the jackets and three without were taken as illustrating the results obtained. The conditions under which this series of trials had been made were, if anything, more favorable to economy than were any which prevailed in practice; and although the purpose of these engines was to elucidate the causes of inefficiency rather than to realize the utmost economy, yet it was very desirable that the results obtained should not, whether on account of the comparatively small sizes of the engines, or from other causes, fall greatly behind what might be expected from high-class engines in actual practice. Hence it was eminently satisfactory to find that, notwithstanding the drawbacks already mentioned, the economic results compared favorably with anything yet obtained in practice, even with the largest engines. The pounds of coal and of water per indicated horse power per hour were:

Although these results were extremely good, the sources and extent of the various losses were clearly shown. Thus of the total heat received by the engines, exclusive of radiation, with jackets 19.4 per cent. had been converted into work, and without jackets 15.5 per cent.-the

greatest amount which would have been converted had there been no secondary actions being 23 per cent.; so that with steam jackets there were losses through secondary actions amounting to 17 per cent., and without jackets to 34 per cent. The manner of the distribution of these losses was also apparent. One important source of loss, which with jackets accounted for 5 per cent. of the loss, had been brought to light for the first time. This was the heat carried away from the surfaces of the cylinder and passages, in consequence of the expansion after release. The effects of cylinder condensation were clearly shown in the mean diagrams taken from the trials. Although these trials were not in themselves sufficient to determine anything like a complete law of this action, they exhibited in a striking manner its dependence on certain circumstances. One circumstance in particular, which had not previously received much attention, was here shown to be of primary importance in the action of steam jackets. These diagrams showed that with the temperature of steam in the jackets of No. I engine the same as that of the initial steam, the effect of the jackets on the cylinder condensation was very small. In No. II engine, with 80° Fah. difference in the temperature of the jackets and that of the initial steam, the condensation was reduced from 30 per cent. to 5 per cent.; and a difference of tem

perature of 180° Fah. between the jackets and the initial steam in engine No. III entirely prevented condensation. Thus, in these trials, with steam at boiler pressure in the jackets, lowpressure diagrams had been obtained, apparently for the first time, in which the curve of expansion coincided exactly with the curve for saturated steam.

CHAPTER IV.

TRIPLE EXPANSION MARINE ENGINES. *

THE last few years may be regarded as a transition period in the history of marine engineering, as the high-pressure triple-expansion engine has now proved the successful rival of the double expansion compound. The object here is to bring forward the results of experience with this new type of engine, and to consider briefly the various points which have a direct bearing on its efficiency, as well as the most suitable design for marine purposes.

Position of Cylinders.-There has been great difference of opinion regarding the best method of placing the three cylinders in a triple expansion engine of ordinary size; and some very crude plans have been adopted since the introduction of the system by Mr. Kirk in the Propontis. The high pressure cylinder placed on the top of either the intermediate pressure or the low pressure has been tried, together with many schemes to lessen the difficulties of overhauling; while the main objections to this design appear to have been overlooked. With the exception

* Wyllie.

of what was being done in the matter by Mr. Kirk, no further attempt seems to have been made to construct small power expansion engines on three cranks until about three years ago, when the writer undertook to build one of 700 indicated horse power which should fulfill the condition that no more space was to be occupied than would have been taken up with an ordinary compound. Engines of small power had previously been constructed on the tandem principle; but experience had shown the writer that in order to take full advantage of the triple expansion system, an engine must be built on three cranks placed at equal angles; and the best proof that this was the correct solution of the problem is the fact that the arrangement is now almost universally followed, although, at the time, it was considered by the advocates of the tandem engine and others to be a step entirely in the wrong direction.

General Conditions of Efficiency.-The most important conditions to be considered in order to obtain an efficient engine are that there should be approximate equality, first, in the range of temperature in each cylinder; secondly, in the initial stress on each crank; and thirdly, in the indicated horse power of each engine. What may be termed the complements to these three essentials are: 1, steam jacketed cylinders; 2, cylinder ratios; 3, velocities of initial and ex