form of fluid pulp to the revolving distributor launders and deposited on top of the charge, again sinking down to the bottom to be subjected to the same action. In this way a vertical circulation of the ore particles through the comparatively stationary solution takes place giving (at least theoretically) a more rapid individual motion relatively to the solution than occurs in systems where the whole pulp is in agitation simultaneously. While this is going on a continuous stream of pulp is flowing out at a point on the top of the tank opposite the inflow and maintaining a constant level of pulp in the tank.

Another feature claimed by the inventor is a "selective agitation" of the more sandy part of the pulp, the theory being that the larger and heavier particles, those therefore that require longer time for treatment, sink more rapidly and circulate more often than the lighter particles and also remain longer in the tank. The latter point is a little puzzling when it is remembered that just as much coarse material must be found in the outflow as there is in the inflow in order to prevent a gradual but continuous increase of sand in the tank, but this is not incompatible with there being continuously present in the tank a larger percentage of sand than is to be found either in the inflow or outflow. It is difficult to say how much practical importance is to be attached to this phenomenon but that it has an actual existence is claimed to have been sufficiently proved. According to Dorr the mechanical power taken varies from 14 to 2 or 3 hp. according to the size of the tank, the speed, and the nature and dilution of the pulp handled. The air required is from 8 to 40 cu. ft. per minute at 10 to 20 lb. pressure.

Among other devices for agitation which have attained to some prominence are the Trent agitator, the Parral tank, the Hendryx agitator, and the Devereux agitator. The first of these consists of a submerged mechanism somewhat on the principle of the Butters distributor. A watersealed chamber on ball bearings is placed on the bottom of the tank at its centre and several pipes of different lengths radiate from it, whose ends are turned at a right angle. A centrifugal pump is placed outside the tank

with its suction connecting with the pulp near the surface and its discharge entering the central chamber and finding its way out through the radial pipes, the tangential force causing the whole to revolve slowly. One of the chief objections lay in the inaccessibility of the mechanism in the event of stoppages or blocking of pipes, but this has been largely overcome by J. A. Carpenter who has devised a method of elevating the whole of the moving parts for repairs by means of a tackle and replacing them without emptying the tank.

The Parral tank, devised by Bernard MacDonald and named after the town in Mexico where it was first used, is the result of a desire to apply the good points of the Pachuca to a tank of more normal proportions and less head room consumption. In any tank of ordinary dimensions are placed 4 air lifts spaced at equal distances around the periphery and discharging tangentially through a tee or ell, so that the flow will be in the same direction. This action of the discharges imparts a rotary motion to the pulp which it is claimed is sufficient to hold in suspension an admixture of a considerable amount of sandy material. The centre being a dead point there is, of course, a tendency to build up a cone of solid material there and in more recent installations the tank is fitted with an artificial cone to take the place of this accumulation of solids.

The Hendryx agitator is somewhat like the Pachuca except that instead of using an air lift, the ascending motion is imparted to the central column of pulp by means of mechanically driven screw propellers.

The Devereux Agitator is operated simply by a comparatively small propeller traveling at a high speed. The tank used is of the usual type, but with a slightly conical bottom. The following dimensions are given in an article by Roscoe Wheeler in the Engineering and Mining Journal of July 15th, 1916; a 34-foot diameter tank with 5-ft. propeller revolving at 80 to 100 r.p.m. handles from 80 to 165 tons of dry slime at a charge with a dilution of 2 to 1. The maximum horsepower under these conditions is 12.

A 30-ft. diameter tank with 4-ft. propeller revolving at 80 to 100 r.p.m. at a maximum horsepower of 10 will handle 138 tons of dry slime at a dilution of 1.7 to 1.

It is claimed that there is no dead point in the tank and therefore no deposit of slime in any part of it and that there is no difficulty in getting the whole of the charge mixed and in circula

tion in a short time even after a prolonged shut down. An important feature is said to be the baffle which is used to break the vortex which would otherwise retard the velocity of the pulp and decrease the scouring action. Sufficient air is said to be drawn into the charge by the motion of the pulp to supply all the oxygen needed for chemical purposes, the amount being regulated by adjustments of the baffle. A subsidiary baffle for use

during charging is provided, the same operating on hinges and being drawn up out of the pulp as soon as the tank is full.

Cyanide Strength.—The best conditions under which to conduct the treatment can only be determined for each ore by careful experiment both in the laboratory and later in the working plant. Cyanide strength varies greatly; on the Rand the most usual strength is from 0.005% to 0.02% KCN or 0.1 lb. to 0.4 lb. per ton of solution. All-gold ores can usually be treated with much weaker solutions than silver-gold and silver ores. For the latter the writer has found that the maximum extraction can rarely be obtained with a treatment solution of less than 0.1% KCN or 2 lb. per ton of solution, while if time of treatment is unduly limited, even higher strengths are advisable. In Pachuca strengths of solution up to 0.4% KCN and over are frequently used.

Dilution of Pulp.-On this point also wide variations occur in practice. Theoretically the thinner or the more dilute the pulp is the more rapid will be the action on the precious metals. Experiments by Julian and Smart' showed that taking the rate of dissolution of gold in cyanide solution alone as 100 a pulp density of 6 of solution to 1 of slime gave a rate of only 58.6, and so on in gradually diminishing amounts till a 1:1 dilution showed only a rate of 20. They account for this by the retarding influence of viscosity on the movement of the K and CN ions. In practice, however, the increased cost of handling and precipitating large volumes of solution has to be offset against the increased extraction and an economic basis arrived at.

In any case the theory above stated does not always work out in practice, and at present the ratio of solution to ore in greatest favor lies between 114 and 2 to 1.

It may be noted that the pulp from an all-agitation mill, that is, a pulp with a high percentage of sandy material will call for a lower ratio of solution to ore than a slime pulp from which the granular material has been separated for leaching. A separated slime may need a ratio of 2 or 3 to 1 to give the best extraction, 1 Cyaniding Gold and Silver Ores, page 224 (Second Edition),

while a general mill pulp may give as good an extraction with less mechanical troubles if treated in a 1 to 1 pulp.

The usual practice where Pachuca tanks are in use is to make a dilution for the agitation treatment that does not require any subsequent thickening prior to the process in use for displacing the value-bearing solution, and such dilution is usually placed between 1 and 1.2 to 1.

Time of Agitation. This may vary from a few hours in the case of free milling gold ores to several days for silver sulphide ores, and has to be determined by experiment. As already stated the time necessary may often be shortened by increasing the strength of the solution in cyanide.

Increasing the Dissolving Effect by a Succession of Washes.—It seems to be a very widely observed fact that an increase in extraction is often obtained by settling and decanting several times during treatment and substituting fresh barren solution each time as against conducting the whole period of agitation in the same solution. This is a point the writer has never been able to verify, personally. Even in some plants where the phenomenon has been reported as proved by experiment, he has been able to obtain the same extraction without changing the solution during agitation. In the majority of instances where this fact is observed the writer is inclined to believe that if an increased cyanide strength were used for the agitation a result could be obtained similar to that shown by a succession of washes. Even if this be so, it may happen in any particular case that a succession of weak washes may turn out to be more economical than a smaller amount of stronger solution, and this is a point to be decided by the metallurgist in charge.

Continuous Treatment. The system of continuous treatment consists in allowing the stream of pulp to flow continuously through a connecting series of agitator tanks, as distinguished from the intermittent system where a charge is placed in a tank and agitated therein until the dissolution of the precious metals is complete. It may be applied to almost any form of agitator, whether the old stirring-geared kind, the Pachuca, or the Dorr.