At the close of these experiments, the metal was, in many places, but little short of a red heat, visible in day-light.

The precise state of things in a boiler, of which parts are unduly heated, was represented in these experiments; there was the surcharged steam, and heated metal ready to give up its heat to replace what might be absorbed in the conversion of the water injected into steam. This latter circumstance renders the case different from that contemplated in the deductions of theory which have been brought to bear upon the question. The greater or less intensity of the heat afforded by the top and sides of the boiler would necessarily modify the effects observed, by the injection of any given quantity of water; this is observable in the numbers given in the table, where although the greater quantity of water injected does not fail in two consecutive experiments to show a greater depression of the gauge, yet in distant experiments the same is not the case. We see that in no case was an increase of elasticity produced by injecting water into hot and unsaturated steam, but the reverse, and in general that the greater the quantity of water thus introduced, the more considerable was the diminution in the elasticity of the steam. The quantity of water injected was from 3.5 to 24.3 cubic inches. The immediate rise of the gauge after each experiment, shows how rapidly heat was supplied by the sides of the boiler to the steam within.

That the steam was highly surcharged with heat, appears by comparing the pressures corresponding to th etemperatures with those

Rose again immediately to 21.90. Fell nearly 2 inch. Note. 5330 is, according to the formula of Arago and Dulong, the temperature of saturated steam of more than sixty atmospheres.

The

given by Dulong and Arago for saturated steam. For example, the pressure shown by the gauge when the steam was at 506° Fah. was 6.15 atmospheres, while the table just referred to gives for this temperature a pressure of forty-eight atmospheres. temperature was carried to 533° Fah. when the pressure shown by the gauge was 6.82 atmospheres, while saturated steam at that temperature would have had a pressure of more than sixty atmospheres.

In order to ascertain whether the thermometer relied upon to give the temperature of the steam was affected, if at all, in excess or defect by the conducting power of the metal; the temperature of the boiler just beyond the tubes was taken, as nearly as was practicable, by a thermometer R, fig. 1, dipping into a clay receptacle, upon the top of the boiler. This thermometer did not rise above 4050 Fah.; its distance from this source of heat was ten inches, and that of the iron tube inclosing the thermometer, six and a half inches. Supposing the temperature stationary on top, the temperature of the metal of the top of the boiler near the tube of the thermometer would have been 4790,* show

* If we suppose the heat of a small bar of metal, cut from the top of the boiler, to have been derived by the conducting power of the metal alone, the heating effect of the steam within being neglected, and further, that the temperatures of the bar had become constant, then the ratio of the excess of the temperature (y) of auy point at a distance (a) above the temperature of the air, to that (y') of any point at a distance (x'), is given by the proportion,

y' y: log. x: log. x'.

In the case before us, y 405-80 325°, x=

102

ing that it tended to carry off heat from the thermometer, which, if at all affected by the metal above it, showed too low a temperature for the steam in contact with it. The tem-. perature of the source of heat would have been from these data, 582° at the extreme end of the part covered with fuel, which was of course at a lower temperature than the middle portion

On examining the apparatus after the conclusion of the last day's experiments, it was found that some of the putty used in tightening the lower joint of the thermometer in the water had been softened by the heat, and had flowed into the tube, thus affording a direct communication between the steam and the bulb of this thermometer: this circumstance accounts for the instrument being affected in this day's experiments and not in the preceding ones.

IV. The next query may be thus stated: when steam, surcharged with heat, is produced within a boiler by the contact with heated metal, does this steam remain surcharged, or does it take up water from contact with that in the boiler, and become saturated steam? If the latter supposition be correct, at what pressure and temperature with regard to the temperature of the surcharged steam, and to that of the water on which it rests?

The answer to this question is given by the experiments just detailed; and as they established the negative in relation to the surcharged steam becoming saturated, there was no necessity for a repetition of the experiments to ascertain the precise temperature of the water in the boiler. When fire was applied to the top of the boiler, the water within was at 3180 Fah.; a moderate fire was kept up below, and one so nearly uniform, that great variations from that temperature could not have taken place, and which the results satisfactorily show did not occur. If we assume that during the experiments the temperature was 30840 Fah., a remarkable correspondence will be found in the observed pressures, and in those calculated on the supposition that this steam was expanded by heat, as a gas would have been, without any addition of water. The table below gives the temperatures of the surcharged steam observed at different times during the course of the experiments; the pressure shown by the gauge at that temperature; the pressure which would have been produced by heating steam at 308 to the temperatures given in the first column by the mere effect of expansion; and the pressures of saturated steam at the different temperatures.

acquiring from it the water necessary to convert it into saturated steam, but retaining its surcharged state. There is nothing to warrant the belief that any of the surcharged steam was condensed by the water.

V.—Inquiry in relation to Plates of
Fusible Alloys.

It is well known that one of the most scientific nations of Europe relies, particularly, as a means of safety for steam boilers, on the use of plates of fusible metal. The plates are alloys of tin and lead, or of these two metals with bismuth, the proportions of the component metals regulating the point at which they fuse. In France these alloys are prepared at the royal mint, where plates made from them, or ingots of the alloys, may be purchased for use. The examinations which must have been made to determine the proportions of the metals necessary to produce an alloy fusing at a given temperature, and the circumstances of fusion, have not, as far as the committee know, been made public. A table of the fusing points of different alloys of tin, lead, and bismuth, &c., was drawn up by Parke, from experiment, and is contained in his chemical essays, vol. ii. p. 615. This table was made the basis of the investigations* undertaken by the committee, but they soon found it convenient to depart, more or less, entirely from it.

The method employed by Parke for deter mining the fusing point of a metal, or rather the solidifying point of the melted metal, was ingenious. On melting a metal, and allowing it to slowly cool to the point of congelation, and observing a thermometer plunged in it, a rise of temperature, and then a stationary point, is observed; this is a point where a change is going on, by which the heat given out in the change is equal to that of which the metal is robbed by the surrounding medium. This point usually coincides with the passage of the metal to the solid state, from what may be either the liquid state, or a semi-fluid state, similar in aggregation to sand; sometimes the alloy is solid throughout, before the stationary point arrives; and sometimes there is more than one such point.

The stationary point is not that at which the alloy, when used as a fusible plate for a boiler, gives way; the plate being covered by a perforated brass disk, to prevent its being pressed outwards before fusion, and so reduced in thickness as to burst, the metal is not forced out through these openings until perfectly fluid; if any part of the metal be

At the time these experiments were made, the paper of Rudberg Ann. de Chimfe. et de Phys., vol. 48, had not appeared.

comes fluid before the rest, and gives way, the rest being in the sandy state, just spoken of, the particles seem to act like those of sand in a similar case, and to oppose an effective resistance to the pressure of the steam; these facts will be further developed in the examination of the application of these plates.

The stationary points, when taken with due reference to the state of the metal at the time, afford so many approximate marks by which to compare together the fusibilities of the plates, and to ascertain whether they bear a due relation to each other, when fused, in place upon the boiler; and to study the alloys themselves. In composing alloys of the metals, before referred to, the tin was fused first at as low a temperature as possible, then the bismuth and lead added, the heat being kept up; these metals were readily taken up by the liquid tin, and were thus little exposed to oxidation: the surface of the alloy was always protected by a stratum of oil. The metal was constantly stirred to promote the uniform diffusion of the different metals throughout each other.

The alloy being liquid, a thermometer, of which the errors had been carefully ascertained, was plunged into it, and the fall noted until it reached the lowest point; the rise to the stationary point followed, and at this the thermometer usually remained for such a length of time, often some minutes, as to render any error of observation unnecessary. Some of the alloys have no stationary point, properly so called, and the beats of a second's pendulum were used to determine the rate of their loss of heat. When the quantities of metal used were inconsiderable, the heat was observed to be carried off so rapidly as to lower, or entirely to destroy, the stationary point: to avoid this, the crucible containing the alloy was placed in a second one, the edges of the former resting on the middle of the sides of the latter. The quantity of metal used was never less than between five and six ounces, troy.

The stationary point being at the passage of the liquid metal to the solid state, or at some interior change of the solid itself, the thermometer was entangled in the metal; and in moving the alloy, to re-melt it, the instrument was endangered.* This was remedied by the use of a small cylinder of very thin sheet-iron, containing mercury. This cylinder was placed in the alloy, and filled up to the surface of the metal with mercury, and the thermometer could now be readily placed

Although the instrument was frequently used in determining the stationary points, no permanent changes in the indications of the instrument, such as was noticed by M. Rudberg, took place.

104

and removed. Care was taken to ascertain that the stationary point, given in the cylinder, was the same with that shown by the naked thermometer. As some of the alloys expanded considerably on congealing, it was supposed that the cylinder might prevent error from the compression of the bulb of the thermometer, but no such compression in the instrument used was detected by frequent trials.

As the alloys were intended for ordinary use, it was deemed advisable to ascertain how far the impurities of the metals, as they

occur in commerce, would cause a variation in the fusing point. Tin has a very uniform purity in commerce, the grain or stream tin being always accessible. The bismuth of commerce being obtained principally from the native bismuth, is probably not very variable. The lead contains variable quantities of silver, copper, and iron. The first experiments were made on the fusing point, on various specimens of common tin: this tin showed, by re-agents, a trace of iron and of copper, as impurities. The fusing point of grain tin is 442° Fah.

An attempt was next made to ascertain what effect the impurities shown by the fusing point of lead would have upon the fusing points of alloys, into which it entered. Alloys, in atomic proportions, were selected, as much was expected from them in the way of avoiding the slow passage from the liquid to the solid state, which was observed to be the property of certain mixtures of the metals. Alloys of tin and lead were therefore made in atomic proportions; first, of grain tin and the lead already spoken of, from the Paris mint; the second, of block tin and common lead. The tin was employed in multiple proportions, as being the more fusible metal, it would probably enter more largely than the other, into the composition of fusible plates for steam-boilers. equivalent of lead is 104; of tin, 58; the first alloy was composed of the two metals, united in this proportion, the total weight of the components being about ten ounces, troy; a new equivalent of tin was next added, and

Equivalents of

The

Mean,

604

so on, through the series: the results are given in the following table.

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 3° for the first series, and 51° 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 on next page.