slight loss from condensation in the high-speed engine, but it shows in engines moving at very slow speeds more heat wasted by condensation than there is utilized.

Many of the rolling-mill-engines of the country are of good type, but most of them will use 6 pounds of coal per horse-power, despite their comparatively high speed. What shall we say of mill-engines without cut-offs or condensation, where coal is dear, and where steam is made by firing under boilers?

Now, why does this general wastefulness of steam go on from year to year? Mill-managers have a ready answer. They want "a strong, simple engine, that will stand the rough attendance it gets in a mill, and that won't break down." The first reason is ridiculous; yet, you can find plenty of mill-engines with leaky pistons, out of line, and wasting their cost every year, to save a dollar a day more pay to an engine-driver. The second reason is perfectly sound. An engine may better waste a few thousand tons of coal, if it only makes a regular business of it, than to go to pieces without notice, in the middle of a time-contract. But the inference that an engine which is economical of steam, is, for that reason, more likely to break down, is totally without foundation. The commerce of the Atlantic Ocean is largely done by compound engines, which are just twice as complicated" as single-cylinder-engines, and which run with 2 to 3 pounds of coal, and which are subject, at every pitch of the ship -many thousands of times every voyage-to the violent strains of the severe plunging in and out of water-strains which often shake the ship from stem to stern; but we rarely hear of these engines breaking down. I have recently taken time to examine, in detail, the working of eleven specimens of one of the most economical, highly-finished, and delicately-adjusted engines ever built-the Porter-Allen engine and I find them notably free from breakdowns and abnormal wear. They had been running from 2 to 6 years, and in some cases the working parts had not been readjusted at all. It is true that some of the earlier Corliss mill-engines were too light, but this had nothing whatever to do with their economical steam-distribution. Probably there are more successful engines of this type than of any other.

The stability and durability of an engine is simply a question of good proportions, materials and workmanship. Of course, a piston that takes the boiler-pressure at the first opening of the valve, receives more force than one which gets only 60 or 70 per cent. of it, and that not till quarter-stroke, of which there are many examples.

But on the Porter-Allen demonstration, this force is required to start the very heavy reciprocating parts, which, in turn, give it out to the crank during the decrease of cylinder pressure by expansion, so that the strains on the crank-pin are not necessarily greater with good than with bad steam-admission. If the sudden impact of steam is hard on machinery, what should be the result of a 10-ton hammerhead, falling 6 or 8 feet upon an 80-ton anvil? And yet, steamhammers are made to run for years. Some of the most economical valve-gears are not complex; the complex ones a boy can work by hand, so that they are not under dangerous strains, and rarely break down. Finally, the mill-engine, by means of the interposed flywheel, is relieved from all violent shocks, while the marine-engine has no such protection.

The objection, therefore, that engines are liable to breakdowns and delays, because they are economical of steam, will not hold. If the builders of such engines do not make them strong enough, the engineer of an iron or steel-works ought to know it beforehand, and to know what changes to specify. There are numerous examples of sturdy engines which he can study. I will venture the opinion that more breakdowns are caused by accepting the lowest bid for engines, than by all other causes combined.

There is a great temptation to save the heavy cost of radical change, by patching up old engines, attaching condensers, applying cut-off valves, enlarging cylinders to get more expansion, running faster by means of change of gearing or belts. All these things may save fuel, but do they not stand in the way of a greater economy, by constantly adding cost to the old engine, and so giving it a long lease of life? No man ever throws out an engine which he has just rebuilt, however bad it may be. The way to get the greatest possible profit out of an engine of bad type, is to melt it down in a cupola. The excuse is often made that a new engine cannot be afforded. One would suppose that an establishment which can afford to waste $20,000 a year in steam-coal, might afford to lay up half that sum every year to invest in better engines.

The blowing-engine presents a larger problem than mere perfection of valve gear. The air-piston should not run fast; the steam-piston must not run slow. Gearing a pair of small engines, making, say, 150 revolutions, to a large air-piston making 25, entirely overcomes the objectionable features of gearing, which only works harshly when the small wheel is driven. There is a possibility by this arrangement, of saving more steam than can be done even by compounding,

which would cost about the same. Compound engines have certainly achieved a very great success for marine, pumping, blowing and other large scale uses. They, however, involve condensation, and either a large water-supply or extensive cooling-ponds, both of which are often very costly, and in some cases impracticable.

The use of the indicator-not the vile instrument made only to sell, but Elliott Bro's. Richards indicator-should be a matter of regular practice. I will venture to say that a high degree of steamengine economy cannot be maintained without the regular application of the indicator, to show whether or not the condition and functions of the machine are normal and healthy. Beyond a few hundred dollars cost in preparation, the expense of taking cards once a week is absolutely nothing.

There are vast numbers of bad boilers, as any one will conclude who takes pains to observe the innumerable proportions of grate, combustion-chamber and heating-surface in use for the same kind of fuel. Fortunately for mill-owners, great improvements rarely require such sweeping changes in boilers as in engines. The changes, however, should be scientifically devised. Altering the style of a grate when its surface is insufficient, or increasing draft when combustion-room is too small, will not promote much economy. The employment of a commission of experts to ascertain the real nature of the defect would be a paying investment in a great many cases. The diseases of boilers and their setting are often obscure, and the diagnosis of a cheap engine-driver is not infallible. Perfect combustion, while boilers are being fired by hand, is probably impracticable, but bad combustion all the time is entirely unnecessary. Steam-induced air-jets sweeping over the surface of the fire, where there is heat enough to ignite the combustible mixture thus formed, is an effective and cheap arrangement. Mechanical firing, as done at Barrow and other works abroad, promotes almost continuously perfect combustion. I cannot dwell upon this subject; it would require a treatise by itself.

II. Improved Heating Furnaces.-The commercial importance of this subject is no less than that of the preceding. The coal-fired reverberatory furnace, however skilfully managed, must be wasteful of both fuel and metal. Much of the coal placed on the grate is lost through the grate; the very irregular volume of air passing unconsumed through the solid fuel is at one time insufficient to unite with the combustible gases over it, and at another time great and undistributed enough to go bodily over the bridge and consume the metal

on the hearth; combustion cannot be perfected when firing is done by hand, without an excess of air which will waste the metal, so that the flame must always be smoky. The heat passing out of the furnace may be utilized under boilers, but not as economically as the same quantity of heat can be under coal-fired boilers, nor any better than waste heat can be utilized in regenerators.

Do iron-makers realize the enormous loss due to oxidation? An iron-rail mill making 40,000 tons of product, heating all the material twice, and oxidizing not less than 8 per cent. of it at each heat, would, at present prices, burn up more than $200,000 worth of iron in a year. Upon averaging a number of results, I find the saving in the oxidation of iron in regenerative gas-furnaces, as compared with coal-furnaces, to be over 3 per cent. In one case of first-rate practice with both furnaces, on small iron billets, it is 3.32 per cent.; in another case of good average practice on large iron piles, it is 4.45 per cent. In heating iron piles for plates, the waste in the ordinary furnace has been in some cases as high as 15 per cent., while in Siemens' furnaces, which have been substituted in the same works, it has been as low as 4 per cent. The smaller of these savings would amount, in the rail practice we are considering, to some $70,000 per year, which would pay for half the labor on rails, or it would pay above 20 per cent. on the cost of a rail-mill. The oxidation of steel is somewhat less in either furnace, because the required temperature is lower; but the proportion of loss appears to be about the same, so that the economy of the gas-furnace is also very important in heating steel.

The amounts of fuel used in gas-furnaces are surprisingly various in different works. They run from 350 to 650 pounds per ton of rail-piles and blooms, and steam-coal varies similarly. These figures indicate very bad working of furnaces in some cases; and the absence of all definite data in some other cases, as to steam and gas-fuel used, indicates, at least, that bad working may be going on without the knowledge of the management. The fuel for heating rail-piles and blooms in ordinary coal-furnaces also varies from 700 to 1200 pounds per ton, the differences not being wholly due to plant or management, but to quality of coal. With ordinarily bad boilers and engines, the saving of fuel, including steam-coal, by the use of the gas-furnace, varies from 5 to 25 per cent. A fair comparison of the two furnaces should be based on the best practice in both cases. If coal-furnaces lose heat enough up chimney to make steam for wasteful engines, no economy of fuel is attained by using economical ones. But as

naces.

regenerative gas-furnaces, which make no steam, are necessary to the minimum oxidation of the metal heated, steam must be generated by firing under boilers, so that good engines in this case have an unlimited opportunity to save fuel. Thus the gas-furnace economizes, first, and chiefly, in oxidation; second, in fuel directly; third, in fuel indirectly, by giving economical engines a chance; and fourth, it also saves largely in quality of fuel, as cheap slack and small coal may, in most cases, be burned in gas-producers, but not in coal-furTo these advantages must be added decreased loss of coal due to burning it in one concentrated nest of producers, rather than all over a rolling-mill, and the general convenience and economy of avoiding the handling and storage of coal in various parts of the works. Still another important advantage of the regenerative gasfurnace is the facility it affords for varying the chemical character of the flame, and more especially for maintaining the neutrality and perfect uniformity of the flame. This furnace seems essential to the production of the high temperatures required in the Martin steel process. The Siemens furnace has, within the last few years, been so highly perfected in its proportions and details, that its adaptation and working are no longer subject to any unusual risk or embar

rassments.

The most imperfect feature of the gas-fuel system, at least in the English and American practice, is the gas-producer, which has not been materially improved for years. As it is simply a fire-box with a grate below, and is subject to variation in firing, in stirring, in, the caking of the coal, in the holes in the fire through which carbonic acid gas and air may pass up into the chamber, and in air-admission, due to the irregular formation and removal of clinkers, the most careful attendance cannot insure the regular production of good gas, nor prevent the burning of some gas above the coal-bed. The Ponsard producer, on the continent, and in some experimental practice here, promises well. It dispenses with the grate, promotes uniformity of combustion, and furnishes gas to the furnace at a higher temperature. The gazogen of Mr. Tessié du Motay is said to be very successful abroad; fourteen are running at the works of M. de Wendel, in Lorraine. The same producer is employed for the production of illuminating gas at the new works in Buffalo, Troy, and New York. There are many other schemes for the improvement of gas-generation, which I cannot even refer to within the limits of this paper. It thus seems probable that a considerable