(e) A 1" vent-pipe 7' long was then attached to the opening. The first discharge nearly broke the seal; the second not only broke it but left the water standing " below the mouth of the inlet-pipe. Three other tests gave the same results.

(/) The vent-pipe was then lengthened to 17'. A single discharge broke the seal, and swept nearly all the water out of the trap. Three repetitions of the test produced the same results.

The vent-opening was then reduced to one inch, and no vent-pipe attached. Two discharges broke the seal.

(g) The 1" common cast-lead trap was then tested.

Tested with the 1" vent open at the crown, but without any vent-pipe attached to it, a single discharge lowered the seal ". A second discharge produced no further effect.

(h) With a 1" vent-pipe 7' long attached to this opening seven discharges sufficed to break the seal.

(i) With the vent-pipe increased to 17', two discharges not only broke the seal but nearly emptied the trap. Substantially the same result was obtained upon repeating the experiment twice.

(j) A 1" pipe 20' long was attached to the vent opening. Two discharges broke the seal.

The vent-opening was then reduced to 7/8" +, and no vent-pipe attached. Two discharges broke the seal.

(k) The middle vent was then tested. With a vent-pipe 1" in diameter (inside measure in all cases being understood) and only 8" long, the water-seal was lowered " by two discharges. A third discharge did not increase this loss.

(/) With a 1" pipe 7' long attached a single discharge broke the seal.

(m) A 1" pipe 20' long was attached; a single discharge broke the seal.

(n) The lower vent was then tested. A 1" pipe 20' long was attached. Two discharges broke the seal. A repetition of the experiments produced substantially the same result.

It must here be noted that the connections between the traps to be tested and the branches on the soil-pipe were made by means of 1" copper tubes soldered into a lead cap on 4" Y-branches, as shown in Figure 13. The outlet arms of the several traps tested were slid over this 1" copper tube, and made tight with putty. The connection being only 1", when 1* traps were tested, the suction on their seal W2S evidently somewhat restricted by the contraction. Hence the records of the experiments on traps having 1" outlets may safely be accepted as well within the limits of actual power of the siphonage produced at this point on the 4" Y-branches.

The above recorded tests were all made on the second floor at a distance of 11' below the water-closet trap. Backpressure was here hardly perceptible. Tested on the first floor the siphonic action was much weaker, but a slight back-pressure was observable. On the basement floor the siphonic action was quite feeble, but back-pressure was exceedingly strong.

Deductions from the Experiments with the Trapped

Plunger-Closet.

From the experiments thus far recorded we learn that the siphonic action which may be produced by a trapped plunger water-closet under certain simple conditions which are likely to be encountered in plumbing, is sufficient to unseal small S-traps, such as are ordinarily used for lavatories, though they be ventilated either at or below the crown in the manner prescribed by the plumbing regulations with vent-pipes of the full size of the trap, and that it makes no material difference as to siphonage whether the vent-pipe be applied immediately at the crown or at a considerable distance below it. This action takes place even when the pipes are clean and new. When partially closed or clogged with sediment the results would be even more serious.

(2) Experiments on the Siphonage of Ventilated S-Traps by a Trapless Plunger Water-Closet. (a) The tests were made first with the water-closet alone. A 1" ordinary cast-lead S-trap having a vent-opening at the crown the full size of the trap (1") was tested without a vent-pipe. The first discharge of the closet reduced the seal "; the second "; the third produced no further effect.

The same trap was then tested with 17' of 1" pipe attached to the vent-opening. Four discharges of the closet destroyed the seal.

On another occasion the same trap with 7' of 1" pipe attached to the vent-opening lost its seal in four discharges.

A 1" S-trap was then tested at the outlet on the first floor. With the water-closet alone, and a 1" vent-pipe 20' long attached to a 1" opening in the crown of the trap three discharges removed 5/8" of the seal.

The same with a 1" vent-pipe 7' long lost its seal in ten discharges.

(b) The tests were next made with the water-closet and bath-tub discharging together.

A 1" S-trap with 20' of 1" vent-pipe lost its seal in five discharges. With 17' of 1" pipe a single discharge broke the seal; with 7' of 1" pipe seven discharges sufficed and on a second trial only three discharges.

Tests were then made on the floor below (first floor). The 1" S-trap with 1" vent at the crown and 17' of 1" pipe lost its seal in three discharges. In a second and third trial the seal was destroyed in two discharges.

The same, with a 1" vent-pipe 20' long, lost its seal in four discharges.

A 1" S-trap vent-opening at the crown, and a 1" vent-pipe 17' long lost its seal in four discharges, and on a second trial in a single discharge.

Deduction from the Experiments with he Trapless Plunger-Closet.

From these tests we find that the effect of siphonage produced by the discharge of a trapless plunger-closet is not appreciably severer than that produced by one having a trap, provided the trap is constructed of smooth material, has a shallow seal, and is placed near the water in the bowl. The increase of friction is in this case so slight that the manner in which the water discharges into the soil-pipe is not materially modified.

(B) Experiments on Pot and Other Traps Unventilated.