Upon this table we remark, first, that at all the stationary points, except in the alloy of 1 lead to 2 tin, the metal was solid at the stationary point; second, that although the proportion of tin varied from one to six, and even to seven, the stationary point was not changed more than 34° for the first series, and 54° for the second; third, that in the proportion of one of lead to four of tin, a second stationary point appeared at the point at which the metal began to lose its entire fluidity, and was found in the higher parts of the series, rising with the increased proportion of the more fusible metal, with difficulty detected at times, and disappearing by agitation of the alloy; fourth, that the tin and lead of commerce give, for the lower stationary points in the same alloys, quantities nearly the same. A comparison of the upper stationary points appears below.

The variable nature of the results, seems to point out rather the difficulty of detecting the upper point and the effect of accidental circumstances, than that it is affected materially by the impurity of the metals as found in commerce. This upper point rises with the proportion of the less fusible metal. The number of degrees between it and the corresponding point, for the solid state of the metal shows one difficulty to be obviated in the use of the fusible alloys. For example, the first in the table, just given, has 103° between the point at which it begins to lose fluidity, and that at which it is solid; the second has 114°, and the third 293°; these defects, it was hoped, would not have been found in alloys in definite proportions. They indicate that the variety of combinations in definite proportions is not considerable, if even it exceeds a single one; and that when the metals are mixed in definite proportions, the alloys are in fact combinations, or mixtures, of one or more chemical compounds with the metals themselves. If this be the case with alloys in which the proportions are in the ratio of the equivalents or in multiple ratios, it would seem to follow certainly, that in alloys made in proportions not definite, the same fact would appear. That this is so, and that its effects are of importance in practice, will appear subsequently.

The second part of the inquiry relates to the action of fusible plates when in place on the boiler; it supposes proper alloys, fusing at required temperatures, to be composed, and then studies the causes, modifying the action of them when placed on the boiler. In the first apparatus for the use of these plates, the attempt was made to introduce them within the boiler, but the difficulty of replacing a plate which had fused by another plate, led to the abandonment of this apparatus. The opening made in the boiler, when the plate was withdrawn, was so great that the contents of the boiler were violently discharged through it, before the operation of replacing the plate could be effected. This observation has a bearing upon the plans for making large openings in boilers of full size, to avoid explosions.

The apparatus finally used was a sliding plate, moving in a groove upon the upper side of the boiler, as shown in Plates 1 and 3, where s represents the slide moved by the lever r; in the middle of the slide was an aperture slightly conical, for receiving the fusible plate, this aperture was eight-tenths of an inch in diameter. By means of the lever, the plate could be brought over an opening in the top of the boiler, or the solid part of the slide might be made to cover the same opening. The fusible plate was covered by a disk of brass, the edge of which projected over the plate, and rested upon the slide. There

T

of an inch in diameter. To retain the slide in its place, when pressed from below, and to retain the fusible plate when in a similar situation, the forked stem, Lpressed in the former position by one leg, upon the slide, 8, in the other by the other leg upon the disk covering the fusible plate; the upper end of the stem entered a cavity in an adjusting screw, t, passing through the gallows, u; by this means allowance could be conveniently made for expansion. The lever for moving the slide rested, when the aperture in this latter for receiving the fusible plate coincided with the opening in the boiler, against an upright, projecting from the top of the boiler, and serving as a stop. By the use of this apparatus, the plates were applied very readily, were removed when fused, and the opening into the boiler closed with so much despatch as to prevent the foaming within from taking effect. The disk which covered the fusible plates, prevented in part the loss of heat from the upper surface.

The plates which were first cast, were intended for low pressures as most convenient for experiment, they were fifteen-hundredths of an inch in thickness. The observations made upon the manner in which they acted when in place upon the boiler, led to the question of the effect of varying the thickness upon their use. When a plate of sufficient thickness to prevent its giving way to pressure, verges towards its point of fusion; the top part, which is in contact with the metal disk, melis, and flows over the holes in the disk; sometimes it accumulates until the liquid rolls off the plate. The temperature rising, a small pellet of the more perfectly fused parts is thrown out by the steam, the flow of which is instantly checked; this is repeated frequently, until a breach through the plate is made, and the uninterrupted flow of steam takes place. If the plate be removed at once, a very small hole appears, which would gradually have been widened by the action of the escaping steam, probably before the entire fusion of the plate. The under surface of the plate appears oxidized and the fusion to have taken place at the top: the plate has contracted in its dimensions, and the periphery of the upper surface has lost its circular figure, which is tolerably well preserved by the lower surface. To give some idea of the extent to which a plate such as just supposed, may lose its substance before giving way, two measurements are subjoined. Before fusion, the diameter of the upper surface of the plates was eighty-four hundredths of an inch; the lower diameter eighty-two hundredths of an inch; the thickness of the plates, fifteen hundredths. After the plate had given way, the diameter of the hexagonal figure into which both the surfaces had passed, was about seventy-nine hundredths for the first, seventy-four hundredths for the second; the diameter of the lower surface, which was still nearly circular, was, for the first, seventy-six hundredths; for the second, sixty-nine hundredths; the thickness of the first was about twelve-hundredths, of the second, one-tenth of an inch; the thickness not being uniform in all parts. The first plate had lost, therefore, nearly three-tenths, and the second half of its substance, without allowing the passage of steam.

The observed oxidation of the lower side of the plate, led to the supposition that it might retard the fusion of the plate, but no confirmation of this view was given by comparative experiments, with plates of which the lower surface brightened, and of others in which the same surface was highly oxidated, the thickness in each case being the same.

was

In the course of the experiments on the effect of oxidation, the plates were much reduced in thickness by filing away the under surface, and the fusion of the thinner plates took place at points so much lower than those at which the thicker plates of the same alloy gave way, so as to require an examination of

the cause.

Before proceeding with further detail, it may be well to state the general method of experimenting upon the plates. The stationary point of an alloy

were cast from it; and one of these being placed in the opening in the slide of the apparatus already described, was covered with the pierced disk, and the slide moved so as to bring the plate over the opening in the boiler. The steam was now raised, the temperature being noted from time to time, until the plate gave way; steam was then let off to keep the temperature from rising; the plate, which had just fused, removed, and one of an alloy, fusing at a higher tempe. rature, substituted. The steam was again retained, and allowed to rise in temperature, the new plate pushed to its place, and the operation renewed. This course was continued until the alloy, fusing at the highest point of those prepared, had been used, or until the limit of the elasticity of steam, which could be produced in the actual condition of the boiler and state of the fire, was attained; steam was then let off, water thrown into the boiler, and a new series commenced. The tables which will be given, required many days of trial, and of close attention.

To try the effect of thickness on the fusion of the plates, three different thicknesses were cast of each of the alloys used; the first, or thickest, was fifteen-hundredths of an inch thick; the second, eight-hundredths; and the third, four-hundredths. There were five different alloys of tin, lead, and bismuth, composed; the stationary points of which, and the gave way in the boiler, appear below.

points at which they

The plates of experiments 1, 2, and 3, were exposed to pressures tending to render them of less than one atmosphere; 1 and 3, the two extremes of thickness, show a great uniformity in the point at which they give way, and render it probable that some flaw, in casting the plate number two, caused its fusion at a lower point than that of either of the others: we to that at

these low pressures the fusing point in the boiler coincided, very nearly, with the point at which the alloy was a soft solid in the crucible. In this case the thinnest plates, when properly cast, were probably thick enough to withstand the small pressure to which they were exposed, and therefore did not give way at lower temperatures than the thickest, each attaining the temperature at which they were soft solids.

inex

The next series, Nos. 4, 5, and 6, with a less fusible alloy, show, first, that the thinnest plate was too feeble to resist the pressure of steam, and gave way before the metal lost its solidity; second, that the plate, eight-hundredths of an inch, was probably defective, as it gave way at a lower temperature than the thinner plates of No. 4. No. 6 presents a curious fact; the point of yielding of the plate, given by four experiments, is actually above the point at which appear the alloy from which it was composed became liquid: this would plicable to one who had not attentively observed the mode of fusion of these thick plates, and would lead to the suspicion of error. The observation of the point at which the alloy became liquid, was however, deduced from three trials; and four experiments, with the plates in place, are given, the extremes differing but three degrees. The explanation is to be found in the mode of fusion already spoken of; the more liquid parts of the alloy are forced out, the less fusible remain, and if strong enough to resist the pressure, the process goes on; this takes place, unequally, in different alloys. The importance of attending to such indications is obvious.

In the next series, the first thickness seems to have been decidedly too weak, and the second to have been hardly sufficient, while the third exhibits a point of fusion when the metal was in a softened state.

In the remaining experiments, both thicknesses were too inconsiderable to sustain the pressure, as appears by comparing the points at which the plates gave way, with the stationary points. Something of this kind seems to have been deduced from practice in the use of these plates in France, for the last royal ordinance, in relation to the means of safety to steam-boilers, prescribes for the plates a thickness of not less than nine-sixteenths of an inch, making of them fusible plugs rather than fusible plates.

Experiments were subsequently made on plates of greater thickness, the use of which led to an interesting termination to this series of experiments. Before, however, stating the results thus obtained, some further experiments with the plates just considered, will be described. These inquiries were directed to the effect which would be produced by the mode of casting the plates, upon their fusibility; it being not improbable that rapid cooling might so modify the physical properties of the alloy, as to change the fusing point from that of the same alloy when slowly congealed.

As low pressures afforded the most easy means of determining this point, plates were cast from the alloys of series No. 1, No. 2, and No. 3, and tried, in place, upon the boiler. Some of the plates were cast from metal at a high temperature, and the mould as cold as the perfect casting permitted; others, of the same alloy, in a heated mould allowed to cool slowly; and others from metal heated to a temperature as little above the point of fusion as possible. In the case of the higher temperatures, care was taken not to raise the heat so far as to oxidate rapidly either of the constituents of the alloys, thus changing the fusing point. A comparison of the results obtained, showed no greater differences than those which have been seen to occur between plates similarly that the mode cast, and from the same alloy; and the conclusion derived was, of casting has no effect on the fusing point, which is appreciable in this mode of applying them. The French instructions expressly recommend to those

10*

who make or use boilers, to obtain plates in preference to the fusible metal in ingots; on the ground that it will be found difficult to procure plates of the same fusibility with the ingot, from that form of the alloy. This remark led to the experiments just referred to, and the explanation of it which they give, re fers to the undue heating of the alloy in the casting of the plates, since they show that the particular mode of casting produces no difference worthy to be regarded in a practical point of view.

Plates were now cast quite thick, viz.-one-fourth of an inch, of the alloy of eight parts of bismuth, eight of tin, and seven of lead; this alloy being in tended to give way at a temperature corresponding to about one atmosphere of bursting pressure. The alloy was completely liquid at 275° Fah. and solid at 254° F.,* when examined in the crucible. The heat was raised as slowly as possible, so as to permit the full effect of the temperature indicated by the thermometer; the observations recorded are as follows:

Plate 4th of an inch. Metal stands fused in the holes of the brass disk covering the fusible plate.

Steam issues in a very small stream through chinks between the fused metal and an unmelted part within. Steam issues as before; no clear passage through the plate. Steam kept for a long time at this temperature. Six minutes elapsed in raising temperature 44 degrees. Plate gave way, affording a free passage to steam. A second plate of the same composition and thickness, put in place; fluid metal stands in the holes of the cap. Metal which has oozed out remains in a fluid state upon the sliding plate of the apparatus.

Plate gives way, torn in a thin part.

The thermometer on top of the boiler dipped into the mercury in a small cistern, made by inclosing a space on the top of the boiler, with clay; so that the top of the boiler formed the bottom of the cistern. The first plate having failed to give way until the temperature within the boiler was twenty-four degrees above that at which the alloy, of which it had originally been composed, had been fluid, was examined with great care. The plate had decidedly given way to pressure, and not by fusion; it had lost its metallic lustre at the side where it was torn; yielded readily to the nail, which scraped off small particles. A piece of the plate being cut off and laid upon the top of the boiler, remained solid, though the portion which had oozed out, was perfectly fluid, near the same spot. The same remarks apply, generally, to the second plate. They afford a solution of the perplexing circumstances, which had occurred throughout these experiments, and which had led to so many trials to discover their cause. The portions of the metal which oozed out from these two plates had their fusing points taken, by gradually raising their temperature in a bath of oil, while the alloy rested on a small shelf of copper, wholly immersed in the oil. The first portions of fluid metal which had oozed from both the first and second plates, melted at betwen 221 and 223 degrees, Fah., being solid at the lower, and perfectly fluid at the higher, of these two temperatures. The

* This alloy had no stationary point in passing from the liquid to the solid state; but some internal change in the solid at about 206°, produced a rise and stationary point at 208°.