These experiments are divided into: - (1) Those in which the siphonic action was produced by a trapped closet.

(2) Those in which a trapless plunger-closet was used.

(3) Those in which a flush-tank was used.

(1) Experiments on the Siphonage of Unventilated Pot and other Traps by a Trapped Plunger Water-Closet.

(a) The tests were first made with the water-closet and bath-tub discharging together. The pot-traps had 1" or 1" inlet and outlet arms.

A 2" pot-trap had its seal broken, and the water lowered " below the top of the inlet mouth by a single discharge. Five discharges lowered the water nearly to the bottom of the mouth (see Fig. 2).

A 2" pot-trap lost its seal in two discharges (see Fig.

3).

A 3" pot-trap lost its seal in four discharges (see Fig. 4).

A 3" pot-trap lost its seal in seven discharges (see Fig.

5).

A 4" pot-trap lost its seal in seven discharges (See Fig. 6). A 5" pot-trap lost its seal in twenty-two discharges (see Fig. 7). A 6" pot-trap lost its seal in twenty-seven discharges (see Fig. 8).

An 8" pot-trap lost 1" of its seal in twenty-four discharges (see Fig. 9).

A 4" bottle-trap lost its seal in fifteen discharges (see Fig. 10).

A 4" Holland's trap retained 1/16" seal after forty discharges (see Fig. 11).

A "Sanitas" anti-siphon trap retained over " after fifty discharges (see Fig. 12).

The loss of water in the Holland's trap in the last ten discharges was exceedingly slow, showing this trap, which is similar in outward appearance to the 4" bottle-trap, to offer much greater resistance to siphonic action than a bottle-trap of the same general dimensions.

The rate of loss in the "Sanitas" trap constantly diminished after the first few discharges. Several experiments were made on this as on the other traps. Figure 12 shows the effect of sixteen discharges. Figure 25 represents in diagram the record of another experiment on the same trap where the test was prolonged to fifty discharges. It will be observed that the loss towards the end was scarcely perceptible. In the first ten discharges in this experiment the seal was lowered 1 7/8". In the next ten the loss was only one-eighth of an inch, which is equivalent to one-eightieth of an inch for each discharge. In the third ten discharges i. e., one whose top stood 6" above top of seal was first tested it was still further reduced to one-sixteenth of an inch. In the fourth to less than one-sixteenth of an inch, and in the last ten to still less, or about one thirty-second of an inch. As there still remained over " seal the trap may be considered as practically unsiphonable.

Figures 14 to 24 inclusive represent in diagram the result of the experiments already described on "pot" and "bottle" traps.

The perpendicular lines represent the depth of seal of the traps.

Fig. 14 (568). Fig. 20

Fig. 14 (568). Fig. 20

(574).

Fig. 15 (569). Fig. 21 (575).

Fig. 15 (569). Fig. 21 (575).

Fig. 16 (570). Fig. 22 (576).

Fig. 16 (570). Fig. 22 (576).

The Siphonage And Evaporation Of Traps Report To T 607

Fig. 17 (571).

Fig. 23 (577).

The Siphonage And Evaporation Of Traps Report To T 608

Fig. 18 (572).

Fig. 24 (578).

Fig 19 (573). Fig. 25 (579).

Fig 19 (573). Fig. 25 (579).

The circles indicate conventionally the outlet and inlet mouths of the traps, and the horizontal lines the loss of water at each discharge.

(b) A 4" pot-trap was then tested with the water-closet alone. Its seal was broken by sixteen discharges (See Fig. 22).

(c) a 3" pot-trap was then tested with the flush-tank. The first discharge almost and the second entirely broke the seal. Nevertheless, in other experiments made with the flush-tank the siphonic action proved less severe than that produced by the simultaneous discharge of the water-closet and bath-tub.

Deductions from the Experiments on Pot and other Traps.

From the above experiments we learn that the power of resistance of "pot-traps" depends upon their size, and more particularly upon the diameter of the body. It will be observed that the depth of seal of the 4" pot-trap is only 2", while that of the 3" pot-trap is 3". This accounts for the similarity in the results of the tests on these two sizes. A half an inch excess of diameter of the body offsets, in the first series of tests, one inch excess in depth of seal.

It will be observed that, had the seal of all the traps tested been the same in depth, i. e., 2", the resistance would have been in direct and regular proportion to the diameter. Thus, in the first series of tests -

A 2" pot would have lost its seal in less than one discharge.

A 2" pot would have lost its seal in one discharge. A 3" pot would have lost its seal in two discharges. A 3" pot would have lost its seal in three discharges. A 4" pot would have lost its seal in seven discharges. A 5" pot would have lost its seal in ten discharges. A 6" pot would have lost its seal in twenty-eight discharges.

An 8" pot would probably have resisted over one hundred discharges.

Hence (1) the resisting power of pot-traps of equal depth of seal is in direct proportion to the diameter of the body.

(2) No pot-trap whose body does not exceed in sectional area 15 times that of each of its arms or connecting pipes can be accepted as anti-siphonic under all conditions likely to be encountered in plumbing.