At 32° Fahrenheit it will absorb 4.9 per cent. of its own bulk At 50° Fahrenheit it will absorb 3.8 per cent. of its own bulk At 68° Fahrenheit it will absorb 3.1 per cent. of its own bulk *Stromeyer states that under 150 pounds pressure, cold feed water absorbs 3.2 pounds of oxygen per ton. With independent feed pumps there is less opportunity for air to get into the boilers than when the pumps are worked off the engines. Air or oxygen is most corrosive in its action, and this is the reason for the boiler feed delivery pipes being fixed either in the steam space or near the water line. PREVENTIVE.-Where possible, the hot well water should be pumped to a filter tank situated eight to ten feet above the feed pump suction valves. By so doing a large amount of air rises and is liberated from the surface of the water, and a head of water at the suction valves of the pump is assured. REMEDY.-Salt water absorbs more air than fresh water. Care should be taken to keep the pump glands tight, and to efficiently entrap free air in the air vessels. GALVANIC ACTION Formerly, nearly all corrosion in boilers was attributed to this cause, and zinc slabs were suspended everywhere possible within the water space. The position of zinc relative to that of iron in the scale of electro-positive metals, causes it to be attacked instead of the metal of the boiler when galvanic action takes place. PREVENTIVE. To afford protection by the use of zinc, however, there must be positive metallic contact between the zinc and iron. Practically, it is impossible to maintain this contact with the usual methods of installation, and it has been shown that no galvanic current exists after a few hours of steaming, in the arrangements ordinarily employed. REMEDY.-The use of zinc, however, should not be abandoned on this account, as it appears still a very important element of protection against corrosion due to air in feed water. Its suspension in drums, and points within the boiler near the entrance of the feed, is recommended as of positive benefit, and, indeed, as long as zinc slabs continue to disintegrate and oxidize in a boiler, they deflect to themselves from the iron just that amount of harmful action. METHOD OF TESTING WATER FOR CORROSIVENESS The first thing in testing, as is well known, is to see that the color of the water, as shown in the gauge glass, is neither black nor red. The only color * "Corrosion of Boiler Tubes in U. S. Navy," Lt. Com. Walter F. Worthington, U. S. N., Journal of the American Society of Naval Engineers, Vol. XII. admissible is slightly dirty gray or straw color, unless the water is transparent. So long as the water is red or black, corrosion is going on, and it must immediately be neutralized by freely using lime or soda, and frequently scumming and blowing off, the make-up being provided by the evaporator. The salinometer is not a very accurate instrument for determining the quantity of sea water in boiler water, but the apparatus here described gives OWNERS: PACIFIC WHALING Co. BABCOCK & WILCOX BOILERS, 450 INDICATED HORSE-POWER a convenient and accurate method of ascertaining the exact number of grains of chlorine per gallon in the water tested. It is based on the scheme for the volumetric determination of chlorine devised by Fr. Mohr, an eminent chemist, and requires one graduated bottle, one bottle of silver solution containing 4.738 grams of silver nitrate to 1000 grams of distilled water, and one bottle of chromate indicator, which is a 10 per cent. solution of pure neutral potassium chromate. TO MAKE TEST.-Fill the graduated bottle to the zero mark with the water to be tested; add one drop of the chromate indicator; then slowly add the silver solution; keep shaking the bottle. On nearing the full amount of silver solution required, the water will turn red for a moment, and then back to yellow again when shaken. The moment it turns red and remains red, stop adding the silver. The reading on the graduated bottle at the S. S. "AUGUSTUS B. WOLVIN" LARGEST COARSE CARGO CARRIER IN THE WORLD OWNERS: ACME STEAMSHIP Co., DULUTH, MINN. BABCOCK & WILCOX BOILERS, 2500 INDICATED HORSE-POWER, AND SUPERHEATERS level of the liquid will then show the amount of chlorine in grains per gallon. For example, if a permanent red color is shown when the level is midway between 150 and 200, there are 175 grains of chlorine per gallon. The principle of the process depends upon the fact that if some of this silver solution be dropped into water containing a chloride, a curdy white precipitate of chloride of silver will be formed. If there is also present in the water enough potassium chromate to give a yellow color, the white precipitate will continue to form as before, owing to the silver having a greater affinity for chlorine than for the chromic acid in the chromate. But, at the moment when all the chlorine in the sample has been converted, the silver will attack the yellow potassium chromate, and chromate of silver will be formed, which is red in color. The amount of chlorine present is, therefore, shown by the amount of silver solution required to convert it all to silver chloride, and the determination of the exact point at which the chloride precipitate ceases to form is greatly facilitated by observing when the chromate indicator turns from yellow to red. It is not necessary to add the silver solution until the color becomes very red, as the delicacy of the reaction would be destroyed, but the change from yellow to yellowish red must be distinct and must not change on shaking. The sample of water to be tested should be neutral, as free acids dissolve the silver chromate. If it should be acid, neutralize by adding sodium carbonate. Slight alkalinity does not interfere with the reaction, but should the sample be very alkaline, it may be neutralized with nitric acid. GRADUATED BOTTLE Should it happen that the color does not change within the limits of the graduations, the sample may be tested by diluting with distilled water. For example, add three parts of distilled water to one part of the sample. If then, on testing the mixture, the color changes at 200, the number of grains per gallon in the original sample will be four times this reading, or 800 grains. The chlorine should be kept down to the least possible amount-say below 50 grains per gallon as the nearer the boiler water is to fresh water the safer the boilers are against corrosion. If the water is so corrosive as to be acid, blue litmus paper, which has not been allowed to become deteriorated through exposure to the atmos |