It is pretty generally agreed that an average of 50 gallons of water per day for each individual is a liberal allowance for all purposes.

The enormous quantities (150 to 300 gallons per head per day) registered by some of the large cities of the United States and paid for by the people, indicate a very large percentage of loss. The chief causes of such loss are leaky mains, bad plumbing and carelessness, the last item being greatly reduced by metering the service.

Authorities* place the average loss through leakage in the joints in the mains at a shockingly high figure, equalling half of the entire supply, or considerably more in many cases than the total amount allowed as a liberal average requirement for each consumer.

*See American Architect and Building News for Jan. 13, 1900, already quoted in these articles.

Assuming this liberal allowance for each individual to be 50 gallons per diem, and that a public supply is the most economical, then those of the 84 million of people of the United States enumerated in the last census who pay for the public supply are actually paying for something near double what they use. If we assume, furthermore, an average meter rate of from 10 to 20 cents per 100 cubic feet, they pay for 50 gallons per day of wasted water at the rate of say 15 cents per 750 gallons, or 1 cent for 50 gallons, which amounts to paying \$3.65 per year for an unscientific system of pipe jointing.

Inasmuch as even this tax is, as we have said, more economical than to obtain water through individual effort without public supply, we may assume that ultimately all the citizens of the United States will be glad to enjoy the privilege of paying 50 per cent, for leakage, unless some more scientific form of jointing is adopted, and the average, \$3.65 loss per citizen per year, will aggregate the trifling sum of over three hundred millions of dollars a year.

The United States census for 1900 gives the total value of gas sold in the United States for the census year as \$69,-432,582.00. Calculating a proportionately similar waste from bad gas joints we have for the annual leakage in the mains of both water and gas in this country not far from half a billion dollars.

Therefore even a small improvement in the matter of pipe jointing is worth while. A joint which would save only 10 per cent, of this loss would mean an annual money saving alone of many millions of dollars and corresponding sanitary advantages.

*See very interesting article by James C. Baylies on Gas Leakage in "Domestic Engineering" for July and August, 1902.

Author's Attempts to Improve Pipe Jointing.

Our first experiments in pipe jointing were made in 1883, and were conducted on the inadequate idea that if a perfectly rigid joint could be economically constructed the problem would be solved.

It was assumed that if such a joint could be made adjustable in setting up, and approximately as rigid when assembled as the pipe itself, the utmost attainable would be accomplished.

The joint then devised, and originally called by the writer the "Sanitas" joint, as constituting part of his "Sanitas" system of plumbing appliances, is shown in Figs. 522 to 535, inclusive. Recently, however, the small spigot on one of the flanges has been added, forming an important improvement in this joint, and, by way of distinction, the name "Securitas" has been given to the improved form, including it with his other recent improvements in sanitary appliances going under this name.

The "Securitas" Flanged Joint.

To avoid the difficulties connected with the threading and screwing together of wrought iron pipe, and to permit of the use of cast iron without the defects involved in hand calking, this joint was devised and put into successful use in house building by the writer. In general terms it may be described as an adjustable flanged joint with lead gaskets for packing forced in place by bolts after the manner of flanged steam pipes without the employment of skilled labor. It is a steamfitter's joint with improvements which adapt it for plumbing, gas and water carriage where a rigid joint is desired.

The pressure is applied by two ratchet wrenches, Figs. 529 and 530, constructed for the purpose and used simultaneously, one working left and the other right handed, as shown in Fig. 531. This avoids the necessity of securing the pipes while the nuts are being screwed up, and causes both sides to be compressed alike, since the wrench which has given and received the greatest pressure ceases temporarily to turn until the other has caught up with it. This permits the joint to be made up in very contracted places, as shown in the figure, and by a single ordinary unskilled workman in less than twenty seconds after the pipes are once set in place. To calk an ordinary bell-and-spigot joint in the usual defective manner is estimated by good authorities as requiring, on the average, the pipes being in place, as many minutes. Moreover, two men instead of one are required for it; one, the plumber, to do the calking, and the other, the helper, to handle the fire and melt the lead.

Fig. 521.

Fig. 522.

Fig. 523.

Fig. 524.

Fig. 525.