from the sewage itself, from electrodes, or from salt water. The chemicals thus appear in the sewage in their nascent state, in which condition they are considered to be most active.

Sewage has, by the "Webster" process, been decomposed by causing it to flow between electrodes placed an inch or so apart in a trough, after which it was allowed to settle for an hour or two. By this method chlorine and oxygen were carried to the positive electrodes as a hypochlorite, at the rate of 2 grains per gallon treated. It was estimated that it required 0.25 ampere-hours of current for each gallon treated. There was effected a 95.3% reduction of the sus pended matter.

Another process uses electricity to decompose sea-water, or a solution of magnesium and sodium chlorides. This is an antiseptic, not a purifying process, sodium hypochlorite or some oxygenated compound of chlorine being produced. This method has been used in this country, under the name of the Woolf process, at Brewsters, N. Y., and at Danbury, Conn.; but in 1895 the latter place was enjoined from discharging the effluent from this treatment into the Still River and is now filtering its sewage. At Brewsters 1000 gallons of water containing 160 pounds of salt was subjected to an electric current of about 700 amperes and 5 volts, the positive electrode being of copper plated with platinum, and the negative of carbon; a 4-H.P. dynamo being used. One part

of this solution was used in 100 parts of sewage; or $3.20 of salt to each 1,000,000 gallons. Practically the same process was used in Bombay in 1897, but abandoned after four months' trial, it being found that the same amount of free chlorine could be obtained in chloride of lime at half the cost. Up to the present time, at least, it has been found that any desired chemicals can be purchased more cheaply than they can be manufactured from sewage, whether they be precipitants or fertilizing precipitates.

Considering the practical application of the above ideas, we see that we must prepare the chemicals, introduce them into the sewage, permit the latter to deposit the insoluble matter, draw off the effluent, and dispose of this and of the deposit (called "sludge ").

The chemicals are ordinarily obtained as crystals or in powdered form. As such they would not readily or quickly mix with the sewage, and they are usually dissolved, better in sewage than in water, to form a more or less saturated solution, in which form they are introduced into the sewage. In Glasgow the lime-mixer consists of a cast-iron box, through which passes a vertical shaft driven by belting, to the shaft being attached four horizontal radial bars at different elevations and of different lengths. Pieces of chain are used as agitators which drag along the bottom to prevent deposit. A horizontal grating with 7 X 14-inch spaces fills the interior at 2 feet 8 inches from the top, through which grating the lime must percolate. The depth of water in the mixers is usually 3 feet 3 inches. The alum is mixed in four wooden. vats 3 X5 X 10 feet, the agitation being effected by exhaust air from the sludge-lifts which is led into the bottoms of the

vats.

In East Orange the mixers were in the form of cylindrical cast-iron vats 4 feet in diameter, with conical bottoms, each overlaid with a perforated plate. The chemicals were placed on the plates and air blown in from the bottom as in the Glasgow plant.

At Worcester the mixing-tanks are 8 X 16 X 8 feet deep, of iron in brick masonry. Two and one-half tons of lime can be mixed at a time in each. Compressed air is used here also as an agitator.

The concentrated solution thus prepared is admitted to

the sewage and should be thoroughly mixed with it. This should be done before the sewage is pumped, if pumping is necessary; both because this assists in the mixing, and because less suspended matter in the sewage has been taken into solution, in which form but little of it can be removed by chemicals. To obtain thorough mixing with the sewage it is better to maintain a continuous flow of precipitant than to introduce a certain amount at intervals of one to fifteen minutes; although the latter is generally the simpler plan. The amount of chemical introduced per minute should be proportioned to the amount of sewage flowing and to its chemical composition. For this purpose analyses should be taken about once an hour; and the flow at any moment should be ascertainable by observing a weir inserted in the sewage channel, or otherwise. A gate or cock can be provided with an index or gauge by which the amount of chemical required from time to time can be caused to flow into the sewage. In very small plants, however, it may be found. cheaper to introduce the chemical at such a fixed rate during the day, and such another during the night, as has been found to produce the desired purification with the highest rate of flow of the strongest sewage; thus avoiding the expense of keeping a chemist constantly on the work. To effect the mixing of the chemical and the sewage, the former is generally introduced while the latter is flowing along an open channel, which is provided lower down with baffle-boards forming a "salmon-ladder," or with a small under-shot wheel.

From this channel-after being pumped, if this is necessary the sewage flows to tanks in which the insoluble matter precipitates, forming sludge. There are three general styles. of tanks: the continuous-flow and the intermittent-flow horizontal tanks and the upright tank. In the intermittent-flow one tank is filled and the sewage is then allowed to stand at rest for from half an hour to three hours, another tank being

meantime filled. Three tanks at least are here necessary; several small ones being better than a few large ones, as allowing longer rest for sedimentation with equal storage capacity. Since each tank must be emptied of the clarified sewage, they must either be pumped out, or the stream or land to which the effluent is led must be several feet lower than the sewer-outlet. The Glasgow tanks require 7 minutes to fill, 45 minutes of rest, and 7 minutes to empty.

In the continuous-flow tank there is no absolute rest, but the sewage is continually moving at a rate of .02 to .006 of a foot per second through the tank from inlet to outlet. Two tanks only are necessary; and, the effluent leaving but slightly lower than the sewage enters, pumping of sewage, which might be necessary with intermittent tanks, is avoided. The cross-section of the continuous tank can be calculated by dividing the maximum flow per second in cubic feet by the required velocity-.02 to .01. The length will depend upon the time the precipitant will require for settling. From 2 to 8 hours is the more general practice; 2 hours should generally be sufficient if based on the maximum flow, for the ordinary flow will then have about 4 hours. Upon this basis, if the velocity is .02 of a foot per second, a tank 144 feet long will be required. The Worcester tanks are 1663 feet long.

In some plants the continuous-flow tanks have cross-walls over which the sewage is required to flow, but this is by some considered harmful rather than beneficial; except that there should be such a wall at or near the upper end to reduce agitation in the tanks by the entering sewage.

Tanks must have some arrangement for removing the sludge. In intermittent-flow tanks the effluent is generally drawn off by a hinged pipe, its free end being maintained, by a float, about 3 to 6 inches below the surface. When the effluent begins to run cloudy the remaining contents of the tank (or the sludge) is drawn off into a sludge-well or pit, to

facilitate which the tank bottom slopes to a middle channel, which itself slopes toward a sludge-gate. In the continuousflow tanks the sludge may be lifted by a pump whose movable suction passes just above the bottom; or in some cases it may be drawn off through an opening in the bottom of the tank; the sludge being preferably forced to this opening (which is at one end of the tank) by a squeegee' reaching across the tank and travelling its full length pushing the sludge before it along the bottom, and scraping from the sides the colonies of bacteria which are likely to grow there. In some cases the supernatant sewage is pumped out before the sludge is removed.

Intermittent-flow tanks have been constructed in this country at the Mystic Valley works, and continuous-flow at East Orange and Long Branch, N. J., Worcester, Mass., New Rochelle, N. Y., and Canton, Ohio.

Tanks should have smooth walls and should be watertight. There is little danger from frost, except in the most northern States, as the sewage retains considerable heat and the tanks are generally open to the air. The East Orange ones were roofed over as a concession to the prejudices of the citizens living quite near the works. On account of this popular prejudice, as well as to reduce the cost of the considerable area occupied by horizontal tanks, they will generally be placed as far as possible from built-up sections. Where this cannot be done the area required can be reduced by use of a vertical tank.

In the vertical tank the sewage flows upward and the precipitant collects on the bottom, which is, in the "Dortmund" tank, conical in shape. Fig. 37 shows the Chicago vertical-flow tanks, modelled after those at Dortmund. If the sludge-pipe is made to discharge 1 to 2 feet below the level of the sewage in the tank, this head will be sufficient to force it out without pumping, providing it contains 90% to