practice I have used partly one and a quarter inch pipe in lines over forty or forty-five feet long. In this line of perforated pipe fifty-five feet long, the jets nearest the supply-pipe would rise about twelve feet, and at the extreme end eight feet, under our assumed head (forty feet). In this connection an important fact should be noted, which is that the height of jets under lower pressures diminishes rapidly and about in proportion to the pressure, while the quantity of water discharged was reduced but little. For instance, under twenty feet head, or one-half that in the case above recited, the quantity of water discharged would be reduced but about ten per cent, while the height of the jets at the extreme end of the pipe would be reduced from eight feet to four feet. From what has now been stated, it will be seen that if an effective head of forty feet be maintained in the supply-pipes of a system of sprinklers at the points where the perforated pipes are connected with them, we may expect the jets to rise a minimum height of eight feet, and that the discharge of water will be at the rate of one-tenth cubic foot from each lineal foot of pipe. If we now suppose our lines of perforated pipe to be placed fifteen feet apart (which I will suggest should be the extreme limit), we shall discharge one cubic foot of water over one hundred and fifty square feet of floor, or eight one-hundredths of an inch in depth, per minute. This quantity is somewhat in excess of that mentioned by Mr. Francis as the amount which would be discharged by the sprinkler arranged by him for the Tremont & Suffolk Mills. He states the quantity at onefourteenth, or say seven one-hundredths, of an inch in depth per minute over the floor protected. I do not wish to tax your patience with details and figures, and have so far as possible kept them from this paper; but I wish to impress you with the fact that an effective sprinkler is not a lot of pipe with a lot of holes in it, and that it is necessary to consider what size the orifices should be, how much water they will discharge, and where the water is coming from to supply them. I have made experiments to determine the amount of water discharged from different-sized orifices drilled in a piece of ordinary wrought-iron pipe; and the result for our purpose may be stated as follows: Taking the quantity discharged by our one-fourteenth of an inch orifice at one hundred, the quantity discharged by a onetwelfth of an inch orifice will be one hundred and seventy-two, and by a one-eighth of an inch orifice three hundred and thirteen. It will thus be seen that it makes a vast difference, in providing a supply of water for sprinklers, as to the exact size of the orifices. Soon after deciding upon one-fourteenth of an inch as the size of orifice, I arranged a working system of sprinklers consisting of five lines of perforated pipe, each about one hundred and ten feet long, which was exhibited to a large number of manufacturers and gentlemen connected with your mutual insurance companies. Mr. William B. Whiting, in an article written upon "Sprinklers for Mills" in January, 1875, after alluding to the arrangement of pipes says, "We have seen a sprinkler constructed on these proportions, and temporarily erected in a building on the premises of the Albion Company in Rhode Island, at work, and in common with all others who saw the same we feel sure that no fire in its earlier stages could live a moment under such a shower-bath as it afforded." It is necessary to state, however, that Mr. Whiting at the time of writing supposed the orifices in this case were one-twelfth of an inch, whereas they were but one-fourteenth of an inch in diameter. Before considering the size of supply-pipes, it is necessary to state that the perforated pipe has usually been arranged in sections, covering the entire width of a mill, and not to exceed one hundred feet of its length. Each of these sections has independent supply pipes which lead from the "distributer" so called, and through which they receive the water from the supply-main. At the distributer there is a gate valve in each supply-pipe, so that water may be admitted to either section of the sprinklers, independent of the others. Each valve is marked in a manner to show the location of the section to which it is attached. The distributer should be placed where it is easy of access, and where fire cannot possibly prevent approach to it. No general rule can be laid down to govern the size of supplypipes. If the water is to be supplied by a steam or rotary pump, it will be found in general practice that a supply-pipe having an area one-fourth greater than the aggregate area of the orifices will deliver the required amount of water without im pairing the action of the pump by reason of friction in the pipe, but of course the work to be done by the pump depends upon the length of the supply-pipes, the height of the sprinklers above the pump, the length and size of pipe connecting the pump to the distributer, and other details. On the other hand, if we are to make a reservoir supply available, the area of our supply-pipes may require to be two or three times the area of the orifices, in order to have the loss of head by friction so small that an effective head will be maintained in them. It would be tedious to make any calculations or work out any particular examples to illustrate this. In fact, the best service I can render in connection with this whole matter of sprinklers is to make it apparent that it is necessary to have the work carefully and intelligently planned by persons having the requisite time and information to enable them to do so. Having arranged our sprinklers from, and including the distributer, let us next inquire as to the total quantity of water necessary to render them effective. In doing this we encounter the disappointing feature of the sprinkler system; the fact being, that only a limited portion of the sprinklers placed in a mill can be supplied at one time with the pumps in general use, or the available reservoir supplies. The erroneous idea has prevailed, that only one section, or, at most, all of the pipe in any one story, would be used at one time; and the aim has been to have parties provide pumps of sufficient capacity to supply the sprinklers in at least one story of their mills; but we shall find, as our investigation proceeds, that even this has not usually been done, while disastrous experience shows that there should be a supply for two or more stories, in order to have a really efficient fire-extinguishing apparatus. We have seen that one-tenth of a cubic foot of water is required for each lineal foot of perforated pipe. In a mill 350 feet long and 72 feet wide we should have in each story, say, 1,750 feet of perforated pipe, which would require 175 cubic feet, or 1,312 gallons, per minute. To supply this quantity we should require a rotary-pump of the largest size, delivering say eight hundred gallons per minute, and a steam-pump having a water-cylinder eight-inch diameter, eighteen-inch stroke, with a steam-cylinder say sixteen-inch diameter, which would deliver five hundred or six hundred gallons per minute, if making one hundred and fifty strokes per minute. From this it will be seen that it is necessary to have accurate information as to the amount of sprinkler-pipe the pumps will supply, and that care and judgment are required in the use of them in time of fire. The inconsiderate opening of too many valves would be certain to result in dispelling the water over too much surface, and fail to extinguish the fire. Let us now investigate the matter of the water pressure required at the pumps, and the consequent steam pressure required in the boiler in an assumed system of sprinklers. Let us suppose our sprinklers to be in the fifth story of a mill, say sixty feet above the pump, and the pump connected to the distributer with a six-inch pipe one hundred feet long. We will suppose the sprinklers in this story to be divided into four sections, and each section connected to the distributer by a three-inch pipe. Say the average length of the supply-pipe, from the distributer to the perforated pipe, is one hundred feet. Assume the quantity of water to be distributed at one hundred and seventy-five cubic feet per minute, and the head required at the orifices forty feet as heretofore. Each of the four three-inch supply-pipes will have to carry one-quarter of one hundred and seventy-five, or say forty-four cubic feet, of water per minute. We shall then have as the head or pressure required at the pump as follows: Head at the sprinklers 40 feet. Height of sprinklers above the pump. 60 66 Loss by friction in the three-inch supply-pipes, 100 feet long 25 66 13 " 3.5" 3.5" Loss by friction in the six-inch main, 100 feet long Loss by influx into six-inch pipe from pump and into the threeinch pipe from the distributer, is 1.75 feet for each change, or for both. Friction in the distributer, say Total head required Or, reduced to pounds pressure, we have 65.145 pounds. We have thus to maintain a water-pressure, at the delivery-nozzle of the steam-pump, of 65.145 pounds, or say sixty-six pounds. The area of the steam-cylinder, of the pump we have assumed, is four times the area of the water-cylinder: hence the steampressure would have to be 66÷4, or but 16.5 pounds greater than would be required to run the pump, say, one hundred and fifty strokes per minute, and deliver the water through the delivery-nozzle of the pump, without any pipe attached. The steam pressure required to run the pump, and deliver the water through an open nozzle, would depend upon its construction and condition, and will be found to vary from ten to twenty pounds or more. Then we have the lift and loss by friction in the suction-pipe (which, of course, should be made as little as possible), and if, say, eight pounds in the water-cylin- · der, would call for two pounds in the steam-cylinder. Further, a small allowance should be made for difference in pressure in the boiler and pump-cylinder. Summing these results, we can say that it will take from thirty to forty pounds boiler-pressure to discharge water effectively from sprinklers in the fifth story of a mill, with the arrangement we have here assumed. While this result is very satisfactory, and we find that our work in this particular case is to be done with a reasonably low steam-pressure, it nevertheless shows that proper knowledge should be had of each particular system of pumps and sprinklers, that we may not attempt to use them with twenty-five pounds boiler pressure when fifty might be required. As briefly as possible let us now examine into the matter of supplying our fifth-story sprinkler from a city water-main, and also from a local reservoir built for the protection of the mill or mills. The result just worked out shows that if the six-inch pipe which connected the pump to the distributer, and which we assumed to be one hundred feet long, was connected to a city water-main one hundred feet distant from the distributer, we should require 151.5 feet head to be maintained in the city main, while one hundred and seventy-five cubic feet of water were being drawn from it. There are but very few localities where this supply can be maintained. A partial remedy of course would be to enlarge our pipes, but it would be found that we should require about one hundred and fifteen feet head at the city-main with the largest pipes it would be practicable to use in the mill. If we could build a reservoir one thousand feet away from the distributer, and say at an |