]
2	On standing	24	400	400	0.25	—	1	52.6	52.6
2	On standing	48	400	400	0.25	—	1	52.6	52.6
2	On standing	24	400	400	0.25	—	1	52.6	52.6
3	” ”	24	400	400	0.25	2	1	52.6	52.6
3	” ”	48	400	400	0.25	—	1	52.6	52.6
3	” ”	24	[[unclear]]	400	0.25	—	1	[[unclear]]	[[unclear]]

In some cases, as the figures of Table XI show, the maximum extraction is attained in a comparatively short time.

TABLE XI. Limit of duration of treatment.

Duration of treatment, hours	1	6	12	18	24	36	Maximum
Tailings of the Bernikul mine, extracted, %	60.39	73.99	71.55	72.91	72.91	72.91	76.20
” of the II Ioannovsky mine, ” ”	—	—	66.36	87.87	92.42	92.42	93.94
” of the Varvarinsky mine, ” ”	—	28.00	40.00	—	54.00	66.00	72.60
” of the Udaly mine, ” ”	—	—	—	—	92.10	—	—

For example, for the Bernikul tailings 6 hours is sufficient; for the tailings of the Udaly and Ioannovsky mines 24 hours are required; but for the tailings from amalgamation of ore from the Varvarinsky mine even 36 hours of agitation is not sufficient for all the soluble gold to be extracted.

On the basis of these data it may be said that the limit of the duration of treatment for achieving maximum extraction should lie beyond 36 hours, but should not greatly exceed this period, since in most cases the maximum extraction has been attained, or is very close. See Diagram II.

The foregoing served as sufficient grounds for adopting in our laboratory, as the standard in determining maximum extraction, 48-hour agitation. As was mentioned above, tailings from amalgamation are tested without any preliminary grinding; the ore subject to testing, however, is ground to 150.

Determination of maximum extraction is carried out as follows: into a glass vessel of 1½–2 liters capacity, measure 500 cu. cm of a 0.3% KCN solution; add to it 0.5 g CaO (0.2%) and 0.5 g BaO₂H₂O (0.2%); after stirring, add 250 g of the material to be tested.

…which are placed in a rotating apparatus of some kind, such as an agitational stand. In Fig. 4 is shown the agitator used in the laboratory, consisting of two wooden disks fastened at the ends of a horizontal axle; by means of a hinged bracket, 6 tumblers can be attached to each disk, the sizes of which may be quite varied. The agitator is set in motion by a caloric engine of \(1/10\) h.p., at a speed of 20–25 revolutions per minute.

It is recommended, after 24 hours have elapsed, to remove the samples and open the tumblers for several minutes, after which the tumblers are sealed again and agitation continues for another 24 hours.

With very poor ores, the charge is taken twice as large, with a corresponding increase in the quantity of reagents.

After the agitation is finished, the tumblers are removed and left to settle for 5–10 minutes; after this, the quantity of liquid needed for the determination of gold can be filtered off very quickly, by decanting the clear solution through a dry filter into a dry beaker.

Determination of the extraction of gold, both in this case and in the others, is more convenient and more correct when carried out on the solution, since in this way the troublesome work of washing the tailings is entirely eliminated; the results obtained agree with one another better than can be achieved by determining the extraction from the tailings. The point is that from solutions it is very easy, after mixing them, to take a sample whose gold content corresponds exactly to the entire mass, whereas the tailings either must be fused in their entirety (which is too much for crucible fusion), or else must first be finely pulverized and charges selected with all possible care, by quartering or dividing into squares. But in the latter case, the chances that the sample will be completely correct are still smaller than when solutions are sampled.

Determination of gold in solutions. Usually, for determining gold, 200 cubic cm of solution are taken, corresponding to 100 g of material. The solution is measured with a graduated cylinder or, more accurately, with a 200 cubic cm volumetric flask, graduated for delivery.*)

The measured quantity of solution is placed in a beaker of 400 cubic cm capacity; 3 g of zinc dust, about 1 g of potassium cyanide, and 5–6 drops of a 10% solution of lead acetate are added. The beaker is placed on a sand bath and heated to boiling.

As is known, zinc precipitates gold from solution according to the equation:

\[ 2\mathrm{KAu(CN)_2}+\mathrm{Zn}=\mathrm{K_2Zn(CN)_4}+\mathrm{Au}. \]

*) Such flasks have two marks on the neck, of which the lower indicates the volume when filling, and the upper when delivering.

When lead acetate is added, the lead is displaced from the solution by zinc:

\[ \mathrm{Pb(CH_3COO)_2 + Zn = Zn(CH_3COO)_2 + Pb} \]

and forms with the latter a couple that electrolytically promotes the precipitation of gold. The precipitation of gold is also accelerated by the addition of potassium cyanide.

The assay is boiled for 20 minutes, stirring from time to time, after which 25 cubic cm of a 10% solution of lead acetate is poured into the assay; it is stirred, and after 3–4 minutes 15 cubic cm of concentrated hydrochloric acid are added.

When lead acetate is added, according to the equation given above, zinc dust displaces lead, which separates out in the form of a loose spongy mass, in most cases forming, after the addition of hydrochloric acid and boiling, a single lump.

Boiling is continued until the excess zinc dissolves and the solution becomes clear, after which it is decanted through a filter of lead foil and assay lead. Such a filter is obtained by rolling, in the same way as an ordinary paper one, a piece of thin lead foil 6×6 cm and straightening it in a glass funnel. Into the cone thus obtained, 15 g of fine assay lead are poured and 10 small holes are pierced with a needle at its apex. Through such a filter the solution filters with great speed, especially if filtering is done under vacuum. The lead sponge is washed by decantation, transferred to the filter, and pressed with a glass rod. Then a piece of silver for quartation is placed in it, the edges of the filter are bent inward, and, placing it on a scorifier, 10 g of assay lead, 1–2 g of borax, and ¼ g of powdered glass are added; after this the scorifier is roasted, cupelled, and the bead obtained after parting is dissolved and the remaining gold is weighed. From the weight of the latter it is not difficult to calculate the percentage extraction, if the gold content in the material is known.

The results of determinations of the maximum extraction by the method described are placed in the last column of Table XI.

XIII.

Determination of the Most Advantageous Conditions of Treatment

The most advantageous degree of grinding. As is known, the size of the particles plays an outstanding role in the cyanide process. The smaller the particles, the more easily contact of the gold with the solution is achieved, the sooner and more completely extraction proceeds, and the more perfect the process as such. But grinding is, in gold work, the most expensive operation: the finer the final product of grinding, the more complicated the installation and the greater the expenditures for wear of parts.

In relation to this, the question now under consideration was, as is often the case, posed as a requirement of production technique—to grind the material as finely as possible. This faces the greatest possible limitation of the degree of grinding, dictated by economic considerations.

Consequently, the metallurgist’s task is reduced to finding that point in production at which further grinding is not paid for by the value of the gold extracted as a result of this additional grinding. It is obvious that the degree of grinding corresponding to this point is the most advantageous.

In order to be able to establish this degree rigorously, one must have at hand all the data, both of a technical and of an economic character, such as: the cost of power, the type of machinery, its wear, the amount of wages, and the like, in addition to data that make it possible to judge how much the extraction of gold increases with a successive decrease in the size of the particles. Of course, none of the data listed first can be supplied by any laboratory investigation; therefore this question belongs to the category of those whose solution in full is impossible in the laboratory.

What, then, can a laboratory investigation provide under such conditions? The answer to this question varies, depending on what material is to be investigated.

If it is ore, then it is necessary to find at what particle size the gold associated with them dissolves in potassium cyanide just as completely, or almost as completely, as does the gold associated with smaller particles. This particle size is obviously the limit below which grinding will not give any additional extraction—this is the limit of technically expedient grinding, which, we repeat, may not coincide with the economically possible limit.

If the investigation is to be carried out on tailings from amalgamation, then the laboratory is assigned the task of determining whether the tailings have been ground finely enough, and whether, after sizing, too much gold remains in the set of coarse particles. Of course, such a question most often has to be decided for sands.

As regards determining the most advantageous degree of grinding of fresh ore, the ideal solution to the question would be achieved if it were possible to prepare a series of samples whose particle size successively decreased. Treatment of these samples with potassium cyanide under identical conditions would at once give the desired result, since, beginning with a certain sample, the amount of gold extracted would no longer increase. The grain size of this sample would be the desired degree of grinding. Unfortunately, there is no apparatus that grinds ore only to particles of a definite size. Even rolls, along with particles equal in size ...

according to the distance between the bands, give many particles of smaller size, down to the finest dust. It is therefore necessary to resort to an indirect method, which we shall also apply to the study of tailings.

Assaying gold by size classes. The indirect method mentioned above consists in the following: the material subject to investigation is crushed—if it is ore—to a certain specified grain size, or is taken as such if it is tailings, and then a 100 g portion is sieved on standard sieves; from the sieve analysis thus obtained, it is calculated what quantity of material must be taken so that, after sieving, each class will contain an amount sufficient for determining the gold (100 g, if two fusions are to be made). It is necessary to note that the classes up to \(+200\) must be ground before assaying, since otherwise the separate assays do not agree. As a result of the determinations, a series of figures is obtained which gives an idea of how the gold is distributed among particles of material of different sizes. However, these figures will be indicative only if the separation into classes was carried out under definite conditions. It is possible to screen dry material, and it is possible to separate the classes while working in water. In the latter case the individual classes are obtained incomparably more homogeneous and cleaner than in the former, when finer particles, incrusting grains of larger ones, remain on sieves through which they could freely pass. Finally, separation of the classes, especially the class \(-200\), can be achieved by suspending the material in water and decanting through a sieve until the decantate becomes completely clear. In this way, in the laboratory one can obtain results similar to those obtained on factory hydraulic classifiers, the prototype of which is the simple Spitzkasten, where the light particles are carried away by the stream of water and the heavy ones settle. It is obvious that there is a profound difference between a method in which the particles are separated according to weight and methods of dry or wet screening, where the particles are separated exclusively by size. In the latter case one must expect enrichment of the fine classes in gold and sulphides at the expense of the coarse ones, since sulphides, as more brittle minerals, are ground finer than the other components of the ore, while gold, in most cases, occurs in the materials studied—especially in amalgamation tailings—in the form of particles indistinguishable even under strong magnification. Conversely, in the former case one may expect that gold and sulphides, as the heavier components not carried away by the stream of water, will be concentrated in the coarser classes.

What has been said is confirmed by the following experiments carried out in our laboratory. Two samples of efel from the Berikul mine, subjected—

[[unclear: beginning of sentence]] were subjected to dry and wet screening on standard sieves; the screening was carried out to completion. The results obtained after sampling the classes are summarized in Tables XII and XIII.

TABLE XII. Screening by the dry method.

Classes	Quantity, in g	Quantity, in % (a)	Gold content per 100 poods of ore (b)	Total amount of gold in the class \(\left(\dfrac{a \cdot b}{100}\right)\)	Distribution of gold by classes, in %
+ 80	865	35,3	1,54	0,54	16,1
+ 100	274	11,2	1,92	0,22	6,6
+ 120	202	8,3	2,69	0,22	6,7
+ 150	155	6,3	3,46	0,22	6,7
+ 200	207	8,4	3,84	0,32	9,9
− 200	747	30,5	5,76	1,76	53,6
Total	2450	100,0	3,69	3,28	100,0
Screening loss	50

TABLE XIII. Screening by the wet method.

Classes	Quantity, in g	Quantity, in % (a)	Gold content per 100 poods of ore (b)	Total amount of gold in the class \(\left(\dfrac{a \cdot b}{100}\right)\)	Distribution of gold by classes, in %
+ 80	967	39,3	1,54	0,61	18,2
+ 100	245	10,0	1,92	0,19	5,8
+ 120	145	5,9	2,30	0,14	4,1
+ 150	112	4,5	3,46	0,26	4,8
+ 200	143	9,5(13,8)	3,46(8,45)	0,33(1,02)	9,9(35,9)
− 200	759	30,8(26,5)	6,14(3,84)	1,84(1,16)	57,2(31,2)
Total	2462	100,0	3,69	3,42	100,0
Screening loss	38

As can be seen from a comparison of the figures in both tables and diagrams III, IV, and V, on which the results placed in the tables are plotted, both methods give, for the distribution of gold by classes, a picture of the same order. Quite different results are shown by the figures placed in Table XIII in parentheses. These figures were obtained in the following way: after the separation of the coarse classes, up to \(+150\), there remained the task of separating the \(+200\) and \(-200\) classes, for which purpose the remaining material was stirred up with water, and the suspended slime was decanted through a 200-mesh sieve until further stirring could no longer wash out anything, and completely clear water was decanted; the entire material was not transferred onto the sieve. With this method, a considerable part of the sulphides and gold remained in the \(+200\) class, and the gold content in it proved to be 8.45 zolotniks per 100 poods, while in the \(-200\) class it was 3.84. After assaying, the \(+200\) class was placed entirely on the sieve and washed to exhaustion. Almost exclusively sulphides passed through the sieve; these were added to the remaining \(-200\) class. A second assay of both classes showed that the gold content in the \(+200\) class had fallen to 3.46 zolotniks per 100 poods, whereas the \(-200\) class had become enriched to 6.14 zolotniks. Of course, the figures placed in parentheses would have been closer to the plant results if the Berikul tailings had been subjected to classification, for example on a Dorr classifier.

In order that the results of the investigations might be more or less uniform, dry screening to exhaustion has been adopted in our laboratory.

Having determined the distribution of gold by classes in the material under investigation, the latter, in a quantity of 2–3 kg, is subjected to cyanidation either in a percolator, if the material filters satisfactorily, or in 2–4 separate beakers, as described for the bottle test. In the latter case, the duration of agitation should not be made excessively long, since with too complete an extraction of gold the separate classes become so impoverished that the determination of gold in them cannot be made with sufficient accuracy. Agitation for 6–12 hours is usually quite sufficient.

In order to avoid the necessity of opening the bottle from time to time to admit air, \(0.1\%\) of barium peroxide is added to the samples. To prevent excessive consumption of KCN, CaO (\(0.2\%\)) is always added. The KCN solution is usually prepared at a strength of \(0.3\%\) and is taken in a ratio of 1:1 by weight of the material.

After the end of agitation, the extraction achieved is determined from the solution. The tailings are transferred to one large, thick-walled beaker, the KCN solution is allowed to settle, and then it is decanted by means of a siphon. If the diameter of the siphon tube is not very large, then the solution

it can be decanted almost clean. The beaker is then filled to the top with tap water, 15–20 cu. cm of a 10% alum solution is added to accelerate settling, the contents are vigorously stirred, and, after settling, decanted by siphon. The tailings are washed in this way 5–6 times, after which they are transferred onto a sheet of filter paper laid on canvas stretched over a frame fixed above some vessel, for example a bucket. After 10–12 hours the tailings are transferred to pans (round cast-iron pans are convenient) or sheet-iron trays and then dried either on a muffle or by heating the pans from below with a burner until a test tube with cold water, placed above the tailings, ceases to be covered with dew.

The dried tailings, usually caked into lumps, are rubbed with a porcelain pestle through a coarse sieve (20–30), after which the required portion is weighed out and sizing is begun. If the material contains many fine particles, it is better first to screen off the −200 class, since in that case there is less dusting.

In the screened classes the gold content is determined by the usual methods of fire assay.

Table XIV (p. 42) gives an example of fixing the gold by classes before and after treatment of the tailings, after amalgamation, with potassium cyanide in the laboratory, using ore from the Varvara mine.

The treatment was carried out over 48 hours (22 hours of agitation and 26 hours of settling with open stoppers) in three bottles. Into each were placed 250 g of tailings and 500 cu. cm of a 0.5% KCN solution. In this case the duration of treatment (48 hours instead of 24) and the strength of the potassium-cyanide solution (0.5% instead of 0.3) were increased against the standard, since earlier experiments had shown that the tailings were difficult to extract. Another portion was subjected to cyanidation in a laboratory percolator.

From the figures in the last column of the table and in column 11, where the extraction of gold in percent from each class is shown, it is evident that the classes +100, +120 and +150 give an extraction of about 60% in the bottle test and from 40 to 60 in the percolator, whereas the extraction from the −150 class reaches 83.5 and, respectively, up to 78%. Hence it follows that the degree of grinding of the coarse classes, which, as is evident from the table, contain about 30% of all the gold, is insufficient, and technically it would be correct to grind the entire material to the size of a grain passing through a 150 sieve.

What has been said sufficiently explains the essence of the method for determining the most advantageous degree of grinding.

Table XIV. Flotation tests by size classes.

Classes	Tails from grinding of the ore — Quantity, g	Tails from grinding of the ore — Quantity, %	Tails from grinding of the ore — Gold content, in 100 p.	Tails from grinding of the ore — Distribution of gold by mass, %	Tails from the butara test — Quantity, g	Tails from the butara test — Quantity, %	Tails from the butara test — Gold content, in 100 p.	Tails from the butara test — Distribution of gold by mass, %	Tails after concentration — Quantity, g	Tails after concentration — Quantity, %	Tails after concentration — Gold content, in 100 p.	Tails after concentration — Weight of gold grains	Tails after concentration — Distribution of gold by mass, %	Tails after concentration — Gold recovered, %
+100	470	58.1	3.33	68.9	725	58.6	1.59	68.0	1308	60.8	1.85	1.15	63.6	[[unclear]]
+120	70	8.6	1.90	6.0	65	9.2	0.92	5.0	176	7.7	1.07	0.80	5.8	[[unclear]]
+150	58	7.2	1.54	4.0	50	7.0	0.61	3.9	107	4.7	0.69	0.03	[[unclear]]	[[unclear]]
−150	205	25.3	12.30	24.1	180	25.0	0.38	8.5	582	25.5	0.50	0.13	9.2	[[unclear]]
Total	805	100.0	4.01	100.0	720	100.0	0.84	100.0	2280	100.0	1.42	1.29	100.0	[[unclear]]
Free gold	1	[[unclear]]	[[unclear]]	2.5	[[unclear]]	[[unclear]]	[[unclear]]	11	12	[[unclear]]	1.4	1.5	[[unclear]]	[[unclear]]

It should, however, be pointed out that the drawback of this method consists in the fact that, when the ore is ground, the coarser minerals are comminuted less than the finer remaining ones, and consequently the composition of the various classes will differ to a greater or lesser degree.

For this reason, the figures showing the extraction of gold from the various classes are not entirely comparable with one another, since they refer, strictly speaking, to materials of different composition. If it were possible to crush the ore into completely identical particles, then the extraction, for example, from particles passing through a 100-mesh sieve but remaining on a 120-mesh sieve, would differ from the extraction obtained for class \(+120\), separated from the same ore ground in the usual manner, because in the second case the composition of the class would differ somewhat from the composition of the ore in the first.

But in mill mechanisms the comminution of ore to a definite grain size is always accompanied by the production of finer particles of all classes, so that determining the most advantageous degree of grinding on the basis of data obtained by the method described above gives results which, although not fully coinciding with reality, are nevertheless sufficiently accurate for practical purposes.

Determination of the most advantageous strength of solution. Samples of the material under examination, 300 g each, are agitated for 24 hours with 300 cu. cm of KCN solution of various strengths: 0.05%, 0.1%, 0.2%, and so on. To each sample 0.3 g of CaO is added.

After the agitation is finished, the samples are allowed to settle; the clear solution is decanted through a dry filter into a dry beaker, and from the filtrate 2 samples of 100 cu. cm each are taken, which serve for determining the amount of gold extracted. With 300 g of material and 300 cu. cm of solution, it is possible without difficulty to filter 200 cu. cm even when investigating finely ground material that retains much moisture.

The concentration of potassium cyanide at which the greatest extraction of gold is obtained will be the most advantageous.

As an illustration of the method, we give the results of determining the most advantageous strength of solution for treating tailings from the Berikul mine (Table XV).

TABLE XV. Determination of the most advantageous strength of solution.

Concentration of KCN, in %	Extraction of gold, in %
0.05	61.1
0.10	73.5
0.205	77.1
0.30	72.6

From the figures in Table XV it is evident that, when treating the tailings with a 0.265% solution, an extraction of 77.1% was obtained, which is too sharply different from the extraction obtained with both stronger and weaker solutions. Therefore this result should be attributed to accidental causes, always possible in assaying, and 0.1% KCN should be recognized as the most advantageous strength.

Determination of the most advantageous ratio of liquid to solid. The determination is carried out according to the same scheme as that described above: charges of 300 g of material are treated, in the presence of 0.3 g CaO, with various quantities of 0.2% KCN solution. The most advantageous ratio of liquid to solid corresponds to the maximum extraction of gold from the solutions obtained after assaying.

As an example, Table XVI gives the figures obtained in the study of the tailings from the II Ioannovsky mine.

TABLE XVI. Determination of the most advantageous ratio of liquid to solid.

Tailings charge, g	0.2% KCN solution, cu. cm	Ratio of liquid to solid	Extraction, %
200	200	1:1	92.42
200	300	1:1½	84.28
200	400	1:2	84.84

The most advantageous ratio of liquid to solid for these tailings should be recognized as 1:1, since at this ratio 92.42% of the gold was extracted, whereas at higher ratios the extraction proved lower.

Let us note that determining the correct ratio of liquid to solid is of great importance, since with too small a ratio the extraction decreases, while with too large a ratio the productivity of the plant falls.

The most advantageous duration of treatment. An indirect method of solving this question was indicated in the place where the determination of the maximum extraction of gold was discussed. On page 34, figures were presented which were obtained when treating certain tailings for 1, 6, 12, 18, 24, and 36 hours. Considering the figures in Table XI, we see, for example, that for the tailings of the II Ioannovsky mine, 24 hours is the most advantageous duration of treatment, since with a further increase in the time of agitation the extraction does not rise; for the Berikul tailings, a 6-hour treatment will be the most advantageous, and so on.

To carry out the determination, charges of material, 300 g each—

then agitated for 6, 12, 18, 24, and 36 hours with 300 cc of a 0.2% KCN solution in the presence of 0.3 g of CaO. The solutions are treated as usual.

The most advantageous alkalinity of the solution. The most advantageous alkalinity can be determined by agitating ordinary samples (300 g; 300 cc of 0.2% KCN) with various amounts of calcium oxide (0.1, 0.2, 0.3 g, etc.). The alkalinity to which the greatest extraction corresponds is considered the most advantageous.

The test acquires great importance because, according to the observations of Clark and Cherwood, no more than 0.3 pound of lime per 1 ton (0.15%) should be taken, since a large excess of lime retards the dissolution of gold. ) The observations of the authors cited are confirmed by the experiments of Edmans, who dissolved gold in KCN and NaCN solutions in the presence of various amounts of CaO. In Table XVII some of the results obtained by Edmans are given. *)

TABLE XVII. Effect of lime on the solubility of gold in KCN.

KCN %	CaO %	Dissolved Au, mg	Could have dissolved, mg	Of the available KCN used, %
0.026	—	69.2	78.8	87.8
0.030	0.020	83.8	90.9	92.2
0.042	0.041	31.5	127.3	24.6
0.136	traces	243.3	412.2	58.9
0.136	0.074	0.5	412.2	0.1

Instead of the described test, in our laboratory the alkali consumption for a given material is determined; moreover, it is assumed that the amount of alkali corresponding to this consumption will be the most advantageous in treating it.

Alkali consumption ***) for the material under study is determined by agitating a sample of the material for 1 hour in 100 g with 100 cc of a 2% NaOH solution. It is assumed that during this interval of time all reactions possible between the alkali and the constituents of the ore have time to come to completion.

) Clark & Cherwood, Trans. Inst. Min. & Met., 1913, vol. XXII, p. 106.
) According to the abstract in Journ. Chem., Metall. and Min. Soc. of South Africa, 1917, vol. XVII, p. 160.
**) In practice, lime is used exclusively as the alkali; the use of caustic soda, preserved, unfortunately, at some of our cyanide plants, should be considered an irrational and unjustified waste of an expensive reagent. In the laboratory the use of caustic soda is more convenient than the application of sparingly soluble CaO.

At the end of the agitation, the solution is filtered off through a dry filter into a dry beaker. From the filtrate, 10 cubic cm of solution are taken for titration.

Knowing the content of NaOH in the solution before treatment and determining it after it, it is not difficult to calculate the consumption of alkali.

The content of NaOH in the solution is determined by titration with an acid of known titer in the presence of phenolphthalein or methyl orange. Most often sulfuric or oxalic acid is used for this purpose. In the first case the titer of the acid must be established exactly, for example, against a decinormal solution of soda; in the second, an acid of the corresponding concentration may be prepared directly by dissolving a weighed portion of chemically pure crystalline oxalic acid \((\mathrm{C}_2\mathrm{H}_2\mathrm{O}_4 \cdot 2\mathrm{H}_2\mathrm{O})\)*. If methyl orange was used in establishing the titer, then this same indicator should also be used in the subsequent titrations.

In most cases decinormal acids are used. For determining the consumption of alkali, however, it is more convenient to take an acid twice as strong. In preparing sulfuric acid of such strength, concentrated \(\mathrm{H}_2\mathrm{SO}_4\) of specific gravity 1.84 is used, containing, as is known, in 1 liter 1759 g of chemically pure monohydrate \((\mathrm{H}_2\mathrm{SO}_4)\), 5.57 cubic cm of which are diluted to a liter. 5.57 cubic cm contain 9.8086 g of \(\mathrm{H}_2\mathrm{SO}_4\), which is \(1/5\) of the equivalent weight.

The titer of the acid thus prepared is established against an \(n/10\) solution of \(\mathrm{Na}_2\mathrm{CO}_3\), for the preparation of which 5.3 g—\(1/10\) of the equivalent weight—of soda, obtained by ignition to constant weight of \(\mathrm{NaHCO}_3\), are weighed out; the weighed portion is dissolved in water and made up to one liter. Placing 10 cubic cm of the solution in a porcelain dish, 1–2 drops of methyl orange are added (0.02 g per 100 cubic cm), and it is titrated with acid until the yellow coloration changes to orange.

In the same way the NaOH solution is titrated before and after treatment of the latter with the material under investigation. A numerical example will explain how, from the results of the titrations, the consumption of alkali is calculated.

Results of titration:

1. For 10 cubic cm of \(n/10\) \(\mathrm{Na}_2\mathrm{CO}_3\) — 5.2 cubic cm of \(\mathrm{H}_2\mathrm{SO}_4\) were required.
2. For 10 cubic cm of NaOH solution — 23.2 cubic cm were required.
3. For 10 cubic cm of filtrate — 20.1 cubic cm were required.

Calculation:

5.2 cubic cm of \(\mathrm{H}_2\mathrm{SO}_4\) will be consumed for 10 cubic cm of \(n/10\) NaOH
(equivalent to 10 cubic cm of \(n/10\) \(\mathrm{Na}_2\mathrm{CO}_3\)).

23.2 cubic cm would be consumed for \(x\) cubic cm of \(n/10\) NaOH.

20.1 cubic cm would be consumed for \(y\) cubic cm of \(n/10\) NaOH.

* In accurate determinations the titer of oxalic acid should also be established.

hence

\[ x=\frac{1000\cdot ?}{5,2}=44,6\ \text{cc of } \frac{n}{10}\,\mathrm{NaOH}, \]

\[ y=\frac{10\cdot 20,1}{5,2}=38,65\ \text{cc of } \frac{n}{10}\,\mathrm{NaOH}. \]

But 1000 cc of \(\frac{n}{10}\,\mathrm{NaOH}\) contain 4 g of \(\mathrm{NaOH}\); therefore

\[ 44,6 \quad\text{“}\quad\text{“}\quad\text{“} \]

\[ 38,65 \quad\text{“}\quad\text{“}\quad\text{“} \]

\[ \text{hence}\quad n=\frac{4\cdot 44,6}{1000}=0,178\ \text{g } \mathrm{NaOH}, \]

\[ m=\frac{4\cdot 38,65}{1000}=0,155\ \text{g } \mathrm{NaOH}. \]

Consequently: 10 cc of the solution contain 0,178 g \(\mathrm{NaOH}\), or 1,78%;

10 cc of the filtrate contain 0,155 g \(\mathrm{NaOH}\), or 1,55%.

The concentration of the \(\mathrm{NaOH}\) solution after treatment decreased by 0,23%. Hence it follows that 100 cc of solution lost 0,23 g \(\mathrm{NaOH}\), attributable to 100 g of ore.

For 100 g of ore, 0,23 g \(\mathrm{NaOH}\), or 0,23%, is consumed.

For 100 poods, 0,23 pood, or \(0,23 \times 40 = 9,2\) lb, is consumed.

This is the alkali consumption for the given material.

Taking, instead of \(\mathrm{NaOH}\), the equivalent amount of \(\mathrm{CaO}\), one obtains the alkali consumption, expressed in pounds of lime, required to neutralize the material under investigation.

If \(\frac{n}{5}\) oxalic acid is used for titration (12,6 g per liter), then 1 cc of it is equivalent to 0,008 g \(\mathrm{NaOH}\), or 0,0056 g \(\mathrm{CaO}\). If, however, 10 cc of solution are taken each time for titration, then 1 cc of oxalic acid corresponds to 0,08% \(\mathrm{NaOH}\), or 0,056% \(\mathrm{CaO}\).

In those cases where one has to make many determinations of the alkalinity of solutions daily, expressing it as percentages of \(\mathrm{CaO}\) or \(\mathrm{NaOH}\), it is more convenient to use oxalic acid of such strength that each cubic centimeter of its solution is equivalent to 0,01 g \(\mathrm{CaO}\), or 0,01 g \(\mathrm{NaOH}\). To prepare the first solution, it is necessary to dissolve in a liter 22,5 g of oxalic acid (an amount equivalent to 10 g \(\mathrm{CaO}\)); to prepare the second, 15,75 g (equivalent to 10 g \(\mathrm{NaOH}\)). When titrating 10 cc, each cc of the first solution will correspond to 0,1% \(\mathrm{CaO}\), and of the second—to 0,1% \(\mathrm{NaOH}\).

As an indicator in titration with oxalic acid it is convenient to use phenolphthalein (1 g per 100 cc of alcohol).

XIV. Percolation Test.

Despite the brilliant development of the technique of the cyanide process and

one of the cheapest methods of treating gold-bearing materials is percolation, provided only that the material filters sufficiently well or can be made to do so after appropriate classification, and provided that this method gives a satisfactory extraction of gold from the material being treated. In this sense percolation may be compared with amalgamation. Both of these methods of extracting gold are, from the technical point of view, less perfect than, for example, the full slime process; nevertheless, owing to their cheapness and simplicity, they are fundamental in the gold industry.

In connection with what has been said, the percolation test acquires great importance among laboratory tests.

Of course, this test is unnecessary in those cases where preliminary investigations have shown that satisfactory extraction is obtained only when the entire material is ground to −150 and finer, or where it has turned out that the filtering capacity of the material is small and cannot be improved.

A special role in the latter case is played by sieve analysis of the material and by microscopic study of its separate classes, especially the fine ones. If the latter consist of fragments of quartz crystals, etc., then it may be assumed that, even with a considerable content of these classes in the material being treated, the rate of percolation will be satisfactory. Conversely, a small percentage of clayey particles very greatly lowers the filtering capacity, at the same time worsening the extraction of gold.

Thus, some idea of the filtering capacity of a material can be obtained by studying it under the microscope. A more exact idea of this capacity is given by determining the amount of moisture retained by it after drainage, and its porosity, i.e., the volume occupied in the freely poured material by voids, because both of these quantities depend on the sizes of the particles, on which, in turn, the filtering capacity depends.

Determination of the porosity of a material may be performed in several ways.

1. Into a vessel with a capacity of 100 cu. cm. (a graduated cylinder), filled to the mark with the material under investigation, water is poured from a burette until it rises to the level of the material. The reading of the burette gives the porosity of the material in percent. The results of the determination depend on the method of filling the vessel. In addition, the air contained in the pores hinders the penetration of water, reducing the results of the determination.

As an example, let us give the figures obtained in determining

porosity of the unclassified tailings of the II Ioannovskii mine, and of the same tailings from which class +200 had been separated.

For the unclassified tailings from the burette there went 39.8 cu. cm.

— Porosity 39.8%

For the classified tailings there went 38.5 cu. cm.

— Porosity 38.5%

1. If the specific gravity of the material—\(a\)—and its bulk density—\(b\) (the weight of 1 cu. cm.) are known, then the porosity—\(p\)—may be calculated by the formula:

\[ p=\frac{a-b}{a \cdot b}\cdot 100 \]

The specific gravity is most simply determined in the following manner: a measuring flask filled with water up to the mark, of 500 cu. cm capacity, is weighed. The water is then poured out, a weighed quantity of material (100 g) is poured in, and, after adding water up to the mark, it is weighed again. The specific gravity is calculated by the following formula:

\[ \text{Specific gravity}= \frac{\text{weight of material}} {\text{weight of flask with water}+\text{weight of material}-\text{weight of flask with water and material}} \]

For example,*) the flask with water weighs 600 g; the flask with water and the Nava ore weighs 660 g; the Nava ore weighs 100 g.

\[ \text{Specific gravity of the ore}= \frac{100}{600+100-660}=2.5, \]

The method described gives, of course, approximate results, but in most cases they are sufficiently accurate.

For determining specific gravity in our laboratory we use the Schumann apparatus, intended for determining the specific gravity of cement and shown in Fig. 5.

The determination is carried out as follows. The apparatus, with the ground stopper \(A\) tightly greased, is filled, by means of funnel \(C\), with water up to the lower mark, or slightly above it, in such a way as not to wet the walls of the graduated tube. It is then placed in a bath with water of 10–15 liters capacity, the temperature of which has been completely established and is checked by an accurate thermometer with divisions allowing tenths of a degree to be read.

After 2–3 hours the division at which the meniscus has stopped is read. The reading is repeated after an hour; if the meniscus has shifted, this indicates that the apparatus has not yet assumed the temperature of the bath.

Having taken the apparatus out of the bath, the weighed material, about 50 g, prepared with an accuracy to 1 mg, is poured into it through the funnel in small portions. In the event that material gets stuck in the tube, the latter is lightly tapped with a little mallet (a wooden stick on the

) MacFarren H. W., Text Book of Cyanide Practice*, N. Y. 1912, p. 88.

…into which water has been drawn (a thick-walled rubber tube). When the entire charge has been poured in, carefully striking the vessel, in an inclined position, against a felt pad placed on the table, all air bubbles are removed. While tapping, it is necessary to hold the lower part of the apparatus with one hand, and with the other the tube, ground to it, with divisions. After all bubbles have been removed, the apparatus is again placed in the bath with water; after making sure that the apparatus has taken on the temperature of the bath (the meniscus no longer shifts), the division to which the water has risen in the tube is read off. The difference between the second and the first readings gives the volume occupied by the material.

For the tailings of the Ioannovskii mine the following readings were obtained:

\[ \begin{aligned} \text{before pouring in the charge} &\quad 3.00,\\ \text{after } \quad » \quad » &\quad 28.70,\\ \text{volume of the charge in 70 g} &\quad 25.70\ \text{cu. cm.} \end{aligned} \]

Specific gravity of the tailings:

\[ \frac{70}{25.7}=2.72 \text{ at } 17.6^\circ\text{ C.} \]

Bulk weight is determined by weighing a tared vessel of known capacity, filled with the material under investigation. A measuring cylinder of 100 cu. cm may be used, filling it up to the mark.

Of course, the results of the determination will depend very strongly on the filling of the vessel. To eliminate the influence of this circumstance, for filling the vessel one uses the apparatus shown in Fig. 6, consisting of two inclined planes along which the material slides before entering the vessel placed beneath, whose volume is equal to one liter. The vessel is filled from above, and then the excess material is removed by means of a thin steel plate level with the edges of the vessel. The difference in the weights of the filled vessel and the empty one gives the weight of one liter.

For the tailings of the Ioannovskii mine the following figures were obtained (Table XVIII).

TABLE XVIII. Determination of bulk weight.

Tailings	Unclassified, 1835 g	Classified, 1842 g
Weight of vessel with tailings	1835 g	1842 g
Weight of empty vessel	563 g	563 g
Weight of 1 liter of tailings	1272 g	1279 g
Weight of 1 cu. cm	1.272 g	1.279 g
Weight of 1 cu. cm	1.380 g	1.369 g

In the last line of the table are placed the bulk weights determined by weighing a measuring cylinder, which considerably exceed the weights found by the method described above; this was to be expected with an arbitrary method of filling the cylinder.

Substituting into the formula given on p. 49 the specific and volumetric weights, we obtain the porosity of the material. Thus we find the porosity:

\[ \text{unclassified tailings} = \frac{2.72 - 1.272}{2.72 \cdot 1.272}\cdot 100 = 41.8\%. \]

\[ \text{classified tailings} = \frac{2.72 - 1.279}{2.72 \cdot 1.279}\cdot 100 = 41.5\%. \]

If the volumetric weights determined by the cruder method are substituted into the formula, the porosity of the unclassified tailings is found to be 35.7%, and of the classified tailings 36.8%.

1. Finally, the porosity of the material may be determined in the apparatus proposed by Julian and Smart*) and also serving for determining the moisture content of drained tailings and the rate of percolation.

The apparatus, shown in Fig. 7, consists of a glass tube A, 4.5′ long and 1¼″ (32 mm) in diameter, mounted vertically in a stand. At the bottom the tube is closed with a stopper, through which passes a glass tube B, opening inside flush with the stopper. In order that the material poured through the funnel from above to a height of 3′6″ (one-half of our height) should not clog the opening of the tube, the latter is covered with a wad of glass wool and, above it, with a circle of cloth of somewhat larger diameter than the tube. When the stopper is inserted, the edges of the circle are pressed tightly by it, and in this way a good filter is obtained. By means of a long rubber tube a bottle with a tubulure C is connected to tube B, serving to fill the tube loaded with material with water. During filling the bottle should be raised very slowly, so as not to increase the water pressure excessively, since otherwise voids form in the charge and the accuracy of the determination is considerably reduced.

If the volume of the part of the tube filled with material, \(V\), and the volume of water, \(v\), which has filled the pores between the particles and risen to its level in the tube, are known, then the porosity of the material, in percent, is expressed by the formula

\[ p = \frac{v \cdot 100}{V}. \]

The volume occupied by the material is not difficult to calculate, knowing the internal diameter of the tube (3.2 cm) and the height of the column of material in it (3′6″ = 106.7 cm):

\[ V = \frac{\pi \cdot 3.2^2}{4}\cdot 106.7 = 847\ \text{cu. cm}. \]

The volume of the water occupying the pores is determined by successive weighing of the tube filled with dry material, and of the

) Julian H. F. and Smart E. Cyaniding Gold and Silver Ores*. 1910, p. 27, 28.

same tube with water, raised to the level of the material filling it. In Table XIX are given the figures obtained in determining, by this method, the porosity of classified and unclassified tailings from the Ioannovsky mine.

TABLE XIX. Determination of porosity.

Tailings	Unclassified	Classified
Weight of tube with material	1927.5 g	1975.5 g
Same, with water	2269.5 “	2331.0 “
Weight of water	342 “	355.5 “
Porosity by the formula	40.37%	41.97%

The results of determinations of the porosity of both tailings from the Ioannovsky mine by various methods are given in Table XX.

TABLE XX. Results of porosity determination by various methods.

Methods	1	2	3
Porosity, unclassified tailings	39.8	41.8	40.37
“ classified “	38.5	41.5	41.97

Of these methods, only two of the latter may be considered sufficiently accurate, but they do not give results entirely consistent with one another. According to the second method, the porosity of both materials proves to be almost the same, whereas the third method gives for the porosity of the classified material a value 1.60% greater than for the unclassified material. The latter results should be recognized as more correct, since the classified tailings, consisting of coarser particles, should settle more loosely than the unclassified, and therefore their porosity should be greater.

Determination of the moisture content of the tailings after drainage, as indicated above, also makes it possible to draw some conclusion regarding the rate of percolation. This determination may be easily made with the aid of the same apparatus that served for determining porosity. By successively weighing the empty tube, the tube with dry material, and the tube with drained material after the voids in it have been filled with water, we obtain all the data necessary for calculating the amount of retained moisture. In Table XXI are given, by way of example, the figures obtained for the classified and unclassified tailings of the Ioannovsky mine.

TABLE XXI. Determination of the amount of moisture retained in the tailings.

	Unclassified	Classified
Weight of empty tube	743.5 g	760.5 g
Weight of tube with material	1927.5 g	1975.5 g
Weight of material \((V)\)	1184.0	1215.0
Weight of tube after drainage	2234.5	2210.0
Moisture in the tailings \((a)\)	307.0	235.5

Moisture content of the tailings:

\[ \frac{a}{V}\cdot 100 = 26.01\% \]

\[ 19.39\% \]

Below, on page 58, the interested reader will find one more method for determining the percentage of moisture retained by the material.

Considering the figures obtained in determining porosity and the volumes of sand that served us for the experiment, and the amount of moisture retained by it, we see that the classified sand, with greater porosity—we take the results obtained by the third method—retains 6.6% less moisture than the unclassified sand, whose porosity is lower. These results were to be expected, since the unclassified sand, as is evident from the sieve analysis placed in Table XXII, contains 32.7% more fine particles than the classified sand.

TABLE XXII. Sieve analysis of unclassified sand.

Classes	%
+60	12.12
+80	9.66
+100	11.00
+120	10.86
+150	7.03
+200	14.27
−200	32.73
Loss	2.33
Sum	100.00

On the basis of the data obtained, it may be said that the latter sand will percolate better than the former. However, an exact idea of the rate of percolation is given neither by determining porosity nor by the amount of retained moisture; therefore both these methods should be regarded only as qualitative.

Determination of the rate of percolation. As has already been mentioned above, the rate of percolation plays an outstanding role in the process, since a poorly filtering material remains too long in the cha—

… in which the solutions themselves flow off in thin separate streams, reducing the productivity of the entire plant; at the same time, gold extraction falls to a considerable degree, owing to the nonuniform passage of the solutions through the entire thickness of the material.

In practice it is usually considered that material above which the level of the solution in the vat drops by less than \(3/4''\) per hour is not advantageous to treat by percolation. A percolation rate of \(3/4''\) per hour is considered still poor; \(1\frac{1}{2}''\)—satisfactory, and \(3''\)—good.*) This rule is confirmed by the following Table XXIII, borrowed from H. W. MacFarren,**) in which, for charges of various depths, the percolation rates, in inches, are given that are necessary for the solutions to be able to descend in the interval of time allotted for the process.

TABLE XXIII. Dependence between the percolation rate and its duration.

Duration of percolation, in hours	Depth of charge, in feet	Depth of charge, in feet	Depth of charge, in feet	Depth of charge, in feet	Depth of charge, in feet	Depth of charge, in feet
Duration of percolation, in hours	4	6	8	10	12	14
Duration of percolation, in hours	Percolation rate, in inches	Percolation rate, in inches	Percolation rate, in inches	Percolation rate, in inches	Percolation rate, in inches	Percolation rate, in inches
96	0.80	1.20	1.60	2.00	2.40	2.80
120	0.64	0.96	1.28	1.60	1.92	2.24
144	0.53	0.80	1.07	1.33	1.60	1.84
168	0.46	0.69	0.91	1.14	1.37	1.60
192	0.40	0.60	0.80	1.00	1.20	1.40
216	0.36	0.53	0.71	0.89	1.06	1.24
240	0.32	0.48	0.64	0.80	0.96	1.12

From the table it is evident that, at a charge depth of 6′, the most frequently encountered, and a percolation rate of \(3/4''\), 144 hours are required in order for the solutions, taken in an amount equal to the weight of the material being treated (the usual ratio of liquid to solid), to have time to pass through it—144 hours, or 6 days—a duration of treatment the increase of which is impractical. Thus, the above-mentioned norms agree with the data of the table.

The percolation rate, for one and the same material, varies to a considerable degree depending on the method of filling the vat. For example, with dry charging the tailings occupy 10–20% more space than with wet charging; consequently, they lie more loosely…

*) Julian H. F. & E. Smart, p. 27.
**) loc. cit., p. 93.

...the layer and percolates more rapidly.*) For this reason, when determining the rate of percolation, attention should be paid to the method by which the apparatus is charged with material: it should be filled approximately under the same conditions as is customary, or will be adopted, at the plant. Otherwise, results are obtained that are not comparable with reality.

Above, in setting forth the methods for determining the porosity of materials, an apparatus was described that serves this purpose and was borrowed from Julian and Smart (see p. 52). By means of the same apparatus the authors mentioned propose to determine the rate of percolation. Mac Farren, recommending this apparatus in turn, adds that the tube should be of as large a diameter as possible, and that its height should equal the assumed depth of the vat charge.

The rate of percolation by means of the apparatus is determined as follows: having filled the tube with the material under investigation and with water, as described on page 52, enough water is poured on top so that a column of 5–6″ forms above the material; then a vertical strip of paper divided into tenths of an inch is pasted onto the tube and, after opening the clamp below, one waits until the water level coincides with the zero division of the scale. At this moment the time is noted, and after a while it is read how many divisions the water level has fallen. This reading gives the rate of percolation. More accurate results are obtained if one measures, in seconds, the time required for the level to drop by 1″ or by 5″.**)

For unclassified tailings from the Ioannovsky mine, the rate of percolation determined by this method proved to be 1.25″ per hour, and for classified material—9.83″. The determinations were made with a column of material 3′6″ high, since no longer tube was available in the laboratory.

The described method for determining the rate of percolation, properly speaking, would give results in full agreement with the operation of a plant vat if the walls of the tube did not exert a significant influence on the rate of downward flow of the solution, which does not occur under the conditions existing at a plant. It is possible that with tubes 3–4″ in diameter, the influence of the walls would not make itself felt to the same degree as occurred when working with a tube 1¼″ in diameter.

For this reason, in our laboratory, for such determinations a percolator is used, which also serves for determining gold extraction.

Laboratory percolator. Figure 8 shows such a

*) loc. cit., p. 93.

**) Mac Farren recommends proceeding to determine the rate of percolation after the apparatus prepared for the determination has stood for several days. loc. cit., p. 81.

a percolator,* ) mounted in a wooden stand. The percolator is fitted below with a stopper with a glass tube inserted into it, onto which a piece of rubber tubing with a clamp is put. A piece of sponge is placed above the stopper, and then a false bottom made of wire mesh with canvas stretched over it is inserted. If the latter is new, it must first be washed in order to free it from sizing. On top of the canvas a layer, 3–5 mm thick, of coarse-grained quartz is poured, of course containing no gold. Onto the filter thus prepared the material to be investigated is poured.

In most cases the rate of percolation is determined incidentally with the determination of the percentage extraction of gold; therefore the quantity of material charged into the percolator, as well as the magnitude of the latter, depends on the gold content in the material. Usually 3 kilograms are taken, which are thoroughly mixed with the required quantity of calcium oxide, determined according to its general consumption for the given material.

Since both the completeness of extraction and the rate of percolation depend on the correctness of the charging, special attention should be paid to this circumstance. The coarse and fine particles of the material are most uniformly distributed if it is charged through a sieve with sufficiently large openings—for example, a No. 40 sieve—placed above the percolator.

Procedure for pouring in the solutions. The first solution, of strength 0.2% KCN, which in most cases is sufficient, is introduced from below by means of a flask with a tube connected to the lower end of the percolator by a long rubber tube. When the solution is introduced from below, the charged material is saturated much more uniformly; no voids or channels are formed in it that would impair proper percolation. To avoid the same, the leveling flask should be raised, as the solution rises, in such a way that the levels of liquid in it and in the percolator do not differ too much.

Usually the amount of the first solution is 40% of the weight of the charge, or 1200 cubic cm if 3 kg have been taken. This amount is more than enough to fill all the pores between the particles of the material. When the solution has risen in the percolator to the level of the charge, the clamp below is closed, and the remainder of the solution is poured in from above, directing the stream onto a glass plate placed on top of the charge, in order to avoid washing out its upper layer.

In this condition the percolator is left for 24 hours. After this time has passed, the clamp is opened and the solution begins

*) Vessels of this form—of various sizes—are manufactured by many American firms, among which we shall indicate Denver Fire Clay Co, Denver, Colo, U. S. A.

flow into the measuring cylinders set aside for sampling, while adding no more. Over the course of 12 hours the first solution will have time to drain off as much as possible, and the entire mass charged into the percolator will become saturated with air. This is the aeration period, followed by treatment with the second solution (30% by weight of the material; 900 cubic cm at 3 kg), poured from above onto the glass plate mentioned above. The clamp is left open until the first drops of filtrate appear. At this moment the clamp is closed, and the solution remains in contact with the material under study for another 12 hours.

After 12 hours, having opened the clamp, there is let down into the measuring cylinder, where the first filtrate was located, such a quantity of solution that in all a volume is collected equal to the volume of the first solution poured in, in our case—1200 cubic cm. As is known, the solution poured into the percolator does not mix with the preceding stronger solution, but displaces it. This is confirmed, for example, by the figures in Table XXIV, where the quantity of solutions poured into the percolator during the treatment of the ore of the Varvarinsky Mine and the potassium cyanide content in them before and after treatment are shown.

TABLE XXIV. Concentration of KCN in solutions during a percolation test before and after treatment

Solutions	Poured in, cubic cm	% KCN	Drained, cubic cm	% KCN
I strong	1000	0.24	1000	0.13
II weak	900	0.09	900	0.07
III wash water	1000	—	1000	0.02

Instead of the cylinder with the first filtrate, another is placed, and percolation is continued until the second solution descends below the level of the material. At this moment the first wash water (30%, or 900 cubic cm) is poured in, which begins to displace the second solution. When the latter has collected to 900 cubic cm (or, in general, to an amount equal to that poured in), a third cylinder is placed, and, continuing to add wash water, another 900 cubic cm (30%) is collected. Wash water is continued to be passed through the charge until its total amount reaches 300% by weight of the material. Such material is considered sufficiently well washed, and the tailings after drying may be assayed for gold in order to determine extraction from the tailings.

The rate of percolation is usually determined while the wash waters are being drained. By this time the material has finally settle—

depends, and accidental causes will not affect the rate of percolation, as could have occurred when the first solution was being passed through (see p. 55). The rate of percolation is determined in the same manner as was described on p. 55, i.e., a strip of paper divided into tenths of an inch is pasted onto the percolator, and it is noted by how many divisions the liquid level falls over the course of one hour.

For the unclassified tailings of the Ioannovsky mine, a percolation rate of 2.2″ per hour was obtained. The same tailings in the Julian and Smart apparatus percolated at a rate of only 1.25″ per hour.

Probably the first figure will be closer to the percolation rate observed at the tailings plant at the Ioannovsky mine. The latter conclusion is based on the following data. At the Ioannovsky mine, freshly mined tailings are treated over the course of 4 days in vats with a charge depth of 7′, holding 113 tons. During this comparatively short interval of time, 200 tons of solutions manage to pass through the tailings. Thus, the ratio of liquid to solid here is 1.8 to 1. If we turn to MacFarren’s table, given on p. 54, we find that with a percolation duration of 96 hours the rate of percolation for a 6′ charge should be 1.2″, and for charges 8′ deep—1.6″ per hour, if the ratio of the weight of solutions to the weight of tailings is equal to unity. If, for a charge depth of 7′, we take the arithmetic mean of the rates required for charges of 6′ and 8′ and multiply it by 1.8, corresponding to the increase in the liquid-to-solid ratio, we obtain the percolation rate required for the case under consideration, namely—2.5″ per hour, which agrees quite well with the rate determined in the laboratory percolator. Unfortunately, we have no data at our disposal on the percolation rate observed directly in the vats at the Ioannovsky mine.

In the examination of the Berikul tailings, the percolation rate was determined by the same method and proved to be 1.25″ per hour, whereas the average from many observations made at the Berikul mine on the rate of percolation in the plant vats gives 1.48″.

Thus, this simple method gives results that agree sufficiently closely with the plant figures.

After draining the first solution, it is easy to determine the amount of moisture retained in the tailings (see p. 53). In this very case, if 1200 cubic cm of the first solution had been poured in, and after 12 hours of draining 260 cubic cm had collected in the graduated cylinder, ...

6 cm, then 3 kg of material retained 940 cubic cm, or

\[ \frac{940}{3000}\cdot 100 = 31\% \]

(for the Berikul sands) of moisture.

The moisture content of the same tailings was determined by drying 500 g of raw tailings just taken from the percolator. After drying, the weight of the tailings proved to be 400 g. Consequently, 400 g retained 100 g of moisture, or 25%.

It may be supposed that the result of the first determination (31% moisture) corresponds more accurately to the average moisture content of the entire charge than the result obtained by drying the sample, since the moisture content of the charge in the percolator is not the same in the different layers.

This latter conclusion may be confirmed by the following data, obtained in determining the moisture content of the tailings from percolation of sand from the Ioannovsky mine.

In 3 kg of tailings there remained 960 cubic cm of solution (determined from the difference between the volumes of the solution poured in and filtered off). The moisture content of the tailings was 32%.

The percolator was then overturned onto an iron sheet; the tailings that fell out of it in the form of a cone were quickly divided with a knife, according to height, into 3 parts corresponding to the upper, middle, and lower layers of the charge. From each part one sample of 200 g was taken. These were dried in pans and weighed. The results are given in Table XXV.

TABLE XXV. Moisture content in the different layers of the charge.

Parts of the charge	Weight after drying, g	Weight of evaporated water, g	Moisture, %
Upper	156.7	42.3	26.99
Middle	154.8	45.2	29.19
Lower	153.5	46.5	30.30

It is interesting to compare the data obtained by this method with the method described on page 53. There the moisture content of the tailings was determined as 26.1%; here it is 32%. Such a large difference may be explained by the different conditions under which both determinations were made. In our laboratory the moisture content of the tailings is determined by the method just described.

It was indicated above that determination of the moisture content of tailings makes it possible, to some extent, to judge the filtering capacity of the material. This determination is of incomparably greater importance because the quantity of dissolved gold remaining in the tailings, and consequently also the quantity of wash waters required for sufficiently complete washing of the charge in the vat, depends on the amount of moisture retained in the tailings.

The consumption of potassium cyanide for the material being treated may be established in a percolation test by drawing up a balance of the amount of KCN entering with the solutions into the percolator and leaving with them from the latter. With the same purpose, the KCN content is accurately determined both in the solutions poured in and in the filtrates collected separately.

Let us dwell somewhat on the method of determining the content of potassium cyanide in solutions.

Of the existing methods, titration of KCN with a solution of AgNO₃ is used almost exclusively. This method is based on the following well-known reactions:

\[ \mathrm{KCN + AgNO_3 = KNO_3 + AgCN}. \]

As long as potassium cyanide is present in excess, the silver cyanide dissolves in it with the formation of a complex:

\[ \mathrm{AgCN + KCN \rightleftarrows KAg(CN)_2} \]

until all the KCN has been bound; an excess drop of AgNO₃, reacting with this complex, decomposes it with the separation of AgCN, which, as is known, is insoluble in water:

\[ \mathrm{KAg(CN)_2 + AgNO_3 = KNO_3 + 2AgCN}. \]

Thus the end of the reaction is detected by the appearance of turbidity. If several drops of a 5% KI solution are added to the solution being titrated, then instead of silver cyanide, upon the addition of an excess drop of AgNO₃, the less soluble silver iodide is precipitated:

\[ \mathrm{AgNO_3 + KI = KNO_3 + AgI} \]

\[ \mathrm{AgCN + KI \rightleftarrows AgI + KCN}. \]

Silver iodide is soluble, like cyanide, in an excess of KCN, and therefore is not precipitated before the end of the titration:

\[ \mathrm{AgI + 2KCN = KI + KAg(CN)_2}. \]

The yellow color of the precipitated AgI makes it possible to distinguish the end of the reaction better, especially if the solution being titrated is somewhat turbid.

When large amounts of AgNO₃ solution are added, and also at the end of the titration, curdy flakes of AgCN form in the titrated solution; they dissolve rather slowly, which delays the determination. The addition of several drops of dilute ammonia, in which AgCN dissolves very readily, but AgI is almost completely insoluble, eliminates this inconvenience. It is also recommended to dilute somewhat with water the sample measured off for titration, since then AgCN dissolves more rapidly.

The titration should be carried out against a black background, for example on black glazed paper placed under the beaker containing the sample. The reaction is better noticed if the observer’s eye is raised to the same level as the solution and if an already titrated sample is placed nearby.

In most cases, both in laboratory investigation and at plants, it is necessary to make a very large number of KCN determinations. Therefore the AgNO₃ solution used for titration is prepared with such a calculation that 1 cu. cm of it contains an amount of salt equivalent to 0.01 g of KCN. Since, according to the equations given above, one molecule of AgNO₃ is equivalent to two molecules of KCN, it is not difficult to calculate that 1 cu. cm must contain 0.013046 g of AgNO₃, or 13.046 g per liter \((169.89 : 2 \times 65.11)\). For very exact work it is necessary to check the titer of this solution against an \(n/10\) solution of NaCl; for practical purposes this is unnecessary.

If, for taking samples, a pipette of 10 cu. cm is used, then the burette reading divided by 10 gives the KCN content in the solution in %, because 1 cu. cm of AgNO₃ is equivalent to 0.01 g of KCN, and a solution containing 0.01 g in 10 cu. cm is a 0.1% solution.

Returning to the determination of potassium cyanide consumption in a percolation test, we shall give an example that best explains the essence of the method. Table XXVI gives figures obtained in the investigation of Berikul tailings, taken in an amount of 3 kg, on the basis of which the KCN consumption was calculated.

TABLE XXVI. Determination of potassium cyanide consumption in a percolation test

Solutions	Taken, cu. cm	For 10 cu. cm taken, cu. cm AgNO₃ solution	KCN content in solutions, %	KCN content in solutions, g	Taken, cu. cm	For 10 cu. cm taken, cu. cm AgNO₃ solution	KCN content in filtrates, %	KCN content in filtrates, g	Loss or addition of KCN, g
I solution	1200	2.00	0.2	2.40	1200	1.30	0.130	1.560	−0.840
II, black	900	0.50	0.05	0.45	900	0.45	0.045	0.405	−0.045
III wash water	900	—	—	—	900	0.05	0.005	0.045	+0.045
Total in solutions taken				2.85	Total in percolated filtrates			2.010	−0.840

Thus, for 3 kg of tailings, 0.84 g of KCN was consumed, or 0.028% \((0.84 : 30)\), which amounts to 0.028 pood, or 1.12 pounds per 100 poods of tailings.

Determination of the percentage extraction of gold during percolation can be made both from the tailings and from the solutions. It was stated above that the latter is more convenient and expedient.

When assaying tailings, the question arises whether it is sufficient to...

they were washed free of any already dissolved gold that might have remained in them together with the moisture retained by them. To clarify this question, the following experiment was made. From the tailings after processing the sands of the Ioannovsky mine, 3 kg were placed in a percolator and, after drying, 750 g were selected by quartering. After grinding to −200, two charges of 100 g each were taken from this quantity for crucible smelting. All of the gold obtained from 100 g was 0.225 mg (the average of two determinations), which corresponds to 0.84 zolotnik per 100 poods.

The portions of 200 g selected from the same sample were washed with water by agitation and decantation and, after drying, were likewise assayed. It turned out that the gold content in the washed tailings was the same as in the tailings assayed immediately after drying, i.e., 0.84 zol. per 100 poods (0.225 mg in 100 g).

It follows from this that washing the tailings in the percolator is sufficient to wash out all dissolved gold.

When determining extraction from solutions, the latter are thoroughly mixed, acidified with several cc of hydrochloric acid to dissolve the usually precipitating CaCO₃, and then several samples of 250 or 500 cc are taken, depending on the gold content; these serve for determining the quantity of it (and, if necessary, of silver) that has passed into solution. Since the total volume of the solutions is known, it is not difficult to calculate, from the gold content in, say, 250 cc, how much gold has been extracted from all the material treated in the percolator. If the weight of the latter and its gold content are also known, then the percentage extraction is easily found.

Example. To determine the extraction from the sands of the Varvarinsky mine, two samples of 100 cc each, corresponding to 1 kg of sand each, were taken from the mixed solutions, since the mixed solutions had been evaporated to 300 cc, and 3 kg had been charged into the percolator.

In 100 cc there proved to be 4.62 mg of gold in one sample and 4.58 mg in the other, and on average 4.60 mg, which had been extracted from 1000 g of sands. Consequently, from 100 g, 0.46 mg of gold is extracted, and from 100 poods—1.77 zolotnik.

Since the sands contain 3.07 zol. per 100 poods, the extraction of gold is

\[ \frac{1.77}{3.07}\cdot 100 = 57.6\%, \]

which agrees quite well with the determination of extraction from the tailings, which gave 53.8% (1.42 zol. per 100 poods remained in the tailings).

As is known, percolation is not a universal process. Its applicability is limited only to such gold-bearing materials as do not need to be ground very finely in order to obtain good extraction, and as filter satisfactorily. All materials not satisfying these conditions must be treated by agitation—vigorous stirring of the finely ground material—with a KCN solution.

Besides special cases, when all the ore is converted into slime, in every grinding of ores, along with sands, a greater or lesser amount of slimes is obtained, by which term is understood everything that passes through a 200-mesh sieve. Therefore, in the laboratory it is necessary to investigate slimes just as often as other materials.

Of course, all the general methods are applicable to the study of slimes; their description formed the content of the preceding chapters. However, first of all it is advisable to test what the extraction of gold will be when treating the given material under approximately the same conditions as are most often encountered in practice—that is, with agitation.

For tests of this kind in the laboratory, either a self-made apparatus is used, in which a glass vessel is installed and into which, through a glass tube lowered almost to the very bottom, compressed air is passed, or the apparatus shown in Fig. 9, which is a small-scale reproduction of the Brown–Pachuca agitation vat.

The essential part of the apparatus is a glass vessel (a quart bottle with the bottom cut off and the neck shortened by half), mounted on a stand. In the center the vessel is closed by three glass supports resting with their lower ends on the conical part of the vessel, and with their upper ends fastened in a cork, through which also passes a glass tube that does not reach downward by several millimeters to the lower, closed part of the vessel opening, and at the top is connected with the source of compressed air. Through the same cork passes a glass rod carrying stirring vanes. These vanes partly serve the same purpose as the circulation pipe in the Brown–Pachuca vat, causing the pulp particles to move gradually upward in a spiral.

Unfortunately, in most laboratories there is as yet no compressor permitting air of the required pressure to be obtained. The pressure of the jet of a water-jet pump, which we use at present, does not exceed 2 cm of mercury column. With this pressure it is possible to agitate pulp only at a liquid-to-solid ratio of 8–10 to 1. With thicker pulp, the circulation of particles ceases.

After the agitation is finished, the charge is left in the vessel until the slime has completely settled; after this it is not difficult to decant and filter the liquid quickly, in order to determine the amount of gold dissolved in it.

Below we shall give examples of calculations of cyanidation by the described method (see p. 81).

XVI.

Concentration

When there is a large quantity of sulfides in the material being studied, it is advisable to separate them from the remaining components of the ore by the methods described on page 14, and to test them for gold.

If it proves that a significant part of the gold contained in the entire material is concentrated with the sulfides, or if they are generally rich, then it should be determined whether, in order to attain a sufficiently complete extraction, separate treatment of them is necessary.

To resolve this question, it is first of all necessary to treat, by the methods set forth above, under the most favorable conditions, the material as a whole.

Having treated several kilograms of material under these conditions, depending on the sulfide content, the gold content in the tailings is determined, and then, having separated the concentrates from them, the latter are tested. If it proves that the concentrates contain a still considerable quantity of gold, then this indicates that it is necessary to find conditions under which the gold from the concentrates is extracted satisfactorily, and in this way to arrive at the conclusion that the most advantageous condition for treating the concentrates is their enrichment to one degree or another.

If it proves that the conditions under which the concentrates give good extraction differ greatly from the conditions for treating the remaining mass of the ore, then separate treatment is to be preferred; if, however, the concentrated material presents practically the same conditions as the ore, then the entire ore should be treated in its entirety according to a scheme worked out for the concentrated material, although a more complicated and expensive one, or, having first leached the gold remaining unextracted in the concentrates, the whole ore should be treated according to the simpler scheme established for it. Which of these two ways will be the most advantageous—this question is not resolved in the laboratory for reasons indicated above.

XVII.

Operational determination of sodium cyanide

On page 61 a method was set forth for determining the consumption of sodium cyanide, [[unclear: continuation at bottom of page]].

KCN in a percolation test. In this case the material is always treated in the presence of alkali—CaO. When solving the problem in a general form, however, one usually determines both the maximum consumption of KCN for the given material when treating it without the addition of alkali, and the consumption of KCN in its presence, which gives results close to those that should be expected in practice.

The difference between the two values makes it possible to judge the relative amount of absorbers of potassium cyanide (cyanicides) in the material under investigation.

These determinations are carried out as follows. With a measuring cylinder, 100 cubic cm of a 0.5% KCN solution are measured into 2 dry bottles. To one of them 0.5 g of caustic soda is added and stirred until it is completely dissolved. Then 100 g of material are poured into each bottle, the bottles are stoppered and agitated for 6 hours. After this time has elapsed, the samples are removed, allowed to settle, and then the clear solutions are decanted through dry filters into dry beakers. From the filtrates, 10 cubic cm are taken for titration.

Example. 1. Test without alkali. 100 g of Berikul tailings were agitated for 1 hour with 100 cubic cm of KCN solution.

For 10 cubic cm of solution, 4.8 cubic cm of AgNO₃ solution were consumed.

the solution contains 0.48% KCN

For 10 cubic cm of filtrate, 3.9 cubic cm of AgNO₃ solution, N/3, were consumed:

the filtrate contains 0.39% KCN

The concentration of the solution decreased by 0.09%.

Or the whole solution lost 0.09 g of KCN, consumed by 100 g of tailings. The consumption of potassium cyanide is \(0.09 \times 40 = 3.6\) pounds per 100 poods of tailings.

1. Test with alkali. 100 g of tailings; 100 cubic cm of 0.48% KCN solution; 0.5 g NaOH; agitation for 1 hour.

For 10 cubic cm of filtrate, 4.6 cubic cm of AgNO₃ solution were consumed;

the filtrate contains 0.46% KCN

The concentration of the solution decreased by 0.02%.

The consumption of potassium cyanide is \(0.02 \times 40 = 0.8\) pound per 100 poods. Thus, when treating tailings with alkali, the consumption of potassium cyanide is reduced by \(3.6 - 0.8 = 2.8\) pounds per 100 poods, owing to the fact that the alkali, reacting with the absorbers of potassium cyanide, paralyzes the harmful action of some of them.

Useful consumption of alkali. In connection with the above, a useful consumption of alkali is distinguished, by which is understood that quantity of it equivalent to the quantity of potassium cyanide that is saved as a result—

addition of alkali. In the example given above, the useful consumption of alkali will be the amount of CaO or NaOH equivalent to 2.8 pounds of KCN, i.e., 2.41 pounds of CaO or 1.72 lb. NaOH per 100 poods of tailings.

The useful consumption of alkali should not be confused with its total consumption, the determination of which was discussed in detail on p. 46. Of course, in most cases the total consumption of alkali is greater than the useful consumption, since in the treatment of any material the alkali may react with substances that do not act on potassium cyanide, or with substances whose reaction products with alkali are the same absorbers of KCN as they themselves are. For example, FeSO₄ with alkali gives Fe(OH)₂, which dissolves in potassium cyanide.

\[ \mathrm{FeSO_4 + 2NaOH = Na_2SO_4 + Fe(OH)_2,} \]

\[ \mathrm{Fe(OH)_2 + 6KCN = K_4Fe(CN)_6 + 2KOH} \]

For Berezovsky tailings the total consumption of alkali amounts to 16.8 lb. CaO per 100 poods of tailings, whereas as useful consumption, out of this entire amount, we count 2.41 pounds.

XVIII.

Acidity of the Ore.

A distinction is made between the free acidity of the ore, the total acidity, and the combined acidity, found from the difference between the second and the first.

Free acidity (free acidity) is determined by titration with n/10 NaOH in the presence of phenolphthalein of 10–50 cubic cm of filtrate obtained after treating 100 g of material with 100 cubic cm of water for 1 hour. As always in such cases, the aqueous extract is filtered through a dry filter into a dry beaker. From the number of cubic cm of NaOH used to neutralize the sample, it is calculated what amount of caustic soda or calcium oxide is needed to neutralize the free acid contained in 100 poods of the material under investigation. It should be noted that, to determine the free acidity, a weighed portion must be taken either from freshly taken, usually moist, material, or from a sample kept in a well-stoppered bottle, since drying may spoil the sample by changing the conditions. Usually the determination of free acidity is made simultaneously with the determination of moisture in the material, and the results are recalculated to dry weight. *)

Example. For 50 cubic cm of filtrate, 2.1 cubic cm of n/10 NaOH were used.

\[ 2.1\ \text{cubic cm n/10 NaOH contain } \frac{4.2,1}{1000} = 0.0084\ \text{g NaOH.} \]

) Clennell J. E. The Cyandie Handbook.* N. Y. 1910, p. 484.

Thus 50 cc of filtrate neutralize 0.0084 g of NaOH,

and 100 “ “ “ 0.0168 “ “

Since 100 cubic centimeters were used to treat 100 g of moist material, 0.168 g of NaOH is required to neutralize the free acidity in it.

If the material contains 13.7% moisture, then the free acidity, recalculated to dry material, corresponds to

\[ \frac{0.0168}{100-13.7}\cdot 100 = \]

\[ =0.0195\% \text{ NaOH.} \]

For 100 poods of material this amounts to \(0.0195 \times 40 = 0.79\) pounds of NaOH, or

\[ \frac{0.79\cdot 56.09}{40}=1.09 \]

pounds of CaO.

In most cases, however, the reaction of the aqueous extract is neutral, and only rarely—weakly alkaline, as, for example, when treating tailings from the Udaloi mine. For 100 cc of filtrate, corresponding to 25 g of tailings (100 g were treated with 400 cc of water), in this case 1 cc of \(n/10\) (exact) \(H_2SO_4\) was used in the titration; consequently, the quantity equivalent to 0.004 g of NaOH. For 100 g of tailings this amounts to \(0.004 \times 4 = 0.016\) g of NaOH, or 0.0112 g of CaO. The free alkalinity of the tailings is therefore \(0.016 \times 40 = 0.64\) pounds of NaOH, or 0.45 pounds of CaO, per 100 poods.

The total acidity (total acidity) of an ore or of any material is expressed by the number of pounds of NaOH or CaO consumed per 100 poods. Thus the total acidity is identical with the total consumption of alkali, the determination of which was discussed on p. 46.

The bound acidity (latent acidity), as already mentioned, is determined as the difference between the total and the free acidities.

XIX.

Specific and Bulk Gravity.

We became acquainted above with the determination of the specific and bulk gravities of the materials investigated (see pp. 50, 51). Here it will be appropriate only to point out the importance of these constants in designing a new plant or in determining the capacity of an existing one. As an example we shall cite the statement of A. Clark and E. Lead, who treated at Homstake an ore with a specific gravity of about 3. The volume of one ton of this ore was 10% less than usual, since in most cases the specific gravity of ore varies between 2.5 and 2.7. Correspondingly, with the reduction in volume, the costs of treatment per 1 ton also decreased. On the other hand, the high specific gravity of the ore caused a number of difficulties, for example, in discharging pulp from the mortar and distributing it over the sluices, and in transporting the pulp through troughs*).

*) Trans. Am. Inst. Min. Eng, Vol. LII, 1916, p. 3.

XX.

The Necessity of Auxiliary Operations.

In the foregoing, methods of investigation have been described that are applicable, in general, to all gold-bearing materials.

For most of them, the results obtained by these methods make it possible to work out a rational scheme for the plant process. In some cases, however, complications of two kinds must be encountered: first, an excessive consumption of reagents, especially potassium cyanide; second, a low extraction of gold. In both cases an additional problem confronts the investigator—to determine the causes of these complications and to find measures for eliminating them.

The causes of a large consumption of potassium cyanide when treating some ore in the presence of a sufficient quantity of alkali usually consist in the presence in the ore of components that dissolve in potassium cyanide with the formation of complexes, as was already mentioned at the beginning of Chapter X (see p. 27).

1. Most often such a component is ferrous salts—the product of oxidation of pyrites during prolonged exposure of the material to air, and also during dry crushing of pyritic ores. Ferrous salts, usually sulfates, under the action of alkali yield ferrous hydroxide, soluble in potassium cyanide, as indicated in the equation given on page 66. It is evident that, if even a few tenths of a percent of ferrous oxide or its salts are present in the material, the consumption of potassium cyanide in its treatment will be very great.

It is not difficult to detect ferrous oxide in the material by a qualitative analysis of the filtrate after treatment of the material with potassium cyanide.

If it is found that ferrous oxide is present, the metallurgist has at his disposal two means of eliminating its harmful effect: preliminary washing of the material with water to remove soluble salts, and oxidation of the ferrous oxide to ferric oxide, the hydrate of which is insoluble in potassium cyanide. Both of these means should be tested by the investigator in the laboratory.

Preliminary washing with water is carried out by agitating the material with a double or triple quantity, by weight, of water and decanting the settled clear solution. After washing, the material is dried, and by the general method the consumption of potassium cyanide and alkali for it is determined. If the consumption has fallen to normal limits, then preliminary washing should be introduced into the process scheme; otherwise the investigation should be continued.

Oxidation of ferrous oxide is in practice achieved either by the natural method, by leaving the material exposed to the air, or by blowing air through the charge, if it is a slime, or by sucking air through (if it is tailings). An analogous procedure is used in the laboratory, aerating either in an agitation apparatus or by sucking air through moist material located in a Büchner funnel. After such treatment, the consumption of KCN and alkali is determined.

As an example of the practical application of aeration for the oxidation of ferrous oxide, let us cite the Homstake Mining Co. plant, where laboratory experiments established that one ton of tailings absorbs 75 cubic feet of oxygen. For this reason, after the wet charge has been drained, air is blown from below through the vats under a pressure of 4 pounds per square inch (21.6 cm Hg) for 16 hours. Such aeration is repeated twice more in the intervals between solution charges.

1. The presence of sulphide or oxidized copper compounds in the ore is also a cause of the consumption of a large quantity of KCN, because both the former and the latter dissolve in it with the formation of complexes. If analysis shows that there is enough copper in the ore to pay for smelting, and if the latter is possible by the nature of the ore and by local conditions, then, of course, smelting should be preferred to all other processes. Otherwise, it is necessary first, before cyanidation, to remove the copper, and only then to extract the gold from the tailings.

If the copper is present in the form of sulphide compounds, the latter can be separated by flotation, or the copper can be extracted after roasting the ore by one of the hydrometallurgical methods. The latter are also used in the case of oxidized copper ores.

As an example, let us cite the treatment of poor cupriferous ores containing gold at Goldfield, where the ore was concentrated on slime tables; the concentrates obtained, in which almost all the copper contained in the ore was enclosed, after roasting were agitated in Pachuca vats with sulphuric acid, and then, after separation of the solution into which the copper had passed, the material was sent for the extraction of gold to the slime plant, where it was treated together with the rest of the ore. Recently, flotation has been introduced instead of concentration on tables.*

At the Koktas-Dzhartass mine in the Semipalatinsk region, oxidized copper ores were worked, from which copper was extracted with sulphuric acid at the Dzheltav plant. The tailings were intended to be cyanided in order to extract the gold remaining in them. Ac—

* Trans. Am Inst. Min. Eng., Vol LII, 1916, p. 11.

** According to a private communication from mining engineer N. S. Penna.

the latter was not carried out, since the enterprise soon ceased to exist *).

If, in the laboratory, it becomes necessary to investigate ores similar to those indicated, then it is necessary, first of all, to work out methods for extracting copper, and only then, for the tailings, to establish the most advantageous conditions for treatment with potassium cyanide; as for the development of methods for extracting copper, their exposition does not enter into the program of the present article. We shall only point out that, for ores in which copper occurs in the form of sulfide compounds, flotation must be tried, by means of which it is sometimes possible to extract gold together with the concentrates; for oxidized ores, hydrometallurgical processes are still the cheapest, and their applicability to the given ore must be determined in the laboratory.

1. Much more rarely, other absorbers of potassium cyanide are present in ores, such as, for example, arsenic (usually mispickel \(FeAsS\)), \(MgSO_4\) **), \(Al_2(SO_4)_3\), \(Ni\), \(Co\), etc.

The harmful influence of some of them—\(MgSO\), \(Al_2(SO_4)_3\)—is paralyzed by the addition of alkali, since the corresponding hydroxides are harmless to the process. In large quantities, hydroxides greatly reduce the filtering capacity, which is why it is sometimes necessary to introduce a preliminary washing of the material.

The causes of insufficient extraction of gold in the majority of cases consist either in insufficiently fine grinding, necessary in order to expose the particles of gold, or in a deficiency of oxygen, which, as is known, is an essential factor in the process of dissolving gold.

A deficiency of oxygen is caused most often by the presence in the ore of sulfide compounds which absorb oxygen and convert \(KCN\) into \(KCNS\). Such compounds should include, first of all, alkaline sulfides formed by the action of alkalis on sulfides, especially minerals containing sulfide arsenic and antimony, then copper sulfides. Pyrite and marcasite are very weakly subject to the decomposing action of alkalis and are sufficiently harmless in themselves.

The formation of alkaline sulfides can be prevented by the addition either of lead acetate, or of litharge, which is often practiced at cyanide plants. Lead sulfide oxidizes rather slowly, and therefore does not absorb oxygen.

In most cases arsenopyrite is decomposed with difficulty by alkalis, and the extraction of gold from ores containing it even in—

) Mining Journal, 1912, vol. III, p. 137.
*) According to certain experiments in our laboratory, not yet fully completed, it may be supposed that \(MgSO_4\) does not affect the consumption of \(KCN\).

in significant quantities, presents no difficulty. However, ores sometimes prove so refractory that, in order to achieve complete extraction of the gold, excessively prolonged treatment is required. In such cases a preliminary roasting should be tried; it is also used for the removal of antimony.

Preliminary roasting is also applied to telluride ores, from which the gold, although it may be extracted, is obtained only after very prolonged treatment*).

As an auxiliary operation applied to refractory telluride ores, bromocyanidation should be mentioned.

However, both bromocyanidation and preliminary roasting are so rare in the practice of the cyanide process that it is unnecessary to dwell here on methods of studying ores in cases where their application is required.

In the following chapters the reader will find examples of the application of the methods set forth to the study of plant materials: amalgamation tailings and ores.

XXI.

Study of Tailings from the Berikul Mine.

At the 6th Berikul area of the mine, during the summer campaign of 1917, it was proposed to treat, at the tailings plant, a dump of tailings left over from the operation of the amalgamation mill and not used in previous years.

The purpose of the study was:

1. determination of the quantity of tailings in the dump;
2. taking an average sample;
3. determination of the gold and silver content;
4. systematic study of the tailings in the laboratory, with the aim of determining the possible extraction and establishing the most advantageous treatment conditions.

Thus, the entire study was divided into two unequal parts: 1) work at the mine and 2) work in the laboratory.

Work at the Mine.

1. Volume of the dump. As can be seen from the drawing showing a schematic view of the dump from above and in cross section, the dump was divided into 7 parts by parallel lines drawn at a distance of 5 sazhens from one another. Each such part could be

*) Clevenger G. H. Met & Ch. Eng. Vol XIV, 1916 p. 203.

regard it as a truncated pyramid with trapezoidal bases, the volume of which is expressed by the formula:

\[ V = (F + f + \sqrt{F \cdot f})\, \frac{h}{3}, \]

where \(F\) is the area of the lower trapezoid,
\(f\) is the area of the upper trapezoid,
\(h\) is the height of the pyramid.

The sum of the volumes of all seven parts gives the required volume of the dump.

As an example, let us present the calculation of the volume of the part marked on the drawing with the Roman numeral II.

The height of this part is determined by multiplying the arithmetic mean of the lengths of the two adjacent slopes \(l_{II}\) and \(l_{III}\), measured with a tape, by the sine of the slope angle, determined with a mining compass and found everywhere to be \(40^\circ\). Thus, the height \(h_{II}\) is equal to

\[ h_{II} = \frac{l_{II} + l_{III}}{2} \cdot \operatorname{Sn} 40^\circ = \frac{2{,}60 + 2{,}90}{2} \cdot 0{,}64 = 2{,}46\ \text{sazhens}. \]

The area \(f_{II}\) of the upper trapezoid is equal to half the sum of the parallel sides \(a_{II}\) and \(a_{III}\), multiplied by the height, i.e., by the distance between the separate lines delimiting adjacent parts of the dump, equal to 5 sazhens.

\[ f_{II} = \frac{a_{II} + a_{III}}{2} \cdot 5 = \frac{4{,}40 + 4{,}20}{2} \cdot 5 = 21{,}50\ \text{sq. sazhens}. \]

The area of the lower trapezoid \(F_{II}\) is determined by the formula:

\[ F_{II} = \frac{a_{II} + l_{II}\operatorname{Cos} 40^\circ + a_{III} + l_{III}\operatorname{Cos} 40^\circ}{2} \cdot 5 = \]

\[ \frac{4{,}40 + 60 \cdot 0{,}77 + 4{,}20 + 2{,}90 \cdot 0{,}77}{2} \cdot 5 = 32{,}08\ \text{sq. sazhens}. \]

From these data we calculate the volume of the part of the dump:

\[ V_{II} = (21{,}50 + 32{,}08 + \sqrt{21{,}50 \cdot 32{,}08}) \cdot \frac{2{,}46}{3} = 65{,}48\ \text{cub. sazhens}. \]

Summing the volumes of the remaining parts of the dump, found in the same manner, we find that the volume of the entire dump is equal to 306.47 cubic sazhens, or 2976.9 cubic meters.

The weight of 1 liter of air-dry tailings, determined by means of the apparatus shown in Fig. 6, is 1.252 kg.

Hence, the weight of the dump is equal to \(2976{,}9 \times 1{,}252 = 3727{,}1\) tons, or 227,271 poods; in round numbers—230 thousand poods.

1. To take an average sample, 6 places were marked on the top of the dump, along the middle of each of the 6 lines dividing it into sections, and 10 places at equal distances from one another along a line running along the middle of the slope. In Fig. 10 the places where the samples were taken are marked with circles.

Since the sample had to be taken in winter, when the frozen tailings constituted a very hard mass, difficult to work with a pick, it was necessary to confine ourselves to the indicated 16 points, at which holes were made with the aid of a hand drill. Unfortunately, these latter were made too deep, so that even two dynamite cartridges could not break through the upper layer. With the aid of a crowbar and pick it was possible to drive the pits down to the depth at which, by the action of the explosion, the tailings had been sufficiently broken up.

From each pit one wheelbarrow of tailings was taken (8–10 poods each); in all, the total weight of the sample amounted to about 100 poods.

The sample was arranged in the form of an annular layer on a swept platform. By shoveling over and breaking the frozen lumps with a hammer, the sample was crushed, and then twice quartered after being poured into a cone. From the final sample, in all about 10 poods, approximately 1 pood was taken for determination of moisture, while the remaining mass, after drying, was packed in a tight box and sent to the laboratory.

In taking the sample, 7 workers and 1 foreman were occupied for the whole day (10 hours).

After thawing in a room at ordinary temperature, the sample taken for determination of moisture was screened through a coarse sieve; by dividing into squares, 500 g were taken and dried in a pan until the moisture was completely removed. It proved that, after drying, the charge had lost 70.5 g. Consequently, the moisture content of the tailings was 14.1%.

Work in the laboratory.

1. From the sample brought to the laboratory, about 10 kg of air-dry tailings were taken and passed through a sieve with openings of 1 mm. Only 0.8% by weight of the material remained on the sieve, consisting of chips, small pebbles, etc., accidental impurities which could therefore be disregarded. The sieve analysis of the tailings is given in Table XXVII.

By means of a Jones divider, from the screened 10 kg about one kilogram was taken, ground subsequently to 150, and used for determining the content of gold and silver by crucible fusion.

TABLE XXVII: Sieve analysis of the Berezovskii tailings.

Classes	%
+40	7.45
+50	8.30
+60	9.90
+70	7.20
+80	7.00

+90	7.60
+100	3.40
+120	6.90
+150	6.15
+200	8.20
−200	26.00
Screening	1.80

As the average of 5 determinations, the gold content in the tailings was established as 3.69 zolotniks of gold, and 4.85 zolotniks of silver per 100 poods.

The ratio of silver to gold in the tailings is equal to 1.3 : 1, or a gold fineness of 432, counting silver and gold without ligature.

1. The further investigation was considerably simplified by the fact that, under the conditions existing at the mine, any possibility of additional classification of the material, of any further comminution, concentration, etc., was completely excluded; it had been established in advance that the tailings, in one form or another, but in the same condition in which they are found in the dump, would be treated at the existing tailings plant. Thus the problem was reduced to: a) finding the maximum extraction, b) determining the most advantageous treatment conditions, c) determining extraction by means of percolation, d) establishing the consumption of reagents, e) summarizing the results of the investigation.

a. Maximum extraction. The samples were agitated for 48 hours. There were taken 250 g of tailings, 250 cu. cm. of a 0.3% KCN solution, 0.3 g of CaO and 0.3 g of barium peroxide.

Gold extraction—70.20%—2.59 zol. per 100 poods.
“ silver—53.33%—2.59 ” ” ” ”

b. The most advantageous strength of solution. The results were given on p. 86 in Table XV, from the figures of which it follows that the most advantageous strength of solution is 0.1%.

The necessary duration of treatment was determined by agitating samples of 300 g with 300 cu. cm. of a 0.2% KCN solution for 1–36 hours. The results are given in Table XXVIII.

TABLE XXVIII. Determination of the duration of treatment.

Duration of agitation in hours	KCN content in solution, % before treatment	KCN content in solution, % after treatment	KCN consumption, lb. per 100 p.	Extraction, %%
1	0.2	0.175	1.0	60.39
6	0.2	0.12	3.2	73.99
12	0.2	0.08	4.8	71.55
18	0.2	0.16	1.6	72.91
24	0.2	0.08	4.8	72.91
36	0.2	0.08	4.8	72.91

As is evident from the figures in the last column, extraction is completed within 6 hours.

c. Percolation test. Into the percolator were charged 3 kg of tailings mixed with 6 g of CaO. The first solution, in an amount of 1200 cu. cm (40% by weight of the tailings), was left in contact with the charge for 12 hours. After 12 hours of aeration, 900 cu. cm (30%) of a second solution, which had been in contact with the tailings for 12 hours, were poured in. Then enough wash water was added to obtain the full quantities of both solutions and a further 900 cu. cm (30%) of wash water. On p. 61, in Table XXVI, the results of titrating the solutions before and after treatment were given, and the consumption of KCN was calculated; it proved to be equal to 1.12 pounds per 100 poods.

At the same time there were determined: the rate of percolation, which proved to be equal to \(1\frac{1}{4}''\) per hour, and the amount of moisture in the tailings after draining the first solution—31% (of the 1200 cu. cm of the 1st solution, 260 cu. cm descended:

\[ \frac{1200-260}{3000}\,100 = 31 \]
).

As a result of the treatment, from the tailings there were extracted 2.73 zol. of gold and 2.36 zol. of silver from 100 p., or 74.0% of the gold and 48.7% of the silver.

d. Consumption of reagents. The consumption of potassium cyanide, determined in the percolation test, proved to be 1.12 lb. per 100 poods. In treating the tailings in the presence of 0.5 g NaOH, the consumption of KCN was equal to 0.8 lb. per 100 poods (concentration of the solution before treatment 0.48%, after treatment 0.46%).

Consumption of alkali. 100 g of tailings + 100 cu. cm of a 1.59% NaOH solution. Concentration of the solution after treatment, 1.30%.

Consumption \(1.59 - 1.30 = 0.29\%\), or 11.6 lb. NaOH per 100 poods of tailings, which is equivalent to 16.3 lb. CaO.

The reaction of the tailings proved to be neutral.

In view of the fact that the consumption of alkali proved high, a preliminary washing of the tailings with water was tested. After washing, the sample of tailings was dried and then the consumption of alkali and potassium cyanide was determined.

Consumption of KCN after washing—0.8 lb. per 100 p. Consumption of NaOH after washing, 10.4 lb. per 100 poods, which is equivalent to 14.6 lb. CaO. Washing gives only a small saving in reagents, and therefore it makes no sense to introduce it as a preliminary operation.

The cause of the large amount of alkali consumed is apparently \( \mathrm{Al_2(SO_4)_3} \), which passes in appreciable quantities into solution when the tailings are treated with water.

e. Summary of the results of the investigation. 1. By measurements, the volume of the tailings dump must be taken as 306.5 cubic sazhens.

1. The moisture content in the tailings is 14.2%.

2. The bulk weight of the air-dry tailings is 1.252 kg (weight of 1 liter).

3. The weight of the tailings in the dump (without moisture), in round figures, is 230,000 poods (the approximate accuracy of this figure is 15%).

4. The content in the average sample is 3.69 zolotniks of gold per 100 poods, and 4.85 zolotniks of silver.

5. The total content in the dump (230,000 poods) is: gold—2 poods 8 pounds 39 zolotniks, and silver—2 poods 25 pounds 75 zolotniks.

6. At a potassium cyanide concentration of 0.1%, 73.5% of the gold is extracted, a figure close to the greatest extraction obtained in the investigation; therefore the customary strength for the first solution, 0.2% KCN, appears quite sufficient.

7. At a solution strength of 0.2%, dissolution of the gold is completed in 6 hours, which could be considered a sufficient duration of treatment.

8. When treating the tailings in a laboratory percolator, 74% of the gold and 48.7% of the silver are extracted—the maximum extraction figures of all those obtained in the investigation. In practice one should expect the same extraction, perhaps with some deviation upward.

9. With the extraction indicated (in 9), the total amount of gold that can be extracted from the dump will be 1 pood 25 pounds 39 zolotniks, and 1 pood 16 pounds 52 zolotniks of silver.

10. The ingots of the ligature gold obtained will contain 53.6% Au and 46.4% Ag, if the ligature is not counted and the sum of silver and gold is taken as 100.

11. For treating the tailings, 1.2 pounds of KCN, or 0.92 pounds of NaCN, will be required per 100 poods (we take the KCN consumption value obtained in the percolation test), and in all about 72 poods of KCN, or 53 poods of NaCN.

12. For 100 poods of tailings, together with this, 16.8 pounds of CaO will be required, and in all 966 poods.

13. Preliminary washing with water gives no special results in the sense of saving potassium cyanide or lime.

14. According to laboratory data, the percolation rate is determined to be \(1\frac{1}{4}''\) per hour. With a vat depth of 2 arshins 6 vershoks, the layer of percolating tailings will be about 6 feet. According to MacFarren’s Table XXIII (see p. 54), for a 6-foot layer at a percolation rate of \(1.2''\) per hour, 96 hours, or 4 days, are required for the passage of the solutions, with a ratio of the weight of solutions to the weight of the charge of 1:1. Since the tailings retain about 30% moisture after draining, the quantity of solutions will have to be increased, relative to the weight of the charge, to \(1\frac{1}{2}\) times the amount. Then the passage of the solutions will require \(96 \cdot 1.5 = 144\).

hours, or 6 days. Allowing 1 day for loading and unloading, we obtain 7 days for the turnover of one vat.

1. The low rate of percolation is explained by the presence in the slime of a considerable quantity—up to 26%—of silts, as may be seen from Table XXVII.

On studying the size fractions under the microscope, it turned out that both the coarse and the fine particles are crystalline; therefore the slime, despite the large content of the latter, nevertheless filters more or less satisfactorily.

Unfortunately, it proved impossible to compare all the results obtained in the laboratory with the results of the summer campaign of the slime plant at the Berikul mine, since, along with the treatment of slime from the dump, fresh slimes of the most varied character were also processed in the vats. The latter was explained by the fact that throughout the summer of 1917 the mill worked almost exclusively on miners’ ore.

Table XXIX gives the results of the summer campaign of the slime plant.

TABLE XXIX. Results of the campaign of the slime plant.

Item	Quantity
Slimes treated	452,000 puds
Ligature gold obtained	5 puds 10 lb. 07 zol.
Pure gold obtained	2 puds 35 lb. 90 zol.
Pure gold obtained per 100 puds	2.24 zol.
Silver obtained	1 pud 32 lb. 29 zol.
Silver obtained per 100 puds	1.53 zol.
Assay of gold (without ligature)	585.2
NaCN consumed	182.5 puds
NaCN consumed per 100 puds	1.61 lb.
1.61 lb. NaCN equivalent to KCN	3.14 lb.
NaOH consumed	169 puds
169 p. NaOH equivalent to CaO	274.4 puds
CaO consumed	1950.0 puds
Total alkali as CaO	2224.4 puds
CaO alkali per 100 puds	19.8 lb.

XXII.

Study of the slimes of the Ioannovsky mine.

A more complicated problem was presented by the study of the slimes of the Ioannovsky mine. In this case it was necessary to find, independently of the equipment available on site, the most rational methods for extracting gold from the slimes produced at the amalgamation mill.

A sample of slimes weighing about 8 puds was delivered to the laboratory; from it, by successive reduction, about 1.5 kg was selected for further work.

The chemical analysis of the tailings proved to be as follows:

SiO₂—86.4%; Fe—3.53%; S—0.39%; CaO—2.21%; CO₂—1.72%.

In accordance with this analysis, the mineralogical composition of the tailings is represented as follows:

SiO₂—86.4%; Fe₂O₃—3.89%; FeS₂—1.74%; CaCO₃—3.93%.

Thus, the tailings consist of almost pure quartz with small admixtures of limestone and pyrite, and with an admixture of iron oxide, which is partly a product of the oxidation of pyrite and partly the oxidation of iron particles—the result of wear of the crushing mechanisms, the crusher and the runners.

The screen analysis of the tailings is presented in Table XXX.

TABLE XXX. Screen analysis of the tailings of the Ioannovsky mine.

Classes	%
+60	12.12
+80	9.66
+100	11.00
+120	10.86
+150	7.03
+200	14.27
−200	32.73
Spill-through	2.33

Study of the classes under the microscope showed that the mineralogical composition given above corresponds to reality, since in the microscope’s field of view there were occasional grains of pyrite, with characteristic striation on the faces, and limestone, which, upon the addition of dilute hydrochloric acid, gave off bubbles of carbon dioxide, as well as quartz particles strongly colored brown by iron oxides.

The −200 class appeared under the microscope to consist of crystalline particles, which indicated the possibility of good filtering capacity.

To determine the average gold content, several samples were fused; on average they gave a gold content of 6.34 zolotniks per 100 poods and silver 4.93 zolotniks. Gold fineness—563.

The data of the screen analysis given above show that the content of slimes, understanding by the latter the −200 product, is very significant. Therefore, it remained to determine subsequently the most advantageous conditions of treatment for the unclassified material, and separately for the sands and slimes. The latter was especially important, since in the slimes, as is evident from the figures of the last column of Table XXXI, there is contained half of all the gold present in the material.

TABLE XXXI. Gold fixation by classes.

Classes	Quantity, in g	Quantity, in %	Gold content in 100 poods, zolotniks	Total amount of gold in the class, zolotniks	Distribution of gold by classes, in %
+ 60	181,83	12,12	5,490	0,666	9,5
+ 80	145,00	9,66	5,340	0,516	7,4
+ 100	165,00	11,00	5,375	0,591	8,5
+ 120	162,90	10,86	4,839	0,523	7,6
+ 150	105,40	7,03	5,183	0,364	5,3
+ 200	214,00	14,27	5,868	0,795	11,7
− 200	491,00	32,72	10,556	3,455	50,00
Sweepings	34,77	2,33	—	—	—
Total	1500,00	100,00	6,34	6,910	100,0

To obtain the starting materials for the investigation, 6 kg, selected from the general sample, were thoroughly washed on a 200-mesh sieve. That which remained on the sieve was investigated separately, as were the tailings, which had passed through separately, and the slime.

The maximum extraction for all three products was determined by 48-hour agitation of samples with a 0,3% KCN solution in the presence of CaO and barium peroxide. The results given in Table XXXII show that all three products can give good extraction.

TABLE XXXII. Determination of maximum extraction.

Material investigated	Gold content, zol. in 100 poods	Gold extraction, zolotn. from 100 poods	Gold extraction, in %
unclassified	6,34	6,05	95,45
tailings \((\Sigma +60+200)\)	6,31	6,05	95,45
slime \((−200)\)	6,53	5,34	81,76
[[unclear: material name]]	8,35	8,02	96,10

The most advantageous treatment conditions. 1. The strength of solution sufficient to attain satisfactory extraction was determined by agitation of a series of samples from all three products in the presence of 0,3 g CaO, with KCN solutions of various strengths. The results are summarized in Table XXXIII.

TABLE XXXIII. Determination of the most advantageous strength of solution.

	0,31	0,21	0,10	0,05
KCN concentration, in %	0,31	0,21	0,10	0,05
Extraction, in %, from unclassified	93,93	92,42	93,93	—
Extraction, in %, from tailings	80,88	79,41	77,94	76,47
Extraction, in %, from slime	95,49	91,21	91,00	88,53

From the figures given it follows that, in both of the latter materials, the greatest extraction is obtained with a 0.31% solution; but since, with an increase in the concentration of KCN, its consumption increases, for tailings it would be necessary to limit oneself to a strength of 0.2%, and for slimes to 0.1% or even 0.05%. For unclassified material, an increase in the strength of the solution does not entail an increase in the quantity of gold extracted; therefore the solution strength customary in practice, 0.2%, appears quite sufficient.

1. The most advantageous duration of treatment for unclassified material and for tailings proved to be 24 hours; for slime, 6 hours is sufficient, as is evident from the figures in Table XXXIV. The method of determination is set forth on p. 44.

TABLE XXXIV. Determination of the most advantageous duration of treatment.

Duration of agitation, hours	6	12	18	24	36
Extraction, in %% from unclassified	—	86.36	87.87	92.42	92.42
Extraction, in %% from tailings	—	66.17	74.23	79.41	—
Extraction, in %% from slime	92.87	92.87	—	94.71	92.81

1. The most advantageous ratio of liquid to solid, as follows from the data given in Table XXXV, is, for slimes, 2:1, and for unclassified material—1:1. For tailings the ratio was not determined, since, in general, in this case this constant does not especially affect extraction, and the results of the determination would be similar to the results found for unclassified material.

TABLE XXXV. Determination of the most advantageous ratio of liquid to solid.

Ratio of liquid to solid	1:1	1:1½	2:1
Extraction in %% for unclassified	92.42	84.28	84.84
Extraction in %% for slime	91.29	92.87	95.49

Consumption of reagents. Consumption of KCN without addition of alkali for unclassified material: 2.4 lb per 100 poods.
\((0.42 - 0.36 = 0.06\%\), or \(0.06 \times 40 = 2.4\) lb per 100 poods).

Consumption of KCN in the presence of alkali: 0.6 lb per 100 poods.
\((0.42 - 0.405 = 0.015\%\), or \(0.015 \times 40 = 0.6\) lb).

Useful consumption of alkali: 1.11 lb NaOH per 100 poods.
\(\left(\dfrac{2.4 - 0.6}{65.11} \cdot 40 = 1.11\right)\), or 0.78 lb CaO.

Total consumption of alkali (total acidity): 3.2 lb NaOH per 100 poods.
\((1.96 - 1.88 = 0.08\%\), or \(0.08 \times 40 = 3.2\) lb), or 2.24 lb CaO.

There is no free acidity or alkalinity; the reaction of the aqueous extract is neutral.

The consumption of reagents on the unclassified material does not exceed what is normally permissible, so that no auxiliary operations are required. It is evident that the same results are also applicable to the products of classification, i.e., to the slimes and to the sands.

Percolation test was carried out in parallel with the unclassified material and with the sands. In both cases 3 kg were taken. The quantities of solutions poured in, and their strength before and after treatment of both materials, are given in Table XXXVI.

For the unclassified material the extraction proved to be 86.1%; and for the classified material, 80.3%. In the first case, 6.34–0.88 zolotniks per 100 poods remained in the tailings; in the second, 6.53–1.29. The latter figure indicates that the grinding of the sands is not entirely sufficient.

The rate of percolation, even for the unclassified material, proved quite satisfactory, namely 2.2″ per hour.

Agitation of slimes. Into the laboratory pneumatic agitator shown in Fig. 9, 200 g of slime were charged and agitated with a ninefold, by weight, quantity of 0.32% KCN solution for 12 hours. The extraction, determined from the solution, proved to be 93.1%, and out of 8.35 zolotniks per 100 poods only 0.58 zolotniks remained in the tailings. In the agitation test the strength of the solution was increased somewhat as compared with the most advantageous one, since during aeration in small volumes of solutions the KCN contained in them falls rapidly. The duration of treatment was also somewhat increased.

Conclusions. On the basis of the results obtained one may draw a conclusion as to whether separate treatment of slimes and sands is necessary.

We have seen that in the tailings from treatment in the percolator of unclassified material there remains 0.88 zolotniks per 100 poods.

With separate treatment of the same 100 poods, consisting of 66.5 poods of sands and 33.5 poods of slime (according to the screen analysis, with correction for sprinkling), there will remain in them

\[ \frac{66.5 \times 1.29}{100} + \frac{33.5 \times 0.58}{100} = 0.86 + \]

\[ + 0.19 = 1.05 \text{ zolotniks,} \]

i.e., somewhat more than with combined treatment. Thus, classification is not needed, and the whole material may be treated entirely in percolation vats.

If additional grinding were introduced for the sands, the extraction would rise by \(93.1 - 86.1 = 7.0\%\), or by \(0.88 - 0.58 = 0.30\) zolotniks per 100 poods of material. Whether these 0.3 zolotniks would repay the cost of grinding 66.5 poods of sands, and what material benefit would thereby be obtained, can be decided only with all data of an economic and technical character in hand.

TABLE XXXVI. Solutions and their KCN content before and after treatment.

Unclassified material

Order of charging the solutions	GOLD, cm³	GOLD, % by weight of material	KCN content in solutions, wt. %	KCN content in solutions, g	Passed through, cm³	KCN content in filtrates, wt. %	KCN content in filtrates, g	Loss or decrease in KCN, g
I (contact with material for 24 h.)	1440	48	0.205	2.95	1440	0.165	2.37	−0.58
(Aeration 12 h.)	—	—	—	—	—	—	—	—
II (contact with material for 4 h.)	1080	36	0.052	0.56	1100	0.047	0.51	−0.05
III washing water	1080	36	—	—	1015	0.0028	0.03	+0.03
IV washing water	1080	36	—	—	—	—	—	—
Total in solutions				3.51		Total in filtrates	2.01	−0.60 (0.8 lb per 100 p.)

Tailings \((\Sigma + 60 + 200)\)

Order of charging the solutions	GOLD, cm³	GOLD, % by weight of material	KCN content in solutions, wt. %	KCN content in solutions, g	Passed through, cm³	KCN content in filtrates, wt. %	KCN content in filtrates, g	Loss or decrease in KCN, g
I (contact with material for 24 h.)	1440	48	0.205	2.95	1440	0.175	2.52	−0.43
(Aeration 12 h.)	—	—	—	—	—	—	—	—
II (contact with material for 4 h.)	1080	36	0.052	0.56	1080	0.047	0.51	−0.05
III washing water	1080	36	—	—	1113	0.004	0.04	+0.04
IV washing water	1080	36	—	—	—	—	—	—
Total in solutions				3.51		Total in filtrates	3.07	0.44 (0.59 lb per 100 p.)

XXIII.

Study of the Ores of the Khotimskaya Vein at the “6th Berikul Area” Mine

The “6th Berikul Area” mine is located in the Mariinsk district of Tomsk Province, 70 versts south of the Tyazhin station of the Tomsk Railway.

Among the numerous veins at this mine, the Khotimskaya vein has long attracted the attention of the mine owners, both because of its considerable thickness and because it is gold-bearing. From the article by P. P. Gudkov* devoted to the description of the “6th Berikul Area” mine, we borrow some data concerning the Khotimskaya vein.

The vein lies at the contact of limestone with porphyrite. In the vein mass the chief role is played by coarsely crystalline calcite; quartz, however, is contained in separate cavernous pieces only in the decomposed part of the vein, near its outcrop on the slope. Of the ore minerals in the vein, the dominant importance belongs to pyrite, which sometimes forms almost continuous accumulations throughout the entire thickness of the vein; arsenical pyrites is almost entirely absent.

Despite its composition (almost entirely of pyrite alone), the Khotimskaya vein was distinguished by a fairly considerable gold content, namely: in the decomposed part it contained from 20 to 40 zolotniks per 100 poods, while in the pyritic part it yielded 2–3 zolotniks of mechanical gold and 5–6 zolotniks of chemical gold; meanwhile, in other veins at the same mine, the pyrite apparently contains almost no gold.

The aim of the present investigation was to determine as accurately as possible the average gold content in the ores of the Khotimskaya vein and to establish a method of extracting gold from them that would be technically and economically expedient under the conditions of the Berikul mine.

Special attention was paid to obtaining a proper average sample. This task was considerably facilitated by the extensively developed underground workings, which made it possible to take samples along almost the entire dip of the vein.

Average sample. From fifteen faces, the location of which is indicated in Table XXXVII, separate samples were taken alternately, now closer to the footwall side (limestone), now closer to the hanging-wall side (porphyrite).

In doing so, chiefly the material of the vein itself was taken; but, in addition, separate samples were also taken from the side rocks. From each face 30–40 poods of ore were obtained, and the total weight of the first sample amounted to about 500 poods of lump ore with a small amount of fines.

* Izvestiya Tomskogo Tekhnologicheskogo Instituta, vol. XXI, 1911, No. 1.

TABLE XXXVII. Distribution of workings in the Hot Zone, residential.

Working No.	Position of the face	Predominant minerals
15	In the Khotimskaya adit, 18 sazhens above the level of the Sukhoi Beregulya	Exclusively pyrite; pyrite II predominates.
14	Same	Pyrite with inclusions of chalcopyrite.
13	In the eastern adit, connecting with the Petropavlovskaya adit	Pyrite, with inclusions of zinc blende.
12	In drift No. 4 toward the contact	Exclusively pyrite.
11	In the lower Petropavlovskaya adit, near the Razvalistaya vein	Same.
10	Beginning of the raise at the Petropavlovskaya adit	Same.
9	In a stope near the limestone	Pyrite with inclusions of zinc blende.
8	In a stope near the porphyry	Pyrite and mispickel; mispickel predominates; zinc blende and galena occur as inclusions.
7	At the end of crosscut No. 3 into vein No. 4	Pyrite with a considerable admixture of zinc blende.
6	At 2 sazhens below the third level of the western adit and at 22 sazhens below the base	Pyrite.
5	In a stope of vein No. 4, near the limestone	Pyrite with loose inclusions of zinc blende.
4	Same, near the limestone	Same.
3	Same, near the limestone	Exclusively pyrite.
2	Same	Same.
1	Near the hanging wall of vein No. 4	Pyrite with loose inclusions of zinc blende.

At 8 sazhens below face No. 1, work ceases.

Reduction of the first sample. For the laboratory investigations it was sufficient to have 6–8 poods of ore. Therefore the first sample was subjected to reduction. All the ore was passed through a Blake crusher, \(410 \times 310\) mm, with a gap of 25 mm, after which the ore passed along an inclined chute into rolls \(890 \times 290\) mm, set to a gap of 6 mm. The ore thus crushed was quartered twice; in this operation the mixing of the ore was performed by throwing it onto a cone around an iron rod serving as the axis of the cone. The reduced sample, amounting to \(1/4\) of the original quantity of ore, was crushed in rolls set to a gap of 2 mm until the entire sample had passed through a sieve with openings of 4 mm. The crushed sample was reduced by successive quartering to \(17 \tfrac{1}{2}\) poods, from which two laboratory samples of 3 poods each were taken by fractions. The remainder of the sample was left in the ore bin. The degree of crushing of the latter laboratory sample is characterized by the sieve analysis given in Table XXXVIII.

TABLE XXXVIII. Sieve analysis of 1000 g of ore.

Classes	Quantity, g	Quantity, %
\(+\,3\) mm (sieve II)	87	8.7
\(+\,2\) mm (sieve III)	242	24.2
\(+\,0.75\) mm (sieve IV)	170	17.0
\(+\,40\)	222	22.2
\(+\,50\)	21	2.1
\(+\,60\)	41	4.1
\(+\,80\)	34	3.4
\(+\,100\)	36	3.6
\(+\,120\)	36	3.6
\(+\,150\)	11	1.1
\(+\,200\)	24	2.4
\(-\,200\)	76	7.6
Total	1000 g	100%

This degree of comminution must be recognized as sufficient and as conforming to the following rule of R. Richards,* derived by him from practical data: for each given ore, the quantity of the latter taken as a sample must be proportional to the square of the diameter of the largest piece, or the weights of the sample in successive reductions must stand in the same ratio as the squares of the diameters of the largest pieces. On the basis of this rule Richards compiled the following table—see Table XXXIX.

TABLE XXXIX. Richards’s Table.

Weight of the first sample, in English pounds	Ores of average content: size of the largest piece, mm, from	Ores of average content: size of the largest piece, mm, to
20000	76.2	50.8
10000	53.9	35.9
5000	38.1	25.4
2000	24.1	16.1
1000	17.0	11.4
500	12.0	8.0
200	7.6	5.1
100	5.4	3.6
50	3.8	2.5
20	2.4	1.6
10	1.7	1.1
5	1.2	0.8
2	0.76	0.51
1	0.54	0.36
0.5	0.36	0.25
0.2	0.24	0.16
0.1	0.17	0.11

In the first vertical column of Richards’s table is shown the weight of the first sample; in the second column, the size of the pieces for ores of average gold content. When using the table in reducing a sample, there is, of course, no need to comminute the sample successively to all the sizes of pieces indicated in the second column. The ore may be comminuted to any practically convenient size of pieces, and then only the quantity corresponding to that size and indicated in the first column may be taken for the sample; and conversely, if it is desirable to take for the sample the quantity of ore indicated in the first column, then all the ore must first be comminuted to the corresponding size of pieces indicated in the second column. Thus, for example, if the weight of the first sample amounts—

* R. Richards. Ore Dressing, Vol. II, p. 287.

amounts to 20,000 English pounds (about 500 poods), and the maximum size of the pieces of ore is 76 mm, then, when crushing all the ore to 3.6–5.4 mm, the sample may at once be reduced to 100 pounds (about 3 poods), which was in fact done in reducing the sample of the ore under investigation.

At concentration mills and plants in America, when sampling ordinary gold-bearing ores, the following table, compiled from practical data,* is used—see Table XL.

TABLE XL. Minimum weight of sample for ores of various degrees of crushing.

When crushing the ore to	Minimum weight of sample, lb
2 in.	10000
1½ in.	5000
1 in.	2000
¾ in.	1000
½ in.	400
⅜ in.	300
¼ in.	200
⅛ in.	100
⅛ in.	75
6 openings per 1 linear in.	50
10 openings per 1 linear in.	25
18 openings per 1 linear in.	10
30 openings per 1 linear in.	4
50 openings per 1 linear in.	1

Content of gold and silver in the average sample. Half of the sample, weighing 3 poods, was ground in a ball mill, screened through a sieve with openings of 0.75 mm, and reduced by means of Braun’s automatic sampler to 16 kg. The sieve analysis of this sample is given in Table XLI.

TABLE XLI.

Classes	%
+100	56.2
+150	8.6
+200	4.0
−200	30.00
Sum	98.8

* J. B. Woodbridge. Ore sampling conditions in the West. Technical Paper 86, Bureau of Mines U. S. A. Washington, 1916.
D. W. Brunton. Theorie and practice of ore sampling. Trans. A. I. M. E., vol. 25, 1897, 826.

From the 16 kg, by fractional division three samples were taken:

1 kg for analyses,
10 kg for amalgamation,
5 kg for cyanidation.

The last fraction, of 5 kg, was subjected to fine grinding in a ball mill. The sieve analysis of the ground material was as follows—see Table XLII.

TABLE XLII

Classes	%
+100	0.8
+150	4.7
+200	7.7
−200	86.8
Total	100.0 *)

In this material the content of gold and silver was determined by means of crucible assay, both with preliminary roasting and without roasting. The results of the crucible assays are given in Table XLIII.

TABLE XLIII. Results of crucible assays.

Sample No.	mg of gold in 100 g of ore	mg of silver in 100 g of ore	Note
1	2.2	3.99	samples with preliminary roasting
2	2.3	—	samples with preliminary roasting
3	2.3	—	samples with preliminary roasting
4	2.3	3.41	samples with preliminary roasting
5	2.35	—	samples with preliminary roasting
6	2.20	—	samples with preliminary roasting
7	2.27	—	samples with preliminary roasting
Average	2.28	3.70	samples with preliminary roasting
8	2.30	—	without roasting
9	2.40	—	without roasting
Average	2.35	—	without roasting

1 mg Au in 100 g of ore = 3.84 zolotniks in 100 poods.

*) Subsequently, in the results of the sieve analyses, a correction was introduced for loss by spilling, which was distributed proportionally among the classes +150, +200, and −200.

Average content from seven samples:
Au—8.76 zol. (with roasting)

" " from two "
Au—9.02 " (without roasting)

" " "
Ag—14.21 "

Since in the subsequent assaying was carried out only with roasting of the ore before crucible smelting, the following values were adopted in all calculations for the average content:

gold—8.76 zol. in 100 pud.
silver 14.21 " " "
Fineness of the gold . . . 381.4

Chemical and mineralogical composition of the ore. To establish the method of treating the ore and to clarify the difficulties that might be encountered, it was necessary to know the complete chemical and mineralogical composition of the ore, as well as the gold content both in the separate minerals making up the ore and in the accompanying rock.

The quantitative analysis of the ore gave the following results:

Component	Content
Fe	32.30%
S	37.52 "
Zn	2.56 "
Cu	0.10 "
Pb	traces
Mn	0.20 "
As	0.11 "
CaO	4.73 "
MgO	1.66 "
CO₂	5.40 "
SiO₂	13.22 "
total	97.80 "

The rational mineralogical composition of the ore, calculated from the data of the quantitative analysis, is as follows:

Mineral	Content
pyrite (FeS₂)	69.39%
zinc blende (ZnS)	3.86 "
copper pyrites (CuFeS₂)	0.29 "
mispickel (FeAsS)	0.22 "
limestone (CaCO₃)	8.44 "
magnesite (MgCO₃)	3.47 "
quartz (SiO₂)	13.22 "
total	98.89 "

Among the ore minerals in the vein, pyrite predominates; second place is occupied by zinc blende, the quantity of which is quite small. Mispickel, galena, and copper pyrites occur as accidental admixtures.

A somewhat noticeable accumulation of these minerals was found in stopes No. 8 and No. 14, as is evident from Table XXXVII.

From the first general sample of lump ore, individual pieces were selected in which one or another mineral predominated. By gradual crushing and sorting with sieves, the individual minerals were isolated and taken in quantities from 100 to 1000 g. All the minerals were crushed to a No. 150 sieve and subjected to assaying and quantitative analysis. In the mass of the ore it was easy to distinguish two varieties of pyrite.

One variety—pyrite I—with a clear silvery luster; the other—pyrite II—somewhat decomposed, finely crystalline, with a gneiss-yellowish tarnish.

The composition and content of the individual minerals are given below:

Component	Pyrite I	Pyrite II
Fe	43.54%	44.52%
S	48.60%	49.40%
Au	14.22 zolotn. in 100 poods	4.84 zolotn. in 100 poods
Ag	2.67 zolotn. in 100 poods	28.19 zolotn. in 100 poods

Component	Zinc blende	Galena
Zn	56.0%	—
S	32.57%	15.03%
Fe	8.23%	—
Pb	—	64.72%
Au	0.70 zolotn. in 100 poods	13.24 zol. in 100 poods
Ag	5.38 zolotn. in 100 poods	692.90 zol. in 100 poods

Component	Mispickel	Copper pyrites
Fe	34.50%	31.66%
As	42.17%	—
S	21.22%	—
Cu	—	29.92%
Au	12.06 zolotn. in 100 poods	trace
Ag	10.98 zolotn. in 100 poods	53.76 zol. in 100 poods

Mineral	Composition	Gold and silver
Quartz	almost pure SiO₂	none
Limestone	CaCO₃	none

Specific gravity of the ore. The sample was crushed to a No. 200 sieve. The specific gravity was determined in kerosene, in Schumann’s apparatus. Kerosene was used because, when the ore was ground—as further investigations showed—ferrous sulfate was formed, which, when the ore was wetted with water, reacted with the limestone contained in the ore and gave gypsum, soluble in water.

Specific gravity at 15.9°—3.953.

On the basis of the analytical data, we may characterize the ore investigated as gold-bearing pyrite with insignificant inclusions of zinc blende, mispickel, copper pyrites, and galena. Gold is associated only with the sulfides and is absent in the barren rock. The only valuable constituent in economic terms—

in relation to the metal in the ore is gold. In silver content the ore belongs to the category of very poor ores.

General remarks. The ore investigated is pyrite with an average gold content, and the aim of all methods of treating the ore that are technically and economically feasible under the conditions of the Berikul mine is the most complete possible extraction of the gold.

The chemical and mineralogical composition of the ore and the local conditions exclude the possibility of applying to this ore the methods of smelting, chlorination, and amalgamation.

For smelting, the ore contains no collectors of the noble metals, such as copper or lead. Neither at the mine nor in the nearest surroundings are there any known deposits of copper or lead ores of any significance*). Moreover, there is no suitable mineral fuel at the mine, and the delivery of such fuel, owing to the absence of a railway, is very difficult. The comparatively low gold content, the low fineness of the latter (excess of silver), the high content of limestone and magnesite, the absence of cheap fuel, and the high cost of chlorination completely exclude the economic possibility of applying this process for extracting gold from the ore investigated.

The association of gold exclusively with sulfides and the high specific gravity of the ore predetermine the question of the inapplicability of amalgamation as the principal process for extracting gold. With ordinary amalgamation, for example in a bowl, besides the very small percentage of gold extraction, the silting-up of the sluices and a large loss of mercury due to spraying and flouring of the latter are inevitable.

Despite the obvious unsuitability of amalgamation as the principal process for extracting gold from this ore, it was nevertheless necessary to carry out an experimental verification of this assumption in order to ascertain whether amalgamation would be necessary at least as an auxiliary or intermediate stage in the process of gold extraction. To this end, amalgamation tests were carried out at the mine and in the laboratory.

Amalgamation.

Amalgamation at the mine. From an average sample of ore, crushed in rolls, 450 poods were subjected to amalgamation in one of the Berdan bowls at the factory.

Principal dimensions of the bowl and Berdans.

Dimension	Value
External lower diameter of the bowl	9′8″
“ upper ” ”	12′5″

*) The copper outcrops examined by us in the summer of 1917 on the left bank of the Kundat River, about 15 versts above its confluence with the Kiya River and 20 versts from the mine, consist of a small veinlet of quartz stained with copper green and blue, among destroyed slates.

Height of the bowl	2′11″
Distance between the axes of the millstones	9″
Diameter of the tire	4′9″
Thickness of the tire (badly worn)	2″
Width of the tire	11″
Height of the sill	12½″
Length of the screen	2′2″
Width of the screen	9″
Number of openings in the screen	3400
Diameter of an opening	2 mm
Number of runner revolutions per minute	14

Sluice. The fixed sluice is assembled from copper, amalgamated sheets. Width of the sluice—2′2″; length—16′8″; slope—5°.

Below are given the details and results of the trial amalgamation.

Ore treated	450 poods
Duration of the experiment	11 hours
Loaded per hour	45 poods
Lime added per 100 poods of ore	22 pounds
Ratio of liquid to solid	20:1
Mercury used for rubbing the sheets	800 g
Squeezed out of the amalgam	540 g
Loss of mercury	260 g
Dry amalgam obtained	56 g
Gold retorted	21.5 g = 5 zolotniks
Pure gold obtained	4 zolotniks

The sluice became heavily flooded; the sheets turned green, and considerable spraying of mercury was observed. From the data of the experiment performed it follows that:

1. the loss of mercury amounted to 360 g, or 61 zolotniks, which per 100 poods gave 13.6 zolotniks. The normal loss of mercury in external amalgamation is considered to be 2–5 zolotniks per 100 poods;

2. with a gold content in the ore of 8.83 zolotniks (see p. 93), the percentage extraction of gold in the experiment amounted to 10%.

During the trial amalgamation, every ½ hour samples of tailings were taken in front of the head sheet of the sluices across the entire width of the screen, and in the pocket behind the last sheet along the entire length of the latter. The samples were collected in separate vessels; the clarified water that had settled was drained off with a siphon, and the samples were squeezed and dried until the moisture was completely removed.

TABLE XLIV. Sieve analysis of tailings upon discharge from the bowl

Classes	%
+ 100	17.5
+ 150	21.5
+ 200	15.0
− 200	45.5
sum	100%

The gold content in the slimes proved to be 8.83 zolotniks per 100 poods; the gold content in the tailings from amalgamation was 8.06 zolotniks per 100 poods. Consequently, the percentage extraction of gold amounted to 8.7%.

The extraction ~~from the tailings~~ proved to be 1.3% less than according to the amount of gold obtained from the amalgam. This is explained by the difficulty of obtaining a correct average sample of the tailings. In view of the small slope of the sluices, the high specific gravity of the ore, the great difference in the specific gravities of the individual components of the ore, and the excessively great height of the sill, classification of the slimes by grain size and by individual minerals took place both partly in the bowl itself and, chiefly, on the sluices, which had a very great length of 16′8″. Therefore, in the present case, for calculating the extraction it appears more correct to proceed from the average gold content in the ore before amalgamation and from the quantity of gold obtained during amalgamation. In any case, the experimental amalgamation in the bowl proved the complete inexpediency of this process for treating the given ore.

Laboratory amalgamation experiments. To verify the results of the experimental amalgamation at the plant, laboratory amalgamation experiments were carried out, in which it was possible to vary the working conditions.

10 kg of the average sample of the sieve analysis, given on p. 85, were ground in a ball mill and subjected to amalgamation on a laboratory sluice and in a ball mill. From the ground ore before amalgamation, a sample was taken for sieve analysis and for determining gold by classes, the results of which are given in Table XLV on p. 94.

Amalgamation on a sluice. A pulp consisting of 2 kg of ore, 24 kg of water, and 20 g of CaO was mixed in an agitator and passed over the sheet 10 times; the tailings were washed twice from mercury, dried, and assayed.

Amalgamated ore . . . . . . . 2000 g
Tailings washed . . . . . . . 1710 ”
Ore lost . . . . . . . . . . 290 g = 14.5%
Gold content in 100 poods of tailings—7.68 zolotniks.
Percentage extraction . . . . . 12.3%.

TABLE XLV. Sieve analysis of the ore and assaying of gold by size classes.

Classes.	%	Gold content in 100 poods of each class.	Amount of gold in all classes.
+ 100	14.4	9.21	1.33
+ 150	12.1	7.68	0.93
+ 200	9.0	5.95	0.53
− 200	64.5	9.41	6.10
sum	100%	8.76	8.89

Amalgamation in a ball mill. Into the mill were charged 2000 g of ore of the same degree of comminution as for the preceding experiment, 500 cc of water, 10 g of CaO, and 205 g of mercury, which amounted to 10.2% by weight of the ore. The duration of amalgamation was 2 hours. The ore was washed free of mercury, a considerable part of which was slimed.

Washed-out mercury . . . . . . . . . . 186 g.
Lost ” . . . . . . . . . . . . . . . . . 19 g.—9%
Weight of dried tailings . . . . . . . 1940 g.
Lost ore . . . . . . . . . . . . . . . . 60 ” or 3%
Gold content in 100 poods of tailings: 6.72 zolotniks.
Percentage extraction of gold . . . . 23.3%

In view of the very low extraction of gold, the assaying of the latter by classes in the tailings was not carried out.

Comparing the results of all the amalgamation experiments, we obtain the following:

Percentage extraction of gold during amalgamation at the plant—10.0%
” ” ” ” ” in the sluice—12.3 ”
” ” ” ” ” in the ball mill—23.3 ”

It follows from this that, for this ore, amalgamation, even as an auxiliary or intermediate process, is entirely inexpedient and should not be applied in the process of ore treatment.

Cyanidation.

Of the possible variants of the cyanide process, percolation and siltation with the application of mechanical agitation were tested.

Sieve analysis and the distribution of gold by classes in the material sent for amalgamation (see Table XLV) do not yet provide sufficient grounds for deciding the question of the suitability or unsuitability of the percolation method for this ore. But, proceeding from the fact that the gold in the ore is associated only with sulfides, which under any method of comminution will give a very large quantity of slime, it could be said in advance that the percolation method for this ore is economically unsuitable.

Microscopic study of the individual classes of the sieve analysis showed that the material contains much non-crystalline slime, which encrusted the larger crystalline particles, both of the rock and of the sulfides. This promised a low rate of percolation and, consequently, a long duration of the process with low gold extraction.

The laboratory tests were to determine the rate of percolation, the percentage extraction of gold, and the consumption of potassium cyanide.

For the cyanidation tests there were four degrees of ore comminution, designated as ores I, II, III, and IV. These materials were obtained by grinding separate portions of the average sample in a ball mill for various intervals of time, from 3 to 12 hours.

Sieve analyses of the above-mentioned degrees of comminution are given in Table XLVI.

TABLE XLVI. Sieve analyses.

Classes	Ore I	Ore II	Ore III	Ore IV
	in percent	in percent	in percent	in percent
+ 100	56.2	14.4	0.8	1.2
+ 150	8.6	12.1	4.4	1.7
+ 200	4.0	9.2	9.2	4.0
− 200	31.2	65.5	85.6	93.1
--	100.0	100.0	100.0	100.0

Percolation. For the percolation method, ore I was taken, as promising the highest rate of percolation. Into the laboratory percolator were charged 3000 g of ore, mixed with 15 g CaO (0.5% by weight of the ore), and solutions of potassium cyanide and water were poured in the order and quantities shown in Table XLVII.

TABLE XLVII.

Solution poured in, cu. cm	KCN, %	KCN, g	Solution obtained, cu. cm	KCN, %	KCN, g
I 1100	0.33	3.63	I 1100	0.14	1.54
II 1100	0.14	1.54	II 1100	0.105	1.155
III 1000	0	0	III 1000	0.013	0.13
IV 500	0	0	—	—	—
Total 3700	—	5.17	3200	—	2.83

The total quantity of solutions for 3000 g of ore amounted to 3700 cu. cm, or 123.3% by weight of the ore.

Consumption of KCN . . . . . . . . . . . . \(5.17 - 2.83 = 2.34\) g.
    “     “    per 100 g of ore . . . \(2.34 : 30 = 0.078\) “
    “     “    per 100 poods “ . . . \(0.078\text{ p.} = 3.12\) lb.

Tailings retained moisture . . . \(\dfrac{3700 - 3200}{30} = 17\%\)

The rate of percolation was . . . . . . . . \(0.5—0.6''\) per hour.
Gold extracted in the solutions . . . . . . 2.42 zol. per 100 poods.
    “      silver                “ . . . . . . 0.36 “ “ “ “
Percentage extraction of gold . . . . . . . 27.6% “ “ “ “

The very low rate of percolation, lying below the economic minimum permitting the use of percolation and amounting to \(1 \tfrac{1}{2}—2\) inches per hour, and the very insignificant extraction of gold confirm the assumption made above a priori as to the unsuitability of the percolation method for treating the ore under investigation.

Slime process. The second variant of cyanidation tested in the laboratory was the slime process with the use of agitation.

The laboratory experiments were to determine:

1. the maximum extraction of gold,
2. the shortest duration of the process,
3. the most advantageous degree of grinding, and
4. the consumption of potassium cyanide for treatment of the ore.

Maximum extraction was determined for ore III. Three charges of ore of 250 g each + 500 cu. cm of KCN solution (0.5% KCN) + 0.5 g CaO + 0.5 g barium peroxide were agitated in bottles for 48 hours. The extraction of gold and silver was determined in the solutions, and the following results were obtained:

average maximum extraction of gold from 100 poods — 7.95 zol.

» » » silver » » » 7.45 »

Consequently, the maximum percentage extraction proved to be:

\[ \text{for gold } \frac{7.95}{8.76}\,100 = 90.76\% \]

\[ \text{for silver } \frac{7.45}{14.21}\,100 = 52.43\%. \]

Thus, 90.76% is the laboratory maximum of extraction for degree of grinding III and an agitation duration of 48 hours.

The next question, concerning the shortest duration of treatment, was resolved by means of a series of analogous experiments, in which the duration of agitation varied from 1 to 84 hours, while all the other conditions remained identical, namely:

amount of ore — from 250 to 600 g.
amount of KCN solution — from 500 to 1200 cu. cm.
strength of the solution — from 0.49 to 0.51%
ratio of liquid to solid — 2:1
amount of CaO — 0.2% by weight of the ore
barium peroxide — 0.2 » » »

The results of this series of experiments are given in Table XLVIII.

TABLE XLVIII. Duration of treatment.

Duration of agitation, hours	Consumption of KCN, funts/100 poods	Gold extraction, zolotniks from 100 poods	Gold extraction, % of gold in the ore	Silver extraction, zolotniks from 100 poods	Silver extraction, % of silver in the ore
1	10.8	7.29	83.22	4.34	30.54
3	14.4	7.41	84.6	4.92	34.64
6	16.8	7.68	87.7	8.10	57.0
12	24.0	7.95	90.76	7.95	58.95
24	26.8	7.49	85.62	—	—
48	27.76	7.95	90.76	7.45	52.43
84	32.0	7.84	89.50	8.30	58.40

The experiments show that:

1. the ore very readily gives up gold to a potassium cyanide solution;
2. agitation for 6 hours gives an extraction close to the maximum,

3, after 12-hour agitation the extraction of gold does not increase. Hence it follows that in practice, on a plant scale, with sufficient aeration, one may expect the maximum extraction after 6–8 hours of agitation.

The most advantageous degree of comminution. To clarify this condition, a series of experiments was carried out with ore of five degrees of comminution. The results of these experiments are given in Table XLIX.

TABLE XLIX. The most advantageous degree of comminution.

Duration of agitation, h	Classes	Ore I, %	Ore V, tailings from the bowl, %	Ore II, %	Ore III, %	Ore IV, %	Ore I, gold	Ore I, %	Ore V, gold	Ore V, %	Ore II, gold	Ore II, %	Ore III, gold	Ore III, %	Ore IV, gold	Ore IV, %
	+100	56,2	17,5	14,4	0,8	1,2
	+150	8,6	21,5	12,1	4,4	1,7
	+200	4,0	15,0	9,0	9,2	4,0
	−200	31,2	45,5	65,5	85,6	93,1
1							—	—	—	—	6,91	78,9	7,29	83,22	—	—
3							—	—	—	—	7,22	82,42	7,41	84,60	—	—
6							—	—	6,91	78,51	7,53	85,96	7,68	87,70	—	—
8							—	—	6,91	78,51	—	—	—	—	—	—
12							—	—	—	—	7,99	91,22	7,95	90,76	—	—
24							6,14	70,1	6,7	76,2	7,99	83,34	7,49	85,62	7,70	87,90
48							—	—	—	—	7,53	85,96	7,95	90,76	—	—

If we allow that the extraction at 24-hour agitation for ore III proved, merely by chance, to be lower than the extraction for the same ore at 6 and 48 hours of agitation, i.e., 85.96%, then, comparing the figures of percentage extraction for the five degrees of comminution, we may conclude that extraction increases with an increase in the amount of the −200 class, and degrees of comminution III and IV give almost the same extraction. Hence it follows that, in work on a plant scale, there is no need to grind more than corresponds to a degree of comminution lying between II and III; i.e., the amount of −200 material in the ground ore may vary within the limits from 70 to 85% without any significant difference in the percentage extraction.

Another method for determining the most advantageous degree of ore grinding for cyanidation is to fix the gold by classes before and after cyanidation. In view of the fact that after agitation even for 6 hours the amount of gold remaining in the tailings is very small and does not exceed 1—1½ zolotniks per 100 poods, for fixing the gold by classes ore II was agitated for only 3 hours. The results obtained thereby are given in Table L.

TABLE L. Fixing of gold by classes.

Classes	Gold content in 100 poods of the class. Before cyanidation. Zolotniks.	Gold content in 100 poods of the class. After cyanidation. Zolotniks.	Gold extracted from 100 poods of each class. Zolotniks.	Gold extracted from 100 poods of each class. %
+ 100 . . . .	9,21	9,21	0	0
+ 150 . . . .	7,68	2,41	5,25	68,36
+ 200 . . . .	5,95	2,30	3,65	61,34
− 200 . . . .	9,41	1,15	8,26	87,78

Fixing the gold by classes in the cyanidation tailings shows that the greatest extraction is given by the −200 class, as could have been expected.

Behavior of the separate minerals entering into the composition of the ore during cyanidation. A very important question was the role of the separate minerals entering into the composition of the ore during cyanidation, with respect to the extraction of gold from them and the consumption of potassium cyanide in the process.

For this purpose, experiments were carried out on cyanidation of these minerals, ground to −150 mesh and taken in quantities from 50 to 250 g. The experiments were performed by agitating the material with a solution of potassium cyanide of strength 0.4—0.5% at a liquid-to-solid ratio of 2:1, in the presence of CaO in an amount of 0.2% by weight of the material, for 24 hours. The results of the experiments are given in Table LI. (See p. 100).

Sphalerite, being very poor in gold, was not investigated. The most refractory proved to be arsenopyrite, which gave an extraction of only 25%. It is necessary to note the small consumption of KCN for the treatment of the separate sulfides in comparison with the consumption of the same reagent for treating the ore. The reasons for this phenomenon will be clarified below.

Since pyrite predominates in the ore, which, according to the produc—

TABLE LI. Cyanidation of minerals.

Minerals	Gold content in 100 poods, zolotniks	Gold extraction, zolotniks per 100 poods	Gold extraction, %	KCN consumption, pounds per 100 poods
Pyrite I	4.22	3.84	91	4
Pyrite II	4.84	4.61	95.3	7.2
Galena	13.24	9.22	70	4
Mis pickel	12.06	3.10	25	0.8

...by the experiments given, an extraction of 91–95%, this makes it possible to conclude that the extraction of gold from the ore on a factory scale will be no lower than 91%.

Consumption of potassium cyanide for treating the ore. In the laboratory experiments the consumption of potassium cyanide proved to be very high. The exception is ore V, or the slimes from the bowl before amalgamation, carried out at the factory.

The results of experiments to determine KCN consumption are given in Table LII.

TABLE LII. Consumption of potassium cyanide.

Duration of agitation, hours	Added CaO, % by weight of ore	Ore I	Ore II	Ore III	Ore IV	Ore V (slimes)
Duration of agitation, hours	Added CaO, % by weight of ore	KCN consumption in pounds per 100 poods of ore	KCN consumption in pounds per 100 poods of ore	KCN consumption in pounds per 100 poods of ore	KCN consumption in pounds per 100 poods of ore	KCN consumption in pounds per 100 poods of ore
1	0.2	—	4.8	10.8	—	—
3	0.2	—	5.6	14.8	—	—
6	0.2	—	8.8	17.6	—	3.2
8	0.2	—	—	—	—	3.44
12	0.2	—	10.4	24.0	—	—
24	0.2	4.8	14.8	26.8	32.8	2.4
24	—	—	25.6	33.6	36.8	—
8	0.3	—	—	—	—	3.2
8	0.6	—	—	—	—	2.8
8	0.8	—	—	—	—	2.8
24	0.3	—	—	—	—	2.4

From Table LII we see that the consumption of KCN increases both with the duration of agitation of the ore with the KCN solution and with a decrease in the size of the ore grain. Thus, for example, when the ore was agitated for one hour the consumption of KCN was 48 pounds, while with 24-hour agitation of the same ore it reached 14.8 pounds per 100 poods. The influence of the degree of grinding is clearly seen from a comparison of the KCN consumption for ores of different degrees of grinding under 24-hour agitation and otherwise identical conditions. In this case the KCN consumption rose from 4.8 pounds for ore I to 32.8 pounds for ore IV. It is obvious that, in grinding the ore, substances were formed that decomposed KCN or reacted with it. To clarify the question, aqueous extracts were made from ore of various degrees of grinding. In these extracts only SO₃ and CaO were found; iron compounds, both ferric and ferrous, were entirely absent. In ore freshly mined there was neither gypsum nor other sulfates soluble in water. This led to the conclusion that, when the ore is ground, the sulfide predominant in it—pyrite—is oxidized. This fact is known in the literature and has also been observed repeatedly in our laboratory. Thus, in the article “The exact determination of sulphur in pyrite and marcasite,” E. Allen and J. Johnston*) state that, for the accurate determination of sulfur in pyrite, the latter should be ground only coarsely, approximately to 20 meshes per linear inch, since, when pyrite is finely ground in an agate mortar, the odor of the sulfurous gas evolved is clearly perceptible, which, understandably, lowers the results of the sulfur determination. In an aqueous extract of ground pyrite the authors of the above-mentioned article found Fe and SO₃, and the quantities of the latter corresponded exactly to the formula FeSO₄. To prevent the oxidation of FeSO₄ during washing, Allen and Johnston carried out the washing, in an atmosphere of carbon dioxide, with boiled water cooled in carbon dioxide, and obtained the following results.

Duration of grinding: 1 hour.

Charge of pyrite	Fe in the filtrate	Equivalent amount of FeS₂	% of pyrite oxidized
4.11 g	0.0042 g	0.0090 g	0.22%
5.77 g	0.0059 g	0.0125 g	0.22

Duration of grinding: 6 hours.

Charge of pyrite	Fe in the filtrate	Equivalent amount of FeS₂	% of pyrite oxidized
4.21 g	0.0211 g	0.0451 g	1.07
6.13 g	0.0317 g	0.0679 g	1.11

Hydrochloric acid, as is known, does not act on pure pyrite either in the cold or when heated, and it is used for cleaning pyrite. When leaching ground pyrite with 20% hydrochloric acid, Allen and Johnston found in the filtrate only ferrous oxide; ferric oxide [[unclear: text continues on next page]]

*) Journ. Industr. Eng. Chemistry, Vol. 2, (1910), p. 197.

was absent. All these facts prove that \(\mathrm{SO_2}\) and \(\mathrm{FeSO_4}\) are the only products of the oxidation of pyrite during its grinding, and the process may be expressed by the reaction:

\[ \mathrm{FeS_2 + 3O_2 = FeSO_4 + SO_2} \]

Analogous experiments were carried out in our laboratory with Karabash pyrite purified by boiling in 20% hydrochloric acid*). From the purified pyrite, 4 weighed portions of one gram each were taken; two of them were finely ground in an agate mortar. All the weighed portions were heated with dilute hydrochloric acid (1:5) on a water bath, washed with water, and Fe and S were determined in the filtrates obtained. The determinations gave the following results:

	Fe	S
\(\mathrm{FeS_2}\), not ground	was not found	was not found
\(\mathrm{FeS_2}\), ground	4.2 mg	0.6 mg

The same observations—oxidation and formation of \(\mathrm{FeSO_4}\)—were made in grinding pyrrhotite, chalcopyrite, and iron-bearing zinc blende; and it may be accepted as a rule that all sulfides containing ferrous iron, when ground, are partially oxidized and give \(\mathrm{FeSO_4}\).

It is evident that the same phenomena took place in the ore investigated, and the amount of \(\mathrm{FeSO_4}\) in the latter must have increased with decreasing grain size and with increasing duration of grinding. In order to clarify the quantitative side of the question, cold and hot aqueous extracts of ore of various degrees of grinding were prepared, and in the solutions obtained \(\mathrm{SO_3}\) and \(\mathrm{CaO}\) were determined. No iron at all was found in the solutions. This is explained by the fact that, when the ground ore was moistened with water, the \(\mathrm{CaCO_3}\) and \(\mathrm{MgCO_3}\) present in the ore reacted with \(\mathrm{FeSO_4}\), forming \(\mathrm{CaSO_4}\), \(\mathrm{MgSO_4}\), and \(\mathrm{Fe(OH)_2}\); ferrous hydroxide, however, is insoluble in water.

The aqueous extracts were prepared as follows: 25 g of ore were shaken with 400 cc of cold water for 10–12 hours. The mixture was filtered through a dense filter, and in separate portions of the filtrate, 50 cc each, corresponding to 3 g of ore, the contents of \(\mathrm{SO_3}\) and \(\mathrm{CaO}\) were determined. In determining \(\mathrm{SO_3}\), the solution was first oxidized with bromine, since calcium thiosulfite \((\mathrm{CaS_2O_3})\) was found in it, formed during the grinding of the ore from calcite, \(\mathrm{SO_2}\), and the sulfur of pyrite in the presence of air.

To obtain the hot aqueous extract, 25 g of ore were boiled with water for \(1/2\) hour in a 250 cc volumetric flask. After cooling, the contents of the flask were brought up to the mark, and in separate

*) From the Karabash plant near Kyshtym in the Urals.

in samples of 100 cubic cm of filtrate (= 10 g of ore), the contents of SO₃ and CaO were determined.

The results of the determinations are given in Table LIII.

TABLE LIII. Content of SO₃ and CaO in the ground ore.

Ore No.	% SO₃. By cold aqueous extraction.	% CaO. By cold aqueous extraction.	% SO₃. By hot aqueous extraction.	% CaO. By hot aqueous extraction.	Equivalent SO₃ by CaO, %.
I	0,355	—	0,357	0,284	0,405
II	0,579	—	0,583	0,388	0,553
III	0,635	—	0,915	0,672	0,959
IV	0,726	—	0,947	0,546	0,779
Slimes V	0,445	0,296	—	—	—
—	—	require, theoretically,	—	—	—
—	—	0,428% SO₃.	—	—	—

The amounts of SO₃ leached by hot water did not correspond to the amounts of CaO in the same solutions. This is explained by the fact that, in addition to CaSO₄, part of the CaCO₃ and MgSO₄ also passed into the solutions; their amounts were not determined.

Comparing the consumption of KCN per 100 poods of ore with the SO₃ content in it, we obtain the figures in Table LIV.

TABLE LIV. Consumption of KCN in pounds per 100 poods of ore.

(Agitation—24 hours).

Ore.	SO₃ content, %	Product −200, %.	KCN consumption, lb/100 poods. At 0.2% CaO.	KCN consumption, lb/100 poods. without CaO.
I	0,357	31,2	4,8	15,6
II	0,583	65,5	14,8	25,6
III	0,915	85,6	26,8	33,6
IV	0,947	93,1	32,8	36,8
Slimes V	0,445	45,5	2,4	—

The consumption of KCN therefore increases in direct proportion to the amount of \(\mathrm{SO_3}\) in the ore; the amount of \(\mathrm{SO_3}\) increases with decreasing grain size, as is shown by the figures in the column “%—200,” where the amount of product—200 is taken as an indicator of the degree of grinding. These relationships are still more clearly seen from the following diagram 6.

The slimes from the bowl constitute an exception. With a content in them of \(0.445\%\ \mathrm{SO_3}\), the consumption of KCN per 100 poods of slimes amounts to only 2.4 lb. This proves that the compounds which destroy KCN and cause a considerable consumption of it are not the sulfates \(\mathrm{CaSO_4}\) and \(\mathrm{MgSO_4}\), but the base that was combined with \(\mathrm{SO_3}\), i.e., ferrous oxide. The amount of \(\mathrm{SO_3}\) in the aqueous extracts of the ore is an indirect indicator of the corresponding amount of ferrous iron formed during oxidation of the ore as a result of grinding.

In the slimes, all the ferrous iron passed into ferric iron, neutral with respect to KCN, whereas in the ore ground dry in the laboratory all the iron was present in the form of ferrous sulfate, reacting with KCN according to the equation:

\[ \mathrm{FeSO_4 + 6KCN = K_4Fe(CN)_6 + K_2SO_4.} \]

To determine the amount of ferrous sulfate formed during grinding of the ore, cold extracts with \(2\%\) hydrochloric acid were made from samples taken from several degrees of grinding of the ore. In the hydrochloric-acid solutions, qualitative analysis showed only ferrous iron, \(\mathrm{SO_3}\), \(\mathrm{CaO}\), and \(\mathrm{MgO}\), and traces of ferric iron. Quantitative determination of iron in the hydrochloric-acid extracts gave the following results:

Ore	% Fe
I	1.31
II	1.66
IV	2.21
slimes V	1.16

As we see, the amount of soluble iron increases with decreasing grain size of the ore. In the hydrochloric-acid extract from the slimes, all the iron was in the ferric form. When the ore was ground in a bowl with an excess of water, pyrite was oxidized to \(\mathrm{FeSO_4}\); this salt, owing to the oxidizing action of the air in the presence of an excess of \(\mathrm{Ca(OH)_2}\), was completely converted into \(\mathrm{Fe(OH)_3}\), neutral with respect to KCN.

Similar investigations were carried out with pyrite I, selected in pure form.

Cold and hot aqueous extracts of ground to—150-mesh pyrite showed a neutral reaction, the presence of \(\mathrm{SO_3}\), \(\mathrm{CaO}\), and the complete absence of iron compounds. In the hydrochloric-acid extract there were only \(\mathrm{SO_3}\), \(\mathrm{CaO}\), and ferrous iron.

According to the analysis of the hot aqueous extract, the content of \(\mathrm{SO_3}\) in the ground...

in valuable pyrite proved to be 0.165%; the amount of ferrous oxide, according to the analysis of the hydrochloric-acid extract, was 1.34%. Such a low content of SO₃ in the pyrite accounts for the insignificant consumption of KCN, amounting to 4 lb per 100 poods of pyrite, and is explained, in the first cases, by the fact that the pyrite was crushed only to 150 openings, and, in the second, by the fact that the duration of grinding, carried out in a disk mill owing to the small quantity of material (1000 g), did not exceed 15–20 minutes, whereas grinding of the ore in a ball mill continued, in order to obtain various degrees of comminution, from 12 to 24 hours.

The quantity of pyrite oxidized is proportional to the duration of grinding.

During the grinding of the ore in a ball mill, the formation of a vacuum in it was observed, which indicates an intensive absorption of oxygen, owing to the oxidation of the pyrite, from the volume of air that was in the hermetically closed mill.

The investigations and observations carried out prove that the grinding of pyritic ores rich in iron sulfides is connected with considerable oxidation of the latter into ferrous sulfate. The latter is partly oxidized into ferric sulfate, or else reacts, when the ore is moistened with water, with bases such as CaCO₃ and MgCO₃ present in the ore. In treating the ground ore with a KCN solution, the iron sulfates give with KCN either complex compounds or ferric oxide hydrate. In the case of FeSO₄ we have:

\[ 6\ \mathrm{KCN}+\mathrm{FeSO}_4=\mathrm{K}_4\mathrm{Fe}(\mathrm{CN})_6+\mathrm{K}_2\mathrm{SO}_4, \]

in the case of Fe₂(SO₄)₃:

\[ \mathrm{Fe}_2(\mathrm{SO}_4)_3+6\ \mathrm{KCN}+3\mathrm{H}_2\mathrm{O}=2\mathrm{Fe}(\mathrm{OH})_3+3\mathrm{K}_2\mathrm{SO}_4+6\ \mathrm{HCN}. \]

All these reactions cause a very considerable loss of KCN.

The addition of lime reduces the consumption of KCN, since lime neutralizes free H₂SO₄ and converts Fe₂(SO₄)₃ into Fe(OH)₃ hydrate, which is indifferent toward KCN; but lime does not render FeSO₄ harmless, since FeSO₄ and Fe(OH)₂ react with KCN in the same manner.

Preliminary washing of the ground ore, associated with the removal of FeSO₄, may somewhat reduce the consumption of KCN, as is seen from the following experiment. Ores II and III were washed with warm water until the reaction for SO₃ disappeared in the wash waters. The washed ores were dried and agitated with a KCN solution for 24 hours, without alkali and with alkali. The consumption of KCN proved to be as follows:

KCN consumption, in pounds per 100 poods:

	before washing	after washing
Ore II without CaO	25.6	16.0
Ore II with 0.2% CaO	14.8	8.0
Ore III without CaO	33.6	28.8
Ore III with 0.2% CaO	26.8	12.8

The consumption of KCN even after washing the ore with water proved still very considerable, which is explained by the insufficient oxidation of FeSO₄ and by the absence of free alkali, necessary for the decomposition of iron sulfates.

Oxidation of iron sulfides takes place both during dry and wet crushing of the ore, and wet crushing by itself does not reduce the consumption of KCN. This may be concluded from the following experiment: 2 kg of ore I were crushed with 2 kg of water in a ball mill for 12 hours. Sieve analysis of the crushed ore gave 100% of the product—200. The crushed ore was filtered off, dried, but not washed. Samples of the ore were agitated with a KCN solution for 24 hours, with alkali and without alkali. The KCN consumption in this case proved very large, namely:

	KCN consumption
for ore I before crushing	4.8 lb per 100 poods
for ore I after wet crushing with alkali	30.8 lb per 100 poods
for ore I after wet crushing without alkali	34.4 lb per 100 poods

In the crushed ore, analysis of the hydrochloric-acid extract showed 2.66% iron in the ferrous state, which accounted for such a high consumption of KCN.

Thus, in order to reduce as much as possible the consumption of KCN in the treatment of this ore, in addition to wet crushing, intensive oxidation, or aeration of the pulp in the presence of alkali, is necessary in order to convert all iron salts into Fe(OH)₃—a compound neutral with respect to KCN*).

In view of all that has been set forth, the exact consumption of KCN for the treatment of this ore, as depending on the conditions and degree of crushing of the ore, can be established only during operation at the plant. Alternatively, the crushed laboratory sample must first be subjected to aeration in the presence of an excess of water and alkali; then the sample should be washed with water, and subsequent operations carried out with the dried material.

*) A comparison of the results of wet crushing in a laboratory ball mill and in a vat at the factory shows that, in the latter case, the product contains fewer KCN absorbents than in the former. This is understandable. In the ball mill, toward the end of crushing, a considerable vacuum was formed, i.e., almost all the oxygen was used up, and further oxidation of ferrous iron to ferric could not take place; at the factory, however, during crushing in an open vat in the presence of an excess of air, ferrous iron was oxidized to ferric, which was converted by lime into harmless ferric hydrate.

The results of the laboratory investigations show that, in plant treatment, the ore should not be ground in a KCN solution; rather, the latter should be added to the pulp only after sufficient aeration in the presence of alkali.

Conclusion.

The results of the laboratory investigation of the ore of the Khotimskaya vein, and consideration of local conditions, make it possible to draw the following conclusions regarding the most rational scheme for treating the ore on a plant scale.

Slime process. Two stages of crushing. After preliminary screening (2″), the ore enters a crusher, e.g., Blake or Gates, where the first stage of coarse crushing is carried out to 1—1½″. The ore is then fed to rolls, where it is reduced to ½—¼″ in the presence of water and lime.

Two stages of grinding. From beneath the rolls the ore enters a Hardinge conical mill. The mill product is fed to a Dorr hydraulic classifier, from which the slimes are led to a Dorr thickener, while the sands are reground in a tube mill or a second Hardinge mill for fine grinding. From there the material enters the Dorr classifier. The overflow product from the classifier is led to the thickener, and the underflow is returned to the tube mill. This creates a closed circuit, and the material entering the thickener should contain not less than 70% of product—200.

Slime plant. The slimes enter a Dorr thickener for preliminary removal of excess water, which is returned to the rolls. The thickened slime is fed to a collecting tank, where the required quantity of potassium cyanide is added in the form of a strong solution from the precipitation of gold. From the collecting tank the pulp passes successively into two or three Dorr agitators, where the gold is dissolved during agitation for 8–12 hours. From the agitators the pulp is directed to a second collecting tank, feeding mechanical filters or, for example, Merrill filter presses, where the gold-bearing solution is separated and the tailings are washed. The gold-bearing solution enters a filter tank and then proceeds to precipitation with zinc dust in Merrill presses. The barren solution is returned to the first collecting tank before the Dorr agitators. The precipitates are sent to the laboratory.

Thus, the treatment scheme includes the following operations:

1. Crushing in two stages:
   a. coarse, in a Blake- or Gates-type crusher,
   b. intermediate, in rolls.

2. Grinding in two stages:
   a. preliminary, in a Hardinge mill,
   b. final, in a tube mill or Hardinge mill.

1. Classification of the pulp in a closed cycle.

2. Thickening of the pulp.

3. Agitation

4. Filtration ”

5. Precipitation of the gold.

6. Treatment of the precipitates.

Table LV shows the sequence of operations; Table LVI—the scheme of operations, with an indication of the chain of apparatus.

The composition and character of the ore of the Khotimskaya vein, together with the observations made in the laboratory investigation, make it possible, in addition to the scheme given, to present certain considerations concerning the conditions and possible results of operation on a plant scale.

The predominance in the composition of the ore of such brittle minerals as pyrite and zinc blende, together amounting to 74%, and of calcite with magnesite, which account for 13%, ensures a high capacity of the crushing and grinding machines with a comparatively small expenditure of power. In view of this same circumstance, the second stage of grinding may be carried out not in a tube mill, but in a Hardinge conical mill, operating on pebbles. The advantages of the Hardinge mill over the tube mill are lower cost, smaller bulk, and lower power consumption, with a sufficiently high capacity.

Since the best results in fine grinding are obtained with the lowest possible water content in the pulp (40–50%), the additional quantity of water required for classification will have to be added to the product as it leaves the mill. Grinding should be carried out in water, and not in a solution of potassium cyanide, since in this way an excessive consumption of KCN will be avoided, as the laboratory experiments showed. The addition of KCN should be carried out in the collector after the thickening of the slimes.

The ease with which the ore gives up gold to a KCN solution makes a prolonged contact of the ore with the solution superfluous. Agitation in the collector and in Dorr agitators will ensure the maximum extraction.

The thickening of the ore is likely to present little difficulty, since its high specific gravity will bring about a fairly rapid settling of the slimes and a low moisture content in the thickened product.

The same high specific gravity of the ore excludes the possibility of using frame vacuum filters for separating the solution from the slime. Rapid stratification of the pulp, uneven deposition of the cakes, and poor washing—such are the inevitable consequences of using frame vacuum filters for this ore. More rational are mechanical filters of the Oliver type or filter presses.

Merill’s with automatic discharge of tailings. The precipitation of gold will present no difficulties, owing to the purity of the solutions. It is preferable to use zinc dust for precipitation, since the consumption of zinc is thereby less than with zinc shavings and, moreover, the gold is realized more quickly, which is especially important given the comparatively small scale of production and the considerable expenditures for equipping the mine and plant.

Conclusions.

1. The ore of the Khotim vein is gold-bearing pyrite, the amount of which in the mass of the ore is 70%, with 26% gangue consisting of quartz, calcite, and magnesite. The gold is associated with the pyrite and other sulfides present in the ore in very small quantities.

Average content in 100 poods of ore:

Component	Content
gold	8.76 zol.
silver	14.21 "

In the accompanying rock, gold is absent.

1. The most rational process for extracting gold from the ore, from the technical and economic points of view, is a complete slime process with the use of mechanical agitation of the slime pulp.

2. According to laboratory investigations, the percentage extraction of gold will in practice amount to not less than 91%. The extraction of silver, which is of only secondary importance, may reach 50–60%.

3. The ease with which the ore gives up gold to a solution of potassium cyanide excludes the possibility of encountering any particular difficulties in carrying out the process in practice. The consumption of potassium cyanide for the treatment of the ore can be established only in practice when operating under plant conditions, but according to laboratory data the consumption of potassium cyanide will not exceed 3–5 pounds per 100 poods of ore.

4. On the basis of the data from the laboratory investigation of the ore, a flowsheet has been proposed for the process that is most expedient under the conditions of the Berikul mine.

The present article is dedicated to the memory of the late professor of the Tomsk Technological Institute, Lev Lvovich Tov. Lev Lvovich was an outstanding authority on the theory and practice of gold metallurgy in general, and in Russia in particular; and it is natural that his attention was attracted by investigations of gold-bearing ores conducted in Metallurgical Laboratory I of our Institute. By the possibility of carrying out—

We owe the implementation of these studies to a considerable degree to the assistance of Lev Lvovich, who helped us both with his extensive knowledge and with moral support. Thanks to Lev Lvovich’s personal intercession, the laboratory received the necessary funds for its initial equipment.

The creation at the Tomsk Technological Institute of an exemplary laboratory for the gold industry was always Lev Lvovich’s dream; and if this dream is destined to be realized, then a worthy perpetuation of the memory of the deceased would be to give this laboratory the name of Professor Lev Lvovich Tove.

January 1918.

Tomsk.

For the article by V. Ya. Mostovin and V. A. Pazukhin

Fig. 1. Ore pan.
Scale: 1/10.
Visible dimensions/marks: Ø70; 20.

Fig. 2. Ball mill.
Scale: 1/20.

Fig. 3. Laboratory sluice.
Scale: 1/10.
Visible labels: A, B.
Visible dimensions: 110; 36.

Fig. 4. Amalgamation machine.
Scale: 1/10.
Visible dimensions/marks include: 95; 90; 65; 55; 36; Ø20; 69.

Fig. 5. Fuhrmann apparatus for determining specific gravity.
Scale: 1/10.
Visible labels: A, B, C.

Fig. 6. Apparatus for determining volumetric weight.
Scale: 1/6.

Fig. 7. Apparatus for determining the rate of cyanidation, according to [[unclear: handwritten names, possibly “Ostwald” and “Siemens”]]. Scale \(1/8\).

A — glass tube
B — narrow glass tube
C — [[unclear: handwritten label, “Woulff bottle”/flask]]
D — funnel for [[unclear: liquid/reagent]]
E — scale for readings
G — hose for mercury
H — glass wool

Detail shown at full size. Scale \(1/1\).

Fig. 8. Laboratory percolator. Scale \(1/5\).

Detail of the bottom. Scale \(1/2\).

Labels in the drawing: ore; [[unclear: layer of glass wool / openings in the glass]]; tap; outlet tube; ventilator.

Fig. 10. Gold-bearing [[unclear: dump; caption handwritten]].

Labels in the figure: layers I, II, III, IV, V, VI, VII; layer 3.7; layer 4.4; layer II 4.2; layer IV 4.4; layer V 3.36; layer VI 5.7; layer VII 6.1; layer VIII 6.3; layer 2.4; layer II 2.6; layer III 2.20; layer IV 2.40; layer V 2.3; layer VI 2.3; layer VII 2.01; layer VIII 1.6.

Cross-section label: groove along line C1 B1.

Dimensions and labels in the cross-section: layer 4.4; 1.4; 4.50; 40 degrees; 4.50.

Fig. 9. Apparatus for elutriating slimes [[unclear: in/with settling funnels]], scale 1/5.

Label in the figure: water.

Dimensions shown: 135; 35; 300.

Diagram I. Effect of treatment duration on extraction.

Vertical axis: %

Horizontal axis: Hours of treatment

Curve labels visible on the graph:

• Effect of [[unclear: cyanide]]
• Effect of bromocyanide
• Effect of Cl? [[unclear]]

Diagram II. Preliminary treatment.

Vertical axis: %

Horizontal axis: [[unclear: hours]]

Curve labels visible on the graph:

• Effect of bromocyanide
• Effect of [[unclear]]
• Effect of [[unclear]]

Diagram III. Sieve analysis.

Vertical axis: %

Horizontal axis: classes: +80, +100, +120, +150, +200, −200

Legend:

• solid line — washed with water
• dashed line — [[unclear]]
• dash-dot line — [[unclear]]

Diagram IV. Distribution of gold by classes in %%.

Vertical axis: %

Horizontal axis: classes: +80, +100, +120, +150, +200, −200

Curve labels visible on the graph:

• [[unclear: fine gold]]
• [[unclear: coarse gold]]

Diagram V. Dissolution of gold by classes, in gold per 100 mgr.

Vertical axis: milligrams

Horizontal axis: classes: +80, +100, +120, +150, +200, −200

Diagram VI. Consumption of KCN solution as a function of the degree of grinding of the ore.

Vertical axis: lb KCN per 100 tons of ore

Horizontal axis: 100% −200