What is the capillary effect of large and small quantities of the sediment found in traps, and how can the loss of water by this agency be prevented?

It is evident, first, that the substances exerting the capillary action can conduct the water only to a certain limited distance above and beyond its surface, and that the rapidity of the removal of the water in a closed vessel will be in proportion to the shortness of the distance required to be raised; second, that capillary action in an open vessel greatly increases the loss by evaporation, and that the rapidity of the removal of the water in vessels of similar form but exposed to different degrees of change of air will be in proportion to the velocity and hygrometric condition of the air currents. Hence, if we use a trap having a seal of proper form, and do not allow the air above the trap to be changed in such a manner as to cause evaporation, aggravated by the spreading out of the water through capillary action, the trap will be secure against loss of seal through this agency.

To ascertain the distance which water will travel above the seal of a trap under the influence of capillary action, a number of experiments were made with various materials, such as are liable to collect in traps in practical use, including among them those which are found to have the maximum of effect in conducting the water by capillary action. The experiments were made both in ordinary open glasses and in different kinds of traps, both open and closed.

The first tests were to ascertain the perpendicular distance. Figs. 283 to 286 show the manner in which I made the tests with ordinary goblets. A number of these glasses were supported on blocks and filled with water. Over their edges were hung the different substances to be tested, one end extending below the bottom of the water in the goblet, and the other to the top of a tumbler placed below it, as shown in the figures, to receive the water drawn from the goblets. The substances tested were matted hair-felt, lamp-wicking, both with and without its covering; jute; hemp-cord, unraveled and separated into fine fibres; hemp-cord in its natural braid, and pieces of string. Of all the substances tested, jute, such as is used by plumbers in caulking joints, proved to be the most effective in removing the water by capillary action, and this is the substance I shall use in our tests to-night. Different amounts of each substance were used in making the experiments. In each case experiments were made both with a small quantity of each material and with a mass large enough to completely fill the waste pipe. As might be expected, the water was most quickly removed when the quantity of the substance tested was large enough to completely fill the pipe. These experiments showed the limit of the carrying power in an upward direction with these materials under capillary action to be 4 inches. The water was rarely lifted more than 3 inches or 3 inches. The majority of the water carried into the lower glass was moved during the first ten or twelve minutes, the rapidity of the action gradually decreasing as the water descended, becoming extremely slow when the level was reduced about 2 inches, and generally ceasing altogether and the fibres becoming entirely dry at the top when the water in the upper vessel was reduced about 3 or 4 inches.

Fig. 284. Capillary Action Forming Longer Siphon.

Fig. 284. Capillary Action Forming Longer Siphon.

Fig. 286. Capillary Action with Horizontal Extension.

Fig. 286. Capillary Action with Horizontal Extension.

Fig. 285. Capillary Action with Long Siphon.

Fig. 285. Capillary Action with Long Siphon.

In emptying a vessel of water by means of a bent tube forming a siphon, the excess of length of the outer over the inner limb of the siphon governs, as I have shown, the rapidity of the flow. With capillary action, however, though the outer limb must always be longer than the inner, yet beyond a certain fixed point the excess appears to have no marked influence. A small predominance of the outer limb, as shown in Figure 283, or just enough to overbalance the column of water in the inner limb, carries off the water as rapidly as the long limb shown in Figure 285.

The results of the experiments with ordinary open vessels were as follows:

(a) Hair-felt. This is a material which closely resembles the matted deposit of short hairs which form so large a proportion of the deposit in traps and waste pipes. Strips of this felt one-quarter of an inch thick and of various widths and lengths were tested under different conditions. Tested with the glasses arranged as in Figure 283, with a strip three-quarters of an inch wide, it lowered the water in the glass 2 inches in the first ten minutes, but required four hours and a half to diminish the level another inch. After this no more water was carried' over into the lower vessel, but the rate of evaporation of the water in the goblet was considerably increased by the hair-felt, which lifted the water to a certain height and distributed it over the fibres of the felt, thereby exposing a very large surface to the action of the air. A piece of felt 2 inches wide tested in a similar manner gave similar results. An increase in the width or length of the felt did not make, in this case, an increase in the amount of the water transferred from the upper to the lower vessel corresponding to the increased size, but increased the velocity.

(b) Lamp-wicking. A material closely resembling the soft, porous lint formed in traps and waste pipes. This was tested both with and without its cylindrical fibrous covering, as it comes prepared in the market for use in lamps. The wicking was cylindrical in form and about inch in diameter. It was first tested without its cover. Placed as shown in Figure 283, the wicking lowered the water only 1 inches in seventeen and one-half hours, after which no further transfer took place. Placed as in Figure 284, 1 inches were transferred in the same time. With the wicking covering on, only half the amount of water was transferred in the same time.