This subject has aroused within the last ten years more comment and controversy than any other connected with the blast-furnace. In dealing with the thermal principles of the furnace I shall endeavor to show that the claims for economy made on behalf of the dry blast are well founded, and in treating with the question of operation I shall endeavor to discuss the considerations other than those of fuel economy which affect the subject and must be considered before reaching a correct decision in regard to this matter. Here we have only to do with the design and construction of the plant if its installation be decided upon.

It is well known that the atmosphere contains moisture, but the laws under which the moisture exists are probably less widely understood than any others dealing with a subject which affects our daily life so much.

According to Dalton's law two or more vapors or gases can, in fact and in principle, occupy the same space at the same time, and each acts exactly as though the other were not there. That is to say, if we have two cylinders holding one cubic foot each, one of them containing nitrogen sufficient to exert a pressure of twelve and a half pounds per square inch on the inner surface of the cylinder (that is, not pressure above the atmosphere, but absolute pressure) and another containing oxygen sufficient to exert a pressure of two and a half pounds per square inch absolute, and pump the contents of the second cylinder into the first, each gas will continue to exert the same pressure that it did before, but in this case they will both be exerted on the walls of the same cylinder, and we shall have a nitrogen pressure of twelve and a half pounds, plus an oxygen pressure of two and a half pounds, or a total of fifteen pounds per square inch. Similarly if we have a cylinder containing a cubic foot of water-vapor at an absolute pressure of half a pound per square inch, and pump that into the first cylinder, we shall have a total pressure of fifteen and a half pounds per square inch, and if the cylinder were opened part of the mixture would flow out until the total pressure was down to atmospheric pressure, but the gases and the vapor would be present in the same proportions as before.

This on the assumption that the temperature is high enough to produce a vapor pressure of half a pound, and here we come into complications because the pressure of the vapor depends entirely upon the temperature, or at least is limited absolutely by it. It can never be more than the pressure corresponding to the given temperature, but it may be, and in the atmosphere generally is, less. For our present purpose it seems to me to simplify matters to consider the air not as containing vapor but as acting simply as a carrier for it. That is to say, if we handle ten thousand feet of air the water-vapor in that air must go with it, though not in the chemical sense dissolved therein.

The determination of the quantity of moisture actually present in the atmosphere may be made in several ways. The simplest is to cool very slowly, by some form of artificial refrigeration, a glass or metal cup or vessel containing a thermometer and note the temperature at which visible moisture begins to settle upon it. This is known as the dew point, and knowing this we can easily take the weight of vapor per cu. ft. directly from the vapor tables, which the United States Weather Bureau publishes in very comprehensive and satisfactory form. This, however, is not the method generally recommended for obtaining the dew point. As the result of very extensive investigations it has been stated by practically all the authorities that the wet-and-dry bulb thermometer furnishes by far the most satisfactory and reliable information. The theory of this instrument and the tables for its use have been set forth very thoroughly both in government publications and in a paper by Mr. W. H. Carrier of Buffalo before the American Society of Mechanical Engineers, in Vol. 33, 1911, the latter being particularly valuable.

The quantity of moisture in air is usually given in grains per cubic foot, but this is an exceedingly unsatisfactory unit for blast-furnace work because the other quantities with which we are concerned are measured in tons and pounds, but certainly never in grains, and as the effect of water-vapor is purely a quantitative one it is very desirable to have it expressed in the same units as those used for other quantities involved in furnace operation. For this purpose pounds per thousand cubic feet of air are by all means the most desirable, and as there are exactly seven thousand grains in a pound avoirdupois we have only to divide the grains per cubic foot of the ordinary tables by seven to convert the results to pounds per thousand cubic feet.

Here a point of much importance arises. Air is very commonly measured at 32° F., but when its temperature is raised to say 70° F. its volume is increased about 8 per cent., and as the quantity of moisture increases directly as the volume, the moisture content of a cubic foot of saturated air at 70 is greater than that of the same weight of air at freezing, not only because of the increased density of the water-vapor but because the carrying space in which this vapor exists is simultaneously increased.

In 1906, soon after the first announcement of the success of the dry blast, I published before the (British) Iron and Steel Institute, a paper entitled "Different Modes of Blast Refrigeration, and their Power Requirements," in which was given a chart showing the weight of water-vapor per 1000 cu. ft. of air at temperatures from minus 20° to plus 80° F., but as this question of cubic feet is very troublesome, because we must always state the temperature at which the air is supposed to be measured, I converted this to 75 lbs. of air, which is the weight of atmospheric air at 70° F. and standard barometer. This chart is shown in Fig. 239, from which it will be seen that the moisture content at saturation per 75 lbs. of air rises from 0.07 lb. at zero F. to 1.14 lb. at 70° F.