Branion & Friday, Schenectady, write: "We inclose a rough sketch of a dwelling having hot and cold water supply for two families, each floor having independent stops in the cold-water pipe in the cellar. Each boiler has a stop cock in the cold-water pipe. The hot-water pipe on the first floor leads down by the side of the boiler to the cellar and branches are taken out for the sink and wash-trays. The second-floor hot-water pipe leads down beside the boiler through a partition to the bathroom, which adjoins the kitchen. The complaint made is that the water in the boiler syphons out through the cold-water pipe back into the main. Both boiler tubes are vented as shown in the sketch. The house is situated in a high part of the city and the pressure is very light at times, or not enough to supply the house, hence the syphon. "We understand that boilers sometimes syphon out, but do not see how they can below the vent in the boiler tube. Will you kindly suggest a remedy for the trouble, as we find nothing in your ' Plumbing Problems' that explains the case ?"

[Assuming, to begin with, that there is a good pressure on A, your city water main, that both hot-water tanks are full and that the vent holes B B are free, so long as there is head enough to send water into No. 1 it can be drawn from at any tap below it. The same is true of No. 2. When the head has been so much reduced that the water will flow only to C or D, or just sufficient to fill No. 1, the drawing of cold water on the lower floor will, if a hot-water tap is opened on the top floor, start a syphonage from boiler No. 2, down E and into G at the point of connection F, and so to the point of delivery; or if a hot-water tap was opened on the lower floor, other conditions remaining the same, the water from No. 2 would pass down E, making the circuit at F, pass up H into No. 1, out of No. 1 into I, and to points of distribution through J, though in either case the syphonage would be broken when the water in No. 2 had lowered to vent hole B in circulating tube inside boiler. When the head had been still further lowered to a point below any hot-water tap on the lower floor, upon opening it the syphonage would start from boiler No. 1 down H into G and out to the main through L. But, as before, it would be broken by the supply of air when the water had been lowered to B or earlier, should the hot-water tap be closed. In no case will the water in the boilers be exhausted below B B, but the entire contents of the boilers can be evaporated through B B into the pipes E and H. If a continuous service is required, place independent open tanks in the attic if there is room; if not, then on the roof, housing them well to prevent freezing. Carry the main supply pipe to tanks ending with a ball cock, which will shut off when the tanks are sufficiently full. Then from the bottom of the tanks connect back to the two boilers. If the water will not rise to the open tanks at all times, set independent pumps to supply them, unless it is some one person's duty to see that they are kept full. If so, one pump will suffice. This is the practice in New York, and we think will govern your case.]