Tawas attached, passed through the half-collars 51andio, which served as guides. On the - point terminating the upper end of the stem, s pressed the lever lm, for receiving the weights. -x The fulcrum of the lever was a knife-edge, monesting in a curved opening in the upright Janey which was screwed into the boiler, (See

also Ky plate 1.) The apparatus was graduated by trial, in order that the friction iqmight be duly estimated. The valve, lever, -1&os, being in place, a delicate balance-beam eff graduated to inches and tenths, on each -side of the edge g, which was its axis, was so placed, fig. B, that one of the divisions was vertically above a small hole made in the to this division a cord hh passrough the hole, was attached. A scaleing pan was hung to a division k, of the opposite arm of the balance, and the string and pan made to balance by weights on the arm fh. The string being now stretched by a weight rather greater than the probable amount

would have to bear in the graduation

of

valve, the beam was made horizontal. Weights were then placed in the scale-pan, until the weight of the lever, the stem and disk, and the friction, being overcome, the valve rose from its seat. This rise was indicated at the end 7 of the valve-lever, by an upright placed there for the purpose. Next a convenient weight was placed in the pan, and the point ascertained on the arm of the

lever at which a small weight, assumed, balanced it. Thus several points were determined. In graduating the first safety-valve gh, was 44 inches and gk 11 inches, so that the weight raised at was two and a half that applied at k. The weight of the lever, &., balanced 24 oz. troy at k, or 24 × 2=69 oz. 5lbs. troy at h. This weight rested on an effective surface of 515 inches in diameter, or 232 square inches in area. It was equivalent, therefore, to 24 04lbs. troy to the square inch, or to 1978lbs. avoirdupoise; that is, to a pressure of 135 atmospheres of 14:68lbs., corresponding to 30 inches of merculy.

A weight of 6 oz. being at d, near the end of the lever I m. was balanced by 56 oz. at k on the arm of the balance. To determine. from this the ratio of dm to nm, we liave 56 × 2 -60 80 13, which coincided 6

nearly with the ratio given by measurement, being a little in excess.

At the mark d, 6 oz., together with the weight of the lever, &e, and friction, produced a pressure of 314 atmospheres and the weight used in the experiments, 9:369 oz. troy, produced a pressure of 3.27 atmospheres, exclusive of atmospheric pressure.

At the mark c, called in the experiments. the third mark, 9 oz. balanced 56 at k, and

212

the weight, 9-369 oz., together with that of the lever, &c, produced a pressure of 2 63 atmospheres.

At the mark b, the second mark, 12 oz. were required to balance 56 at k; the pressure produced by 9.369 oz. &c, was, therefore, 2-31 atmospheres. Ata, the first mark, 15 oz. balanced 56 at k, and the pressure from 9.369 oz. &c., was 2 12 atmospheres.

The experiments with these valves will show why the Committee are so particular in giving the data upon which the graduation was founded. The details in relation to the second valve were as follows:

The string from the valve-stem was attached to the mark 5, on the balance-beam, and the scale to the mark 10 on th opposite arm; when 29 oz. troy at 10, balanced the weight of the lever, &c. at 5. This weight was, therefore, 59 oz. troy, or 4.04lbs. avoirdupoise; pressure upon a surface of 515 inches in diameter, or 208 square inches in area, it was equivalent to 19.44lbs. to the square inch, or to 1.32 atmospheres of 14 68 lbs. At the last mark, near the end of the lever, a weight of 23 oz. balanced 49 oz. at the mark 10 on the beam of the balance; the weight 3.76 oz. which was used in the experiments, together with the weight of the lever, &c., produced, therefore, a pressure of 2.52 atmospheres.

At the mark next to the end, called in the experiments the second mark, the same weight, together with the lever, &c., produced a pressure of 2.10 atmospheres.

Weights of 264 oz. and of 5 28 oz. were also prepared to be attached to the mark near the end, the weight of 3.76 oz. being prepared with hooks to adapt it to this purpose. The smaller weights produced an additional pressure of 84 of an atmosphere, and the larger of 1.68 atmospheres.

On the Performance of the Safety-Valves. The committee, with a view to test the performance of the safety-valve, selected the form calculated, in their opinion, to be most serviceable in practice, a disk valve. One of these valves was attached to the boiler. This was in action in the different experiments, and was tested by comparing with the temperature of the water within the boiler and with the steam gauge, when this latter was in use; the results were rendered more valuable from the fact that one of the valves having been injured in the experiments on highly heated steam, a second one had to be put in its place and to be graduated anew. The pains bestowed in this graduation enabled the committee, not only to compare the pressure at which the valve opened when similarly weighted in different experiments, but to compare the pressure given by calculation from the area of the valve, &c. with the actual

pressure at which the valve allowed a free passage to the steam. Any adhesion of an unusual character could not have escaped detection. The valves were kept in good working order, but no unusual pains were bestowed upon them, and the experiments ranged through more than two years of interval, irregularly continued. They there! fore give quite as severe, if not a more severe, trial, than in a working apparatus. The following tables contain the results of the ex+ periments and the comparisons, necessary to give an entire view of the subject. (See p. 213.)

The first column of the first table gives the observed temperatures of the steam, the pressure of which raised the safety valve, weighted as expressed in the third column; the pressures corresponding to these temperatures are given in the second column. An abstract of the first table given in the second; it contains the mean, the highest and lowest observed presures, and the range of pressure. The pressure at which the valve should have opened, according to calculation, and the ratio of the calculated pressure to the mean is also furnished by the last two columns of the same table.

The performance of this valve is shown by the table to have been so far satisfactory that the greatest range of pressure between the highest and lowest observations, at any pɔsition of the weight, is .6 of an atmosphere. While, therefore, no accurate investigation. of the elasticity of steam could result from the use of such a valve, it is found to answer, so far as these results can show, the requirements of practice. In regard to the ratios of the average observed pressures to the calculated pressures, they range from .98 to 1.11, and at a mean the ratio is 1 to 1.034, a ratio which would produce, at higher pressures, marked differences between the calculated and observed pressures.

In a paper by M. Garnier, in the Annales des Mines,* vol. viii. the results of his comparisons of safety valves having a considerable projection over their valve seat, with others having a small projection, are given, and are very worthy of attention. He found that a valve, the disk of which projected upon the valve seat beyond the opening of the seat 4 of an inch, opened at a pressure within the boiler, of only two-thirds of that shown by a mercury gauge, and that the distance to which the weight upon the lever of the valve required to be placed from the fulcrum, in another case, for a pressure of four atmospheres, was that corresponding to five atmospheres, if the pressure of the air on the upper part of the valve were estimated. With a disk projecting but .02 of an inch the ratio

• Second Series.

of the calculated pressure to be observed wes 1.06 to 1.

boThe experiments above quoted, with a disk projecting.1 of an inch, give a less ratio than is indicated by M. Garnier, and show besides that the phenomena are not constant, the individual experiments varying more from each other than the mean varics from the calculated result. The fact developed by M. Garnier is attributed by him to a want of perfect contact between the disk and its seat, and the variations observed by the committee go to strengthen this opinion. Different positions of the disk, different states of its surface, and the interposition of small particles of dirt will satisfactorily account for the want

This,

of uniformity in the pressure at which the valve opened when all circumstances were apparently the same. Further evidence on this head will be gathered from the results furnished by the second safety valve. it will be recollected, was of the same form and dimensions with the first, and similarly graduated. In the following table the observed temperatures and pressures are given, at which the valve, circumstanced as stated in the column headed remarks, opened. The mean pressure, the highest and lowest pressures, and range, are given in the next four columns. The calculated pressure, and the ratio of the mean pressure to the calculated, are in the next two columns.

Corresponding pressure.

Fah.

The weights used in the experiments were verified after their conclusion, and found to be ancurate. The ratios of the calculated pressure to the observed increase, and after. wards decrease, showing that the results are not attributable to inaccuracy in the unit of weight which was successively applied. They are not explicable by a defect in measuring the valve seat, as is shown by the nearer coincidences at the lower pressures, when the ratios are much nearer to unity than in the higher.

This valve was fitted, by grinding, twice during the experiments, but that the contact with the seat was not perfect, is shown by the number of times at which leaking is noticed before the rise of the valve; these leaks at the high pressures became very sensible, and increased rapidly, from the instant of their first appearing, to that at which the valve opened.

The mean ratio of the calculated to the observed pressure is 1.10 to 1, but taking for comparison with the first valve, the same range of pressures as was calculated for the first, the ratio is as 104 to 1, differing but little from the ratio obtained for the first valve.

This defect in the safety-valve, which affects its use as a means of judging of pressure in proving boilers, is increased as the pressure increases, and may cause an inadequate proof by the forcing pump, or the hydraulic press. No adhesion of an undue kind is shown, by these experiments, to have taken place with the second valve, the difference of pressure of opening and of undue leakage. under similar circumstances, being moderate. The experiments, it should be remembered, apply only to disk valves, and not to those of a conical form.

XI. To ascertain, by direct experiment, the effect of deposits in boilers.

The committee have been fortunate on this head in obtaining the account of the results, on a large scale furnished by the deposits in the boilers of boats on our western waters; to these they refer, as showing,

1. That deposits, consisting of sedimentary matter, carbonate of lime, and other salts, collect in particular parts of boilers, and, preventing the communication of heat to the water, are baked hard, becoming "as hard as brick," when the water is low.*

2. That these collections of mud, &c. may, by causing the undue heating of the bottom of a boiler, produce exfoliations of oxide, which gradually reduce the thickness of the metal;

Explosion of boiler of steam-boat Caledonia, W. Littlefield, p. 310, vol. viii, Jur. Frank. In-t. Anonymous, v. 31), ib. Matthew Robison, p. 311, ib. E. W. Benton, p. 314, ib. Also L. Hebert, p. 379. ib. Thos. W. Bak well, p. 386, ib. Thos. J. Halderman, p. 28, vol. ix,

or they may allow the temperature of the metal to be raised so high, that it swells out by the ordinary pressure of the steam, and finally bursts. Thus leading, gradually or suddenly, to the weakening the boiler, and to the discharge of its contents.

The committee have also examined cases of similar deposits in iron boilers, using spring (hard) water. They consisted chiefly of the carbonate of lime and iron, mixed with oxide of iron, containing. besides, the earthy salts from the water. Unless removed at short intervals they form in the ordinary use of the boilers, and without undue heating, exceedingly hard crusts upon the bottom of the boilers, requiring the aid of the chisel for their removal. Retarding the passage of heat into the boiler, they lead to a great waste of fuel; and exposing the bottom to be unduly heated, they destroy the boiler gradually, by wear, if not suddenly, by explosion.

The nature of these deposits, and the ra pidity of deposition, varies, of course, with the kind of water which is used.

XII. On the elastic force of steam at norking pressures. ›

The

The committee, determined to put the apparatus which was necessary for other experiments, to the best use possible, in de termining the elastic force of steam at different temperatures; and accordingly great pains were bestowed upon the graduation of the gauge, the regulation of the temperature of its parts, &c., the comparison of the thermometers, the maintenance of the scales at about the same temperature, &c‡ small size of the boiler, and the various openings required to be made in it for the experiments which were the immediate ob jects of the committee, were unfavourable to the attainment of considerable pressures, but the discrepancies, even at working pressures, of the different tables of the elastic force of steam, made it important to push those trials as far as could be done without material changes. They succeeded without much difficulty in reaching ten atmospheres, which is but one atmosphere less than the reputed working pressure of our high-pressure engines, and as the experiments on the safetyvalves have rendered probable, is very near the true working pressure.

A series of results obtained in the trials of the fusible plates, is given below in the tabu

lar form.

The table contains the temperature ob served by the thermometer in the water, corrected for the error of the graduation; the

Co'. S. H. Long, pp. 244-5, vol. viii. Prof Johnson, at monthly meeting of Institute.

See description of apparatus, &c., M. M. pp. 83-86, &c.