the solid matters become macerated and finely divided by fermentation, before entering the distributing pipes. Moreover, if such a tank be allowed to overflow constantly into the porous pipes by a dribbling discharge, they would become choked after a while even then. In order to keep them free, the flow must be intermittent and take place with such a rush as to fill the whole system of distributing pipes at once, and brush away slight obstructions which may have been left by former discharges.

There are two methods of obtaining this result: First, by providing a stop-gate in the outlet-pipe where it leaves the tank, to be opened by hand when the tank is full and closed again when empty. Second, by providing a siphon or float to discharge the tank automatically whenever filled.

Of course the latter method is far more satisfactory, if made reliable, but it is somewhat more costly, and the siphon, as heretofore constructed, is not always reliable, i.e., the apparatus for setting the siphon in action does not always work as expected.

Field's siphon has generally been used for this purpose, and works well with pure water; but if used with sewage on a small scale, it is liable to miss, unless the cup into which it discharges is occasionally brushed out. When used on a large scale, as for a combination of several houses, this difficulty disappears.

Several other devices have been used for starting a siphon. In one a tumbler tank is placed in the upper part of the flush tank, which upsets when filled, turning on brass trunnions, and righting itself at once when empty. A tumbler tank of two gallons capacity will start a siphon of two inches diameter, if the lower or discharging end dips in water as soon as the flow through it begins. It is important that it should not so dip when no water is flowing through it, for in that case the flush tank would never be filled again after once discharging; but it can be so arranged that a very slight flow would seal the discharging end.

The device patented by Mr. Field was to accomplish this result.

The tumbler tank accomplishes the result well enough when filled with pure water, but if used for sewage, the solid matters are apt to adhere more or less to its sides and thereby destroy its poise, on which its certainty of action depends.

Other devices have sought to accomplish the same end by a float actuating a valve in the bottom of the flush tank, or by a bucket in a side chamber which can be filled by the overflow of the tank, and thereby becomes heavy enough to open the valve, while a small leak in the bottom allows the valve to raise it again after the tank is empty.

All of these contrivances are subject to failure from wear of moving parts except Field's siphon, which it is hoped may be perfected still further. Fig. 110 shows the two tanks and the Field siphon as modified by the writer, the changes being at least a partial remedy for the imperfections complained of.

If the siphon were to be applied as the outlet of the same tank which receives the sewage for maceration, etc., it would often become choked by solid matter. It is therefore best to provide a second tank between this and the siphon. The first is called the "settling basin and the latter the "flush tank" on Fig. 110, for the last is alternately filled and emptied, while the first remains always full.

Siphons are often constructed inside of the flush tank, but it is better to make them accessible at all times by placing them outside, as here shown.

A two-inch siphon is as large as can be surely set in action by the drainage of a single house.

The surest method of starting such a siphon would undoubtedly be to provide a copper float in the tank with an ordinary brass cock, such as is

B

A

used on water-supply pipes
for the automatic filling of
house tanks; but the float
must be adjusted so that
the cock should be opened
instead of closed, when the
sewage rises to the over-
flow line of the siphon.
Supply this cock with pure
water under pressure from
some outside source by a
small pipe, and let the dis-
charge be delivered by a
tight connection through
the arch of the siphon, in
such a way that it would
fall free of the sides of the
tube into the cup or basin
where the siphon is made

FIG. 111.-A, Section across Branch. B, to discharge. A very small stream of water, thus ap

Section lengthwise of Branch.

plied, would start the siphon with certainty, and with scarce a moment's delay, if its discharging end is so placed as to be sealed with water by a slight flow.

The outlet pipe for this apparatus should be four or five inch stoneware pipe for a two-inch siphon. If over one hundred feet in length, and on an inclination less than two in one hundred, the larger size may be preferable.

The distributing pipes should be cylindrical and two inches bore. The quantity needed will vary with the porosity of the soil and the size of the flush tank. The tank here shown contains 37 cubic feet when filled. The pipe should be of sufficient length to contain about one-half this amount, say 19 cubic feet. Since the sectional area of a two-inch pipe is 3.14 square inches, it will require about 46 linear feet of the pipe to contain one cubic foot. They should then be about 46 times 19, or 874 feet in length. If the soil is extremely porous, a smaller quantity would doubtless answer the purpose, since a larger proportion of the water would soak away while the siphon is discharging the tank.

FIG. 112.-Per

glazed 2-inch Tiles

These distributing pipes should be laid with a perfectly uniform slope of not over six or eight inches in one hundred feet. Even less than this will often answer. If the slope exceeds this amount, the water may burst up at the lower spective View of Unend and make a nuisance. They should not be covered over with Troughs and eight or ten inches below the surface of the ground. In order to combine these two conditions, the ground must be somewhat smoothly graded, and the lines for the pipes must be laid out to conform to the contour of the surface, i.e., the trenching must follow lines which have a surface slope limited as above stated. The trenches may be at intervals of

Covers.

five or six feet, dividing the field in a gridiron fashion. The tiles must be laid one-fourth of an inch open at joints.

The branches where the two-inch pipes leave the main should be so made as to allow no fluid to be retained in the main, but lead from its bottom, as shown in Fig. 111.

The branches can be made rights and lefts to suit the places where used, and can be either of the Y or T form. Ordinary pipe-branches are intended for combining two streams into one, but these are for dividing one stream into two. They should, therefore, be formed as shown in the drawings, having no socket on the small hole, which is to abut against the first piece of two-inch tiles. These porous tiles have sometimes been laid upon boards bedded in the trench, to secure a more uniform gradient and prevent dislocation of joints; but the decay of the boards soon allows the pipe to become displaced by settlement, and it is better to place terra cotta troughs under the pipe, breaking joints therewith. Similar pieces, about four inches long, are placed over the joints as covers (see Fig. 112). It is often objected to this system that the pipes would become filled with ice in Northern winters unless buried deeper than eight inches. But experience has shown that such is not the fact near Boston. If buried

FIG. 113.-Disposal of House Drainage by Surface Irrigation.

deeper, the roots of grass and the air do not get so good access to the sewage, which is therefore likely to accumulate-a result which we wish by all means to avoid. When laid near the surface no such accumulation occurs, and the frequent flow of warm water from the house prevents the ground from freezing under and around the pipes, if covered with sod.

In places where a sufficient slope exists, the sewage of single houses may be distributed on the surface with advantage. The above cut shows such a method, which has been in use for several summers at a house used only as a summer resort, and has been attended with complete success. There is an overflow pipe from the cesspool, which is perfectly tight, indicating when it is full. The gate in the outlet-pipe is then opened and the whole contents distributed on the kitchen garden in ten minutes. If an opportunity is selected when the wind is blowing from the house to the garden, no offensive odors are perceived, and the growing crops soon absorb the fluid, much to their advantage (see Fig. 113).

DISCONNECTING HOUSE DRAINS FROM SEWERS.

The complete separation of the air-space in our house drains from that in the public sewers is now insisted on by most European municipal authorities where regulations are enforced on such subjects. This separation is even more important in cases where a cesspool is used in place of a sewer, for the gases of the cesspool are much the more foul, and no

amount of ventilation will prevent their becoming so. The cesspool is but a retort where such gases are constantly evolved from the decomposition of the fluids and solids retained in them.

The devices for attaining this separation which are applied in England are not so well adapted to the climate of New England. Here the severity of the winters requires the separating trap to be placed several feet under the ground, and as it should be kept accessible, a man-hole or well is gen

erally constructed over it in brick-work, with a perforated iron cover at the top, as shown in the annexed cut (Fig. 114). If this is used in a snowy climate, a bent pipe should be extended from the interior of the chamber to a point a few feet above the surface, where it may be protected by a wire-basket. Little or no offence is likely to arise from such an opening.

This subject has attracted much attention of late, and deserves careful consideration. The liability of the interior surfaces of house drains to become foul by the accumulation of solid matter, or from the formation of a slimy coating of organic matter on their interior walls, is well known, and the high temperature of the drains as compared with that existing in the sewers conduces to a more rapid decomposition of such matter than in the sewers themselves. A constant change or current of air through every part of the drain, is therefore essential to avoid the concentration of such gases. If such concentration occur, the risk of harm is multiplied, by the well-known law of diffusion of gases, by which they penetrate very small cracks, and through imperfection of workmanship or inefficiency of the water seal in the traps, might enter and mix with the air of the house. When houses are artificially heated this tendency is increased by the diminished density of the air within the house, causing a slight inward pressure through all such fissures.

Our best municipal codes now require the house drains to extend up through the roof over every separate "stack" or vertical line of soil pipe, of a size not less than four inches. This size is found necessary in order to avoid accumulation of frost by condensation at the top, and other rea

sons.

In order to provide a draft through these pipes, the fresh air should be freely admitted to the drain at its lower end, next to the trap by which it is separated from the outside drain and sewer.

Such an opening as referred to in connection with Fig. 114, can be made

to terminate anywhere above ground, and within a few feet of the surface, without risk of offence, for the draft is almost always inward. It is customary to terminate them by a pipe extending some three feet above ground against the yard fence, with some protection at the orifice, to prevent stones, etc., from being thrown in by children.

The need of a draft through every part of the house drain is not the only reason for their being open to the air at the top. It is impossible to retain the water seal in the best forms of traps-the most simple ones-un

FIG. 115.-Vented S-Traps.

VENT

less the atmospheric pressure is freely admitted just below the water seal. The partial vacuum that follows a charge of water as it descends a vertical line of pipe is such that, without a free admission of air below every trap, the pressure on the house side of the trap would force the water through them and leave them unsealed.

It therefore becomes necessary to apply branch vents to each separate tube (see Fig. 115). This is particularly needed in connection with waste pipes smaller than four inches in diameter.

In order to ascertain the amount of risk incurred by lack of air in the outlets of traps, the writer was employed by the National Board of Health, in the summer of 1882, in connection with E. W. Bowditch, Esq., to make a series of experiments with apparatus constructed for the purpose. The following is an abstract of the report upon these experiments, published in August, 1882, in columns of the New York Sanitary Engineer:

A typical stack of four-inch, and another of two-inch waste pipes were erected in a vertical position, extending fifty-seven and a half feet above the basement floor, with branches in the basement and on the floors above, and having the upper ends open to the air eight and one-fourth feet above the upper floor.

"The vertical pipes extended down to a point two feet above the basement floor, from which they turned to a slope, falling this two feet in a horizontal distance of thirty-two feet.

At

The two-inch waste joined the four-inch pipe at a branch on this slope five and a half feet from the lower end of the vertical portion. A fourinch running trap was attached to the lower end of the slope, having a four-inch air-hole close above it, to represent the foot ventilation. a point on the slope six feet below the lower end of the vertical reach, a branch was inserted with a four-inch trap in it, marked (n) on the plate, to represent an inlet for a basement water-closet, or for any other desired fixture at that level. The relative position of these branches and flows is shown on the accompanying plate (Fig. 116).

"Traps may lose their water seal by siphon action, i.e., by lack of airpressure, either when water is poured through them, or when poured through the main waste pipe into which they discharge from fixtures at a higher level. These two processes are, therefore, examined separately, as described below.

"Since it is now generally admitted that the upper ends of all soil or waste pipes should be wide-open to the air, all experiments were tried under this condition, except where otherwise noted for special purposes there explained.

"We will first consider the loss of the water seal in traps caused by water flowing through them from the fixtures attached thereto.