Two Thompson and two Tabor indicators were used, and interchanged occasionally; and all the springs and pressure-gauges were tested at the close of the trial. The exponent of the expansion curve

in the expression, P =

; or, PV" = C, a constant, was found,

by the logarithmic method, to be n = 1.118 about 10.

The engines were examined for tightness at the beginning and end of each run. All clearance spaces were accurately ascertained. Diagrams were taken simultaneously every 10 minutes, and, during the first 10 minutes of each run, once a minute.

PERFORMANCE OF THE ENGINE, DEDUCED FROM A SELECTED SET OF INDICATOR DIAGRAMS TAKEN IN THE AFTERNOON OF EACH DAY, ABOUT EQUAL TO THE MEAN FOR EACH DAY, CORRECTED FOR ERRORS OF INDI CATOR-SPRINGS AND ALL SOURCES OF INACCURACY.

The quantity of condensing-water was measured approximately, and its temperature, as well as that of the out-flow, kept; and by this method 87.2 per cent of the heat rejected by the engine was accounted for. Had the performance of the boilers equalled 10,000 B. t. u. usefully imparted to the water per pound of coal (a result easily attainable), the consumption of coal per indicated horse-power per hour would have been:

Condensing, 3.3 × .868.

Non-condensing, 4.25 × .837

2.86 pounds

3.56 pounds

The superheating, while favorable to the engines, reduced the apparent efficiency of the boilers.

THE ATLAS-CORLISS ENGINE,

Manufactured by the Atlas Engine Works, Indianapolis, Ind., justly acknowledges its indebtedness to Mr. Corliss, while it presents some peculiar features of considerable interest.

The four valves, situated as in the Corliss engine, are cylindrical segments actuated by wrist-plates; and the steam-valves are tripped by mechanism controlled by the governor, much as in the Corliss engine.

These steam-valves are placed in steel bushings having two portopenings, which give a double length of opening at the beginning of admission, and also provide for repair without disturbing the cylinder itself. Two eccentrics are used, in place of the single eccentric of Mr. Corliss, set on the shaft, the exhaust eccentric nearly in the same angular position as in the Corliss engine, but the steameccentric about 90 degrees behind it, so that the cut-off has a range all the way from 0 to .875 stroke.

These eccentrics are not placed on the engine crank-shaft, but on a special shaft near the cylinder driven by a pair of equal gears from the crank-shaft. The governor is driven by gears from this eccentric shaft, thus avoiding the use of a belt.

The eccentrics are small, and the connections are short and direct. The Porter governor is used.

Flat slides are used, instead of the V-slides of Mr. Corliss; and the cross-head pin is placed in the middle of the length of the wearplates, the wear of which is taken up by adjustable wedges.

The main bearing and the girder frame are cast in one piece. The stroke ranges from 3 to 5 feet; the speed, from 60 to 80 revolutions; and the piston-speed, from 480 to 600 feet per minute. No tests of these engines have come to the knowledge of the writer.

THE PORTER-ALLEN ENGINE,

Manufactured by the Southwark Foundry and Machine Company, Philadelphia, has been completely transformed since its precursor, the "Allen Engine," was exhibited in England, and for a while manufactured by Whitworth & Co.; and considerably improved in matters of detail since it was described in the supplementary volume of Tolhausen's Uhland, in 1882.

Among these improvements, are the adjustable "pressure-plates" back of the steam-valves, which greatly facilitate the keeping of these valves tight, and also serve as very efficient relief-valves for

water in the cylinder, whenever greater pressure exists in the cylinder than in the steam-chest.

The workmanship of these engines, and the facilities of all kinds provided for accurate fitting and systematic inspection of their parts, are, probably, not surpassed in any mechanical establishment. The system of inspection requires the exact determination of the finished. dimensions of every part, and the disposal of the inspected parts according to their degree of conformity to standard dimensions within certain prescribed limits of error.

If within the limit, the piece is accepted: if beyond the limit, it is laid aside for more accurate fitting, if the nature of the defect admits of remedy; if otherwise, it is inexorably rejected.

This engine, by reason of its four-ported steam and exhaust valves, its unbroken valve-connections, and the compensation afforded by its peculiar link for the inequalities of piston-speed near the two extremities of the stroke, by the solidity and excellence of its frame, and by the skilful adaptation of the weight of reciprocating parts to the rotatory velocity, is peculiarly well adapted to high speed.

The normal speeds range from 350 revolutions per minute for 12-inch stroke, to 80 revolutions per minute for 5 feet 6 inch stroke, giving piston-speeds from 700 to 880 feet per minute. Three pairs of these engines, 11.25 inches diameter of cylinder, 16-inch stroke, condensing, 200 to 250 indicated horse-power, have been running several years at the mill of the Willimantic Linen Company, Willimantic, Conn., at 350 revolutions per minute, giving a piston-speed of 933 feet; and cards taken from one of them by the writer show excellent distribution of steam.

The Porter-Allen engine has been extensively used for driving dynamos for electric lighting, but perhaps its most signal triumphs have been won in rolling steel plates and rails.

At the Cambria Iron-works, Johnstown, Penn., a Porter-Allen engine, 40-inch diameter of cylinder and 48-inch stroke, several years ago replaced a throttling, slide-valve engine, running at 82 revolutions per minute with empty rolls, but falling to 65 or 60 at every pass, so that by the time a rail was finished its temperature was very sensibly reduced, and its plasticity greatly lessened.

The Porter-Allen engine, running at the unalterable speed of 90 revolutions per minute, finished a rail with 11 passes instead of 15, without sensible diminution of temperature, and fully doubled the production (2,400 tons per week against 1,200), with greatly reduced

breakage, and a smaller proportion of defective rails. Of course there were other incidental advantages.

The Cambria Iron-works have in use, it is said, 14 of these engines of various sizes. An engine of 40-inch cylinder, 48-inch stroke, coupled directly to a train of rolls at the Otis Steel-works, Cleveland, O., rolls plates 10 feet in width, and of less width 50 feet in length, reducing 3 inch at a pass.

The Troy Steel-works, Troy, N.Y., have three of these engines, now several years in use, driving roll-trains; and there are many others. One of these engines, 12-inch cylinder, 24-inch stroke, supplying power to the Clinton Wire-cloth Mill, Clinton, Worcester County, Mass., now more than ten years in use under the observation of the writer, recently ran six months at about 50 horse-power, as computed from daily indicator-diagrams, with a mean hourly consumption of 3 pounds of coal per horse-power, all the coal burned during the whole period being divided upon the number of hours

run.

-

The coal was a good quality of anthracite, egg-size. The engine is non-condensing, and not steam-jacketed, and during cold weather works against 3 or 4 pounds back-pressure, as exhaust-steam is used for warming the mill.

THE HOADLEY PORTABLE ENGINE.

An engine designed, patented, and for several years manufactured by the writer, and now made by Messrs. George T. McLaughlin & Co., Boston, Mass., should be mentioned as the pioneer in this country of the large class of steam-engines with automatic cut-off controlled by a centrifugal governor placed on the crank-shaft, of which we have already noticed several examples, and shall meet others.

The genesis of this device appears to be as follows: Richard Wilcox of Bristol took out letters-patent of Great Britain, No. 2,574, of 1802, June 21, for several improvements in steam-engines and boilers, among which was a centrifugal regulator attached to the arms of a fly-wheel, complete as an automatic variable cut-off in all respects, except that no provision was made for centripetal action, by springs or otherwise.

The description is imperfect; but the drawing is clear, and the idea of regulating the speed by causing the speed itself to proportion the period of steam-admission to the requirements of the load, is clearly expressed. The device shown a conical cam shifted endwise on the crank-shaft by the governor-is substantially the same