Thoughts on the Constituent Parts of Water and of Dephlogisticated Air; With an Account of Some Experiments on That Subject. In a Letter from Mr. James Watt, Engineer, to Mr. De Luc, F. R. S.

XXV. Thoughts on the constituent Parts of Water and of Dephlogisticated Air; with an Account of some Experiments on that Subject. In a Letter from Mr. James Watt, Engineer, to Mr. De Luc, F. R. S. Read April 29, 1784. DEAR SIR, In compliance with your desire, I send you an account of the hypothesis I have ventured to form on the probable causes of the production of water from the deflagration of a mixture of dephlogisticated and inflammable airs, in some of our friend Dr. Priestley's experiments. I feel much reluctance to lay my thoughts on these subjects before the public in their present indigested state, and without having been able to bring them to the test of such experiments as would confirm or refute them; and should, therefore, have delayed the publication of them until these experiments had been made, if you, Sir, and some other of my philosophical friends, had not thought them as plausible as any other conjectures which have been formed on the subject; and that though they should not be verified by further experiments, or approved of by men of science in general, they may perhaps merit a discussion, and give rise to experiments which may throw light on so important a subject. I first thought of this way of solving the phenomena in endeavouring to account for an experiment of Dr. Priestley's, LEY'S, wherein water appeared to be converted into air; and I communicated my sentiments in a letter addressed to him, dated April 26, 1783*, with a request that he would do me the honour to lay them before the Royal Society; but as, before he had an opportunity of doing me that favour, he found, in the prosecution of his experiments, that the apparent conversion of water into air, by exposing it to heat in porous earthen vessels, was not a real transmutation, but an exchange of the elastic fluid for the liquid, in some manner not yet accounted for; therefore, as my theory was no ways applicable to the explaining these experiments, I thought proper to delay its publication, that I might examine the subject more deliberately, which my other avocations have prevented me from doing to this time. 1. It has been known for some time, that inflammable air contained much phlogiston; and Dr. Priestley has found, by some experiments made lately, that it "is either wholly pure phlogiston, or at least that it contains no apparent mixture of any other matter." (In my opinion, however, it contains a small quantity of water and much elementary * This letter Dr. Priestley received at London; and, after shewing it to several Members of the Royal Society, he delivered it to Sir Joseph Banks, the President, with a request that it might be read at some of the public meetings of the Society; but before that could be complied with, the author, having heard of Dr. Priestley's new experiments, begged that the reading might be delayed. The letter, therefore, was reserved until the 22d of April last; when, at the author's request, it was read before the Society. It has been judged unnecessary to print that letter, as the essential parts of it are repeated, almost verbatim, in this letter to M. De Luc; but, to authenticate the date of the author's ideas, the parts of it which are contained in the present letter are marked with double commas. Parts of Water and of Dephlogisticated Air. "He found, that by exposing the calces of metals to the solar rays, concentrated by a lens, in a vessel containing inflammable air only, the calces of the softer metals were reduced to their metallic state;" and that the inflammable air was absorbed in proportion as they became phlogisticated; and, by continually supplying the vessel with inflammable air, as it was absorbed, he found, that out of 101 ounce measures, which he had put into the vessel, 99 ounce measures were absorbed by the calces, and only two ounce measures remained, which, upon examination, he found to be nearly of the same quality the whole quantity had been of before the experiment, and to be still capable of deflagrating in conjunction with atmospheric or with dephlogisticated air. Therefore, as so great a quantity of inflammable air had been absorbed by the metallic calces; the effect of reducing them to their metallic state had been produced; and the small remaining portion was still unchanged, at least had suffered no change which might not be attributed to its original want of purity; it was reasonable to conclude, that inflammable air must be the pure phlogiston, or the matter which reduced the calces to metals. 2. "The same ingenious philosopher mixed together certain proportions of pure dry dephlogisticated air and of pure dry inflammable air in a strong glass vessel, closely shut, and then set them on fire by means of the electric spark," in the same manner as is done in the inflammable air pistol. The first effect was the appearance of red heat or inflamma- * Previous to Dr. Priestley's making these experiments, M. Kirwan had proved, by very ingenious deductions from other facts, that inflammable air was, in all probability, the real phlogiston, in an aerial form. These arguments were perfectly convincing to me; but it seems more proper to rest that part of the present hypothesis on the direct experiment. tion in the airs, which was soon followed by the glass vessel becoming hot. The heat gradually pervaded the glass, and was dissipated in the circumambient air, and as the glass grew cool, a mist or visible vapour appeared in it, which was condensed on the glass in the form of moisture or dew*. When the glass was cooled to the temperature of the atmosphere, if the vessel was opened with its mouth immersed in water or mercury, so much of these liquids entered, as was sufficient to fill the glass within about \(\frac{1}{60}\)th part of its whole contents; and this small residuum may safely be concluded to have been occasioned by some impurity in one or both kinds of air. The moisture adhering to the glass, after these deflagrations, being wiped off, or sucked up, by a small piece of sponge paper, first carefully weighed, was found to be exactly, or very nearly, equal in weight to the airs employed. In some experiments, but not in all, a small quantity of a footy-like matter was found adhering to the inside of the glass, the origin of which is not yet investigated; but Dr. Priestley thinks, that it arises from some minute grains of the mercury that was used in order to fill the glass with the air, which being super-phlogisticated by the inflammable air, assumed that appearance; but, from whatever cause it proceeded, "the whole quantity of footy-like matter was too small to be an object of consideration, particularly as it did not occur in all the experiments." I am obliged to your friendship for the account of the experiments which have been lately made at Paris on this subject, * I believe that Mr. Cavendish was the first who discovered that the combustion of dephlogisticated and inflammable air produced moisture on the sides of the glass in which they were fired. with large quantities of these two kinds of air, by which the essential point seems to be clearly proved, that the deflagration or union of dephlogisticated and inflammable air, by means of ignition, produces a quantity of water equal in weight to the airs; and that the water, thus produced, appeared, by every test, to be pure water. As I am not furnished with any particulars of the manner of making the experiment, I can make no observations on it, only that, from the character you give me of the gentlemen who made it, there is no reason to doubt of its being made with all necessary precautions and accuracy, which was farther secured by the large quantities of the two airs consumed. 3. "Let us now consider what obviously happens in the case of the deflagration of the inflammable and dephlogisticated air. These two kinds of air unite with violence, they become red-hot, and upon cooling totally disappear. When the vessel is cooled, a quantity of water is found in it equal to the weight of the air employed. This water is then the only remaining product of the process, and water, light, and heat, are all the products," unless there be some other matter set free which escapes our senses. "Are we not then authorised to conclude, that water is composed of dephlogisticated air and phlogiston, deprived of part of their latent or elementary heat; that dephlogisticated or pure air is composed of water deprived of its phlogiston, and united to elementary heat and light; and that the latter are contained in it in a latent state, so as not to be sensible to the thermometer or to the eye; and if light be only a modification of heat, or a circumstance attending it, or a component part of the inflammable air, then pure or dephlogisticated air is composed of water deprived of its phlogiston and united to elementary heat?" 4. "It 4. "It appears, that dephlogisticated water," or, which may be a better name for the basis of water and air, the element you call humor, "has a more powerful attraction for phlogiston than it has for latent heat, but that it cannot unite with it, at least not to the point of saturation, or to the total expulsion of the heat, unless it be first made red-hot," or nearly so. "The electric spark heats a portion of it red-hot, the attraction between the humor and the phlogiston takes place, and the heat which is let loose from this first portion heats a second, which operates in a like manner on the adjoining particles, and so continually until the whole is heated red-hot and decomposed." Why this attraction does not take place to the same degree in the common temperature of the atmosphere, is a question I am not yet able to solve; but it appears, that, in some circumstances, "dephlogisticated air can unite, in certain degrees, with phlogiston without being changed into water." Thus Dr. Priestley has found, that by taking clean filings of iron, which, alone, produce only inflammable air of the purest kind, and mercurius calcinatus per se, which gives only the purest dephlogisticated air, and exposing them to heat, in the same vessel, he obtained neither dephlogisticated nor inflammable air, "but in their place fixed air." Yet it is well known, that a mixture of dephlogisticated and inflammable air will remain for years in close vessels in the common heat of the atmosphere, without suffering any change, the mixture being as capable of deflagration at the end of that time as it was when first shut up. These facts the Doctor accounts for, by supposing that the two kinds of air, when formed at the same time in the same vessel, can unite in their nascent state; but that, when fully formed, they are incapable of acting upon one another, unless they are first Parts of Water and of Dephlogisticated Air. first set in motion by external heat. "Phlogisticated air seems also to be another composition of phlogiston and dephlogisticated air;" but in what proportions they are united, or by what means, is still unknown. It appears to me to be very probable, that fixed air contains a greater quantity of phlogiston than phlogisticated air does, because it has a greater specific gravity, and because it has more affinity with water. 5. "For many years I have entertained an opinion, that air was a modification of water, which was originally founded on the facts that in most cases, wherein air was actually made," which should be distinguished from those wherein it is only extricated from substances containing it in their pores, or otherwise united to them in the state of air, "the substances were such as were known to contain water as one of their constituent parts, yet no water was obtained in the processes," except what was known to be only loosely connected with them, such as the water of the crystallization of salts. "This opinion arose from a discovery," that the latent heat contained in steam diminished in proportion as the sensible heat of the water from which it was produced increased; or, in other words, "that the latent heat of steam was less when it was produced under a greater pressure, or in a more dense state, and greater when it was produced under a less pressure, or in a less dense state;" which led me to conclude, "that when a very great degree of heat was necessary for the production of the steam, the latent heat would be wholly changed into sensible heat; and that, in such cases, the steam itself might suffer some remarkable change. I now abandon this opinion in so far as relates to the change of water into air, as I think that may be accounted for on better principles." 6. "In 6. "In every case, wherein dephlogisticated air has been produced, substances have been employed, some of whose constituent parts have a strong attraction for phlogiston, and, as it would appear, a stronger attraction for that substance than humor has; they should, therefore, dephlogisticate the water" or fixed air, and the humor thus set free should unite to the matter of fire and light and become pure air. Dephlogisticated air is produced in great abundance from melted nitre. The acid of nitre has a greater attraction for phlogiston than any other substance is known to have; and it is also certain, that nitre, besides its water of crystallization, contains a quantity of water as one of its elementary parts, which water adheres to the other parts of the nitre with a force sufficient to enable it to sustain a red heat. When the nitre is melted, or made red-hot, the acid acts upon the water and dephlogisticates it; and the fire supplies the humor with the due quantity of heat to constitute it air, under which form it immediately issues. It is not easy to tell what becomes of the acid of nitre and phlogiston, which are supposed to be united," as they seem to be lost in the process. Dr. Priestley has lately made some experiments, with a view to ascertain this point. He distilled dephlogisticated air from pure nitre, in an earthen retort glazed within and without. He employed 2 oz = 960 grains of nitre: the retort was placed in an air furnace, and, by means of an intense heat, he obtained from the nitre in one experiment 787, and in another experiment 800 ounce measures of dephlogisticated air; and he found that, upon weighing the retort and nitre before and after the process, they had suffered a loss of weight equal to the weight of the air, and to the water of crystallization of the nitre, but nothing more. He remarked, that the air had a pungent smell, smell, which he could not divest it of by washing; and that the water in which the air was received had become slightly acid. I examined a portion of this water, which he was so kind as to send me, and found by it that the whole of the receiving water had contained the acid belonging to 2 drams = 120 grains of nitre. I also examined the residuum and the retort in which the distillation had been performed, and found the residuum highly alkaline, yet containing a minute quantity of phlogisticated nitrous acid. It had acted considerably upon the retort, and had dissolved a part of it, which was deposited in the form of a brownish powder, when the saline part was dissolved in water. This earthy powder I have not yet thoroughly examined, but have no doubt that it principally consists of the earth of the retort. This experiment, and all others tried in earthen vessels, leave us still at a loss to determine what becomes of the acid and phlogiston. They seem either to remain mixed with the air, in the form of an incoercible gas; or to unite with the alkali, or with the earth of the retort, in some manner so as not to be easily separated from them; or else they are imbibed by the retorts themselves, which are sufficiently porous to admit of such a supposition. All that appears to be conclusive from this experiment is, that above one half of the weight of the nitre was obtained in the form of dephlogisticated air; and that the residuum still contained some nitrous acid united to phlogiston. 7. Finding that the action of the nitre on the retort tended to prevent any accurate examination of the products, I had recourse to combinations of the nitrous acid with earths from which the dephlogisticated air is obtained with less heat than from nitre itself. As these processes have been particularly described by Dr. Priestley, by Mr. Scheele, and others, I shall not enter into any detail of them; but shall mention the general phænomena which I observed, and which relate to the present subject. The earths I used were magnesia alba, calcareous earth, and minium or the red calx of lead. I dissolved them in the respective experiments in nitrous acid dephlogisticated by boiling, and diluted with proper proportions of water. I made use of glass retorts, coated with clay; and I received the air in glass vessels, whose mouths were immersed in a glazed earthen basin, containing the smallest quantity of water that could be used for the purpose. As soon as the retort was heated a little above the heat of boiling water, the solutions began to distil watery vapours containing nitrous acid. Soon after these vapours ceased, yellow fumes, and in some of the cases dark red fumes, began to appear in the neck of the retort; and at the same time there was a production of dephlogisticated air, which was greater in quantity from some of these mixtures than from others, but continued in all of them until the substances were reduced to dryness. I found, in the receiving water &c., very nearly the whole of the nitrous acid used for their solution, but highly phlogisticated, so as to emit nitrous air by the application of heat; and there is reason to believe, that with more precaution the whole might have been obtained. 8. As the quantity of dephlogisticated air produced by these processes did not form a sufficient part of the whole weight, to enable me to judge whether any of the real acid entered into the composition of the air obtained, I ceased to pursue them further, having learned from them the fact, that however much the acid and the earths were dephlogisticated before the solution, the acid always became highly phlogisticated in the process. Parts of Water and of Dephlogisticated Air. In order to examine whether this phlogiston was furnished by the earths, some dephlogisticated nitrous acid was distilled from minium till no more acid or air came over. More of the same acid was added to the minium as soon as it was cold, and the distillation repeated, which produced the same appearance of red fumes and dephlogisticated air. This operation was repeated a third time on the same minium, without any sensible variation in the phenomena. The process should have been still farther repeated, but the retort broke about the end of the third distillation. The quantity of minium used was 120 grains, and the quantity of nitrous acid added each time was 240 grains, of such strength that it could dissolve half its weight of mercury, by means of heat. It appears from this experiment, that unless minium be supposed to consist principally of phlogiston, the source of the phlogiston, thus obtained, was either the nitrous acid itself, or the water with which it was diluted; or else that it came through the retort with the light, for the retort was in this case red-hot before any air was produced; yet this latter conclusion does not appear very satisfactory, when it is considered, that in the process wherein the earth made use of was magnesia, the retort was not red-hot, or very obscurely so, in any part of the process; and by no means luminous, when the yellow and red fumes first made their appearance. 9. As the principal point in view was to determine whether any part of the acid entered into the composition of the air, I resolved to employ some substance which would part with the acid in a moderate heat, and also give larger quantities of air than had been obtained in the former processes. Mercury was thought a proper substance for this purpose. 240 grains of mercury were put into a glass retort with 480 grains Mr. Watt's Thoughts on the constituent of diluted dephlogisticated nitrous acid, which was the quantity necessary to dissolve the whole of the mercury, a gentle heat was applied, and as soon as the common air contained in the retort was dissipated, a vessel was placed to receive the nitrous air proceeding from the solution, which was 16 ounce measures. When it had ceased to give nitrous air, the neck of the retort became hot from the watery steams of the acid. The air receiver was taken away, and a common receiver was luted on, with a little water in it, to condense the vapours, and a quantity of dilute, but highly phlogisticated, acid was caught in the receiver. When the watery vapours had nearly come over, and yellow fumes appeared in the neck of the retort, the common receiver was removed, and the air receiver replaced; about four ounces of very strong nitrous air passed up immediately, the fumes in the retort became red, and dephlogisticated air passed up, which, uniting with the nitrous air in the receiver, produced red fumes in the receiver; and the two kinds of air acting upon one another, their bulk was reduced to half of an ounce measure. At this period the fumes in the retort were of a dark red colour, and dephlogisticated air was produced very fast. After a short time, some orange-coloured sublimate appeared in the upper part of the retort, and extended a little way along its neck, the red colour of the fumes gradually disappeared, and the neck of the retort became quite clear. At the same time that this happened, small globules of mercury appeared in the neck of the retort, and accumulated there until they ran down in drops. The production of the air was now very rapid, and accompanied with much of the white cloud or powdery matter, which passed up with the air into the receiver, and mixed with the water, but did not dissolve in it. After giving about 36 ounce measures of dephlogisticated air, it suddenly ceased to give any more; and the retort being cooled, the bulb was found to be quite empty, excepting a small quantity of black powder, which, on being rubbed on the hand, proved to be mostly running mercury. The orange-coloured sublimate was washed out of the neck of the retort, and what running mercury was in it was separated, and added to that which had run down into the basin among the water. The whole fluid mercury, when dried, weighed 218 grains; therefore 22 grains remained in the form of sublimate, which, I believe, would also have been reduced if I could have applied heat in a proper manner to the neck of the retort, as some of it, to which heat could be applied, disappeared. 10. The 16 ounce measures of nitrous air, which had been produced in the solution of the mercury, and had remained confined by water in the receiver, was converted into nitrous acid by the gradual admission of common air, and was taken up by the water; this water was added to that in the basin, which had served to receive the dephlogisticated air. The whole quantity was about two quarts, was very acid to the taste, and sparkling with nitrous air. It was immediately put into bottles, and well corked, until it had lost the heat gained in the operation. In order to determine the quantity of acid in the receiving water and in the sublimate, I dissolved, first, alkali of tartar in water, and filtered the solution. 352 grains of this alkaline solution saturated 120 grains of the nitrous acid I had employed to dissolve the mercury, and 1395 grains of the same alkaline solution saturated the orange-coloured precipitate, and all the acid liquor obtained from the process: therefore we have the proportion as $352 : 120 :: 1395 : 475$, from which it appears, that all the acid employed was recovered again in the form of acid, excepting only five grains; a smaller a smaller quantity than what might reasonably be supposed to be lost in the process by the extreme volatility of the nitrous air. In order to ascertain the exact point of saturation, slips of paper, stained by the juice of the petals of the scarlet rose, were employed, which were the nicest test I could procure, as litmus will not shew the point of saturation of any liquor containing much phlogisticated nitrous acid, or even fixed air, but will turn red, and shew it to be acid, when the test of those leaves, violets, and some other of the like kind, will turn green in the same liquor, and shew it to be alkaline. But the exact point of saturation of so dilute a liquor is so very difficult to ascertain, than an error might easily be committed, notwithstanding the attention bestowed upon it. Supposing this experiment to be unexceptionable, the conclusions which may be drawn from it are very favourable to the hypothesis I endeavour to support. Thirty-six ounce measures of dephlogisticated air were obtained, and only five grains of a weak nitrous acid were lost in the process. Two hundred and eighteen grains of mercury out of two hundred and forty were revived, and all the dephlogisticated nitrous acid employed is found to be highly phlogisticated in the process. It appears, that the nitrous acid does not enter into the composition of dephlogisticated air; it seems only to serve to absorb phlogiston from the watery part of the mercurial nitre. 11. As this last process proved very tedious and complicated on account of the necessity of ascertaining the quantity of acid in the receiving water, by means of an alkali which afforded a double source of error in the point of saturation, I resolved to try the distillation of dephlogisticated air from cubic nitre in a glass vessel, and to draw from it only such a quantity of air as it would yield without acting much upon the retort, which latter latter circumstance is essentially necessary to be attended to. An ounce of the crystals of mineral alkali were dissolved in nitrous acid, and the mixture brought to an exact saturation by the test of litmus; 30 ounce measures of air were distilled from it, which, during the latter part of the process, was accompanied with slightly yellow fumes; the receiving water was found to be acid, and the residuum alkaline. The residuum being dissolved in the receiving water, the solution was neutral, or very nearly so, by every test; for in this case litmus might be used, as the acid was very slightly phlogisticated. On adding a few drops of a very dilute nitrous acid, the tests shewed the liquor to be acid. 12. Encouraged by the success of this experiment, I took an ounce = 480 grains of pure common nitre, and put it into a flint-glass retort, coated, which was placed in a furnace. It began to give air about the time it became red-hot, and during the latter part of the process this air was accompanied with yellowish fumes. I stopped the process when it had produced 50 ounce measures of air. The receiving water, and particularly the air, had a strong but peculiar smell of nitrous acid. The air was well washed with the receiving water, but was not freed from the smell. The receiving water, which was 50 ounces, was slightly acid, and the residuum alkaline. I dissolved the latter in the former, and found the mixture alkaline. 10 grains of weak nitrous acid were added to it, which saturated it, and 105 grains of this spirit of nitre was found to contain the acid of 60 grains of nitre; therefore the 10 grains contained the acid of 5.7 grains of nitre, which, by Mr. Kirwan's experiments is equal to two grains of real nitrous acid. We have, therefore, 34 grains weight of dephlogisticated air produced, and only two grains of real acid missing; and it is not certain certain that this quantity was destroyed, because some portion of the glass of the retort was dissolved by the nitre, and some part of the materials employed in making the glass being alkali, we may conclude, that the alkali of the nitre would be augmented by the alkali of that part of the glass it had dissolved. As the glass cracked into small pieces on cooling, and some part of the coating adhered firmly to it, the quantity of the glass that was dissolved could not be ascertained. From this experiment it appears, that if any of the acid of the nitre enters into the composition of the dephlogisticated air, it is a very small part; and it rather seems, that the acid, or part of it, unites itself so firmly to the phlogiston as to lose its attraction for water. 13. "The vitriolic salts also yield dephlogisticated air by heat; and in these cases the dephlogisticated air is always attended with a large quantity of vitriolic acid air or sulphureous vapour," even when the salts used are not known to contain any phlogistic matter. Mr. Scheele mentions his having obtained dephlogisticated air from manganese dissolved in acid of phosphorus, and also from the arsenical acid: from whence it appears, that these acids, or perhaps any acid which can bear a red heat, can concur to the production of dephlogisticated air. It is necessary to remark, that no experiments have been yet published shewing that dephlogisticated air can be produced from salts formed by the muriatic acid. The acids which produce salts suitable for this purpose, have all a strong affinity with phlogiston; and the marine acid has either a very small affinity with it, or else is already saturated with it, at least so far saturated as not to be able to attract it from the humor. 14. "The dephlogisticated air obtained from the pure calces of metals may be attributed to the calces themselves, attracting the phlogiston from water which they have imbibed from Parts of Water and of Dephlogisticated Air. "the atmosphere, or from dephlogisticating the fixed air which they are known to contain." It is very probable, that the dephlogisticated air extruded from growing vegetables may be owing to their dephlogisticating the water they grow in; but it appears more probable, that the plants have a power of dephlogisticating the fixed, or phlogisticated, air of the atmosphere. "When dephlogisticated and nitrous air are mixed, the dephlogisticated air seizes part of the phlogiston of the nitrous air." The water contained in the nitrous air, and the other part of the phlogiston, unite with the nitrous acid, which then assumes a liquid form, or at least that of a dense vapour; "and that part of the latent heat of the two airs not essential to the new combination is set at liberty *." In the combustion of sulphur the same thing happens, but in a greater degree; for the vitriolic acid, having a much weaker attraction for phlogiston than air has, abandons it almost entirely to the latter, which is thereby converted into water, and in that form attracts the vitriolic acid, and reduces it to a liquid state. The same reasoning may be applied to the combustion of phosphorus, which is attended with similar effects. * I cannot take upon me to determine, from any facts which have come to my knowledge, whether any part of the dephlogisticated air employed in this experiment is turned into fixed air; but I am rather inclined to think that some part is, because the quantity of heat, which is separated by the union of the two airs, does not seem to be so great as that which is separated when the dephlogisticated air is wholly changed into water: yet some water appears to be formed, because when the mixture is made over mercury, the solution of the mercury in the nitrous acid assumes a crystallized form, which, however, may be due to the watery part of the nitrous air. 15. I shall not make, at present, any further deductions from what I myself consider still in the light of a conjectural hypothesis, which I have perhaps dwelt upon too long already. I shall only beg your attention to some general reasoning on the subject; which, however, may possibly serve more to shew the uncertainty of other systems on the constituent parts of air, than the certainty of this. Some of those systems suppose dephlogisticated air to be composed of an acid and something else, some say phlogiston. If an acid enters into the composition of it, why does not that acid appear again when the air is decomposed, by means of inflammable air and heat? And why is the water which is the product of this process pure water? And if an acid forms one of its constituent parts, why has nobody been able to detect any difference in the dephlogisticated air, made by the help of different acids, when compared with one another, or with the air extruded by vegetables? These airs, of such different origins, appear to be exactly the same. And if phlogiston be a constituent part of air, why does it attract phlogiston with such avidity? Some have, on the other hand, contended that air is composed of earth, united to acids or phlogiston, or to both, or to some other matter. Here we must ask, what earth it is which is one of the component parts of air? All earths which will unite with the nitrous or vitriolic acids, and with some others, such as the phosphoric and the arsenical acids, will serve as bases for the formation of air, and the air produced from all of them appears by every test to be the same, when freed from accidental impurities. To this argument it is answered, that it is not any particular species of earth which is the basis of air, but elementary or simple earth, which is contained in all of them. If this were the matter matter of fact, would not that earth be found after the decomposition of the air? Mr. Scheele has formed an hypothesis on this subject, in which he supposes heat to be composed of dephlogisticated air united to phlogiston, and that this combination is sufficiently subtile to pass through glass vessels. He affirms, that the nitrous and other acids, when in an ignited state, attract the phlogiston from the heat, and set the dephlogisticated air at liberty; but he does not seem to have been more successful than myself in explaining what becomes of the acid of nitre and phlogiston in the case of the decomposition of nitre by heat. And since we know, from the late experiments, that water is a composition of air, or more properly, humor and phlogiston, his whole theory must fall to the ground, unless that fact be otherwise accounted for, which it does not seem easy to do. 16. To return to the experiment of the deflagration of dephlogisticated and inflammable air, "it appears from the two airs becoming red-hot on their union, that the quantity of heat contained in one or both of them, is much greater than that contained in steam; because, for the first moments after the explosion, the water deposited by the air remains in the form of steam, and consequently retains the latent heat due to that modification of water. This matter may be easily examined by firing the mixture of dephlogisticated and inflammable air in a vessel immersed in another vessel containing a given quantity of water of a known heat, and after the vessel in which the deflagration is performed is come to the same temperature with the water in which it is immersed, by examining how much heat that water has gained, which being divided by the quantity of water produced by the decomposition of the airs, will give the whole quantity of eleme- Mr. Watt's Thoughts on the constituent "tary or latent heat which that water had contained, both as air and as steam; and if from that quantity we deduct the latent heat of the steam, the remainder will be the latent or elementary heat contained more in air than in steam." This experiment may be made more completely by means of the excellent apparatus which Messrs. Lavoisier and De la Place have contrived for similar purposes. Until direct experiments are made, we may conclude, from those which have been made by the gentlemen just named, on the decompositions of air by burning phosphorus and charcoal, that the heat extricated during the combustion of inflammable and dephlogisticated air is much greater than it appears to be; for they found that one Paris ounce (= 576 Parisian grains) of dephlogisticated air, when decomposed by burning phosphorus, melted 68,634 ounces of ice; and as, according to another of their experiments, ice, upon being melted, absorbs 135° of heat, by Fahrenheit's scale, each ounce of air gave out $68,634 \times 135^\circ = 9265^\circ,590$; that is to say, a quantity of heat which would have heated an ounce of water, or any other matter which has the same capacity for receiving heat as water has, from $32^\circ$ to $9265^\circ$: a surprising quantity! (It is to be understood, that all the latent heats mentioned herein are compared with the capacity of water). And when an ounce of dephlogisticated air was changed into fixed air, by burning charcoal, or by the breathing of animals, it melted 29,547 oz. of ice; consequently we have $29,547 \times 135^\circ = 3988^\circ,845$. the quantity of heat which an ounce of dephlogisticated air loses when it is changed into fixed air. By the heat extricated during the detonation of one ounce of nitre with one ounce of sulphur, 32 ounces of ice were melted; and, by the experiment I have mentioned of Dr. Priestley's (6), it appears that nitre nitre can produce one half of its weight of dephlogisticated air. When the nitre and sulphur are kindled, the dephlogisticated air of the nitre unites with the phlogiston of the sulphur, and sets its acid free, which immediately unites to the alkali of the nitre, and produces vitriolated tartar. The dephlogisticated air, united to the phlogiston, is turned into water, part of which is absorbed by the vitriolated tartar, and part is dissipated in the form of vapours, or unites to the nitrous air, or other air, produced in the process. As half an ounce of dephlogisticated air is, in this process, united by inflammation to a quantity of phlogiston sufficient to saturate it, and no fixed air is produced, it should melt a quantity of ice equal to the half of what was melted by the combination of an ounce of air with phlogiston in burning phosphorus; that is, it should melt $34,317$ ounces of ice; and we find, by Messrs. Lavoisier and De la Place's experiment, that it actually melted $32$ ounces of ice: the small difference may be accounted for by supposing, that the heat produced by the combustion might not be quite so great as that Dr. Priestley employed in his experiment; or that the nitre might be less pure, and consequently not so much air formed. The two facts, however, agree near enough to permit us to conclude, that dephlogisticated air, in uniting to the phlogiston of sulphur, produces as much heat as it does in uniting with the phlogiston of phosphorus. 17. According to Dr. Priestley's experiments, dephlogisticated air unites completely with about twice its bulk of the inflammable air from metals. The inflammable air being supposed to be wholly phlogiston, and being $\frac{1}{9}$ of the weight of an equal bulk of dephlogisticated air, and being double in quantity, will be $\frac{1}{4}$ of the weight of the dephlogisticated air. Mr. Watt's Thoughts on the constituent it unites with. Therefore one ounce (576 grains) of dephlogisticated air, will require 120 grains of inflammable air, or phlogiston, to convert it into water. And supposing the heat extricated by the union of dephlogisticated and inflammable air to be equal to that extricated by the burning of phosphorus, we shall find, that the union of 120 grains of inflammable air with 576 grains of dephlogisticated air, extricates 9265° of heat. 18. In the experiment on the deflagration of nitre with charcoal, by Messrs. Lavoisier and De la Place, an ounce of nitre and one third of an ounce of charcoal melted twelve ounces of ice. Supposing the ounce of nitre to have produced half an ounce of dephlogisticated air, it ought to have consumed 0.1507 ounces of charcoal, and should have melted 14,773 ounces of ice; and I suppose it fell short of its effect by the heat not being sufficiently intense to decompose the nitre perfectly. 19. By the above gentlemen's experiment an ounce of charcoal required for its combustion 3,3167 ounces of dephlogisticated air, and produced 3,6715 ounces of fixed air; therefore there was united to each ounce of air, when changed into fixed air, 61.5 grains of phlogiston, and 3988° of heat were extracted. It appears by these facts, that the union of phlogiston, in different proportions, with dephlogisticated air, does not extricate proportional quantities of heat. For the addition of 61.5 grains produces 3988°, and the union of 120 grains produces 9265°. This difference may arise from a mistake in supposing the specific gravity of the inflammable air Dr. Priestley employed to have been only $\frac{1}{9.6}$ of that of dephlogisticated air; for if it be supposed that its specific gravity was a little more than $\frac{1}{7}$ of that of the dephlogisticated air, then equal additions of phlogiston would have have produced equal quantities of heat*: this matter should therefore be put to the test of experiment, by deflagrating dephlogisticated air with inflammable air of a known specific gravity, or by finding how much dephlogisticated air is necessary for the combustion of an ounce of sulphur, the quantity of phlogiston in which has been accurately determined by Mr. Kirwan; or by finding the quantity of phlogiston in phosphorus, the quantity of dephlogisticated air necessary for its decomposition being known from Messrs. Lavoisier and De la Place's experiments. On considering these latter gentlemen's experiments on the combustion of charcoal, a difficulty arises, to know what became of the remainder of the ounce of charcoal; for the dephlogisticated air, in becoming fixed air, gained only the weight of 0.3548, or about \( \frac{1}{3} \) of an ounce; about \( \frac{2}{3} \) of an ounce are therefore unaccounted for. The weight of the ashes of an ounce of charcoal is very inconsiderable; and, by some experiments of Dr. Priestley's, charcoal, when freed from fixed air, and other air which it imbibes from the atmosphere, is almost wholly convertible into phlogiston. The cause of this apparent loss of matter, I doubt not, these gentlemen can explain satisfactorily, and very probably in such a manner as will throw other lights on the subject. * Or it may arise from my being mistaken, in supposing that the same quantity of heat is disengaged by the union of dephlogisticated air with phlogiston, in the form of inflammable air, as is by its union with the phlogiston of phosphorus or sulphur; and there appears to be some reason why there should not; because in these latter cases the water, being united to the acids, cannot retain so much elementary heat as it can do when left in the form of pure water, which is the case when the inflammable air is used. It is also worthy of enquiry, whether all the amazing quantity of heat let loose in these experiments was contained in the dephlogisticated air; or whether the greatest portion of it was not contained in the phlogiston or inflammable air. If it was all contained in the dephlogisticated air, "the general rule is not fact, that elastic fluids are enlarged in their dimensions in proportion to the quantity of heat they contain;" because then, inflammable air, which is ten times the bulk of dephlogisticated air, must be supposed to contain no heat at all; "and it is known, from some experiments of my friend Dr. Black's, and some of my own, that the steam of boiling water, whose latent and sensible heat are only 110°, reckoning from 60°, or temperate, is more than twice the bulk of an equal weight of dephlogisticated air." It seems, however, reasonable to suppose, that the greater quantity of heat should be contained in the rarer fluid. It may be alledged, that in proportion to the quantity of phlogiston that is contained in any fluid, the quantity of heat is lessened. But if we reason by analogy, the attraction of the particles of matter to one another in other cases is increased by phlogiston, and "bodies are thereby rendered specifically heavier;" and we know of no other substance besides heat which can be supposed to separate the particles of inflammable air, and to endow it with so very great an elastic power, and so small a specific gravity. On the other hand, if a great quantity of elementary heat be allowed to be contained in inflammable air, on account of its bulk, the same reasoning cannot hold good in respect to the phlogiston of phosphorus, sulphur, charcoal, &c. But all these substances contain other matters besides phlogiston and heat. The acids in the sulphur and Parts of Water and of Dephlogisticated Air, and phosphorus, and the alkali and earth in charcoal, may attract the phlogiston so powerfully that the heat they contain may not be able to overcome the adhesion of their particles, until, by the effect of external heat, they are once removed to such a distance from one another as to be out of the sphere of that kind of attraction*. If it be found to be a constant fact, that equal additions of phlogiston to dephlogisticated air do not extricate equal quantities of heat, that may afford the means of finding the quantities of heat contained in phlogiston and dephlogisticated air respectively, and solve the problem. Many other ideas on these subjects present themselves; but I am not bold enough to trouble you, or the public, with any speculations, but such as I think are supported by uncontroverted facts. I must therefore bring this long letter to a conclusion, and leave to others the future prosecution of a subject which, however engaging, my necessary avocations prevent me from pursuing. I cannot however conclude, without acknowledging my obligations to Dr. Priestley, who has given me every information and assistance in his power, in the course of my enquiries, with that candour and liberality of sentiment which distinguish his character. I return you my thanks for the obliging attention you have paid to this hypothesis; and remain, with much esteem, &c. JAMES WATT. * On the whole, this question seems to involve so many difficulties, that it cannot be cleared up without many new experiments.