gives as high a figure as the condensing engines, a fact due, however, rather to the exceptionally low total than to exceptionally high friction on the main shaft. The second highest item is, in all cases apparently, the friction of piston and rod, the rubbing of rings and the friction of the rod packing. This is a very irregular item, as would have been naturally anticipated, and amounts to from a minimum of 20 per cent. to some higher but undetermined quantity. The third item, in order of importance, is the friction of valve, in the case of the engines having unbalanced valves. This is seen to be hardly a less serious amount than the frictions of shaft and of piston. But it is further seen at once that this is an item which may be reduced to a very small amount by good design, as is evidenced by the fact that, in the Straight Line Engine, it has been brought down from 26 to 2.5 per cent. by skillful balancing. Ninety per cent., therefore, of the friction of the unbalanced valve is avoidable or remediable. The importance of this fact is readily perceived when it is considered that not only is it a serious direction of lost work and wasted power and fuel, but that the ease of working of the valve is a matter of supreme importance to the effective operation of the governing mechanism in this class of engines. No automatic engine can govern satisfactorily when the valve is unbalanced, and is certain to throw much load on the governor. The frictions of crank-pin, of crosshead, and of eccentrics, are the minor items of this account; they are comparatively unimportant.

Studying these facts with a view to further improvement of the steam-engine, certain inferences are at once obvious. The improvement of the steam-engine has to-day reached a point beyond which, in its thermodynamic relations, but little advance can be anticipated. Under usual conditions of operation of our very best engines, they are so near the efficiency of the ideal engine, working under precisely similar conditions, that the range of possible gain left to us is too small to permit us to look in that direction for rapid or important changes in further increase of efficiency and economy. Where the ideal engine would consume 10 pounds of steam per horse-power per hour, we have actually reached as little as fourteen, if the latest and best reports of the best of modern engines may be accepted as substantially correct; and even this thirty per cent. margin is reduced by practical conditions restricting expansion. If it were to be asserted that we may hope to bring the consumption of steam in good engines of the best

type down to as low as twelve pounds per hour per horse-power, it is probable that the most experienced and best informed engineers would think it a somewhat rash statement; but, in the opinion of the writer, that is what the tendency and rate of recent improvements would seem to promise for the immediate future, assuming that no very great increase in pressures and temperatures of steam may be expected. Practically, also, it is now known that the highest duty is not the most desirable, nor, on the whole, the most advantageous condition of operation of the engine, and we are restricted to lower duties and reduced efficiencies whenever we consider financial relations. It is, nevertheless, the fact, that the conditions of improvement are those which also give higher ratios of expansion for the best point of cut-off and most advantageous ratios of expansion. The duty to seek further means of improvement and higher efficiency becomes all the more imperative when we study the practical conditions under which our engines must be employed. Having, however, as just remarked, so nearly reached the limit of possible gain on the thermodynamic side, it becomes advisable to seek the more carefully for opportunities of improvement in other directions. We have, in the work outlined in this paper, both the directions shown us and the specific method of procedure suggested.

The real, final efficiency of the steam-engine, or of any heatengine, as has been somewhat fully shown in earlier papers by the writer,* and later by others, is composed of the resultant of several distinct efficiencies, as the thermodynamic efficiency, the efficiency of the engine as a heat preserver and user, the efficiency as a machine, and the efficiency of a whole considered from a commercial standpoint. Of these several efficiencies, we have in this investigation the means of studying the efficiency of the machine as a division of the whole within which to seek the best means of securing a gain of total efficiency. The real and final efficiency is certain to be increased if we can effect an improvement at this point, whatever the extraneous conditions of operation. Finding little chance of gain thermodynamically, it becomes our duty to ascertain what are the probabilities of securing progress elsewhere. It is at once seen that the difference here between the real and the ideal engine is greater than in the domain

*On the Several Efficiencies of the Steam Engine. Trans. Am. Soc. M. E., Vol. III., 245.

p.

of thermodynamics, the best cases being in both instances taken. Those engines which are most nearly perfect thermodynamically, are undoubtedly often least perfect, or at least of the least perfect types, when the efficiency of the engine as a machine is studied. Few of them have less than a total of twenty per cent. friction; while they are sometimes probably nearer the ultimate limit of improvement, practically, as converters of heat into work. We are now, for the first time in the history of the theory of the steamengine, in a position to say just where the losses of the machine are in detail, how we are to endeavor to reduce them, in what degree we may hope for such gain, and where it is to be found if effected at all.

The first and most remarkable fact to be noted is the extraordinary amount, absolutely and relatively, of the friction of the crank shaft. This amounts to nearly one-half of the whole waste, and to from five to ten pèr cent. of the whole power of the engine, in the cases here examined. It is remarkable not only for its amount, but also because of the fact that we had begun to believe that, under similar conditions of pressure, speed of rubbing, and of lubrication, it was perfectly practicable to bring down the coefficient to less than one per cent and perhaps to as little as onetenth of one per cent. Here, however, we find, on examination of Table X., that this co-efficient rises, in the unloaded engine, to about 0.30 as a maximum, and, as a minimum, to at least 0.09; while it only falls to 0.04 in the best case, with the increase of pressure on the bearings due to full load and power. This is the more astonishing when it is considered that, on the axle of the car-wheel, it has been found often that the friction is a fraction of one per cent. and often as low as one-tenth per cent. Here is evidently the first place in which to seek further improvement. If this item can be brought down as low as in car-axle journals, the efficiency of the engine as a machine will be increased by about five per cent. in the very best cases, and by ten per cent. in ordinary engines. How this is to be done can be best ascertained when it is found just what are the causes of this extraordinary and previously unsuspected loss. The only conditions apparent tending to exaggerate this waste are the continuous rotation in one direction and the unintermitted pressure of the journal in its bearing. It would appear probable that it is a case of commonly imperfect lubrication. Could the oil-bath system in method and in results be secured here, it would seem probable that the friction

might be enormously reduced. It would even in many cases, if not in all, pay well to have a thoroughly reliable system of lubrication by means of a forcing pump that should insure the support of the journal upon a cushion of lubricant, thus making its action analogous to that of the "palier glissant" of Giffard and the "water bearing" of Shaw and of others.

The second and most obvious conclusion is that the valve should be balanced and so connected as to cause the least possible waste by friction through its motion or that of its moving connections. There is evidently no probable line of improvement so certain to yield a large and profitable result as this. The balancing of the valve has been accomplished, and frequently, during many years past, and so successfully that there is no excuse for neglecting this point in even the cheapest classes of engines. No engine can be considered as belonging to the best class which is not either provided with a balanced valve or which has not a system of valvegear as with some of the "drop cut-off" engines, in which the loss in this direction is rendered insignificant. Here lies an opportunity to raise the efficiency of mechanism of ordinary engines at least five per cent., and of the best of engines with unbalanced valves two or three per cent. It is evidently better, in many cases, to have a valve which is balanced, though slightly leaky at times, than to use an unbalanced valve, though absolutely tight at all times. The simple fact, here revealed, that nine-tenths of this friction may be avoided is very important.

The third item in order of importance is the friction of piston and its rod. This is as great as that just referred to, and is vastly more variable with the class of engine, and probably in the same engine with differences in handling, and especially in setting up packing and springs, where they exist. The writer has often known the power of an engine to be sensibly affected by the carelessness or inexperience of the attendant, who had screwed up his packing in the rod stuffing-box too tightly, and has, on more than one occasion, had a similar experience where the rings were set out too hard. The metallic packings and the unpacked pistons and rods now coming slowly into use will unquestionably do much to remedy this defect of the average engine. Meantime, with the older design, it is perfectly possible to keep piston and stuffingbox tight without wasting much power, or by slowing down the engine by conversion of heat into work at points where the operation is likely to produce serious harm as well as waste. Rings

are much oftener too tight than too loose, and a stuffing-box should only be set up when the engine is running, and then only with fresh packing and not more than is sufficient to check any visible leakage. New packing in a well-made box never needs much compression, and when it becomes necessary to screw it down hard, it is time to replace it by new. Any packing that compels severe compression when new should be promptly condemned.

The remaining items are of minor importance as bearing upon the efficiency of the machine, and they are all obviously easily taken care of by a good designer and a good engineer in charge. Here, if anywhere, it is the fact that freedom of lubrication is the essential consideration, and the more nearly most absolutely flooded the parts can be, and the more absolutely certain lubrication can be made, the better, and irrespective, also, to a great extent, of the cost of the lubricant. Any lubricant freely used can be filtered and cleansed in such manner and so effectively that its more or less free supply to the bearing is a matter of no consequence as a matter of first cost; while the cost of wasted power and fuel, and of repairs due to excessive friction and wear, will usually enormously exceed any apparent gain in that direction. This latter consideration has been very fully treated by the author elsewhere,* and is probably also too familiar to engineers of experience to make it advisable to extend the limits of this paper so as to include more of detail in this and other matters. The importance of the work of which this is the history is sufficient, however, probably, to justify the length to which the paper has already extended.

By reducing the several items of waste and loss above enumerated to their minimum amounts, in the various ways pointed out, and by other less obvious expedients, it is evident that the efficiency of the machine may be so far advanced, in the case of the ordinary engine, as to give us the power of applying from five to ten per cent. of the total indicated power of the engine to useful, instead of to wasteful purposes, and thus to effect a gain of no small amount by improving both sides of the account.

Messrs. Carpenter and Preston have done other work of value, to which it may be possible to give attention later, and Professor

*Friction and Lost Work in Machinery and Mill- Work. J. Wiley & Sons : N. Y., 1885.