Part 1 — APPENDIX.

Since the publication of the second part of the first volume, (1810) some important essays on the subject of heat have appeared, which have a direct bearing upon some points of the doctrine on that subject inculcated in the said volume. It may be proper to state the results, with such remarks and reflections as have occurred in the consideration of them.

In the Annales de Chimie for January 1813, also in the Annals of Philosophy, vol. 2, we find a Memoir on the specific heat of different gases, by M. M. De la Roche and Berard. This exhibits a most laborious and refined series of experiments on this most difficult subject. Great merit seems to be due to them, both for invention and execution.

It is unnecessary to describe the particulars of the apparatus and the mode of conducting the experiments, as a description may be found as above referred. It is sufficient to observe that the calorimeter used was a copper cylinder of 3 inches diameter and 6 in length, filled with water, and having a serpentine tube 5 feet in length, running through the interior and opening at both ends on the outside of the vessels. By means of this tube a regular current of any gas of a given temperature (212°) might be passed through the vessel so as to part with its excess of temperature to the water. The quantity of water and the capacity of the vessel for heat were previously determined; and the quantity of heated gas passed through the calorimeter was determinable at any time, as well as the temperature of the water, from the judicious arrangements.

It is easy to see that when an apparatus of this kind is at work, the gas will impart heat, more or less according to its capacity, to the water; and that the temperature of the calorimeter will gradually ascend till it arrives at a maximum; that is, till the refrigerating effect of the surrounding atmosphere upon the calorimeter is equal to the heating effect of the current of gas.

The following Table exhibits the results of their experiments.

Specific Heat

Of the same bulk. Of the same weight.

Air 1.0000 1.0000

Hydrogen 0.9033 12.3401

Carbonic Acid 1.2583 0.8280

Oxygen 0.9765 0.8848

Azote 1.0000 1.0318

Nitrous Oxide 1.3503 0.8878

Olefiant Gas 1.5530 1.5763

Carbonic Oxide 1.0340 1.0805

Aqueous Vapour 1.9600 3.1360[24]

They found the specific heats of equal volumes of air of the pressures 29.2 and 41.7 inches of mercury to be nearly as 1 ∶ 1.2396, differing from the ratio of the pressures or densities, which is 1: 1.358.

The above table of the specific heat of the permanent gases (excluding aqueous vapour) was corroborated by the results of another series of experiments in which the principle was varied a little: namely, to find how many cubic inches of each gas at a given temperature were required to raise the temperature of the calorimeter a given number of degrees, and inferring the capacities for heat to be inversely as the quantities of gas employed. The differences in the results were from 1 to 10 per cent., which may be considered small, in experiments of such delicacy.

The ratios of the specific heats of several gases being found, it was highly expedient to find the ratio of the specific heat of water, and that of some one gas, as common air. This was effected by passing a small current of hot water through the calorimeter, and comparing the effect of this current with that of the larger one of air, the requisite care being taken to ascertain the quantity of water passing in a given time and its temperature at the ingress. The result of this experiment was that the specific heat of water is to that of common air as 1 ∶ .25 nearly. By two other experiments, varied from the above, results not much differing were obtained, so that the average of the three gave, water to air, as 1 ∶ .2669.

Reducing the specific heats of the gases to the standard of water as unity, we have the following Table of the specific heats of equal weights of the respective bodies:

Water 1.0000

Air 0.2669

Hydrogen 3.2936

Carbonic Acid 0.2210

Oxygen 0.2361

Azote 0.2754

Nitrous Oxide 0.2369

Olefiant Gas 0.4207

Carbonic Oxide 0.2884

Aqueous Vapour 0.8474

Before we animadvert upon these results, it will be expedient to give an abstract of the not less interesting experiments of Messrs. Dulong and Petit, on heat, as given in the Annales de Chimie and de Physique, vol. 7 and 10.

These gentlemen begin by an investigation of the expansion of air by heat. The absolute expansion of air from freezing of water to boiling had been previously determined by Gay Lussac and myself to be from 8 to 11 nearly: they however extended the enquiry above and below these points of temperature, namely to those of freezing and boiling mercury. From the temperature of freezing mercury or thereabouts, to that of boiling water, they find the expansion of air to keep pace with that of mercury, as indicated by the common thermometer; but from the boiling point of water to that of mercury, the latter expands somewhat more in a proportion gradually increasing: as by the following Table.

TABLE I.

Temperature by Mercurial Thermometer. Corresponding volume of a given mass of air. Temperature by an air Thermometer, corrected for expansion of glass.

Fahrenheit. Centigrade. Centigrade.

-33° -36° 0.8650 -36.8

32 0 1.0000 0

212 100 1.3750 100

302 150 1.5576 148.70

392 200 1.7389 197.05

482 250 1.9189 245.05

572 300 2.0976 292.70

680 M. boil 360 2.3125 350.00

The absolute dilatation of mercury claims their attention. They quote nine authorities for the expansion from freezing to boiling water temperatures; the extremes of these nine are, Casbois ¹/₆₇ of original volume, and mine ¹/₅₀ of the same. They determine it to be ¹/₅₅.₅. By doubling and tripling the elevation of the temperature, they made observations from which are deduced the results of the following Table. The dilatations are for each degree of the thermometer centigrade, to which I have added the corresponding ones for Fahrenheit’s.

TABLE II.

Temperature by an air Thermometer. Mean absolute dilatations of mercury. Temperatures indicated by dilatation of mercury, supposed uniform.[25]

Fahr. Cent. Fahr. Cent. Fahr. Cent.

32° 0° 0 0 32° 0°

212 100 ¹/₉₉₉₀ ¹/₅₅₅₀ 212 100

392 200 ¹/₉₉₄₅ ¹/₅₅₂₅ 400.3 204.61

572 300 ¹/₉₅₄₀ ¹/₅₃₀₀ 597.5 314.15

By a series of observations on the apparent dilatation of mercury in glass vessels, compared with the results in the above tables, they deduce the absolute dilatation of glass for each degree of the thermometer, and the temperature that would be indicated by supposing the uniform expansion of a glass rod adopted as the measure of temperature as under:

TABLE III.

Temperature by an air Thermom. Mean apparent dilatation of mercury in glass. Absolute dilatation of glass in volume. Temperature by a Thermometer made of glass.

Fahr. Cent. Fahr. Cent. Fahr. Cent. Fahr. Cent.

212° 100° ¹/₁₁₆₆₄ ¹/₆₄₃₀ ¹/₆₉₆₆₀ ¹/₃₈₇₀₀ 212 100

392 200 ¹/₁₁₄₃₀ ¹/₆₃₇₈ ¹/₆₅₃₄₀ ¹/₃₆₃₀₀ 415.8 213.2

572 300 ¹/₁₁₃₇₂ ¹/₆₃₁₈ ¹/₅₉₂₂₀ ¹/₃₂₀₀₀ 667.2 352.9

The absolute dilatations of iron, copper, and platina were investigated with great address, from 0° to 100° and from 0° to 300° centigrade; and were found as per Table below, for each degree of the centigrade thermometer.

TABLE IV.

(A) = Temp. by the air Therm.

(B) = Mean dilatation of iron, in volume.

(C) = Temp. by iron rod Therm.

(D) = Mean dilatation of copper in volume.

(E) = Temp. by copper rod Therm.

(F) = Mean dilatation of platina in volume.

(G) = Temp. by platina rod Therm.

(A) (B) (C) (D) (E) (F) (G)

Cent.

100° ¹/₂₈₂₀₀ 100° ¹/₁₉₄₀₀ 100° ¹/₃₇₇₀₀ 100°

300 ¹/₂₂₇₀₀ 372.6 ¹/₁₇₇₀₀ 328.8 ¹/₃₆₃₀₀ 311.6

Connected with this subject was another important enquiry, whether the capacities of bodies for heat remain constant at different temperatures, or whether they diminish or increase as the temperature advances. In other words, does a body that requires a certain quantity of heat to raise it from 0° to 100° centigrade, require the same quantity to raise it from 100° to 200°, and from 200 to 300°, &c.; or does it require less or more as we ascend? This enquiry involves that of the measure of temperature. They adopt the uniform expansion of air, or the air thermometer, as the proper measure, and find the capacity of iron,

From 0° to 100° = .1098

0 to 200 = .1150

0 to 300 = .1218

0 to 350 = .1255

the capacity of an equal weight of water being 1.

The following Table exhibits the capacities of seven other bodies according to their results.

TABLE V.

Mean capacity between 0° and 100° Mean capacity between 0° and 300°

Mercury .0330 .0350

Zinc .0927 .1015

Antimony .0507 .0549

Silver .0557 .0611

Copper .0949 .1013

Platina .0335 .0355

Glass .1770 .1900

According to this table the capacities of bodies increase with the temperature in a small degree: and the increase, though it would still exist, would be less, if the common mercurial thermometer were the measure of temperature.

Also supposing that thermometers made of these bodies and graduated by immersion in freezing and boiling water into 100°; if these were all immersed in a fluid in which an air thermometer stood at 300°. Then the relative temperatures of the several thermometers would be as under, if measured by the absolute quantity of heat acquired, namely,

Iron 322.2°

Silver 329.3

Zinc 328.5

Antimony 324.8

Glass 322.1

Copper 320.0

Mercury 318.2

Platina 317.9

From these observations they infer that the law which has been promulgated for the refrigeration of bodies, cannot be strictly true: namely, that bodies part with heat in proportion as their temperature exceeds that of the surrounding medium.

Some animadversions on the general laws relative to the phenomena of heat, announced in my elements of Chemical Philosophy (page 13) then follow, together with a table drawn up to show the discordance between the air thermometer and the mercurial thermometer, both being graduated in the manner I proposed in the said elements. On these points I may have to remark in the sequel.

The first part of the Essay concludes with some remarks to shew why a preference should be given to the air thermometer, or more strictly, the thermometer whether of mercury or any other body, supposed to be graduated so as to correspond with an air thermometer of equal degrees.

The Second Part of the Essay is on

The Laws of Refrigeration.

Adopting the air thermometer as the most eligible measure of temperature, Messrs. Dulong and Petit proceed to investigate the laws of the refrigeration of bodies, under a great variety of circumstances, in vacuo and in air or gases of different kinds and densities. The inquiry abounds with experiments and observations evincing great skill and acuteness; but which it will not suit our purpose to detail. It may suffice for us to give a general summary of the Laws deduced by them from their experiments, at the same time recommending all those who feel sufficient interest in the subject to peruse the essay at large, which exhibits a profound philosophical train of experiments, the results of which are illustrated by the aid of mathematical generalization.

“Law 1. If one could observe the cooling of a body placed in a vacuum, and surrounded by a vessel absolutely destitute of heat, or otherwise deprived of the power of radiating heat, the velocities of cooling would decrease in geometrical progression when the temperatures diminished in arithmetical progression.”

“Law 2. The temperature of a vessel containing a vacuum being constant, and a body being placed in it to cool, the velocities of cooling for excesses of temperature in arithmetical progression, decrease as the terms of a geometrical progression diminished by a constant number. The ratio of this progression is the same for the cooling of all kinds of bodies, and is equal to 1.0077.”

“Law 3. The velocity of cooling in a vacuum for the same excess of temperature, increases in geometrical progression, the temperature of the vessel circumscribing the vacuum increasing in an arithmetical progression. The ratio of the progression is the same as above, namely 1.0077 for all kinds of bodies.”

“Law 4. The velocity of cooling due to the sole contact of a gas is entirely independent of the nature of the surface of the cooling bodies.”

“Law 5. The velocity of cooling due to the sole contact of a gaseous fluid varies in a geometrical progression, while the excess of temperature itself varies in a geometrical progression. If the ratio of this second progression be 2, that of the first is 2.35, whatever be the nature of the gas and its elastic force.”

“This Law may be likewise announced by saying that the quantity of heat carried off by a gas is in all cases proportional to the excess of the temperature of the heated body raised to the power whose index is 1.233.”

“Law 6. The cooling power of a gaseous fluid diminishes in a geometrical progression, when its tension itself diminishes in a geometrical progression. If the ratio of this second progression is 2, the rate of the first is 1.366 for atmospheric air; 1.301 for hydrogen; 1.431 for carbonic acid; and 1.415 for olefiant gas.”

“This law may also be presented as follows: The cooling power of a gas, all other things being alike, is proportional to a certain power of the pressure. The exponent of this power depends on the nature of the gas, and is for air 0.45; for hydrogen 0.315; for carbonic acid 0.517; and for olefiant gas 0.501.”

“Law 7. The cooling power of a gas varies with its temperature in such a manner that if the gas can dilate so as to preserve the same uniform tension, the cooling power will be as much diminished by the rarefaction of the gas, as it is increased by its augmentation of temperature; so that definitively it depends only on its tension.”

Another ingenious Essay was published by Messrs. Dulong and Petit, in the Annal. de chimie et de physique, vol. 10, namely, “Researches on some important points of the theory of heat.”—One object is to ascertain the specific heats of bodies with superior precision. A table of the specific heats of certain metals, found by their method, is given, together with the weights of the atoms of those metals, and the products of the specific heats and weights of the atoms, as under:

Specific heats, that of water being 1 Weights of the atoms, that of oxygen being 1 Product of the weight of each atom by the corresponding capacity.

Bismuth 0.0288 13.300 0.3830

Lead 0.0293 12.950 0.3794

Gold 0.0298 12.430 0.3704

Platinum 0.0314 11.160 0.3740

Tin 0.0514 7.350 0.3779

Silver 0.0557 6.750 0.3759

Zinc 0.0927 4.030 0.3736

Tellurium 0.0912 4.030 0.3675

Copper 0.0949 3.957 0.3755

Nickel 0.1035 3.690 0.3819

Iron 0.1100 3.392 0.3731

Cobalt 0.1498 2.460 0.3685

Sulphur 0.1880 2.011 0.3780

The inference intended from this Table is pretty obvious, namely, that the atoms or ultimate particles of the above bodies contain or attach to themselves the same quantity of heat, or have the same capacity. This principle the authors think will apply to the simple atoms of all bodies, whether solid, liquid, or elastic; but they hold it does not apply to compound atoms. It differs therefore essentially from a suggestion of mine, made eighteen years ago, (see Vol. I. page 70,) that the quantity of heat belonging to the ultimate particles of all elastic fluids, must be the same under the same pressure and temperature. They seem to apprehend, from experience, that a very simple ratio exists between the capacities of compound atoms and that of the elementary atoms. They draw another inference from their researches, that the heat developed at the instant of the combination of bodies, has no relation to the capacity of the elements; this loss of heat, they argue, is often not followed by any diminution in the capacity of the compounds. They seem to think that electricity developes heat in the act of combination; but they do not deny that a change of capacity may sometimes ensue, and heat be developed from this cause.

Remarks on the above Essays.

Results nearly agreeing with those of De la Roche and Berard, on the capacity of certain elastic fluids for heat, were about the same time obtained by M. M. Clement and Desormes. (See Journal de Physique, Vol. 89—1819.) Such results, impugning some of the most plausible doctrines of heat, could not be admitted but upon very good authority. I remained doubtful, in some degree, till satisfied by my own experience. I procured a calorimeter of the construction of De la Roche’s, and to simplify the experiment, instead of forcing a given volume of hot air through the calorimeter to impart heat to the water, I drew, by means of an air-pump, a certain volume of atmospheric (or other air) of the common temperature, through the calorimeter filled with hot water, in order to find how much this process would accelerate the cooling. From several experiments of this kind, I am convinced that the capacity of common air for heat is very nearly such as the above ingenious French chemists have determined. That is, it is about ¹/₇ part only of what Dr. Crawford deduced from his experiments, and nearly the same part of what I inferred from my theoretic view of the specific heats of elastic fluids. (See Vol. I. pages 62 and 74.)

Indeed M. M. De la Roche and Berard appear to have been puzzled with the admission of their own results. The combined heats of oxygen and hydrogen gases give only .6335 for the specific heat of water; whereas by experiment the heat of water is found to be 1, notwithstanding an immensity of heat is evolved during the combination of these gases.[26]

“It is necessary therefore,” they observe, “to abandon the hypothesis which ascribes the evolution of heat in cases of combination to a diminution of specific heat in the bodies combined, and admit with Black, Lavoisier, and Laplace, and many other philosophers, the existence of caloric in a state of combination in bodies.” I am not aware of any writer that denies the existence of caloric in a state of combination of bodies. Dr. Crawford, who would be thought the most likely to err in this respect, maintains, “that elementary fire is retained in bodies partly by its attraction to those bodies and partly by the action of the surrounding heat,” and that “its union with bodies will resemble that particular species of chemical union wherein the elements are combined by the joint forces of pressure and of attraction.” (On animal heat, 2d edition, page 436.) He is perhaps somewhat unfortunate in his instance in the combination of carbonic acid and water; muriatic acid or ammonia and water would have been more in point.

The truth is, these important experiments shew that in elastic fluids the increments of temperature are not proportional to the whole heat, compared with the like increments of temperature and whole heat in those bodies when in the liquid and solid states.

The specific heats of bodies, it is well known, are determined by means of the relative quantities of heat necessary to raise the temperature of those bodies a certain number of degrees. They are expressed by the ratios of those quantities. If the capacities of the same bodies for heat were permanent at all temperatures, then these ratios would also express those of the whole quantities of heat in bodies. In fact, most authors represent the specific heats as expressing both the ratio of the total quantities of heat in bodies, and of the relative quantities to raise their temperature a given number of degrees; but it is the latter only which they accurately represent, and the former only hypothetically.

In regard to bodies in the solid and liquid forms, all experience shews that their capacities for heat are nearly if not accurately constant within the common range of temperature; it seems therefore not unreasonable to infer that the whole quantity of heat in each is proportional to their increments. When, however, a solid body by an increase of temperature assumes a fluid form, and absorbs heat without any increase of its temperature, its total quantity of heat is thus increased; and it is contended by the writers on capacity, that the increments of heat afterwards are increased in the same proportion as the total quantities. This is probable enough; but it ought to be proved in several instances by direct experiment before it can safely be admitted as a general principle; more especially now since the analogy in the case of a liquid becoming an elastic fluid is found to fail in this particular. As an instance of uncertainty, the capacity of ice to water has been found as 9 to 10 by one person, and as 7.2 to 10 by others; such wide difference in the results shows there must be a difficulty in determining the specific heat of ice, and that it may even be doubted whether the specific heat of ice or water is greatest.

From the foregoing detail of experiments on elastic fluids, it appears evident that such fluids exhibit matter under a form in which it has the greatest possible capacity for heat, when capacity is understood to denote the total quantity of heat connected with the fluid; but if the capacity or specific heat is meant to denote the quantity of heat necessary to raise the body a given number of degrees of temperature, then the elastic fluid form of matter is that which has the least capacity for heat of any known form of the same matter. When therefore we use the terms specific heat as applied to elastic fluids we should henceforward carefully distinguish in what sense they are used; but the terms may still be indifferently used in the one or the other sense as applied to liquids and solids, till some more decisive experiments shew that a distinction is required. Probably the anomalies that have occurred in investigations of the zero of cold, or point of total privation of heat, are in part due to the want of accordance between the ratio of the total quantities of heat in bodies, and the ratio of the quantities producing equal increments of temperature.

The greatest possible quantity of heat which a given weight of elastic fluid can contain is when the dilatation of the fluid is extreme. For, condensation, whether arising from mechanical pressure or from increased attraction of the atoms of matter for each other, tends to dissipate the heat, by increasing its elasticity. Hence increase of temperature, at the same time that on one account it increases the absolute quantity of heat in an elastic fluid, diminishes the quantity on another account by an increase of pressure, if the fluid be not suffered to dilate. This is well known from the fact that condensation produces increase of temperature in elastic fluids.

When it is considered that all elastic fluids expand the same quantity by the same increase of temperature, it might be imagined that all of them would have the same capacity, or require the same quantity of heat to produce that expansion. The results of De la Roche and Berard do not seem to admit of this supposition, though the differences of the capacities of elastic fluids of equal volumes are not very great. There is a remarkable difference too between their results and those of Clement and Desormes, in regard to hydrogen gas: namely, .9033 and .6640; also in carbonic acid gas, 1.2583 and 1.5. The subject deserves further investigation.

In reference to the experiments of Dulong and Petit, on the relative expansions of air and mercury by heat, I have no doubt their results are good approximations to the truth. My former experiments were chiefly made in temperatures between 32° and 212°, and I found, as General Roi had done, the expansion of air to be somewhat greater in the lower half than in the upper half of that interval, compared with mercury. On a repetition of the experiments, I think the difference is less than I concluded it to be, and I find that the like coincidence of the air scale and mercurial, continues down to near freezing mercury; at least the difference will not be so great as my new table of temperature makes it at page 14. I have made some experiments on the expansions of air above 212°, which lead me to adopt the results of Dulong. On a comparison of the air and mercurial thermometer upon the laws which I pointed out, namely, the former expanding in geometrical progression to equal intervals of temperature, and the latter expanding as the square of the temperature reckoned from its freezing point, it appears that in the long range of 600° from freezing water to boiling mercury, the greatest deviation of the two thermometers does not exceed 22°. However, the great deviation of the scales between the temperatures of freezing water and freezing mercury, is sufficient to shew, as Dulong and Petit have observed, that their coincidence is only partial. Like the scales of air and mercury, which are so nearly coincident from -40° to 212° that scarcely any difference is sensible, though no one doubts of its existence; yet afterwards the differences become obvious enough, and the greater the farther we advance.

Expansion of Mercury. See page 34, vol. I. I have overrated the expansion of glass bulbs (as will be seen presently,) and hence that of mercury; my expansion of mercury corrected on account of the glass, will be ¹/₅₃ nearly, which leaves it still greater than Dulong’s. The 2nd table of Dulong is valuable, on account of its affording us information of the rate of expansion in the higher degrees of temperature, from a given or standard air thermometer.

Iron, Copper, and Platina.

Expansion of Glass.—By the 3rd Table of Dulong and Petit, it appears these ingenious chemists found the expansion of glass for 180°, or from 32° to 212°, very nearly the same as had been determined previously by Smeaton and others. It also expands increasingly with the temperature, whether it is estimated by the air or mercurial standard. This was observed by Deluc, but more extensively by the present authors. The expansions of iron, copper, and platina, from 32° to 212° as detailed in the 4th table, agree nearly with the results of others; but the expansions in the higher part of the scale manifest some remarkable facts not before known. Platina not only expands the least of the above bodies, but its expansion is almost equable; iron expands more than glass and less than copper, but the most unequally of any one, the expansion increasing rapidly as the temperature advances. These facts explain some others which have fallen under my observation. I was formerly surprised to find glass and iron expand so nearly alike (see vol. I. page 31); but it now appears that iron increases more slowly in proportion than glass about the freezing point. More recently I procured a small thermometrical vessel of platina to contain water like those described at page 31, vol. I, and having filled it and treated it as the other metallic vessels, I was again surprised to find that the apparent greatest density of water in this vessel was at 43°, whereas I expected to have found it below 42°, the point for glass vessels. This observation, in conjunction with Dulong’s, shews, that platina expands more than iron at low temperatures, though for a range of 300° the whole expansion of the platina is to that of the iron as 2 to 3 nearly. Hence the error (for I now consider it as such) which I was led into with respect to the expansion of glass bulbs, (see vol. I. page 32) and subsequently into that of the expansion of mercury abovementioned. It is not the expansion of glass which approaches that of iron, but it is the reverse, which occasions the two bodies to meet so nearly in the table, page 31. This consideration will affect the point of greatest density of water also; for, the less the expansion of iron and glass, the nearer will be the points of real and apparent greatest density of water, contained in vessels of those materials. My observations on brown earthenware are scarcely to be relied upon from the difficulty of making such vessels water tight: but the common white ware I have verified repeatedly since the publication of that table, and am satisfied the point of apparent greatest density, is at or near 40° in such vessels; hence the real maximum density of water must be below 40°. I am inclined to adopt 38° as the most proximate degree.

Capacities of bodies for heat. In the 5th table of Dulong, we have the specific heats of glass and of six metals, determined between freezing and boiling water: that of iron is given before. So far the question does not involve that of the measure of temperature. Their results afford no striking differences from those previously determined; however, it is desirable to find a greater accordance amongst philosophers in this respect. The experiments which give the specific heats between 0° and 300° centigrade, are original and interesting. The results go to shew that the capacities of bodies increase in a small degree with the temperature. But supposing that these results may be relied upon as accurate (which can scarcely be affirmed of any former ones) still the character of them may be changed by adopting a different measure of temperature.

The Essay of M. M. Dulong and Petit, in the 10th vol. of the An. de chimie (see An. of Philosophy, vol. 14th, 1819) manifests great ingenuity. It does not appear, however, so fortunate either in theory or experiment as the former one. It would be difficult to convince any one, either by reasoning or by experience that a number of particles of mercury at the temperature of -40°, whether in the solid,, liquid, or elastic state, have all the same capacity for heat. Indeed the experiments of De la Roche and Berard, if they are to be credited, demonstrate the inferior capacity of condensed air to rarefied air; and if the same body changes its capacity in the elastic form, it may well be concluded that all the three forms have not the same capacity. M. M. Dulong and Petit have themselves shewn, in their former essay, (see page 276) that solid bodies vary in their capacities for heat, and that scarcely any two bodies, vary alike; hence it is impossible that the product of the weight of the atom and specific heat of the body should be a constant quantity. Their specific heat of certain metals differ greatly from what is found by others. For instance, they make the specific heat of lead .0293; the lowest authority I have seen is Crawford, .0352, and the highest Kirwan .050; from repeated trials I have lately found it, upon an average, .032. The weights of some of the atoms in their table, differ materially from what are commonly received; for instance, bismuth is 13.3 instead of 9; also copper, silver, and cobalt, are only half the weights of some authors. The gases too are unfortunate examples. Oxygen gas gives a product of .236 instead of .375; azotic gas gives a product .1967, if oxygen be to azote as 7 to 5, but a product of .393 if oxygen be to azote as 7 to 10: by Dr. Thomson’s ratio of oxygen to azote, 4 to 7, the product will be .482, very different from .375. Hydrogen will give a product of .47 or .41 instead of .375. All these differences, it may be said, are occasioned by errors in the specific heats of the gases; but if errors of this magnitude can still subsist after all the care that has been taken, we shall scarcely know what to trust in experimental philosophy.

If M. Dulong would assume all his simple elements in an elastic state and under one uniform pressure, the hypothesis would then make a part of mine (vol. 1. page 70), and there is great reason to believe it would be either accurately true or a good approximation; but to suppose that some of the bodies should be in a solid state, having their particles united by various degrees of attraction, others fluid, and others in the elastic state, without any material modifications of their heat arising from these circumstances, appears to me to be in opposition to some of the best established phenomena in the mechanical philosophy.

Their observations on the specific heat of compound elements, on the relation of the heat developed by combination, as compared with the heat of the elements before and after the combination, &c., are not supported by a detail of actual experience. Heat given out by chemical changes they suppose not to have been previously in a state of combination with the elements. As an argument, the heat given out by charcoal kept in a state of ignition, by a current of galvanic fluid, is adduced. It is true this case is most easily explained, by allowing that the galvanic fluid is in such circumstances converted into heat. But the charcoal does not undergo any chemical change, and therefore this is not a case in point.

All modern experience concurs in shewing that the heat of combustion is primarily dependent on the quantity of oxygen combining. The heat evolved by the combustion of phosphorus and hydrogen is very nearly, if not accurately, in proportion to the oxygen spent. The heat by the combustion of charcoal is not in a much less ratio: and I find the heat in burning carbonic oxide, carburetted hydrogen and olefiant gas is the same as in burning hydrogen gas, provided the combining oxygen is the same.

One difficulty seems to have occurred to M. M. Dulong and Petit. They all along conceive that the specific heats of bodies, that is, the heats producing equal increments of temperature, must necessarily be proportionate to their whole heat. This is purely hypothetical, till established by experiment. The generality of writers on specific heat had conceived it almost confirmed by experiment. The results of Delaroche and Berard have shewn that in elastic fluids the increments of heat are not proportional to the whole quantities, but on the contrary are less when a body is elastic than when liquid. Indeed some writers have argued this should be the case; because a body nearly saturated with another has less affinity for it left.[27] It is plain then that oxygen gas or any other elastic fluid, may have a small specific heat in the sense above defined, and yet have an almost unlimited quantity of heat. I am not aware of any one established fact that does not admit of an explanation upon the hypothesis that heat exists in definite quantities in all bodies, and is incapable of any change, except perhaps into one of the other equally imponderable bodies, light or electricity.

NEW TABLE OF THE Forces of Vapours in Contact with the Generating Liquids at Different Temperatures.

(A) = Temperatures by the common thermometer.

(B) = Ether vapour, ratio 2 Spec. Gravity .72

(C) = Sulphuret of carbon vapour, ratio 1.978

(D) = Alcohol vapour, ratio 2.7 Spec. Gravity .82

(E) = Acetic acid vapour, ratio 2.57

(F) = Water, ratio 2.602()

(A) (B) (C) (D) (E) (F)

Inches of M. Inches of M. Inches of M. Inches of M. Inches of M.

7° 3.75 3.134 .193 .11

35 7.5 6.20 .560 .27 .29

65+ 15. 12.26 1.51 .69 .75

97 30 24.26 4.07 1.77 1.95

133 60 48. 11.00 4.54 5.07

173 120 29.70 11.7 13.18

220 240 80.2 30. 34.2

272 88.9

340 231

This is an improved and extended table of the force of vapour, similar to that at page 14, vol. I. It shews that the different vapours increase inforce in geometrical progression, to certain intervals of temperature, the same to most or all liquids. These intervals of temperature were presumed in the former table, to be in reality equal to one another; but the accuracy of this last notion has been questioned.

TABLE Shewing the expansion of air, and the elastic force of aqueous and ethereal vapour, at different temperatures.

TABLE III.

Temperat. Vol. of air. Utmost force of Weight of 100 cubic inches of aqueous vapour.

Aqueous vapour. Ethereal vapour.

-28° 420 Inches of Inches of

-20 428 Merc. Merc.

-10 438 Grains.

0 448 .08

10 458 .12

20 468 .17

30 478 .24

32° 480 .26 7.00 .178

33 481 .27 7.18 .184

34 482 .28 7.36 .191

35 483 .29 7.54 .197

36 484 .30 7.73 .203

37 485 .31 7.92 .209

38 486 .32 8.11 .216

39 487 .33 8.30 .222

40 488 .34 8.50 .229

41 489 .35 8.70 .235

42 490 .37 8.90 .245

43 491 .38 9.10 .255

44 492 .40 9.31 .267

45 493 .41 9.52 .275

46 494 .43 9.74 .284

47 495 .44 9.96 .293

48 496 .46 10.18 .303

49 497 .47 10.41 .313

50 498 .49 10.64 .323

51 499 .50 10.87 .329

52 500 .52 11.10 .341

53 501 .54 11.34 .354

54 502 .56 11.59 .366

55 503 .58 11.85 .378

56 504 .59 12.12 .384

57 505 .61 12.39 .396

58 506 .62 12.66 .402

59 507 .64 12.94 .414

60 508 .65 13.22 .420

61 509 .67 13.51 .432

62 510 .69 13.80 .444

63 511 .71 14.10 .456

64 512 .73 14.41 .468

65 513 .75 14.72 .480

66 514 .77 15.04 .492

67 515 .80 15.36 .509

68 516 .82 15.68 .521

69 517 .85 15.90 .539

70 518 .87 16.23 .551

71 519 .90 16.56 .569

72 520 .92 17.00 .580

73 521 .95 17.35 .598

74 522 .97 17.71 .610

75 523 1.00 18.08 .627

76 524 1.03 18.45 .645

77 525 1.06 18.83 .662

78 526 1.09 19.21 .680

79 527 1.12 19.60 .700

80 528 1.16 20.00 .721

Applications of the above Table.

These tables will be found of great use in reducing volumes of air from one temperature or pressure to any other given one: also in determining the specific gravities of dry gases from experiments on those saturated with or containing given quantities of aqueous or other vapours.

As several writers, and some of considerable eminence, have given erroneous or imperfect formulæ on these subjects, more particularly with regard to the effect of aqueous vapour in modifying the weights and volumes of gases, it has been thought proper to subjoin the following precepts and examples for the use of those who are not sufficiently conversant in such calculations.

The 5th column of the above table, or weight of aqueous vapour, is new, and may therefore require explanation. Gay Lussac is considered the best authority in regard to the specific gravity of steam; but it would be well if his results were confirmed or corrected, as they are of importance. According to his experience, the specific gravities of common air and of pure aqueous vapour, of the same temperature and pressure, are as 8 to 5, or as 1 to .625. Now I assume that 100 cubic inches of common air, free from moisture, of the temperature 60° and the pressure of 30 inches of mercury, weigh 31 grains nearly. It is an extraordinary fact that philosophers are not agreed upon the absolute weight of a given volume of common air. Most authors now assume the weight of 100 inches = 30.5 grains, whilst according to my experience it is more than 31 grains. If common air be assumed 31 grains, steam would be 19⅜ grains for 100 cubic inches, at the same temperature and pressure, could it subsist; but as it cannot sustain that pressure at the temperature of 60° we must deduct according to the diminished pressure, the utmost force of steam at 60° being .65 parts of an inch of mercury, we have 30 inches ∶ 19⅜ grains ∷ .65 ∶ .420 grains = the weight of 100 cubic inches of aqueous vapour at 60° and pressure .65 parts of an inch; which is the number given above in the table. The like calculation is required for any other pressure: but in addition to this, there is to be an allowance for the temperature from the 2d column: Thus, let the weight of 100 cubic inches of steam at 32° be required. We have 30 inch. ∶ 19⅜ grs. ∷ .26 inch. ∶ .1679 grs.; the weight of 100 inches of steam at 60°; then if 480 ∶ 508 ∷ .1679 ∶ .178 grs. = weight of 100 cubic inches of steam at 32° and pressure .26 parts of an inch, the tabular number required.

Examples.

1. How many cubic inches of air at 60° are equivalent in weight to 100 cubic inches at 45°?

By the column headed volume of air we have this proportion, if 493 ∶ 508 ∷ 100 inch. ∶ 103.04 inches, the volume required.

2. How many cubic inches of air with the barometer at 30 inches height, are equal in weight to 100 cubic inches when the barometer stands at 28.9 inches?

Rule. The volume of air being inversely as the pressure, we have, 30 ∶ 28.9 ∷ 100 inches ∶ 96⅓ inches the answer.

3. How many cubic inches of dry air are there in 100 inches saturated with aqueous vapour, at the temperature of 50°, and pressure 30 inches of mercury?

Here the formula

(p - f)

p

applies, where p denotes the atmospheric pressure at the time, and f denotes the utmost force of vapour in contact with water at the temperature. Hence p = 30, f = .49 per table, and we have

(p - f) = (30 - .49) = 29.51 = 98¹¹/₃₀, or

p 30 30

98¹¹/₃₀ percent dry air.

& 1¹⁹/₃₀ vapour.

100

If the vapour of ether is assumed, then f = 10.64, and we have

(p - f) = (30 - 10.64) = 19.36 = .645, or ...

p 30 30

64½ per cent dry air.[28]

35½ per cent ethereal vapour.

100

4. Suppose we find by trial the weight of 100 cubic inches of common air saturated with vapour at 60°, the barometer standing at 30 inches to be 30.5 grains, and the weight of hydrogen gas in like circumstances to be 2.118 grains; query the weights of 100 cubic inches of each gas free from vapour, and their specific gravities, the temperature and pressure being as above?

If 30.5 ∶ 2.118 ∷ 1 ∶ .0694 = sp. gr. of vapourized hydrogen, that of vapourized air being 1. Subtracting .42 grs. (weight of vapour per table) from 30.5 grs., leaves 30.08 grains; and subtracting .65 parts of an inch from 30 inches, leaves 29.35 inches. Hence 100 cubic inches of dry air at the pressure of 29.35 inches, weigh 30.08 grains; and we have 29.35 ∶ 30 ∷ 30.08 ∶ 30.746 grains, the weight of 100 inches of dry air. Again, subtracting .42 grs. from 2.118, leaves 1.698 grains = weight of 100 cubic inches of hydrogen of 60° and sustaining the pressure of 29.35 inches; whence if 29.35 ∶ 30 ∷ 1.698 ∶ 1.736 grains, weight of 100 inches of dry hydrogen; and 30.746 ∶ 1.736 ∷ 1 ∶ .05645 = sp. gr. of dry hydrogen, that of dry air being unity. Or the results may be exhibited as under:

Weight of 100 cubic inches. Sp. Gravities.

Vap. air 30.5 grains 1 14.4

Vap. hydrogen 2.118 — .0694 1

Dry air 30.746 grains 1 17.7

Dry hydrogen 1.736 — .05645 1

FORMULÆ FOR DETERMINING THE PROPORTIONS OF COMBUSTIBLE GASES IN MIXTURES.

It frequently happens, especially in the decomposition of vegetable substances by heat, that the product consists of several combustible gases in mixture, and it is desirable to determine the proportions of each of those which collectively constitute the mixture. The following forms will be found useful for this purpose.

1. Carbonic oxide and hydrogen.

Let x = the volume of carbonic oxide, y = that of hydrogen, w = that of mixture, and a = that of carbonic acid, produced by exploding the mixed gases with oxygen over mercury.

Then the carbonic oxide, or x = a, and the hydrogen, or y = w - a.

2. Sulphuretted hydrogen and hydrogen.

Let x = the volume of sulphuretted hydrogen, y = that of hydrogen, w = that of the mixture, and g = the oxygen spent in the combustion of w.

Then because x + y = w,

and 1½x + ½y = g;

we have x = g - ½w,

and y = 1½w - g.

3. Phosphuretted hydrogen and hydrogen; also carburetted hydrogen and hydrogen, and carburetted hydrogen and carbonic oxide.

The notation being as above, we have x + y = w, and 2x + ½y = g (see page 171): and,

x = 2g - w , and y = 4w - 2g .

3 3

4. Olefiant gas and carburetted hydrogen.

The notation being as above, we have

x + y = w, and

3x + 2y = g; whence

x = g - 2w and

y = 3w - g.

5. Carburetted hydrogen, carbonic oxide and hydrogen.

Let x = carburetted hydrogen, y = carbonic oxide, z = the hydrogen, g = the oxygen spent in the combustion of w volumes of mixed gas, and a = the carbonic acid produced.

Then x + y + z = w,

x + ½y + ½z = g,

and 2x + y = a.

whence we have

x = 2g - w , y = 3a - 2g + w

3 3

and z = w - a.

6. Olefiant gas, carburetted hydrogen and carbonic oxide.

Let a = the olefiant gas, y = the carburetted hydrogen, and z = equal the carbonic oxide, g = the oxygen entering into combination, and a = the carbonic acid produced; also w = the whole volume as before.

Then we have x + y + z = w,

3x + 2y + ½z = g,

and 2x + y + z = a.

Whence x = a - w,

y = 4w - 5a + 2g,

3

and z = ⅔(w + a - g).

7. Superolefiant gas,[29] carburetted hydrogen, and carbonic oxide.

Let x = volume of superolefiant, y = volume of carburetted hydrogen, z = volume of carbonic oxide, g = the oxygen combining, a = carbonic acid produced, and w = volume of mixed gas.

Thenx + y + z = w,

4½x + 2y + ½z = g,

and 3x + y + z = a.

Whence x = a - w

2

y = 3w - 4a + 2g

3

and z = 3w - 4g + 5a

6

8. Superolefiant gas, carburetted hydrogen, carbonic oxide, and hydrogen.

This is the mixture of gases obtained by a red heat from coal and oil, after being freed from carbonic acid, &c., by the usual means.

This mixture requires a very complicated formula, in consequence of the specific gravities of the gases entering into the calculus. The importance of the subject however may be an apology for the labour.

Let x = vol. of superolefiant, S its sp. gr.

y = vol. of carb. hydrogen, f its sp. gr.

z = vol. of carbonic oxide, c its sp. gr.

& u = vol. of hydrogen, s its sp. gr.

C = specific gravity of the mixture, g = oxygen, a = carbonic acid, and w = whole volume of mixture as before.

Then we have x + y + z + u = w

4½x + 2y + ½z + ½u = g

3x + y + z = a

and Sx + fy + cz + su = Cw.

Whence u =

(3S + 5c - 8f)a - (4c - 4f)g - (3S + 6C - 6f - 3c)w

8f + c - 3S - 6s.

The value of the hydrogen being obtained, it may be subtracted from w, and the remainder will be best divided into three portions, by the preceding formula.

HEAT PRODUCED BY THE COMBUSTION OF GASES.

Subsequent experience to that detailed at page 77, Vol. 1. has furnished the following more correct results of the heat produced by the combustion of pure gases.

Hydrogen, combustion of it raises an equal volume of water 5°

Carbonic Oxide 4½

Carburetted Hydrogen, or Pond Gas 18

Olefiant Gas 27

Coal Gas (varies with the gas from 10° to) 16

Oil Gas (varies also with the gas from 12° to) 20

Generally the combustible gases give out heat nearly in proportion to the oxygen they consume. See note at the end of Vol. 4, new series of the Manchester memoirs.

ABSORPTION OF GASES BY WATER, &c.

This curious subject has attracted much less attention than it deserves. Very little has been published relating to it since the time of Dr. Henry’s essays and my own, now more than twenty years ago. The only author I remember is M. Saussure of Geneva, who published a similar essay about twelve years afterwards. See Thomson’s Annals of Philosophy, Vol. 6. He investigates the quantities of gases absorbed by various solid bodies, in a manner which I do not fully comprehend; he then treats of the absorption of gases by liquids, adverting at the same time to Dr. Henry’s experiments and mine. My enquiries were principally confined to one liquid, water; but I made a few trials with others, such as weak aqueous solutions of salts, alcohol, &c., and observing no remarkable differences, I concluded somewhat too hastily that “most liquids free from viscidity, such as acids, alcohol, &c., absorb the same quantity of gases as pure water.” Manchester memoirs, new series, Vol. 1. M. Saussure however asserts that there are considerable differences in liquids in this respect. He finds sulphuretted hydrogen to be more absorbable by water than Dr. Henry and I did; in this I find he is right. Water takes about 2½ its bulk of this gas when pure; and it seldom had been obtained unmixed with hydrogen when Dr. Henry and I made our experiments upon its absorption. In regard to carbonic acid, nitrous oxide, and olefiant gas, M. Saussure nearly agrees with us; but his results with oxygen gas, carbonic oxide, carburetted hydrogen, hydrogen and azote, would prove that water absorbs twice the quantities of each that we have assigned. I have no doubt he is wrong in the less absorbable gases. In the case of absorption of mixed gases, Saussure has given four examples, in which he finds the results to militate against my theoretic view, as stated at page 201, Vol. 1.; namely, that water takes the same quantity of each in a mixed state as it would do if they were separate, and in other respects in like circumstances. But I have shewn in the Annals of Philos. Vol. 7, 1816, that his results coincide as near as any one can expect with the views which I have all along taken of this subject.

It will be seen, page 173, that another gas has been found to coincide with olefiant gas in absorbability; namely, phosphuretted hydrogen.

FLUORIC ACID.—DEUTOXIDE OF HYDROGEN.

In treating of Fluoric acid, (Vol. 1, page 277) we came to the conclusion that this acid was probably constituted of two atoms of oxygen, and one of hydrogen, and have figured it accordingly (Plate 5, fig. 38). Subsequent experience however has shewn that deutoxide of hydrogen, though it can be formed synthetically, is not the same thing as fluoric acid. We are indebted to M. Thenard for the discovery of this curious compound, the deutoxide of hydrogen or oxygenated water. An ingenious memoir on the subject was published by him in 1818, in which the formation and the properties of this compound are fully detailed. I had no small satisfaction in 1822, when at Paris, in being obligingly favoured by M. Thenard with a view of the process of the formation, and of the more distinguishing properties of this singular liquid.

The nature of fluoric acid is still enveloped in obscurity. My experience led me to adopt the composition of fluate of lime to be 40 acid and 60 lime per cent. I had not then seen Scheele’s admirable essay on the subject. From the 5th section of his 2d. essay on fluor mineral, 1771, it may be deduced that fluate of lime is composed of 72.5 lime and 27.5 acid per cent. In 1809 Klaproth, and near the same time, Dr. Thomson found about 67½ lime and 32½ acid per cent. in fluor spar. They both erred, no doubt, as I did, by not repeating the treatment of the mineral with sulphuric acid often enough. Since then most authors, as Davy, Berzelius, Thomson, &c., agree with Scheele nearly, in assigning 27.5 acid, and 72.5 lime, in 100 parts of fluate of lime. My experience in 1820 gave me 1 per cent. less of lime; and Dr. Thomson now finds about 1 per cent. more of lime than Scheele’s analysis gives.

If we estimate the atom of lime at 24, that of fluoric acid must be about 9, according with the above proportion; this is much below 15, the weight of an atom of deutoxide of hydrogen.

Should Sir H. Davy’s view of fluate of lime be found correct, its atomic constitution would be one atom of calcium, the metallic substance of which lime is the protoxide, and one atom of fluorine, the name he has assigned to the other element, which with hydrogen is supposed to constitute the fluoric acid. The atom of fluor spar would then be 1 atom of calcium, 17, united to one atom of fluorine 16.

MURIATIC ACID.—OXYMURIATIC ACID, &c.

From the articles muriatic acid and oxymuriatic acid in the former volume, published now 16 years ago, as well as from the appendix to said volume, in which sundry animadversions are found on the fluctuating opinions entertained in regard to these acids, the reader will not be surprised to find some further addition.

Three notions have been submitted to the public in the last twenty years in regard to the nature of muriatic acid. First, the gas detached from common salt by sulphuric acid has been thought to be the acid in a state of purity, and constituted of a certain base or radical united to oxygen; this was the notion inculcated in the articles alluded to above. Second,—it is stated as a fact that when oxymuriatic acid and hydrogen in equal volumes are united by the electric spark, a volume of muriatic acid gas is the result equal to the sum of both the other volumes, and that this gas perfectly agrees with the gas obtained from common salt by sulphuric acid; this suggested the idea that muriatic acid gas is a compound of what has been called real or dry muriatic acid one atom, and water one atom. And, third, it is argued, that the element we have called oxymuriatic acid gas, is, for aught that appears, a simple body, and consequently, that muriatic acid gas is the real acid, and is constituted as above, of one atom of hydrogen, and one atom of oxymuriatic acid (now called chlorine.) It is not intended here to enter into a discussion of the arguments and facts adduced in support of the different conclusions. More experience must be had before all the doubts and difficulties are removed from the subject. But it will be proper to illustrate these different positions by an example. For instance, common salt, muriate of soda or chloride of sodium. By the first notion 50 parts of dry common salt will consist of one atom of muriatic acid gas, 22, and one atom of caustic soda, 28. By the second notion the same salt will be formed of 30 parts of muriatic acid gas, and 28 of caustic soda; but 8 parts of water evaporate when the salt is dried. By the third view common salt consists of oxymuriatic acid, or chlorine and sodium, or the metal of which caustic soda is the protoxide; and 50 parts of salt will consist of 29 chlorine and 21 sodium, or one atom of each.

NITRIC ACID—COMPOUNDS OF AZOTE AND OXYGEN.

Since the account of nitric acid (Vol. 1, page 343) was printed, a change has universally taken place in estimating the weight of the nitric acid atom, and of the proportion of azote and oxygen in the same. This has been effected chiefly by a more correct analysis of nitre than existed at that time. Nitre is now found to consist nearly of 52 parts acid and 48 parts potash per cent. Hence if the atom of potash be 42, that of nitric acid must be 45; for, 48 ∶ 52 ∷ 42 ∶ 45, nearly. That is, the nitric acid atom consists of 10 azote + 35 oxygen by weight; or of 2 atoms of azote (according to my estimate) and 5 of oxygen. There appear to be two nitrous acids; namely, the one which I have designated by that name, which may now be called subnitrous, or as Gay Lussac terms it pernitrous; and the other what I considered as nitric acid in the former volume, composed of 1 atom azote, and 2 of oxygen.

Real nitric acid then is that combination which is effected by uniting oxygen with a minimum of nitrous gas; or 1 measure of oxygen with 1.3 nitrous gas, (See Vol. 1, page 328). The oxynitric acid, which I was led to infer from the last mentioned combination, (1 azote with 3 oxygen) does not appear to exist. The Table of nitric acid (Vol. 1, page 355) will require some correction. An increase of about 4 per cent. should be made, I apprehend, on the quantities of acid corresponding to the several specific gravities.

Since my former volume of Chemistry was printed, several essays on the compounds of azote and oxygen have been published, with some new and some additional experiments, the chief of which may be seen in Sir H. Davy’s Elements of Chemical Philosophy, the Annales de chimie et de physique, Vol. 1; Annals of philosophy, Vol. 9 and 10; and the Manchester Society’s Memoirs, Vol. 4, second series; also Dr. Thomson’s first principles of Chemistry. Notwithstanding all that has been written on the subject, there still appears uncertainty as to the number of combinations formed by these two elements, their relative weights, and the number of atoms in the several compounds.

The results of an experiment I lately made on the decomposition of nitrate of potash by heat seem to be worthy of record, as I am not acquainted with those of any other person who has pursued the experiment to the same extent.—I took an iron retort of 6 cubic inches capacity, and cleaned it as well as I could from carbonaceous matter which it had previously contained, first by heating nitre to redness for an hour or more in it, and then washing it repeatedly with water till the liquid came out tasteless, and only mixed with a little red rust; I then put in 480 grains of purified nitre, and having secured a copper tube to the retort so as to be air tight, the retort was put into a fire and gradually raised to a red heat, and the fire was occasionally urged with a pair of bellows, in order to keep up a glowing red on the retort for nearly two hours; the air was received over water in jars; the first 4 or 5 inches were thrown away, and the rest was preserved and transferred to a graduated jar; the products were examined in successive portions as under, namely,

Inches.

1 produce, 85 cubic inches, 83 per cent. pure = 70.5

2 5 77 = 3.85

3 25 50 = 12.5

4 6 30 = 1.8

Total 121 Oxygen 88.65

Oxygen 88.65 = 30 grains.

Residue 32.35 = 10 grains.

About 1 per cent. on the whole gas was carbonic acid, the rest oxygen and azote, the weights of which would be nearly as above.

Towards the last the gas came very slowly, and being of inferior quality, the operation was discontinued.

The remaining contents of the retort were diluted with water, and well washed till the water ceased to shew alkali; the liquid was then concentrated and gave 1600 water grain measures of the sp. gr. 1.153. There were obtained also 64 grains of red oxide of iron from the washing of the retort, containing 19 grains of oxygen.

The liquid was divided into portions and examined; the original nitre consisted of 250 grains of nitric acid united to 230 of potash = 480 grains. After the process there appeared to be,

10 grains of carbonic acid united to 21 grains potash.

62 grains of subnitrous acid to 84 ””

134 grains nitric acid to 125 ””

230

The quantity of carbonic acid was determined by lime water: the quantity of potash uncombined with nitric acid was found by precipitating it by tartaric acid, and manifested 105 grains of potash in the bitartrate = that combined with the carbonic and subnitrous acids; from which subtracting 21, it was inferred the remainder 84 must have been in union with subnitrous acid, or else with nitrous acid; the rest of the potash, not being acted upon by tartaric acid, was understood to be combined with nitric acid.

The quantity of subnitrous acid given above, appeared somewhat hypothetical, till it was confirmed by treating a portion of the liquid with oxymuriate of lime solution of known strength; it was found that 32 grains of oxygen were required to be combined with the subnitrous acid, in order to restore it to the state of nitric acid; that is, when oxymuriate of lime, containing that quantity of oxygen, was added to the liquid, and this was afterwards rendered acidulous by the addition of sulphuric acid, neither nitrous vapour nor oxymuriatic gas was perceptible; but a greater or less quantity of the oxymuriate being applied, and the liquid made acidulous, the fumes of the one or the other were abundantly manifest.

It remains to account for the oxygen. There were 250 grains of nitric acid at first in the nitre; of which 200 grains were oxygen and 50 azote, nearly. One-fifth part of the oxygen = 40 grains, corresponds to 1 atom of oxygen. Now the whole of the oxygen derived from the nitre in the course of the experiment, seems to be 30 grains in gas, 7 grains in the carbonic acid, and 19 grains in the iron oxide, together equal to 56 grains. Now the azote and oxygen in the gas collected, were very nearly in the proportion of those elements in nitric acid; so that a portion of the acid (about ⅙) might be considered as completely decomposed, whilst the rest was only losing a small part of its oxygen: this is remarkable, and I think indicates that the carbonic acid (formed from the carbon of the retort, or from the adhering carbon) unites to the potash, expelling the nitrous acid, which is immediately decomposed into its elements azote and oxygen. This would not however account for the whole of the azote: for, 40 grains of nitric acid would be united to 37 potash; whereas we find only 21 potash with carbonic acid; and I cannot believe that an error in the estimate of carbonate of potash could exist to that amount. The fact, however, was, that the elements of 40 grains of nitric acid were found in the evolved gas, and hence we have to account for the remainder 210 grains. From this there appears to have been expelled 26 grains of oxygen, nearly 19 and 7 as related above; of which the 19 grains cannot be correctly estimated by reason of the uncertainty as to the real quantity of oxide formed during the operation: there might be some left adhering to the retort, or on the other hand there might be more than the due share, derived from former experiments. Supposing then, that 26 grains of oxygen were extracted from the nitric acid, the remaining acid would require the same to be added to re-form the nitric; but by the experiments with oxymuriate of lime it seemed to require 32 grains of oxygen. This difference wants an explanation; I believe the greater error must belong to the 26 grains; perhaps the truth might be approximated best by supposing both to be 30 grains.

When the liquid decomposed nitre is treated with any acid, a gas is instantly expelled which produces red fumes in the air; it is pure nitrous gas, which joining with the oxygen of the atmosphere, generates nitrous acid vapour. At the same time, no doubt, the subnitrous acid is disengaged from the potash, but that part of it which is real nitrous acid (1 atom azote to 2 of oxygen) is retained by the water, whilst the other part, (1 atom azote and 1 of oxygen) assumes the gaseous form. In order to be satisfied respecting this point, I made several experiments with the liquid over mercury: taking a given portion of the liquid, and sending it to the top of a graduated tube filled with mercury, I passed up as much muriatic acid as was sufficient to engage the potash; immediately there was a disengagement of nitrous gas and carbonic acid gas, and afterwards a slow evolution of gas, evidently arising from the liquid in contact with the mercury. Wishing to ascertain the quantities, I sent up 25 grain measures of liquid, and to that nearly half its bulk of muriatic acid; in 2 or 3 minutes there was,

1.1 cubic inch of gas. H. M.

1.4 in 0 45

1.5 1 5

1.7 2 45

1.75 7 45

1.78 9 45

The gas was washed in lime water, and lost .33 parts of an inch of carbonic acid; the rest, 1.45 cubic inch, was nitrous gas. It is obvious that ½ of the nitrous gas, together with the carbonic acid, was liberated instantly; the rest of the nitrous gas was due to the nitrous acid, slowly acting upon the mercury. At the end of the process, there was a little black oxide floating upon the mercury. Calculating from this, the whole quantity of nitrous gas would be 31 or 32 grains, whereas it ought to have been 48 grains to constitute 62 of subnitrous acid. It is probable that whilst a portion of the subnitrous acid is oxidizing the mercury, another portion may be forming nitric acid and dissolving the oxide.

From some trials, I have reason to think that even carbonic acid will expel nitrous gas from the liquid sub-nitrite of potash.

In the essay of Dr. Henry, already alluded to, published in the 4th Vol. of the Manchester Society’s Memoirs, a new and interesting discovery is made; namely, that a mixture of nitrous and olefiant gases, though not explosive by an electric spark, may still be exploded by the more powerful impetus of a shock from a charged jar. Dr. Henry has adduced the results obtained in this way, as corroboratory of those which shew the constitution of nitrous gas to be 1 volume of azote and 1 of oxygen united to form 2 volumes of nitrous gas. (See page 507 of the Memoirs.)

Some time ago in repeating these experiments of Dr. Henry, I found some extraordinary circumstances attending them. After determining that 1 volume of olefiant gas may be fired with from 6 to 10 volumes of nitrous, I found a shock from a jar sometimes inadequate to fire the mixture, which, however, when repeated a second or third time, succeeded. This is not a novelty; for, mixtures of olefiant gas as well as other gases and vapours, with a minimum of oxygen, frequently require several sparks before the explosion: but this case occurs at times with nitrous and olefiant gas, when they are mixed in the most favourable proportions for exploding. The most remarkable circumstance, however, was, that when a phial was filled with the mixture of the two gases in the proportion of 1 volume olefiant to 6 or 7 nitrous, (exclusive of small portions of azote), the decomposition of the nitrous gas and the combustion of the olefiant were scarcely ever perfect; and what increased the perplexity more, was, the results obtained from the same mixture scarcely ever agreed one with the other. After about 30 experiments, I was inclined to adopt the conclusion, that the uncertainty was occasioned by the oblong form of the eudiometer. The spark or shock, in my eudiometer, is imparted at one extremity of a column of air, which is often 10 times as much in length as in diameter: it mostly was found that the larger the quantity of mixture exploded at once, the more imperfect and incomplete was the combustion. I imagine the intensity of heat is not sufficient to carry on the combustion through the length of the column, owing, perhaps, to the cooling power of the sides of the tube. Hence it was, I apprehend, that in one or two instances, when a small quantity of gas was used, I got nearly complete results, as Dr. Henry reports his; but in the majority both gases were found in the residue after the explosion.

In pursuing this enquiry into the decomposition of nitrous gas by combustible gases, I found that it might be effected by any combustible gas or vapour: at least it succeeded in all I tried. The method I pursued, and which was suggested by the known properties of phosphuretted hydrogen, is this: it has been shewn (page 181) that a mixture of phosphuretted hydrogen and nitrous gas exploded by an electric spark, the former gas being completely burned in case the proportions are duly adjusted; now, it occurred to me, that as the above combustible gas is usually a mixture of pure phosphuretted hydrogen and of hydrogen, and that the latter of these is also burned as well as the former, the effect must be produced through the heat occasioned by the combustion of the former. Having some old phosphuretted hydrogen by me, at the time, which on examination, I found to be 91 per cent. combustible gas, and 9 azote; and the 91 combined with 156 of oxygen, consequently was 74 pure, and 17 hydrogen; I tried this mixture with nitrous gas, when it exploded by the spark, as usual; but on trying it with an excess or defect of nitrous gas, the spark was inefficient, but the shock instantly fired the mixture. As there did not appear to be any of the pure hydrogen left unburned in these experiments, I proceeded to mix the old phosphuretted hydrogen with hydrogen; and then this new mixture with nitrous gas. The first experiment was made with 4 parts of old phosphuretted hydrogen + 16 hydrogen + 36 nitrous gas = 56 total. On this mixture the spark, of course, had no effect; but it exploded the first trial by the jar, and left 20 measures, of which 2 were found to be oxygen, and the rest azote. This experiment succeeding so well, I next tried mixtures of phosphuretted hydrogen, with carbonic oxide, carburetted hydrogen, and ether vapour successively, along with nitrous gas; and found that all these mixtures refused combustion by the spark, but were instantly exploded by the shock, yielding carbonic acid and water, the same as if the combustion had been effected by free oxygen. In some instances the combustion was complete, leaving neither combustible gas nor nitrous gas; but generally there was a residue of one or both of the gases.

From these experiments it may be concluded that the heat, produced by the combustion of phosphuretted hydrogen and nitrous gas or oxygen gas, disposes other gases, accidentally in the mixture, to chemical changes. In conformity with this view, I mixed phosphuretted hydrogen and oxygen, in the proportion of mutual saturation; and taking a small proportion of this mixture, and as much ammoniacal gas as would saturate the phosphoric acid to be formed, I found that causing an explosion over mercury, the phosphoric acid combined with the ammonia, and nearly the whole gas disappeared. In this case, the heat was not sufficient to decompose the ammonia. But in another experiment, with a portion of the same explosive mixture and a less proportion of ammonia, after the firing a residue of azote and hydrogen was found, amounting nearly to the quantity due from the decomposition of the ammonia. Here the heat produced, had evidently decomposed the ammonia.

ON AMMONIA.

The constitution of ammonia still remains undecided. The latest experiments on this article are those of Dr. Henry, in his essay on the analysis of the compounds of nitrogen. (Memoirs of the Manchester Society, vol. 4, 1824.) By electrifying ammoniacal gas over mercury, as carefully as could be devised, Dr. Henry found results as under:

1st experiment 44 measures became 88+

2d 157 320

3d 60 122

4th 120 240

The evolved gases carefully analysed by combustion with oxygen, were found to consist of 3 volumes of hydrogen and 1 of azote. The analysis of ammonia was also effected by exploding it with nitrous oxide, with the requisite precautions. The results confirmed the previous ones by electricity, both in regard to doubling the volume of ammonia, and establishing the ratio of 3 to 1 in the volume of hydrogen and azote.—These experiments are highly interesting as far as regards the question of ammonia, as they exhibit the latest investigations of one who has previously shewn uncommon skill and perseverance in this kind of analysis. (See Philos. Transact. 1809, &c.)

Dr. Henry’s analysis of ammonia, in 1809, has been adverted to in our article on the subject, vol. 1, page 429. The results of that Essay are given in a tabular form; and the mean of six experiments was nearly as we have stated, namely, that ammonia consists of 27¼ measures of azote, and 72¾ hydrogen. To this it may be proper to add, that the two extremes were, 26.1 azote and 73.9 hydrogen, and 28.2 azote with 71.8 hydrogen; also that a small error has crept into the table, which being corrected, the average results are reduced to 27 and 73, very nearly. Subsequently, both Dr. Henry and Sir H. Davy concurred in assigning 26 and 74 for the most approximating numbers. (See Nicholson’s Journal, 25, page 153). The true quantity of gases procured by the decomposition of ammoniacal gas by electricity, was concluded by both these authorities, to be 180 for each 100 of ammonia, when the requisite precautions were taken, as we have related in vol. 1.

From what is stated above, it is evident the subject is one which requires extraordinary skill and attention. This I can attest from my own experience, which has been frequently renewed and varied; but the results have not been sufficiently accordant to yield me satisfaction.

About ten years ago, I made several experiments on the decomposition of ammonia, which, though they are not convincing, deserve, perhaps, to be recorded in their results.—Some more recent experiments are incorporated with them.

Decomposition of ammonia by nitrous oxide.—I made many experiments, by exploding mixtures of nitrous oxide and ammoniacal gases over mercury. The excess of gas was mostly on the side of ammonia, but the proportions were varied in the different experiments, from 10 vol. nitrous oxide to 11 ammonia or to 5, which are about the extremes capable of being fired by the electric spark.

When 10 parts nitrous oxide and 5 of ammonia are exploded over mercury, the residuary gas contains some free oxygen and some nitrous acid derived from the decomposition of the excess of nitrous oxide used; with 6 parts of ammonia there is rarely any free oxygen. When 10 parts of nitrous oxide, and 7 of ammonia are fired, I never found any free oxygen or hydrogen; but when the ammonia is at or near 8 parts, I find from ¹/₂₀ to ⅒ of the hydrogen from the ammonia in the residuary gases. The two gases appear to be completely decomposed; the oxygen of the nitrous oxide, as far as it can, unites with the hydrogen of the ammonia, without forming any portion of nitrous acid or of free oxygen, and the residue contains the azote of both gases, and the unburnt hydrogen from the ammonia, as Dr. Henry first observed. This continues to be the case till the ammonia becomes 11 parts, when the hydrogen amounts to about ⅓ of the whole quantity which the ammonia yields.

From the above it would seem that the proportions for mutual saturation must be 10 nitrous oxide with from 7 to 8 parts of ammonia. This agrees with the deduction in Dr. Henry’s first essay that 13 nitrous oxide require 10 of ammonia; or that 10 require 7.7: but according to the theory of volumes 10 would require 6⅔; and Dr. Henry recommends in his late essay 10 nitrous oxide to 7.7 or 8⅓ parts of ammonia, in order to secure a small excess of the last, and consequently some free hydrogen after the explosion. The former of these proportions would have nearly ⅐ of the residue hydrogen, and the latter nearly ⅕, supposing the gases pure originally. This gives more hydrogen than I have ever found; but the azote in my experience nearly agrees with the doctrine of multiple volumes.

Decomposition of ammonia by nitrous gas.—About 30 experiments carefully made on mixtures of nitrous gas and ammoniacal gas gave very discordant results. At one time 10 parts nitrous gas with 14 ammonia gave ⅓ of hydrogen in excess, and another time 10 nitrous with 12 ammonia gave excess of hydrogen = ⁹/₂₀; generally 10 parts with 6 or less gave oxygen, and 10 with 8 or more gave hydrogen in the residue.

Decomposition of ammonia by oxygen.—The limiting proportions of oxygen and ammonia which I have fired, are 10 oxygen to 4 ammonia for the minimum, and 10 oxygen to 22 ammonia for the maximum. When 10 oxygen were fired with 4 ammonia, there were ²⁵/₃₇ of the oxygen left, and there was a deficiency of azote amounting to ¹/₁₂ of what was expected from the ammonia, owing no doubt to nitrous acid generated by the explosion. When 10 oxygen to 1.8, or from that to 2.2 ammonia are used, there is a surplus of about ¼ or ⅓ of the hydrogen contained in the ammonia, left in the residue of the gas. When the ammonia is between 13 and 14 there is usually a trace of oxygen or hydrogen as it approaches either of these limits. By the theory of volumes, 10 oxygen should saturate 13⅓ of ammoniacal gas. I have not any instance of hydrogen being left when 14 ammonia were used, though there ought to be ¹/₂₀ of the whole left; and much smaller quantities than that are appreciable by well known methods. The azote resulting from the decomposition of ammonia is usually very nearly ½ the volume of the ammonia.

On the whole the results from firing ammonia and oxygen gas appear to me more satisfactory than those obtained from nitrous oxide and nitrous gas, as they are more simple and less perplexed with any theoretic views.

It may be proper to remind the reader that when we speak of 10 parts of one gas uniting with 8, 10, or more, of another in the above and other cases, it is to be understood of gases absolutely pure; not that we ever obtain them in that state, but approximating as near as we can to it, we mix given portions of such gases as we can obtain, and then in our calculations of results deduct for the impurities.

One source of uncertainty in these experiments on firing mixtures of ammonia, is that the real quantity of ammoniacal gas operated upon is not known. If a certain measure of ammonia be transferred through mercury ever so dry, some portion of it gets entangled in the mercury, and 100 measures become perhaps 95: now in the explosion it is a question whether any part of the 5 measures absorbed is decomposed. I have marked this attentively, and am persuaded that generally speaking, little if any of that portion is decomposed; but some trace of it appears mostly afterwards in the residue as it is liberated from the pressure of its own kind of gas, and hence easily rises into the gaseous mixture. Notwithstanding, when the loss of gas by transfer amounts to 10 or 20 per cent., I have reason to believe that some part of it suffers combustion occasionally.

Volume of gases from the decomposition of ammonia.—It has been observed (vol. 1. Ammonia) that Sir H. Davy obtained 180 measures of gases, by means of electricity, from 100 of ammonia as the maximum when the operation was performed with great care, and Dr. Henry in like circumstances, produced 181, whilst I found 187 measures; since that, as has been related, Dr. Henry has found 200 measures. It is not easy to account for these differences; I am inclined to the opinion that the volume of gases is very nearly doubled, but probably rather less than more. I find the experiments on the rapid combustion of ammonia agree best with that opinion.

Decomposition of ammonia by a red heat.—A short time since I repeated the decomposition of ammonia by passing the gas through a red hot copper tube. The proportion of azote to hydrogen, due allowance being made for a minute portion of atmospheric air, was upon the average of a number of experiments, 26 of the former to 74 of the latter.

Decomposition of ammonia by oxymuriatic acid.—I have made several experiments on this mode of decomposition since the results published in vol. 1, page 435. It is well known that a solution of oxymuriate of lime decomposes ammoniacal salts; water and muriatic acid are produced, azote liberated, and the acid previously combined with the ammonia is evolved. But this is not all; an excessively pungent gas or perhaps vapour is produced, exciting sneezing, and inducing catarrh; the constitution of this vapour is not well understood; it is never formed, as far as I know, without the presence of both oxymuriatic acid and ammonia. The results of such mixtures are of course complicated and likely to be unsatisfactory; it may notwithstanding be useful to relate some of them.

When clear oxymuriate of lime solution, and a salt of ammonia are mixed together with a little excess of oxymuriate, the ammonia is mostly decomposed, the oxymuriate being converted into muriate of lime by the hydrogen of the ammonia, whilst the azote is evolved, and the acid previously combined with the ammonia is liberated; hence oxymuriatic acid gas is also liberated along with the azote; and it is required to be taken out before the azote can be estimated. This circumstance may be obviated by previously adding the requisite quantity of pure potash or soda, to engage the acid, or by leaving a little undissolved lime in the oxymuriatic solution. I could never obtain a volume of azote equal to half that of the ammonia (supposed to be in a gaseous state) though it is universally allowed not to be less than that, if the whole of the azote be evolved; on one occasion only I got so much as ¹⁴/₁₅ of that quantity. The residue of liquid has the extremely pungent smell; but the azotic gas after passing through pure water has no smell. When this experiment is made over mercury, the oxymuriatic acid acts upon it, and hence the excess of oxymuriate should be such as to leave a portion of that undecomposed at the conclusion.

When the object is to ascertain the hydrogen in ammonia, a portion of salt known to contain a given weight of ammonia is to be treated with oxymuriate of lime solution, the strength of which is accurately determined by means of green sulphate of iron, or otherwise. The ammoniacal salt in solution is then to be mixed with a moderate redundance of the oxymuriate liquid, and with a few drops of caustic potash, and the mixture must be repeatedly agitated for some time. At length the liquid must be tested by the green sulphate of iron, and hence the quantity of acid spent upon the ammonia will be determined. I have mostly found the hydrogen this way below the common estimate, allowing the ammoniacal salts to be correctly determined.

SULPHURET OF CARBON.

Since the article at page 462, vol. 1, was written, an excellent essay on the sulphuret of carbon has been published in the Philosophical Transactions, (1813) by Professor Berzelius and Dr. Marcet. After an extensive series of experiments, they infer the atom of the sulphuret to consist of 2 atoms sulphur and 1 of carbon. The investigation did not seem to warrant their including hydrogen in the atom. I have made several experiments on the combustion of the vapour of sulphuret of carbon in oxygen gas by electricity. My method generally was, to vapourize a given portion of atmospheric air over mercury, taking care that the vapour was below the maximum for the temperature; this is easily effected by putting the liquid into a phial of air, drop by drop, and inverting it over mercury till the liquid is evaporated. This vapourized air, I find may be transferred through mercury with very little loss, and even through water several times, without a total condensation of the vapour. The vapour of ether is much more condensible by water than that of sulphuret of carbon. A given portion of this vapourized air is to be mixed with oxygen gas, in Volta’s eudiometer, and then exploded by the electric spark over mercury. One volume of vapour combines with nearly 3½ of oxygen, and therefore requires 4 or 5 times its bulk of that gas before firing, in order that the combustion may be complete. The results of the combustion are carbonic acid and sulphurous acid; and I suspect a small portion of water; though Professor Berzelius and Dr. Marcet could not detect any.

By evaporating a given weight of the sulphuret of carbon, in a given volume of atmospheric air, at the temperature of 60°, I find the specific gravity of the vapour to be 2.75 nearly, air being 1. Now if we assume the atom of vapour to be nearly of the same volume as that of hydrogen, and to consist of 1 atom hydrogen, 2 sulphur, and 1 carbon, it will require 7 atoms of oxygen to form water, sulphurous acid, and carbonic acid, which will accord very well with my experience. When vapourized hydrogen gas is electrified for some time, there is no change of volume, though there is some appearance of decomposition. Probably the hydrogen of the sulphuret is liberated. It is difficult to conceive how so volatile a liquid as the one in question, could be constituted out of sulphur and carbon without the addition of hydrogen.

POTASSIUM, SODIUM, &c.

Two views of the nature of these bodies have been given in vol. 1, (see pages 260, and 484, &c.). In the former they are considered as simple metals; in the latter, as compound bodies resulting from the abstraction of oxygen from the hydrates of potash and soda; or as being constituted of 1 atom of hydrogen united to 1 atom of pure potash or soda respectively. Those who have had the most experience on these elements, Sir H. Davy, and M. M. Gay Lussac and Thenard, seem now to concur in the former view, and it has been adopted by most chemists. Part of the objections which we made to this view have been obviated, it should seem, by establishing the fact, that oxymuriatic gas and hydrogen gas united, form muriatic acid gas. There are still, however, difficulties to remove before this view can be considered perfectly satisfactory; but they are not greater perhaps than would attach to any other explanation of the facts connected with the subject. Besides potassium and sodium, experience as well as analogy would seem to render probable, if not to establish, the existence of barium, strontium, and calcium as metals, of which barytes, strontites, and lime are the protoxides, as potash and soda are of the other two metals; (other oxides of potassium and sodium are stated, see page 55-57); barium has a deutoxide, and probably calcium likewise. The rest of the earths, as magnesia, alumine, silex, &c. are by analogy considered by most chemists as oxides of particular metals, but the proportions of their elements have not been determined.

ALUM.

At page 531, vol. 1, we have given the constitution of this important salt, as under: since that time Mr. R. Phillips has announced another view of it; and Dr. Thomson has published one differing from both of these. They are as follow:

Dalton— 1 atom sulphate of potash.

4 atoms sulphate of alumine.

30 atoms water.

Phillips— 1 atom bisulphate of potash.

2 atoms sulphate of alumine.

22 atoms water.

Thomson— 1 atom sulphate of potash.

3 atoms sulphate of alumine.

25 atoms water.

Notwithstanding these differences, there is a near approximation in all three, in regard to the quantities of acid, alumine, potash, and water in the salt. This is accounted for partly in the different relative weights of the atoms, as estimated by the different analysts, but chiefly in that of alumine.

Some very curious results occurred to me about 10 years ago in analysing alum; they were new to me, but I have since found they had been previously discovered by Scheele. (See his essay on silex, clay, and alum, 1776.) As his observations are not to be found in any of our elementary books that I have seen, I shall give the particulars of my own experiments here.

I take 24 grains of alum and dissolve them in water; of these 8 grains may be allowed for sulphuric acid, ⅕ of which = 1.6 grain = 1.1 grain of lime = 880 grains of lime water, such as I commonly use. To the solution of alum I put 880 grains of lime water; a slight precipitate appears which soon becomes redissolved almost completely. The liquid is then acid by the colour test.

To this liquid I put 880 more of lime water, and agitate; a copious precipitate appears and continues; after subsidence the clear liquid is still acid by the colour test.

Another 880 grains are added, and the whole is then well agitated; the agitation is repeated two or three times after the precipitate has partly subsided, so as to diffuse it equally again through the liquid; finally, the clear liquid is found to be neutral by the colour test, and to contain no alumine; for, lime water produces no precipitate when poured into it.

Another 880 grains being added, and the whole stirred well, the clear liquid after the subsidence of the precipitate is still neutral by the colour test.

The fifth portion of 880 grains being then added, and the mixture well agitated, a considerable portion of the precipitate will evidently disappear, and the mixture become semitransparent; after a time the clear supernatant liquid is found strongly alkaline; a little of it touched with an acid becomes milky, and adding more acid clears it again. The liquid is now 1.0025 sp. gr., or a little heavier than lime water.

The sixth portion of 880 grains being now added to the whole mixture, and agitated, the precipitate rather diminishes, and an increase of specific gravity takes place in the liquid; it is now 1.003.

The seventh and last portion of 880 grains being added to the mixture, and agitation being continued for some time, a dense bulky precipitate is formed, which falls with great celerity, carrying with it the greatest part of the acid, the alumine and the lime, and leaving the liquid of the sp. gr. 1.0012. It is a subsulphate into which acid, potash, lime and alumine enter, as will be shewn.

These phenomena appear to me to be best explained by adopting a constitution of alum, such as to make it consist of 1 atom bisulphate of potash and 3 atoms of sulphate of alumine; after which the following explanation will apply.

The first portion of lime water saturates the excess of acid.

The second portion throws down a correspondent portion of alumine. The clear liquid is acid, because it contains sulphate of alumine, which is essentially acid by the colour test, because alumine is not an alkaline element.

The third portion throws down another portion or atom of alumine; but by continued agitation the two atoms of alumine liberated, join the remaining atom of sulphate of alumine, and the whole compound falls down, being then the common subsulphate of alum. Hence the liquid, containing nothing but sulphate of lime and sulphate of potash, is neutral by the test, and yields no alumine by the addition of lime water.

The fourth portion of lime water being put in and duly agitated, the atom of sulphuric acid is drawn from the subsulphate to join the lime, and then the floating subsulphate of alumine becomes pure alumine, and the clear liquor is still neutral.

The fifth portion of lime water tries to decompose the sulphate of potash, but is unable of itself; however, the floating alumine assists it, and by double affinity the potash leaves the acid to join the alumine, and the lime takes the acid. Hence as ⅓ of the alumine enters into solution with the potash, the precipitate is less copious, and the liquid is alkaline; a small portion of acid put into the clear liquid engages the potash, and liberates the alumine, but a larger portion redissolves the alumine also.

The sixth portion of lime water seems to complete the effect which the fifth commences, and hence the density of the liquid increases, whilst the precipitate rather diminishes.

The seventh portion of lime, together with the sixth, after due agitation and some time, unite the lime with the alumine, one atom of each, and form a precipitate which would fall together, were no other compound present, as I found, and Scheele before me; but if sulphate of lime be present, each compound atom of lime and alumine, unites with one of sulphate of lime, and the whole descends together, forming a subsulphate resembling that of alum, only two atoms of lime are found as substitutes for two atoms of alumine. This subsalt is very little soluble in water.

According to this view, if 2 atoms of alum were decomposed, 4 atoms of subsulphate would be formed, each consisting of 1 acid, 2 lime, and 1 alumine; also 2 compound atoms of potash and alumine, and 6 atoms sulphate of lime. But in the final arrangement, it would seem, that 2 atoms of sulphate of lime are again decomposed, and sulphate of potash formed, the 2 atoms of lime combining with the 2 of alumine, and then two more atoms of subsulphate are formed, and the final arrangement is 6 atoms subsulphate precipitated, and 2 atoms sulphate of potash, and 2 sulphate of lime remain in solution.