This syphon, in size suitable for a single house, costs about $12 delivered, and will always be available to secure an intermittent dosing of the bed or pipe line. Usually the chamber in which this syphon is placed holds about one hour's flow, so that it may be estimated that this syphon will discharge on the bed every sixty minutes. The exact interval of time is not essential nor, perhaps, important, although it may be noted that the coarser the material,--that is, the nearer uniform all the sand particles are to the largest size passing the ten-mesh size,--the smaller must be the dose applied, but the more frequently must the application be made. This has been very thoroughly studied in Massachusetts, and the views of experts on this subject may be found in the report of that Board. Such an intermittent discharge may be made and often is made by a hand valve leading out from this chamber in institutions or in private houses where some one constantly is available for the purpose. Thus it becomes the duty of the man in charge every hour or perhaps three times a day to pull the valve and allow the sewage to discharge (see Fig. 71). An overflow pipe should always be provided, so that if he forgets to pull the valve, the sewage will still find its way into the system rather than out on the ground. [Illustration: FIG. 71.--Plan and section of a septic tank with valve.] _Sedimentation._ As a matter of economy of operation, it has been found desirable to take out from the sewage before the treatment already described as much of the solid matter as may be reasonably done, and for this purpose sedimentation is made use of. Most of the solids in sewage are slightly heavier than water, so that if they be allowed to stand in the water for a short length of time, they will settle to the bottom of the tank and allow the liquid above to pass on, considerably clarified. It has been found worth while to do this, since all three processes described are interfered with if the solids taken out by sedimentation are allowed to be deposited either upon the surface of the ground, giving rise to odors as well as to objectionable appearances, or onto the surface of the sand beds, which they clog up, or in the three-inch tile drain, which may be filled in a short time. It has been further found by experience that if these sedimentation tanks are made large, really larger than necessary for sedimentation, in some way a large proportion of the matter accumulating in the tank will disappear, so that the amount of sediment to be taken out of the tank is not as large as might be expected. In fact it is usual for such tanks to run one or two years without cleaning, although the amount of solids shown by chemical analysis to have been removed from the sewage would fill the tank twice over.