Thus, steam boils at 212°, ammonia at minus 27°. Steam requires about twice the heat for the generation of a pound of its vapor that ammonia does. Again, ether, which boils at 95°, requires but 0.55 as much heat to change the temperature of its liquid one degree as does liquid water, and but about one-fifth as much heat to generate a pound of its vapor. Evidently there must be some very general neutralizing element underlying the effect of such widely differing properties upon the efficiency of a vapor, when used as a medium for the transformation of heat into work in an engine, and my object is to suggest a view of the subject which will bring this element into prominence.

The useful effect or "efficiency of fluid" given in column 6 is simply the quotient of the mean effective pressure of the diagram ABCDA, divided by the heat possessed by as much vapor as

will fill the volume AB, or the displacement of the engine-piston up to the point of cut-off.

Now, the expansion line BC, and the compression line DA, differ so little for all these vapors, that the mean effective pressure differs no more than is due, say, to the variation of expansion lines in indicator cards of various steam-engines, from the Mariotte law of expansion or compression.

For practical purposes, therefore, the mean effective pressure, or the numerator of the efficiency value is the same for all the vapors for equal ratios of expansion. But the heat which must be expended is far less in the cases of some of the vapors than in the case of steam for equal weights of substance, such heat being given in column 5 of the table.

If, for example, the space or piston displacement AB, when filled with ether vapor at 100 pounds pressure, contained as much

weight of ether as it would of steam, then the efficiency of ether would be represented by a fraction whose numerator would be the same as that for steam, but whose denominator would be less in the proportion of 195 to 1,002. In other words, the ether would be about five times as economical as steam. But the weight of a cubic foot of ether is nearly as much greater than steam as is 1,002 greater than 195. That is, to fill the volume AB'with vapor of ether would take about as many times more pounds of ether as the heat to vaporize the latter is less than the heat to vaporize

water.

Hence, to obtain a given horse-power from an engine, both the numerator and denominator of the efficiency value are practically equal for steam and ether, and the same principle is the cause of the close identity of the values in column 7 for other vapors, and extends also to air.*

As a general principle, therefore, we may say that substances more volatile than water-in the sense of requiring less heat for the vaporization of a given weight-produce vapors so much more dense than steam, that, to produce equal horse-powers in a given engine, as much greater weight must be used of a vapor as is roughly represented by the ratio of the heat of its vaporization to that of steam; whereby the economy, neglecting differences of cylinder condensation, is practically the same for all common volatile substances as for steam. Experiments with naphtha and ammonia indicate that there may be less cylinder condensation with ammonia and naphtha than with steam-when the former are used so that they expand from a highly superheated condition, and the latter is used without superheating, so that a loss of upwards of 33 per cent. of the theoretical consumption of steam takes place through cylinder condensation. Line 3 of the table applies to the case of ammonia used as a superheated gas, distilled from an aqua ammonia solution. A practical application of ammonia for motive power purposes is being carried out on this principle. The use of ammonia in its saturated condition

*The air cycle covered by the table is what is known as Joule's Engine. See Art. 276, Rankine's Steam Engine. The air taken from the atmosphere is compressed along DA to 465° Fahr. and 100 pounds presure. It is then passed through the furnace, and heated at constant pressure to 607° Fahr., the volume thereby becoming increased by the amount AB. Expansion then occurs along BC, whereby the temperature falls to 160°, and at this temperature the air is exhausted to the atmosphere.

is, of course, impossible if it is to be condensed, and its waste thereby prevented. No means of condensing it at minus 27° Fahr. exists. It was proposed to use ammonia in such a cycle, however, some years since in the Gamgee or zero motor, which offered to boil the ammonia by the natural heat of the sea-water, and expand it to -27°, and return it to the boiler by some means never clearly defined, and yet unknown.

Experiment also has shown, in a satisfactory manner, that the rate of conduction of heat from the furnace to the liquid in the boiler is greater per square foot of heating surface per hour in the case of highly volatile substances, such as ammonia, naphtha and bisulphide of carbon, than for steam. This fact makes it possible to produce a given horse-power with less boiler capacity or heating surface than with steam; but it should be noted that the total heat expended per hour, or the economy, is not affected by this circumstance. The first cost of boiler plant will be less, but the heat which must be supplied to the vapor per horse-power will be practically the same as for steam, as per columns 6 and 7 of the table. Should steam be superheated, so that cylinder condensation may be eliminated, its economy will exceed that of any of the substances in the table, except air.

The difficulties of preventing superheaters for steam from deteriorating have thus far so offset the extra economy known to result from their use that no general use of superheated steam is now being attempted. The high temperatures involved in the use of air-engines (which are known to actually realize most of the high economy promised by theory) have proved an insurmountable obstruction to their competition with steam. The use of volatile vapors, as substitutes for steam, involves objections quite as serious as these. To maintain a back pressure equal to that of the atmosphere requires an amount of water about equal to half that which affords 27 inches vacuum, with steam.

Hence, to compete with the non-condensing steam engine in cities where water is worth, say, $1.50 per 1,000 cubic feet, the vapor-engine will require an expense for water about equal to that of fuel. Besides this fact, the difficulties of controlling the escape of vapor, so that no offensive odors result, and the unknown difficulties due to boiler deposits and corrosion, are such as to make it extremely improbable that the world will ever permit itself to be educated to the use of a substance of this character in place of its elected favorite,-steam.

Mr. F. II. Ball.-I will say that I think this is very interesting indeed, and as I understand the matter, these substances, ammonia, ether, bisulphide of carbon and so on, would be just as much more expensive to use as the difference between their cost and the cost of water, unless, of course, they were condensed and used again. And then there would be all the difficulties of handling which are encountered, which make them more expensive than water. Is that correct?

Prof. Denton.-Yes, sir. In attempting to run a condenser with bisulphide of carbon, they never get the back pressure below the atmosphere.

Prof. Wood.-In regard to the efficiencies, I have verified all except that for ether. In regard to the condensation that Prof. Denton referred to, you all know better than I the great difficulties involved in it, and its intricate character. But I would like to point out still more exactly, if possible, the fact that we have to consider first initial condensation, that is, the condensation in the cylinder up to the point of cut-off. Then, what Prof. Denton said in regard to 10 per cent. of condensation after that is a low figure compared with many investigations. For instance, I went through an investigation-theoretically-where the initial pressure was 100 pounds, and I found for ten expansions there was nearly 14 per cent. of condensation. Now, in order to get this result, the walls of the cylinder must be non-conducting, no heat going in and none going out. This condition is essential. If I remember correctly, Prof. Rankine, after he discovered this theory, made a computation in which the expansion was very large, carrying it down to actual atmospheric pressure, and he found that 18 per cent. would be condensed if the steam at the point of cut-off was dry saturated steam. Now, I wish to emphasize the fact, that without any cooling from the walls of the cylinder, without refrigeration, after cut-off if we have pure saturated steam doing work against pressure, we get this large amount of condensation. When Prof. Denton spoke of it, I raised the question again, How is it, if there was so much condensation, that the adiabatic line rises above the conventional equilateral hyperbola? It is due to the fact that water is present with the steam at the point of cut-off, and the formula which I have presented are intended to cover all these cases. The results deduced from the formulæ involved the condition of wet as well as dry steam at the point of cut-off.