FIRST PRINCIPLES OF CHEMISTRY. UT potero, explicabo ; nee tamen ut Pytliius Apollo, certa ut sint et fixa, qua? dixero : sed, ut homunculus, probabilia conjectura sequens. Tusc. QU^EST. Lib. I. c. 9. AN ATTEMPT TO ESTABLISH THE FIRST PRINCIPLES OF CHEMISTRY BY EXPERIMENT. BY THOMAS THOMSON, M. D., REGIUS PROFESSOR OF CHEMISTRY IN THE UNIVERSITY OF GLASGOW ; F. R. S. LONDON AND EDINBURGH ; FELLOW OF THE LINN^EAN SOCIETY; OF THE GEOLOGICAL SOCIETY; OF THE CAMBRIAN SOCIETY FOR THE ENCOURAGEMENT OF GEOLOGY, MINERALOGY, AND NATURAL HISTORY ; OF THE IMPERIAL MEDICO-CHIRURGICAL AND PHARMACEUTI- CAL SOCIETIES OF ST. PETERSBURGH ; OF THE ROYAL ACADEMY OF SCIENCES OF NAPLES; OF THE MINERALOGICAL SOCIETY OF DRESDEN ; OF THE MINING SOCIETY OF FRANKFORT ON THE MAINE; OF THE CAESARIAN NATURAL HISTORY SOCIETY OF MOSCOW; OF THE LITERARY AND PHILOSOPHICAL, AND NATURAL HISTORY SOCIETIES OF NEW YORK ; &C. IN TWO VOLUMES. VOL. I. LONDON: PRINTED FOR BALDWIN, CRADOCK, AND JOY. 1825. GLASGOW: ANDREW & JOHN 'M. DUNCAN, Printers to the University. TO y MR. JOHN DALTON, F.R.S., &c., M. GAY-LUSSAC, F.R.S., MEM. DE L' INSTITUT, &c., SIR HUMPHREY DAVY, BART., P.R.S., &c., JACOB BERZELIUS, M.D., F.R.S., KONG. VET. ACAD, L., &c., WILLIAM HYDE WOLLASTON, M.D., V.P.R.S., &c., AND WILLIAM PROUT, M.D., F.R.S., &c., TO WHOM THE ATOMIC THEORY IS INDEBTED FOR ITS ORIGINAL DEVELOPMENT AND FOR ALL THE IMPROVEMENTS WHICH IT HAS HITHERTO RECEIVED; THIS ATTEMPT TO BRING THE SUBJECT TO A STILL GREATER DEGREE OF SIMPLICITY AND PRECISION AS TO THE BEST JUDGES OF ITS IMPORTANCE, AND OF THE DIFFICULTIES TO BE SURMOUNTED, AND MOST DISPOSED, IN CONSEQUENCE, TO VIEW WITH INDULGENCE ANY UNAVOIDABLE IMPERFECTIONS WHICH THEY MAY PERCEIVE IN IT IS PRESENTED BY THEIR SINCERE FRIEND AND HUMBLE SERVANT, THE AUTHOR. 542 TO THE STUDENTS OF MEDICINE AND CHEMISTRY IN THE UNIVERSITY OF GLASGOW. GENTLEMEN, A CONSIDERABLE portion of the work which I now offer to the public, is already known to such of you as attended my course of lectures during the last three or four sessions. For I was in the habit of exhibit- ing to you a table of the atomic weights of the different bodies, as I succeeded in determining them, and of stating to you the experiments by means of which the results had been attained. I observed with pleasure the interest which the better informed and more intelligent students took in these investigations while the numerous ap- plications from various quarters for copies of my atomic tables, satisfied me that many of my students had un- derstood the discussions, and had fully appreciated their importance. a 4 Vlll I have been induced to lay the result of my investiga- tions before the public, chiefly for two reasons, with which I think it requisite to acquaint you. The first of these is, the great advantage which medical practitioners will derive from a knowledge of the atomic weights of bodies, and of the weights of the integrant particles of the salts, &c., which they have occasion to employ in their prescriptions. This knowledge will be easily acquired by a perusal of the following pages; and il will enable those who possess it to avoid some very awkward blunders, into which I have observed too many practitioners, even of considerable celebrity, frequently to fall, to the no little inconvenience of their patients. Indeed, till the atomic theory was pretty completely investigated, even chemists themselves were not exempt from such mis- takes; and I could point out, if I thought proper, some remarkable examples of men who would take it amiss if the world did not give them credit for possessing much chemical science, falling into mistakes that no student of mine, who took the trouble to follow and understand the lectures delivered on chemistry in the College of Glas- gow, would run any hazard of committing. It was a wish to enable my students to recollect with ease this im- portant collection of indispensable information, and to put it in the power of other medical students, who are strangers to the lectures which I deliver, to avail them- selves of the benefits to be derived from the facts and the tables in the following pages, that constituted one of IX the inducements which led me to lay the work before the world. The second consideration which induced me to resolve upon publishing my atomical investigations, was one which more particularly concerns the medical students in the University of Glasgow. It was a wish on my part to shorten the discussions connected with the atomic theory, and with the composition of the salts, which I have hitherto thought it incumbent on me to introduce into my course of lectures. Henceforth, I shall have it in my power to refer you to the present publication for the proofs of the different fundamental points of importance. This will leave me much more time for illustration, and will enable me to make my future courses of lectures more entertaining and varied, without any diminution of the quantity of information which I have hitherto con- sidered it as a point of duty to communicate. I shall be able too, hereafter, to enter more into detail respecting those parts of chemistry which are more intimately con- nected with the theory of medicine ; for I need not observe to those gentlemen who have paid attention to the subject, that by far the most likely means of improving physiology, is a cautious application of chemistry to the investigation of the different constituents of the human body. An accurate knowledge of the chemical composition of every organ, and of the alterations produced on that organ by disease, would probably throw new and unex- pected light upon the nature and treatment of many diseases. Physiology hitherto has been handled almost exclusively by the anatomists. These gentlemen have acted with a zeal and industry that cannot be too much admired. They have examined the anatomical structure of every organ of the human body; ascertained the changes to which it is liable from disease ; and compared it with the analogous organs in all the inferior animals. A great deal of important knowledge has been the result of this active examination : yet, after the anatomists have exhausted all their ingenuity, it must be admitted that our knowledge both of the organs and functions is still very imperfect. A new and more subtile species of anatomy remains still to be applied. Where the labours of the anatomist terminate, those of the chemist should begin. This chemical investigation of the animal body may be just said to be commencing at present : for it was not till the atomic theory was brought to considerable simplicity and perfection, that such an investigation was possible. It is easy to see that it must contribute pro- digiously to the advancement of physiological knowledge. I have been long anxious to lay before my students the general results already attained by these chemical investi- gations. But the preliminary discussions into which I was obliged to enter, left us too little time at the end of the session, to enter upon them ; and after having been at no little pains in collecting and arranging the different facts from sources not easily accessible to students in XI general, f have hitherto had the mortification to be obliged to pass them over, or at least to treat them in so general a way that my views could scarcely be under- stood. By referring my students hereafter to the facts contained in the following sheets, I shall have it in my power to shorten considerably the elementary part of my course, and be enabled to introduce at least a certain number of physiological lectures ; by pointing out the general facts already ascertained, and the physiological consequences legitimately deducible from these facts. This is a branch of medical knowledge which has been much more attended to on the continent than in Great Britain. But from the intelligence and the zeal which the medical students in this country in general possess, I am satisfied that if the subject were once fairly started, it would be- come a favourite object of pursuit, and that it would advance to perfection with a degree of rapidity of which we can at present form but an imperfect idea. PREFACE. THE subject treated of in the following pages is of more importance than any other which can at the present mo- ment occupy the attention of the chemist ; since it is no other than an attempt to establish the fundamental princi- ples of the science upon a sure and steady foundation. The analytical labours of the chemist and mineralogist depend upon these principles, and can only be accurate after their truth has been fully established. All the operations of the chemical manufacturer are equally dependant on them, and can only be prosecuted with perfect certainty and success by those who are acquainted with them. Several chemical manufacturers in this coun- try have already availed themselves of the atomic theory to rectify their processes. This they have done to such an extent, that unless all the other manufacturers follow their example, they will find it impossible to compete with their more skilful rivals. One of the great objects of the present work is to reduce the whole doctrine of atoms to the utmost degree of simplicity and accuracy; so as to put it in the power of chemical manufacturers in general, XIV PREFACE. to avail themselves of it in order to bring their processes to the requisite degree of precision. Nor will the atomic theory prove less important to apothecaries and druggists, and preparers of medicines. It will point out the exact proportions of every substance to be used, and will enable them to attain their object in the most simple and direct manner, without any unprofit- able waste by employing more or less of the different substances than is just requisite for the intended object. To them, and indeed to all medical men and medical students, the knowledge of the atomic theory is of the utmost consequence, and they will find the present publi- cation of no little service to them. The tables contained in this work ought to occupy a place in every laboratory, and to lie upon the shop of every druggist, that he may have it in his power to have recourse to them to regulate all his processes. The work which I now present to the public contains the result of many thousand experiments, conducted with as much care and precision as it was in my power to employ. They have occupied the whole of my time (except what was necessarily devoted to the duties of my situation,) for the last five years. All those experiments which I considered as fundamental, were repeated so often, and varied in so many ways, that I repose the most perfect confidence in their accuracy. I may therefore, PREFACE. XV perhaps, venture to express a hope, that the reader will not immediately judge them to be inaccurate, because they do not exactly coincide with the analytical results of other chemists. The processes of chemists have been so much rectified and improved of late years, and so much care is now generally bestowed upon experiments, that we may rely upon the result (provided the data employed be accurate) certainly within less than one per cent. I believe indeed, that the error in many of Berzelius' analy- ses does not exceed one thousandth part. But it is much more difficult to obtain substances in a state of complete purity than chemists in general are aware : it was in reducing the different salts which I employed, to the greatest possible degree of purity, that the greatest part of my time was wasted. I have in all cases in which it was in my power, deduced the atomic weights of bodies from the rigid analysis of the neutral salts into which they enter; because it is much easier to obtain neutral salts pure, than any of the metallic bodies which constitute their bases. Indeed, not a few of the metals have never yet been exhibited in a state of absolute purity. Berzelius has in general deduced the atomic weights of metals from experiments on these bodies themselves. This I be- lieve to be the principal reason why his atomic weights so frequently differ a little from mine. For example, the atom of zinc deduced from his experiments is 4*03225 ; while the atomic weight obtained by me from the analysis of sulphate of zinc is 4'25. The reason of this difference I 16 XVI PREFACE. believe to be that the sulphate of zinc which I used was a pure salt, while zinc in the metallic state is never quite free from iron and some other impurities, all of which have a tendency to diminish its apparent atomic weight. - As the analysis of sulphate of zinc constitutes in some measure the foundation of the whole superstructure which I have endeavoured to rear, I made the analysis of it with the greatest possible care. My first analysis of it was inserted in the Annals of Philosophy, (second series) I. 247. But by some inadvertency, the quantity of water is stated at six atoms instead of seven, which is the true quantity. I cannot very well account for this mistake; for on looking back to my original analysis, I find it merely stated that 17 grains of sulphate of zinc were just decomposed by 13*25 grains of chloride of barium. No- thing is said about the state of the salt, which would enable us to account for the absence of an atom of water. But I have repeated the analysis of sulphate of zinc many times since the publication of that paper, and have made several of my practical students repeat it, and the uniform result has been seven atoms of water, and not six, MS I there stated. If solutions of 17 grains of sulphate of zinc and 13*25 of chloride of barium be mixed, we shall find that the supernatant liquid, after the deposition of the sulphate of barytes, will become milky when mixed with sulphate of soda, showing that it still retains barytes, PREFACE. XV11 and consequently, that 17 grains of sulphate of zinc do not contain so much as 5 grains of sulphuric acid. I have been in the habit, ever since the atomic numbers given in the present work were determined, to calculate the experiments of other chemists by substituting for the data which they employed, the more correct data which I myself possessed, and I have frequently found that the experiments, when so calculated, came much nearer the truth than the experimenters themselves were aware of. This was remarkably the case with respect to Mr. Fara- day's experiments to determine the constituents of the chlorides of carbon, for a knowledge of which we are indebted to that very ingenious and accurate chemist. The imperfect data which he employed has given his experiments the appearance of greater inaccuracy than the truth : for when the specific gravities and atomic numbers given in the following pages are substituted for those which he used, his experiments almost coincide with the theoretic quantities. In this country, the atomic weights which I here pre- sent to the public, have been already very generally adopted, so far as they are known. But on the continent, the numbers and tables of Berzelius are in common use. Berzelius' numbers are in general very near approxima- tions to the truth ; though I am persuaded that in very few instances he has actually reached it. His tables VOL. I. b XV111 PREFACE. therefore, may be used without the risk of any great error, except those of the arseniates and phosphates, where the error is considerable ; but they consist of much more complex numbers than mine, and the atomic weights are all kept out of view by his method of giving the constituents in 100 parts of the compound bodies. On the score of facility and simplicity, I consider my numbers as much preferable : indeed, the theory of Berzelius is so complicated and intricate, that it would be surprising if it were a true representation of what takes place in nature. A few of my determinations of the atomic weights of various bodies, were published in the Annals of Philo- sophy, during the years 1820 and 1821. But on per- ceiving that I had occasionally fallen into mistakes by stating results deduced from particular views of a subject, before I had been able to consider it in all its bearings, I thought it safer to refrain, till I had brought the in- vestigation to a conclusion, and then to lay the whole at once before the public. This was the reason that induced me to discontinue my original plan of publishing my labours in periodical essays, in proportion as the ato- mic numbers were determined. ADVERTISEMENT. I HAVE been led by my desire of avoiding circumlocution, to make use of some terms in the following work, which may require explanation. 1. When the simple name of a compound occurs, it is to be understood that it is a compound of one atom of each constituent. Thus, oxide of zinc is a compound of one atom zinc and one atom oxygen. Sulphuret of bis- muth, a compound of one atom bismuth and one atom sulphur. Sulphate of potash, a compound of one atom sulphuric acid and one atom potash. 2. When the acid which a salt contains is doubled, the doubling is denoted by prefixing the Latin word bis to the name. Thus, bisulphate of potash means a compound of two atoms sulphuric acid and one atom of potash. Bi- borate of soda is a compound of two atoms boracic acid and one atom soda. 3. When the base is doubled, this is denoted by pre- fixing the Greek syllable dis. Thus, disulphate of alu- mina is a compound of one atom sulphuric acid and two atoms alumina. 9 XX [ ADVERTISEMENT. 4. Three, four, &c. atoms of the acid are denoted by prefixing the Latin numerals, ter, quater,* &c. Three, four, &c. atoms of the base, by prefixing the Greek nume- rals, tris, tetrakis, &c. 5. When a metal is combined with oxygen in the pro- portion of two atoms metal to one atom oxygen, the com- pound is called a suboxide. Thus, suboxide of manganese is a compound of two atoms manganese and one atom oxygen, or, of one atom manganese and a half atom oxygen. All the other terms employed will be understood with- out any farther explanation. * The affix quadro having been already applied for this purpose by Dr, Wollaston, and having come into general use, I have thought it better to con- tinue it than to use guater. CONTENTS THE FIRST VOLUME. CHAP. PAGE I. Historical introduction, . * . . i II. Of the atomic theory, . 29 III. Of the specific gravity of oxygen and hydrogen gases, 49 SECT. 1. Of the composition of oxide of zinc, . 51 2. Of the specific gravity of oxygen gas, . 62 3. Of the specific gravity of hydrogen gas, 67 IV. Of the atomic weights and specific gravities of chlo- rine and iodine, . . 77 SECT. 1. Atomic weight of chlorine, and specific gravity of chlorine gas, ib. 2. Of the atomic weight of iodine, . . 88 V. Of the atomic weight of azote, and the specific gravity of azotic gas, . . Qb VI. Of the atomic weights of the acidifiable combustibles, 133 SECT. 1. Of the atomic weight of carbon, . ~ 134 2. Of the atomic weight of boron, . 157 3. Of the atomic weight of silicon, . . 167 AN ATTEMPT TO ESTABLISH THE FIRST PRINCIPLES OF CHEMISTRY BY EXPERIMENT. CHAP. I. HISTORICAL INTRODUCTION. CHEMISTS appear at all times to have admitted that the constituents of compound bodies are constant, whether we consider the elements themselves, or their proportions. But it was not till Chemistry had made considerable progress, that any thing like an analysis of the salts could be attempted. Homberg, so far as I know, was the first person who thought of determining the composition of these bodies. He contrived a hydrometer for ascertaining the specific gravity of acids. This instrument was similar to the weighing bottle at present in common use for the same purpose. It enabled him to judge, from the specific gravity, how much stronger one acid VOL. I. A 2 HISTORICAL INTRODUCTION. was than another. But to estimate the quantity of real acid contained in each of those liquids, another process was necessary. And he pro- ceeded in this way : He took an ounce of salt of tartar (carbonate of potash of the shops) and saturated it wiih each of his acid liquids, deter- mining the weights of each necessary to produce the required effect. The liquid was then eva- porated to dryness, and the dry salt left was weighed. The augmentation of weight which the ounce of salt of tartar had sustained, marked, in his opinion, the weight of acid which had combined with it. For example : 1 ounce of salt of tartar when saturated with nitric acid becomes 1'45 ounces. Hence he concluded that saltpetre is a compound of Potash 100 * Nitric acid 45 145 In the same way he estimated the composition of muriate of potash, sulphate of potash, acetate of potash. It is unnecessary to give the results which he obtained, because Homberg, not having been aware that salt of tartar contained carbonic acid, increased the amount of the potash, and di- minished that of the nitric acid by the whole quan- tity of carbonic acid previously united with the * Mem. Paris, 1699, p. 48. 3 HISTORICAL INTRODUCTION. 3 potash. Of course, his results convey no idea whatever of the true composition of the salts which he attempted to analyze. We might, indeed, by subtracting the carbonic acid from the potash, and adding it to the acid, deduce the composi- tion of these salts according to his experiments. These deductions being made, it will be found that saltpetre according to his experiments, is composed of Potash 6 , Nitric acid 6-68 12-68 Numbers which, though not correct, constitute a pretty good approximation, and come much nearer the truth than the analysis of Kirwan, or even the more careful analyses of Berard or Wenzel. The experiments of Plummer were made in the same way, and were subject to the same error. In Dr. Black's celebrated Essay on Magnesia and Quicklime, he first pointed out the mistake of Homberg, and from some of his experiments deductions might be made respecting the con- stituents of some of the magnesian salts; but not being professed analyses, they have not attracted much of the attention of succeeding chemists. Soon after the middle of the eighteenth cen- tury, attempts were made by different chemists A2 4 HISTORICAL INTRODUCTION. to analyze various chemical substances, especially by Bergman, who had acquired so great a repu- tation, that his opinions carried with them a high degree of authority, and his analyses were for some time considered as models of accuracy a character* to which they are by no means entitled. J. ]? or the first accurate set of experiments covers the relative on the composition of salts, we are indebted to adds and Wenzel, a German chemist, who published in 1777> a small work entitled, Lehre von der Ver- wandschaft der Korper, or Theory of the Affinities of Bodies. He had been struck with a phenome- non which had already drawn some attention from chemists ; namely, that two neutral salts preserve their neutrality after mutually decom- posing each other. Wenzel proved by a long series of very accurate analyses, that all the al- kalies and earths bear the same ratio to all the acids. For example : if a given quantity of sul- phuric acid be saturated by 9-75 barytes, 6-5 strontian, 6 potash, 4 soda, 3-5 lime, 2-5 magnesia, the respective quantities of these earths and al- kalies, which saturate a given quantity of each of the other acids, will be to each other as the HISTORICAL INTRODUCTION. 5 numbers 9*75, 6-5, 6, 4, 3-5, and 2-5. When nitrate of lime, and sulphate of potash decompose each other, the nitrate of potash, and sulphate of lime formed, are neutral, because the relative quantities of potash and lime, which saturate a given weight of nitric and sulphuric acid, are the same. If we represent .nitric acid by 6*75, then 6 of potash and 3 %5 of lime will just satu- rate that quantity of it ; and 6 potash and 3*5 lime will just saturate 5 of sulphuric acid. Ni- trate of lime (abstracting the water) is composed of Nitric acid 6*75 Lime 3-5 10-25 and sulphate of potash of Sulphuric acid 5 Potash 6 11 When 1O25 nitrate of lime and 11 sulphate of potash are mixed, a double decomposition takes place, and there are formed Nitrate of potash -j Nitric acid 6-75 Potash 6 12-75 A 3 6 HISTORICAL INTRODUCTION. f Sulphuric acid 5 sulphate or lime -I _ . \ Lime 3-5 8-5 And these new salts are neutral, because nitric acid are^neutralized both by 6 potash and 3%5 lime, and 5 sulphuric by 6 potash and 3*5 lime/ It would appear a very extraordinary circum- stance, if we were ignorant of the almost despo- tic power of fashion and authority over the minds even of men of science, that this luminous explanation of Wenzel, confirmed by the most accurate experiments which chemistry possessed, I do not say merely at the time when they ap- peared, but for many years after experiments, which will even bear a comparison with those of Berzelius, made on purpose to demonstrate the truth of the very same principle which Wenzel first suggested it would appear very extraordi- nary indeed, that neither the principle of Wen- zel, nor the experiments by which it was con- firmed, are noticed by any of the contemporary chemists, or seem to have attracted the smallest attention. Bergman was at that time in the height of his reputation. He published analyses of several of the salts, greatly inferior in point of accuracy to those of Wenzel. He, of course, * I have not employed Wenzel's numbers to illustrate his opinions ; but the more accurate numbers determined in the succeeding chapters of this work. HISTORICAL INTRODUCTION. 7 did not perceive the existence of the principle which Wenzel had investigated. Yet the analy- ses of Bergman and of other chemists, though less accurate, and inconsistent with the fact which had led Wenzel to his investigation, were adopted by chemists in general, and thus their attention was withdrawn from one of the most important steps which the science has ever made, and the great improvement itself was delayed for more than thirty years. II. In the year 1792, Richter, who during The same the latter part of his life had the superrnten- supported dence of the porcelain manufactory at Berlin, byli published a work entitled, Anfangsgrllnde der Stocky ometrie, oder Messkunst Chymischer Ele- mente, Elements of Stochyometry, or the Mathe- matics of the Chemical Elements. This work con- sists of four thin volumes, which were printed at Breslau, during the years 1792, 3, and 4. It was followed by a periodical work, entitled, Ueber die Neuern Gegenstdnde der Chymie, on the New Objects of Chemistry. This work was be- gun in 1792,* and continued in twelve different numbers, or volumes, till the year 1807, the period of Richter's death. For he died on the 4th of May of that year.t In these works, he * In my copy, the date of the first part is 1799, but it is a second edi- tion. The second part was printed at Breslau in 1793. Hence I infer, that the first part appeared in 1792. f Gehlen's Jour, (second series) IV. 127. A 4 8 HISTORICAL INTRODUCTION. endeavoured to give Chemistry a mathematical form. He examined the phenomenon observed by Wenzel, and explained it in the same way. He endeavoured to determine the relative capa- city of saturation of the acids and bases, and attached a number to each, indicating this capa- city. The following table exhibits these num- bers. Alumina 525 Fluoric acid 427 Magnesia 615 Carbonic . 577 Ammonia 672 Sebacic 706 Lime 793 Muriatic . 712 Soda . 859 Oxalic 755 Strontian 1329 Phosphoric 979 Potash 1605 Formic 988 Barytes 2222 Sulphuric 1000 Succinic 1209 Nitric 1405 Acetic 1480 Citric 1683 Tartaric . 1694 Richter's labours in this important field pro- duced as little attention as those of Wenzel. Berzelius ascribes this to two causes. 1. The great inaccuracy of Richter's experiments, which indeed cannot for a moment be put in competition with those of Wenzel. 2. The vio- lent discussion between the antiphlogistic and phlogistic system, which had originated about the year 1787> and had not completely subsided at the time of Richter's death. The last of these HISTORICAL INTRODUCTION. 9 reasons had, in all probability, the greatest in- fluence. Bat in Great Britain, where the importance of Richter's opinions were likely to have been early appreciated, another cause operated to prevent them from gaining much ground. The French revolution was at its acme, when Rich- ter began his Stochyometrical investigations, and all Europe was plunged into the bloodiest and most inveterate war, that has almost ever afflicted this part of the world. Great Britain soon became involved in the dispute, and gra- dually, not only bore the brunt of the war, but was, by degrees, deserted by all her allies, and at last left alone to wage war against all the world. Hence the intercourse between the men of science in Great Britain and on the continent was gradually interrupted. This, together with the little attention paid to German literature in this country, prevented us from being aware of the labours of Richter. My first knowledge of them was derived from a note in Berthollet's Statique Chimique, published in 1803, and I found it impossible to procure a copy till after the battle of Waterloo. III. The notice in Berthollefs Statique Chi- mique, would probably have speedily drawn the attention of our countrymen to this most impor- * f first views tant department of chemistry, had not the genius respecting r T\ i A. i i i .the atomic oi Dalton led him to a much more general view theory. 10 HISTORICAL INTRODUCTION. of the subject, which has been usually called the atomic theory. I do not know when the ideas first occurred to him. In all probability they struck him by degrees, and were adopted in consequence of his own experimental investiga- tions. I first visited him at Manchester, on the 26th of August, 1804, and on that day, he ex- plained to me his notions respecting the com- position of bodies. The ultimate particles of all simple bodies are, in his opinion, atoms incapable of farther division. These atoms (at least viewed along with their atmospheres of heat) are all spheres, and are all of them possessed of particular weights, which may be denoted by numbers. And he represented the atoms of the simple bodies by symbols. The following may be taken as a specimen of his method : CHARACTERS. RELATIVE WEIGHTS. O Hydrogen ... 1 Azote .... 5 O Oxygen .... 6-5 Carbon 5 BINARY COMPOUNDS. O Water .... 7-5 (DO Nitrous gas . . . 11-5 Olefiantgas ... 6 (DO Ammonia ... 6 O Carbonic oxide . . . 11 '5 TERNARY COMPOUNDS. Carbonic acid . . . 18 HISTORICAL INTRODUCTION. 11 CHARACTERS. (DO Nitrous oxide Ethers Carburetted hydrogen OCDO Nitric acid RELATIVE WEIGHTS. 16-5 7 18 QUATERNARY COMPOUND. CD Oxv-nitric acid O QUINQUINARY COMPOUND. G)Q(DO Nitrous acid SEXTENARY COMPOUND. Alcohol O CD 00 SEPTENARY COMPOUND. Nitrate of ammonia 24-5 29-5 23-5 31-5 Such was the view of the subject, which Mr. Dalton had taken before the month of August, 1804. Unless my recollection fails me, Mr. Dalton's theory was originally deduced from his experiments on olefiant gas and carburetted hydrogen. He found both composed of carbon and hydrogen ; but the carbon in each being considered as the same, he found the hydrogen in the latter gas twice as much as in the former. This induced him to determine the ratios of these constituents in numbers, and to consider the olefiant gas as a compound of 1 atom carbon 12 HISTORICAL INTRODUCTION. and 1 atom hydrogen ; and carburetted hydro- gen of 1 atom carbon and 2 atoms hydrogen. The idea thus conceived was applied to car- bonic oxide, water, ammonia, &c. and numbers representing the atomic weights of oxygen, azote, &c. dedifced from the best analytical ex- periments which chemistry at that time pos- sessed. In the third edition of my System of Chemistry, published in 1807, I introduced a short sketch of Mr. Dalton's theory, so far as I understood it.* The same year a paper of mine on oxalic acid was published in the Phi- losophical Transactions, in which I showed that oxalic acid unites in two proportions with stron- tian, forming an oxalate and a binoxalate ; and that supposing the strontian in both salts the same, the oxalic acid in the latter is exactly twice as much as in the former.t About the same time Dr. Wollaston showed, that bicar- bonate of potash contains just twice the quan- tity of carbonic acid that carbonate of potash does ; and that there are three oxalates of potash, viz. oxalate, binoxalate and quadroxalate, the acids in which are to each other as the numbers * Vol. III. p. 424. f Berard, in a paper on the oxalates, which he published by way of cor- rection of mine, though his results are in general more erroneous than my own, denies the existence of binoxalate of strontian. I have often made the salt, and regularly exhibit it to my students. It differs very much from the oxalate. HISTORICAL INTRODUCTION. 13 1, 2, 4. These facts gradually drew the atten- tion of Chemists to Mr. Dalton's views, and induced several to adopt them, who were at first inclined to consider them as visionary. Among these I may mention Sir Humphrey Davy, who contributed so materially, a few years after, to confirm and extend this important theory. In 1808 Mr. Dalton published the first volume of his New System of Chemical Philo- sophy. In the last chapter of that volume, consisting only of six pages, he gives us a con- cise view of his atomic theory ; and in a plate at the end of the volume, he exhibits the sym- bols and atomic weights of 37 bodies, 20 of which were then considered as simple, and the other 17 as compound. In 1810, the second volume of this work appeared, in which he examined the combinations of oxygen with hy- drogen, azote, carbon, sulphur, and phosphorus ; and the combinations of hydrogen with carbon, sulphur and phosphorus ; and showed that all these combined according to the laws of the atomic theory. He also examined the atomic weights and combinations of the alkalies and earths. The third volume of this most impor- tant work, though it has been printed for several years, the author, for reasons with which I am unacquainted, has thought proper to withhold from the public and by so doing, has doubtless lost the right which he would have enjoyed, to 4 HISTORICAL INTRODUCTION. many curious discoveries, respecting the me- talic combinations which he treated of in that volume. He made Mr. Dalton made choice of the atom of hy- the atom of . _ , . . , . . . hydrogen drogen for his unity. And in this he has been followed by Dr.* Henry of Manchester, and by one or two chemical gentlemen in London. But this method has been rejected by almost all the British chemists, and by all the chemists, without exception, in Europe and America. The choice was unhappy, for very obvious reasons. 1. Because the atom of hydrogen is the most difficult of all to determine ; and chemists are not yet all agreed about its weight. According to Professor Berzelius, it is only O06&5; while the greater number of British chemists consider it as 0-125, or twice as much as Berzelius's estimate. Now, if we reckon the atomic weight of hydrogen by unity, and commit an error respecting its relation to that of other bodies ; this error will affect the atomic weights of all other bodies, and will make them all either too heavy or too light ; whereas, if we make choice of oxygen for our unity, any error re- specting the atom of hydrogen will be confined to that atom, and will not affect the accuracy of the atomic weights of other bodies. 2. Hydro- gen, so far as we know at present, combines with but few of the other simple bodies ; while oxygen unites with them all, and often in vari- HISTORICAL INTRODUCTION. 15 ous proportions. Consequently very little ad- vantage is gained by representing the atom of hydrogen by unity ; but a very great one, by representing the atom of oxygen by unity. For it reduces the greater number of arithmeti- cal operations, respecting these bodies, to the addition of unity; and we see at once, by a glance of the eye, the number of atoms of oxygen which enter into combination with the various bodies. Thus, if the atom of manganese be re- presented by 3%5, and the weight of the various oxides of that metal be as follows : 1 Suboxide . . 4 2 Protoxide . . 4-5 3 Deutoxide . . 5 4 Tritoxide . 5-5 5 Manganesous acid . 6" 5 6 Manganesic acid . 7'5 It is obvious at once that the Suboxide contains | atom oxygen, Protoxide 1 Deutoxide 1 J Tritoxide 2 Manganesous acid 3 Manganesic acid 4 whereas, if we were to make the atom of hydro- gen unity, these weights would be as follows : Manganese . . 28 Suboxide 32 16 HISTORICAL INTRODUCTION. Protoxide . . 36 Deutoxide , . 40 Tritoxide . . 48 Manganesous acid . 56 Manganesic acid . 64 Numbers, which would not suggest the number of atoms of oxygen contained in each, with the same facility. But the ra. 3. All the knowledge that we are likely ever tins should . . .. . to acquire or the atomic weights of bodies, is bie. si tl ~ merely the ratios of these weights. Now, it is ob- vious, that ratios are much more easily observed, and much more easily employed, the smaller the numbers are, by which they are denoted. Thus the ratio of 38741 to 116223 is the same as that of 1 to 3. But how very inconvenient would the great number of figures be, compared with the simple number 1 : 3. It is therefore highly important in chemistry, that the atomic weights of bodies should be represented by the smallest possible numbers. Now, when we make the atom of oxygen unity, the atomic weights of all bodies are represented by numbers eight times less than they would be if the atom of hydrogen be unity. It has been represented by some persons, in favour of the atom of hydrogen being unity, that in that case the atomic weight of every body is a whole number; whereas, when oxygen is the unity, fractions are frequently unavoidable. And these fractions, it is said, ren- HISTORICAL INTRODUCTION. 17 der the atomic weights more complicated and unmanageable, than the greater size of the whole numbers deduced from hydrogen as a unity. But this view of the subject is quite erroneous. It will be shown in this work, that even when oxygen is made unity, nearly one half of all the atomic weights of bodies are whole numbers. And when fractions exist, they are always either 0*25, 0*5, or 0-75 ; except in four or five cases, when the fraction is 0-125, 0-375, or 0-625. If hydrogen be unity the atom of manganese is 28, while it is 3i if oxygen be unity. The atom of barytes on the former supposition is 78, on the latter 9f. Now surely it will not be said that the fractional numbers are more unwieldy, or more unmanageable than the whole numbers ; while in all cases of whole numbers the advan- tage on the side of the latter method is very great. Thus, if hydrogen be unity, the atom of uranium is 208, while if oxygen be unity it is only 26. It was Dr. Wollaston that first suggested the representing of oxygen by unity. Berzelius, in his important investigations, adopted this me- thod ; and I did the same thing in the papers which I wrote on the Atomic Theory in the first series of the Annals of Philosophy. In the second series of the Annals, Mr. R. Philips has given a table of atoms according to the hydrogen scale. VOL. I. B 18 HISTORICAL INTRODUCTION. IV. For the next step in the progress of the sac's theory , , , .~ T of volumes, atomic theory, we are indebted to Gay-Lussac. In the second volume of the Memoires D'Ar- cueil published in 1809, there appeared a paper by that acute chemist, On the Combination of Gaseous Substances with each other.* In this paper he shows that the gases, considered in re- spect of their volumes, unite with each other in a very simple manner one volume of one gas either combining with one volume, with two vo- Inmes, or with half a volume of the other. Thus one volume of carbonic acid gas combines with one volume and with two volumes of ammonia- cal gas ; fluoboric acid gas combines with one volume and with two volumes of ammoniacal gas; protoxide of azote is a compound of one vo- lume azotic, and half a volume oxygen gas ; deutoxide of azote of one volume of azotic, and one volume of oxygen gas, &c. He proved also in the same paper, that when gases unite, the alteration of volume which they sustain, is al- ways very simple. Deutoxide of azote is com- posed of one volume of azotic, and one volume of oxygen gas united without alteration of vo- lume. Carbonic acid is composed of one volume carbonic oxide, and half a volume of oxygen gas condensed into one volume, &c. This very important view of the way in which gaseous bodies unite, threw a new and unex- * Pag. 207. HISTORICAL INTRODUCTION. 19 pected light on chemical combinations. It has contributed very materially to simplify and per- fect the mode of analyzing compound gases. It agreed admirably with the views of Mr. Dalton, as Gay-Lussac hints in his paper j and might have been employed by him in his investigations ; and would have enabled him, had the specific gravity of gaseous bodies been more accurately known, at the time when he wrote, to have determined the atomic weights of the acidifiable combustibles with much more accuracy than he did. But Mr. Dalton's first volume was pub- lished, and probably the greater part of his se- cond volume printed, before he had an opportu- nity of seeing Gay-Lussac's paper. And when he did see it, instead of welcoming it as an im- portant addition to his own views, he set him- self to show that Gay-Lussac's opinions were ill founded. But the subsequent researches of chemists have left no doubts about their accura- cy ; and if Mr. Dalton still withholds his assent, he is, I believe, the only living chemist who does so. V. When Professor Berzelius of Stockholm thought of writing his elementary treatise on discoveries Chemistry, the first volume of which was pub- itJ s Berze ~ lished at Stockholm in 1808, he prepared him- self for the task by reading several chemical works, which do not commonly fall under the eye of the chemist. Among others, he read the B8 20 HISTORICAL INTRODUCTION. writings of Richter, and was very much struck with the explanation there given, of the compo- sition of salts, and the precipitation of metals by each other. It followed from the researches of Richter, that if we are possessed of good analy- ses of certain salts, we may, by means of them, calculate with accuracy the composition of all the others. Professor Berzelius formed imme- diately the project of analyzing a series of salts with the most minute attention to accuracy. While employed in putting this project in exe- cution, Davy discovered the constituents of the alkalies and earths ; and Mr. Dal ton's theory came to his knowledge. All this enlarged his views as he proceeded, and induced him to em- brace a much wider field than he had originally contemplated. His first analyses were unsatis- factory ; but by repeating them, and varying the methods, he detected errors, improved his processes, and finally obtained results, which agreed exceedingly well with the theoretical calculations. These laborious investigations seem to have occupied him several years. The first knowledge which we obtained of them in this country, was in consequence of the publica- tion of a letter addressed by Berzelius to M. Berthollet, in the 77th volume of the Annales de C/iimie, about the beginning of the year 1811. In this letter he gives an account of his methods of analyses, together with the composition of 47 HISTORICAL INTRODUCTION. 21 compound bodies. He shows that when a me- tallic proto-sulphuret is converted into a sul- phate, the sulphate is always neutral ; that an atom of sulphur is twice as heavy as an atom of oxygen 5 and that when sulphite of barytes is converted into sulphate, the sulphate is neutral, there being no excess either of acid or base. From these, and many other important facts, he finally draws this conclusion : " In a compound formed by the union of two oxides, the one which (when decomposed by the galvanic bat- tery) attaches itself to the positive pole (the acid, for example,) contains 2, 3, 4, 5, &c. times as much oxygen, as the one which attaches itself to the negative pole, (the alkali, earth, or me- tallic oxide). 5 ' Berzelius' Essay itself appeared in the third volume of the Afhandtingar i Fysik, Kemi och Mineralogie, printed at Stockholm in 1810, where it occupies 114 pages. A transla- tion of it was inserted almost immediately, in Gil- bert's Annalen der Physik, and in the Annales de Chimie ; but no English translation of this most important paper has ever been given to the public. In 1815, Berzelius applied his views to the mineral kingdom. His Essay was pub- lished in the fourth volume of the Af hand- lingar, and an English translation of it by Mr. Black was soon after printed. And some years after, Berzelius himself got a French translation of it published in Paris. The fifth B3 22 HISTORICAL INTRODUCTION. volume of the Afhandlingar, published in 1818, is almost filled with papers on the same subject by Berzelius. His papers on the subject in that volume, amounting to nine and all of them, especially the two last, are exceedingly important. The whole of "these papers have appeared in the German Journals; and the greater number of them in the French and British Journals. In the 6th volume of the Afhandlingar, and in several of the latter volumes of the Memoires of the Stockholm Academy, this indefatigable ex- perimenter has prosecuted the subject still far- ther. To him we are indebted for a much greater number of accurate analyses than to the whole chemical world put together. It is quite needless to observe how much the labours of Berzelius have contributed to the establishment of the atomic theory, and the immense progress which the science has in consequence made. In fact, we may almost say that it was by him that the theory was established. His analyses, in point of accuracy, infinitely surpass all those which had preceded him. He has embraced every department of chemistry with equal ardour, and every department lies under the deepest obligations to his indefatigable industry and pro- found sagacity. VL In the J ear 1814 > there appeared in the P m l so P mca l Transactions, a description of a synoptical scale of chemical equivalents by Dr. HISTORICAL INTRODUCTION. 23 Wollaston. In this paper we have the equiva- lents or atomic weights of 73 different bodies, deduced chiefly from a sagacious comparison of the previous analytical experiments of others, ahd almost all of them exceedingly near the truth. These numbers were laid down upon a sliding rule by means of a table of logarithms, and over against them the names of the sub- stances. By means of this rule, a great many important questions respecting the substances contained on the scale may be solved. Hence the scale is of great advantage to the practical chemist, and of course ought to be found in every laboratory. It will give, by bare inspec- tion, the constituents of all the salts contained on it, the quantity of any other ingredient necessary to decompose, and the quantities of the new constituents that will be formed. The contrivance of this scale, therefore, may be con- sidered as an important addition to the atomic theory. It rendered that theory every where familiar to all those who employed it. To it chiefly we owe, I believe, the currency of that theory in Great Britain ; and the prevalence of the mode which Dr. Wollaston introduced, namely, of representing the atom of oxygen by unity, or at least by ten, which comes nearly to Prout . sdis the same thing . c L over y of c the relations VII. In the year 1815, a paper appeared in between the the 6th volume of the Annals of Philosophy weights. 84 24 HISTORICAL INTRODUCTION. by Dr. Prout, " On the Relation between the Spe- cific Gravities of Bodies in their Gaseous State and the Weights of their Atoms. 55 This paper displays a degree of sagacity that has seldom been exceeded in chemical investigations, arid shows clearly liiat the author, if he chose, might rise to the highest eminence as a chemical phi- losopher. He proves, in the first place, that if air be a compound of 20 volumes of oxygen and 80 of azotic gas, then the specific gravities of these gases are Oxygen I'llll Azotic 0-9722 He shows also, that the specific gravity of hydro- gen gas is 0*0694 : that if we reckon the atomic weight of hydrogen 0*125, that of oxygen is unity. The following table exhibits the num- bers which he determined for several simple bodies and most of his determinations were the result of direct and very accurate experiments : 1 Oxygen . 1 2 Azote . 1-75 3 Hydrogen . 0-125 4 Chlorine . 4-5 5 Iodine . 15-5 6 Carbon . 0-75 7 Sulphur . 2 8 Phosphorus . 1*75 9 Calcium . 2-5 10 Sodium 3 HISTORICAL INTRODUCTION. 25 11 Iron . 3-5 12 Zinc . 4 13 Potassium . 5 14 Barium 8-?5 He observed that all these atomic weights are multiples of the atomic weight of hydrogen : indeed, that all of them are multiples of twice hydrogen, or 0*25, and most of them of 4 hydro- gen, or 0*5. He observed also, that in general the specific gravity of the body in the gaseous state, may be obtained by multiplying its atomic weight by O5555, or half the specific gravity of oxygen gas ; because the atom of oxygen is re- presented by half a volume, but that of most other substances by a whole volume. VIII. It was this admirable paper that ,.,... tions of ned me that new analytical investigations were author, still necessary to determine the atomic weights of bodies with perfect accuracy ; and I formed the project of attempting the investigation my- self by direct experiment. This project it was not in my power to commence till the summer of 1819, because it was not till the end of the preceding summer that the college of Glasgow, to which I had recently been attached, could furnish me with a laboratory. Since that period I have devoted the whole of my spare time to the investigation ; and have resolutely withheld my- self from entering upon several rather inviting fields, which have attracted general attention. My 26 HISTORICAL INTRODUCTION. experiments have been very numerous, amounting to many thousands. At first my progress was very slow ; because I was obliged to try many methods of investigation before I hit upon one sufficiently accurate for my purpose. But I found the subject become easier as I proceeded, and at last I acquired so much facility, that, during the last four months, I have been able to determine with tolerable accuracy the consti- tuents of at least a hundred salts. My first ob- ject was to satisfy myself whether Dr. Prout's opinion, that the atomic weights of all bodies are multiples of that of hydrogen, was correct j because the establishment of its truth would at once give a simplicity to the atomic numbers, which had not been suspected, and would place the science of chemistry in a new and much more advantageous situation than it had ever oc- cupied. The very numerous investigations, which will be exhibited in the following pages, will, I flatter myself, fully establish the truth of Dr. Prout's sagacious conjecture. For every substance, of which I could procure a sufficient quantity to enable me to examine it fully, has been found not only a multiple of the atomic weight of hydrogen ; but, if we except a few compounds into which a single or odd atom of hydrogen enters, they are all multiples of 0-25 or of two atoms of hydrogen. After determining the atomic weights of all HISTORICAL INTRODUCTION. 27 the simple substances, and of many of their com- pounds, I next wished to examine the composi- tion of a considerable number of salts, to see whether any law would develope itself respecting the quantity of water of crystallization which these bodies frequently contain. \Vith this view a great number of salts were subjected to expe- riment ; but the only obvious result is, that the number of atoms of water in salts, depends more upon the base than upon the acid : for the same base, in general, affects nearly the same number of atoms of water in its salts. But to this law there are many exceptions. I now present the result of my investigations to the chemical public with considerable diffi- dence. For though I have been at the utmost pains to ensure accurate results ; and often re- peated what I considered as fundamental expe- riments eight or ten times, before I felt myself finally satisfied ; yet the number of experiments gone through is so great, that it will be won- derful if, in some cases, I have not fallen into mistakes. I have made a point of relating my experiments (where I thought them funda- mental) nearly as they were made : because, as I had no person to go over my calculations, which have been of necessity very numerous, though all of the simplest kind ; it is very likely that I have not been always able to prevent the occurrence of some arithmetical errors in them. 3 28 HISTORICAL INTRODUCTION. And in all such cases the knowledge of the ori- ginal experiments will put it in the power of the reader to calculate for himself, and to correct any errors into which I may have inadvertently fallen. The want of this attention in other writers has sometimes been severely felt by me, when an arithmetical error was conspicuous, without the requisite data for correcting it, or the means of knowing upon what number the error had fallen. Even the writings of Berzelius, though in gene- ral his numbers are very accurate, are not always free from this provoking uncertainty. CHAP. II. OF THE ATOMIC THEORY. Two opinions have divided philosophers respect- Opinions ing the nature of the ultimate elements of bodies, the ultimate ever since they began to speculate on the subject. s According to one party, these ultimate elements consist of atoms, or particles, incapable of farther division ; according to another, matter is infinite- ly divisible, and consequently no such atoms can exist. These two opinions divided the ancients ; and, after the revival of science in modern Eu- rope, we find them equally dividing the philoso- phical world, and splitting it into sects. About the beginning of the last century, the infinite di- visibility of matter was a favourite subject of dis- cussion in Great Britain, and seems to have been universally adopted. Dr. Keil, in his Introduc- tion to Natural Philosophy, devotes three lectures to this subject. In the first, he demonstrates the infinite divisibility of matter ; in the second, he answers the objections which have been advanced against the truth of the doctrine ; and in the third, he gives a number of examples of the ex- treme subtilty of matter, by showing how very minute a portion of certain substances may be made visible to the eye, or distinguished by the 30 OF THE ATOMIC THEORY. Leibnitz sense of smelling. Leibnitz, on the other hand, considered . . . . them as adopted the opinion, that the ultimate elements of bodies are atoms ; and, to obviate the mathe- matical objections of those who had embraced the doctrine of the infinite divisibility of matter, he considered these atoms as unextended. But numerous absurd consequences were still deduci- ble from this opinion of Leibnitz, which were finally obviated by Boscovich ; who considered the ultimate elements of bodies to be composed of unextended atoms ; but demonstrated, in a most ingenious manner, from the law of conti- nuity, that mutual contact between these atoms, or the bodies which they constitute, is impossi- His view ble. According to the theory of Boscovich, the improved . iiy BOSCO. latest and most complete and ingenious which has been hitherto offered to the public, bodies are composed of unextended atoms, indivisible and homogeneous, possessing forces, which al- ternate from repulsive to attractive, according to the distance. This theory has been viewed in a favourable light by some of the most eminent philosophers who have appeared since the The- oria Naturalis of that mathematician was pub- lished. The opinion, that the ultimate principles of bodies are atoms which are incapable of far- ther division, seems, chiefly in consequence of the authority of Boscovich, to have become the prevailing opinion among the philosophers of the present time. OF THE ATOMIC THEORY. 31 With respect to the notion entertained by Bos- covich, that the ultimate atoms of bodies are homogeneous, we are incapable at present of deciding whether it be well or ill founded. It is not likely that any of these ultimate elements has ever come under our inspection. All our simple bodies are most probably compounds. It is possible that the ultimate elements of bodies may be very few it is even conceivable that they may be reduced to two ; but in what way all the variety of bodies with which we are ac- quainted, could be produced from one single kind of elementary body or atom, I cannot, for my own part, form any conception. If the atoms of bodies be incapable of farther division or diminution, it seems requisite that we should adopt the opinion of Leibnitz, that these atoms are unextended. This notion involves us in a maze of metaphysical subtilties, through which Boscovich has piloted his way with infinite address. But such discussions being quite unsuit- able to a system of chemistry, or to the science of chemistry in its present state, it will be better to avoid them altogether. By atom, Meaning of J the word then, in the following pages, I would be under-a/omintim stood to mean, the ultimate particles of which any body is composed, without considering whe- ther the farther division of these particles be pos- sible or not. They differ from each other in weight ; but whether this difference be owing to 32 OF THE ATOMIC THEORY. a difference in their size or specific gravity, or of both together, I do not enquire. The weights attached to the atoms of bodies, and deduced from the combinations into which they enter, merely express the ratios of the atomic weights, and hav" e no reference whatever to the size or specific gravity of the atoms to which they are attached. If a body be simple, it seems reasonable to be- lieve, that its atoms or ultimate particles, are in- capable of farther diminution or division : and consequently, that the supposition of the exis- tence of half atoms involves an absurdity. Not- withstanding this, I have in certain cases, in the following pages, admitted the possibility of one atom of one body, uniting with one and a half, or with half an atom of another body. This ad- mission has been viewed by many persons as im- proper and absurd, and as even involving an impossibility. I need hardly observe, that we can get rid of these half atoms with the greatest ease, by merely doubling the numbers represent- ing the constituents of the compound. Thus, if an atom of iron weigh 3-5, and an atom of oxy- gen 1, and if we consider it as absurd to view peroxide of iron as a compound of 1 atom of iron, and 1^ atom of oxygen ; we have only to double 3-5, and 1-5, which are the two consti- tuents of peroxide of iron. By so doing we get 7=2 atoms of iron, and 33 atoms of oxygen, OF THE ATOMIC THEORY. 33 and the peroxide of iron will be a compound of 2 atoms iron, and 3 atoms oxygen. It is by no means unlikely, that this last method may represent the true composition of this oxide, and of the other compounds into which half an atom of oxygen, or of an acid or base, appears to enter. Nevertheless, I have been induced to represent the compound in the first way rather than the second ; because it gives us smaller and simpler numbers, and enables us better to see the connection between the com- pound in question, and other compounds of the same constituents. If we represent the atomic weights of iron and its oxides as follows, Iron . . . 3-5 Protoxide of iron . 4-5 Peroxide of iron 5*0 we see at a glance that the protoxide is a com- pound of 1 atom iron + 1 atom oxygen ; and that the oxygen in the peroxide is to that in the protoxide, as 3 to 2. But if we represent these atomic weights by Iron . . 3'5 Protoxide of iron . 4-5 Peroxide of iron . 10.0 by doubling the number of atoms in the perox- ide, the relation between the oxygen, in the two oxides of iron, is much less evident. For the VOL. I. C 34 OF THE ATOMIC THEORY. sake of these advantages, which I consider as considerable, I have been induced to introduce half atoms in certain cases. But I would not be understood to support the abstract idea of the possibility of the existence of half atoms ; but to have been actuated by the arithmetical facilities derived from the method which I have adopted. Those who prefer theoretical considerations to all such facilities, have only to double the atomic weight of all bodies containing half atoms, and they will get rid of the anomaly. The substances containing a half atom of oxygen do not amount to more than 7- A few occur contain- ing half an atom of sulphur, and a few contain- ing half an atom of acid. Though the word atom, according to etymo- logy, and indeed according to its usual accepta- tion, implies a particle incapable of farther diminution or division, I have been induced, for the sake of shortness, to apply it in the following pages with more latitude. When it is applied to the ultimate particles of simple bodies, it is taken in its usual and strict acceptation ; but when it is applied to a compound body, it is to be understood as synonymous with integrant particle. Thus, an atom of iron, may be con- sidered as implying an ultimate particle of that metal ; but an atom of protoxide of iron means merely an integrant particle of protoxide of iron, which is a compound of 1 atom of iron and 1 . OF THE ATOMIC THEORY. 35 atom of oxygen ; and consequently contains in reality two atoms, I have seldom applied the term atom to any of the salts ; but if it were so applied, it would denote a substance containing a considerable number of atoms. Thus, an atom or integrant particle of sulphate of ammonia, is composed of 1 atom sulphuric acid 5 1 atom ammonia . 2-12.5 1 atom water . 1-125 8-25 | An integrant particle of sulphuric acid contains 4 atoms ; an integrant particle of ammonia con. tains 4< atoms ; and an integrant particle of water contains 2 atoms. So that an atom, or integrant particle of sulphate of ammonia con- tains no fewer than 10 atoms. I doubt not, that this loose mode of applying the term atom will appear to many of my readers an improper liberty, and quite inconsis. tent with the precision necessary in chemical discussions. But my reason for doing so was, that I am of opinion, that we are not at present acquainted with any truly simple bodies. All our simple bodies are most probably compounds ; and many of them may be afterwards decom* pounded and reduced to more simple principles by the future labours of chemists. If this opinion be true, I could not have employed the term C2 36 ^ OF THE ATOMIC THEORY. atom at all, or at least I could not have applied it to any of the bodies with which we are at present acquainted. This consideration appears to have induced several of the most distinguished chemists, both in this country and in France, to discard the teAn atom altogether. Thus, Dr. Wollaston, instead of atom, employs the phrase chemical equivalent, and Sir H. Davy employs the term proportion. But I did not see any reason for being so squeamish. The word atom is more convenient, shorter, and more distinct than any other word which I could think of. And when it is understood to signify merely integrant particle of the substance to which it is applied, it cannot, I think, have any tendency to mislead the reader. 2. Mr. Dalton, in the account of his atomic theory, which he gave to the public at the end of the first volume of his New System of Chemical Philosophy, employs some terms which have gradually got into general use, and which there- B8n * l J tBr * fore require to be explained here. These terms explained, are binary, ternary, quaternary, &c. 1 atom of A -|- 1 atom of B = 1 atom of C, a binary compound. 1 atom of A -\- 2 atoms of B = 1 atom of D, a ternary 2 atoms of A -f- 1 atom of B = 1 atom of E, a ternary 1 atom of A -\- 3 atoms of B = 1 atom of F, a quaternary 3 atoms of A -j- 1 atom of B = 1 atom of G, a quaternary, &c. Thus, protoxide of iron is a binary compound, composed of 1 atom of iron + 1 atom of oxygen. OF THE ATOMIC THEORY. 37 Carburetted hydrogen is a ternary compound, composed of 1 atom carbon and 2 atoms hydrogen : deutoxide of azote is a ternary -com- pound, composed of 1 atom azote and 2 atoms oxygen. Ammonia is a quaternary compound, consisting of 1 atom azote and 3 atoms hydro- gen ; sulphuric acid is a quaternary compound, composed of 1 atom sulphur and 3 atoms oxy- gen. Nitrous acid is a quinquenary compound, composed of 1 atom azote and 4 atoms oxygen. Nitric acid is a septenary compound, composed of 1 atom azote and 5 atoms oxygen. After these illustrations the terms septenary, octonary, &c. will be understood without any farther explanation. Mr. Dalton lays down the following rules, by Mr. Dai- which he thinks we should be guided in our opinions respecting chemical synthesis. (1.) When only one combination of two bodies can be obtained, it must be presumed to be a Unary one, unless some cause appear to the contrary. (2.) When two combinations are observed, they must be presumed to be a binary and a ternary. (3.) When three combinations are obtained, we may expect one to be a binary and the other two ternary. (4.) When four combinations are observed, C3 OF THE ATOMIC THEORY. we should expect one to be binary, two ternary, and one quaternary, &c. (5.) A binary compound should be always specifically heavier than the mere mixture of its two ingredients. (6.) A ternary compound should be specifi- cally heavier than the mixture of a binary, and a simple, which would, if combined, constitute it, &c. 5th and 6th With respect to these rules, the first four not accu- . , i i i i i rate. appear unexceptionable, unless the third should be objected to ; but the fifth and the sixth do not appear to be warranted by a review of the compounds with which we are acquainted, com- pared with their constituents. It is in general true, that when two substances unite, they form a compound whose specific gravity is greater than the mean of that of the two constituents. But to this there are many exceptions. Deutoxide of azote has precisely the mean specific gravity of the two constituents. It has been demon- strated by Gay-Lussac, that it is composed of a volume of oxygen gas and a volume of azotic gas, united together without alteration of bulk. One volume protoxide of azote + I volume of ox- ygen gas united, will form 2 volumes of deutoxide of azote ; so that in the act of uniting, an ex- pansion instead of a contraction takes place. The specific gravity of an alloy of 2 atoms tin OF THE ATOMIC THEORY. 39 and 1 atom lead ought to be 9*321 ; but it amounts only to 8'817, showing that an expan- sion has attended the union of these metals, in- stead of a contraction. 3. Berzelius, in his different papers upon the atomic theory, and especially in his Essay on a the Theory of Chemical Proportions, has made us aquainted with the rules that guided him in the conclusions which he has drawn respecting the atomic weights of the different bodies, and the number of atoms of which the compound bodies are composed. The most important of these laws are the following : (1.) One atom of A may combine with 1, 2, 3, 4, &c. atoms of B. We do not know the limit ; but it rarely passes 5 atoms. (2.) Two atoms of A may combine with 3 atoms of B. But this is a much rarer case than the preceding. (3.) When an acid and a base unite, the oxygen in the acid is a multiple of that in the base by 1, 2, 3, &c. The only exceptions to this law, exist in the combinations of phosphoric, arsenic, and nitric acids with bases. (4.) Oxides containing only 1 atom of oxygen, united to 1 atom of base, have very weak affi- nities for acids. (5.) Oxides with 2 atoms of oxygen have very strong affinities for acids. (6.) Oxides with 3 atoms of oxygen have weak c 4, 40 OF THE ATOMIC THEORY. affinities for acids. Some of them possess rather the properties of acids than of bases. (7.) Most acids contain 3 atoms of oxygen ; but some contain 2, 4, 5, 6, 8 atoms. The first two of these laws are unexcep*- tionable ; bein merely a statement of what every person, who has attended to the combina- tions of atoms with each other, must have ob- served. The third The third law is not so obviously just, and will lawnotum- f t . . * versaiiy require some consideration before it can be adopt- ed. The greater number of the alkaline bases, as will appear from the following pages, contain only 1 atom of oxygen : and the neutral salts being compounds of 1 atom acid, and 1 atom base, it is clear that in all of them, the oxygen of the acid will be a multiple of the oxygen of the base ; because every number is a multiple of unity. Hence every acid, whether it contain 1, 2, 3, or more atoms of oxygen, must, in fact, con- tain a number of atoms of that principle, equal to the oxygen in the base, multiplied by the number of atoms of oxygen in the acid. This must necessarily be the case, because the oxygen in every acid always amounts to a whole number of atoms. Berzelius, indeed, considers almost all the salifiable bases as containing two atoms of oxygen. But he represents the neutral salts as compounds of 1 atom base, and two atoms acid. Now, whatever the number of atoms of oxy- OF THE ATOMIC THEORY. 41 gen may be which an acid contains, it is obvious that the double of it must constitute an even number, or a number divisible by 2, and conse- quently a multiple of 2. As far then as the neutral salts are concerned, Berzelius's law expresses nothing more, than that the salifiable bases are mostly protoxides a pretty remarkable fact, though JBerzelius himself was not aware of it. Even in the neutral phosphates, arseniates, and nitrates, and in the biphosphates, binarseniates, and binitrates, the same law will hold. It is only in the subsalts, and in those salts that contain peroxides united to an acid, that we can see whether this supposed law of Ber- zelius be general or not. For example, sulphuric acid contains 3 atoms of oxygen : now, it would be a very remarkable circumstance indeed, if no protoxide exist, of which 2 atoms are capable of uniting with 1 atom of sulphuric acid. There are but few subsalts known at present; yet when I come to treat of the salts, I think I shall be able to show examples of subsalts, which are in- consistent with Berzelius's supposed law. Per- oxide of iron is a compound of 1 atom of iron, and H atom of oxygen; consequently, whether it combine with 1 atom, 2 atoms, or 3 atoms, of sulphuric acid, it is obvious that the oxygen of the acid must be a multiple of the oxygen of the base: For 3=1 -5 x < 2. But phosphoric acid is a compound of 1 atom phosphorus, and 2 atoms OF THE ATOMIC THEORY. of oxygen. Hence, Berzelius's law could not hold with perphosphate of iron, unless it con- sisted only of 1 atom peroxide of iron united to 3 atoms of phosphoric acid. We are acquainted at present with but very few compounds of peroxides with acids ; yet, "when I come to treat of the salts, I think I shall be able to point out some which are not combined according to Berzelius's law. It would, therefore, be extremely hazardous to employ this law as a guide to enable us to deter- mine the quantity of oxygen united to bases. This, Berzelius is frequently in the habit of do- ing, and I think it has misled him in a variety of cases. These I may perhaps notice more parti- cularly hereafter. kh,5th,and Berzelius's fourth, fifth, and sixth laws, are ' on founded entirely upon his own peculiar views re- specting atomic combinations, and cannot, there- ore ^ serve as guides to us. So far from being of opinion that the combinations of 1 atom base, and 1 atom oxygen, constitute weak alkaline bodies, I consider them as the strongest of all : for the fixed alkalies, alkaline earths, and most of the metallic oxides which constitute salifiable bases, belong to this class of bodies. To the oxides which Berzelius considers as composed of 1 atom base, and 1 atom oxygen, I give the name of suboxides, and believe them to be compounds of 2 atoms base, and 1 atom oxy- gen. They have but a weak affinity for acids ; OF THE ATOMIC THEORY. 1<,'3 though most of them combine, at least, with muriatic acid. Hitherto, only a few of these suboxides have been noticed by chemists, though it is likely that they will increase considerably hereafter. It is unnecessary to make any observations 7th upon Berzelius's seventh law ; it expresses mere- ly his opinion respecting the composition of acids. m If the views which I have given in the following pages be entitled to any confidence, it will be seen that eleven of the acids contain three atoms of oxygen ; namely, 1 Oxalic 4 Arsenic 7 Succinic 10 Tungstic 2 Sulphuric 5 Molybdic 8 Benzoic 11 Fluoboric. 3 Chromic 6 Acetic 9 Formic Six contain only one atom of oxygen ; namely, 1 Silica 8 Hyposulphurous 5 Columbic 2 Phosphorous 4 Oxide of tellurium 6 Fluoric. Twelve contain two atoms ; namely, 1 Carbonic 4 Sulphurous 7 Selenic 10 Uranitic 2 Boracic 5 Titanic 8 Antimonic 11 Uric 3 Phosphoric 6 Arsenious 9 Molybdous 12 Fluosilicic. Two contain four atoms ; namely, Manganesic and citric. Three contain five atoms ; namely, Hyposulphuric, tartaric, chloric. 44 OF THE ATOMIC THEORY. One contains seven atoms ; namely, Perchloric ; And one eight atoms ; viz. Saclactic. Rules by 4. I shall now briefly notice the considerations which the . investiga- by which 1 have been regulated in the following author were investigations, and by which I have been led to regulated. x U p On ^ Q a tomic numbers contained in this work. (1.) The only useful information respecting the atoms of bodies, which we are likely to ac- quire, is that of the ratios which these weights bear to each other. Hence, the smaller the numbers employed to represent these atoms, the more useful will they be to the practical chemist. (2.) Neutral salts (that is, salts which are nei- ther acid nor alkaline) are probably composed of 1 atom of acid united to 1 atom of base. When the salt has acid qualities, it is probably a com- pound of 1 atom base, and more than 1 atom of acid. When it possesses alkaline qualities, it is probably a compound of 1 atom acid, and more than 1 atom of base : but to these rules there are many exceptions. Some acids have not the property of neutralizing the bases, in however great a proportion they are combined with them. This is the case with boracic acid ; even a com- pound of 8 atoms of boracic acid, and 1 atom of OF THE ATOMIC THEORY. 45 soda, possesses alkaline characters. We cannot, therefore, determine the proportion between the atoms of acid and base in the borates, by their neu- trality. The same thing holds with the alkaline car- bonates, which always possess alkaline properties whatever be the number of atoms of carbonic acid they contain. On the other hand, there are several bases, as the oxides of copper and zinc, which do not destroy the acid properties of acids in how great a proportion soever they be combined with them. Such salts alway redden vegetable blues, and possess certain characters denoting the presence of an acid. (3.) The weights of certain acidifiable bases were determined by their relation to oxygen. Thus it was found that 2 of sulphur is capable of combining with 1, 2, and 3 of oxygen. Hence, it was concluded, that 2 represents the atomic weight of sulphur, and that the atomic weights of the compounds of sulphur and oxygen are, ATOMIC WEIGHT. 1. 2 sulphur + 1 oxygen = 3 hyposulphurous acid. 2. 2 +2 = 4 sulphurous acid. 3. 2 +3 =5 sulphuric acid. In like manner, O75 carbon unites with 1 and 2 of oxygen. Hence 0*75 represents an atom of carbon, and the atomic weights of the com- pounds of carbon and oxygen are, 1. 0-75 carbon + ] oxygen = 1'75 carbonic oxide. 2. -f 2 = 2-75 carbonic acid. 46 OF THE ATOMIC THEOJIY. 1*75 azote combines with 1, 2, 3, 4, 5 of oxy- gen. Hence 1. ^1-75 azote + 1 oxygen = 2-75 protoxide of azote. 2. 1-75 + 2 =3-75 deutoxide of azote. 3. 1-75 4" 3 = 4'75 hyponitrous acid. 4. 1*75 +. 4 = 5-75 nitrous acid. 5. 1-75 +5 = 6-75 nitric acid. (4.) After determining in this way, the atomic weights of some of the most common and impor- tant acids, I examined the neutral salts which these acids formed by uniting with the salifiable bases, and determined the proportion between the weights of the acid and base in each. From this I deduced the atomic weight of the base. Thus, knowing that the atomic weight of sul- phuric acid is 5, and that sulphate of potash is a compound of Sulphuric acid 5 Potash 6 - 11 I drew as a conclusion, that the weight of an integrant particle of sulphate of potash is 11, and that an atom of potash weighs 6. By this method, I determined the atomic weights of most ofthe salifiable b as es. (5.) I now combined some .of the most com- mon and important of these salifiable bases with the other acidsthus forming a new set of neutral salts. These were analyzed with great OF THE ATOMIC THEORY. 47 care, and the ratio between the acid and base in them determined. From this I deduced the atomic weights of the other acids. Thus, know- ing that the atomic weight of soda is 4, and that neutral anhydrous acetate of soda is composed of Acetic acid 6-25 Soda 4 10-25 I drew as a conclusion, that 1O&5 represents the weight of an integrant particle of anhydrous acetate of soda, and that 6 '25 is the atomic weight of acetic acid. In this way, the atomic weights of all the acids and bases were determined from neutral salts, and they represent the ratios in which the acids and bases saturate each other, expressed in the smallest numbers consistent with the no- tion, that the atomic weight of oxygen is unity. The methods employed to determine the ato- mic weights of the simple bodies, of which these acids and bases are composed, will be under- stood by perusing the different chapters of the following work. CHAP. III. OF THE SPECIFIC GRAVITIES OF OXYGEN AND HYDROGEN GASES. IT is only of late years, that the great impor- tance of an accurate knowledge of the specific gravities of the gases has been fully appreciated. Three different sets of experiments on the sub- J ments al- ject have been laid before the public. The first ready made was by JBiot and Arago, which was published in mine these the Memoir -es de V Institute, for 1806, p. 320. gravities ' They give a table of the specific gravity of a few of the most important gases, and describe the processes by which they were determined. The next was a paper inserted by me in the 16th volume of the Annals of Philosophy. It contained the specific gravities of 23 gases. The experiments were made with every precau- tion which I could think of, to secure me against mistakes ; and I believe them to come as near the truth as we can reach, with our pre- sent means of operating. The third set was by Messrs. Berzelius and Dulong, and published in the Annales de Chimie et de Physique^ for 1820.* They determined the specific gravity of oxygen, hydrogen, carbonic acid, and azotic * Vol. XVI. p. 386. Vot. I. D 50 OXYGEN AND HYDROGEN GASES. gases with great care, and gave a table of the specific gravity of 21 bodies, calculated from these data, and the doctrines of Gay-Lussac, respecting the mode in which gases combine. These results of Berzelius and Dulong differed so much from 'mine as to excite my surprise ; while the well known accuracy and dexterity of these gentlemen naturally induced me to dis- trust my former results, and to endeavour to verify them by new experiments. After a good many fatiguing trials made with every pre- caution that I could think of, I was unable to find any flaw in my former results, or to re- concile them with those of Berzelius and Du- long. I was satisfied that I could not approach any nearer the goal by the direct road. But another method of proceeding occurred to me, susceptible, I thought, of an unlimited approxi- mation to the truth. This method applied only to oxygen and hydrogen gases. But these bo- dies constitute the keystone of the building. And were their specific gravities once deter- mined with precision, we are enabled, by means of the atomic theory, to determine with accu- racy the specific gravities of most of the other gases. Method of I propose in this chapter to lay before the the author. , r . * J reader, a set or experiments which seem to me to settle the specific gravities of these two gases, without any error which can much exceed COMPOSITION OF OXIDE OF ZINC. 51 the ten thousandth part of the whole. The accuracy of the process, indeed, is only limited by the delicacy of the balance made use of in the experiments. And my balance enabled me to weigh to within less than one hundred thou- sand part of the weight. It has been demonstrated by Gay-Lussac and Humboldt, that water is a compound of oxygen and hydrogen gases, in the proportion of one volume of the former, to two volumes of the latter. It occurred to me that by a very careful analysis of water, the relation between the vo- lumes and the weights of these gases might perhaps be determined. These experiments were made by means of zinc, and before their accuracy can be appreciated, it will be necessary to know the exact composition of the oxide of zinc. I shall therefore divide this chapter into three sections. In the first section, we shall de- termine the composition of oxide of zinc ; in the second, the specific gravity of oxygen gas ; and in the third, the specific gravity of hydrogen gas. SECT. I. OF THE COMPOSITION OF OXIDE OF ZINC. I SHALL be under the necessity to state in this section a variety of facts, which may seem at first sight foreign to the subject in view ; but D2 OXYGEN AND HYDROGEN GASES. which are requisite, in order to enable the rea- der to judge of the justice of the conclusions which are to be drawn. Some of the details may appear needlessly minute ; but I must re- quest the reader's attention to them, because they constitute* the foundation of the whole fa- bric which I mean to build. The zinc of commerce does not answer for the experiments which I am going to describe. It is never quite free from carbon, which has a material effect in increasing the specific gravity of the hydrogen gas, obtained by dissolving it in sulphuric acid. It contains, likewise, small quantities of several other metals, and does not dissolve completely in sulphuric acid. The zinc which I used, was obtained by exposing common zinc to a white heat in a stone-ware retort, luted to a receiver nearly filled with wa- ter. At this temperature, the zinc is sublimed, and freed from all its impurities, except a trace of cadmium too minute to occasion any sensible error in the result of the experiments. The zinc thus distilled over, should be melted in a crucible, and poured upon the surface of a clean, smooth sandstone. We thus obtain it in a thin sheet, which can be easily broken into small pieces to suit the object in view. 21-25 grains of this zinc were introduced into tion of oxide of zinc. a small green glass retort, the weight of which had been previously determined. As much COMPOSITION OF OXIDE OF ZINC. 53 'very weak but pure nitric acid as was fully suf- ficient to dissolve the whole zinc was introduced into the retort. As soon as the metal was all dissolved, a receiver was applied to the retort which was placed in hot sand, and allowed to re- main till the whole liquid contents had distilled over. The liquid in the receiver was saturated with carbonate of ammonia, and boiled for some time to see whether it contained any zinc ; but no precipitate appeared. The retort, thus freed from its liquid contents, was placed upon a charcoal fire, and kept for about an hour in a dull red heat. This com- pletely decomposed the nitrate of zinc, which existed at first in the retort, dissipated the nitric acid, and left nothing behind but the 21*25 grains of zinc, now converted into an oxide. Being placed in the scales, the zinc, by being oxydized, was found to have increased in weight 5 grains. Thus, it appears that oxide of zinc is composed of Zinc 21-25 Oxygen 5 26.25 ' or, dividing each constituent by 5 to bring them to their lowest terms, of Zinc 4-25 Oxygen 1 5-25 D3 54 OXYGEN AND HYDROGEN GASES. As zinc combines with oxygen in only one proportion, it is reasonable to conclude that the oxide is a compound of 1 atom zinc and 1 atom oxygen. If this supposition be admitted, the atom of zinc is, obviously 4^ times heavier than the atom of oxygen. And if we represent the atomic weight of oxygen by unity, then the atomic weight of zinc will be 4*25 and the in- tegrant particle of oxide of zinc will weigh 5*25 ; or in other words, 5'25 of oxide of zinc will just saturate an integrant particle of any acid with which it is capable of combining. The law holds with all acids ; but I shall at present verify it only by referring to three acids ; namely, the sulphuric, carbonic, and arsenic acids. It will be shown in the 5th, 1st, and 7th sections of the sixth chapter of this work, that the atomic weights of these acids are as follows : Sulphuric acid . 5 Carbonic acid . 2-75 Arsenic acid . 7*75 Now, 5*25 of oxide of zinc will just saturate 5 of sulphuric acid, 2'75 of carbonic acid, and 7*75 of arsenic acid; or in other words, the sulphate, carbonate, and arseniate of zinc, sup- posing them anhydrous, are composed as follows : ]. SULPHATE OF ZINC. Sulphuric acid . 5 Oxide of zinc 5-25 COMPOSITION OF OXIDE OF ZINC. 55 2. CARBONATE OF ZINC. Carbonic acid . 2 '75 Oxide of zinc . 5-25 3. ARSENIATE OF ZINC. Arsenic acid . 7*75 Oxide of zinc . 5-25 We shall verify this statement by a rigid analysis of these three salts. 1. SULPHATE OF ZINC. Sulphate of zinc, or white vitriol, is a well A ^ a iy sis f sulphate of known salt which crystallizes in beautifully tran- zinc - sparent four-sided prisms, terminated by four- sided pyramids ; or if the crystallization be rapid, it assumes the form of long slender white needles. However carefully prepared, it always reddens vegetable blues. But it crystallizes also with an excess of acid. In that case, the crys- tals assume the form of large flat rhomboids, which are not altered by exposure to the air, and may be raised to the temperature of at least 500 without being deprived of their excess of acid.* * I am doubtful whether this salt constitutes a truly distinct species. I found its composition 1^ atom acid 6-66 1 atom oxide 5-25 8 atoms water 9-00 20-91 or, it consists of one atom of the protohydrate of sulphuric acid, combined with three integrant particles of the neutral salt. D4 56 OXYGEN AND HYDROGEN GASES. grains of the neutral sulphate of zinc were dissolved in distilled water, and precipitated by a solution of carbonate of soda. The pre- cipitate was separated from the liquid by means of a double filter, consisting of two pieces of unsized and well washed printing paper, made previously of the same weight. The carbonate of zinc on the filter was washed with distilled water, till the liquid which passed through ceased to render muriate of barytes muddy. The filter was suspended in a linen bag above the sand bath, in a temperature of 85, till it was apparently dry ; and it was finally left for some time in a temperature of about ^12. It was afterwards weighed by placing the two filters in opposite scales, and adding weights till the two scales were exactly coun- terpoised. The weight of the carbonate of zinc was exactly 8 grains. It was put into a pla- tinum crucible, which was covered with a lid, and exposed for half an hour to a red heat, There now remained 5*%5 grains of oxide of zinc. Thus it appears, that 18*125 grains of sul- phate of zinc, in crystals, contain 5*25 grains of oxide of zinc. It is still easier to prove that the sulphuric acid contained in this quantity of the salt weighs precisely 5 grains. But it will be necessary to state, in the first place, that when 15-5 grains of crystallized COMPOSITION OF OXIDE OF ZINC. 57 muriate of barytes are exposed to a red heat, they lose 2-25 grains of their weight, and are converted into 13-25 grains of chloride of ba- rium. When 13-25 grains of chloride of barium are dissolved in water, the solution contains just 9*75 grains of barytes. Now, 9*75 represents the atomic weight of barytes, which requires for saturation exactly 5 of sulphuric acid. And 14*75 parts of sulphate of barytes contain just 5 parts of sulphuric acid. A little distilled water was put into two small cylindrical glass jars, and 18-125 grains of crystallized sulphate of zinc were dissolved in the one of these, and 13-25 grains of chloride of barium in the other. The two liquids being mixed, a double decomposition took place, and sulphate of barytes precipitated in the state of a white powder. This precipitate being col- lected, washed, dried, and heated to redness, weighed 14*75 grains, and therefore contained exactly 5 grains of sulphuric acid. A few drops of the colourless liquid from which this sulphate had precipitated, were put into a watch glass, and a small quantity of muriate of barytes mixed with it ; but it did not become in the least muddy or even opal coloured, showing that it contained no sensible quantity of sulphuric acid. For I find that 1 grain of glauber salt (containing 0-2469 grain of sulphuric acid) dis- solved in 12,000 grains of water, is sensibly 58 OXYGEN AND HYDROGEN GASES. precipitated by muriate of barytes. Hence it is clear, that in the present experiment, the quantity of sulphuric acid remaining in the liquid could not amount to jojo2 of a grain. Indeed, as the quantity of liquid never amounted to 1000 grains,"the sulphuric acid remaining in so- lution, could not be so much as %&&&-$ of a grain. Another portion of this residual liquid being mixed with sulphate of soda, was not in the least altered, showing that the liquid contained no sensible portion of barytes. Thus it appears, that the barytes from 13*25 grains of chloride of barium, or 15*5 grains of muriate of barytes, just saturate the sulphuric acid in 18*125 grains of sulphate of zinc. Hence, this acid just weighs 5 grains. We see therefore, that in neutral sulphate of zinc, 5*25 parts of the oxide are combined with 5 parts of the acid; which was the thing to be proved. These two constituents added together make 10*25, which subtracted from 18*125 leave 7*875 for the quantity of water of crystal- lization. It will appear afterwards that this is equivalent to 7 atoms of water. Sulphate of zinc in crystals, then, is composed of ] atom sulphuric acid 5 1 atom oxide of zinc 5-25 7 atoms water . 7*875 18-125 COMPOSITION OF OXIDE OF ZINC. 59 So that 18*125 is the weight of an integrant particle of the crystallized salt. It was the knowledge of this that induced me to make choice of 18*125 grains of the salt. My first experiments were made upon a much larger quantity ; but I have often repeated them with 18*125 grains, and obtained the result just stated. If the solution of neutral sulphate of zinc be White vi - triol. concentrated too much, it deposites, while still hot, an opaque crust upon the bottom of the vessel. This crust is very white, not so soluble in water as the crystals, and has not the least appearance of crystallization. It is neutral, though it reddens vegetable blues like the pre- ceding salt; and being analysed in the same way as the crystallized sulphate, was found composed of 1 atom sulphuric acid 5 1 atom oxide of zinc 5*25 3 atoms water . 3-375 13-625 So that an integrant particle of it weighs 13*625, and it differs from the crystals in containing only 3 atoms of water, instead of seven. The white vitriol of the Germans consists of this second salt, while the sulphate of zinc man- ufactured in Great Britain consists of the for- mer. The German white vitriol contains 24'77 OXYGEN AND HYDROGEN GASES. per cent, of water, while the British sulphate of zinc contains 43-44 per cent. 2. CARBONATE OF ZINC. One of the Experiments related while giving the analysis of sulphate of zinc, is sufficient to show the proportion of carbonic acid, which combines with 5-25 parts of oxide of zinc. For 18-125 grains of sulphate of zinc, when decom- posed by carbonate of soda, yielded a precipitate which, after being well washed and dried in the temperature of 212, weighed 8 grains, and when exposed to a red heat, was reduced to 5*25 grains. These 8 grains were carbonate of zinc. The carbonic acid was dissipated by the red heat, and the oxide remained pure. We see that 5-25 of oxide of zinc are converted into carbonate by uniting with 2*75 grains of carbo- nic acid. When carbonate of zinc is dried in the open air,- without exposure to any artificial heat, I have obtained it composed as follows : 1 atom carbonic acid 2 '75 1 atom oxide of zinc 5-25 2 atoms water 2*25 10-25 so that an integrant particle of it weighs 10-25. But the water is very easily dissipated indeed, COMPOSITION OF OXIDE OF ZINC. Gl so much so, that it is not easy to dry this car- bonate without sustaining at least a partial loss of the water. But I have more than once suc- ceeded in getting it with the exact atomic proportions. The carbonate of zinc occurs in the mineral kingdom, sometimes anhydrous, and sometimes containing 1 atom of water. 3. ARSENIATE OF ZINC. This salt was obtained by the following pro- ofarsemate cess: 18-12,5 grains of the crystallized sulphate of zmc * of zinc, and 20*75 grains of arseniate of soda were dissolved in separate portions of distilled water, and the two solutions, after being well mixed, were evaporated to dryness, and the dry mass digested in distilled water. A dou- ble decomposition took place, and arseniate of zinc separated in the state of a gelatinous looking matter. Being washed and dried in the open air, it constituted a white powder, weighing 26-5 grains. But when heated to redness, the weight was reduced to 13 grains. By a careful analysis of this salt, I found that its constituents were 1 atom arsenic acid 7*75 1 atom oxide of zinc 5-25 8 atoms water 9'0 It would be easy to increase the number of these examples j but those given are sufficient I OXYGEN AND HYDROGEN GASES. to show that the atomic weight of oxide of zinc is 5&5, and that it is a compound of 4'2<5 zinc and 1 oxygen. And this being the only object in^view, we need not at present proceed any far- ther. SECT. II. OF THE SPECIFIC GRAVITY OF OXYGEN GAS. VARIOUS statements of the specific gravity of this gas have been published by different che- mists. The following table exhibits those which have been generally considered the most to be depended on. g!arities of 1 ' 114 Saussure, Ann. de Chim. LXXI. p. 269. oxygen gas. 1*103 . Biot and Arrago, Mem.de 1'Institut. 1806, p. 320. 1-089 Allen and Pepys. Phil Trans. 1807. 1-1117 Thomson, Annals of Philosophy, XVI. p. 1 63, 1-1026 . Berzelius and Dulong, Ann. de Chim. et de Phys. XV. 386. It would be impossible to decide which of these determinations deserves the preference. The experiments were all made with care, and no doubt, sufficient precautions were taken to have the gas in a state of purity. I shall there- fore proceed to relate the attempts which I have made to determine the specific gravity of oxy- gen gas in quite a different way. SPECIFIC GRAVITY OF OXYGEN GAS. ( A hundred grains of distilled zinc were put Method of r>ii i i detennin- into a small retort, nearly filled with a mixture i ng the true of one part of sulphuric acid, and four parts distilled water. The beak of the retort was im- mediately plunged into a water trough, and the hydrogen gas evolved, was received into a gra- duated glass jar. After an interval of twenty four hours, when the whole zinc had been long dissolved, and the apparatus and its contents had acquired the temperature of the room, the hydrogen gas evolved was measured. The ex- periment was twice repeated, and the quantity of gas obtained in both cases was exactly the same. The thermometer during both experi- ments stood at 48. The barometer, during the first experiment, was at 30 inches ; but during the second, at 29*6 inches. The volume of gas extricated in the first ex- periment was 137*08 cubic inches. In the se- cond experiment, 100'18 grains of zinc were dissolved, and the volume of the gas extricated was 139 '284 cubic inches. Now, when we make the necessary corrections for the height of the barometer, and the excess of zinc dissolved, the quantity of gas evolved is reduced to 137*08 cubic inches, as in the first experiment. These two experiments give us the volume of hydrogen gas evolved during the solution of 100 grains of pure zinc in dilute sulphuric acid. But the gas was saturated with moisture at the 64 OXYGEN AND HYDROGEN GASES. temperature of 48 b . Had it been dry, its volume would have been only 135-4*755 cubic inches.* At the temperature of 60, 135-4755 cubic inches of gas become 138-7551 cubic inches.t Thus it appears, that when 100 grains of zinc are dissolved i dilute sulphuric acid, the hydro- gen gas evolved, supposing it perfectly dry, and supposing the barometer to stand at 30 inches, and the thermometer at 60, amounts to 138*7551 cubic inches. Now, it has been de- monstrated long ago, by Lavoisier, that the hy- drogen gas thus evolved is derived from the water with which the sulphuric acid is diluted. A portion of this water is decomposed ; its oxy- gen unites to the zinc, and converts it to an oxide, while the hydrogen makes its escape in the state of gas. The other constituent of the water, then, the oxygen has united to the zinc. * Let p = pressure of atmosphere = SO inches mercury. J'= elasticity of vapour, at 48 = 0-351 inches mercury. x = volume of dry gas at 48. It has been shown by Dal ton, that pae P-f 137-08 /, 137-08 f x = and (substituting for p and f their = 137-08. Consequently mbstit values, ) 4064 -265 x = 30 = 135-4755 cubic inches. f From the experiments of Guy-Lussac we learn, that air and all gases increase in bulk _^_ part for every degree of Fahrenheit's thermometer. Hence at 60, the volume of the gas is 2 greater than at 48. Now, ^ of 135-4755 is 3-2796. Adding this to 135-4755, the volume be- comes 138-7551 cubic inches. SPECIFIC GRAVITY OF OXYGEN GAS. 65 But if it were in the gaseous state, we know that its volume would be just half that of the hydrogen gas, or 69*3775 cubic inches. Thus it appears, that 100 grains of zinc, when con- verted into an oxide, unite with 69*3775 cubic inches of oxygen gas. But from the facts established in the last sec- tion, it is evident, that 100 grains of zinc, when converted into an oxide, combine with 23*5294 grains of oxygen.* Hence, the weight of 69*3775 cubic inches of oxygen gas is 23*5294 grains. Consequently, 100 cubic inches weigh 33*915 grains.t From the experiments of Sir George Shuck- burgh Evelyn we learn, that 100 cubic inches of dry common air, under the same pressure, and at the same temperature, weigh 30*5 grains. And the specific gravity of common air being always reckoned unity, it is obvious that the specific gravity of oxygen gas must be l*llll.t Thus we have obtained by a very simple pro- cess, which does not seem liable to any uncer- tainty, the very specific gravity of oxygen gas which Dr. Prout deduced from the supposition, that air is a mixture of four volumes azotic, and one volume oxygen gas ; and which I adopted in * 4-25 grains of zinc unite with 1 grain of oxygen, and 4-25 : 1 : : 100 : 23-5294. f For 69-3775: 23-5294 : ; 100 : 33 915. t For 30-5 : 33-915 : : 1 : Mill nearly. VOL. I. E 66 OXYGEN AND HYDROGEN GASES. the last edition of my System of Chemistry, and in my Essay on the Specific Gravity of the Gases. If we compare the table at the begining of this section, with the specific gravity now found, we will perceive the results of the different experi- menters as follows : Saussure's, too high by . . -^\ G th part. Biot and Arago's, too low by . T ^th part. Allen and Pepys's, too low by . j^d part. Thomson's, too high by . . YS^^d part. Berzelius and Dulong's, too low by y^^th part. There seems no reason to doubt that Biot and Arago, and Allen and Pepys, took the specific gravity of oxygen gas standing over water, and consequently saturated with moisture. Let us suppose the temperature at which their experi- ments were made to have been 60, and let us see what influence the presence of moisture must have had upon the specific gravity of their gas. Let p pressure of the atmosphere, suppose 30 in. mercury. y^elasticity of vapour at 60.=O524 inch mercury. rt=sp. gr. of dry oxygen gas=l-llll. 6==sp. gr. of vapour at 6'0.=0.00772. .r=sp. gr. of moist oxygen gas at 60. It is well known that x ^V~L(a + &)= 1-0992. This is nearly the mean between the results of SPECIFIC GRAVITY OF HYDROGEN GAS. 67 Biot and Arago, and Allen and Pepys ; Saus- sure's errors and my own, being in excess, can only be explained by a mistake in the weighing. My own, indeed, is accurate as far as my balance could enable me to go. The quantity of oxygen gas weighed, amounted to 20 grains, and the smallest weight I used was riuth of a grain ; con- sequently, I could only estimate the weight with- in aoVoth part : the three first decimal places I could depend on, but not upon the fourth ; ac- cordingly the fourth decimal place was too high, SECT. III. OF THE SPECIFIC GRAVITY OF HYDROGEN GAS. I GOT a thin glass flask blown, with a flat bot- Method f ' e determin- tom and a narrow cylindrical mouth, into which ing the Sp2 , the extremity of a bent glass tube was ground, of hydro- 7 so as to fit it accurately. The capacity of the gen gas ' flask was about 18 cubic inches ; the tube was 18 inches long, and the diameter of its interior cavity about the fifth of an inch. Its two extremities were bent nearly at right angles with the body, and turned different ways as here represent- . , . _ -/J ed ; a being the i end which was ground into the mouth of the flask. The whole E 2 68 OXYGEN AND HYDROGEN GASES. length of this tube was filled with small fragments of fused muriate of lime, except the two extremi- ties, which were stuffed full of amianthus, in order to keep the muriate of lime in its place. This flask was nearly filled with a mixture of sul- phuric acid and distilled water, in the proportion of about four parts of the latter to one part of the former. It was accurately weighed by means of an excellent balance, and the weight of the tube with the muriate of lime, was also separately determined. I then introduced into the flask, as rapidly as possible, a known weight of distilled zinc, in fragments ; the tube with the muriate of lime was fixed in its mouth ; and it was immedi- ately plunged up to the neck in a large trough of water of the temperature of 48, and left in that situation till the zinc was all dissolved. The flask was then wiped dry, put again in the bal- ance, and the loss of weight which it had sus- tained was ascertained ; at the same time, the augmentation of weight of the tube with the muriate of lime was determined. The object of this experiment was to ascertain the weight of a given volume of hydrogen gas ; for the experiments related in the last section give us the volume of hydrogen gas evolved dur- ing the solution of a given weight of zinc. It was attended with so many unforeseen difficul- ties, that I was obliged to repeat it ten times be- fore I became aware of all the precautions SPECIFIC GRAVITY OF HYDROGEN GAS. 69 necessary to ensure success. Unless the gas be extricated very slowly, the muriate of lime does not dry it sufficiently indeed, I was never able to free it completely from moisture ; but a very simple calculation gave me the weight of the vapour which still adhered to it. The tempera- ture of the liquid should be kept as low as possi- ble, not merely to moderate the fermentation, but to avoid the risk of sulphuric acid escaping along with the gas. When the flask was left in the air, the temperature continued to rise for more than an hour after the experiment began. In one experiment it rose from 50 to 87, in another, from 48 to 81 : the greater the quantity of zinc dissolved, the more accurate (caeteris paribus) was the experiment ; but the quantity was unluckily very much limited by the small size of the flask. It will be sufficient if I relate one experiment, during which, all the precautions that I could think of were taken ; and which, I have reason to believe, was very nearly accurate. The quantity of zinc dissolved was 143 grains the effervescence continued for 24 hours the temperature never rose higher than 49 the barometer stood at 3O1 inches, and the thermo- meter at 49 the loss of weight was 4*3 grains. It will be, perhaps, more perspicuous, if we suppose that 100 grains of zinc only were dis- solved. On that supposition, the loss of weight E s 70 OXYGEN AND HYDROGEN GASES. sustained was 3 grains, and the tube containing the muriate of lime, had increased in weight 0-163 grains. From the experiments related in the last sec- tion, it follows, that the volume of hydrogen gas given out during the solution of 100 grains of zinc, supposing the barometer to stand at 30*1 inches, and the thermometer at 49, amounts to 136-88 cubic inches. The specific gravity of vapour at 49> under a pressure of 30*1 inches. of mercury, is 0-00533 ; and the weight of the vapour contained in 136-83 cubic inches of moist gas, is 0-00533 x 136-88 x 0-305 gr.=0-2225 grain ; but the moisture re- tained by the muriate of lime was only 0*163 grain. It is obvious from this, that the hydro- gen still retained 0-059 grain of moisture. If from the weight lost, amounting to 3 grains, we subtract this 0-059 grain for moisture, the remainder, amounting to 2*941 grains, gives the true weight of the hydrogen gas exhaled, sup- posing it perfectly dry. Now, from the experi- ments related in the last section, we know that the volume of this gas, under the pressure of 30 inches of mercury, and at the temperature of 60, is 138-7551 cubic inches. Thus it appears, that 138*7551 cubic inches of ust sixteen . times heavi- dry hydrogen gas (barometer 30 inches, thermo- drogen gL meter 60) weigh 2-941 grains: consequently, 100 cubic inches must weigh 2*119 grains. But SPECIFIC GRAVITY OF HYDROGEN GAS. 71 we have seen in the last section, that 100 cubic inches of oxygen gas weigh 33*91.5 grains ; now 2-119 : 33-915: : 1 : 16-005. This approaches so nearly the ratio of 1 : 16, that it leaves no doubt that the specific gravity of oxy- gen gas is exactly 16 times greater than that of hydrogen gas. Consequently, the specific gravity of hydrogen gas is Ye 1 -, or 0*0694; or exactly the specific gravity already deduced by Dr. JProut, from other considerations, and which I obtained experimentally, as may be seen in my Essay on the Specific Gravity of Gases. That the proportion between the specific gravities of oxygen and hydrogen gases, is that of 16 : 1, is farther confirmed by an experiment of Berzelius and Dulong. They made dry hy- drogen gas pass through a determinate weight of oxide of copper, heated to redness. The water thus formed was collected by means of dry chloride of calcium. The loss of weight sus- tained by the oxide of copper gave the weight of the oxygen in the water, while the weight of the hydrogen was denoted by the difference be- tween the weight of the water and the oxygen : the mean of three trials gave the oxygen 889, and the hydrogen 111. Now 111 : 889 : : 1 : 8*009 ; and as water is a compound of 2 volumes hydrogen to 1 of oxygen, it follows, that the E4 OXYGEN AND HYDROGEN GASES. specific gravity of hydrogen is to that of oxygen gas as 1 : 16-018.* In my experiment, the oxygen was more than 16 times heavier than the hydrogen gas by smooth part in that of JBerzelius and Dulong it was a little more. The mean of both (which must come very near the truth) gives us the specific gravity of hydrogen to that of oxygen gas, as 1 : 16*011. That the true specific gravity of oxygen gas is I'll 11, will be shown in the fifth chapter of this work.t The following table exhibits the most recent specific gravities of hydrogen gas obtained by experiment : SP. GR. OF HYDROGEN GAS. 0-0732 . . Biot and Arago. 0-0694 . . Thomson. 0-0688 . . Berzelius and Dulong. It is obvious, when we compare these with the above found specific gravity of hydrogen gas, which must be exceedingly near the truth, that * Ann. de Chim. et de Phys. xv. 389. f I ought to mention that the weight of my flask filled with the dilute sulphuric acid and zinc, was about 3000 grains, and that of the tube filled with muriate of lime, about 1000 grains. My scales, loaded with double that weight, turned sensibly with ^o^th of a grain. Hence, I consider myself as certain of the weight of the hydrogen gas to as many decimal places as I have given in the text. I weighed so as to make myself sure of the accuracy within yy^th of a grain : Berzelius and Dulong deter- mined the composition of water by an experiment nearly similar. Their result does not tally exactly with mine, nor can I institute a comparison, because they do not state the delicacy of their balance. SPECIFIC GRAVITY OF PIYDROGEN GAS. 73 Biot and Arago's is too high by rath part, Ber- zelius and Dulong's too low by rfjth part, while my results, owing probably to a fortunate accident, are, as far as they go, quite correct. Water being; a compound of 1 volume of oxy- C . . f tion of wa- gen gas, and 2 volumes ot hydrogen gas, it is ob- ter. vious, that it is composed of 8 parts, by weight, of oxygen, and 1 part of hydrogen. There can be no doubt that water is a neutral compound, and that it consists of an atom of oxygen united to a determinate atomic quantity of hydrogen ; but chemists are not yet quite agreed whether the hydrogen be 1 atom or 2 atoms. Sir H. Davy and Professor Berzelius consider the hy- drogen as amounting to two atoms, because there are 2 volumes of it united to 1 volume of oxy- gen ; for they consider the volumes of these gaseous bodies as representing atoms. On the other hand, Mr. Dalton, Dr. Wollaston, Dr. Prout, and indeed by far the greater number of chemists, consider water as a compound of 1 atom oxygen, and 1 atom hydrogen, founding their opinion upon the well known fact, that these two gases cannot be directly united in any other proportions than those which exist in water. If water contain only 1 atom of hydrogen, the atomic weight of this substance is obviously ath of that of oxygen, or 0-125 ; but if that liquid contain 2 atoms of hydrogen, then the atomic weight of hydrogen is only 0-0625. 74 OXYGEN AND HYDROGEN GASES. It is of no great consequence which of these opinions be adopted ; yet if we take a view of the various compounds into which hydrogen enters, we can have little hesitation in fixing upon 0-125 for its atomic weight. A few of these compouuds may be noticed in this place. The evidence of the facts here stated will be laid before the reader in the subsequent chap- ters of this work. Atom of i. It is now known that oxygen and hydrogen hydrogen / & '. weighs unite in two proportions. Water contains the smallest quantity, while the deutoxide of hy- drogen, discovered by Thenard, contains the greatest quantity of oxygen. The composition of these bodies is as follows : Water consists of 1 + 0-125 Deutoxide of hydrogen of 2 + 0-125 According to this way of stating the consti- tuents, water is a compound of 1 atom oxygen, and 1 atom hydrogen ; and the deutoxide of 2 atoms oxygen and 1 atom hydrogen. But if we state the constituents according to Berze- lius's views, OXYGEN. HYDROGEN". Water consists of 1 + 0-0625 X 2 or 2 atoms. Deutoxide of . 1 _}. 0-0625 or 1 atom. Or, water is a compound of 1 atom oxygen and 2 atoms hydrogen ; and the deutoxide of 1 atom oxygen and 1 atom hydrogen. Now, would it SPECIFIC GRAVITY OF HYDROGEN GAS. ^5 not be a singular anomaly, if water, which is decomposed with considerable difficulty, should consist of 1 atom oxygen united to 2 atoms hydrogen, while the deutoxide, which parts with half its oxygen very readily, consists of 1 atom of oxygen united to 1 atom of hydrogen. This is not the case with other compounds. Bi- sulphate of potash may be deprived of half its sulphuric acid by exposure to a strong red heat ; but the other half (which constitutes with potash the neutral compound) resists the highest tem- perature which we can apply. 2. Hydrogen unites with azote, carbon, phos- phorus, and sulphur, as well as with oxygen, and with some of these bodies it unites in more than one proportion. But with none of these bodies does it ever combine in a smaller pro- portion than 0-125. Now, it would be a most extraordinary phenomenon, if never less than two atoms of this substance were to unite with an atom of other bodies. The following table ex- hibits the composition of a variety of bodies containing hydrogen. HYDROGEN. Ammonia, composed of 1-75 azote +0-125 X 3 Olefiant gas . . 75 carbon + 0-125 Carburetted hydrogen . 0-?5 + 0-125 X 2 Phosphuretted hydrogen . 1-5 phosphorus + 0-125 Bihydroguret of phosphorus 1-5 +0-125X2 Sulphuretted hydrogen . 2 sulphur +0-125 1-75, O75, 1-5, and 2 represent the atomic 76 OXYGEN AND HYDROGEN GASES. weights of azote, carbon, phosphorus, and sul- phur respectively, on the supposition that an atom of oxygen weighs unity. It is obvious from the table, that 0*125 hydrogen is the smallest quantity of hydrogen that unites with an atom of aeote, carbon, phosphorus, or sul- phur. Now, if 0*0625 were the atomic weight of hydrogen, how comes it that it never unites to any of these bodies in so small a quantity ? But if 0*125 be the atomic weight of hydro- gen and 1 of oxygen, since water is a compound of 1 oxygen and 0*125 hydrogen, and is at the same time a compound of 1 volume oxygen and 2 volumes hydrogen gases, it follows as a con- sequence, that a volume of oxygen contains twice as many atoms as a volume of hydrogen. So that if we represent an atom of hydrogen by a volume, we must represent an atom of oxygen by half a volume. It is obvious that the weight of an integrant particle of water is 1125; nor is this weight altered, whether we consider it as a compound of 1 atom oxygen and 1 atom hydrogen, or of 1 atom oxygen and 2 atoms hydrogen. For on either supposition the atomic quantity of hydro- gen must be represented by 0*125. CHAP. IV. OF THE ATOMIC WEIGHTS AND SPECIFIC GRAVITIES OF CHLORINE AND IODINE. THESE two substances resemble each other in so many respects, that they ought to be classed together. Their analogy with oxygen is also very striking ; though they differ likewise in some circumstances ; the most remarkable of which is the property that they have of uniting with oxygen, and constituting with it an acid very similar in many respects to the nitric. I shall treat of each of these bodies in a distinct section. SECT. I. OP THE ATOMIC WEIGHT OF CHLORINE, AND THE SPECIFIC GRAVITY OF CHLORINE GAS. IT has been demonstrated, that muriatic acid gas is a compound of equal volumes of chlorine gas and hydrogen gas, united together without any change of bulk. Hence, the knowledge of the specific gravity of dry chlorine gas at the temperature of 60, and when the barometer stands at 30 inches, will give us the ratio be- 78 CHLORINE AND IODINE. v tween the atomic weights of chlorine and hy- drogen. Srrit^of Now, I took the specific gravity of this gas chlorine three times successively, and found it always 1-500.* Many other considerations lead to the conclusion th#t this is very near the truth. Of these it may be sufficient to mention the fol- lowing : 1. The specific gravity of muriatic acid gas was found by me from an average of three ex- periments, to be 1-284366. Now, the specific gravity of this gas is a mean between that of chlorine and hydrogen gases. Let x = sp. gr. of chlorine gas a = sp. gr. of hydrogen gas We have -+-" = 1-284366. Consequently x 2-568733 0-0694 = 2-499333. Now, 2-499333 is only irrVr part less than 2-5. This is a coincidence between the two experi- ments as near as the nature of the apparatus employed will admit of. 2. Gay-Lussac, in a table inserted in the first volume of the Annales de Chimie et de Physique, states the specific gravity of chlorine gas to be 2-470. Now, if his experiments were made witli gas prepared over water, as was undoubtedly the case, his result approaches very nearly to mine, if we estimate the specific gravity of the * Annals of Philosophy, xvi. 169. CHLORINE. 79 dry gas, as will be obvious from the following simple calculation. Let x = sp. gr. of dry chlorine gas a = sp. gr. of vapour at 60 = 0'00772 p = 30 inches mercury f elasticity^of vapour at 60 = 0-524 then f ' + 0) = 2-47 P The solution of this equation gives us x = Thus it appears, that if Gay-Lussac made his ex- periments at the temperature of 60, the true specific gravity of his chlorine gas was 2-507. If the temperature (as is most likely) was a few degrees above 60, then the true specific gravity of the gas would be exactly 2-5, and his experi- ments would coincide with my own. We may therefore, without hesitation, adopt 2*5 as the true specific gravity of dry and pure chlorine gas. As muriatic acid is a compound of equal volumes of chlorine gas and hydrogen gas, and wei s hs 4>5 - as these gases unite in no other proportions, we cannot hesitate to consider muriatic acid as a. compound of 1 atom chlorine and 1 atom hy- drogen. The atom of chlorine, therefore, must bear the same relation to the atom of hydrogen, that the specific gravity of chlorine gas does to that of hydrogen gas. But chlorine gas is 36 times heavier than hydrogen gas ; for 0-0694 x 80 CHLORINE AND IODINE. 36 = 2-5. The atomic weight of hydrogen is 0-125. Consequently that of chlorine is 0-125 X 36 = 4-5. It is obvious too, that the volume of chlorine gas bears the same relation to the atom of chlo- rine that the volume of hydrogen gas does to the atom of hydrogen. If, therefore, we may substi- tute a volume of hydrogen gas for an atom, we may, in like manner, substitute the volume of chlorine gas for the atom of chlorine. -^ -^ * s easv t show by simple and decisive chloride of experiments, that 4-5 is the true atomic weight of chlorine. Of these, it may be worth while to state one or two in this place. 1. An atom of copper, as will appear here- after, weighs 4 ; an integrant particle of sub- oxide of copper weighs 9 ; and an integrant particle of black oxide, 5. Each of these oxides combines with muriatic acid. The protomuriate of copper is white, and the permuriate green. Both of these muriates crystallize, and both con- tain water of crystallization. When the mu- riates of copper are heated without the contact of air, they not only lose their water of crystalli- zation ; but the oxide of copper, and the muria- tic acid lose respectively, the oxygen and the hydrogen which they contain. These two con- stituents unite together, and make their escape in the form of water, leaving behind a com- pound of chlorine and copper. In the suboxide CHLORINE. 81 of copper there are two atoms of copper united to one atom of oxygen, and the protomu- riate of copper is a compound of one integrant particle of muriatic acid, and two integrant parti- cles ofsuboxide of copper, and the permuriate of one atom muriatic acid, and one atom black oxide of copper. Now, that this is the true constitution of these two salts, I have ascertained by a care- ful analysis of them. These facts being understood, we are pre- pared for examining the chlorides of copper. Of these chlorides, there are two the protochlo- ride, obtained by heating the protomuriate in a close vessel ; and the deutochloride, formed when the permuriate is treated in the same way. Now, Dr. Davy analyzed the protochloride of copper, and found it a compound of Copper 8 Chlorine 4-5 12-5* Two atoms of copper weighing eight, 4%5 must represent the weight of an atom of chlorine. I dissolved 5 grains of black oxide of copper in muriatic acid, and evaporated the solution to dryness. The chloride of copper was obtained in the state of a brownish yellow powder, and it weighed exactly 8-5 grains. Now, 5 grains of * Phil. Trans. 1812. p. 170. Vor. I. F CHLORINE AND IODINE. black oxide are equivalent to 4 grains of metal- lic copper. So that the chloride was a com- pound of Copper 4 Chlorine 4' 5 4*5 is evidently one atom of chlorine, and 4 one atom of copper. In this chloride one atom of copper is united with one atom of chlo- rine. Thus it appears, that an atom of chlorine unites with either one atom or two atoms of copper; and that an atom of chlorine weighs exactly 4*5. rideofmer- % Jt wil1 be shown hereafter, that an atom of cur y> mercury weighs &5. Now, mercury, like copper, combines with two proportions of chlorine, for- ming a protochloride, usually called calomel, and a deutochloride, usually called corrosive sublimate. Calomel is composed of Mercury 25 Chlorine 4-5 29-5 and corrosive sublimate of Mercury 25 Chlorine 9 34* * See Annals of Philosophy, (second series) II. 127. 9 CHLORINE. 83 We see that the first is a compound of one atom mercury and one atom chlorine, and the second of one atom mercury and two atoms chlorine. Here, as well as with copper, 4'5 represents exactly the weight of an atom of chlorine. 3. Silver is a metal, the atomic weight . , chloride of which is 13 tr ?5. Dissolve 13*75 grains of pure silver. silver in nitric acid ; evaporate the solution gently to dryness, and re-dissolve the nitrate of silver in water ; precipitate the silver by pouring into the liquid a solution of 7'5 grains of pure and dry common salt ; collect the pre- cipitate, wash it well, and fuse it on a piece of glass : if the experiment be properly conducted, the chloride of silver, thus obtained, will weigh 18*25 grains.* Now, it contains exactly 13*75 grains of silver. Hence, it is composed of Silver 1375 Chlorine 4-5 or in this compound, an atom of silver is com- bined with an atom of chlorine. In all these examples, the atomic weight of chlorine is exactly 4%5. 4. It has been long known, that when chlo- * The chloride of silver, while on the filter, must be washed with distilled water. This water should be very slightly acidulated with nitric acid. For want of this precaution, I was long unable to obtain the exact weight of chloride. The reason is, that distilled water often contains ammonia, which dissolves a little chloride. F2 84 CHLORINE AND IODINE. rine gas is collected over water in temperatures chlorine, below 40, a number of green coloured crystals form. These crystals were considered as solid chlorine, till Sir. H. Davy showed, in 1810, that they were a compound of chlorine and water, or a hydrate of cftlorine. During the severe winter of 1822-3, this hydrate was analyzed by Mr. Faraday, who found it to be composed of Chlorine 27'7 or 4-5 Water 72.3 11-745 100-0 * 11*745 exceeds 10 atoms of water by 0'495, which is rather more than ^ of an atom. It was doubtless hygrometrical water. These crystals, then, must be considered as constituting a deca- hydrate of chlorine composed of 10 atoms water 11-25 1 atom chlorine 4-5 15-75 Chlorine II. Chlorine, according to the experiments of x= 5 and x = 3-8888* ~ = 0-9722 == specific gravity of azotic gas.* * The specific gravity of oxygen and azotic gases might be deduced directly, from the known specific gravity of air ; viz, 1 -000, and the pro- portion of these two gases, viz. 80 azotic and 20 oxygen gas in 100 volumes of air, as Dr. Prout has already done, (Annals of Philosophy, VI. 322.) Let the atom of oxygen = 1 azote =1-75 G2 100 AZOTIC GAS. Atomic 2. I have been thus particular in determining weight of . -. * nitric acid the specific gravity of azote, because it is a ne- ' cessary element in the investigation of its atomic weight. We shall now determine the atomic weight of nitric? acid, which is known to be a compound of azote and oxygen ; and it contains the greatest proportion of oxygen with which an atom of azote is capable of uniting. Nitric acid Air is a compound of one atom of oxygen and two atoms azote ; or of Oxygen 1 or 22-22 Azote 3-5 77-77 100 Let x = specific gravity of oxygen 22-22 = a y = azotic gas 77-77 = b then *+ 4 ff 1 and x = 5 4 y o . 4 a V and x : 4 y : : a : b and x = - Hence 5 _4 = and y = = 0-9722 and* = 5 4z/ = 5 3-8888= Mill I have thought it better to proceed as I have done in the text, because the atomic weight of azote is most easily deduced from the specific gravity of azotic gas. Berzelius and Dulong, in their paper on the specific gravity of gases, give us the specific gravity of oxygen and azotic gases,' as follows : Oxygen 1 -1026 Azotic 0-9760 Now, as air is a compound of 4 volumes azotic and 1 volume of oxygen 1-1026 4- 0-9760 X 4 gas, it is obvious that - - p - = specific gravity or com- o mon air. If we resolve the equation, we find the specific gravity of com- mon air not 1, as it ought to be, but 1-00132; showing that the specific gravities of oxygen and azotic gases, as determined by these chemists, can- not be correct. AZOTIC GAS. 101 has the property of combining with the various bases, and of forming salts distinguished by the generic name of nitrates. Several of these nitrates contain no water of crystallization, but consist of the anhydrous acid united to the base ; this is the case with nitrate of lead and nitrate of po- tash, I shall make use of nitrate of lead to de- termine the atomic weight of nitric acid. It will be shown in a subsequent chapter, that From ni - the atom of lead weighs 13, and the atom of pro- toxide of lead, which constitutes the base of ni- trate of lead, 14. This salt is neutral, and con- sequently consists of one atom of nitric acid uni- ted to one atom of protoxide of lead. It is white, opaque, has a strong lustre, and crystallizes in octahedrons. Sometimes it may contain a little water mechanically lodged between its plates ; but this is not the case when it is properly crys- tallized ; and it may be always rendered anhy- drous by exposure for a few hours to a heat of Sulphate of soda is a salt composed of one atom sulphuric acid, one atom soda, and ten atoms water : but it is with great facility freed from its water of crystallization. A few days' exposure in a dry room is sufficient for the pur- pose ; or we may place it in the vacuum of an air-pump, in a vessel raised an inch or two above a flat plate filled with sulphuric acid. Twenty- four hours of such a situation renders it quite G2 102 AZOTIC GAS. anhydrous ; a moderate heat drives off the water of crystallization with great facility ; and we may afterwards fuse it, to be absolutely certain that all traces of water are removed. When this salt is anhydrous, it is composed of 1 atom sulphuric acid = 5 1 atom soda = 4 9 so that nine grains of it contain exactly five grains of sulphuric acid. The nature and constitution of these two salts being understood, we are prepared to investigate the weight of an integrant particle of nitric acid. I took two small glass vessels, containing each a little distilled water ; in the one, I dissolved 20*75 grains of pure nitrate of lead, and in the other, 9 grains of anhydrous sulphate of soda. These two solutions being mixed together, a double decomposition took place ; the sulphuric acid and oxide of lead uniting together, and forming an insoluble sulphate of lead, which fell rapidly to the bottom, while the nitric acid and soda remained in solution in the colourless li- quid. As soon as the sulphate of lead had left the liquid quite clear and transparent, a few drops of it were put into two different watch glasses y into one of these, I let fall a little of AZOTIC GAS. 108 the solution of muriate of barytes, and into the other, a little of the solution of sulphate of soda. But no precipitation, or even opalescence, was produced by either of these re-agents ; showing that the liquid contained no sensible quantity of sulphuric acid, or of oxide of lead. From this experiment it follows, that the sulphuric acid in nine grains of anhydrous sulphate of soda, is just sufficient to saturate the oxide of lead in 20*75 grains of nitrate of lead ; but nine grains of anhydrous sulphate of soda contain five grains of sulphuric acid, which re- quires just fourteen grains of protoxide of lead to saturate it. Accordingly, if the sulphate of lead precipitated be washed, collected, and dried, without any loss, it will be found to weigh exactly nineteen grains. Thus, it is proved, that 20*7<5 grains of nitrate of lead contain 14 grains of protoxide of lead. So that nitrate of lead is composed of Protoxide of lead 14 Nitric acid 6-75 20-75 The salt being neutral, and 14 denoting an integral particle of protoxide of lead, it is clear that the atomic weight of nitric acid is 6*75. 3. The importance of an accurate knowledge A d nitrate . , . . , of silver. ot the atomic weight of nitric acid is so great, G 4 104 AZOTIC GAS. that it may not be amiss to give an equally rigid proof of what it really is, drawn from the com- position of nitrate of silver, a salt very frequently employed in chemical analyses. Nitrate of silver is a white transparent salt, which may be crystallized in beautiful rhom- boids ; but it most commonly assumes the form of thin plates, which have been compared to sword blades. It contains no water of crystalli- zation, though sometimes a little may be me- chanically lodged between the plates of the crys- tals. From this it is easily freed, by fusing it in a moderate heat, and keeping it for some time in that state. The small cylinders of it sold by apothecaries,' under the name of lunar caustic, are quite anhydrous. But I have never met with them perfectly pure ; they usually contain a little gold, which separates in flocks, when the cylinder of nitrate of silver is dissolved ; and not uncommonly, other impurities (particularly nitre) are likewise present. For the experiment which I am going to relate, the salt must not only be anhydrous, but quite pure. The atomic weight of silver, as wfll be shown hereafter, is 13'75, and an integrant particle of oxide of silver weighs 14<'7<5. Consequently, if the atomic weight of nitric acid be 6'75, as we have found it to be from the analysis of nitrate of lead, the constituents of nitrate of silver must be AZOTIC GAS. 105 Nitric acid 6-75 Oxide of silver 14-75 21-5 When common salt is fused, it is converted into chloride of sodium, which is composed of Chlorine 4*5 Sodium 3 When 7*<5 parts of chloride of sodium are dis- solved in water, they combine with 1-125 of wa- ter, and are converted into muriate of soda, composed of Muriatic acid 4-625 Soda . 4 8-625 These facts being understood, we may pro- ceed to the experiment. I dissolved 21 %5 grains of fused nitrate of silver, and J-5 grains of fused common salt, in separate portions of distilled water, and mixed the solutions. A double de- composition took place ; the muriatic acid and oxide of silver united together, and precipitated in the form of a heavy, white, curdy matter ; while the nitric acid and the soda remained in solution, in the state of nitrate of soda. When 106 AZOTIC GAS. the precipitate had subsided, and left the super- natant liquid quite clear, a few drops of it were put into two separate watch glasses ; into the one a little nitrate of silver was let fall, and into the other a little common salt. The liquid was not precipitated nor rendered opalescent by ei- ther of these re-agents, showing, that it con- tained neither muriatic acid nor oxide of silver. From this experiment we see, that the muriatic acid in 8*625 grains of muriate of soda, just sa- turates the oxide of silver in 21 '5 grains of nitrate of silver. But 4-625 of muriatic acid require just 14*75 of oxide of silver for saturation. This is, of course, the quantity of oxide in 21*5 of fused nitrate of silver the remainder, a- mounting to 6*75, must be nitric acid. Thus we have found the composition of nitrate of sil- ver exactly agreeing with our hypothesis, de- monstrating that 6*75 is the atomic weight of nitric acid. When the muriate of silver, obtained in the preceding experiment, is washed, dried, and fused, it is converted into chloride of silver; consequently, 4-625 muriatic acid become 4>'5 chlorine. 14-75 oxide of silver become 13-75 silver. 18-25 Hence, if it be collected without any loss, which AZOTIC GAS. 107 is a difficult process, the weight, instead of be- ing 19-375 grains, will only be 18-25 grains. From these two experiments, we are entitled to affirm, that the atomic weight of nitric acid is 6-75. 4. To demonstrate the true composition of4 nal y sisof 1 nitre nitric acid, I shall employ nitrate of potash. This salt, when pure, contains no water of crystallization ; but a little water is often me- chanically lodged between the plates of the crys- tals. This water is easily dissipated by keeping the salt for some time in a moderate heat ; or simply, by bringing it into fusion. The salt upon which the following experiments were made was pure, and perfectly anhydrous. There is a particular temperature, higher than By heat; 600, but I do not know how much, in which, if saltpetre be kept, it allows a certain portion of oxygen gas to escape and no more. When 127J- grains of nitre are kept in this temperature dur- ing the requisite time, there are extricated 29 \ cubic inches of pure oxygen gas ; the weight of the nitre is reduced to 117i grains, and the salt is no longer nitrate of potash, but nitrite of pot- ash ; for when acetic acid is poured on it, red fumes are extricated, which is not the case with saltpetre. Now, it will be shown in a subsequent part of this chapter, that the atomic weight of ni- trous acid (the acid in nitrite of potash) is 5*75. It will appear immediately, that nitric acid is com- 108 AZOTIC GAS. posed of 5 oxygen and 1-75 azote ; consequent- ly, 29^ cubic inches, weighing 10 grains, is the fifth part of the whole oxygen in 127J grains of nitre. This makes the whole oxygen in 127J grains of nitre to amount to 147*5 cubic inches, or 50 grains. But though I have made many attempts to obtain the whole oxygen and azote in a given weight of nitre, in the gaseous state, I have only met with partial success. I never succeeded in decomposing nitre in a porcelain retort, without destroying the vessel before the decomposition was completed. The potash, as it is evolved, acts so powerfully upon the vessel, at the high temperature which must be kept up, that it soon makes its way through it, and enables the gaseous products to escape. When the nitre was mixed with twice its weight of siliceous sand, I decomposed it with facility in a common stone-ware retort ; but the retort, in the high temperature to which it was necessary to raise it, was pervious to the air. This prevented me from determining with preci- sion the proportion between the volumes of the two gases evolved and Wedgewood retorts, though air tight, always cracked before the pro- cess was completed. In general, the gaseous products consisted of 1 volume azotic gas 2-3 volumes oxygen gas AZOTIC GAS. 109 And the proportion of azotic gas was always found to increase with the length of the process. I came nearer the truth with a small quantity of nitre than with a large quanity but I was never able to obtain the whole volume of gas which nitre is capable of giving out. I next tried to decompose nitre by heating it, By char- mixed with charcoal powder, in a copper tube. c The tube was heated to redness, and the gaseous products collected over mercury. By this pro- cess, I obtained a greater volume of gas than nitre itself is capable of giving out ; in some cases, the volumes of the azotic gas and carbonic acid gas, evolved, were to each other as fol- lows : 1 volume azotic 2^ volumes carbonic acid. But in general, the gas, unabsorbable by potash, bore a greater ratio to the carbonic acid gas than the preceding ; and the longer the heat was con- tinued, the greater was the quantity of unab- sorbable gas evolved. There was always evolved, likewise, a very considerable quantity of water. It was impossible to doubt, that the charcoal (though I had previously heated it strongly in a wind furnace) contained both oxygen and hy- drogen ; and, when I substituted Kilkenny coal for charcoal, my results were not more satisfac- tory. With Kilkenny coal, the proportion of 110 AZOTIC GAS. water formed was greater, and of carbonic acid smaller, than when charcoal was used. By iron My next process was to mix together nitre and clean iron filings to introduce the mixture into a copper tube and to expose it to a red heat. In this process, the oxygen of the nitre united with the iron filings, while the azote was evolved in the gaseous state. Four experiments made in this way very nearly agreed with each other. I found, that 8*65 grains of nitre gave out four cubic inches of azotic gas, supposing the baro- meter to stand at 30 inches, and the thermome- ter at 60. It will be more convenient to divide these quantities by 4 ; this will give, as a result, 1 cubic inch of azotic gas from 2-1625 grains of nitrate of potash. Now, since the specific gravity of azotic gas is 0-9722, a cubic inch of it weighs 0-296527 grain ; and this is the weight of azote in 2-1625 grains of nitre. But nitre is anhydrous : the atomic weight of nitric acid is 6*75, and of potash 6 ; conse- quently, nitrate of potash is composed of Nitric acid 6- 75 Potash 6 12-75 And 2-1625 grains of nitre contain 1*144852 grain of nitric acid. The preceding experiments AZOTIC GAS. Ill give us the weight of azote in this quantity of nitric acid, and the remainder of the weight must be oxygen ; consequently, nitric acid is composed of Azote 0-296527 or 1-7477 Composi- Oxygen 0-848824 - 5 % 1-144852 6-7477 But there is obviously a slight inaccuracy in this analysis, for it gives us the atomic weight of ni- tric acid 6 '7^77 > instead of 6*75, which we found above to be the true quantity. It is clear, that this error lies in the azote, which is a little too low, and not in the oxygen, which must consti- tute a whole number of atoms. Had I obtained from 8-65 grains of nitre 4*004 cubic inches of azotic gas, instead of 4 cubic inches, this error would not have existed. But my apparatus was not delicate enough to measure the gas evolved so exactly I may, in reality, have obtained 4*004 cubic inches, without perceiving the slight difference in volume. There can be no doubt, from these experi- ments, that the true composition of nitric acid is Azote 1-75 Oxygen 5 6 ' 75 Weight of ^ the atom of Ur it is composed of 1 atom of azote and 5 azote. 112 AZOTIC GAS. atoms of oxygen ; and an atom of azote weighs 1-75. It appears from an experiment, which I have stated above, that 8-65 grains of saltpetre con- tain 4-004 cubic; inches of azotic gas. Now, as nitre is a compound of 6 potash and 6*75 nitric acid and as nitric acid is a compound of 1*7^ azote and 5 oxygen it is obvious, that 8 '65 grains of nitre contain Azote 1-187254 gr. or 4*00386 cubic inches, Oxygen 3-392155 - - 10-00960 cubic inches. 4-579410 Now, 4-00386 : 10-00960 : : 1 : 2-5. So that the constituents of nitric acid, when converted into gases, consist of 1 volume azotic gas 2^ volumes oxygen gas. But we have seen already, that half a volume of oxygen gas is equivalent to an atom. It is ob- vious, that a volume of azotic gas is equivalent to an atom ; in this respect, azotic gas agrees with hydrogen gas, chlorine gas, and the vapour of iodine. The specific gravity of azotic gas is obtained by multiplying its atomic weight, 1*75, by 0-5555. The product is 0-9722, which we have found to be the true specific gravity of this gas. AZOTIC GAS. 113 Nitric acid is the compound of azote and oxy- gen, which contains the maximum quantity of oxygen. It is obtained by heating a mixture of Method of sulphuric acid and nitre, in a retort, to which receiver is luted. When 6 parts of the strong- est sulphuric acid of commerce are mixed with 12f parts of nitre, thoroughly freed from water, and perfectly pure, the acid which comes over is the strongest that can be procured. Its speci- fic gravity is about 1-55, and it is a compound of 1 atom acid and 1 atom water, or it contains Real acid 8 Water 14-286 parts 100-000 When 12i parts of sulphuric acid are mixed with 12f parts of pure anhydrous nitre, the whole ni- tric acid may be obtained without any loss ; but its specific gravity is only 1-4855, and it is a compound of 1 atom real acid, arid 2 atoms water. When nitric acid is quite free from nitrous gas, it is colourless, and nearly as limpid as wa- ter ; but the presence of this gas gives it a yellow, a red, or a brown colour, according to its quan- tity. This gas greatly increases the volatility of the acid, but its effect upon the specific gravity is not so great as I had anticipated the volume of the acid being increased, nearly in proportion to the quantity of gas absorbed. VOL. I. H 114 AZOTIC GAS. Table exhi- biting its strength. The following table, from a set of experiments which I made with great care, exhibits the spe- cific gravity of various atomic compounds of real nitric acid and water. The experiments were made, by adding given quantities of water to acid, of the specific gravity 1'4855, and taking the specific gravity of the mixture. The strong- est acid which I was able to obtain from nitre, had the specific gravity 1%534. By determining the composition of this acid, and observing the rate at which the specific gravity diminishes with the proportion of water added, I concluded, that acid composed of 1 atom real acid, and 1 atom water, has the specific gravity 1-55. Atoms of Acid. Atoms of Water. Acid in 100. Specific Gravity. 1 85-714 1-55 2 75-000 1-4855 3 66-668 1-4546 4 60-000 14237 5 54-545 1-3928 6 50-000 1-3692 7 46-260 1-3456 8 42-857 1-3220 9 40-000 1-3032 10 37-500 1-2844 11 35-294 1-2656 12 32-574 1-2495 13 31-579 1-2334 14 30-000 1-2173 15 28-571 1-2012 The acid, composed of 1 atom of acid and 4 atoms of water, seems to be the most intimate, for its boiling temperature is the highest, being 248 while acid, both stronger and weaker than AZOTIC GAS. this, boils at a lower temperature. This is the strength of the acid usually met with in com- merce ; it contains exactly 60 per cent, of real acid, and 40 per cent, of water. Acid, of the specific gravity 1*3692, contains just half its weight of water ; it boils at the temperature of 543. Acid of the specific gravity 1-3032, may be considered as just the reciprocal of acid of the specific gravity 1*4237, for it consists of 40 acid + 60 water ; whereas, the latter consists of 60 acid + 40 water. Its boiling point is 236, or 1 2 lower than that of the acid, whose boiling point is a maximum. These boiling points were determined by Mr. Dalton. I shall now take a view of the remaining; com- Other com - pounds of pounds ot azote and oxygen ; these are protox- azote and ide of azote, deutoxlde of azote, and nitrous add. * The knowledge of the atomic weight of azote, and the specific gravity of azotic gas, will enable us to make out the composition of these com- pounds without difficulty. 1. Protoxide of azote, the nitrous oxide of Analysis of -rx 11 ...protoxide Davy, is usually obtained by exposing nitrate of of azote. ammonia to a heat of about 400, or a little high- er it is a colourless gas, having a sweetish taste, and water absorbs about three fourths of its volume of it. If you mix 100 volumes of pure dry protoxide of azote, with 100 volumes of dry hydrogen gas, and pass an electric spark through the mixture, a detonation takes place, H 2 116 AZOTIC GAS. and some water is deposited upon the inside of the eudiometer. If the experiment be made when the barometer stands at 30 inches, and the thermometer at 60, the residual gas will amount to 101 '77 volumes. But this gas is sa- turated with humidity if it were quite dry, its volume would be 100.* its consti. Thus it appears, that when protoxide of azote tution and atomic is mixed with its own volume of hydrogen gas, and fired, the gaseous residue, which is azotic gas, is just equal to the original volume of the gas the hydrogen gas has disappeared and been converted into water it must, therefore, have combined with a quantity of oxygen, which, had it been in the state of gas, would have amounted to half the bulk of the hydrogen gas. The con- stituents of protoxide of azote then are 1 volume of azotic gas 1 _ V condensed into one volume. volume or oxygen gas ) But a volume of azotic, and i a volume of oxy- gen gases, are each equivalent to an atom ; hence this gas is a compound of 1 atom azote + 1 atom oxygen, so that its atomic weight is 2*75. * Let the volume of dry gas be x, and let;> = 30 ; and /= 0*524. We have =101.77 The solution of this equation will give .r = 100. 9 AZOTIC GAS. 117 Its specific gravity is obviously obtained by adding together the specific gravity of azotic gas, and half the specific gravity of oxygen gas. Specific gravity of azotic = 0-9722 J specific gravity of oxygen = 0-5555 1-5277 I took the specific gravity of this gas three times successively, and obtained each time 1%5269 This is as near an approximation to the truth as I could reach with my balance and is amply sufficient to satisfy us, that the specific gravity above deduced, from the known specific gravities, volumes, and condensation of its two constituents, is correct. 2. Deutoxide of azote, the nitrous air of Analysis of Priestley, has been long known, and for many O f azote, years constituted a favourite object of research among chemists. Sir H. Davy was the first per- son who attempted an accurate analysis of it. From his experiments .(Researches, p. 122.) it was evident, that it consisted very nearly of equal volumes of oxygen and azotic gases. Gay-Lus- sac afterwards demonstrated, that its constituents are 1 volume of oxygen, and 1 volume of azotic gases, united together without any condensation. He heated some potassium in 100 volumes of this gas ; after the combustion was at an end, and the whole apparatus cooled down to the H3 118 AZOTIC GAS. temperature of the apartment, the residual gas, which was pure azote, amounted to exactly 50 volumes,* the potassium had absorbed the oxy- gen, which obviously constituted half the volume of the original *gas. Now, since j a volume of oxygen gas is equivalent to an atom, and 1 vol- ume of azotic gas to an atom, it is evident, that deutoxide of azote is a compound of 2 atoms oxygen, and 1 atom azote ; consequently, its atomic weight is 3*75. its consti. 'phg specific gravity of this gas is the mean tution and J . weight. between the specific gravity or oxygen and azotic gases, or 1/0416. For Sp. gr. of oxygen gas 1-11 11 Sp. gr. of azotic gas 0-9722 2 ) 2-0833 1 -04-16' = sp. gr. of deutoxide of azote. I took the specific gravity of this gas, prepared with as much care, to secure its purity, as possi- ble. The mean of three trials, differing from each other only by unity in the fourth decimal place, was 1-04096. The small difference be- tween this result and the preceding, deduced from theory, which does not exceed y^Wth part, is probably to be ascribed to the presence of a little azotic gas, with which it is extremely diffi- * Mem. d'Arcueil. II. 216. 9 AZOTIC GAS. 119 cult to prevent it from being contaminated, when the gas is prepared over water. Berzelius and Dulong, in their paper on the Specific Gravity of Gases, state its specific grav- ity to be only 1-0010. But^this statement can- not be accurate ; for it is inconsistent with the specific gravities of oxygen and azotic gases, which they themselves have given. According to them, the specific gravity of Oxygen gas = 1 1 026 Azotic gas = 0-9760 2 ) 2-0786 1-0393 = sp. gr. of deutoxide of azote, de- duced from these data. These gentleman have rated the specific gra- vity of deutoxide of azote a great deal too low. If 1 volume of protoxide qf azote, and half a volume of oxygen gas, were to unite and be re- duced to fth of a volume, we should obtain a compound, having the constituents and specific gravity of deutoxide of azote. For Sp. gr. of protoxide of azote = 1-5277 sp. gr. of oxygen gas =0-5555 2 ) 2-0833 1-04-16= sp. gr. of deutox- ide of azote. H 4 120 AZOTIC GAS. Analysis of 3. Nitrous acid is an orange coloured liquid, acid. which may be obtained in a state of purity, by exposing dry nitrate of lead to heat in a retort, while the receiver is surrounded with a mixture of snow and sak. It is a liquid, which, when newly prepared, has an orange colour ; but I have never been able to prevent it from acquir- ing a green colour a change produced by the absorption of a very small quantity of water. This acid was first obtained in a separate state by Berzelius Gay-Lussac first ascertained its na- ture Dulong analyzed it, by passing it through red hot copper wire. The oxygen united to the copper, while the other constituent, the azote, passed off in the state of gas. By this method of proceeding, he obtained its constituents as , follows : Azote 1-75 Oxygen 4-1 78 This is rather more than 4 atoms of oxygen united to 1 atom of azote. But that the true composition of nitrous acid, is 1 atom of azote and 4 atoms of oxygen, is obvious from the fol- lowing experiment, which I have often repeated. Introduce into a cylindrical glass tube, 18 inches long and 0*9 inch in its internal diameter, shut at one end, filled with water, and standing inverted over a water trough, 100 volumes of common air ; then let up to this air 100 vol- AZOTIC GAS. umes of deutoxide of azote ; the gaseous mix- ture assumes an orange colour, and diminishes rapidly in volume. When the gas has become again colourless, and has ceased to diminish, its bulk will be found to amount to 138 volumes. To be able to understand what really takes place in this experiment, we must recollect that common air consists of 20 volumes oxygen gas, and 80 volumes azotic gas ; and that the yellow colour is occasioned by the combination of the oxygen of the common air, with the deutoxide of azote and the formation of nitrous acid. The whole oxygen of the air disappears, and the amount of the residual gas will enable us to judge of the volume of deutoxide, with which it com- bines. The experiments were made over water at the temperature of 60. Had the air and deutoxide been dry, instead of saturated with moisture, each would have occupied only 98*253 volumes instead of 100 ; so that the whole gaseous matter used, had it been dry, would have occupied 196*506 volumes. The 138 volumes of residual gas, sup- posing them dry, would be reduced to 135*589 volumes. This last quantity subtracted from 196*506 (the original volume) leaves 60*91? vol- umes for the quantity absorbed or converted into nitrous acid. Now 98*253 volumes of common air, contain 19*65 volumes of oxygen, which, in order to be 122 AZOTIC GAS. converted into nitrous acid, must have combined with 41 '267 volumes of deutoxide of azote. According to this experiment nitrous acid is a compound of 19'65 volumes of oxygen gas 41'267 volumes deutoxide of azote Or, 20 volumes oxygen gas 42 volumes deutoxide of azote 42 exceeds 20 x 2 only by 2 volumes and these 2 volumes are absorbed by the air in the wa- ter through which it was passed. This air al- ways causes a little of the deutoxide of azote to disappear, and renders its absorption in all ex- periments made in this way a little greater than the truth. From the preceding experiment, then, we may conclude that nitrous acid is composed of 1 volume oxygen gas 2 volumes deutoxide of azote. But we have seen, in a preceding part of this chapter, that deutoxide of azote is composed of 1 volume of azotic and 1 volume of oxygen gas, united, without any condensation. It is obvious, then, that nitrous acid is composed of 1 volume azotic gas 2 volumes oxygen gas. AZOTIC GAS. But a volume of azotic gas being equivalent to an atom, and half a volume of oxygen gas equi- valent to an atom, nitrous acid is obviously com- posed of 1 atom azote Its atomic 4 atoms oxygen weight. Hence its atomic weight is 5*75. Such are the atomic weights and constitutions Compounds of all the known compounds of azote and oxy- with gen. Azote appears likewise capable of uniting with chlorine and iodine ; the compounds formed possess violently detonating properties, and have not hitherto been applied to any useful purpose. I have not myself made any experiments to de- termine the composition of these substances. Davy made some experiments on the chloride of azote, from which he deduced its constituents to be 4 atoms chlorine == 18 1 atom azote = 1*75 19-75 Gay-Lussac, from theoretical considerations, con- siders the iodide of azote to be a compound of 3 atoms iodine = 36-5 , 1 atom azote =' 1'75 37-25 It would be very difficult to analyze it, for it de- AZOTIC GAS. tonates on the slightest motion ; often, indeed, spontaneously without any agitation whatever. But azote enters into a combination with hy- drogen, of considerable interest in a chemical point of view, a it constitutes ammonia, or vola- tile alkali, so much employed in chemical analy- ses, and so often formed during the decomposi- tion and putrefaction of animal substances. Ammonia Ammonia, in a state of purity, is a gaseous sub- stance, easily obtained by heating a mixture of sal-ammoniac and quicklime. 3i parts of pure quicklime are sufficient to decompose 6| parts of sal-ammoniac ; but it is better to add a consider- able excess of quicklime. Indeed, I always fill the small retort in which the sal-ammoniac is put with quicklime, which is of service by pre- venting the two speedy extrication of moisture, which stops the evolution of the gas. Ammoniacal gas must be received over mer- cury ; for it is the most absorbable of all the gases by water, 1 volume of that liquid absorb- ing 780 volumes of ammoniacal gas. Analysis of When pure dry ammoniacal gas in a glass tube, standing over mercury, is exposed to a continued series of electrical sparks passed through it, decomposition takes place ; its volume is just doubled ; and it is found to consist of a mixture of one volume azotic and three volumes hydrogen gas. This experiment was first made by M. Berthollet senior. It was afterwards re- AZOTIC GAS. 125 peated with great care by M. Berthollet junior, Sir H. Davy, and Dr. Henry.* What I have just stated, is the obvious deduction from these elaborate and difficult experiments. As in azo- tic and hydrogen gases, volumes are equivalent to atoms, it follows that ammonia is a compound of 1 atom azote = 175 3 atoms hydrogen = 0-375 2-125 So that its atomic weight is 2*125, its atomic Ammoniacal gas has been shown to be a com- pound of one volume azotic and three volumes hydrogen gas, condensed into two volumes. Its specific gravity, of course, will be obtained by adding together the specific gravity of azotic gas, and three times the specific gravity of hy- drogen gas, and dividing the sum by two. This gives us O59027 for the true specific gravity of this gas. Sir H. Davy took the specific gravity of this gas with great care, and obtained for a result * Dr. Henry, in a paper on the Analysis of the Aeriform Compounds of Nitrogen, printed in 1824, but not yet published, has repeated the analysis of ammonia with still greater care, and has obtained results which scarcely differ from those in the text. The ammoniacal gas was doubled by the electric sparks, and the gas, thus increased in volume, was found a mixture of one volume azotic and three volumes hydrogen gas. Dr. Henry's analyses of protoxide of azote, deutoxide of azote, and nitric acid, fully confirm the statements in the preceding part of tliis chapter. 126 AZOTIC GAS. 0*590164. A set of experiments to determine the specific gravity of this gas, made with great care in my laboratory, gave (the mean of three trials) '59031. Davy found the specific gra- vity of this gasa little below the theoretical re- sult, while I obtained it a little above it but if we take the mean between these two specific gravities, we obtain 0*590237, which differs from the theoretical result by less than Tooooth part. There cannot be the least doubt that the atomic weight and specific gravity of ammoniacal gas have been determined with precision. But as the exact knowledge of the composi- tion of this substance is a most important point in the theory of chemistry, it will not be amiss to exhibit even a redundancy of evidence. I shall, therefore, add a few additional experiments. Analysis of 1. Oxalate of ammonia is a neutral salt, which oxalate of . . p . ammonia, crystallizes in beautitul transparent prisms. It is not very soluble in water. Its constituents I have found, by a careful analysis, to be 1 atom oxalic acid . 4-5 1 atom ammonia . 2-125 2 atoms water 2-25 8-875 8*875 grains of this salt were dissolved in a small quantity of distilled water. 6'25 grains of pure calcareous spar (equivalent to 3-5 grains of lime) AZOTIC GAS. 127 were dissolved in muriatic acid : the solution was evaporated to dryness, and the dry residual salt, constituting muriate of lime, was re-dissolv- ed in a little water. The two solutions being mixed, a double decomposition took place, and oxalate of lime subsided to the bottom. As soon as the supernatant liquid had become quite clear, it was tested by oxalate of ammonia, and by muriate of lime ; but was not rendered mud- dy by either of these reagents, showing that it contained no lime nor oxalic acid. From this it is obvious, that 8-875 grains of oxalate of ammo- nia contain just the quantity of oxalic acid re- quisite to saturate 3*5 grains of lime. Now, 3 '5 being the atomic weight of lime, the oxalic acid in 8-875 grains of the oxalate must be the equi- valent of an atom, or 4-5 ; for it will be shown afterwards that 4*5 is the atomic weight of oxalic acid. The liquid from which the oxalate of lime had precipitated was neutral : hence the muriatic acid in the muriate of lime was just capable of saturating the whole ammonia in the 8 '875 grains of oxalate of ammonia. Now, this muriatic acid weighed exactly 4*625 grains. And it will be shown in the next paragraph, that 4 -625 grains of muriatic acid just saturate 2' 125 grains of ammonia. This, therefore, is the quantity of ammonia in 8-875 grains of oxalate of am- monia. 128 AZOTIC GAS. We have thus determined the weight of acid and ammonia in 8-875 grains of oxalate of am- monia. The surplus weight being undoubtedly water, it is obvious that the constituents of oxa- late of ammonia are 1 atom oxalic acid . 4>-5 1 atom ammonia . 2-125 2 atoms water . 2-25 8-875 The atomic weight of ammonia in this salt is undoubtedly 2-125. g. Sal-ammoniac, when newly sublimed, or . ~ when dried tor some time upon the sand bath, is an anhydrous salt. It is neutral ; and, there- fore, a compound of one atom muriatic and one atom ammonia. 1 atom muriatic acid . 4-625 1 atom ammonia . 2-125 6-75 6-75 grains of pure dry sal-ammoniac were dis- solved in water j 21*5 grains of pure anhydrous nitrate of silver were dissolved in another portion of water, and the two solutions were mixed. A double decomposition took place, and chloride of silver precipitated. As soon as the residual liquid had become clear, it was tested by nitrate of silver and common salt. Neither of these AZOTIC GAS. salts produced any effect, if we except an almost imperceptible opalescence which appeared when the common salt was added ; but there was no precipitate whatever, even after the liquid had stood a week. From this experiment it is obvi- ous, that 6*75 grains of sal-ammoniac contain just 4 625 grains of muriatic acid ; for that is the quantity necessary to saturate the 14*75 grains of oxide of silver present in 21*5 grains of nitrate of silver. Hence the other constituent of the salt, the ammonia, must weigh 2*125, be- cause that is the weight wanting to make up the full quantity of sal-ammoniac employed ; and, as sal-ammoniac is neutral, and 4*625 the atomic weight of muriatic acid, 2*125 must be the atomic weight of ammonia. 3. 13*5 grains of dry sal*ammoniac were wrap- Analysis of ped up in blotting paper, and dropped into a re- tort filled with dichloride of lime, (Mr. Tennant's bleaching powder,) made into a thin paste with water. The whole retort and beak was then filled with water, and the beak of the retort was plunged into a water trough, under an inverted graduated jar, filled with water. As soon as the paper round the sal-ammoniac was sufficiently softened to allow the dichloride to come in con- tact with the salt, an effervescence took place, and azotic gas was disengaged. This is just the effect always produced when chlorine and am- moniacal gas come in contact. The lime which VOL. I. I 130 AZOTIC GAS, was in excess in the salt decomposed the sal- ammoniac ; and the ammonia, as it was evolved, came in contact with chlorine, and was decom- posed ; the hydrogen uniting with the chlorine, and the azote being disengaged in the gaseous state. The action is so violent, if the dry sal- ammoniac be dropped at once into the retort, that it is difficult to collect the whole gas ; but when the salt is wrapped in paper, the action is slow, and the gas may be all collected with the greatest facility. The azotic gas obtained in this process was 11*7 cubic inches, at the tem- perature of 47, and when the barometer stood at 29*93 inches. This is equivalent to 11-853 cubic inches of dry gas, of the temperature 60, and under a pressure of 30 inches mercury. This constitutes the whole amount of the azo- tic gas in 4*25 grains of ammonia, the quantity contained in 13*5 grains of dry sal-ammoniac. Now, 11 '853 cubic inches of azotic gas weigh 3*5147 grains : Hence it follows, that the weight of the other constituent, the hydrogen, is 0*7353 grain. Consequently, ammonia is composed of Azote 1-7573 or 1 volume Hydrogen 0-3676 2'9* 2-1250 The small excess of azote in this experiment was owing to a small admixture of common air with AZOTIC GAS. 131 the azote, in consequence of the gas standing 24 hours over the water. The experiment was repeated seven times, in various ways, and the mean of the whole came exceedingly near 11-8 cubic inches of dry azotic gas from 13*5 grains of sal-ammoniac. This weighs 3*4993 grains, giving us the composition of ammonia as follows ; Azote 1'74965 or 1 volume Hydrogen 0-37535 3-0028 2-12500 This analytical result of direct experiment comes within less than Ttuioth part of the theoretical estimate ; and, taken together with the preced- ing facts, can leave no doubt of the composition of ammonia. CHAP. VI. OF THE ATOMIC WEIGHTS OF THE ACIDI- FIABLE COMBUSTIBLES. THE simple bodies, which I have distinguished by the name of acidifiable combustibles, are the nine following : 1 Hydrogen, 2 Carbon, 3 Boron, 4 Silicon, 5 Phosphorus, 6 Sulphur, 7 Selenium, 8 Arsenic, 9 Tellurium. I have already treated of hydrogen in the first chapter of this work. I shall endeavour in this chapter to establish the atomic weights, and to point out the principal combinations of the re- maining eight. These bodies differ from all the other combustible substances in the property which they have of entering into gaseous com- pounds, either with hydrogen gas or fluoric acid, while none of the other combustible bodies enter into any gaseous combination whatever. An accurate knowledge of their atomic weights will be found important, not merely as connected with chemical analysis ; but because it throws an important light on the relation between the atomic weight and the specific gravity of gaseous bodies. 13 ACIDIFIABLE COMBUSTIBLES. SECT. I. OF THE ATOMIC WEIGHT OF CARBON. BRITISH chemists seem all agreed in opinion respecting the atomic weight of carbon. It is to the atom of oxygen as 6 to 8 ; consequently, if an atom of oxygen be reckoned unity, an atom of carbon must be represented by 0*75. Berzelius, who is followed by a great proportion of the chemists on the continent, considers the atomic weight of carbon to be 0'7533. As this number differs nearly 1 per cent, from 0*75, I have been at great pains to endeavour to deter- mine which of the two is nearest the truth ; and I shall here describe a set of experiments which I made, with that specific object in view, with every possible attention to the most minute ac- curacy. ^ was established by the experiments of La- its voisier, and has been amply confirmed by suc- volumes of . \ J J oxygen gas. ceedmg experimenters, that when carbon is burnt in oxygen gas, the volume of the gas un- dergoes no alteration ; but it is converted into carbonic acid. It is obvious from this that, if we subtract the specific gravity of oxygen gas from that of carbonic acid gas, the remainder will give us the weight of carbon with which a volume of oxygen gas has united ; and, of course, CARBON. 135 furnish us with the requisite data for determin- ing the exact composition of carbonic acid gas. We have three sets of experiments on the spe- cific gravity of carbonic acid gas, which appear to have been made with great care. The follow- ing are the results obtained : 8P. GR. OP CARBONIC ACID GAS. 1-5196 . by Biot and Arago. 1 '52673 . by my experiments. 1-5240 . by Berzelius and Dulong. ' These results differ too much from each other to enable us to deduce from them the composi- tion of carbonic acid, with the requisite degree of precision \ nor have we any criterion by which we can discover which of the three results is nearest the truth. My own determination is the highest, and that of Biot and Arago the lowest. Now, if any error resulted from the impurity of the gas weighed, the probability is, that it would diminish the specific gravity somewhat ; for, common air, the only gas likely to contaminate the carbonic acid gas examined, is a good deal lighter. So far the probability of accuracy ra- ther inclines to my own result ; but it is not sufficient to enable us to decide the point, far less to determine whether my own result be not in reality a little too low. This uncertainty suggest- ed to me a method similar to that by which the specific gravities of oxygen and hydrogen gases 14 13G ACIDIFIABLE COMBUSTIBLES. were determined in the first chapter of this work. I shall now lay the result of these experiments before the reader. i t It is well known, that, when calcareous spar tion of car- bonate of or carbonate of Jime is exposed to a white heat, it loses its carbonic acid, and is converted into quicklime. I exposed given quantities of very pure crystallized calcareous spar, inclosed in a platinum crucible, covered with a lid, to the greatest heat of a wind furnace, in order to de- termine the weight of carbonic acid which would be driven off. The loss of weight in different trials was not rigidly the same. However, by a number of experiments, varied by uniting the lime to sulphuric acid, and determining the weight of sulphate of lime remaining after igni- tion, I satisfied myself, that, when none of the lime is lost, and when the carbonate is anhy- drous, 100 grains of it, when deprived of its car- bonic acid, lose exactly 44 grains. The remain- ing 56 grains are lime. Carbonate of lime, then, is a compound of Carbonic acid . 44 or 1 1 Lime 56 14 100 25 It will appear afterwards, that the atomic weight of lime is 3\5. Now, 14 : 11 : : 3%5 : so that (if carbonate of lime be a com- CARBON. 137 pound of 1 atom lime and 1 atom carbonic acid) the atomic weight of carbonic acid is 2 '75. The following experiment, which I have repeated more than once, will show that this determina- tion is rigidly exact. 2. I must begin by statins*, what will be after- Atomic weight of wards proved, that an atom of potash weighs 6. carbonic This alkali combines in two proportions with car- mined! bonic acid, forming a carbonate, composed of 1 atom potash, united to 1 atom carbonic acid ; and a bicarbonate, composed of 1 atom potash, and 2 atoms carbonic acid. The latter of these salts crystallizes in large transparent and perma- nent crystals, not affected by exposure to the air. It can easily be procured in this state in this country, very nearly pure, being only occasion- ally contaminated by a little lime and a little iron. If we dissolve these crystals in distilled water, filter the solution, evaporate it to dry- ness, and expose the saline residue to a red heat, we obtain anhydrous carbonate of potash in a state of great purity. Dissolve 87i grains of this salt in water ; dissolve in muriatic acid 62i grains of pure anhydrous carbonate of lime ; eva- porate the solution to dryness, in a gentle heat, to drive off the excess of acid ; and dissolve the muriate of lime formed in distilled water. Thus we have two solutions, containing each determi- nate quantities of a particular salt : the first so- lution, 87i grains of carbonate of potash, and the 138 ACIDIFIABLE COMBUSTIBLES. second, a quantity of muriate of lime equivalent to 62i grains of carbonate. When these two so- lutions are mixed, a double decomposition takes place ; the lime and carbonic acid unite together, and precipitate jn the state of carbonate of lime, while the muriatic acid and potash remain in so- lution, constituting muriate of potash. If we examine the clear residual liquid after the car- bonate of lime has precipitated, we shall find in it no uncombined potash, and no carbonate of potash ; for, cudbear paper, dipped into it, is not in the least altered, although a very minute quantity of uncombined alkali, or alkaline car- bonate, gives it a violet stain. Neither does it contain any sensible quantity of lime ; for it is not in the least altered by the addition of a few drops of oxalate of ammonia. When the preci- pitated carbonate of lime is collected and dried, it weighs exactly 62 grains, or is just equal to the original quantity of carbonate of lime em- ployed a clear proof that it contains all the lime and all the carbonic acid of the two salts. This experiment will be more easily under- stood if we divide the quantities employed by 10. We thus get 8 'J5 grains of carbonate of potash, and 6*25 grains of carbonate of lime each constituting an integrant particle of the an- hydrous salt j for carbonate of potash is com- posed of CARBON. 139 Carbonic acid 2-?5 Potash 6 8-75 And carbonate of lime of Carbonic acid 2-?5 Lime 3-5 6-25 The carbonic acid was separated from this car- bonate, and an atom of muriatic acid, weighing 4'6&5, was substituted in its place forming mu- riate of lime, composed of Muriatic acid 4*625 Lime 3-5 8-125 The 3*5 lime and the 2*7<5 carbonic acid united together, and formed 6*25 carbonate of lime. In the same way, the 6 potash and the 4'625 muriatic acid united together ; and these respec- tive quantities of acid and base exactly neutral- ized each other : for both the carbonate of lime and the muriate of potash were perfectly neu- tral. It is clear from this experiment, that the atomic weight of carbonic acid is just 2'75. 3. Two facts of great importance have been as- certained by the preceding experiments ; name- ly, 1. That an atom of carbonic acid weighs 2*75. 2. That 100 grains of pure carbonate of lime contain exactly 44 grains of carbonic acid. 140 ACIDIFIABLE COMBUSTIBLES. next bject which engaged my attention carbonic wa s, to disengage the carbonic acid from 100 acid gas de- . & & grains of calcareous spar, and collect it over mer- cury, in order to ascertain its volume. This ex- periment appeals so easy, that the reader will be surprised to hear, that a considerable time elaps- ed, and many fruitless experiments were made, before I was able to make it with accuracy. The first method that I tried was, to fill a gra- duated glass jar with mercury, to let up into it some muriatic acid, and then to introduce a given weight of calcareous spar. By this method the carbonic acid was extricated with great rapidity ; but the volume of gas obtained was always considerably greater than it ought to have been, judging from my previous knowledge of the specific gravity of carbonic acid gas. The reason I found to be, that a portion of the mu- riatic acid had been disengaged in the form of gas, had mixed with the carbonic acid gas, and had augmented its volume. Thus the gas from 100 grains of calcareous spar sometimes amount- ed to 109i cubic inches ; while I knew that the true volume of the gas, if pure, could not exceed 94i cubic inches. I attempted to prevent this extrication of mu- riatic acid gas, by employing very weak muria- tic acid for dissolving the calcareous spar ; but when the experiment was made in this way, the volume of carbonic acid gas extricated fell al- CARBON. ways a little below the truth. The reason I found to be, that a portion of the carbonic acid remained in the liquid muriatic acid, and did not assume the gaseous form at all. When acetic acid was used, the extrication of gas went on very slowly, unless the carbonate of lime was in powder ; and the gas extricated was always below the true quantity, because it did not all assume the elastic state. Sulphuric acid did not answer well, because the sulphate of lime formed was apt to screen some of the calcareous spar from the action of the acid, and thus to prevent the whole gas from being obtained. Oxalic, tartaric, and citric acids, could not be used, for a similar reason. Nitric acid was ex- cluded, because it acted on the mercury. Foiled in my attempts to determine the vo- lume of carbonic acid gas extricated from a given weight of calcareous spar by the above method, I had recourse to another, which succeeded per- fectly. To the beak of a small retort, A, I cemented a long glass tube, bent at right angles, at a, b, and B, and the upright leg, B, C, of such a length, that it reached very nearly to the top of the graduated glass ACIDIFIABLE COMBUSTIBLES. jar D, fitted at its top E with a stop cock. The capacity of the retort was 36 cubic inches. This apparatus was attached to a mercurial trough, so that the jar D could be placed over the upright tube JB, C. The stop-cock F being shut, I opened the stop-cock E, and with my mouth sucked the air out of the jar till only 5 cubic inches remained in it, over a co- lumn of 8 inches of mercury; the stop-cock E was then shut. I next took out the stopper G, the stop-cock F remaining shut, and poured into the retort a cubic inch of strong nitric acid. 100 grains of calcareous spar, in three or four pieces were then dropped into the retort, and the stopper being at the same instant put into its place, the stop-cock F was opened. The whole calcareous spar was speedily dissolved, and the carbonic acid gas made its way into the cylinder D. \Vhen the experiment was over, I noted the volume of gas in the cylinder, the height of the column of mercury below it (ge- nerally about 1 inch), and the height of the ba- rometer and thermometer at the time. Knowing exactly the volume of air in the apparatus before the experiment began, it was easy, by subtract- ing it from the total quantity of air in the appa- ratus after the experiment, to deduce the volume of carbonic acid gas evolved. All the necessary reductions being made, the volume of carbonic acid gas evolved from 100 grains of calcareous Gl CARBON. 143 spar, supposing the barometer to stand at 30 inches, the thermometer at 60, and the gas to be perfectly dry, amounts, from a mean of two experiments, both made with very great care, to 94*426 cubic inches. Now, we have seen that the weight of this vo- lume of carbonic acid gas is 44 grains. Conse- quently, 100 cubic inches of carbonic acid gas must weigh 46-5973 grains; but 100 cubic inches of air weigh 30%5 grains. We have, therefore, 30*5 : 46-5973 : : 1 : 1 -52778 = speci- fic gravity of carbonic acid gas. Thus the specific gravity of carbonic acid gas, determined by ascertaining the volume of a given weight, amounts to 1 -52778' By means of a nonius and a microscopic sight, I was enabled to measure the volume of gas with precision to the thousandth part of a cubic inch. My result, therefore, must be very near the truth. It is not, however, quite exact. But, by means of a law, which I shall explain more at length in the following chapter, we have it in our power to determine the specific gravity of this gas with mathematical accuracy ; for, its atomic weight, multiplied by 0-5555, (half the specific gravity of oxygen gas,) is equal to its specific gravity. Now, we have seen that the atomic weight of this acid is 2-75 ; and 275 x 0-5555 = 1-5277, which is the true specific gravity of this gas. My estimate, deduced in the way just explained, 144 ACIDIFIABLE COMBUSTIBLES. does not exceed this quantity by so much as Toofeooth part. composi- The S p ec ifi c gravity of carbonic acid gas being tion of car- J bonicacid. thus ascertained with rigid accuracy, we have only to subtraction! it the specific gravity of oxygen gas, to know the weight of carbon united in it, with a volume of oxygen gas. Sp. gr. of carbonic acid gas 1-5277 oxygen gas 1*1111 0-4166 = carbon Thus we learn, that this acid is composed of Oxygen . I'll 11 or 2 Carbon . 0-4166 or 0-75 Atomic 2 denotes 2 atoms of oxygen ; and O75 must weight of , . Ji=) carbon. be the atomic weight of carbon. Thus it ap- pears, that carbonic acid is a compound of 2 atoms of oxygen, and 1 atom of carbon.* The carbon in this gas is evidently in the state of vapour. If we consider the gas as a com- * If the specific gravities of oxygen gas and carbonic acid gas were 1*1026 and 1-5240, as stated by Berzelius and Dulong, an atom of car- bon would be 0-76438, a number considerably higher than the one pitched upon by Berzelius; for, 1-5240 1-1026 = 0-4214 = the quantity of carbon in a volume of carbonic acid ; so that carbonic acid would be a compound of Oxygen 1-1026 or 2 Carbon . 0-4214 or 0-76438 It is obvious from this, that these specific gravities cannot be accurate. CARBON. 145 pound of 1 volume oxygen gas and 1 volume carbon vapour, condensed into one volume, it is obvious that the specific gravity of carbon va- pour is O4166. As is the case with hydrogen, azotic, and chlorine gases, a volume of it is equi- valent to an atom. II. Carbon enters into another gaseous com- C . , , , . . tion of car- bmation with oxygen, namely, the gas distm- bonk- oxide guished by the name of carbonic oxide. The as atomic weight of this gas is not of so much im- portance as that of carbonic acid, because we do not know any compounds of which it consti- tutes a part; but it will be worth while to de- termine it, because the knowledge of it will serve still farther to elucidate and confirm the preced- ing conclusions respecting the atomic weight of carbon. Carbonic oxide was discovered by Dr. Priestley, but its nature was first ascertained by Mr. Cruik- shanks. M. M. Clement and Desormes made a set of experiments on it nearly at the same time, or very soon after Cruikshanks, and came to the same conclusion with him : namely, that it is a compound of oxygen and carbon without any hydrogen. I have never succeeded in obtaining it abso- lutely free from all admixture of carburetted hy- drogen gas. Its specific grayity according to my trials is O9700. But, owing to the carburetted hydrogen which it contained, this result is too VOL. I. K 146 ACIDIFIABLE COMBUSTIBLES. low. Berzelius and Dulong state its specific gravity to be 0*9727. This is rather too high ; but it is a good deal nearer the truth than mine. When 1 00 volumes of this gas are mixed with 50 volumes of oxygen gas, and an electric spark passed through the mixture, combustion takes place, both gases disappear, and a quantity of carbonic acid gas is formed just equal to the ori- ginal volume of the carbonic oxide gas employed. This fact, which has been sufficiently established by experiment, will enable us to determine the true specific gravity of this gas, and likewise its constituents and its atomic weight. 1. Since half a volume of oxygen gas added to one volume of carbonic oxide gas constitutes one volume of carbonic acid gas, it is obvious that, if from the specific gravity of carbonic acid gas, we subtract half the specific gravity of oxy- gen gas, the remainder will be the specific gra- vity of carbonic oxide gas. Sp. gr. of carbonic acid gas = 1-5277 sp. gr. of oxygen gas == 0-5555 0-9722 == sp. gr. of carbonic oxide gas. Thus we learn, that the specific gravity of this gas is 0-9722*, or exactly the same with that of azotic gas. The specific gravity of it as given by Berzelius and Dulong, exceeds the truth by little more than vdtn part. CARBON. 147 2. It was observed in a former part of this chapter, that carbonic acid gas contains just its own volume of oxygen gas. But to convert a volume of carbonic oxide gas into a volume of carbonic acid gas, we must add half a volume of oxygen gas. It is evident from this, that a volume of carbonic oxide gas contains just the volume of carbon that carbonic acid gas does, but only half the volume of oxygen gas. It must therefore consist of one volume carbon va- pour, and half a volume oxygen gas condensed into one volume. 1 volume carbon vapour weighs 0*4166' volume oxygen gas. . . Q-5555 0-9722 specific gravity of carbonic oxide gas. 3. A volume of carbon vapour being equiva- lent to an atom, and half a volume of oxygen gas to an atom, it is obvious that carbonic oxide is a compound of one atom carbon, and one atom oxygen. Hence its atomic weight is 1-75. Thus we have two compounds of carbon and oxygen. CARBON. OXYGEN. Carbonic oxide composed of 1 atom -f- 1 atom Carbonic acid ... 1 +2 The last of these only possesses acid properties. The first has no tendency whatever to unite with alkaline bases. K 2 ACIDIFIABLE COMBUSTIBLES. III. There are several gaseous combinations of carbon and hydrogen, which were long con- founded under the general appellation of heavy inflammable airs. Their nature and constitution deserve to be pointed out, because they consti- tute the gaseous substances employed for giving light during the night; and they differ materially from each other in their illuminating powers. They may be divided into two species, which have been distinguished by the names of olefiant gas, and carburetted hydrogen gas. Analysis of i. Olefiant gas was first examined by the oletiant . i , i TV/T gas. Dutch chemists ; but we are indebted to Mr. Dalton for an accurate knowledge of its com- position. Its specific gravity, as determined by my experiments, is 0-9709 ; but Berzelius and Dulong found it as high as 0-9804*. It will be seen immediately, that my result is about yisth part too low ; while that of Berzelius and Du- long is about Truth part too high. It was shown long ago by Mr. Dalton, and the observation was confirmed by my experiments, that one volume of olefiant gas requires for com- plete combustion, three volumes of oxygen gas, and the gas remaining after the combustion amounts to two volumes of carbonic acid gas. The knowledge of this fact will enable us to de- termine the specific gravity, the composition, and the atomic weight of this gas with rigid accuracy. CARBON. (1.) Three volumes of oxygen gas being ne- cessary to consume one volume of olefiant gas, and only two volumes of carbonic acid gas being formed, it is obvious that only two volumes of the oxygen gas were employed in forming car- bonic acid ; the other volume of oxygen gas must have been employed in forming water, and it must have combined with a quantity of hydro- gen, which, had it been in the gaseous state, and uncombined, would have amounted to just two volumes. The two volumes of carbonic acid gas contain each a volume of carbon vapour. Thus it appears, that one volume of olefiant gas is composed of two volumes of carbon va- pour, and two volumes of hydrogen gas con- densed^into one volume. Its specific gravity, therefore, will be obtained by adding together twice the specific gravity of carbon vapour, and twice the specific gravity of hydrogen gas. Twice the sp. gr. of carbon vapour = 0-8333 Twice the sp. gr. of hydrogen gas = 0-1388 0-9722 = specific gravity of olefiant gas. Thus it appears that azotic gas, carbonic oxide gas, and olefiant gas have the same specific gravity. (2.) In the case of carbon vapour and hydro- gen gas, a volume is equivalent to an atom ; so that olefiant gas is a compound of two atoms K3 150 ACIDIFIABLE COMBUSTIBLES. carbon and two atoms hydrogen. This is the same thing as one atom carbon and one atom hydrogen. Consequently, this gaseous compound ought to be distinguished by the name of car- buretted hydrogen, had not that name, unfortu- ately, been previously appropriated to a different gas. Five gases j$ u t this combination of carbon and hydrogen, exist com- J posed of atom to atom, in the gaseous state, deserves to equal atoms of carbon fix our particular attention. I believe that no gen. y r fewer than five different gases or vapours exist, composed of these two substances in that pro- portion, but coupled together in different ways. First sp e. The first consists of cies. 1 volume carbon vapour l IT., > condensed into 1 volume. 1 volume hydrogen gas. f Its specific gravity is 6*4861. One volume of it requires for complete combustion, 1| volume of oxygen gas. After the combustion there re- mains one volume of carbonic acid gas. This peculiar gas has not yet been met with by chemists ; but I see no reason to doubt its existence. It probably exists occasionally as a constituent of coal gas. Second The second is olefiant gas, composed of species. 2 volumes carbon vapour 1 , , , > condensed into 1 volume. 2 volumes hydrogen gas ) Its specific gravity is O9722. One volume of it requires for complete combustion three vo- CARBON. 1,51 lumes of oxygen gas. After combustion there remain two volumes of carbonic acid gas. This gas is formed when alcohol and sulphuric acid, in the proportion of one part of the former to four of the latter, are heated. It exists also as a constituent of coal gas. The third consists of T^ni spe- cies. 3 volumes carbon vapour 1 . , > condensed into 1 volume. 3 volumes hydrogen gas ) Its specific gravity is 1*4,583. It requires 4i times its volume of oxygen gas for complete combustion ; and the residual gas consists of three volumes of carbonic acid gas. Mr. Dalton discovered this gas as a constitu- ent in the gas extricated, when whale oil is ex- posed to a red heat. This gas is employed for lighting houses, and even streets, under the name of oil gas. The fourth consists of Fourth s pe . cies. 4 volumes carbon vapour I , v condensed into 1 volume. 4 volumes hydrogen gas J Its specific gravity is 1*9444. It requires six times its volume of oxygen gas for complete com- bustion 5 and a volume of it, thus burnt, leaves four volumes of carbonic acid gas as a residue. The vapour of sulphuric ether constitutes a gaseous body of this kind. But in this vapour there exists, likewise, a volume of the vapour K 4. ACIDIFIABLE COMBUSTIBLES. of water. Hence, the specific gravity of ether vapour is 2*569i ; for it consists of 1 volume of quadro-carburetted hydrogen 1 '9444 1 volume of vapour ef water . . 0-6250 2-5694 = sp. gr. of ether vapour. Fifth s P e. The fifth consists of ClCSt 6 volumes carbon vapour ^ ~ T , , > condensed into 1 volume. 6 volumes hydrogen gas J Its specific gravity is 2*9166. It requires nine times its volume of oxygen gas to consume it completely : and when 1 volume of it is con- sumed in this way, there remain behind 6 vol- umes of carbonic acid gas as a residue. I find, that the vapour of naphtha, from coal, constitutes a gaseous compound of this kind. It is not condensed by passing it through water ; and is, I believe, what gives coal gas its peculiar smell. These gaseous bodies are chiefly used for lighting the streets and illuminating houses. Now, it is obvious, that the quantity of light which they will emit while burning, will be pro- portional to the number of volumes of carbon condensed into 1 volume of the gas. Olefiant gas burns with great brilliancy ; but the vapour of naphtha will give three times as much light, CARBON. 153 while the same volume of it is burning. This vapour, then, if we could get rid of the smell, would be the best substance for giving light ; one obvious method would be, to take care that the whole of it is consumed, and that none es- capes under the form of naphtha vapour. 2. Carburetted hydrogen gas can only be ob- tained pure, by stirring the mud at the bottom gen of stagnant waters, during hot weather. It was first accurately examined by Mr. Dalton. Soon afterwards, I published a set of experiments on it, in the first volume of the Wernerian Memoirs. I took its specific gravity at two different times, with great care. Gas, collected at Restalrig, near Edinburgh, had the specific gravity of 0*5554. The mean of the specific gravity of three specimens, collected in the neighbourhood of Glasgow, was 0*5576. It will be proper to state each of these trials, as the gas was collected at three different times. 1st trial 0-5602 2d trial 0-5558 3d trial 0-5570 It is obvious that the gas was not of equal purity in each of these trials ; it was purest in the se- cond specimen, the specific gravity of which comes nearest that of my former experiments. The mean of the two gives us 0*5556, which is undoubtedly very near the truth. Berzelius and Dulong state the specific gravity of this gas to ACIDIFIABLE COMBUSTIBLES. be 0*5590 : this nearly agrees with the specific gravity of the first specimen of gas collected by me near Glasgow. It has been established by accurate experi- ments, that a volume of this gas requires, for complete combustion, two volumes of oxygen gas, and that the residual gas is one volume of carbonic acid. From this fact, it is easy to de- duce the true specific gravity, the composition, and the atomic weight of this gas. (I.) One volume of the oxygen gas must have gone to the formation of carbonic acid gas, and must have united with 1 volume of carbon va- pour 5 the other volume of oxygen gas must have gone to the formation of water, and must have united with two volumes of hydrogen gas. Thus it appears, that the constituents of car- buretted hydrogen gas, are one volume carbon vapour, and 2 volumes hydrogen gas, condensed into 1 volume. Hence, its specific gravity will be obtained by adding together the specific gra- vity of carbon vapour, and twice the specific gravity of hydrogen gas. Sp. gr. of carbon vapour = 0-4166' 2 sp. gr. of hydrogen gas = 0-1388 0-5555 = sp. gr. of carburetted hydrogen gas. Or its specific gravity is exactly half that of oxy- gen gas. CARBON. 155 In the case of carbon vapour and hydro- gen gas, we may substitute atoms for volumes. It is obvious, therefore, that this gas is a com- pound of 1 atom carbon, and 2 atoms hydrogen. And as an atom carbon weighs O75, and an atom of hydrogen O1&5, the atomic weight of this substance is unity, or the same as oxygen. Its specific gravity is only half that of oxygen gas ; because, a volume of oxygen gas is equiva- lent to two atoms, while a volume of this gas is equivalent only to one atom. 3. There exists a third combination of carbon Naphtha- and hydrogen, which has come to our knowledge quic'arbu- since coal gas was employed for lighting tbedrogeo. 7 ' streets. It was first noticed by Mr. Garden, who described some of its most remarkable pro- perties, in the year 1820.* Soon after, I got from Mr. Cooper a considerable quantity of this substance, which I purified by sublimation, and subjected to a chemical examination. A valua- ble set of experiments on this substance was made by Dr. Kid, and published in the Philosophical Transactions for 1821 ; he gave it the name of Naphthaline. I have only to refer the reader to this interesting paper for a satisfactory explana- tion of the formation and properties of this sin- gular substance. But, as Dr. Kid made no ex- periments to determine its constituents, it may * Annals of Philosophy, XV. 1.7. 156 ACIDIFIABLE COMBUSTIBLES. be worth while to state here the results which I obtained. One grain and a quarter of naphtha- line, in crystalline plates, was mixed with a suf- ficient quantity of oxide of copper, and exposed to heat in a copper tube to which was attached a glass tube, filled with fragments of fused muriate of lime. The gas extricated was col- lected over mercury ; it was pure carbonic acid ; and measured under the mean temperature and pressure 8 '85 cubic inches. The increase of weight of the muriate of lime was very nearly 1*125 grain. Now, 8*85 cubic inches of carbonic acid gas weigh very nearly 4*125 grains, and contain 1*125 of carbon ; and 1*125 water contain 0*125 hydrogen. Now, 1*125 carbon = 1J atom ; while 0*125 hydrogen is an atom of hydrogen. Consequently, naphthaline is composed of l atom carbon 1-125 1 atom hydrogen 0-125 1-25 It is, therefore, a sesquicarburet of hydrogen. Dr. Kid states, that the specific gravity of naphthaline is a little heavier than that of water. I found its specific gravity only 0*7812 at 60. But I am disposed to ascribe this apparent light- ness to the pores which it contains ; for when digested in nitric acid, its specific gravity became 1 *23. But this digestion alters its properties, and BORON. 157 converts it into a kind of acid, which has a yel- low colour, and very different properties from the pure naphthaline. SECT. II. OF THE ATOMIC WEIGHT OF BORON THE experiments of Davy, and of Gay-Lussac Atom of rri i i -i boracic acid and Ihenard, to determine the composition ot stm doubt- boracic acid by the augmentation of weight u which takes place when boron is converted into boracic acid, differ so much from each other, that it is impossible to put any confidence in either. Davy found that 30 grains of potassium were requisite to reduce a quantity of boracic acid capable of furnishing 2*37<5 grains of boron. Now, 30 grains of potassium when converted in- to potash, (as they were by the process) absorb 6 grains of oxygen. Hence, we have boracic acid composed of Boron 2 -375 or 1 Oxygen 6 - 2-52 He found farther, that 1 grain of boron, when converted into boracic acid, absorbs 5-125 cubic inches of oxygen gas, or 1*7368 grain. This gives us boracic acid composed of 158 ACIDIFIABLE COMBUSTIBLES. Boron ] Oxygen 1-7368 The mean of these two experiments gives us the composition of boracic acid as follows : Boron 1 Oxygen 2-128 Now, I am persuaded that this comes nearer the truth than any other analytical results which have bee,n given to the public. The statement of Gay-Lussac and Thenard, that boron, when converted into boracic acid, only absorbs half its weight of oxygen, is most certainly extremely inaccurate. Davy's result, I believe to be about th part below the truth -, but the result of the French chemists cannot be much less than fths below the truth. In the second volume of the second series of the Annals of Philosophy, I published a set of experiments on boracic acid, from which I de- duced 2*7,5 as the probable atomic weight of that acid. But my experiments, which consisted in the analysis of borax, were not quite conclusive, owing to the difficulty of collecting and weighing the whole of the boracic acid. I had not at that time examined jluoboric acid gas with the re- quisite attention. I have since found, that this curious combination furnishes us with the means of determining the atomic weight of boracic acid with rigid accuracy. BORON. 159 1. It will be proved in the 15th chapter of Specific this treatise, that the weight of an atom of fluo- fluo ric acid is 1 -25. Fluoboric acid gas may be ob- ac tained by mixing together anhydrous boracic acid and pulverized fluor spar, and exposing the mixture to a red heat, in an iron tube or sim- ply, by heating in a retort (as was first discovered by Dr. John Davy), a mixture of 1 part boracic acid, 2 parts fluor spar, and 12 parts sulphuric acid. I collected a quantity of this gas, and took its specific gravity two different times, and obtained each time the very same result - 9 name- ly 2-3709. On causing the gas under examination to be absorbed by water, I observed that there was a slight deposite of silica, indicating a mixture of fluosilicic gas. By weighing this silica, I deter- mined the volume of fluosilicic gas, and on de- ducting it, the specific gravity of the fluoboric acid gas was 2*38833. A careful examination of the water containing the fluoboric acid gas, enabled me to detect in it the presence of a small quantity of sulphuric acid ; it was present to the amount of 0*8 grain in the hundred cubic inches of the gas. This being also deducted, the specific gravity of pure fluoboric acid gas be- came 2-3622.* * Dr. Davy found the specific gravity of this gas 2-3709. This is the very specific gravity which I got and a little above the truth, owing, doubtless, to a little fluosilicic or sulphuric acid, or both. 160 ACIDIFIABLE COMBUSTIBLES. its atomic 2. Dr. Davy found that fluoboric acid gas and ammoniacalgas, combine with each other in the three following proportions : (1) 1 volume fluoboric gas ~fc > forming a solid salt. 1 volume ammomacal gas J (2) 1 volume fluoboric gas ") _ v forming a liquid. 2 volumes ammomacal gas J (3) 1 volume fluoboric gas ~) > forming a liquid. 3 volumes ammomacal gas \ It will be shown in the 7th chapter of this treatise, that, in the greater number of gases, a volume of the gas represents an atom. But in ammoniacal gas this is not the case ; two volumes of this gas are requisite to constitute an atom. It is reason- able to conclude from this, that the second of the preceding compounds is the one consisting of an atom of each constituent. We have it, there- fore, in our power to determine the atomic weight of fluoboric acid from it. 1 volume fluoboric gas weighs 2'3622 or 4-2519- 2 volumes ammoniacal gas - 1-18055 - 2'125 M25 is the weight of an atom of ammonia ; con- sequently, 4*2519 is the weight of an atom of fluo- boric acid. The law of Dr. Prout, which will be found to hold in the atomic weights of all bo- dies, shows us that this number is a very little too high. The true atomic weight is undoubt- edly 45 5 consequently, the true specific gravity BORON. 161 of fluoboric acid gas is 2-361 i, which differs less than Wooth part from the specific gravity which I actually obtained. 3. 4'&5, then, is the atomic weight of fluoboric ^ t s io c n ompo ~ acid, and 1-25 is the atomic weight of fluoric acid ; consequently fluoboric acid is a compound of Fluoric acid 1-25 Boracic acid 3-00 4-25 3, then, is the weight of an atom of boracic acid ; for Davy's experiments on the composition of boracic acid Berzelius' analysis of borate of ammonia and mine of borax preclude the pos- sibility of either more or less than an atom of boracic acid being united in fluoboric acid, with an atom of fluoric acid. From Davy's analytical experiments, it is Atomof * * boron and clear, that the atom of boron is 1 j and that bo- racic acid is a compound of 1 atom boron 1 2 atoms oxygen 2 4. It occurred to me, that, by heating together Ratio be- anhydrous boracic acid and dry carbonate of weights of potash, some light might be thrown on the c^nic^ atomic weight of boracic acid, by the quantity acids * VOL. I. L ACIDIFIABLE COMBUSTIBLES. of carbonic acid driven off. A quantity of car- bonate of potash was fused in a platinum cruci- ble, which varied in different experiments, from 80 to 100 grains ; and 9 grains of anhydrous bo- racic acid being added, the fusion was usually continued from one hour to an hour and a halfi or in one case, two hours. The loss of weight in the different experiments was not uniform. The smallest loss amounted to 6*8 grains, and the greatest to 8 grains. In these experiments it is obvious, that the greatest loss of weight must be nearest the truth. Even 8 grains, the maximum toss of weight, is a little below the real quantity of carbonic acid which ought to have been displaced. The reason of the variation in these results, and the difficulty of driving off the full quantity of carbonic acid, I take to be, that the boracic acid has a tendency to fall to the bottom of the fused mass, and to remain at the bottom. It is, therefore, difficult to bring the acid and the carbonate into contact, and to keep them in contact till the whole acid enters into combination with the potash and displaces the carbonic acid. I found that I succeeded best when I placed the acid on the surface of the al- kali. But let us see what the atomic weight of bo- racic acid will be, if we suppose 9 grains of it to saturate as much potash as 8 grains of carbonic acid. 8:9: : %'75 : 3-09375 = atomic weight BORON. 163 of boracic acid. This number comes near 3, which we have seen to be the true atomic weight of boracic acid; and serves still farther to show that the experiments from which I de- duced 2*75 to be its atomic weight, were in- accurate. 5. Boracic acid, as it is usually obtained from borax by precipitation, is in white opaque scales, containing a considerable quantity of water. Sir H. Davy, Berzelius, and myself, endeavoured to determine the quantity of water present in these scales. The experiment is attended with a good deal of difficulty. I varied my method after I became aware that my first experiments were wrong, and endeavoured to prevent any volatilization of boracic acid, by mixing the hydrous acid with a considerable quantity of protoxide of lead. But I still found the loss of weight to amount to 44 per cent. The follow- ing are the results of Davy, Berzelius, and my- self. The hydrate is composed of Real acid 57 . . 56 . .56 Water 43 . . 44 . .44 I am satisfied that Davy came nearest the truth, though I have not been able to discover the cause of my own failure. If the hydrous acid contain two atoms of water united to one atom of acid, its constituents should be L2 164 ACID1FIABLE COMBUSTIBLES. 1 atom boracic acid 3 or 57' 143 2 atoms water 2-25 42-85? 5-25 100 which almost coincides with Davy's experiment. 6. Very few of the borates have hitherto been Se of ami analyzed, and very few of them accurately. However, from Berzelius' analysis of borate of ammonia, I think we are entitled to conclude that the salt is composed of 1 atom boracic acid 3 1 atom ammonia . 2-125 2 atoms water . . 2-25 7-375 Berzelius obtained Boracic acid 2-659 Ammonia 2-125 Water 2-25 7-034 Analysis of j t It was not till I had satisfied myself of the true atomic weight of boracic acid from the specific gravity and combinations of fluoboric acid, that I became acquainted with M. Arfwed- son's method of analyzing the borates by means of fluor spar. He mixes the borate to be analyzed with a determinate quantity of fluor spar reduced to the state of a very fine powder, and exposes the mixture with the requisite BORON. 16.5 quantity of sulphuric acid to heat in a platinum crucible. The boracic acid is disengaged in the state of fluoboric acid, and the base of the borate remains combined with sulphuric acid. In the alkaline borates the base, when converted into a sulphate, is soluble in water, and may therefore be obtained in a separate state and its weight determined. I had recourse to this method, in order to determine the weight of soda in borax, my previous analysis of this salt having been obviously incorrect. (1.) 19 grains of borax being slowly raised to a red heat in a platinum crucible till they ceased to lose water, were found to have sustained a loss amounting almost exactly to 9 grains. (2.) 19 grains of borax, and 9%5 grains of very finely pounded fluor spar, intimately mixed together, were made up into a paste with sulphuric acid in a platinum crucible, and after digesting for 24< hours were exposed to a strong red heat. The matter remaining in the crucible was white, and weighed about 26 grains. This matter was digested in water, which dissolved the sulphate of soda, together with a little of the sulphate of lime. This water was mixed with oxalate of ammonia to throw down the lime. It was then filtered, evaporated to dryness, and the residual salt exposed to a red heat. It weighed within rmnyth part of 9 grains ; indicating 4 grains of soda. Thus it appears, that 19 grains L 3 166 ACIDIFIABLE COMBUSTIBLES. of crystallized borax contain 4 grains of soda, and 9 grains of water. The 6 grains wanting to complete the whole weight, must represent the boracic acid in the salt. The reader will observe that the. original weight of the ingre- dients put into the crucible (abstracting the water) was 19 '5, consisting of 10 grains anhydrous borax 9*5 grains fluor spar 19-5 The boracic acid of the borax, and the fluoric acid of the fluor spar, had been driven off to- gether, and these acids had been replaced by sulphuric acid. Now, 9*5 fluor spar are com- posed of 7-0 lime 2-5 fluoric acid 7 of lime require 10 of sulphuric acid for satu- ration. So that the sulphate of lime in the residue amounts to 17 grains. The remaining 9 grains are sulphate of soda, composed of 4 soda 5 sulphuric acid 9 But the original weight of anhydrous borax was 10, and it obviously contained 6 grains of bora- SILICON. 167 cic acid. Thus, all the parts of this experiment confirm eac h other, and show that borax is a compound of 2 atoms boracic acid 6 1 atom soda . 4 8 atoms water . 9 19 8. Arfwedson has analyzed boracite by aofboradte. similar process, and it follows from his analysis, that the mineral is a biborate of magnesia, and consequently a compound of 2 atoms boracic acid 6 1 atom magnesia 2-5 8-5 This salt is perfectly anhydrous. SECT. III. OF THE ATOMIC WEIGHT OF SILICON. THOUGH the experiments of Sir H. Davy have satisfied chemists that silica is a compound of oxygen, and a dark coloured combustible sub- stance, like carbon or boron, to which the name of silicon or silicium has been given ; and though Berzelius has obtained silicon in a separate state } yet as silica is the only known compound L 4 168 ACIDIFIABLE COMBUSTIBLES. into which it enters, the principal object of inquiry in this section will be into the atomic weight of silica. This inquiry is of great im- Proofs that portance ; because upon it the exact knowledge the atom of r silica or the constitution of most stony bodies depends. I am of opinion that the atomic weight of silicon is precisely 2, and I shall proceed to state the evidence upon which that opinion is founded. *' In the y ear 1810j Berzelius published a set of experiments, containing an analysis of silica which he made in the following manner :* He mixed together determinate proportions of iron filings, silica, and charcoal in a crucible, and exposed the mixture, properly covered, to the greatest heat of a blast furnace. The united action of the iron and charcoal decomposed the silica, and the silicon as it was evolved, entered into combination with the iron. The silicious iron thus formed was whiter, and had a smaller specific gravity than common iron. When dis- solved in muriatic acid, it gave out a greater proportion of hydrogen gas than common iron, and this hydrogen gas contained a certain pro- portion of carbon. It left undissolved a quan- tity of silica. He dissolved a given weight of this iron in muriatic acid, collected the hydro- gen gas evolved, and by burning it over lime water determined the quantity of carbonic acid formed, from which he deduced the carbon * Afhandlingar. III. 117. SILICON. 169 which the iron contained. The iron was then peroxydized, precipitated by ammonia, heated to redness, and weighed. The iron and carbon deducted from the original weight of the iron, gave the weight of the silicon which had been united with it. This silicon weighed 127 parts, but the silica obtained when the specimen was treated with muriatic acid, amounted to 335 parts. Now, 335 127 = 208 = the oxygen combined with 127 parts of silicon. But other experiments gave him silica composed of Silicon 5 1-5 and 52-25 Oxygen 48-5 and 47*75 These experiments were repeated by Stromeyer,* who obtained for a result Silicon 46-0069 Oxygen 53-9931 Now, the mean of these two sets of experiments gives us the composition of silica as follows : Silicon 49-9189 or 0-99^6 Oxygen 50-0811 1-0000 100-0000 1-9966 = atomic weight of Gives silica, which approaches exceedingly near 2. ^ silica. Berzelius, in his essay on Chemical Propor- tions, and in the second edition of his Larbok i * Gilbert's Annalen. XXXVIII. 321 170 ACIDIFIABLE COMBUSTIBLES. Kemien, states the composition of silica as follows : Silicon 49-7 or 0-988 Oxygen 50-3 1- 100-0 1-988 = atomic weight of silicon. I do not know whether this new determination, which approaches the ratio of equality much more nearly than the number given by Berzelius in his original paper, be the result of a new set of experiments, or deduced from corrections made upon the old ones. But the atomic weight of silica deduced from it, is only dhrth part less than 2. And I do not believe it to be possible to operate in the way practised by Ber- zelius and Stromeyer, without the risk of a still greater error than yfeth part. I consider, there- fore, this determination of Berzelius as a very strong evidence in favour of 2 being the true atomic weight of silica. The com. XI. Among the minerals which are found in position of natural sin. regular crystals, and which have been subjected to chemical analysis, there are a good many con- taining silica united chemically, and conse- quently in definite proportions with different bases. If these minerals were perfectly pure, an accurate analysis of them would give us suffi- cient data for determining the atomic weight of silica. It will be worth our while to examine a few of these compounds, that we may see how SILICON. 171 nearly they agree in giving the same atomic weight to this substance. 1 . TABLE SPAR. This mineral, which is a bisilicate of lime, or Table spar a compound of two atoms silica and one atom of lime, occurs in the Bannat. It has been ob- served likewise in Ceylon, constituting a part of the cinnamon stone rock. It has a greyish white colour, and a foliated fracture in one direc- tion only, from which it has got its name. We have four analyses of this mineral ; one by Kla- proth, a second by Strom eyer, a third by Rose, and a fourth by Bonsdorff. The specimens analyzed by Klaproth and Stro- meyer were from the Bannat; those by Rose and Bonsdorif were found in Finland. According to Klaproth, the constituents of table spar are Silica 50 Lime 45-9* Water 5 100-9t The constituents according to Stromeyer, are * Klaproth reckons only 45 ; but he obtained 82 of carbonate of lime, which contains 45-9 lime. | Beitrage, III. 289. 172 ACIDIFIABLE COMBUSTIBLES. Silica 51-445 Lime 47'412 Protoxide of iron 0-401 of manganese 0-257 Water 0-076 99-591 * Rose analyzed, in Berzelius' laboratory, a spe- cimen of table spar from the village of Perhe- niemi in Finland, the constituents of which were Silica 51-6 Lime 46-41 Oxide of iron, trace Actinolite 1-11 99'12 t The specimen analysed by BonsdorfT was from Pargas in Finland ; it had been mistaken for tremolite. Its constituents were Silica 52-58 Lime 44-45 Magnesia 0-68 Protoxide of iron 1-13 Volatile matter 0-99 99-83 J Before drawing any conclusion from these ex- periments, it may be proper to observe that in the Bannat, the table spar occurs in a rock con- sisting of garnets, blue limestone, and table spar. * Untersuchungen, I. 356. f Nordenskibld's Bedrage, 1st part, p. 94-. f Annals of Philosophy, (second series) II. 300. SILICON. 173 It is reasonable from this circumstance to expect, that the table spar analyzed by Klaproth and Stromeyer should contain a little too much lime, in consequence of the mixture of a small quan- tity of the blue limestone, with the portion of the mineral subjected to analysis. The iron and manganese, small quantities of which were de- tected by Stromeyer, no doubt originated from the garnets. The loss of five per cent, in Klap- roth' s analysis, which he ascribed to water, might possibly be owing to the presence of some blue limestone in the specimen. The table spar from Perheniemi occurs in a limestone rock. This might lead to an excess of lime in the specimen analyzed by Rose. That from Pargas was mixed with pargasite and other varieties of horneblende, which might occasion a slight excess of silica in the specimen analyzed by Bonsdorff. The iron, manganese, magnesia, &c. which occur in minute quantity in the mineral, may be neglected altogether as foreign matter. We shall attend only to the silica and the lime. Let us suppose an atom of silica to weigh 2, and an atom of lime 3'5 ; and let us suppose table spar to be a bisilicate of lime : its constitu- ents would be Silica 4. or 53% Lime 3-5 46f Now, Klaproth's analysis gives us 6 174 AC1DIFIABLE COMBUSTIBLES. Silica 53J; Lime 48-96 Strom eyer's gives us Silica 53J Lime 49'153 Rose's gives us Silica 53^ Lime 47-969 BonsdorfFs gives us Silica 53 Lime 45-086 In the three first analyses there is an excess of lime ; because the mineral analyzed was found in a limestone rock ; in the last analysis by Bons- dorff, there is an excess of silica, because the table spar was from a horneblende rock. If we take a mean of all these analyses as likely to fur- nish a near approximation to the true composi- tion of table spar, we obtain Silica 53-333 or 3-905 Lime 47-792 3-5 It is obvious that this mean cannot be correct, because three of the specimens examined came from a limestone rock, and therefore were likely to contain an excess of lime, and only one from a horneblende rock, and likely to contain an ex- cess of silica. However, we see that even from this mean, we get 3-908 for the weight of two atoms of silica, or 1-9525 for the weight of one atom. 16 SILICON. 175 If we take the mean of the first three analyses which contain an excess of lime, we get 48*694 for the quantity of lime united with 53^ silica. And taking the mean of this quantity, and of 45-086, the lime united to the silica, according to BonsdorfFs analysis, we get 46 '89 for the lirne. This gives us table spar composed of Silica 53-333* or 3'936 Lime 46-890 - 3-5 By this constitution (which is as near the truth as the accuracy of the analysis will admit) two atoms of silica weigh 3*936, and consequently, the weight of one atom is 1*968. 2. SILICATE OF ALUMINA, OR BUCHOLZITE. Some years ago, a mineral was found in the Tyrol, to which Weiss gave the name ofjibrous quartz, but Brandes, who analyzed it, distin- guished it by the appellation of buchohite. It has a white colour, interspersed with black spots ; its texture is fibrous, and it has never been met with in crystals. Its constituents, according to the analysis of Brandes, are as follows : Silica 46 Alumina 50 Oxide of iron 2-5 Potash 1-5 100-0* * Schweigger's Jour. XXV. 125. 176 AC1DIFIABLE COMBUSTIBLES. Let us calculate the composition of this min- eral, on the supposition that the silica is com- bined with three bases, atom to atom ; and that an atom of silica weighs 2, an atom of alumina 2-25, an atom of protoxide of iron 4-5, and an atom of potash 6. 1 . 2'25 (an atom of alumina): 2 (an atom of silica) : : 50 (alumina in the mineral) : 44-44 = silica united to the alumina. 2. 4-5 (atom of protoxide of iron) : 2 : : 2-5 (oxide in the mineral) : I'll silica united to the iron 3. 6 (atom of potash) : 2 : : 1-5 (potash in the min- eral) : 0-50 = silica united to the potash Total silica in the mineral = 46'05 This exceeds the quantity found by Brandes by only 0-05 grain a quantity so small, that it might very easily be overlooked in an analysis. Let us, however, suppose, that the analysis is perfectly accurate, and let us deduce from it the atomic weight of the silica. About zVth of the silica is united to the oxide of iron and the potash. Now, the 28th part of O05, the excess of silica, is nearly 0-002, which being deducted, leaves about 0-053. If we subtract this excess from the 44-444 of silica, supposed above to be united to the alumina, there will remain 44-391 grains. Hence, by Brandes' analysis of this mineral, silicate of alumina is composed of SILICON. 177 Silica 44-391 or 1-997595 Alumina 50 - 2-25 Thus the atomic weight of silica, from the ana- lysis of bucholzite, comes out below 2 ; but the deficiency does not amount to Ftfoth part a dif- ference greatly within the limits of unavoidable error in such analyses. 3. NEPHELINE. This mineral, which is found on Monte Som- ma, near Naples, has been long known to min- eralogists. It crystallizes in six sided prisms, often bevelled at their extremities ; it is hard enough to scratch glass, and has a specific gra- vity of 2*2741. From the late analysis of this mineral by Arfwedson, its constituents appear to be Silica . 44-11 Alumina 33-73 Soda . 20-46 Water . 0-62 98-92* Let us suppose that in this mineral every atom of alumina is combined with an atom of silica; and every atom of soda with an atom and a half of silica, (this is a common mode of combination of soda with acids ;) and let an atom of silica weigh 2, an atom of alumina 2'25, and an atom of soda 4. Let us calculate the composi- * Kongl. Vetenskaps Acad. Handl. 1821. p. 150. VOL. I. M 178 ACIDIFIABLE COMBUSTIBLES. tion of nepheline according to these supposi- tions : (1.) 2-25 (atom of alumina) : 2 (atom of silica) : : 3373 (alumina in mineral) : 29*98 = silica unfted to the alumina. (2.) 4 (atom of soda) : 3 (1 \ atom silica) : : 20-4,6 (soda in mineral) : ... 15-345 = silica united to soda. Total silica .... 45-325 The silica in the mineral being only 44'11 grains, is less than the amount of the silica de- duced from the above suppositions by 1^1 5 grain. Let us admit the specimen analyzed by Arfwed- son to have been perfectly pure, and the analy- sis to have been accurate, and let us deduce from it the atomic weight of silica. It is obvious that nepheline must consist of 3 integrant particles of silicate of alumina, united to 1 integrant particle of sesquisilicate of soda. Of consequence, id of the deficiency belongs to the sesquisilicate of soda. Hence, silicate of alumina, according to Arfwedson's analysis, is composed of Gives Silica . 29-17 or 1-94582 1-94582. Alumina . 3373 - 2-25 The atom of silica deduced from this mineral is less than when deduced from bucholzite. The reason is, probably, that the substances with SILICON. 179 which the nepheline is mixed, (hornblende, meio- iiite, &c.) introduce a slight excess of alumina, or soda, by destroying the rigid purity of the mineral. 4. HEDENBERGITE, OR TR1SILICATE OF IRON. This mineral occurs in Marmor's mine, Tuna- berg, Sweden, where it forms layers in white cal- careous spar. It was described and analyzed by M. Hedenberg, in the year 1807.* It has a blackish green colour, a fibrous fracture, and it seems to consist of rhombs, having the same fi- gure as calcareous spar ; phosphoresces both by heat and friction ; specific gravity, 3-154. The constituents assigned by Hedenberg, but correct- ed by Berzelius, are Silica . . . 40-62 Protoxide of iron . 32-53 Water . . . 16-05 Carbonate of lime . 4-p3 Oxide of Manganese . 0-?5 Alumina . . . 0-37 95-25t The carbonate of lime may be left out of view, because it is obviously derived from the rock in which the hedenbergite was found. * Afhandlingar, II. 164. f Nouveau Systems, p. 31. M2 180 AC1DIFIABLE COMBUSTIBLES. Let us calculate the composition of hedenber- gite, on the supposition that it is a trisilicate of iron ; that silica weighs 2, and protoxide of iron 4*5. 4-5 (atom of protoxide of iron) : 6 (3 atoms silica) : : 32-53 : 43-37 = silica united to the iron. The quantity of silica united to the iron (on the supposition that an atom of it weighs 2) exceeds the quantity in the mineral by 2*75 grains ; but, in all probability, the oxide of manganese and the alumina supplied the place of the deficient silica. Now, 0*37 alumina will go as far in satu- rating oxide of iron, as 0*328 of silica ; and 0*75 of protoxide of manganese will go as far as 0*333 of silica. We may, therefore, subtract 0*66 from the deficient silica, thus reducing the deficiency to 2'09 grains. If we suppose the analysis perfectly correct, the composition of trisilicate of iron deduced from it will be Silica . 41-28 or 5-7104 Protoxide of iron 32-53 - 4-5 According to this analysis, 3 atoms of silica weigh 5*7104 : hence the atom of silica is ^f* =19035. 5. D10PTOSE, OR SESQUISILICATE OF COPPER. r phis mineral, which was at first confounded SILICON. 181 with the emerald, is usually crystallized in six- sided prisms, terminated by trihedral pyramids. It has a fine green colour, a foliated texture, and a specific gravity of 3-3. It is a conductor of electricity. From an analysis of this mineral by Lowitz, we learn that its constituents are Silica . . 33 Peroxide of copper . 55 Water . . 12 100 Let us calculate the constituents of this mineral on the supposition that it is a hydrated sesqui- silicate of copper ; that an atom of silica weighs 2, an atom of oxide of copper 5, and an atom of water (1.) 5 (atom of oxide of copper) : 3 (1J atom silica) : : 55 : 33 = the silica united to the oxide. (2.) 88 : 12 (water in mineral) : : 8 : 1-0905 = water in the mineral, very nearly 1 atom. Hence the mineral is composed of j silica 33 \ oxide 55 1 atom water . . . 12-37 1 integ. part, of sesquisilicate of copper 4 S I 88 t oxide 55 J 100-37 This mineral gives us exactly 2 for the atomic Gives 2. weight of silica. There is a slight deficiency in M3 zinc ACIDIFIABLE COMBUSTIBLES. the water, which could be probably accounted for, if we knew the previous treatment which the mineral had received. Q. SILICATE OF ZINC. silicate of This mineral was first accurately described and examined by Mr. Smithson. It is easily dis- tinguished from the other calamines, by the pro- perty which it has of becoming electric when heated, which is peculiar to it. It is also very generally, if not always, in crystals. According to Smithson's experiments, publish- ed in 1803, its constituents are Silica . . 25 Oxide of zinc . 68-3 Water . . 4-4 97-7 Berzelius, who made an analysis of it, expressly with a view to elucidate the doctrine of definite proportions, found it a compound of Silica Oxide of zinc Water Carbonic acid Oxides of lead and tin 99-916 * Nouvenu Systeme Mineral, p. 258. > SILICON. 183 An atom of oxide of zinc, as has been proved in the first chapter of this work, weighs 5-25. Now, if silica weigh 2, and this mineral be a silicate of zinc, we have 2 : 5-25 : : 24-893 : 65-344 = ox- ide of zinc united to the silica. The oxide of zinc in the mineral exceeds this quantity by 1-493 grain, according to Berze- lius' analysis. Let us see how far this excess can be accounted for. 1. The water in the mineral amounts, we see, to 7*460 per cent. Now, an atom of water weighs 1-125, and an integrant particle of an- hydrous silicate of zinc 7'&5 5 and 7*%5 : 1-125 : : 90-237 (the silicate of zinc per cent.) : 14-002 = quantity of water, supposing the silicate to contain 1 atom water. Now, 14*002 being near- ly double the quantity of water actually contain- ed in the mineral, it is obvious that the silicate of zinc can contain only half an atom of water, which, in 100 parts of the mineral, amounts to 7*001. The quantity of water in the mineral is 7-460, which exceeds 7-001 by 0-459. 2. Besides the silica, then, there are two other substances contained in the mineral, both of which have the property of uniting to oxide of zinc. There are 0-459 grain of water, and 0-45 grain carbonic acid. Now, 0-45 grain of carbonic acid requires 0-86 grain of oxide of zinc to form anhydrous carbo- nate ; but carbonate of zinc usually contains li 184 ACIDIFIABLE COMBUSTIBLES. atom of water. This, united to a quantity of anhydrous carbonate, weighing 1*31 grain, a- mounts to 0-276 grain. There still remains 0-183 grain of water to be disposed of. It is probably united* with oxide of zinc in the mine- ral, so as to form the common hydrate of zinc, composed of 1| atom water, united to 1 atom of oxide of zinc. Now, 1-6875 (H atom of water) : 5-25 (an atom of oxide of zinc) : : 0-183 (water still unaccounted for) : 0-569 = oxide of zinc united to this water, in the state of hy- drate. 3. Thus it appears, that 1 -429 grain of oxide of zinc may have been in the state of carbonate, or hydrate of zinc, reducing the amount of the surplus to 0*064 grain a quantity greatly with- in the limits of the unavoidable errors of ana- lysis. We are entitled, therefore, to consider silicate of zinc as a compound of Gives Silica . 24-893 or 1-998 Oxide of zinc . 65-408 - 5-25 Here the atomic weight of silica is about TTnjrrth part less than 2. 7. BISILICATE OF MANGANESE. Bisiiicateof Thj s mineral is found at Londbanshytta, in manganese Wermeland, Sweden ; and it is said to occur SILICON. 185 more abundantly at Catherinburgh, in Siberia. It has a rose red colour, and has never been ob- served crystallized ; but it has a foliated fracture, and natural cleavages, indicating a rhomboid for its primitive form. Its specific gravity, as deter- mined by Hisinger, is 3*5384. This mineral was analyzed by Ruprecht, and likewise by Napioni ; but at a time when the chemical knowledge of manganese was not far enough advanced to admit of correct results. It was analyzed by Berzelius in 1806, and the constituents were published by him in the first volume of the Afhandlingar (p. 108). He re- peated the analysis in 1817, with as much care as possible, and obtained the following consti- tuents : Silica . . . 48 Deutoxide of manganese 54-42 Lime . . . 3-12 Magnesia . . . 0-22 With a trace of iron. 105-76 Here the weight of the ingredients exceeds the weight of the original specimen by 5'76. The reason is obviously, that, in the mineral, the man- ganese exists in the state of protoxide ; but it was converted into deutoxide in the course of the analysis. It will be shown hereafter, that protoxide of manganese weighs 4%5, and deutox- 186 ACIDIFIABLE COMBUSTIBLES. ide 5. If we reduce the weight of the manga- nese in that proportion, the constituents of the mineral become Silica. ... 48 Protoxide of manganese 4<8 f 98 Lime . . . 3-10* Magnesia . . . 0-22 100'3 Let us consider this mineral as a bisilicate of manganese ; and let us calculate its composition, on the supposition that an atom of silica is 2, an atom of protoxide of manganese 4*5, an atom of lime 3%5, and an atom of magnesia 2%5. (1.) 4-5 (atom of protoxide manganese) : 4 (2 atoms silica) : : 48-98 : 43-53 = silica united to man- ganese ....... 43-53 (2.) 3-5 (atom lime) : 4 : : 3-1 : 3-54 = silica united to lime 3-54 (3.) 2-5 (atom magnesia) : 4 : : 0-22 : 0-35 = silica united to magnesia .... 0-35 Total silica united to the bases = 47'42 This quantity of silica is 0*58 grain less than what was found in the mineral by Berzelius. The quantity which he found is Tauth part more than it ought to have been, supposing an atom of silica to weigh 2. Let us, therefore, deter- * Correcting Berzelius's estimate. SILICON. 187 mine the atomic weight of silica, on the supposi- tion that the constituents, as determined by Berzelius, are rigidly accurate. The bisilicates of lime and magnesia may be omitted as foreign bodies. They contain about Ath of the whole silica ; consequently, yVth of the surplus silica, or O05, must be subtracted. The remainder is 0%53 : let us add this to the si- lica united to the protoxide of manganese. This gives us bisilicate of manganese composed of Silica . . . 44-06 or 4-04798 Protoxide of manganese 48-98 - 4-5 Two atoms of silica weigh 4O4/798 ; consequent- Gives ly, the weight of an atom will be 2-02399, which exceeds 2 by rather more than 1 per cent. 8. KNEBELITE, OR SILICATE OF IRON AND MANGANESE. THIS ore has been described and analyzed by Dobereiner. It is grey, with spots of various colours, occurs in amorphous masses, and lias a specific gravity of 3'714. According to Do- bereiner's analysis its constituents are Silica . . . 32-5 Protoxide of iron . 32 Protoxide of manganese 35 99-5 188 ACID1FIABLE COMBUSTIBLES. Let us calculate the composition of this min- eral on the supposition that it is a silicate of iron and manganese ; and that the atomic weight of silica is 2 9 of protoxides of iron and manganese, each 4\5. (1.) 4-5 (protoxide of iron atom) : 2 (atom of silica) : : 32 : 14-22 = silica united to oxide of iron 14-22 (2.) 4-5 (atom of protoxide of manganese) : 2 : : 35 : 15-55 = silica united to the oxide of manganese 15-55 Total silica united with the bases = 29-77 The silica found in the mineral by Dobereiner exceeds this quantity by 2*72 grains. This ex- cess might be ascribed to the presence of a cer- tain quantity of trisilicate of manganese. Pro- toxides of iron and manganese having the same atomic weight, the silicates of these two bases should have the same weight. Hence, it is not unreasonable to suppose, that only 32 grains of the protoxide of manganese were in the state of silicate the other 3 grains may have existed in the state of trisilicate. Now, 4%5 (atom of pro- toxide of manganese) : 6 (3 atoms silica) : : 3 : 4 = silica united to the 3 manganese. The con- stituents of the mineral may be thus stated : Silicate of iron . 46-22 Silicate of manganese 46-22 Trisilicate of manganese 7'00 99'44 SILICON. 18 ( J This almost coincides with the analysis of Do- bereiner. But waving this probable explanation of the excess, let us calculate the atomic weight of si- lica, on the supposition that the mineral contains only simple silicates, and that the analysis is quite correct. The excess of the silica must be divid- ed between the two silicates in the ratio of 14*22 : 15*55 ; or we must add 1-42 of it to the silica united to the manganese, and 1*3 to the silica united to the iron. This gives us silicate of iron composed as follows : Gives Silica . . 15-52 or 2-1827 2-1827. Protoxide of iron 32 - 4-5 Let us now take a mean of the weight of an atom of silica, deduced from the analyses of the 8 minerals which we have just examined. 1 Table spar gives the atom of silica 1-96800 2 Bucholzite . 1-997595 3 Nepheline .... 1 94582 4 Hedenbergite .... 1-90350 5 Silicate of zinc . . . 1-99800 6 Dioptase 2-00000 7 Bisilicate of manganese . . 2-02399 8 Knebelite .... 2-18270 Mean Mean . 2-00245 2-00245. This weight exceeds 2 only by T oVo th part. The atomic weight given by Berzelius, from his own 190 ACIDIFTABLE COMBUSTIBLES. experiments, is below %. But the weight just deduced is from a much greater number of ex- periments, and ought, therefore, to be much nearer the truth. Ratio be- jjj j t occurpe d to me, that, as silica, when tween the atom of si- heated with carbonate of potash, has the proper- lica and . r . A . carbonic ty of disengaging the carbonic acid gas and unit- ing with the potash in its place, I might be able to form some estimate of the atomic weight of silica by mixing determinate quantities of it with pure dry carbonate of potash, exposing the mix- ture to a strong heat in a platinum crucible, and noting the loss of weight. I found, in the first place, that carbonate of potash, obtained by exposing crystals of the bi- carbonate to a red heat, lost no sensible weight though kept for a considerable time in a state of fusion. I found, in the second place, that, to make a successful experiment, it was necessary to em- ploy, at least, eight times as much carbonate of potash as of silica. When less was employed, the loss of weight was always much too small. Twelve grains of pure silica in fine powder, and anhydrous carbonate of potash, were mixed to- gether, and kept in a state of fusion for 13 hours in a platinum crucible. The loss of weight amounted to 16*67 grains. It appears from this experiment, that 16'G7 grains of carbonic acid are displaced by 12 grains of silica : but upon examining the cruci- SILICON. 191 ble, I found that it had become lighter by 0*15 grain. Deducting this diminution, there remain 16%52 grains for the true quantity of carbonic acid expelled. Now 16 '52 (the carbonic acid expelled) : 12 (the silica employed) : : 2*75 (the atomic weight of carbonic acid) : 1 '997^8 = the p ives /-.-.. 1 "99 *58 for weight of an atom of silica. the atom of From this experiment, which was several times S1 repeated, the atom of silica appears to weigh 1*99758, which is very nearly -nfooth part less than 2 ; the atomic weight of silica, deduced from the analysis of eight silicates, was a little above 2. As the deviations in these two differ- ent sets of experiments are on different sides, there is a chance that the mean of the two may come still nearer the truth. Atomic weight of silica from the analysis of eight silicates 2-00245 Atomic weight of silica from heating silica and carbonate of potash .... 1-99758 - Mean of Mean of both 2-00001 both gives 2 O0001. This atomic weight is little more than ^^th part greater than 2. IV. Silica enters into definite combinations with water, forming various hydrates. Some of these I have examined with care ; and they fur- nish satisfactory evidence that the true atomic weight of silica is 2. 1. When silica is fused with potash or car- 192 ACIDIFIABLE COMBUSTIBLES. bonate of soda, and the fused mass digested in of silica. hot water, we may either obtain a solution, or a white insoluble flocky residue may be left, ac- cording to the quantity of water which we em- ploy. This floky residue is a hydrate of silica, which may be separated by the filter, washed by repeated effusions of water, and dried in a tem- perature which must not exceed 85 or 90. In summer, when the weather happened to be very warm, I have repeatedly obtained an accurate atomic hydrate of silica by drying it in the open air ; but in general, unless the temperature be artificially raised, the silica retains a small excess of water. This hydrate, when properly prepar- ed, is composed of Silica . . 4 Water . . 1-125 5-1 25 Now, 1-1&5 is 1 atom of water, and 4 must be just 2 atoms of silica ; so that the hydrate is a compound of 1 atom water, and 2 atoms si- lica ; and an atom of silica weighs exactly 2. I shall relate two experiments which I made to obtain this hydrate, the first an unsuccessful, and the second a successful one, that the read- er may be enabled to judge how great a surplus of water the silica may retain. About the middle of May, 1823, I fused a SILICON. 193 quantity of silica, with thrice its weight of an- hydrous carbonate of soda, and digested the fus- ed mass in a small quantity of water, till the si- lica assumed a flocky appearance. The whole was then thrown upon a filter ; and the silica was washed repeatedly with distilled water, till no traces of soda could be found in the washings. In two days the filter with the silica became dry enough to be handled. I placed the filter on several folds of bloating paper, on a table in the middle of my laboratory, where it was allowed to remain for six weeks, without being disturbed. It may be necessary to mention, that the weather during the whole time was uncommonly cold ; and I have reason to believe, that the tempera- ture of the room scarcely ever exceeded 60, if it amounted to so much. When I returned to Glasgow, on the 24th of June, the thermometer in my laboratory stood at 57. The silica, to the eye and the feel, appeared perfectly dry ; it weighed 43*23 grains. By exposure to a red heat, it lost 10*55 grains, and was reduced to 32-68. Now, 32-68 : 10-55 : : 4 : 1-2913 = wa- ter combined with 4 silica. This exceeds 1-125 by 0-1663, which is rather more than fth of an atom. This may be considered the greatest a- mount of the excess which ever remains. I shall now state the circumstances of the successful ex- periment. A quantity of silica obtained, precisely as in Vor., I. N 194 ACIDIFIABLE COMBUSTIBLES. the preceding experiment, after being sufficiently washed was left on the filter till it was dry enough to be moved. It was then placed (still on the filter) in a linen bag, which was suspended at some height above my sand bath, in a tempera- ture of 85, where it was left for several days, till it ceased to lose weight. By this treatment, the silica became a very white, dry powder, which weighed 38*4 grains. Being exposed to a red heat, the weight was reduced to 29*97 grains ; so that it had been a compound of 29*97 silica, and 8-43 water. Now, 29'97 : 8-43 : : 4 : 1*125 = the water united to 4 of silica. Here we see that the silica and water were united in the true atomic proportions. And this is an ex- periment which I have often made successfully. Thirteenth g. If the silicated potash, or soda, be digested hydrate of . . f silica. m a sufficient quantity or water, we may obtain a complete solution, which is neither precipitat- ed by acids nor alkalies. When this solution is concentrated by evaporation, it assumes the form of a transparent, colourless, tremulous jelly, con- sisting of silica and water, united together in the form of a hydrate. If we throw this jelly upon the filter, and wash it with hot water till the al- kali be completely removed, the silica remains in the state of a matter, so transparent and co- lourless, as scarce to be perceptible at first. If we allow this jelly to dry on the filter till it ceases to feel wet, or to moisten bibulous paper, it still SILICON. 195 retains a certain degree of transparency, and is not quite destitute of its gelatinous form. In this state it is a compound of 1 atom silica . 2 13 atoms water . 14-625 16-625 So that in this hydrate, though apparently quite dry and solid, more than fths of the whole weight is water. 3. If we allow this solid mass of ielly to dry Hydrate of _ . . J in the open air, without breaking it m pieces, it gradually assumes a yellowish colour, somewhat similar to that of gum-arabic ; and even after an interval of several years, it is still semi-transpa- rent. In this state it is very brittle, but hard, and breaks with a conchoidal fracture. It has a certain resemblance to common flint ; or rather it may be stated as an intermediate looking sub- stance between flint and calcedony. This hy- drate of silica, when successfully prepared, is composed of 1 atom silica . 2 1 atom water . 1-125 I have succeeded in preparing this hydrate, only by leaving it for a very long time (in general several years) in a warm and dry room. I shall state the result of an experiment made during the summer of 1823. A quantity of gelatinous N 2 196 ACIDIFIABLE COMBUSTIBLES. silica, prepared in the way just described, was left, unbroken, on the filter, for six weeks. Its weight was 13-63 grains, Being exposed to a red heat, it lost 5*13 grains ; consequently, the anhydrous silica amounted to 8 '5 grains, and the water to 5-13. Now, 8-5 : 5-13 : : 2 : 1-207 = water united to 2 silica. This exceeds the atomic proportion by about ftth of an atom, ether 4 There are several other hydrates of silica, hydrates. besides the three which I have just described. I have formed a hydrate, composed of 3 atoms silica = 6 1 atom water = 1-125 7-125 I have obtained another, composed of 4 atoms silica = 8 3 atoms water = 3-375 11-375 This great variety of hydrates renders it always indispensable in the chemical analysis of minerals to expose the silica to a red heat, in order to be sure of its true weight. Conciu- y The details in this section have been some- what tedious ; but they were necessary to esta- blish the true atomic weight of silica. Let us recapitulate the different results obtained. sions. SILICON. 197 1. Berzelius has concluded from his analysis of silicated iron, that the atomic weight of silica is 1-988. But the mean of the experiments of Berzelius and Stromeyer gives us for its atomic weight 1-9966. 2. The mean of the analysis of eight native silicates of different kinds gives us the atomic weight of silica 2-00245. 3. My decomposition of carbonate of potash by silica gives for its atomic weight 1 -997^8. 4<. The analysis of the hydrates of silica gives its atomic weight exactly 2. Even if we were to strike a mean from all these different estimations, the weight of an atom of silica would not deviate so much as roVoth part from 2. But this method of pro- ceeding would not be quite proper. The ex- periments on the hydrates being most easily made, and most susceptible of precise results, should have the greater effect in fixing our no- tions on the subject. Now, the hydrates make the atomic weight of silica precisely 2. Next to the hydrates, the greatest confidence seems due to the mean deduced from the analysis of the eight silicates ; now, this mean differs from 2 by only yornjth part. Upon the whole I am persuaded, that no unprejudiced person who peruses the preceding details can have any doubt that the true atomic weight of silica is 2. . Silica is a compound of silicon and oxygen. N 3 198 ACIDIFIABLE COMBUSTIBLES. Now, from the experiments of Berzelius, men- tioned in the beginning of this section, it is clear that the weights of the silicon and oxygen are nearly equal. Since the atomic weight of silica is 2, and of oxygen 1, it is evident that the atomic weight of silicon is also 1. The true composition of silica is therefore 1 atom silicon 1 1 atom oxygen 1 SECT. IV. OF THE ATOMIC WEIGHT OF PHOSPHORUS. PHOSPHORUS enters into chemical combinations both with hydrogen and oxygen ; with the former it constitutes only gaseous, with the latter, only solid compounds. Its combinations with oxygen are phosphorous and phosphoric acids combinations of great importance in the animal, and not without interest in the mineral kingdom. Let us first investigate the atomic weight of phosphoric acid. It will then be easy to determine the composition of phosphorous acid. Atomic 1. Phosphoric acid forms soluble and cryslal- weight of 1 1 i ^ r phosphoric lizable salts with the alkalies. Or these none is better known than phosphate of soda, which being employed as a medicine, is prepared by PHOSPHORUS. 1 99 trading chemists, and kept for sale in apothe- caries' shops. It crystallizes in rhomboids, has a cooling and somewhat bitter taste, and effloresces when kept in a dry atmosphere. I have met with this salt consisting of 11 atoms, and likewise of 12 atoms of water combined with one integrant particle of anhydrous phosphate of soda. By far the most common proportion of water, however, is 12 atoms. This difference in constitution, as far as the aqueous part is concerned, probably depends upon the way in which it has been kept. It very speedily effloresces when exposed to the air. Now, when got from the shops, we often perceive white effloresced spots scattered through the crystals. It is in such cases that the water of crystallization is sometimes reduced to 11 atoms. We can never depend upon the water of crystallization, unless the salt be quite trans- parent. I think it best to make it anhydrous by heating it red hot, and keeping it for some time in a state of fusion in a platinum crucible. 7*5 grains of anhydrous phosphate of soda, and 2O75 grains of crystallized nitrate of lead, were dissolved in separate portions of distilled water, care being taken not to employ more water than was sufficient to dissolve the salts completely. These two solutions being mixed together, a double decomposition took place ; phosphate of lead precipitated in the slate of a N 4 200 ACIDIFJABLE COMBUSTIBLES. white insoluble powder, while a transparent colourless liquid remained, holding nitrate of soda in solution. A drop of this liquid was put into a watch glass, and mixed with a drop of solution of sulphate of soda. No precipitation or opalescence took place, showing that the li- quid contained no sensible quantity of lead. Another drop of the same liquid being put into another watch glass, was tested in the same manner with nitrate of lead, without being in the least affected, showing the absence of any sensible quantity of phosphoric acid. This ex- periment was repeated several times with exact- ly the same result. From the preceding experiment it is obvious, that the oxide of lead in 20*75 grains of nitrate of lead, just neutralizes the phosphoric acid in 7*5 grains of anhydrous phosphate of soda. In like manner, the nitric acid of the former salt just neutralizes the soda of the latter. But 20'75 grains of nitrate of lead are composed of Nitric acid . 6-75 Protoxide of lead 1 4- 20-75 Numbers which are equivalent to the atomic weights of nitric acid and protoxide of lead. Fourteen grains of protoxide of lead require for saturation, a quantity of phosphoric acid, equivalent to an atom ; 6*75 grains of nitric PHOSPHORUS. 201 acid require exactly 4 grains of soda, which will be shown afterwards to be equivalent to an atom. 7%5 grains of anhydrous phosphate of soda, then, contain just 4 grains of soda. The re- maining 3-5 grains must be the phosphoric acid of the salt. Thus, it appears, that anhydrou phosphate of soda is composed of Phosphoric acid 3-5 Soda . . 4-0 7-5 3*5 of phosphoric acid just saturate 4 of soda, shown to and 14 of protoxide of lead, each of which num- bers, represents the atomic weight of the respec- tive bases ; consequently, 3*5 is the atomic weight of phosphoric acid. 2. Many attempts have been made by che- Attempts mists to determine the composition of phosphoric acid by the rapid combustion of phosphorus. For ac when phosphorus burns with flame, it unites with the oxygen of the atmosphere, and is con- verted into phosphoric acid. Now, as both phosphorus and phosphoric acid are solids, it ap- pears at first sight easy to determine the quan- tity of oxygen with which the phosphorus has combined, merely by weighing the phosphoric acid produced. But in reality, this method, though apparently easy, is not susceptible of perfect ac- curacy. Phosphorus can scarcely be obtained 16 ACIDIFIABLE COMBUSTIBLES. in a state of complete purity, and we have no means of appreciating the proportion of impuri- ties with which it is contaminated. It can hardly be burned completely in close vessels, always leaving behind an orange coloured matter, which deliquesces in the air, and is gradually acidified spontaneously. It would be needless to relate the experiments made in succession in this way, by Lavoisier, Rose, myself, Dulong and Berze- lius. Sir Humphrey Davy had the sagacity to obviate some of the difficulties that had proved most perplexing to his predecessors, and his ex- periments came by far the nearest to the truth, though they did not altogether reach it. It may be worth while, therefore, to mention the result of his researches on this difficult subject. He inclosed the phosphorus in a separate tray, which he introduced into the belly of a glass re- tort j by this contrivance, he was enabled to give a stronger and longer continued heat to the phosphorus. He found that 100 parts of phosphorus, when converted into phosphoric acid, unite with 135 parts of oxygen. Let us deduce the atomic weight of phosphoric acid from this composition. It is obvious, that this acid is a compound of one atom of phosphorus with one or more atoms of oxygen. Were the quantity of oxygen only one atom, the atomic weight of phosphoric acid would be only 174. For 135 (the oxygen in PHOSPHORUS. 203 the acid) : 100 (the phosphorus in the acid) : : 1 (an atom of oxygen) : 0-74074. = atomic weight of phosphorus. But we have just seen that its true atomic weight is 3-5. Let us suppose the quantity of oxygen to be two atoms. This would make the atomic weight of phosphoric acid 3-481. For 135 : 100 : : 2 : 1-4815 = atomic weight of phosphorus. It appears from this, that phosphoric acid is a compound of two atoms oxygen, and one atom phosphorus; and that the atomic weight of phosphorus, according to the result of Davy's experiments, is 1-4815. But we have seen that phosphoric acid weighs 3-5 ; therefore, the atom of phosphorus must in fact weigh 1-5, or nearly 2 per cent, more than results from Davy's syn- thetic experiment. Hence, 100 phosphorus must combine with only 133i of oxygen, to form phosphoric acid ; for 133^ : 100 : : 2 : 1-5. That this is the true composition of the acid will appear from the following experiments. 3. Phosphorus and hydrogen combine in two proportions, and form two gaseous compounds called phosphuretted hydrogen, and bihydroguret of phosphorus ; the former of which was first distinguished as a peculiar substance by M. Gen- gembre, the latter by Sir Humphrey Davy. (1.) Phosphuretted hydrogen gas may be ob- C tained pure, by putting fresh phosphuret of lime phosph into a small retort, filling the retort with water, g e en hydr< 204 AC1DIF1ABLE COMBUSTILBES. and plunging its beak into a trough containing re- cently boiled water, or water in which some deut- oxide of azote has been agitated ; the gas is extri- cated with considerable rapidity ; it is colourless, has a smell somewhat similar to that of garlic, and takes fire spontaneously when it comes in con- tact with common air. Water absorbs about 5 per cent, of its volume of it. The specific gra- vity of this gas from a mean of some very care- ful experiments made in my laboratory, is 0-902735. But I have elsewhere shown that its true specific gravity is 0'90277 5* which exceeds the experimental result by an almost inapprecia- ble quantity. When phosphorus is heated in hydrogen gas, a portion of phosphuretted hydrogen gas is formed ; but the_ volume of the original gas is not increased. It is obvious from this, that a volume of phosphuretted hydrogen gas contains just a volume of hydrogen gas. Hence, if from the specific gravity of phosphuretted hydrogen gas we subtract the specific gravity of hydrogen gas, the remainder will represent the weight of phosphorus united to a volume of hydrogen gas. Sp. gr. of phosphuretted hydrogen gas 0*90277 hydrogen gas 0-06944. Phosphorus = 0-83333 * Annals of Philosophy, XVI, 262. PHOSPHORUS. 205 Thus it appears, that phosphuretted hydrogen gas is a compound of Hydrogen 0-0694 or 0-125 Phosphorus 0-8333 - 1-5 But 0*125 represents the weight of an atom of hydrogen, and 1%5 must be the weight of an atom of phosphorus. The gas is obviously a compound of one atom hydrogen, and one atom phosphorus. If we mix together one volume of phosphu- retted hydrogen gas with either one volume or 1-5 volume of oxygen gas, combustion takes place, and the whole gaseous mixture disappears, being in the first case condensed into water and phosphorous acid, and in the second, into water and phosphoric acid. Let us analyze these com- bustions ; they will give us the composition of both phosphorous and phosphoric acids. A volume of phosphuretted hydrogen gas may be considered as composed of one volume hydrogen gas, and one volume phosphorus va- pour condensed into one volume and a volume both of hydrogen gas and of phosphorus va- pour is equivalent to an atom. The volume of hydrogen, in order to be converted into water, requires half a volume of oxygen gas. It is ob- vious from this, that phosphorous acid is a com- pound of one volume phosphorus vapour and half a volume of oxygen gas ; and phosphoric 206 ACIDIFIABLE COMBUSTIBLES. acid of one volume phosphorus vapour and one volume oxygen gas. But one volume of phos- phorus vapour, and half a volume of oxygen gas being each equivalent to an atom, it is obvious, that the constitution of these two acids is the following : Atom of 1. PHOSPHOROUS ACID, phospho- rous acid. 1 atom phosphorus =1-5 1 atom oxygen . = 1 2-5 = atomic weight of acid. 2. PHOSPHORIC ACID. 1 atom phosphorus = 1'5 2 atoms oxygen . = 2 3-5 = atomic weight of acid. Thus an atom of phosphorous acid weighs 2%5, and an atom of phosphoric acid 3*5, which is pre- cisely the weight deduced in the beginning of this section, from the analysis of phosphate of soda. Subphos. When this gas is allowed to stand for some phuretted hydrogen days over mercury or water, it undergoes no sensible alteration in its volume, but deposites one-fourth of its phosphorus. When thus altered, it does not catch fire when mixed with common air, or oxygen gas ; but combustion may be im- mediately induced, by letting up into it a little deutoxide of azote. When sulphur is sublimed 16 PHOSPHORUS. 207 in it, no alteration takes place in its volume. For complete combustion, 1 volume of this altered gas requires 1*25 volume of oxygen gas. By the combustion, it is converted into water and phosphoric acid. Complete combustion cannot be induced with any smaller volume of oxygen gas. When we pass an electric spark through a mixture of 1 volume of the gas, and O7<5 vo- lume of oxygen gas, a detonation takes place ; but the residual gas occupies very nearly 1 vo- lume. On adding 0*5 volume of oxygen gas, a new combustion takes place, and the whole gase- ous matter disappears. It would seem that, by the first combustion, the whole phosphorus is consumed, leaving the hydrogen very nearly pure. It is obvious, that half a volume of the oxygen gas was employed in converting the hy- drogen of the gas into water. The remaining O75 volume of oxygen converted the phosphorus vapour into phosphoric acid ; but phosphorus requires its own volume of oxygen gas for this purpose. From this it is evident that the altered gas is a compound of 1 volume hydrogen gas . . 0-0694 0-75 volume phosphorus vapour . - 6250 0(5914 Hence it contains just nine times as much phos- phorus as hydrogen, and its specific gravity is 208 ACIDIFIABLE COMBUSTIBLES. 0-6944, or ten times that of hydrogen ; and its atomic weight is 1'25. We may distinguish this altered gas by the name of sub-phosphuretted hy- dro gen. (2.) Bihydrogmret of phosphorus was obtained phosphorus, by Davy, by heating crystallized phosphorous acid. When sulphur is heated in it, the volume of the gas is just doubled, and it is converted in- to sulphuretted hydrogen gas. Now, it will be shown in the next section, that sulphuretted hy- drogen gas contains just its own volume of hy- drogen gas. Hence it is obvious, that bihydro- guret of phosphorus contains twice its volume of hydrogen gas. This gas has a similar smell to phosphuretted hydrogen gas. It does not burn spontaneously when mixed with common air, or hydrogen gas \ but combustion takes place readily, if an electric spark be passed through the mixture. For com- plete combustion, 1 volume of it must be mixed with 1-5, or with 2 volumes of oxygen gas. The gases disappear ; and, with the smaller propor- tion of oxygen gas, we obtain phosphorous acid, with the greater phosphoric acid. From these facts it is obvious that the gas is a compound of 2 volumes hj^drogen gas . condensed into one volume. 1 volume phosphorus vapour ) Its specific gravity, therefore, must be the sum of twice the specific gravity of hydrogen SULPHUR. 209 gas and the specific gravity of phosphorus vapour added together, or 0-9722. For, Twice sp. gr. of hydrogen gas = 0-1388 Sp. gr. of phosphorus vapour = O8333 0-9722 = sp. gr. of the gas. Bihydroguret of phosphorus is the fourth gas which we have met with, possessed of this speci- fic gravity. The other three are azotic gas, car- bonic oxide, and olefiant gas. Thus it appears, that the composition and ato- mic weights of the various compounds of phos- phorus with oxygen and hydrogen are as fol- lows : PHOSPHORUS. OXYGEX. ATOMIC WEIGHT. 1. Phosphorous acid . 1 atom + 1 atom 2-5 2. Phosphoric acid .1 +2 3-5 HYDROGEN. 3. Phosphuretted hydr. 1 + 1 atom 1-625 4. Bihydroguret of phosph. 1 +2 1*75 5. Subphosphuretted hydr. 0-75 + 1 1-25 Bihydroguret of phosphorus has both the specific gravity and atomic weight of azotic gas. SECT. V. OF THE ATOMIC WEIGHT OF SULPHUR. SULPHUR forms three acid compounds with oxygen, distinguished by the names of hyposul- VOL. I. O 210 ACIDIFIABLE COMBUSTIBLES, phurous acid, sulphurous acid, and sulphuric acid. Let us, in the first place, determine the atomic weight of sulphuric acid, which is the most im- portant of these compounds. ^* *' Sulphate of potash is a salt which usu- acid huri a ^ crystallizes in pyramidal dodecahedrons, though its primary form is an oblique four sided prism ; and it contains no water of crystalliza- tion. 11 grains of sulphate of potash, and 13-25 grains of chloride of barium, were dissolved in separate portions of distilled water : the two solu- tions being mixed, a double decomposition took place ; and sulphate of barytes, being insoluble, immediately precipitated to the bottom. The residual liquid, which was transparent and co- lourless, was tested with sulphate of soda, and with muriate of barytes ; but by neither of these reagents was it in the least affected showing that it contained no sensible quantity either of barytes or sulphuric acid. From this experiment it follows, that the sul- phuric acid in 11 grains Of sulphate of potash just saturates the barytes from 13*25 grains of chloride of barium ; but the barytes from that quantity of chloride is just 9*7^ grains, which is equivalent to the atomic weight of barytes. When the sulphate of barytes formed in the pre- ceding experiment was washed, dried, and ex- posed to a red heat, its weight was exactly 14*75 grains; and of this weight, 9'75 grains SULPHUR. 211 were barytes. Hence sulphate of barytes is com- posed of Barytes . . 9-75 Sulphuric acid . 5 14-75 The salt being neutral, and 9'7<5 representing an atom of barytes, 5 must represent the atomic weight of sulphuric acid. The potash in 11 grains of sulphate of potash must weigh 6, which is equivalent to an atom ; for it just saturates 4*625 grains of muriatic acid, which is equivalent to an atom ; and we have seen that the sulphuric acid in it weighs 5. Consequently, sulphate of potash is a compound of Sulphuric acid . 5=1 atom. Potash . 6=1 atom. 11 2. Sulphuric acid is manufactured in this conn- Method of try, by the combustion of sulphur, which con- verts it into sulphurous acid gas. This gas passes into large leaden chambers, the bottom of which is covered with about 6 inches of water ; to- gether with the sulphurous acid gas, a quantity of the vapour of nitrous acid passes into the leaden chamber. These two acids have the pro- perty of uniting together ; and they form minute o 2 ACIDIFIABLE COMBUSTIBLES. crystals, which fall into the water at the bottom of the chamber. The moment they come in contact with that liquid, a decomposition takes place; the sulphurous acid is converted into sulphuric acid, and the nitrous acid into deutox- ide of azote. This last substance makes its es- cape in the gaseous state ; but immediately comes in contact with the common air of the chamber, and is converted into nitrous acid, which, unit- ing with a new dose of sulphurous acid, falls again into the water. These decompositions and new formations continually go on as long as the air of the chamber contains oxygen gas and sulphurous acid. The water, when sufficiently impregnated with sulphuric acid, is drawn off from the bottom of the chamber, and concen- trated as far as possible in leaden boilers. It is then put into glass or platinum retorts, and kept in a heat of at least 600, till as much concentrat- ed as possible.* tioTofihe Acid, thus prepared, is a clear, transparent, acid ot colourless liquid like water, with a certain degree of consistency, so as to adhere to the sides of a * The acidulous water, when taken out of the leaden chambers, has a specific gravity of about 1-22. It is a compound of about 1 atom sulphu- ric acid, and 15 atoms water. Jt is concentrated in the leaden boilers till its specific gravity becomes 1 '75. In this state it is nearly a compound of 1 atom acid, and 2 atoms water ; the boiling point of acid of this strength is 4-35. Beyond this strength it cannot be conveniently carried in leaden vessels, because its corroding action on the lead increases at a great rate as the boiling point rises. 9 commerce. SULPHUR. 213 phial like oil. When quite free from all impurity, its specific gravity (when as strong as possible) is 1*84/7 : if it be heavier than this, it is contami- nated with lead. The liquid, in this state, is a compound of 1 atom acid and 1 atom water. Now 1 atom acid = 5 1 atom water = l-li25 6-125 So that 6*125 parts of it contain 1*125 water and 5 of pure acid. Acid of this strength requires the highest temperature to boil it. Mr. Dalton states that its boiling temperature is 620 but it is extremely difficult to determine such high temperatures with accuracy. I have satisfied myself, however, that the boiling point of this acid is above 600. If to 6-125 parts of this acid we add 1*125 parts of water, so as to form a compound of 1 atom acid and 2 atoms water, we obtain a liquid of the specific gravity of 1'78. Acid of this strength possesses the remarkable property of freezing at a higher temperature than acid of any other strength its freezing point, ac- cording to the experiments of Mr. Keir, being 45. Its boiling point, according to Mr. Dalton, is 435. 3. Protosulphate of iron is a salt composed of O3 214 ACIDIFIABLE COMBUSTIBLES. Formation 1 atom sulphuric acid = 5 l atom protoxide of iron = 4-5 7 atoms water . = 7'875 17-375 So that an integrant particle of it weighs 17*375, and 4 integrant particles weigh 69*5 parts. If we expose 69*5 parts of this salt to a sufficient heat, it becomes white and loses 30*375 parts of water. Now, the whole water contained in 69*5 parts of this salt is obviously 31 '5 parts ; conse- quently, there still remains 1*125 parts of water united to the salt ; and the salt thus heated is a compound of 4 atoms sulphuric acid = 20 4 atoms protoxide of iron =18 1 atom water . . =1-125 39-125 When these 39*125 parts are put into an earthen- ware retort, and exposed to a strong heat, one half of the sulphuric acid, amounting to 10 parts, is decomposed into 8 parts of sulphurous acid, which flies off and is dissipated, and 2 parts of oxygen, which, uniting to the 18 parts of pro- toxide of iron, convert them into 20 parts of peroxide of iron. The remaining 10 parts of the sulphuric acid pass over into the receiver united to the 1-125 of water. This is the process followed aj; Nordhausen in SULPHUR. Germany, by which the fuming sulphuric acid, which answers so well for dissolving indigo, is obtained. Its colour is always dark, and it is opaque. I do not know the reason of this opa- city, but suspect the presence of some selenium in it, from some phenomena which I have ob- served. The specific gravity of this acid is 1'896. I find it, by a careful analysis, to be a compound of 2 atoms acid, and 1 atom water. Now 2 atoms acid =10 1 atom water = 1-125 11-125 So that an integrant particle of it weighs 11 '125, and contains 1/1&5 parts water and 10 parts of pure acid. 4. When this acid is exposed to the air, it smokes, in consequence of its great avidity for moisture. It is much more volatile than the sulphuric acid of this country. When heated in Pure dry . -ii i sulphuric a glass retort to which a receiver is luted, and add. surrounded with a mixture of snow and salt, one half of the acid passes over in a solid state, form- ing a snow white solid acid, like asbest.us in ap- pearance. What remains in the retort is a com- pound of 1 atom acid and 1 atom water or it is precisely similar to the acid of this country. I have not been able to determine the specific gra- vity of this acid : its avidity for water is so great o 4 216 ACID1FIABLE COMBUSTIBLES. that it smokes violently when exposed to the air, and is dissipated so speedily that it cannot be weighed. But, by dropping a piece of it into a known weight of water, and then determining the increase of tlfe weight of the liquid, I was enabled (by throwing down the acid by means of muriate of barytes) to show that, when solid, it consists of pure acid, totally destitute of water. I have obtained it crystallized in flat rhomboids, quite transparent, the large faces deviating only a few degrees from 90. It is converted into vapour at a temperature which must be low, though I have not determin- ed it with accuracy. The specific gravity of this vapour is 2*7777> that of common air being unity. One volume of it is capable of uniting with, and condensing, just 2 volumes of ammo- niacal gas. A volume, therefore, is equivalent to an atom of the acid, for 2 volumes of ammo- niacal gas are just equivalent to 1 atom of am- monia. II. Let us now determine the atomic weight of sulphurous acid. 1. This acid contains less oxygen than sul- tion oi sul- phurous phuric ; for when the alkaline sulphites are dis- solved in water, they are gradually changed into sulphates, while at the same time a quantity of oxygen gas is absorbed. Its atomic weight, then, is less than that of sulphuric acid. 2. Sulphurous acid is a gaseous compound, SULPHUR. 217 which may be easily formed by burning sulphur in oxygen gas. Now, when we convert oxygen gas into sulphurous acid by this process, we do not produce any alteration in its volume, but we produce an increase in its specific gravity. Hence, it is obvious that, if from the specific gravity of sulphurous acid gas we subtract the specific gravity of oxygen gas, the remainder will give us the weight of sulphur united to a volume of oxygen gas. I have shown in a paper on the Specific Gravity of the Gases, inserted in the 16th volume of the Annals of Philosophy, (p. 256) that the true specific gravity of sulphurous acid is '2222* ; and we have seen that oxygen gas weighs 1-1 111. Hence, it is obvious that sul- phurous acid is composed of Sulphur 1-1 111 Oxygen 1-1 11 1 2-2222* Or of equal weights of each constituent. We may consider sulphurous acid gas as a compound of 1 volume oxygen gas 1 , V condensed into 1 volume. 1 volume sulphur vapour ) 3. Sulphuretted hydrogen gas may be obtained c by mixing together sulphuret of antimony and pi muriatic acid in a small retort, and applying the ^ heat of a lamp. The gas may be collected over ACIDIF1ABLE COMBUSTIBLES. mercury. This gas has a fetid odour, and a specific gravity of 1*1805 : it burns with a blue flame ; and 1 volume of it, for complete combus- tion, requires H volume of oxygen gas. Water is formed, and a quantity of sulphurous acid gas, just equal to the volume of sulphuretted hydro- gen gas, consumed. This combustion is sufficient to enable us to determine the composition of sulphuretted hy- drogen gas. It may be proper to mention, in the first place, that when sulphur is heated in hydro- gen gas, the gas is partly converted into sul- phuretted hydrogen ; but no alteration whatever takes place in its volume. From this it is evi- dent, that a volume of sulphuretted hydrogen gas contains just a volume of hydrogen gas. This being understood, let us proceed to ana- lyze the composition of this gas from its combus- tion with oxygen. One volume of it requires li volume of oxygen gas, and forms water and a volume of sulphurous acid gas. The half vol- ume of oxygen gas went to the formation of wa- ter and united with a volume of hydrogen gas : the whole volume of oxygen must have gone to the formation of sulphurous acid, and it must have combined with a volume of sulphur vapour. Thus, we see that sulphuretted hydrogen gas is a compound of 1 volume hydrogen gas ) > condensed into 1 volume. 1 volume sulphur vapour J SULPHUR. And we obtain its composition by stating the specific gravity of each constituent, for it must be a compound of Hydrogen 0-0694 or 0-125 Sulphur 1-1111 - 2-000 1-1805 2-125 No doubt, it consists of an atom of hydrogen united to an atom of sulphur; 0'125 is the weight of an atom of hydrogen, and 2 must, Atom of therefore, represent the atomic weight of sul- su phur. 3. We see from the composition of sulphuret- ted hydrogen gas, that a volume of sulphur va- pour is equivalent to an atom, and a volume of oxygen gas is equivalent to 2 atoms. Now, sul- phurous acid gas is a compound of 1 volume sulphur vapour and 1 volume oxygen gas ; it is, therefore, a compound of 1 atom sulphur and C 2 atoms oxygen. And, since an atom of sulphur Atomic weighs 2, the atomic weight of sulphurous acid is 4. acid - III. It would be easy to show, by a similar process of reasoning, that an atom of hyposul- phurous acid weighs 3. Thus, the atomic weights of sulphur and its acid compounds are as follows WEIGHT OF ATOM. Sulphur 2 ACIDIFIABLE COMBUSTIBLES. Hyposulphurous acid . 3 Sulphurous acid . . 4 Sulphuric acid . . 5 Sulphuretted hydrogen . 2-125 Hyposulphurous acid is a compound of 1 atom + 1 atom Sulphurous .... 1 -|- 2 Sulphuric .... 1 -\- 3 HYDROGEN. Sulphuretted hydrogen . 1 + i The hyposulphuric acid of Gay-Lussac is formed by the union of an integrant particle of sulphu- rous and sulphuric acids with each other ; conse- quently, its atomic weight is 9 ; and it is a com- pound of 2 atoms sulphur with 5 atoms oxy- g&i. SECT. VI. OF THE ATOMIC WEIGHT OF SELENIUM. SELENIUM is so difficult to procure, that I have been obliged to confine my experiments to very minute quantities ; not having had more in my possession than about 40 grains, which did not put it in my power to obtain satisfactory results. For the investigation of its atomic weight, there- fore, I must rely almost entirely upon the exper- iments of Berzelius. Fortunately, his experi- SELENIUM. inents in general are made with so much preci- sion that they furnish very good approximations. Selenium forms with oxygen an acid, which has been distinguished by the name of selenic acid, and with hydrogen a gaseous compound, which has been called selenietted hydrogen gas. 1. Selenic acid combines with most bases in two proportions, and with a considerable number in three proportions, forming with them, salts distinguished by the names of seleniates, biselen- iates, and quadroseleniates. Two seleniates have been analyzed by Berzelius : namely, seleniate of soda, and selenlate of barytes. Let us deduce from each of these analyses, the atomic weight of selenic acid. (1.) Seleniate of soda does not crystallize Analysis of ^ ' f seleniate of readily, though it may be made to deposite crys- soda; talline grains. The biseleniate crystallizes more easily, and assumes the form of needles. Ber- zelius rendered a given weight of each of these salts anhydrous, mixed each with twice its weight of sal-ammoniac, and applied a subliming heat ; the selenic acid was sublimed and decom- posed, and nothing remained but the base of each salt united to muriatic acid, or rather, in the state of chloride of sodium. 100 parts of seleniate of soda furnished 66 parts of common salt, equivalent to 35 55 soda. Hence, this salt is composed of 222 ACIDIFIABLE COMBUSTIBLES. Selenicacid 64-44 or 7'25 Soda 35-55 - 4 100 100 parts of biselejiiate (anhydrous, but deprived also of a little of the acid) when treated in the same way, yielded 41 parts of common salt, equivalent to 22*22 parts of soda.* Hence, the constituents of the salt are Selenic acid 77-77 or 14 Soda 22-22 - 4 100 18 It is obvious that the first of these two salts contains one atom, and the second two atoms of selenic acid, united to an atom of soda. But the atomic weight of selenic acid resulting from the two analyses does not quite agree ; accord- ing to the first analysis it is 7*&5 ; according to the second 7. (2.) Seleniate of barvtes is insoluble in water, ate of ba , . . * . rytes. and can only be obtained in the state or a white powder. But biseleniate of barytes is soluble, and probably contains water of crystallization. Two parts of anhydrous seleniate of barytes dissolved in muriatic acid, and precipitated by sulphuric acid, furnished 1 -76*5 parts of sulphate of barytes equivalent to 1-1657 parts barytes. Hence, the salt is composed of * Ann. de Chim. et de Phys. IX. 258. 16 SELENIUM. 223 Selenic acid 0-834-3 or 6-9782 Barytes 1-165? - 9*75 2-0000 171 parts of anhydrous biseleniate of barytes treated in the same way, yielded 785 parts of sulphate of barytes, equivalent to 518-9 parts of barytes. Hence, biseleniate is composed of Selenic acid 752-1 or 14-132 Barytes 518-9 - 9*75 1271-0* According to the result of the analysis of sele- niate of barytes, the atomic weight of selenic acid is 6*9782 ; according to the analysis of the bise- leniate, it is 7'066. Though these results do not accurately corres- pond with each other, they, upon the whole, concur in rendering it very probable that the true atomic weight of selenic acid is 7 I con* sider the analysis of seleniate of barytes as sus- ceptible of greater precision than that of the analysis of seleniate of soda. Now, the atomic weight of selenic acid, deduced from the mean of the analysis of seleniate and biseleniate of barytes, is 7*02211, which comes extremely near to 7- T . n 1 Attempts 2. 1 may state here a few attempts which I to deter. made to determine the atomic weight of selenic lenic acid. * Ann, de Chim. et de Phys. IX. 262. 224 ACIDIF1ABLE COMBUSTIBLES. acid, though they were not attended with the success which I had anticipated. 21*63 grains of crystallized selenic acid heated till they be- gan to sublime, and therefore considered as an- hydrous, were dissolved in water, and crystallized carbonate of soda in powder was added by small quantities at a time, till the liquid began to give a violet colour to cudbear paper. 43*43 grains of the carbonate were necessary to produce this effect. The liquid was now evaporated to dry- ness, and the dry residuum exposed to a temper- ature of about 600 on the sand bath, till it ceased to give out moisture. It weighed 30-4 grains. As it contained an excess of soda, the salt was again dissolved in water, and a few drops of mu- riatic acid being added, the whole was again evaporated to dry ness. It now weighed 32| grains, and, consequently, had increased in weight 2'266 grains. This increase of weight was occasioned by the union of the muriatic acid with the surplus soda. No effervescence could be perceived, so that I had no evidence that any carbonic acid had been driven off. The salt having been exposed to heat, it is obvious, that the common salt was converted into chloride of sodium. The simplest way of determining its quantity, will be to consider it a compound of 3%5 muriatic acid and 4 soda. Now, 3%5 : 4 : : 2-266 (the augmentation of weight) : 2*59 = soda in the common salt. The common SELENIUM, salt then amounted to 4 '856 grains ; this subtract- ed from 32*666 leaves 27*816 grains of biseleniate of soda. Now, it contains 21 -63 grains of selenic acid. Hence, biseleniate of soda is composed of Selenic acid 21-63 or 14> Soda 6-18 - 4 27-61 This experiment gave therefore 7 for the atomic weight of selenic acid. But I consider this as merely an accidental circumstance, because, on repeating it a second time, the biseleniate was found a compound of Selenic acid 12-353 Soda 4 This would reduce the atomic weight of selenic acid to 6'176. But this mode of experimenting is not susceptible of the requisite degree of pre- cision for deciding so delicate a point. Every time the salt was evaporated to dryness, a little selenium was separated, which prevented the quantity of salt from remaining quite perma- nent. I attempted to determine the atomic weight of selenic acid, by precipitating it from seleniate of soda, by means of muriate of barytes. But the precipitation produced by this salt is only partial a portion of selenic acid remaining in solution even after a considerable excess of muriate of barytes had been added. VOL. I. P ACIDIFIABLE COMBUSTIBLES. Foiled in all these attempts, it would have been requisite in order to determine the atomic weight of this acid, to have had recourse to the direct analysis of some of the seleniates. But, besides that my stock of acid was exhausted, I had no reason to expect more accurate results from such an analysis, than those already ob- tained by Berzelius. Upon the whole, I have little hesitation in fixing upon 7 as the true atomic weight of this acid. ^* ^ e ex P er i men t s m ade by Berzelius on se- lenietted hydrogen gas, will enable us easily to hydrogen. . determine the atomic weight or selenium, from which the composition of selenic acid may be directly deduced. This gas is colourless, has a smell somewhat similar to that of sulphuretted hydrogen gas, is very soluble in water, reddens vegetable blues, and precipitates metals from their solutions. To ascertain its constituents, Berzelius passed a current of it through a solution of acetate of silver ; seleniet of silver was precipitated, which weighed when dry 1*888 parts. This seleniet he dissolved in nitric acid, and threw the solu- tion, while boiling hot, into muriatic acid. By this contrivance, the seleniet was decomposed, and the silver thrown down in the state of chlo- ride of silver. The chloride thus obtained, weighed 1-844. Now, chloride of silver is a SELENIUM. compound of 1 atom chlorine = 4%5, and 1 atom silver = 13*75 ; so that 18*25 grains of it con- tain 13 '75 grains of silver: consequently, 1-844 parts of this chloride must consist of Chlorine 0-4546 Silver 1-3894 1-8440 Now, this 1-3894 of silver constituted one of the constituents of the 1-888 parts of seleniet of silver ; the only other constituent was the se- lenium : consequently, seleniet of silver is com- posed of Silver 1-3894 or 13-75 Selenium 0-4986 - 4-9343 1-8880 13-75 being the atomic weight of silver, 4-9343 must represent the atomic weight of selenium. The silver in the acetate, through which the selenietted hydrogen gas was passed, was in the state of an oxide composed of 1 atom silver = 13-75 1 atom oxygen = 1 14-75 But in the seleniet it was in the metallic state : the reason was, that the selenietted hydrogen and the oxide of silver, had mutually decoin- P2 228 ACIDIFIABLE COMBUSTIBLES. posed each other. The hydrogen of the gas had united to the oxygen of the oxide, and formed wa ter while the selenium and the silver combin- ing together, formed the seleniet of that metal. Seleniet of silver being a compound of 1 atom silver and i atom selenium, and water of 1 atom oxygen and 1 atom hydrogen, it is evident, that selenietted hydrogen must be a compound of 1 atom hydrogen and 1 atom selenium. Its con- stituents, then, from the preceding analysis, must be Selenium 4-9343 Hydrogen 0425 But if 7 be the atomic weight of selenic acid, it is obvious that the atomic weight of selenium must, in fact, be 5 instead of 4'9343 ; for sele- nic acid must be a compound of 1 atom selenium 2 atoms oxygen : And its weight could not be 7> unless the weight of selenium were 5. The true composition of selenietted hydro- gen, then, must be Selenium 5 or 40 Hydrogen 0-125 - 1 5-125 And the atomic weight of this gas must be ARSENIC. 229 It will be shown in the next chapter, that the specific gravity of most of the gases is obtained by multiplying their atomic weights by 0*5555, or half the specific gravity of oxygen gas. Now, 5-125 x 0-5555 = 2-8522, which must be the specific gravity of selenietted hydrogen gas. SECT. VII. OF THE ATOMIC WEIGHT OF ARSENIC. ARSENIC has the property of combining with, at least, two proportions of oxygen, and of forming two well known and very poisonous compounds, distinguished by the names of arsenious and ar- senic acids. It unites, likewise, with hydrogen gas, and forms a gaseous compound, the properties of which have been but imperfectly investigated. An examination of the acids of arsenic will en- able us to determine the true atomic weight of this substance. 1. Arsenic acid is easily obtained, by dissolv- Atomic T . . . . . , 1 i weight of ing white arsenic in nitric acid, mixed with a arsenic small quantity of muriatic acid, and evaporating ac the solution to dryness. It is a white matter, soluble in water ; and the aqueous solution has a very sour taste. With most of the metallic oxides it forms compounds, insoluble in water ; but easily soluble in nitric acid ; with ammonia, P3 230 ACIDIFIABLE COMBUSTIBLES. potash, and soda, it unites in two proportions, and forms soluble compounds with each. Bi- narseniates of potash and soda crystallize readily the former in four sided prisms the latter in rhomboids. Binarseniate of soda contains rather more than 20 per cent, of water, being composed of 2 atoms arsenic acid = 15-5 1 atom soda . = 4-0 5 atoms water . = 5*625 25-125 as I ascertained by a careful analysis.* When this salt is heated to about 500, it loses the whole of its water of crystallization, melting at first into a liquid, and then falling down in the state of a white powder. At a red heat this powder melts into a transparent liquid, which is quite anhydrous. If we add 18 grains of crystallized carbonate of soda to a solution of 25-125 grains of crystals of binarseniate of soda, and agitate the mixture till the whole soda be dissolved, and till the car- bonic acid has made its escape, the liquid will be found incapable of reddening vegetable blues. Now, 18 grains of carbonate of' soda contain just 4 grains of soda, which is equivalent to an atom. * Or rather synthesis. The experiment, which is of importance, because it proves the inaccuracy of Berzelius's opinions respecting the composition of this acid, will be given at large in chapter XVIII. of this work. ARSENIC. 231 It is evident, from this, that the rhomboidal crys- tals contain two atoms of arsenic acid united to 1 atom of soda ; for &5-1&5 grains of them con- tain just 4 grains of soda, which is equivalent to an atom. It has been already explained, that nitrate of lead is an anhydrous salt composed of 1 atom nitric acid = 6-75 1 atom protoxide of lead =14 20-75 Consequently, 41 *5 grains contain 28 grains of protoxide of lead, which is equivalent to 2 atoms. 19*5 grains of anhydrous binarseniate of soda, and 41*5 grains of crystallized nitrate of lead, were dissolved in separate portions of distilled water. The solutions being mixed, a double decomposi- tion took place ; the arsenic acid and protoxide of lead united together, and precipitated in the state of an insoluble arseniate. The clear trans- parent colourless liquid remaining, after this salt had subsided, reddened vegetable blues ; but it was exactly neutralized by the addition of 18 grains of crystallized carbonate of soda. The liquid thus neutralized was tested by sulphate of soda, and likewise by nitrate of lead ; but the solutions of neither of these salts produced any sensible alteration ; showing the absence both of lead and arsenic acid. p 4 232 ACIDIFIABLE COMBUSTIBLES. Now, it is evident from this experiment, that 19*5 grains of anhydrous binarseniate of soda contain exactly 4 grains of soda ; for the 13*5 grains of nitric acid in 41 %5 grains of nitrate of lead require 8 grains of soda to neutralize them. To accomplish this neutralization, (after the oxide of lead had been separated) it was neces- sary to add only 18 grains of carbonate, contain- ing 4 grains of soda ; consequently, 4 grains had been already combined with it, and these 4 grains could only have been derived from the binarseniate of soda. 19*5 grains of anhydrous binarseniate of soda, then, are composed of Arsenic acid 15*5 Soda 4-0 These 1,5*5 grains of arsenic acid neutralized 28 grains of protoxide of lead, which are equivalent to 2 atoms ; consequently, 15*5 grains of arsenic acid must be equivalent* to 2 atoms. Thus, we see that the atomic weight of arsenic acid is 7*75. 2. Let us now investigate the constitution of this tion of ar- . senic acid. acid. Into a small green glass retort, previously weighed and the weight marked upon it with a diamond, 4*75 grains of clean metallic arsenic were introduced, and about an ounce of rather strong nitric acid was poured over it. A moder- ate heat being applied, the arsenic dissolved with ARSENIC. 233 efferverscence. As soon as the solution was com- pleted, a receiver was attached to the retort, the whole nitric acid was not only distilled off, but the heat was continued till the inside of the retort had become quite dry. By this process, the arsenic had been converted into arsenic acid, and its weight was exactly 7*7^ grains. From this experiment it is obvious, that ar- senic acid is a compound of Arsenic 4'75 Oxygen 3 7-75 From this it appears, that the atomic weight of Atomic weight of arsenic is 4 '75, and that arsenic acid is a com- arsenic, pound of 1 atom arsenic and 3 atoms oxygen. 3. Let us now investigate the atomic weight Atomic of arsenious acid. This is the substance known, arsen in common language, by the name of white ar- senic. It reddens vegetable blues, dissolves in water, and unites to bases ; though it does not seem capable of neutralizing them. But the atomic weight, and the composition of this acid, is rendered evident by the following sim- ple and easy experiment. 6*75 grains of pure arsenious acid were put into a small retort, the weight of which had been previously determined ; and there was poured over it, about an ounce of a mixture of 234 ACIDIFIABLE COMBUSTIBLES. 12 parts nitric acid of the specific gravity 1-25, and 1 part of muriatic acid of the specific gra- vity 1'2. The retort being exposed to a moder- ate heat, the arsenious acid was soon dissolved with effervescerfce ; but the heat was continued till all the liquid portion was distilled over into the receiver, and till the interior surface of the retort had become quite dry. The best way of drying the retort is to put it on the sand batib and to cover the whole of it except the extremity of the beak, with sand of a temperature between 500 and 600, and keep it in this situation for several hours. By this process the arsenious acid was converted into arsenic acid, and its weight was found to be exactly 7*75 grains. From this experiment it is evident, that arse- nious acid contains one atom of oxygen less than arsenic acid. Consequently, its atomic weight is 6'J5 ; and it is a compound of one atom ar- senic, and two atoms oxygen. From the preceding investigation it is obvious, that an atom of arsenic weighs 4*75, and that the constitution and atomic weights of the acids of arsenic are as follows : ARSENIC. OXYGEN. ATOMIC WEIGHT. Arsenious acid of 1 atom + 2 atoms 6-75 Arsenic acid - - 1 +3 7-75 So that the constitution of these acids is the same as that of sulphurous and sulphuric acids. 4. It is exceedingly probable that there exists ARSENIC. an oxide of arsenic, composed of one atom arse- nic and one atom oxygen, and, consequently, weighing 5-^5. When very thin crusts of me- tallic arsenic are exposed to the air, they soon lose their metallic lustre, and crumble into a black powder. This black powder is probably the oxide in question ; but it has not been parti- cularly examined by modern chemists. When I was occupied with the experiments related in this section, I made some attempts to investigate its composition, but they were not successful, and I have not again had leisure to resume them. ,5. Arsenic has the property, likewise, of com- Arsenietted i -,i i J J > j? hydrogen. bining with hydrogen gas, and of forming a gase- ous compound, which was first discovered by Scheele, but has never yet been obtained in a state of purity. From our knowledge of the constitution of the other compounds of the com- bustible substances with hydrogen gas, however, there can be no difficulty in stating its leading properties from theory. It is probably a compound of one atom arsenic and one atom hydrogen, and, consequently, its atomic weight must be 4*875. From the analogy of the other gaseous com- pounds of hydrogen and a combustible, there seems little doubt that it is a compound of 1 volume arsenic vapour} 1 volume hydrogen gas Condensed into 1 volume. We will obtain its specific gravity in the gaseous 16 236 ACIDIFIABLE COMBUSTIBLES. state by multiplying its atomic weight by 0-5555. Now, 4-875 x 0-5555 = 2-7083 = specific gra- vity of arsenietted hydrogen gas. The specific gravity of arsenic vapour ought to be 2-6388. One volume of arsenietted hydrogen gas will require for complete combustion, either 1-5 vol- ume, or two volumes of oxygen gas. In the first case, water and arsenious acid, and in the second case, water and arsenic acid will be formed. SECT. VIII. OF THE ATOMIC WEIGHT OF TELLURIUM. THE atomic weights of all the acidifiable com- bustibles have now been investigated, except of tellurium, which is so scarce a substance, that I have never had it in my power to make any ex- periments upon it. But Berzelius has deter- mined the composition of its oxide with his usual accuracy. We may therefore take advan- tage of his experiments to complete the series of the atomic weights of all the simple acidifiable combustibles. According to Berzelius, tellurium unites with only one dose of oxygen, and forms an oxide which has the curious property of forming neu- tral salts, both with oxides and acids so that TELLURIUM. 237 it possesses at once the characters of an acid and an alkali. Davy discovered that it forms a gas with hydrogen, similar in its properties to sul- phuretted, phosphuretted, selenietted, and arse- nietted hydrogen gases. But no experiments have been made to determine the specific gra- vity or composition of this gas. 1. Berzelius made two analytical experiments, from which the atomic weight of tellurium may be deduced. (1.) He found that 100 parts of tellurium be- Composi- tion of ox- came 124*8 parts of oxide ; hence, oxide of tel- ide of teiiu- lurium is a compound of Tellurium 100 or 4-032 Oxygen 24-8 - 1 We see that the oxide of tellurium is a com- pound of one atom tellurium, and one atom ox- ygen. The atomic weight of tellurium by this experiment is 4*032, or rather less than one per cent, more than 4. Now, it is exceedingly dif- ficult to make an experiment similar to the one by which Berzelius determined the composition of oxide of tellurium, without some loss. If the loss amounted to 0*2, which is no more than Trijth part of the whole, then the oxide of tel- lurium is composed as 1 follows : Tellurium 100 or 4 Oxygen 25 - 1 ACIDIFIABLE COMBUSTIBLES. And the atomic weight of tellurium will be ex- actly 4, and that of the oxide 5. (2.) Berzelius found that 201*5 parts of fused tellurate of lead contained just the same weight of protoxide of Iqad, which 157 parts of sulphate of lead do.* But 157 parts of sulphate of lead con- tain 115*684 parts of protoxide of lead ; hence, it follows, that tellurate of lead is composed of Protoxide of lead 115-684 or 14 Oxide of tellurium 85-816 - 10-385 201-500 To reconcile this experiment with the former, the salt analyzed must have been a bitellurate, or a compound of two atoms oxide of tellurium and one atom protoxide of lead. The atomic weight of oxide of tellurium from the above analysis is 5*1925. This is rather higher than the weight deduced from the composition of oxide of tellurium, The former experiment being much simpler, and con- sequently, susceptible of greater accuracy than the latter, there can be no hesitation in pre- ferring it, and, consequently, in fixing upon 4 as the true atomic weight of tellurium, and 5 as the atomic weight of oxide of tellurium. 2. We may deduce the specific gravity and constitution of telluretted hydrogen gas, with a * Annals of Philosophy, HI 250. TELLURIUM. 239 great deal of probability, from the known consti- tution of the other similar compounds. If it be a compound of 1 volume tellurium vapour 1 > condensed into 1 volume, 1 volume hydrogen gas J as is the case with all the others, and if a vo- lume of tellurium vapour be equivalent to an atom, then the atomic weight of telluretted hy- drogen is 4-215, and its specific gravity 2-2916. One volume of this gas will require, for com- plete combustion, one volume of oxygen gas, one half of which will go to the formation of water, and the other, to the formation of oxide of tellurium. CHAP. VII. OF THE RELATION BETWEEN THE ATOMIC WEIGHTS AND SPECIFIC GRAVITIES OF GA- SEOUS BODIES. GAY-LUSSAC, in a paper published in the second volume of the Memoires d'Arcueil, demonstrated, a good many years ago, that when gases com- bine with each other, one volume of the one gas either combines with one volume, or two vo- lumes, or three volumes of the other. Dr. Prout, in a paper inserted in the sixth volume of the Annals of Philosophy, first pointed out the con- nection which existed between the specific gra- vity and the atomic weights of the gases ; and showed that, in general, the specific gravity is equal to the atomic weight multiplied by half the specific gravity of oxygen gas. I propose in this chapter to take a general view of the re- lation between the atomic weights and the spe- cific gravities of all the known gases. 1. In the preceding chapters we have deter- mined the atomic weights of all the simple bodies which are capable of assuming the gaseous form. Of these there are only four, which are gases in the uncombined state : these are Oxygen, Von. I. Q GASEOUS BODIES. Hydrogen, Chlorine, and Azote. There are seven simple bodies which are capable of com- bining with hydrogen, and forming with it a gaseous compound : these are iodine, carbon, phosphorus, sulphur, selenium, arsenic, and tel- lurium. The two remaining bodies, boron and silicon, enter into gaseous combinations with fluo- ric acid,- a substance, the consideration of which is deferred till a future chapter ; though the specific gravity and atomic weights of the gases will be inserted in the following tables, for the sake of completing the series of gaseous bodies. Gases 2. We have seen that water is a compound of dfic gravity one volume of oxygen gas, and two volumes of hydrogen gas ; while, from the proportion in which hydrogen combines with all known bodies, there is every reason to consider two volumes of it as equivalent to an atom ; or if we choose, for the sake of simplicity, to represent an atom of hydrogen by a volume of the gas, then we must represent an atom of oxygen by half a volume of the gas. We have seen that chlorine and azote have the same relation between their atomic weights and their volumes that hydrogen has; this is the case also with iodine, carbon, phos- phorus, and sulphur, when in the state of va- pour. The relation between the atomic weight and the volume of the vapours of selenium, ar- senic, and tellurium, has not yet been determin- ed by experiment ; but there seems no reason = atom X 555. GASEOUS BODIES. to hesitate about extending the same law to these bodies also. Thus it appears, that the atomic weight of oxygen is obtained by dividing the specific gra- vity of oxygen gas by itself j for ^4 = 1 ; but the atomic weights of the other gases and vapours are obtained by dividing the specific gravity of the gas or vapour by half the specific gravity of oxygen gas, or 0*5555.* The specific gravities of oxygen gas, hydrogen gas, chlorine gas, and azotic gas, have been de- termined with rigid accuracy in the preceding chapters. The knowledge of the atomic weights of the other bodies will enable us to determine the specific gravities of the vapours and gases of which they constitute parts, with equal ac- curacy ; for we obtain these specific gravities by multiplying the atomic weights by O5555. The following table exhibits the atomic weights * I need hardly point out to the reader, that the reason why the multi- plicand of the specific gravity of the gases is ^, or ^th the specific gravity of oxygen gas, instead of simply ^ and ^th, is, that the specific gravities and the atomic weights are referred to different standards. The specific gravity of atmospheric air is reckoned unity, and that of every other gas is referred to it ; but the atomic weight of oxygen being unity, and. that of all the other gases referred to it, as far as atomic weights are concerned, it is ob- vious, that when specific gravities are to be converted into atomic weights, we must employ for our divisor half or |th the specific of oxygen gas, in order to change one system into the other. All this trouble would be saved, if we were to make the specific gravity of oxygen gas unity, as well as its atomic weight. 0,2 244 GASEOUS BODIES. and specific gravities of all the gases and vapours coming under this law. I. Gases whose specific gravity is equal to the atomic weight multiplied by 1*1 111, 'ATOMIC WEIGHT. SPECIFIC GRAVITY. Oxygen gas . . 1 1-1111 Fluosilicic acid . . 3-25 3-611 i II. Gases and vapours whose specific gravity = atomic weight multiplied by 0-5555. 1. SIMPLE GASES AND VAPOURS. Hydrogen gas '. . 0-125 0-0694 Carbon vapour . . 0-?5 0-4166 5 Phosphorus vapour . 1-5 0-8333 Azotic gas ... 175 0-9722 Sulphur vapour . . 2 1-1111 Tellurium vapour . . 4 2-2222 Chlorine gas . . 4-5 2-5 10 Arsenic vapour . . 4-75 2-6388 Selenium vapour . . 5 2-7777 Iodine vapour . . 15-5 8-611 i 2. HYDROGEN, COMBINED WITH A SIMPLE BASE. Carburetted hydrogen . 1 0-5555 Phosphuretted hydrogen 1-625 0-9027 15 Subphosphuretted hydrogen 1-25 0-6944 Bihydroguret of phosphorus 1-75 09722 Sulphuretted hydrogen 2-125 1-1805 Telluretted hydrogen . 4-125 2-2916 Arsenietted hydrogen . 4 875 2-7083 20 Selenietted hydrogen . 5-125 2-8522 GASEOUS BODIES. 3. OXYGEN, COMBINED WITH A SIMPLE BASE. ATOMIC WEIGHT. SPECIFIC GRAVITY. Steam . . . 1-125 0-625 Carbonic oxide gas . l-?5 0-9722 Carbonic acid . . 2-75 1-5277 Protoxide of azote . 2-75 1-5277 25 Nitric acid vapour . 6-75 3-75 Sulphurous acid . . 4 22222 Sulphuric acid vapour . 5 2-7777 4. OTHER COMPOUND GASES. Cyanogen . . . 3-25 1-8055 Fluoboric acid . . 4-25 2-361*1 30 Sulphuret of carbon 4-75 2-6388 Chlorocarbonic acid . 6-25 3-4722 The specific gravity of these 29 gases and va- pours being obtained by multiplying the atomic weights of each by 0*5555, or half the specific gravity of oxygen, it is obvious, that in all of them a volume may be substituted for an atom ; or, in other words, the specific gravities and ato- mic weights of all of them bear the same relation to each other. In oxygen gas and fluosilicic acid, on the contrary, an atom is represented by half a volume ; so that, in the first set of gases, a volume contains twice as many atoms as in the second set. 3. There are five compound gases, whose ato- Gases ^ ^ mic weights and specific gravities bear a different cific gravity . = atom X relation to each other from the gases in the pre- 0-2777. Q3 246 GASEOUS BODIES. ceding table. The compound gases in the pre- ceding table may be considered as consisting of a volume, or a determinate number of volumes of each constituent, so condensed, that a volume of the compound gas bears the same relation to a volume of oxygen gas, as a volume of hydrogen gas does to a volume of oxygen gas ; but there are four compound gases, which consist of a vo- lume of each of the constituents united together, without any condensation whatever. Conse- quently, their specific gravities, instead of being equal to the specific gravities of all their consti- tuents added together, are only the means of the specific gravities of the constituents. In reality, therefore, the volumes of these gases may be considered as doubled, with reference to their atomic weights. These gases are deutoxlde of azote, hydrocyanic acid vapour, muriatic acid, and liydriodic acid. (1.) Deutoxide of azote is composed of 1 vo- lume of oxygen gas, and 1 volume of azotic gas, united together, without any alteration of vo- lume ; hence its specific gravity is the mean of the specific gravities of oxygen and azotic gases, or 1-0416. (2.) Hydrocyanic acid, muriatic acid, and hy- driodic acid, are composed respectively of one vo- lume of hydrogen gas, united to a volume of cyanogen, chlorine, and iodine vapour, without any condensation. GASEOUS BODIES. 247 (3.) The constitution of ammoniacal gas is analagous to that of the four gases just mention- ed. It is composed of 1 volume of azotic gas, and 3 volumes of hydrogen gas, condensed into two volumes ; so that its specific gravity, instead of being equal to the specific gravity of azotic gas, added to thrice the specific gravity of hy- drogen gas, is only half that quantity, or O59027- Therefore, with respect to its atomic weight, its bulk is also doubled. From what has been just stated, it must be obvious, that in these five gases the specific gra- vity will be obtained by multiplying the atomic weight of each by O2777, or ith of the specific gravity of oxygen gas. The following table exhibits the atomic weights and specific gravities of these gases : ATOMIC WEIGHT. SPECIFIC GRAVITY. Ammoniacal gas . 2-125 0-59027 Hydrocyanic acid vapour 3-375 0-9375 Deutoxide of azote . 3-75 1-04166 35 Muriatic acid . . 4-625 1 -28472 Hydriodic acid . . 15-625 4-34027 It is obvious that a volume of oxygen gas con- tains four times as many atoms as a volume of any of these five gases. When we compare these five gases with each other, an atom may be sub- stituted for a volume ; but when we compare them with the gases in the preceding table, two Q4 248 GASEOUS BODIES. volumes represent an atom, while 1 volume of the second set of gases represents an atom. Thus, when we wish to combine 1 atom of sulphurous or carbonic acid with 1 atom of ammonia, we must mix 1 volame of the acid gas with 2 vo- lumes of the ammoniacal gas. Protoxide 4. The constitution of protoxide of chlorine ofchiorme. is Different from that of both the preceding sets of gases. It is composed (if Davy's experiments be accurate) of 4 volumes of chlorine, and 2 vo- lumes of oxygen gas, condensed into 5 volumes. We obtain its specific gravity by adding together four times the specific gravity of chlorine gas, and twice the specific gravity of oxygen gas, and dividing the sum by five ; this gives us 2'4444 for the specific gravity of this gas. As it is a compound of 1 atom chlorine and 1 atom oxy- gen, its atomic weight is obviously 5*5. Now, the specific gravity of this gas is obtained by multiplying its atomic weight by 0*4444, which is fths of the specific gravity of oxygen gas ; for 5-5 x 0-4444 == 2-4444. The reason of this is obvious from the peculiar condensation which the constituents of this gas undergo when they combine. 5. Olefiant gas has not been inserted in the pre- ceding tables ; it has been reserved for separate notice, on account of the peculiarity of its na- ture. Its specific gravity is 0'9722 ; and it is obvious, from the phenomena of its combustion, GASEOUS BODIES. that it consists of 2 volumes of hydrogen gas, and 2 volumes of carbon vapour, condensed into one volume. Now, as in these two bodies atoms may be substituted for volumes, this gas is obviously a compound of 2 atoms hydrogen, and 2 atoms carbon. Hence its atomic weight is 1-75 ; but 1-75 x 0-5555 = O9722 ; so that this gas coincides with the compound gases in the second table ; and it might have been insert- ed in the second division of that table, after car- buretted hydrogen. But the singularity of this substance, which is obviously a compound of 1 atom hydrogen and 1 atom carbon, is, that it seems capable of form- ing a variety of gases, differing from each other in the number of integrant particles of carburet of hydrogen, which a single volume contains. Gases seem to exist, containing 1 integrant par- ticle, 2 integrant particles, 3 integrant particles, 4 integrant particles, and 6 integrant particles. Of course, the specific gravity of each of these gases will be different, though each is regulated by the law which determines the specific gravi- ties of the compound gases in the second table. It may be worth while to exhibit the atomic weights and specific gravities of these gases. ATOMIC WEIGHT. SP. GRAVITY. Simple hydrocarburet . . 0-875 0-4861 Olefiant gas . . . . 1-75 0-9722 Oil gas, or tritohydrocarburet . 2-625 1.4583 16 &50 GASEOUS BODIES. ATOMIC WEIGHT. SP. GRAVITY. Sulphuric ether vapour, containing 4 volumes .... 3-5 . 1-9444 Naphtha vapour, containing 6 vols. 5-25 . 2'9l66 The only one> of these five gases not yet de- tected is the first. But it may be looked for among the inflammable gases produced from ve- getable bodies, or from pit coal. CHAR VIII. OF THE ATOMIC WEIGHTS OF THE FIXED ALKALIES AND ALKALINE EARTHS. THE atomic weights of the simple bodies which constitute the bases of the acids, have been de- termined in the preceding chapters. I propose in the five following chapters to investigate the atomic weights of the bases of the alkaline bodies, including under the term alkali, all bases which are capable of neutralizing acids. The substances usually distinguished by the name tfjixed alkalies and alkaline earths, are 1 Potash 4 Barytes 7 Magnesia 2 Soda 5 Strontian 3 Lithia 6 Lime Sir H. Davy first showed that these bodies are compounds of oxygen, with certain white coloured solid metals, which have so strong an affinity for oxygen that they can with difficulty be preserved in the metallic state. Indeed, if we except potassium and sodium, the remainder, as far as utility is concerned, may be considered as still unknown ; for they cannot be procured in such quantities as to be susceptible of useful applications. The fixed alkalies and alkaline FIXED ALKALIES AND ALKALINE EARTHS. earths, on the contrary, which constitute the oxides of these metallic bases, are in constant use, and constitute some of the most important substances known to the chemist. They are the bases of the most common and useful salts, and an accurate knowledge of their atomic weights is of the utmost consequence to the progress of the science. The principal part of our attention will, therefore, be bestowed on the investigation of the atomic weights of these alkaline com- pounds. The knowledge of the atomic weights of the bases will follow as a kind of corollary from that of their compounds. SECT. I. OF THE ATOMIC WEIGHT OF POTASH AND POTASSIUM. THERE are two salts of potash, which, when properly prepared, are destitute of water, and consist respectively of water united to nitric acid and sulphuric acid. These are nitrate of potash, and sulphate of potash. Analysis of 1. Nitrate of potash crystallizes in six-sided potash. prisms, terminated by six-sided pyramids. When kept red hot, the nitric acid is gradually dissipa- ted, and pure potash remains behind. It may appear at first sight easy, by this means, to deter- POTASH AND POTASSIUM. L 253 mine the composition of saltpetre ; for the loss of weight would indicate the quantity of nitric acid, while the residual matter in the crucible would be the potash ; and if we were possessed of any vessel upon which red hot saltpetre would not act, and if potash were absolutely fixed in the fire, this method would be a good one ; but I have not been able to obtain satisfactory results by decomposing this salt by heat. The process, which answered best, was to triturate a given weight of saltpetre with twice as much siliceous sand, and to fuse the mixture in a platinum crucible. When 12'75 grains of saltpetre were employed, the loss of weight approached very near 6*75 grains, though it always somewhat exceeded that quantity. 2. Sulphate of potash affords an easy method of determining the atomic weight of potash with precision. Before relating the experiment by which this object was attained, it may be pro- per to remind the reader that the atomic weight of sulphuric acid is 5 ; and that 13-&5 grains of chloride of barium yield a quantity of barytes just capable of neutralizing 5 grains of sulphuric acid. 11 grains of sulphate of potash, and 13 '25 grains of chloride of barium, were dissolved in separate portions of distilled water. The two solutions being mixed, a double decomposition took place, and sulphate of barytes precipitated FIXED ALKALIES AND ALKALINE EARTHS. in the state of a white powder. The liquid, after it had recovered its transparency from the com- plete subsidence of this precipitate, was tested by sulphate of soda and muriate of barytes ; but neither of theSe reagents produced the least muddiness or opalescence, showing that the liquid contained neither barytes nor sulphuric acid. From this experiment it is evident, that the sulphuric acid in 11 grains of sulphate of potash, just saturates the barytes from 13*25 grains of chloride of barium. Consequently, it amounts to exactly 5 grains. So that sulphate of potash must be a compound of Sulphuric acid 5 Potash . . 6 11 Atomic g u f- s i nce the salt is neutral, and since 5 is the weight ot potash. weight of an integrant particle of sulphuric acid, it is plain that 6 must be the weight of an inte- grant particle of potash. ,L potas ~ 3. The atomic weight of potash being known, it is easy from the experiments made on potas- sium to determine the atomic weight of that metal. From the experiments of Gay-Lussac and Thenard, it appears that when 100 grains of potassium are put into water, the metal combines with the oxygen of that liquid, and is converted into potash, while the hydrogen gas evolved siura. POTASH AND POTASSIUM. 255 measures 116-38 cubic inches. This experiment was made by plunging a given bulk of potassium into water, and inferring the weight from the known volume and specific gravity of this metal. I have repeated it several times, employing various contrivances to enable me to weigh the potassium ; but I was unable to obtain better results than those of the French chemists. I shall therefore employ them, as they are amply sufficient for the object I have in view. The oxygen gas, which must have united with the 100 grains of potassium, must have amounted to ^ = 58-19 cubic inches ; the weight of which is 19*7^ grains. So that potash, according to these experiments of the French chemists, is composed of Potassium 100 or 5-121 Oxygen 19-72 - 1 6-121 The weight of an atom of potash, according to this experiment, is 6-121, and it is a compound of one atom potassium = 5-121, and one atom oxygen. Now, it is obvious, that this atomic weight of potash is a little too high ; it should have been 6, and not 6*121. The reason un- doubtedly is, that the hydrogen gas evolved was a little below the true quantity. Whoever will take the trouble to repeat the experiment, will be at no loss to account for this deficiency. FIXED ALKALIES AND ALKALINE EARTHS. Berzelius took another method of determining the composition of potash. He placed a quan- tity of amalgam of potassium in water ; the amalgam was speedily decomposed, the potas- sium being converted into potash, and dissolving in the water, while the running mercury re- mained behind. The potash thus formed was saturated with muriatic acid, the liquid evapo- rated to dryness, and the residual salt weighed. The difference between the weight of the amal- gam and of the running mercury remaining, gave the weight of the potassium ; while the weight of the muriate of potash formed, fur- nished the data for determining the weight of the potash. It is obvious, that the difference between the weight of the potassium and the potash, was the quantity of oxygen with which the potassium had combined. This experiment gave the constituents of potash as follows : Potassium 100 or 4-889 Oxygen SO'4-5 - 1 According to this experiment, an atom of po- tassium weighs 4-889, and an integrant particle of potash 5-889. This determination is a little below the truth, while that of the French che- mists is a little above it. The mean of the two gives 5-005 for the atomic weight of potassium. The true weight POTASH AND POTASSIUM. 257 is undoubtedly 5 j and potash is a compound of 1 atom potassium = 5, and 1 atom oxygen = 1 ; consequently, its atomic weight is 6, as we found it to be by the experiment related near the be- ginning of this section. 4. Potassium combines with an additional dose Peroxide of of oxygen, and forms a solid yellow substance, pt which has received the name of peroxide of potassium. From the experiments of Gay-Lus- sac and Thenard, it appears, that 100 grains of potassium, when converted into peroxide, com- bine with 177 cubic inches of oxygen gas. Now, this is just three times as much as the quantity necessary to convert potassium into potash ; for 100 grains of potassium were found to absorb 58*19 cubic inches of oxygen gas when con- verted into potash and -^ = 59- Peroxide of potassium must, therefore, be composed of 1 atom potassium = 5 3 atoms oxygen = 3 8 And an integrant particle of it must weigh 8. It deserves attention, that this experiment of Gay-Lussac and Thenard gives the atomic weight of potassium different from their former one. For 59 cubic inches of oxygen gas weigh 20*38 grains ; so that potash is composed, ac- cording to this experiment, of VOL. I. R 258 FIXED ALKALIES AND ALKALINE EARTHS. Potassium 100 or 4-9068 Oxygen 20-38 - 1 5-9068 The atom of potassium, by the first experiment, weighs 5-121 ; by the second 4-9068. The mean of the two is 5-0139; which approaches much nearer the truth than either of the experiments taken singly. SECT. II. OF THE ATOMIC WEIGHTS OF SODA AND SODIUM. THE atomic weight of soda may be determined with the same precision, and nearly in the same manner, as the atomic weight of potash. 1. Sulphate of soda is a well known trans- ptrteof SU ~ parent salt, which crystallizes with great facility. soda. gy a care f u i analysis of this salt I find its con- stituents as follows * 1 atom sulphuric acid 5 1 atom soda . 4 10 atoms water 11-25 20-25 The water of crystallization is not always exactly the same. I have found it amount to 11-28, which exceeds the true atomic quantity by 0-03 SODA AND SODIUM. 59 grain. This excess of water is mechanically lodged between the plates of the crystals ; it may be got rid of by reducing the crystals to powder, and pressing them between two folds of blotting paper. But, as the salt is very easily rendered anhydrous, it is better to employ it for the experiment which I am going to describe, in that state. It may be rendered anhydrous by placing it for some hours, in the vacuum of an air-pump, along with an open vessel of sulphuric acid or by exposing it to a moderate heat in a platinum crucible. Nine grains of anhydrous sulphate of soda, and 13'25 grains of chloride of barium, were dis- solved in separate portions of distilled water. The two solutions being mixed, a double decora position took place, and sulphate of barytes pre- cipitated in the state of a white powder. As soon as the liquid had become transparent, it was tested with sulphate of soda and muriate of ba- rytes, but was not rendered opalescent, nor in the least altered, by either of these reagents. It therefore contained no sensible quantity of ba- rytes or sulphuric acid. From this experiment, it follows, that the sulphuric acid in 9 grains of anhydrous sulphate of soda is just saturated by the barytes from 13-25 grains of chloride of bari. urn. Now, this quantity we know is exactly 5 grains ; consequently, anhydrous sulphate of soda is composed of 260 FIXED ALKALIES AND ALKALINE EARTHS. Sulphuric acid 5 Soda . . 4 9 Atomic ]pj ve being the.atomic weight of sulphuric acid, weight of L and the salt being neutral, 4 must be the atomic weight of soda. 2. The atomic weight of soda being known, it is easy, from the experiments of Gay-Lussac and Thenard, to determine its composition, and to deduce the weight of an atom of sodium. 100 grains of sodium, when plunged into water, Atom of were immediately converted into soda, while 198*3 cubic inches of hydrogen gas were evolved. From this experiment, it follows, that when 100 grains of sodium are converted into soda, they combine with 99*15 cubic inches of oxygen gas, or 33-6 grains. This experiment gives us the composition of soda as follows : Sodium 100 or 2-976 Oxygen 33-6 - 1 3-976 We see that soda is a compound of 2 -976 of so- dium and 1 of oxygen ; or of 1 atom sodium and 1 atom oxygen. The atom of sodium, by this experiment, is 2*976, and the atom of soda 3*976. This is a little below the truth ; for we have just seen that the true atomic weight of soda is 4 ; but the experiment is sufficient to satisfy us that SODA AND SODIUM. 26 1 soda is a compound of 1 atom sodium and 1 atom oxygen ; and that an atom of sodium weighs 3. JBerzelius tried to determine the quantity of oxygen which converts a given weight of sodium into soda, by the same method which he followed in determining the composition of potash. The result obtained was, that soda is composed of Sodium 100 or 2-896 Oxygen 34-52 - 1 3-896 This result deviates a little farther from the truth than that of Gay-Lussac and Thenard. Sir H. Davy made a variety of experiments with the same object in view, and drew as a con- clusion, that soda is a compound of from J5 to 73 sodium, and from 25 to 27 oxygen. One of his experiments gave him the composition of soda as follows : Sodium 100 or 3 Oxygen 33^ 1 Now, this result agrees exactly with the atomic weight of soda, as determined in the first para- graph of this section. There can be no doubt that the atom of sodi- um weighs 3 and that soda is a compound of 1 atom sodium and 1 atom of oxygen. R s 262 FIXED ALKALIES AND ALKALINE EARTHS. 3. Sodium, like potassium, combines with an additional dose of oxygen, and forms a solid of a greenish yellow colour, distinguished by the name of peroxide of sodium. From the expe- riments of Gay^Lussac and Thenard, it appears, that sodium in the peroxide is united with 1 times as much oxygen as in soda.* Hence, the peroxide is composed of Sodium 3 Oxygen 1-5 4-5 So that the atomic weight of this peroxide is 4*5, and it is a compound of 1 atom sodium and li atom oxygen ; or of 2 atoms sodium and 3 atoms oxygen. It is curious that soda has a tendency to combine in the same proportion, with several of the acids, that sodium does with oxygen. SECT. III. OF THE ATOMIC WEIGHTS OF LITHIA AND LITHIUM. THOUGH litliia has been known for only a few years, and though it has been hitherto procured only in small quantities, yet its properties have Uccherches Physico-Chcm. I. 152. 9 V LITHIA AND LITHIUM. 263 been examined with so much care, and some of the salts which it forms have been analyzed with such accuracy, that its atomic weight has been as- certained with as much precision as that of any of the alkaline bases. I have examined several of the salts of lithia, but found the sulphate best adapted for a rigid determination of the atomic weight. 1. Sulphate of lithia crystallizes very well by Analysis of spontaneous evaporation, even when the quantity i. att of it is small. It forms four-sided rectangular prisms with square bases, which are not liable to deliquesce when exposed to the air. Its taste is saline, without any perceptible bitterness. It contains water of crystallization, which is easily driven off by exposing the salt to a heat of about 600. This salt was analyzed by Arfwedson, Vau- quelin, and Gmelin, in an anhydrous state : but Strom eyer has stated both the saline constituents and the water of crystallization. According to Arfwedson,* the constituents of this salt are Sulphuric acid 68 '41 or 5 Lithia 31-59 - 2308 100-00 Vauquelint states its constituents as follows : * Ann. de Chim. et de Phys. X. 90. f Ibid. VII. 287 264 FIXED ALKALIES AND ALKALINE EARTHS. Sulphuric acid 69-2 or 5 Lithia 30-8 - 2-225 100-0 The salt being ueutral, and 5 being the atomic weight of sulphuric acid, the corresponding number for lithia must denote the weight of an atom of that alkali, according to the determina- tion of these chemists. Lithia, consequently, according to the analysis of Arfwedson, weighs 2-308, according to that of Vauquelin, its weight is 2'225. The mean of these two ana- lyses gives us 2*266 for the atomic weight of li- thia. This approaches very near 2'25, which, as I shall show immediately, is the true atomic weight of lithia. Gmelin's* analysis (when corrected) is as fol- lows : Sulphuric acid 6'7'15 or 5 Lithia 32-85 - 2-416 100-00 The atomic weight of lithia derived from this analysis, is higher than that given by Vauquelin, or even by Arfwedson, amounting to 2'4l6. But, if we compare Gmelin's analysis of car- bonate of lithia with his analysis of the sulphate, it will be evident that they cannot be quite ac- * Gilbert'* Annalen, IJX, 238. L1THIA AND LITHIUM. - 265 curate. The fused carbonate he found a com- pound of Carbonic acid 54-46 or 2-75 Lithia 45-54 - 2-2995 100-00 So that the atomic weight of lithia, by his analy- sis of the sulphate, is 2*416, while by the analysis of the carbonate it is only 22995- This dis- cordancy would render it hazardous to employ Gm elm's analysis for the determination of the atomic weight of this alkali. Stromeyer's analysis is more recent than those of the three above named chemists ; and it comes nearer the truth than any of the others.* 100 parts of the crystals contain (according to him) 13 parts of water, and the anhydrous salt (cor- recting the calculations) is a compound of Sulphuric acid 68*977 or 5 Lithia 31-023 - 2-2488 100 The water, when reduced to the atomic propor- tions, is 1*083, or very nearly one atom. 2. I exposed 8*375 grains of very carefully prepared crystals of sulphate of lithia, to the temperature of 600, in a platinum crucible, till * Untcrsuchungen, I. 4-36. 266 FIXED ALKALIES AND ALKALINE EARTHS. the salt ceased to give out moisture. The loss of weight was exactly 1*125 grain, which is equivalent to 1 atom. The 7*&5 grains of an- hydrous salt were dissolved in a minimum of dis- tilled water, and the liquid was mixed with a solution of 13*25 grains of chloride of barium. After the sulphate of barytes had subsided, the residual liquid was tested with sulphate of soda and muriate of barytes ; but neither of these re- agents produced any opalescence, showing that it contained neither barytes nor sulphuric acid. Consequently, the sulphuric acid in 7*25 grains of anhydrous sulphate of lithia amounts to exactly 5 grains. The remaining portion of the salt be- ing lithia, it is obvious that the true composition of this salt is as follows : Sulphuric acid 5 or 1 atom Lithia . . 2-25 - 1 atom Water . . 1-125 - 1 atom 8-375 Atom of Thus, the true atomic weight of lithia is 2*25, and the sulphate contains an atom of each of its constituents. iftSif 3 * ^ was ascer t ame d by Sir H. Davy, that li- thia, like potash and soda, is a compound of a white coloured metal, called lithium, and oxygen. Though no experiments have been made to de- termine the proportion of these two constituents, there is no difficulty in ascertaining the true BARYTES AND BARIUM. 26? composition of lithia, now that we know the atomic weight of this alkali and of oxygen. Potash and soda are each of them compounds of 1 atom of oxygen, with 1 atom of potassium and sodium respectively. We shall find that the composition of barytes, strontian, lime, and mag- nesia, is quite similar. It is reasonable to con- clude from this, that lithia is likewise a com- pound of 1 atom oxygen and 1 atom lithium ; consequently, an atom of lithium must weigh 1-25, and lithia must consist of 1 atom lithium 1*25 1 atom oxygen 1 2-25 SECT. IV. OF THE ATOMIC WEIGHTS OF BARYTES AND BARIUM. BARYTES has been employed in many of the pre- ceding sections to determine the atomic weight of different bodies, in consequence of the pro- perty which it has of forming an insoluble com- pound with sulphuric acid. It has been already explained, that sulphate of barytes is a compound of 268 FIXED ALKALIES AND ALKALINE EARTHS. Sulphuric acid 5 Barytes 9- 7 5 14-75 From which it is obvious, that 9'7<5 must be the atomic weight of barytes ; but a few additional observations may be requisite to prevent any ob- scurity from incommoding the progress of the reader. ** When a current of chlorine gas is made to p ass over red hot barytes, this gas is absorbed, while at the same time oxygen gas is evolved and the barytes is changed into chloride of bari- um. Sir H. Davy, to whom we owe this experi- ment, ascertained, that for every volume of chlo- rine gas absorbed by the barytes, half a volume of oxygen gas was evolved. It is obvious that the oxygen gas was displaced from the barytes by the chlorine, and of consequence, the barytes was converted into barium. Half a volume of oxygen gas and a whole volume of chlorine gas, being each equivalent to an atom, it is clear that barytes must be a compound of 1 atom barium and 1 atom oxygen ; consequently, an atom of barium weighs 8 '75. %. Muriatic acid was shown, in a preceding chapter, to be a compound of 1 atom of chlo- rine = 4%5 ; and 1 atom of hydrogen = O1&5 ; so that its atomic weight is 4 f 6&5. 3. Water was likewise shown to be a com- BARYTES AND BARIUM. 269 pound of 1 atom oxygen and 1 atom hydrogen, and its atomic weight is 1/1&5. 4. An integrant particle of anhydrous muriate Muriate of of barytes weighs 14-375 ; being composed of 1 atom muriatic acid 4-625 1 atom barytes 9-7 5 14-375 5. When muriate of barytes is exposed to a red heat, it undergoes a double decomposition. The hydrogen of the acid and the oxygen of the base uniting together, constitute a particle of water, which is driven off by the heat, while the chlo- rine and the barium remain united together in the state of chloride of barium. For it is obvi- ous, that 14-375 of muriate of barytes may be decomposed into 1 atom water = 1-125 1 atom chloride of barium 13-25 14-375 6. Whenever chloride of barium is dissolved in water it is again converted into muriate of ba- rytes. A portion of the water is decomposed, the hydrogen of which converts the chlorine into muriatic acid, while the oxygen changes the ba- rium into barytes. 7. Thus, it appears, that whenever muriate of barytes is exposed to a red heat, it is converted 270 FIXED ALKALIES AND ALKALINE EARTHS. into chloride of barium ; and that when chloride of barium is dissolved in water, it becomes mu- riate of barytes. We may, therefore, employ 13 '25 grains of chloride of barium as an equiva- lent for 14*375 grains of muriate of barytes, and as capable of yielding 9*75 grains of barytes. These things being understood and recollected, we are prepared for an experimental investigation of the atomic weight of barytes. Atomic 13'25 grains of chloride of barium, and 11 weight of barytes. grains of sulphate of potash, were dissolved in separate portions of distilled water, and the two solutions were mixed together. A double de- composition took place, sulphate of barytes pre- cipitated to the bottom of the vessel, while mu- riate of potash remained in solution. Both of these salts were quite neutral, and the solution contained no traces either of sulphuric acid or of barytes. The sulphate of barytes being collected weighed 14*75 grains ; the solution being eva- porated to dryness, left 9*5 grains of chloride of potassium. Now, the weights of these two salts amount to 24*25 grains, which is just equal to the combined weights of the chloride of barium, and sulphate of potash employed in making the experiment. 11 grains of sulphate of potash contain 5 grains of sulphuric acid, which must exist in the 14*75 grains of sulphate of barytes. Conse- quently, that salt must be a compound of BARYTES AND BARIUM. 271 Sulphuric acid 5 Barytes 9'75] 14-75 The salt being neutral, and 5 being the atomic weight of sulphuric acid, it is plain that 9*7^ must be the atomic weight of barytes. Muriate of barytes is a white, pungent, and disagreeably tasted salt, which crystallizes in very short four-sided prisms with square bases. I find by a careful analysis of these crystals, that they are composed of 1 atom muriatic acid = 4*625 1 atom barytes = 9'750 1 atom water = 1/125 15-5 If 15-5 grains of it be exposed to a red heat, they are reduced to 13-25 grains ; so that the loss sustained, owing to the escape of water, is 2-25 grains. This loss is equivalent to two atoms of water; one of these atoms existed in the salt as water of crystallization ; the other atom was formed by the muriate of barytes being con- verted into chloride of barium. Gay-Lussac and Thenard ascertained that Peroxide of * * banum. when anhydrous barytes is heated in oxygen gas, a portion of the gas is absorbed, and the barytes is converted into a peroxide. This peroxide is FIXED ALKALIES AND ALKALINE EARTHS. grey ; but we have no data to determine how many atoms of oxygen it contains. SECT. V. OF THE ATOMIC WEIGHTS OF STRONTIAN AND STRONTIUM. THE atomic weight of strontian is easily deter- mined by experiments of a similar kind with those related in the last section. It will be suf- ficient if I describe one of them. Atom of i. Nine grains of anhydrous sulphate of soda, strontian J determined, and ten grains of chloride of strontium, (that is, muriate of strontian, exposed to a red heat,) were dissolved in separate portions of water. These two solutions being mixed together, a double decomposition took place. Sulphate of strontian precipitated abundantly, and the residual liquid contained a quantity of common salt, which was perfectly neutral, as no change was produced on paper stained with litmus or cudbear. This residual liquid being tested by muriate of bary- tes, did not give any indication of the presence of sulphuric acid. Nor was it rendered in the least turbid by sulphate of soda, arseniate of soda, or chromate of potash, even when very much concentrated by evaporation. Conse- quently, it contained no sensible quantity of strontian ; the sulphate of strontian being care- STRONTIAN AND STRONTIUM, fully collected and exposed to a red heat, weighed 11 '5 grains. Now, it contained all the sulphuric acid in 9 grains of anhydrous sulphate of soda, which is just 5 grains. The remaining 6%5 grains must be strontian. Thus it appears, that sulphate of strontian is composed of Sulphuric acid 5 Strontian 6-5 11-5 5 being the atomic weight of sulphuric acid, and the salt being neutral, G5 must be the atomic weight of strontian. The soda in 9 grains of anhydrous sulphate of soda amounts to 4 grains, which requires for saturation 4*625 grains of muriatic acid, equiva- lent to 4%5 grains of chlorine and this must be the quantity of chlorine in 10 grains of chloride of strontium. If from 10 we subtract 4\5, the Atom . of strontium. remainder is 5%5, which indicates the quantity of strontium in 10 grains of this chloride. This 5-5 of strontium, when converted into strontian in the sulphate, becomes 6\5. Hence, it is ob- vious that strontian is a compound of 1 atom oxygen, and 1 atom strontium ; and tha an atom of strontium weighs 5%5. Muriate of strontian crystallizes in long needles, Muriate of . . ,. . , , T n strontian. consisting of six-sided prisms, two opposite races of which are broader than the other four. By VOL. I. S 274 FIXED ALKALIES AND ALKALINE EARTHS. a careful analysis of this salt, I have ascertained its composition to be as follows : 1 atom muriatic acid = 4-625 1 atom strontian =6*5 8 atoms water = 9 20-125 So that an integrant particle of it weighs 20'125. When 20*125 grains of these crystals are ex- posed to a red heat in a platinum crucible, they lose 10' 125 grains of water, which is equivalent to 9 atoms ; and there remain behind 10 grains of chloride of strontium, which is equivalent to the weight of an integrant particle of that chlo- ride. Thus, muriate of strontian, like muriate of barytes, loses 1 atom of water more than it contains in the state of water. This atom is produced by the double decomposition of the muriate of strontian into an atom of water, and an atom of chloride of strontium, precisely in the manner explained in the last section, when treating of the conversion of muriate of barytes into chloride of barium by heat. SECT. VI. OF THE ATOMIC WEIGHTS OF LIME AND CALCIUM. THE salts of lime are either insoluble in water, or they contain much water of crystallization, of LIME AND CALCIUM. which they cannot be completely deprived, with- out the risk of losing at the same time a portion of their acid. On this account, the method which I was obliged to employ in determining the atomic weight of lime, was rather more com- plicated than that which gave the atomic weights of the other alkaline bodies contained in this chapter ; though it was equally decisive. 1. 6*25 grains of pure calcareous spar (kept for twenty-four hours at a temperature higher lime deter. than that of boiling water) were dissolved in muriatic acid, and the solution evaporated to dryness in a low temperature, in order to get rid of all excess of acid. The muriate of lime, thus rendered neutral, was re-dissolved in a small quantity of distilled water. In another portion of distilled water I dissolved 8 *75 grains of pure anhydrous carbonate of potash. These two solutions being mixed, a double decomposi- tion took place, the lime and carbonic acid precipitated in the state of carbonate of lime, while the muriatic acid and potash remained in solution in the state of muriate of potash. The clear liquid remaining after the precipita- tion of the carbonate of lime, produced no change in the colour of cudbear pape and con- sequently contained no sensible quantity of carbonate of potash ; and as it was not precipi- tated by oxalate of ammonia, it contained no sensible quantity of lime. S 2 FIXED ALKALIES AND ALKALINE EARTHS. The carbonate of lime precipitated, being washed and dried, weighed just 6-5 grains, or exactly the same as the calcareous spar originally dissolved in the muriatic acid. It was obviously a compound of'all the lime in 6'&5 grains of calcareous spar, and all the carbonic acid in 8 75 grains of anhydrous carbonate of potash. This carbonate of potash had been procured by exposing pure crystals of bicarbonate of potash to a red heat in a platinum crucible. It has been ascertained that this carbonate is an- hydrous, and that it is a compound of 1 atom carbonic acid = 2 '75 1 atom potash = 6 8-75 From this it is obvious, that the 6 '25 of carbon- ate of lime contained just 2'75 of carbonic acid, the remainder being lime, it is evident, that this carbonate is composed of Carbonic acid 2-75 Lime 3-50 6-25 2*75 being the atomic weight of carbonic acid, and the salt being neutral, 3*5 must be the weight of an atom of lime. It is clear likewise, that calcareous spar is a compound of one atom carbonic acid and one atom lime. LIME AND CALCIUM. 277 C 2. If a current of chlorine gas be made to pass Atomic n T i J weight of through a quantity or lime heated to redness, calcium. this alkaline earth is decomposed, oxygen gas is disengaged, and chloride of calcium is formed. Sir H. Davy, to whom we are indebted for this experiment, ascertained that for every half vo- lume of oxygen gas extricated, a whole volume of chlorine gas was absorbed. Now, these two being equivalent, respectively, to one atom, it is obvious that lime must be a compound of one atom oxygen and one atom calcium. Conse- quently, the atomic weight of calcium is 2%5. 3. Muriate of lime may be easily crystallized, by placing a concentrated solution of it under the exhausted receiver of an air pump, over a flat glass vessel containing a quantity of concen- trated sulphuric acid. The crystals are large transparent four or six-sided prisms, usually striated longitudinally. They are composed of 1 atom muriatic acid 1 atom lime . 6 atoms water 14-875 When this muriate is cautiously exposed to a heat gradually raised to redness, it may be fused and converted into chloride of calcium with scarcely any loss of chlorine or muriatic acid. In that case it loses 7 atoms of water, 14*875 grains of the crystallized muriate being by this s 3 278 FIXED ALKALIES AND ALKALINE EARTHS. treatment reduced to 7 grains of chloride of cal- cium. I shall relate one experiment to show the reader how near to these numerical results I have been able to come. 12*5 grains of pure calcareous spar (containing 7 grains of lime) were dissolved in muriatic acid, in a plati- num crucible, the weight of which had been previously determined. The solution was slowly evaporated to dry ness, and the dry salt was fused by exposing the salt to a red heat. The fused salt was white and opaque, and weighed 13*98 grains, as nearly as could be determined. Water being poured upon it, there remained undissolved 0*1 7 grain of a white powder con- sisting chiefly of carbonate of lime ; but contain- ing also one or two flocks of sulphate of lime, apparently derived from a trace of sulphuric acid, in the muriatic acid employed to dissolve the calcareous spar. Now, 0*17 carbonate of lime contain 0*09-5 lime = 0-0678 calcium, which requires for saturation 0*122 chlorine, making together 0*1898 chloride of calcium ; which ex- ceeds the 0*17 grain of carbonate of lime by very nearly 0*02 grain ; making (as near as I could determine) almost exactly 14 grains. Now, these 14 grains contained 2 atoms chlo- rine = 9, and 2 atoms calcium, which must amount to 5. So that the atomic weight of cal- cium is obviously 2*5 ; and lime is a compound of 1 atom oxygen and 1 atom calcium. MAGNESIA AND MAGNESIUM. 279 SECT. VII. OF THE ATOMIC WEIGHTS OF MAGNESIA AND MAGNESIUM. THE atomic weight of magnesia may be deter- mined in the same way as that of the fixed alka- lies. For sulphate of magnesia, which crystallizes in four-sided prisms with square bases, and is per- fectly neutral, may be rendered anhydrous by heat, without losing any of its acid. 1. 7*5 grains of anhydrous sulphate of mag- Atom of nesia, and 13*%5 grains of chloride of barium, were dissolved in two separate portions of dis- tilled water. The two solutions being mixed together and agitated, a double decomposition took place, sulphate of barytes precipitated, while muriate of magnesia remained in solution. The supernatant liquid, after the sulphate of barytes had subsided, was perfectly neutral. Being tested by sulphate of soda, and muriate of barytes, it did not become in the least muddy by the addition of either of these reagents, showing that it contained neither barytes nor sulphuric acid. Hence, the sulphuric acid in 7*5 grains of anhydrous sulphate of magnesia just saturates the barytes from 13*25 grains of chloride of ba- rium so that it amounts to precisely 5 grains. The remainder of the salt consisting of magne- S 4 ton 280 FIXED ALKALIES AND ALKALINE EARTHS. sia, it is evident that anhydrous sulphate of mag- nesia is a compound of Sulphuric acid Magnesia 7-5 The salt being neutral, and 5 being the atomic weight of sulphuric acid, it is plain that 2%5 must represent the atomic weight of magnesia. posi- 2. By a careful analysis of the crystallized of sul- J \ sulphate or magnesia, I have ascertained that it magnesa. . i is a compound ot 1 atom sulphuric acid = 5 1 atom magnesia = 2 5 7 atoms water = 7 '87 5 15-375 So that an integrant particle of it weighs 15-375. When 15*375 grains of these crystals are heated properly, the whole water may be driven off without the loss of any sulphuric acid or magne- sia 5 for the weight of the anhydrous salt is exactly 7*5 grains. 15-375 grains of crystallized sulphate of magnesia just precipitate all the ba- rytes contained in a solution of 13*25 grains of chloride of barium. Muriate of 3. Muriate of magnesia is a deliquescent salt ; magnesia. . i i i but it may be obtained in large transparent crys- MAGNESIA AND MAGNESIUM. 281 tals, which by a careful analysis I find composed as follows : 1 atom muriatic acid = 4'625 1 atom magnesia = 2-5 5 atoms water . = 5'625 12-75 If it could be rendered anhydrous without the loss of its acid, it would doubtless be reduced to chloride of magnesium, weighing only 6 grains ; but this cannot be done ; for when the muriate is heated, not only the water, but a considerable portion of the acid flies off and leaves the magnesia. I may state an experi- ment, however ; because it seems to show that a chloride of magnesium might be formed, if the requisite precautions were taken. It was the only successful trial that I made, though my ex- periments amounted to five or six. 15 '31 grains of pure anhydrous carbonate of magnesia, equiva- lent to 7'29 grains of magnesia, were dissolved in muriatic acid, and the solution being evapo- rated to dryness, the muriate of magnesia was ex- posed to an incipient red heat, with every pre- caution which I could devise to prevent the escape of the muriatic acid. The salt thus treated weighed 13*3 grains, and had a whitish'colour. Water being poured upon it, the greater part was dissolved with the evolution of much heat, FIXED ALKALIES AND ALKALINE EARTHS. and there remained undissolved 3 grains of pure magnesia. Now, if we suppose the 1O3 grains which dissolved, to have been previously in the state of chloride of magnesium, they must have consisted of Magnesium 2'575 Chlorine 7'725 10-300 And 2'575 grains of magnesium are equivalent to 4*29 grains of magnesia. Now, 4*29 + 3*= 7*29, which is the amount of the magnesia ori- ginally present in the carbonate employed. It would appear that in this experiment a chloride of magnesium was obtained. Atom of Jt j s obvious, if the preceding experiment magnesi- um, deserves any confidence, that magnesia is a com- pound of 1 atom magnesium 1-5 1 atom oxygen 3 -0 2-5 And this is corroborated by the following ex- periment originally made by Sir H. Davy. When magnesia is heated, and a current of chlo- rine gas passed over it, oxygen is given out, and the magnesia is converted into chloride of mag- nesium. Now, for every volume of chlorine absorbed, half a volume of oxygen gas is given out. This shows that magnesia contains only 1 MAGNESIA AND MAGNESIUM. 283 atom of oxygen. Consequently, the weight of an atom of magnesium must be 1-5. From the details contained in the different sections of this chapter, it is evident that the fixed alkalies and alkaline earths are all anala- gous substances, being each a compound of one atom of a metallic basis, and one atom of oxy- gen. Potassium, sodium, and barium combine with two doses of oxygen ; though this has not been shown to hold with the other four bases contained in this chapter, yet it is probable, from the striking analogy between the properties of all these bodies, that they will be found hereafter to be in the same predicament. Lithia is the lightest, and barytes the heaviest of these bodies. It may be worth while to sub- join a tabular view of the atomic weights of the metallic bases and the alkalies, arranged accord- ing to the weight. ATOMIC WEIGHTS. Lithium 1-25 Magnesium Calcium 1-5 2-5 Sodium 3 Potassium 5 Strontium 5-5 Barium 8-75 ALKALIES. ATOMIC WEIGHTS. Table of the atoms of the alka- line metals and alka- lies. Lithia 2-25 Magnesia 2-5 Lime 3-5 Soda 4 Potash 6 Strontian 6-5 Barytes 9-75 CHAP. IX. OF THE ATOMIC WEIGHTS OF THE EARTHS PROPER. THERE are four substances, which are now considered as metallic oxides, but which were long distinguished by the name of earths proper. When pure, they have the form of white taste- less powders, insoluble in water, though capa- ble of forming solid hydrates with that liquid. They combine readily with the acids, and form salts, which redden vegetable blues, and are de- composed by exposure to a red heat. These four bodies are alumina, glucina, yttria, and zir- conia. We shall devote this chapter to the in- vestigation of the atomic weights of these bo- dies. SECT. I. OF THE ATOMIC WEIGHT OF ALUMINA. IN a paper inserted in the Annals of Philoso- phy (Second Series), Vol. III. p. 61. I gave the result of a very laborious analysis of alum, from which I drew, as a conclusion, that the atomic weight of alumina was 2*25. Since that time, Mr. 286 EARTHS PROPER. R. Philips has published some experiments on the sulphate of alumina, from which he considers it most probable that the atomic weight of alu- mina is 3*375. This opinion, from a man of Mr. Philips' precision and extensive chemical know- ledge, induced me to repeat the whole of my experiments, and to extend my researches as far as possible, in order to throw as much light as I could upon this very important, but difficult sub- ject. I shall in this section lay the result of my experiments before the reader. There are three methods, by means of which ing the we may hope to come to an accurate conclusion atom of J , -i n i alumina, respecting the atomic weight 01 alumina. 1st. It enters as a constituent into a great number of crystallized minerals, which appear to be definite combinations of acids and bases. By a rigid analysis of these minerals in a state of purity, and by our knowledge of the atomic weights of all the other constituents, we may determine the quantity of alumina which unites with a deter- minate weight of the acid in the mineral, which will give us the atomic weight of alumina, or at least an approximation to it. 2d. Several arti- ficial salts exist, containing alumina as one of the constituents. A precise analysis of these salts, in a state of purity, must lead to an accurate de- termination of the saturating power of alumina, and consequently give us its atomic weight. 3d. Alumina enters into definite combinations with ALUMINA. 287 water, and forms a variety of hydrates ; a precise analysis of which will also throw considerable light upon the atomic weight of alumina. Let us see what number we shall obtain for the ato- mic weight of alumina, deduced in each of these ways. I. The number of minerals containing alumina . . . -111 as an essential constituent is very considerable, sitkm of The acid substance with which it is most com- m monly combined is silica. It occurs, indeed, combined with phosphoric acid, fluoric acid, and mellitic acid ; but the mode of analyzing such combinations is still too imperfect to render it safe for us to trust to them for determining a point of such importance. Beryl would be an excellent mineral for our purpose, were it not that the atomic weight of glucina has not been yet determined with sufficient precision to al- low us to employ it in investigating the atomic weight of alumina. I shall, therefore, satisfy myself with the following mineral species, all of which have been analyzed with much care, and in a state of as great purity as possible ; but I have investigated the atomic weight of alumina from a much greater number of species, and have obtained very nearly the same result as from the species which we are going to examine. 288 EARTHS PROPER. 1. NEPHELINE. It was stated in p. 177 of this work, that from the late analysis of nepheline, by Arfwedson, its constituents are as follows : Silica . 44-11 Alumina . 33-73 Soda . 20-46 Water . 0-62 98-92 It was shown that nepheline is a compound of 3 integrant particles of silicate of alumina, and 1 integrant particle of sesquisilicate of soda. It was shown that the portion of silica united to the alumina was 29*17 grains. Now, 29-17 (the quantity of silica) : 33-?3 (the alumina) : : 2 Gives (atom of silica) : 2-3126 = atom of alumina. 2-3126. 2. KARPHOLITE. Karphoiite This rare mineral occurs in the Schlacken- wald in Bohemia, and was named by Werner, from the resemblance of its colour to straw. It occurs in very fine needles, which usually diverge from a point. Its specific gravity is 2'923 ; it seems to be rather soft. Strom eyer,* from a very careful analysis of this mineral, has given its constituents as follows : * Untersuchungen, I. p. 410. ALUMINA. C 289 Silica . . . 36-154 Alumina . . . 2 8 -669 Deutoxide of manganese IQ-lGO Peroxide of iron . . 2-290 Lime . . . 0-271 Fluoric acid . . 1-470 Water 10-? 80 98-794 It appears from Professor Stromeyer's experi- ments on this mineral, that it is usually mixed with particles of fluor spar. To the presence of this spar in minute quantities, I have no doubt that the lime and the fluoric acid found by Stro- meyer owe their origin. It is true that the lime and fluoric acid were not found in the same pro- portions as they exist in fluor spar ; for I shall show, in a subsequent chapter of this work, that fluate of lime is a compound of Fluoric acid 1-25 Lime . 3-5 4-75 But the accurate analysis of this mineral is at- tended with so much difficulty, that much great- er errors than those here suspected have been committed by the most expert analysts. The manganese was obtained, and weighed by Stromeyer, in the state of deutoxide ; but it must have existed in the mineral, in the state of VOL, I. T 290 EARTHS PROPER. protoxide. Now, the atomic weight of protoxide of manganese is 4-5, and of deutoxide 5. If we reduce the manganese in the mineral in this pro- portion, the quantity of protoxide will be 17 '244 grains. The peroxide of iron is probably foreign to the constitution of the mineral, and may have been introduced by percollation. Now, if we suppose every atom of alumina to be united to an atom of silica, and every atom of protoxide of manganese to be united to H atom of silica, we shall obtain data for calculating the constituents of the mineral. We shall consider an atom of silica as 2, an atom of alumina as 2-25, and an atom of protoxide of manganese as 4-5. (1.) 2-25 (atom of alumina) : 2 (atom of silica) : : 28-669 (alumina in the mineral) : 25-483 = silica united to the alumina. (2.) 4-5 (atom of protoxide) : 3 (atom and half of silica) : : 17-244 (protoxide in the mineral) : 11-496 = si- lica united to the manganese. But 25-483 + 11-496 = 36-979, which exceeds the whole silica found in the mineral by 0-825 grain ; but 3 1 2 of the silica is united to the man- ganese : consequently, 0-257 of this excess must be referred to that portion of the silica. There remains 0-568, which must be subtracted from the 25-483 grains of silica, given by theory tome ALUMINA. as united to the alumina ; so that the quantity of silica actually united with the alumina, ac- cording to Stromeyer's analysis, is 4-915. Now, 24-915 : 28-669 : : 2 (atom of silica) : 2-3013 = atomic Gives weight of alumina. 3. HARMOTOME. This mineral, which occurs in a state of great purity in the lead mine at Strontian, in Argyle- shire, was found, by Klaproth, to be composed of the following constituents : Silica . . 49 Alumina . 16 Barytes . . 18 Water . . 15 98 Let us consider this mineral as a compound of 3 integrant particles of trisilicate of alumina, and 1 integrant particle of bisilicate of barytes. Then, (1.) 2-25 (atom of alumina) : 6 (three atoms silica) : : l6 (alumina in the mineral) : 42-666 = silica united to alumina. (2.) 9-75 (atom of barytes) : 4 (two atoms silica) : : 18 (barytes in the mineral) : 7*384 = silica united to barytes. But 42-666 + 7-384 = 50-05 ; and this exceeds the silica found in the mineral by 1-05 grain. T2 EARTHS PROPER. The silica united to the barytes being -fth of the whole, fth of this excess must be referred to that portion. There remains O9, which must be subtracted from the 42-666 grains of silica, given by theoryas united to the alumina. This being subtracted, leaves 41 '766 for the silica ac- tually combined with the alumina, according to Klaproth's analysis. Now, 4.1-766 : 16 : : 6 : 2-2985 = atomic weight of alumina. 4. MEIONITE. Of this mineral, which occurs most abundant- ly on Monte Somma, near Naples, but which has been found also in other places, and indeed is probably identical with scapolite, we have se- veral very careful analyses.* According to the experiments of Dr. L. Gmelin of Tubingen, its constituents are as fol- lows :t Silica . . . 40-8 Alumina . . 30-6 Lime . . . 22-1 Soda, with some lithia 2-4 Oxide of iron . 1-0 Carbonic acid and loss 3-1 100-0 * The mineral analyzed by Arfwedson, from which Berzelius has taken his notion of the coin position of nieionite, was obviously a leucite. f Sch\veif?ger, XXV. 36. ALUMINA. 293 By comparing this analysis with the two follow- ing, it will appear probable that the alkali, oxide of iron, and carbonic acid, are merely accidental ingredients ; and that silica, alumina, and lime, are the only essential constituents. Let us consider meionite as a compound of 2 integrant particles of silicate of alumina and 1 integrant particle of silicate of lime ; and let us calculate its constitution. (1.) 2-25 (atom of alumina) : 2 (atom of silica) : : 30-6 (alu- mina in the mineral) : 27'2 = silica united to alu- mina. (2.) 3-5 (atom of lime) : 2 : : 22-1 : (lime in the mineral) : 12-62 = silica united to the lime. But 27-2 + 12-62 = 39-82. Now, the silica actually found in the meionite exceeds this by 0-98 grain. The silica united to the lime amounts to ~ 5 th of the whole ; consequently, ^-jth of 0*98 or 0-311, belongs to the silica united to the lime. There remains 0-669, which being added to the 27*2, makes 27-869 for the silica united to the alumina in meionite by Gmelin's analysis. Now 27-869 : 30-6 : : 2 : 2-196 = atomic weight of alumina. Professor Stromeyer analyzed some very pure specimens of meionite from Monte Somma, and found the constituents as follows:* * Untcrsuchungen, I. 580. T 3 294 EARTHS PROPER. Silica . . . 40-531 Alumina . . . 32*726 Lime . . . 24-24-5 Potash with some soda 1-812 Oxide of iron . . 0-182 99.4,96 Here the alkali differs both in quantity and kind from that found by Gmelin, and the oxide of iron is not quite jth of the quantity obtained by that chemist. Let us calculate the composition of this mineral, on the same supposition as we made when calculating the result of Gmelin's. (1.) 2-25 : 2 : : 32-726 : 29-08 = silica united to the alu- mina. (2.) 3-5:2:: 24-245 : 13-854 = silica united to lime. But 29-08 + 13-854 = 42-934, which exceeds the silica found in the mineral by 2-403 grains. The silica united to the lime is g^yth of the whole. Now jfoth of 2-403 is 0*777, which be- ing subtracted from 2*403 leaves 1-626 to be taken from the 29*08, making 27*454 for the amount of the silica actually united to the alu- mina. Now, 27*454 : 32-726 : : 2 : 2-3841 = atomic weight of alumina. Professor Strom eyer likewise analyzed a spe- cimen of meionite from Sterzing in the Tyrol, and found its constituents as follows :* * Untersuchungen, I. 590. 16 ALUMINA. 295 Silica Alumina Lime Potash with some soda Protoxide of iron Protoxide of manganese Water 100-000 The variation in the proportion of alkali and ox- ide of iron is equally conspicuous in this analysis. Let us calculate the constitution of the mineral on the former supposition. (1.) 2-25 : 2 :: 31-97 : 28-417 = silica united to alumina. (2.) 3-5:2:: 23-856 : 13-632 = silica united to lime. But 28-417 + 13-632 = 42-049, which exceeds the quantity of silica present in the mineral by 2-134 grains. The silica united to the lime is ~tli of the whole. Now j^th of 2-134 = 0-69 ; and 2-134 0-69 = 1-444, which must be subtracted from the 28-417, leaving 26-973- And 26-973 : 31-97 : : 2 : 2-3705 == atomic weight of alumina. We have obtained three different atomic weights of alumina, namely, 1. From Gmelin's analysis . . . 2-1960 2. From Stromeyer's analysis of Monte Somma meionite 2-3841 From ditto of Tyrol meionite . . 2-3705 3 ) 6-9506 Gives Mean 2-3168 2-3168. 296 EARTHS PROPER. 5. SCAPOLITE. Scapoiite We have a very careful analysis of this mineral in a state, apparently, of great purity, by Mr. NordenskiolcL* He found its constituents as follows : Silica . . 43-83 Alumina . 35-43 Lime . . 18-96 Water . . 1-03 99-25 The only difference between the composition of meionite and scapolite seems to be, that the former contains only 2 integrant particles of sili- cate of alumina united to 1 integrant particle of silicate of lime, while the latter contains 3. Let us calculate the constitution of this mineral on the same supposition as of meionite : (1.) 2-25 : 2 : : 35-43 : 31-049 = silica united to alumina. (2.) 3-5:2:: 18-96 : 10-834 = silica united to lime. Now, 31-049 + 10-834 = 41-883, which is less than the silica found in the mineral by 1-947 grain. Of this surplus, Jth, or 0.487 must be unit- ed to the lime the remaining 1-460 grain must be added to the 31-049, making 32-409. And Gives 32-409 : 35-43 : : 2 : 2-0580 = atomic weight of alumina. 2 0580. * Bidvag till Narmare Kannedomaf Finland's Mineralier och Geogno- sie, p. 63. ALUMINA. 297 6. SCOLEZITE. This mineral was analyzed with great care by Gehlen and Fuchs, who, from a mean of many trials, state its constituents as follows : Silica . . 46-4 Alumina . 25-7 Lime . 14'2 Water . . . . 13-6 100-0 I have analyzed various specimens of scolezite ; but only one of them was quite free from soda, and its constituents were almost exactly the same as those found by Gehlen and Fuchs. In the other specimens the quantity of soda varied from 0%5 to 1*5 per cent. $ showing that the mineral contained a mixture of natrolite. Let us suppose that scolezite is a compound of 3 integrant particles of silicate of alumina and 1 integrant particle of trisilicate of lime, with 3 atoms of water, and let us see what its constitu- ents will be on that supposition. 2'25 : 2 : : 25-7 : 22-844 = silica united to alumina. 3-5:6: : 14-2 : 24-3428 = silica united to lime. Now, 22-844 + 24-3428 = 47-1868, which sur- passes the silica actually found in scolezite by 0-6868. The silica united to the lime is of the 298 EARTHS PROPER. whole ; and -p^d part of 0-6868 = 0-3557. This being subtracted from 0-6868 leaves 0-3311, which must be taken from 22-844, leaving 22-513 for the silica actually combined with the alumi- na in the mineral. And Gives 2-2831. 22-513 : 25'7 : : 2 : 2-2831 = the atomic weight of alumina. 7 BUCHOLZITE. Buchoizite From the account of this mineral given in page 175 of this work, it appears that its constituents, according to Brande's analysis, are as follows : Silica .... 46 Alumina ... 50 Oxide of iron . . . 2-5 Potash .... 1-5 100-0 And that the silica united to the alumina in this mineral is 44*391 grains. Now, Gives 2-2527.' 44-391 : 50 : : 2 : 2-2527 = an atom of alumina. 8. LEUC1TE. Leucite This mineral, which abounds so much in the neighbourhood of Mount Vesuvius, and which always occurs crystallized, was found by Arfwed- son to be composed of Silica . . . . 56-1 Alumina . . . 23*1 Potash . . . 21. 15 Oxide of iron . 101-3 ALU-I1NA. 299 Let us calculate the composition of this mine- ral, on the supposition that each of the three bases are combined with two atoms of silica, con- stituting as many bisilicates : (1.) 2-25 : 4 : 23-1 : 41-066 = silica united to alumina. (2.) 6:4: :2M5: 14-1= silica united to potash. (3.) 4-5:4:: 0-95 : 0-844 = silica united to iron. Now, 41-066 + 14-1 + 0-844 = 56-01, which is less than the silica actually found in the mineral by 0-09 grain. The silica united to the potash and the iron amounts to 3^Vth of the whole. The 3^43 th of 0-09 is 0-024, which, being subtracted, leaves 0-066 to be added to 41-066, in order to have the silica actually united to the alumina in the mine- ral, it will amount to 41-132 grains. And 41-132 : 23-1 : : 4 : 2-2464 = atomic weight of alumina. Gives 2-2464. 9. THOMSON1TE. ite This mineral, which hitherto has been found Thomson, only in the Kilpatrick hills, near Glasgow, and which had been considered as a mesotype, has been constituted into a new species by Mr. Brooke, on account of the peculiar form of its primary crystal. I made a very careful analysis of this mineral,* and found its constituents as fol- lows: * Annals of Philosophy, XVI. 300 EARTHS PROPER. Silica . 36-8 Alumina 31-36 Lime 15-4 Magnesia Peroxide of iron 02 0-6 Wat* 13-0 97-36 Let us consider it as a compound of three inte- grant particles of silicate of alumina and one in- tegrant particle of silicate of lime, together with 2-J atoms of water ; and let us calculate its com- position according to that opinion. (1.) 2-25 : 2 : : 31-36 : 27'875 = silica united to alumina. (2.) 3-5 : 2 : : 15-4 : 8-8 = silica united to lime. Now, 27-875 + 8-8 = 36-675, which is less than the silica found in the mineral by 0*175 grain. The silica united to the lime is ith of the whole. Now, iths of 0-175 = 0-132, which must be added to the 27-875, in order to have the si- lica actually united to the alumina in the mine- ral ; this makes 28-007. And 23395 28-007 : 31-36 : : 2 : 2-2395 = atomic weight of alumina. 10. NATROLITE. This mineral, which corresponds with the meso- type of Brooke, was analyzed with great care by Gehlen and Fuchs. They found its composition, from a mean of many experiments, as follows : ALUMINA. 301 Silica ... 4-8 Alumina . . 26-5 Soda . . . 16-2 Water . . . 9 -3 100-0 I analyzed a beautifully crystallized and pure specimen, which I received from Mr. Heuland, and found its composition almost identical with the preceding analysis. Let us calculate the composition of this mine- ral, on the supposition that it is a compound of 3 integral particles of silicate of alumina, and 1 integrant particle of trisilicate of soda, with 2 atoms of water. (1.) 2-25 : 2 : : 26'5 : 23-555 = silica united to alumina. (2.) 4 : 6 : : 16-2 : 24-3 = silica united to soda. Now, 23-555 + 24-3 = 47*855, which is less than the silica found in the mineral by O145 grain. One half of the silica being united to the lime, it is obvious that, in order to have the true quantity of silica united to the alumina in natro- lite, we must add O07&5 to the 23-555 grains, given by theory. And 23-6275 : 25-7 : : 2 : 2-1 754 = atomic weight of alumina. 3.1754. 11. FELSPAR. Felspar consists essentially of silica, combined Felspar 302 EARTHS PROPER. with alumina of potash. It has been often ana- lyzed ; but the different results obtained do not well agree with each other. The reason is, that felspar almost always occurs in granite, which is a compound roek ; and the other constituents, doubtless, mix themselves with it in greater or smaller proportions. According to the analysis of felspar by Rose, which is one of the latest and most careful that have been made, the con- stituents of felspar are as follows : Silica . . . 66-75 Alumina . . . 17-50 Potash . . . 12 -00 Lime . . . 1-05 Oxide of iron . . 0-75 98-25 Let us suppose that all the bases are combined with silica, and in the state of trisilicates, except potash, which is a quadrosilicate ; and let us calculate its constitution according to that sup- position. (1.) 2-25 : 6 : : 1 7'5 : 46-666 = silica united to alumina. (2.) 6:8:: 12 : 16 = silica united to potash. (3.) 3-5 : 6 : : 1-25 : 2-143 = silica united to lime. (4.) 4-5 : 6 : : 0-75 : 1 = silica united to oxide of iron. Now, 46-666 + 16 + 2-143 + 1 = 65-809, which is less than the silica in the mineral by 0-941. The silica united to the potash, lime, and iron, ALUMINA. 303 amounts to ^WrrTrth part of the whole. And the y-*T77rrth part of 0-941 = 0-273, which being taken from 0-941 leaves 0-668, which must be added to the 46-666 given by theory. Now 47-334 : 17'5 : : 6 : 2-2183 = atomic weight of alumina. Gives 2-2183. 12. ALBITE. Albite is a name given to a mineral which was Albite long confounded with felspar. Eggertz * first pointed it out as a distinct mineral j it was after- wards analyzed by Stromeyer from Chesterfield in North America, by whom it received the name of IdesdspatliA Mr. Brookes has called it cleavelanditeA Rose has given a minute descrip- tion of the crystals of this mineral, which differ a good deal from those of felspar. Stromeyer found an uncrystallized specimen from North America composed of Silica . . . 70-676 Alumina . . . 19-801 Soda . . . 9056 Lime . . . 0-235 Oxide of iron and manganese 0-111 99-879 Rose found the specific gravity of albite to vary from 2-608 to 2-619. The constituents of the crystals he found * Af hondlingar, V. p. 27. f Untersuchungen, p. 300. \ Introduction to Crystallography, p. 459. 304 EARTHS PROPER. Silica . 68-46 Alumina . 19-30 Soda . 9-12 Lime . 0-68 Oxide of iron O28 97-84* Let us suppose the bases combined with silicon in^the state of trisilicates, and let us calculate the composition on that supposition from Rose's analysis. (1.) 2-25 : 6 : : 19-3 : 51-466 = silica united to alumina. (2.) 4 : 6 : : 9-12 : 13-68 = silica united to soda. (3.) 3-5:6:: 0-68 : 1-165 = silica united to lime. (4.) 4-5:6:: 0-28 : 0-373 = silica united to protoxide of iron. Now, 51-466 + 13-68 + 1-165 + 0-373 = 66-684, which is less than the silica in the mineral by 1-776 grain. It is not improbable that, if the loss, amounting to rather more than two per cent,, and which was more likely to fall upon the bases than the silica, had not taken place, this excess of silica would not have occurred in our calculation. But let us see what the atomic weight of alumina will be, supposing the analy- sis accurate. The silica united to the soda, lime, and oxide of iron is 2 -fr part of the whole ; and 2 -fr part of 1 '77^ = 0-655, which being taken from 1 -776 * Gilbert's Annalen, LXXIII. 18C. 9 ALUMINA. 305 leaves 1-121 to be added to the 51-466 given by theory. Now, 52-587 : 19*3 : : 6 : 2-202 = * io ' mic weight of alumina. The 12 preceding minerals constitute a suffi- cient number to enable us to approach pretty nearly to the atomic weight of alumina. That weight comes out somewhat differently from each mineral. This may be partly occasioned by small errors in the analyses; but it is owing chiefly to impurities existing in the specimens analyzed, the presence of which we have no means of determining. If a mineral contains mixed with it some silica, or a portion of any stone containing more silica than itself, the ato- mic weight of the alumina will come out too high. The contrary will be the case, when the mineral contains an excess of alumina, or of a stone containing more alumina than the speci- men now under examination. An accurate knowledge of the rock from which the specimen was taken, and of the minerals accompanying it, might throw some light upon the foreign matter most likely to be present. But this has not yet been sufficiently attended to ; meanwhile, it is probable, that the errors in the atomic weight of alumina, resulting from these foreign matters, will be on different sides in different analyses ; consequently, the mean atomic weight from all the twelve analyses given above will come near the truth. Let us therefore see what VOL. I. U 306 EARTHS PROPER. atomic weight of alumina will be derived from that mean. 1. Nepheline gives the atomic weight 2*3126 2. Karpholite .... 2-3013 3. Harmotome .... 2-2985 4. Meionite . . . . 2-3168 5. Scapolite .... 2-0580 6. Scolezite .... 2-2831 7. Bucholzite .... 2-252? 8. Leucite 2-2464 9. Thomsonite .... 2-2395 10. Natrolite 2-1 754 i 11. Felspar .... 2-2183 12. Albite , . 2.2020 12)26-9046 Atomic weight of alumina . 2-24205 The reader will observe that the greatest atomic weight of alumina is deduced from meion- ite, and amounts to 2-3168, which is very far short of 3'375, the atomic weight pitched upon by Mr. Phillips. The least weight is 2-0580, which is derived from scapolite. Upon the whole, the atomic weight of alu- mina derived from the constitution of these min- erals, though it would be hazardous to consider it as quite exact, cannot be very far from the truth ; and it differs from 2-25 by a very small 2. From fraction. II. .The aluminous salt most easily obtained ALUMINA. 307 in a state of purity, and in regular crystals fit for examination, is alum. Of this triple salt there are several species ; for it consists of a certain number of integrant particles of sulphate of alu- mina combined with sulphate of potash, or sul- phate of soda, or sulphate of ammonia, or sul- phate of magnesia, and perhaps of some other soluble sulphate which I have not tried. The crystals of alum consist of regular octahedrons, and neither the figure nor the taste of the salt is altered by substituting any of the above named sulphates for the others. The species of alum which I chose as upon the whole of easiest ana- lysis, and capable of being obtained with the greatest facility, was the common alum of this country, composed of sulphuric acid, alumina, potash, and water, or of sulphate of alumina and sulphate of potash, united together with a certain quantity of water of crystallization. As the analysis of this salt was attended with consi- derable difficulty, but was made with the utmost possible attention to precision, it will be proper to give a somewhat detailed account of it. 1. 6O875 grains of pure alum crystals were dissolved in hot water, and the liquid was mixed with a solution of 53 grains of chloride of barium. A double decomposition immediately took place, and sulphate of barytes precipitated to the bot- tom : when the supernatant liquid had become transparent and colourless, it was tested by sul- U2 308 EARTHS PROPER. phate of soda and muriate of barytes, but was not affected by either ; showing that it contained no sensible quantity of barytes or sulphuric acid. It follows from this experiment, that the sulphu- ric acid in 60-875 grains of alum is just saturated by the barytes from 53 grains of chloride of barium. But 53 is equal to 13-25 x 4 ; and 13-25 grains of chloride of barium require just 5 grains of sulphuric acid to saturate all the ba- rytes which they yield. It is obvious from this, that 60-875 grains of alum contain exactly 20 grains of sulphuric acid. 2. 60-875 grains of alum crystals were put into a platinum crucible, which was left for an hour on a hot sand bath. After it had been re- duced to the state of burnt alum by the loss of its water of crystallization, it was introduced into a wind furnace, and kept for an hour in a bright red heat. The loss of weight sustained was 43-62 grains. The alum thus treated, was digested in dis- tilled water, and then thrown upon the filter. The alumina left upon the filter was washed till the liquid which passed through ceased to precipitate muriate of barytes. The liquid which passed through the filter was concentrated on the sand bath, and then precipitated by muriate of barytes. The sulphate of barytes obtained, after being washed, dried, and heated to redness, weighed 13-28 grains, which is equivalent to ALUMINA. 309 4*504 grains of sulphuric acid ; if we add this to the original loss, we obtain 48*124 grains for the amount of the water and sulphuric acid in 60-875 grains of alum. Deducting the 20 grains of sulphuric acid which we know to be present, there remain 28*124 grains for the water of crystallization. But an integrant particle of water weighs 1*125, and 1*125 x 25 = 28*125. From this it is obvious, that 60*875 grains of alum contain 28*125 grains of water, which is equivalent to 25 atoms. 3. 60*875 grains of alum crystals were ex- posed in a platinum crucible to the heat of the sand bath, and afterwards to a dull red heat for an hour. Distilled water was poured upon the residual matter in the crucible in successive por- tions, and digested on it till it ceased to take up any thing 5 this water being evaporated to dry- ness, left 11 grains of sulphate of potash. Hence, 60*875 grains of alum contain 6 grains of pot- ash. We see, too, that one-fourth of the sulphu- ric acid is united to the potash : the remaining three-fourths must be in combination with the alumina. 4. To obtain the alumina, 60*875 grains of alum were dissolved in distilled water, and to the solution was added the quantity of carbonate of potash just necessary to saturate 15 grains of sulphuric acid. The experiment was repeated us 310 EARTHS PROPER. with carbonate of soda and carbonate of ammo- nia. Six experiments made in this way gave the mean quantity of alumina from 6O87-5 grains of alum, equal to 6'746 grains ; but the true quan- tity is obviously*6*7<5 grains : because this small addition is necessary to complete the weight of the alum. From the preceding detail it is evident that alum is composed of Sulphuric acid 20 Alumina . 6-75 Potash . 6 Water . 28-125 60-875 One integrant particle of the sulphuric acid is united to the potash, and three to the alumina, It is obvious from this, that 6*75 must represent Atomic the weight of three atoms alumina. Consequent- 2-25. " ly, the weight of alumina is 6 -r = 2-25. The composition of alum may be stated some- what differently. For it is obviously composed of 3 atoms sulphate of alumina = 21-7^ 1 atom sulphate of potash =11 25 atoms water . . = 28-125 CO-875 The analysis of alum just given was made by ALUMINA. 31 1 me during the course of the summer of 1821, and was repeated so frequently that I considered myself as quite certain of its accuracy. But after Mr. Phillips published his experiment on sulphate of alumina, and his suspicion that 3*375 was the true weight of an atom of alu- mina, I resolved to repeat the whole again. This I accordingly did during the summer of 1823. The result which I obtained may be con- sidered as absolutely identical, for the only dif- ferences took place in the second decimal place ; and the ratios of the constituents of alum de- duced from this second analysis, did not de- viate Toooth part from those of the preceding analysis. If the weight of alumina were 3*375 instead of 2*25, we should obtain from 60-875 grains of alum no less than 10*125 grains of alumina* in- stead of 6*75, which I obtained ; so that I must have lost more than 30 per cent, of the alumina j which, from the care with which the experiment was made, 1 consider as impossible. But my analysis of alum is corroborated and confirmed by the analysis of this salt made pre- viously by Berzelius. When we correct his statements by means of the more accurate data given in this work, his analysis of alum is as follows :* * Ann. de Clilm. LXXXII. 258. U 4 312 EARTHS PROPER. Sulphuric acid 33-85 or 20-6060 Alumina . 10-86 - 6-5958 Potash . 9'90 - 6-0266 Water . 45-00 - 27'3940 99-61 60-6224 Loss . . . . 0-2526 60-8750 According to this analysis, 3 atoms of alumina weigh 6 '5 958. Consequently, the weight of the atom of alumina is 2*1986, which is a little less than the result of my analysis of alum. In a subsequent chapter of this work, when I come to treat of the salts of alumina, it will be seen that the atomic weight of alumina just given, is corroborated by many other aluminous salts which have been subjected to examina- tion. Mr. Phil- Mr. Phillips' opinion, that the atomic weight of lips' experi- ... ment re- alumina is 3*375, seems, as far as I can judge, u! to be founded on the following experiment. He found that when moderately diluted, sulphuric acid was digested for a long time upon an ex- cess of recently precipitated alumina, till it re- fused to take up any more ; the liquid, when filtered, had a tendency to precipitate a peculiar sulphate of alumina in white flocks ; the liquid, when dropped into water, rendered it milky, and a white sediment was deposited. He analyzed ALUMINA. 313 this sulphate of alumina, and found it composed of Sulphuric acid 40 Alumina . 40*83* This he considers as inconsistent with 2*25 being the atomic weight of alumina. But if 3'375 be the atomic weight of that earth, he conceives that his salt was a compound of 2 atoms sulphuric acid 10 or 40 3 atoms alumina . 10-125 - 40-5 20-125 But this is not the only view of the subject which may be taken. I may observe, in the first place, that alumina Explained, is a much weaker base than any of the other earths or alkaline bodies treated of in the pre- ceding chapter. Now, it is a remarkable cir- cumstance, which has been particularly illustrated by Berzelius, that weak bases have the property of combining in a much greater number of pro- portions with acids than strong bases. I have ascertained by experiment that alumina and sulphuric acid may be united in three propor- tions j namely, 1 atom acid + 1 atom alumina 1 atom acid + 2 atoms alumina 1 atom acid -f. 3 atoms alumina * Annals of Philosophy (second series) IV. 280. 314 EARTHS PROPER. The first of these salts constitutes common sul- phate of alumina, and it is an essential consti- tuent of alum. The third of these salts is the mineral found in Sussex and other places, and known by the . name of aluminite. For Stro- meyer has shown that it is a compound of 1 atom sulphuric acid, 3 atoms alumina, and 9 atoms water. If potash be poured into a boiling hot solution of alum, an earthy precipitate falls, precisely similar in its nature with the substance which chemists formerly called alum saturated with its earth. This precipitate is a compound of 1 integrant particle of sulphate of potash, and 3 integrant particles of trisulphate of alumina, as was first proved by Riffault.* I find that the second of these salts, to which I give the name of disulphate of alumina, may be obtained by a process similar to the one followed by Mr. Phil- lips ; but not carried so far. It is soluble in water, and still possesses an astringent and sweetish sour taste. Mr. Phillips' salt was a mixture or compound of disulphate of alumina and trisulphate of alu- mina, in the proportion of 3 atoms of the former salt to 1 atom of the latter. It was the trisul- phate which precipitated when the solution was poured into water ; and it was its tendency to fall down that explains the phenomena which * Ann. de Chim. et de Phys. XVI. 355. ALUMINA. 315 Mr. Phillips considers so unaccountable. I need hardly observe that SULPHURIC ACID. ALUMINA. 3 atoms disulphate contain 15 + 13-5 1 atom trisulphate . 5 + 6'75 Total . . . 20 + 20-25 Now, this is equivalent to Sulphuric acid 40 Alumina 40-5 the very theoretic proportions which Mr. Phillips deduces from an erroneous atomic weight of alumina. III. When alumina is precipitated from an 3. From alkaline solution, collected on a filter, well washed, of alumina. and allowed to dry spontaneously on the filter in a dry warm room, a hydrate is obtained, com- posed of equal weights of alumina and water. This is an experiment which I have frequently made in Edinburgh and in London, where I was in the habit of working in a room elevated con- siderably above the ground ; but in Glasgow, where my laboratory is in a ground floor situated in a clay soil, and consequently rather damp, I have found it much more difficult to obtain cor- rect results, at least during the summer of 1823, which was uncommonly cold and wet, when I was occupied with this investigation. I shall re- late one experiment which I made with consi- 316 EARTHS PROPER. derable care, it will show how near the truth I was able to come. Bihydrate. L A quantity of alumina left upon the filter in a temperature not exceeding 60, for a couple of months, was* found to weigh 36 '66 grains. Being strongly ignited in a platinum crucible, the weight was reduced to 17*85 grains. The loss of weight, therefore, was 18*81 grains, which gives the amount of the water in the hy- drate. Now, 17-85 : 18-81 : : 2-25 : 2-42624, which exceeds 2 atoms of water by 0-17624 ; and this excess of water I find it difficult to get rid of in a damp room. If we were to consider the water in this hydrate as amounting to exactly 2 atoms, and to calculate the atomic weight of alumina on that supposition, we have 18-81 : , 17-85 : : 2-25 : 2-1352 = atomic weight of alu- mina. But the analysis of alum, given in a pre- ceding part of this section, shows that there is really an excess of water. Besides, I have fre- quently obtained a hydrate of alumina composed exactly of equal parts of alumina and water, showing that the atomic weight of alumina is exactly 2-25, and that this hydrate contains 2 atoms water. Hydrate. 2. When the bihydrate of alumina is dried in a temperature of about 100, it is reduced to the state of a hydrate at least very nearly, for I have not always been able to obtain exactly the atomic proportions, though the deviation was ALUMINA. 317 small. I shall state one or two of the experi- ments numerically, that the reader may judge of the approximation. (i.) 18*92 grains of alumina, dried on the fil- ter in a temperature of 100, being exposed to a strong red heat, were reduced to 13*49 grains ; consequently, the water driven off amounted to 5 '43 grains. Now, 13-49 : 5-43 : : 2-25 : 0*9055 = ratio of water. (2.) 20 grains of alumina, which had been left upon a filter for three months, in a temperature which, during the day, was usually 85, though during the night it was lower, as the fire was al- lowed to go out, being exposed to a strong red heat, was reduced to 13'45 grains ; and, conse- quently, it had lost 6*55 grains. Now, 13-45 : 6-55 : : 2-25 : 1-099 = ratio of water. It appears from this experiment, that by a great length of time the alumina may be brought into the state of a hydrate, consisting of 1 atom alu- mina and 1 atom water ; for 1-099 is only a very little less than 1*125, the weight of an atom of water. Such are the results which I have obtained respecting the atomic weight of alumina. They all conspire to show that it is very nearly 2*25, and that it cannot by any means amount to a 318 EARTHS PROPER. number nearly so high as 3 '375. Indeed, from the constitution of alumina, there seems no rea- son to doubt that %'%5 is the true atomic weight of this earth. The probability is, that alumina is a protoxide of aluminum. On that supposition, the atomic weight of aluminum will be 1*25. SECT. II. OF THE ATOMIC WEIGHT OF GLUCINA. THOUGH glucina has been known to chemists for more than twenty years, we are in possession of very few experiments from which we can deduce its atomic weight. Berzelius, so far as my know- ledge extends, is the only person who has ana- lyzed some of the salts of glucina with a view of determining its capacity of saturation, from which the atomic weight is easily deduced. Before re- lating the experiments which I have made to ascertain the atomic weight of this substance, it will be interesting to state the experiments of Berzelius, by which, in fact, my own were di- rected. * Berzelius dissolved a quantity of glucina in a by a n excess of sulphuric acid evaporated the 13erzelius. * solution till the acid began to escape and then washed it in alcohol to cret rid of the excess of GLUCINA. 319 sulphuric acid. The sulphate of glucina thus obtained was dissolved in water. The solution was mixed with carbonate of ammonia to preci- pitate the glucina, and heat was applied till the ammonia ceased to be disengaged. The glucina thus obtained, being well washed and heated to redness, weighed 0*553 parts : the liquid thus freed from glucina was precipitated by muriate of barytes. The sulphate of barytes, after being washed and heated to redness, weighed 5, which is equivalent to 1*6949 of sulphuric acid.* Thus it appears, that the sulphate of glucina, formed by the above process, is composed of Glucina 0-553 or 3-262? Sulphuric acid 1-6949 - 10 Considering this salt as a bisulphate, Berzelius dissolved a portion of it in water and digested it over a quantity of carbonate of glucina free from ammonia. By this digestion he obtained a gum- my looking liquid. Part of this being mixed with water, a white matter precipitated, (doubt- less glucina) and the liquid being filtered, was decomposed by carbonate of ammonia and mu- riate of barytes precisely as the former salt. The glucina obtained weighed 1*001, and the sulphate of barytes 4*549, equivalent to 1-5573 sulphuric acid. Thus it appears, that the new * Berzelius' Attempt, e, p. 134. 16 320 EARTHS PROPER. gummy like sulphate of glucina was composed of Glucina 1-001 or 32139 Sulphuric acid 1-5573 - 5 A portion of the gummy liquid, obtained as above described, was heated over a spirit lamp till it became dry, when it resembled burned alum. After ceasing to give out water from the heat of a spirit lamp, Berzelius exposed 2-5 parts of it to a red heat; the glucina left weighed 1*24 parts. Berzelius, considering the salt as perfectly dry, concludes it was composed of Glucina 1-24 or 4-92 Sulphuric acid 1*26 - 5 2-50 It is obvious that the second of these three salts is neutral, being composed of one atom of acid and one atom of glucina. The first is a bisul- phate, or a compound of two atoms acid and one atom glucina, while the third is a sulphate com- posed of one atom acid, and H atom glucina. These three analyses, however, are not abso- lutely correct; for the atomic weight of glu- cina differs a little in each. By the first it is 3.2627 second 3-2139 third 3-2800 3)9-7566 Mean = 3-2522 GLUCINA. 321 This mean of 3-2522 must be exceedingly near the truth ; as near, probably, as it is possible to come with our present analytical means by a direct analysis. I consider myself entitled to conclude from it, that the true atomic weight of glucina is 3*25. I leave out the two last deci- mal places, because they would destroy the law pointed out by Dr. Prout, which we have seen to hold in all the other atomic weights of bodies; namely, that these weights are all multiples of 0*125, the atomic weight of hy- drogen. 2. Berzelius likewise analyzed the muriate of glucina ; and, although the muriates are not so easily analyzed with accuracy as the sulphates, it will be well worth our while to state the re- sult of the analysis as constituting an exceedingly good approximation. The bimuriate of glucina was a crystallizable salt : but when the excess of acid was expelled, he obtained only a gummy like mass. This matter was dissolved in water, the solution filtered, and the glucina was thrown down by carbonate of ammonia, and the muria- tic acid by nitrate of silver. The glucina weighed 0-626, and the chloride of silver 3-392, which is equivalent to 0-85962 of muriatic acid. Accord- ing to this analysis, muriate of glucina is com- posed of VOL. L X EARTHS PROPER. Glucina 0-266 or 3-3681 Muriatic acid 0-85962 - 4-625* The atomic weight of glucina deduced from this analysis is 3*3681, which is a little higher than that from the sulphate. But the preponderancy of evidence is entirely in favour of 3-25, the weight deduced from the sulphates of glucina. Composi, 3^ T^Q emerald or bend is the mineral from tion of emerald, which glucina was originally obtained, and in- deed, we are scarcely in possession of any other quarry from which this earth can even at present be procured. It is usually crystallized in six- sided prisms, and has all the marks of being a chemical compound. Its constituents are silica, alumina, and glucina, with mere traces of some metallic oxides, which seem to act the part of colouring matters. It will be worth our while to see what atomic weight will be deduced for glucina from the constituents of this mineral. Vauquelin and Klaproth published good ana- lyses of emerald and beryl ; but it will be better to employ for our calculations the analysis by Berzelius, which is much later, and which, from the improvements introduced into the art of analyzing minerals, is probably nearest the truth. According to Berzelius, the constituents of this mineral are as follows : t * Berzelius' Attempt, &c. p. 134. f Afhandlingar, IV. 192. GLUCTNA. 323 Silica . 68-35 Alumina . 17*60 Glucina 13-13 Oxide of iron 0-72 Oxide of tantalum 0-27 100-07 Let us consider the emerald as a compound of two atoms of trisilicate of alumina, and one atom of trisilicate of glucina, and let us see what the quantity of silica will be, supposing the atomic weights of alumina and glucina to be 2-25 and 3-25 respectively. (1.) 2-25 (atom of alumina) : 6 (3 atoms silica) : : 17-6 (alumina in mineral) : 46-933 = silica united to alu- mina. (2.) 3-25 (atom of glucina) : 6 : : 13-13 : 22-422 = silica united to glucina. Now, 46-933 4- 22-422 = 69'355, which ex- ceeds the silica in the mineral by 1 -005 grain. The silica united to the alumina is very nearly two- thirds of the whole ; consequently, two-thirds of this surplus belongs to that portion of the silica. There remain 0-335 grain, which must be deducted from the 22-422. The remainder, 22-087, is the silica actually united to the glucina in the mineral. Now, 22-087 : 13-13 : : 6 : 3-5668 = atom of glucina. 2. EUCLASE. Composi . There is another mineral, containing glucina, X2 324 EARTHS PROPER, namely euclase ; the constituents of which, ac- cording to the analysis of Berzelius, are as fol- lows : Silica . 43-32 Alumina . 30-56 Glucina . 21-78 Oxide of iron 2-22 Oxide of tin 0-?0 98-58 Let us calculate the constitution of this min- eral, on the supposition that it is composed of two atoms silicate of alumina, one atom of sili- cate of glucina, and a small quantity of bisilicate of iron, probably accidentally present. (1.) 2-25 : 2 : : 3O56 : 27-164 = silica united to alumina. (2.) 3-25 : 2 : : 21-78 : 13-403 = silica united to glucina. (3.) 4-5 : 4 : : 2-22 : 1'973 = silica united to iron. Now, 27-164 + 13-403 + 1-973 =42-540, which is less than the silica in the mineral by 0-98 grain. Two thirds of the silica are united to the alumi- na; consequently, only one third of 0'98 or 0-33 is to be added to the 13-403, making 13*733 grains for the silica united to the glucina. Arid 13-733 : 21-78 :: 2 : 3-1719 = atomic weight of glucina. From these two minerals we obtain two dif- ferent atomic weights of glucina, the first too high, and the second too low. The probability GLUCINA. 32,5 is, that the mean of the two will give an atomic weight not very far removed from the truth. Atomic weight from beryl 3-5668 euclase 3-1719 2)6-738? Mean . . = 3-36935 This result, though too high, is much nearer the truth than that derived from the analysis of beryl alone. 4. I shall now relate the experiments by which Analysis of I satisfied myself that 3'&5 is the exact atomic of giucina . , . f . ! by the weight or glucina. author. (1.) I prepared a quantity of bisulphate of glucina precisely in the same way as had been previously done by Berzelius : 18*875 grains of this salt were put into a platinum crucible, and exposed for some time to a strong red heat : there remained 3 f &5 grains of pure glucina. (2.) 18-875 grains of the same salt, and 26\5 grains of chloride of barium, were dissolved in separate portions of distitled water. These two liquids being mixed, a double decomposition took place sulphate of barytes falling to the bottom, while muriate of glucina remained in solution. This liquid, as soon as it had become transparent, was tested by sulphate of soda and muriate of barytes ; but was not affected by either of these reagents, showing that it contained no sensible x 3 326 EARTHS PROPER. quantity either of barytes or sulphuric acid. From this experiment, we see that the quantity of sulphuric acid in 18*875 grains of bisulphate of glucina just saturates the barytes from 26 '5 grains of chloricte of barium ; and consequently, amounts to exactly 10 grains. (3.) Thus it appears, that 18*875 grains of bi- sulphate of glucina contain 3*25 grains of glu- cina and 10 grains of sulphuric acid ; the rest must be water : consequently, the constituents of the salt are Sulphuric acid Glucina Water 18-875 Now, 10 is equivalent to 2 atoms of sulphuric acid, and 5*625 to 5 atoms of water. And corn- Atom of paring this analysis with those of Berzelius, no doubt can remain that 3*25 represents the weight of an atom of glucina. Atom of We have no data to determine the atomic weight of glucinum. But analogy leads to the supposition, that glucina is a compound of 1 atom glucinum and 1 atom oxygen on that supposi- tion an atom of glucinum will weigh YTTRIA. 327 SECT. III. OF THE ATOMIC WEIGHT OF YTTRIA. THOUGH yttria has been known for almost twenty years, the minerals in which it occurs have been hitherto met with only in Sweden ; and even in that country, they are so scarce as hardly to be procured, except in pieces too small to admit of the extraction of a quantity of yttria sufficiently large to enable us to make experi- ments on the composition of the salts which it forms. Thus circumstanced, I have been ne- cessarily confined to experiments made upon a very few grains of this earth ; and though they were conducted with the most scrupulous atten- tion to accuracy, I was not able to draw any satisfactory conclusions from them. I am, therefore, under the necessity of deducing the atomic weight of this earth from the experiments of Berzelius. Fortunately, this illustrious che- mist is so remarkably accurate in his experiments, that they are sure to furnish an exceedingly good approximation. 1. Berzelius put 100 parts of dry carbonate of Analysis of carbonate yttria into a small retort, a glass tube connected of yttria. with the beak of which, was filled with dry mu- 328 EARTHS PROPER. riate of lime. Heat was applied and continued till the retort began to soften. By this process the water was driven off and united to the muriate of lime, the weight of which it had increased by 12-82 parts. The yttria, thus freed from water, was heated to redness in a platinum crucible, in order to drive off the carbonic acid : by this process its weight was reduced to 57*7 parts. From this analysis, it appears that carbonate of yttria is composed of Carbonic acid 29-48 or 2-?5 Yttria . 57-?0 - 5-3825 Water . 12-82 - 1-1959 100-00* This analysis gives us 5'3825 for the atomic weight of yttria. But the water in the salt, which obviously ought to amount to 1 atom, is a little too high. The probability is, that the wa- ter absorbed by the muriate of lime would retain a little carbonic acid. If we divide the 42'3 parts which make up the carbonic acid and wa- ter in the ratio of 2*75 : 1-125 which must have been the true proportion of those bodies, on the supposition that the salt was a chemical com- pound, we shall have the constituents of this carbonate as follows : * Annals of Philosophy, III 359. YTTHIA. Carbonic acid 30-019 or 2-75 Yttria . 5?'700 - 5-28^ Water . 12-281 - 1-125 100-000 This brings the atomic weight of yttria to 5 '285, a number which must come pretty near the truth. 2. Berzelius dissolved 2*8 parts of pure yttria OF sulphate . ' . of yttria. in sulphuric acid evaporated the solution to dryness and heated the residual sulphate till it ceased to give out acid vapours. It now weighed 5'392 parts. From this experiment he con- cluded, that anhydrous sulphate of yttria is com- posed of Sulphuric acid 2-592 or 5 Yttria 2-8 - 5-4012 5-392* By this experiment the atomic weight of yttria comes out 5*4012. But we have no satisfactory evidence that the salt was neutral, except that it was still soluble in water. I am disposed to think that a very small portion of the acid had been driven off: for if we suppose the constitu- ents of the salt to have been Sulphuric acid 2-685 Yttria . . 2-8 5-485 * Annals of Philosophy, 111. 359. 330 EARTHS PROPER. that is to say, had the acid been 0'093 more, then the atomic weight of yttria would have been of hydrate 3. Klaproth found that yttria precipitated from muriatic acid, and dried in the open air after being well washed, was a hydrate composed of Yttria ... 69 Water . . . 31 100 For 100 parts of it when heated to redness lost 31 parts of its weight. Now, 31 : 69 : : 2*25 : 5*008 = atomic weight of yttria, supposing the hydrate to contain just 2 atoms water. Thus we have got three atomic weights of yttria. (1.) From the carbonate . . 5-285 (2.) From the sulphate . 5-4012 (3.) From the hydrate . . 5-008 3 ) 15-6942 Mean .... 5-2314 This mean must be very near the truth : we may, therefore, adopt 5*%5 as the true atomic weight of yttria in order to make that weight accord with Dr. Prout's law. Analysis of 4. Gadolinite, the mineral in which yttria was gadolinite. first discovered, has been likewise analyzed with YTTRIA, 331 great care by Berzelius, who found its constitu- ents as follows : Silica . . 25-80 Yttria . . 45-00 Protoxide of cerium 16-69 Protoxide of iron 10-26 Volatile matter 0-6t) 98-35 * Let us calculate the composition of this mineral, on the supposition that the yttria and oxides of cerium and iron are respectively in the state of silicates, and that the atomic weights of all of these constituents are as follows : Silica . . 2 Yttria . . 5-25 Protoxide of cerium 7*25 Protoxide of iron 4-5 (1.) 525 (atom of yttria) : 2 (atom of silica) : : 45 (yttria in the mineral) : 17-14 = silica united to yttria. (2.) 7-25 (atom protoxide of cerium) : 2 : : 16-69 : 4-302 =s silica united to the oxide of cerium. (3.) 4-5 (atom of oxide of iron) : 2 : : 10-26 : 4-56 = silica united to the oxide of iron. Now, 17-14 + 4-302 + 4-56 = 26-002, which ex- ceeds the silica found in the mineral by 0*2002 grain. The silica united to the oxides of cerium and iron being one third of the whole, it is ob- ' Afhandlingar, IV. 225. rite. 332 EARTHS PROPER. vious, that one third of this surplus, or O0667 belongs to that portion ; we have, therefore, 0*1335 grain to subtract from the 17*14, to obtain the quantity of silica in the mineral, ac- tually united -to the yttria. It amounts to 17-0065. And 17*0065 : 45 : : 2 : 5-2Q21 = atomic weight of yttria. Ofyttroce. 5. There is another mineral, which was dis- meim covered by Berzelius at Finbo in Sweden, and analyzed with great care, namely, the yttrocerite, which he found composed as follows : Lime . . 50 Yttria . . 8-1 Protoxide of cerium 16-45 Fluoric acid 25-45 100-00* Let us suppose the lime and yttria in this min- eral to be in the state of neutral fluates, and pro- toxide of cerium in the state of a bifluate, and let us calculate the composition, on the supposi- tion that the atom of fluoric acid is 1 *25, which will be proved in a subsequent chapter. (1.) 3-5 (atom of lime) : 1-25 (atom of fluoric acid) : : 50 (lime in the mineral) : 17'857 = fluoric acid united to lime. (2.) 5-25 (atom of yttria) : 1-25 : : 8-1 : 1-928 = fluoric acid united to yttria. (3.) 7-25 (atom of oxideof cerium) : 2-5 (2 atoms fluoric acid) : : 16-45 : 5-672 = fluoric acid united to oxide of cerium. * Afhandlingar, IV. 151. YTTRIA, 333 Now, 17-857 + 1-928 + 5-672 = 25-457, which exceeds the fluoric acid found in the mineral by 0-007 grain. But rj& part of the fluoric acid are united to the lime and the cerium, so that only rlhy of 0-007 or 0-00056 grain require to be sub- tracted from 1-928, to have the fluoric acid ac- tually united to the yttria, according to the analysis. This amounts to 1*9224. And 1-9224 : 8-1 : : 1-25 : 5-2668 = an atom of yttria. 6. There is still another mineral, namely, the yttrotantalite, which appears to consist essen- tially of oxide of tantalum and yttria united to- gether. But the other constituents vary so much, that no greatreliance can be put on the analyses hitherto given, the specimens appearing all ex- ceedingly impure. There are three varieties, which differ a good deal from each other in the proportions of their constituents. The first variety has a black colour, and was found by Bcrzelius composed of Oxide of tantalum 57 Tungstic acid 8-25 Yttria 20-25 Lime 6-25 Oxide of iron 3-50 Oxide of uranium 0-50 95-75 * *AfhandlIugar. IV. 907. 834 EARTHS PROPER. The second variety, which is yellow, was com- posed of Oxide of tantalum Tungstic acid Lime Oxide of iron Oxide of uranium 60-124 1-044 29-780 0-500 1-155 6-622 99-225 The third variety, which is dark brown, was composed of Oxide of tantalum Tungstic acid Yttria Lime . Oxide of iron Oxide of uranium 51-815 2-592 38-515 S-260 0-555 1-111 97-848 It must be obvious from a bare inspection of these three analyses, that the ratios between the oxide of tantalum and the yttria differ too much in each, to enable us to deduce from them any information, either respecting the atomic weights of yttria or oxide of tantalum ; not to men- tion the tungstic acid, the lime and the oxide of uranium, which are still more variable. But a comparison of the atomic weight of yttria, de- duced from the analyses of gadolinite and yttro- cerite, with the weights deduced from the car- ZTRCONJA. 335 bonate, sulphate and hydrate of yttria, leave Atom of little doubt that the true atomic weight of this yt earth is 5-25. Analogy is in favour of the notion, that yttria Of yttrium. is a compound of one atom of yttrium and one atom of oxygen ; on that supposition, the atomic weight of yttrium will be 4 '25. SECT. IV. OF THE ATOMIC WEIGHT OF ZIRCONIA. WE have no good data for calculating the atomic weight of zirconia. The best at present in our possession, are derived from the composition of the zircan and hyacinth, which have been repeat- edly analyzed by Klaproth and Vauquelin. The constituents of this mineral, as deduced from their analyses, are as follows : Zirconia 69-0 65 64-5 70 64-5 66 Analyses of Silica 26-5 33 32'5 25 32 31 Oxide of iron 0-5 1 1-5 0-5 2 2 96-0* 99 1 98-5 J 95-5 98-5 1| The great difference in these results renders it probable that none of the analyses is quite cor- * Klaproth Bectrage, I. 222. f Ibid. III. 271. From Norway. \ Ibid. V. 130. From the Circars. Ibid. I. 231. From Ceylon. (I Vauquelin, Jour, de Min. No. XXVI. 106. f Vauquelin, Ibid. 336 EARTHS PROPER. rect. Indeed, neither Klaproth nor Vauquelin succeeded in freeing the zirconia completely from iron. But if we deduce the atomic weight of zirconia from these analyses, supposing the minerals to be simple silicates of zirconia mixed with a little silicate of iron, then the atomic weight of zirconia will be as follows : From first analysis 5-247 From second 3 -9 8 7 From third 4-043 From fourth 5-645 , From fifth 4-134 From sixth 4'370 6)27-426 Mean - 4-537 A careful examination of a collection of hya- cinth crystals, which I was favoured with some years ago by Mr. Heuland, serves, in some measure, to explain the reason of these anomalies. Besides crystals of zircon, I found among them likewise crystals of spinel!, garnet, quartz, and some green fragments having the colour of emerald, besides some black grains, apparently octohedral iron ore. It was doubtless owing to the presence of some of these foreign crystals that the alumina, which I observed in a former analysis of the hyacinth, is to be ascribed. Pure and regular crystals of hyacinth seem composed of 16 ZIRCONIA. 337 Zirconia . . . 75 Silica ... 25 Oxide of iron , . 0'5 100-5 If we suppose the iron foreign, and the mineral a simple silicate, then the atomic weight of zirconia would be 6, which, from experiments immediately to be mentioned, I believe to be its true weight. There is another mineral containing zirconia, C which was discovered by Sir Charles Giesecke, . . . precipitated by ammonia, and the earth, after being well washed, was left for several weeks upon the filter in a temperature, which, during the day, was very nearly 85 ; though it was a good deal lower during the night. The hydrate thus obtained had a greyish yellow colour, a certain degree of semitransparency, and broke with a fracture resembling that of glue. It was exposed to a strong red heat for nearly an hour, in order to drive off the water. I found that this hydrate was a compound of Zirconia ... 6 Water . . . 8-885 Now, 8*885 differs very little from 9, which re- presents the weight of 8 atoms of water. This leads to the inference, that the atomic weight of zirconia is 6, and that the hydrate was a com- pound of 1 atom zirconia and 8 atoms water. (2.) Another quantity of zirconia, precipitated from the muriate, was well edulcorated and dried on the filter by means of steam, at the tempera- ture of 212. The hydrate thus obtained was decomposed in a strong red heat, and was found to be composed of ZIRCONIA. 34 1 Zirconia ... 6 Water . . . 1-125 Now, 1*125 represents an atom of water. Hence, the hydrate was a compound of 1 atom zirconia and 1 atom water ; and this experiment also leads to the conclusion that an atom of zirconia weighs 6. (3.) 11 '85 grains of pure muriate of zirconia, Muriate of not in crystals, but formed by dissolving zirconia in muriatic acid, and evaporating the solution to dryness in a low temperature, were dissolved in water, and the zirconia thrown down by am- monia. When properly washed and dried, and heated to redness, it weighed 4*33 1 grains. The liquid thus freed from zirconia, was' neutralized by nitric acid, and the muriatic acid thrown down by nitrate of silver. The chloride of silver, after fusion, weighed 13-59 grains, equivalent to 3*444 grains of muriatic acid. From this experiment it follows, that muriate of zirconia is composed of Muriatic acid 3-444 or 4-625 Zirconia . 4-331 - 5-815 Water . 4-075 - 5-437 4*625 denotes the weight of an atom of muri- atic acid 5*437 approaches pretty near to 5*625, the weight of 5 atoms of water 5*815 repre- sents the atomic weight of zirconia, according to this experiment : it is a little less than 6, the atomic weight deduced from the hydrates. I Y3 342 EARTHS PROPER. am disposed to ascribe this to a slight excess of muriatic acid in the salt. My quantity of zir- conia was not sufficient to give me muriate of zirconia in crystals. These experiments are not quite decisive ; but they lead to the conclusion, that the atomic weight of zirconia is 6. CHAP. X. OF THE ATOMIC WEIGHTS OF IRON, NICKEL, COBALT, MANGANESE, AND CERIUM. THESE metals are very nearly related to each other, both in their atomic weights and in the proportions of oxygen with which they combine. Each of them forms at least two oxides, the first containing an atom of oxygen, and the second an atom and a half of oxygen united to an atom of the metal; and both of these oxides, at least in the case of iron and manganese, is capable of combining with acids. One of these metals, I mean manganese, unites with two, if not three, additional dozes of oxygen, forming an oxide neither possessing alkaline nor acid properties, and one or two oxides capable of combining with bases, and consequently, pos- sessing the characters of acids. SECT. I. OF THE ATOMIC WEIGHT OF IRON. Analysis of PROTOSULPHATE of iron, or green vitriol, is aprotosui. phate of salt which has been long known, and manutac- iron. Y4 344 IRON, NICKEL, COBALT, &c. tured for the use of dyers and other artists. It crystallizes in four-sided oblique prisms with a rhombic base, is transparent, has a light green colour, and a sweetish astringent taste. It may be obtained perfectly pure from the copperas manufactories ; or by dissolving the salt of com- merce in boiling water, filtering the solution while hot, and allowing it to crystallize by cool- ing. I employed this salt to determine the atomic weight of protoxide of iron, and I shall now state the method by which I proceeded. 1. 17'375 grains of pure dry crystals of proto- sulphate of iron were dissolved in water, and mixed with a solution of 13*25 grains of chloride of barium. A double decomposition immediate- ly took place, and sulphate of barytes precipi- tated to the bottom of the vessel. When the liquid had become clear after the subsidence of this sulphate, it was tested by means of sulphate of soda, and muriate of barytes ; but was not affected by either of these reagents, showing that it neither contained barytes nor sulphuric acid. From this experiment it is evident, that the barytes from 13*25 grains of chloride of barium, just saturates the sulphuric acid in 17*375 grains of crystallized protosulphate of iron. Conse- quently, 17*375 grains of protosulphate of iron contain exactly 5 grains of sulphuric acid. 2. 17*375 grains of crystallized protosulphate of iron were exposed for some time in a plati- IRON. 345 num crucible to a heat of between 500 and 600. The loss of weight was 7*6 grains. Now, as 1 7*375 grains of this salt contain just 5 grains of sulphuric acid, which is equiva- lent to the weight of an atom, the water of crystallization ought likewise to be a determinate number of atoms. Now, 7*875 grains would be equivalent to 7 atoms. And the water actually found was 7*6 grains, which is only 0-275 grains, or ^th of an atom less than 7 atoms. But all the water cannot be separated from this salt by heat, without at the same time de- composing and separating a portion of the sul- phuric acid. Sulphurous acid begins to be disengaged, and sulphuric acid in a liquid state (and therefore containing water) trickles down the sides of the retort. It is easy to show that the water which re- mains behind, and which comes over united to the sulphuric acid, is exactly ith of an atom. For the sake of explanation, I shall take for granted what will be proved in the subsequent part of this section ; namely, that an atom of iron weighs 3 '5, an atom of protoxide of iron 4'5, and an atom of peroxide of iron 5. When protosulphate of iron, freed from as much water of crystallization as possible, is strongly heated, the protoxide of iron which it contains is converted into peroxide at the ex- pense of the sulphuric acid, which gives off one 346 IRON, NICKEL, COBALT, &c. atom of oxygen to the iron, and then flies off in the state of sulphurous acid. Now, 1 atom of oxygen will convert 2 atoms of protoxide of iron into peroxide. It is obvious from this, that one half of the sulphuric acid is decomposed, and that the other passes over in the state of fuming sulphuric acid. Now, this fuming acid I have carefully analyzed, and I find it a com- pound of 2 atoms sulphuric acid = 10 1 atom water . = M25 11-125 Let us suppose that a quantity of protosul- phate of iron, equivalent to 4 integrant particles, or 69*5 grains (17*375 x 4) of the crystals, are treated in this way. By exposing them to a heat of about 600, they lose 3O375 grains, or the whole of the water, except one atom. There remain 39*125 grains, composed of 4f atoms sulphuric acid = 20 4 atoms protoxide of iron = 1 8 1 atom of water = 1-125 39-125 When a strong heat is applied to this residuum, half the sulphuric acid, or 10 grains, is decom- posed into 2 atoms of sulphurous acid 8 2 atoms of oxygen = 2 10 IRON. 347 The 8 grains of sulphurous acid fly off in the state of gas, and the 2 grains of oxygen unite with the 18 grains of protoxide of iron, and con- vert them into 20 grains of peroxide. The re- maining J grains of sulphuric acid, united to the atom of water, are driven off in the state of fum- ing sulphuric acid, and weighing 11 '125 grains. From this statement it is obvious, that when 17*375 grains of protosulphate of iron are thus treated, 7*59375 grains of water (or 7 atoms th atom) will be driven off; 2| grains of sul- phuric acid will be converted into 2 grains of sulphurous acid (which will fly off) and half a grain of oxygen, which will convert the 44 grains of protoxide into 5 grains of peroxide. The remaining 2 grains of sulphuric acid united with 0*28125 grains of water, (the ith atom still remaining) will be driven off in the state of fuming sulphuric acid. From the preceding exposition it is plain, that after all the water has been driven off, which can be made to leave 17*375 grains of protosul- phate of iron by a moderate heat, one fourth of an atom, or 0*28125 grain still remains. Hence, the whole water of crystallization amounts to 7*6 + 0*28125 = 7-88125 grains. This comes so near 7*875 grains, which is the equivalent of 7 atoms, that there can be no doubt that 7 atoms is the true quantity of water present. The small excess (only 0*00625 grain) was owing to a com- 16 348 IRON, NICKEL, COBALT, &c. mencement of decomposition, for the whole of the salt thus treated was not soluble in water. Thus it appears that 17*375 grains of protosul- phate of iron contain 1 atom sulphuric acid = 5 # 7 atoms water . =7-375 12-875 Atom of The 4%5 grains, wanting to make up the whole oflroni 17*375, are obviously protoxide of iron ; and, as the salt is neutral, 4*5 must denote the atomic weight of protoxide of iron. 3. The preceding analysis of protosulphate of iron is rigidly exact ; but it proceeds upon as- sumptions which require to be proved. It will now be requisite to state the experiments by which I satisfied myself that 3%5 is the true ato- mic weight of iron. Atomic 35 grains of pure iron were put into a small iToJfalter- retort, with dilute sulphuric acid. A complete solution took place ; and the volume of the hy- drogen gas evolved (reduced to the mean tem- perature and pressure) was 59 cubic inches. Consequently, the iron must have combined with a quantity of oxygen, which, had it been in the gaseous state, would have amounted to -^, or 29*5 cubic inches, which weigh almost exactly 10 grains. From this it is obvious, that when 3'5 grains of pure iron are dissolved in dilute sul- IRON. 349 phuric acid, they combine with 1 grain of oxy- gen. Now, by this solution, the iron is convert- ed into protoxide ; consequently, protoxide of iron is composed of Iron . 3-5 Oxygen . 1 4-5 1 being the atomic weight of an atom of oxy- gen, 3'5 must be the weight of an atom of iron ; and the atomic weight of protoxide of iron is obviously 4%5. 4. Iron and sulphur unite together in two Sulphuret proportions, forming the protosulphuret and per- f n ana - sulphuret of iron. The former is called by mi- neralogists magnetic pyrites, because, when in powder, it is attracted by the magnet. The lat- ter is called cubic pyrites, because it usually crystallizes in cubes. The protosulphuret of iron may be formed artificially, by mixing together iron filings and sulphur, and exposing the mix- ture for some time to a red heat, in a well cover- ed crucible. The mixture fuses, and the excess of sulphur is gradually driven off, and protosul- phuret of iron remains. This sulphuret, when in crystals, has a yellow colour, the metallic lustre, and a specific gravity of about 4i. When it is reduced to powder, and digested for a sufficient length of time in nitric acid, the 350 IRON, NICKEL, COBALT, &c. sulphur is converted into sulphuric acid, while the iron at the same time is oxydized. If the so- lution, thus changed, be mixed with a sufficient quantity of muriate of barytes, sulphate of ba- rytes precipitates* the weight of which enables us to determine the proportion of sulphur which the protosulphuret of iron contains. 100 grains of protosulphuret, treated in this way by Berze- lius, yielded him 269 grains of sulphate of ba- rytes.* This weight exceeds the truth a little ; because a portion of the oxide of iron is precipi- tated along with the sulphate of barytes, suffi- cient to give it a reddish yellow colour. \Vhen the oxide of iron is precipitated, in the first place, by means of ammonia, and the requisite precau- tions taken to prevent any of the sulphuric acid from remaining in combination with this preci- pitate, the sulphate of barytes is perfectly white, and weighs almost exactly 267*3? grains. Now, 267'37 grains of sulphate of barytes contain al- most exactly 36*36 grains of sulphur. Thus it appears that 100 grains of protosulphuret of iron contain 36 '36 grains of sulphur. Consequently, the constituents of the protosulphuret are Sulphur 36-36 or 2 Iron . 63-64 - 3-5002 100-00 It is obvious that this protosulphuret is n com- * Ann. de Chim. LXXV1II. 12C. IRON. 351 pound of 1 atom sulphur and 1 atom iron ; and, consequently, that an atom of iron weighs 3*5. These observations apply only to protosulphuret of iron formed artificially. The native protosul- phuret, when dissolved in muriatic acid, always deposites a little sulphur. Hence, it contains either a small mixture of persulphuret of iron or of sulphur. Now, it is possible to convert protosulphuret of iron into neutral protosulphate of iron, by int P rot - * sulphate of careful digestion in a sufficient quantity of dilute iron. nitric acid. Hence it follows, that the sulphur and iron have the same ratio to each other in protosulphuret and protosulphate of iron ; name- ly, the ratio of 2 to 3'5. When 2 of sulphur are converted into sulphuric acid, they become 5 ; and when 3 '5 of iron are converted into pro- toxide, they become 4%5, as is obvious from what has been stated above. Hence it is clear, that 17'375 grains of protosulphate of iron must con- tain exactly 4*5 grains of protoxide of iron ; as the quantity of sulphuric acid has been shown to be 5 grains. Thus the proportion of acid, oxide, and water in 17*375 grains of protosulphate of iron has been rigidly made out ; and the com- ponent parts of the salt are as follows : 1 atom sulphuric acid 5 1 atom protoxide of iron 4-5 7 atoms water . 7-875 17-375 iron. 352 IRON, NICKEL, COBALT, &c. Expert 5. Berzelius has fixed the atomic weight of Berzeiius iron at 3'39215, or (as the number should have minftte been) 3 '393. His conclusion is founded upon atom of tne f i] ow i n g experiment : 100 grains of iron were dissolved ifl muriatic acid, and the hydro- gen gas evolved being burned over lime water, a quantity of carbonate of lime precipitated, which weighed 4*1 25 grains. Now, this is equi- valent to 0*495, or very nearly half a per cent, of carbon. The solution was boiled with nitric acid, to peroxidize the iron ; it was then preci- pitated by caustic ammonia. The precipitate, being washed, dried, and calcined, weighed 143*5 grains ; but, if 99'5 grains of iron yielded 143-5 grains, it is evident that 100 grains would have yielded 144-2 grains.* Thus it follows, that pe- roxide of iron is a compound of Iron 100 or 3-393 Oxygen 44-2 - 1-5 But this experiment does not seem susceptible of complete accuracy, unless it were made upon iron in a state of perfect purity ; for we are not certain that all the carbon is converted into car- buretted hydrogen gas. The strong smell of the gas evolved indicates the presence of some oily or bituminous substance, which may be partly deposited before the combustion of the hydrogen gas. Besides, different experimenters have ob- * Ann. de Chim, LXXVIII. 230. IRON. 353 tained different results in this way. I made two experiments with great care. In one, 100 iron were converted into 142*6 of peroxide j in an- other, into 144*75 grains.* Hassenfratz found that 100 grains of iron, calcined on a cupel, were converted into 142*224 grains of peroxide.t In another experiment, 100 iron, when thus treated, became 145 per- oxide. JBucholz, by an experiment similar to that of Berzelius, obtained 142 grains of peroxide from 100 grains of iron.t Gay-Lussac, from 100 grains of iron, obtained 142*35 grains of peroxide. The mean of all these experiments gives 43*303 for the quantity of oxygen which unites with 100 of iron. Hence, peroxide of iron ac- cording to these experiments is composed of Iron 100 or 3-463 Oxygen 43-303 - 1-5 Now, 3*463, the atomic weight of iron de- duced from these experiments, approaches nearer to 3*5, the weight deduced from the experiments above described, than to Berzelius' number. 6. I determined the atomic weight of the Atom of peroxide of iron in the following manner: 7*5 fron deter- grains of cubic pyrites, very carefully selected, m * Nicholson's Journal, XXVII. 378. f Ibid. XXVI. 149. t Ibid. XXV. 353. Ann. de Chim. III. 356. Vor,. I. Z 354 IRON, NICKEL, COBALT, &c. were reduced to a coarse powder, and digested in dilute nitric acid till the whole was dissolved. An excess of caustic ammonia was then poured into the solution, by which the peroxide of iron was thrown down. It was collected on a double filter, washed, dried, and exposed to a red heat. It weighed exactly 5 grains. The residual li- quid, thus freed from iron, was concentrated, in order to drive off the excess of ammonia. It was then mixed with a sufficient quantity of mu- riate of barytes to throw down all the sulphuric acid. The sulphate of barytes, thus obtained, being washed, dried, and heated to redness, weighed 29*5 grains, equivalent to 4 grains of sulphur. From this experiment it is evident, that persulphuret of iron is composed of Sulphur 4 Iron 3-5 7-5 or two atoms sulphur and 1 atom iron. But during the experiment the 3%5 of iron were con- verted into peroxide, which weighed 5. Hence it is obvious, that peroxide of iron is composed of Iron 3' 5 or 1 atom Oxygen 1-5 - 1 atom IRON. 3,55 Consequently, when 100 parts of pure iron are converted into peroxide they become 142*857. 7. Peroxide of iron has the property of com- bining with water, and forming with it a per- hydrate. This compound is found native, and constitutes what the Germans call brauneisen- stein. D'Aubuisson first showed it to be a per- hydrate of iron. Very frequently it is mixed with a little oxide of manganese and silica. This is almost always the case with the variety called hematite. But when the perhydrate is pure, it is a compound of 1 atom water . 1-125 1 atom peroxide of iron 5 6-125 A portion of this water is very easily dissipated : hence, unless very carefully examined, the por- tion of water will turn out rather less than an atom. 8. I found, many years ago, that when 100 parts of iron are oxydized by passing the steam of water over them at a red heat, they combine Lussac with 29'7528 parts of oxygen.* Bucholz ana- lyzed iron by the same process, and found that 100 grains of the oxide thus obtained, when converted into peroxide by nitric acid, became 110 grains. Hence, he inferred that 100 oxy- * Nicholson's Jour. XXVII. 378. Z2 356 IRON, NICKEL, COBALT, &c. dized by steam, combine with 29*87 grains of oxygen.* This is nearly the same result as I obtained. But from the subsequent experi- ments of Gay-Lussac it would appear that iron, oxydized in this* manner, unites with more oxy- gen than was found either by Bucholz or my- self. He found the oxide composedt of Iron 100 or 3-5 Oxygen 37'8 - 1-323 Now, 1*323 is very nearly 1^ oxygen. Hence, it would appear that this supposed oxide is a compound of 1 atom iron + 1^ atom oxygen. Now, this is just the proportion of oxygen which would result from the combination of 1 inte- grant particle of protoxide, and 2 integrant particles of peroxide of iron ; for 1 atom protoxide is composed of 3-5 iron -f- 1 oxygen 2 atoms peroxide of . 7 +3 Total . . . 10-5 4 And l 3 - = 3-5 ; and j = 1 j. So that such a compound would obviously consist of 1 atom of iron, and 1 atom oxygen. When this oxide of Gay-Lussac is dissolved in sulphuric acid, and alcohol poured upon the sulphate, a solution of persulphate of iron is ob- * Nicholson's Jour. XXV. 353. f Ann. de Claim. LXXX. 164. IRON. 357 tained, while protosulphate remains undissolved. From this and several other facts stated by Gay- Lussac in his Memoir, it is evident that this substance cannot be considered as a peculiar oxide, but as a combination of protoxide, and peroxide in the proportions above stated. From the phenomena which appear during Seems to ' exist in the manufacture of Prussian blue, there are Prussian strong reasons for considering the iron which that pigment contains as in the state of Gay- Lussac's oxide, or as a mixture or compound of 1 part of protoxide with 2 parts of peroxide. For if we attempt to make Prussian blue by means of persulphate of iron at once, we get a pigment apparently of the finest colour ; but so intense that it appears quite black when dry. The manufacturers always find it necessary to employ protosulphate of iron. The Prussian blue, when first thrown down, is an exceedingly light blue powder. This powder is continually washed with water for a fortnight or three weeks. The shade gradually deepens during the washings, and when it is judged sufficiently deep, the water is drained off and the pigment allowed to dry. From my knowledge of this process, I am disposed to think that protoxide and peroxide of iron are capable of uniting in the proportion of 1 atom of the former to 2 atoms of the latter. But, with the exception of Prussian blue, I am not aware of any combina- Z3 358 IRON, NICKEL, COBALT, &c. tion into which this compound is capable of entering. SECT. II. - OF THE ATOMIC WEIGHT OF NICKEL. THE atomic weight of nickel and its protoxide may be determined without much difficulty from the sulphate, which I find to be a compound of 1 integrant particle of sulphuric acid, 1 inte- grant particle of protoxide of nickel, arid 7 inte- grant particles of water. This salt has a fine grass green colour, and crystallizes in four-sided rectangular prisms. Its taste is astringent, but not strong, and it is very soluble in water. sulphate i. 17-1^5 grains of the crystals of this salt of nickel TIT i -i i analyzed, were dissolved in water, and mixed with a solu- tion of 13*25 grains of chloride of barium. A double decomposition took place, and sulphate of barytes precipitated. The clear residual li- quor being tested by sulphate of soda and mu- riate of barytes, was not affected by either, showing that it contained neither barytes nor sulphuric acid. It is obvious from this experi- ment, that 17*125 grains of crystallized sulphate of nickel contain just the quantity of sulphuric acid necessary to saturate the barytes from 13-25 grains of chloride of barium ; that is to say, ex- actly 5 grains of this acid. NICKEL. 359 2. When 17-125 grains of crystallized sulphate of nickel are exposed to a heat of 600 for half an hour, they lose 6-25 grains of their weight, or rather more than 5i atoms of water. But at this temperature the salt does not lose its green colour, and consequently, still retains water. When 17-125 grains of the salt are exposed to a low red heat, they lose their green colour, and in general, diminish in weight 8 grains. But the salt which has been thus treated, is not com- pletely soluble in water, though very nearly so : hence, a small portion of its acid has been dis- engaged. From this experiment, which I have often made, it appears that the water of crystalli- zation in 17*125 grains of sulphate of nickel is not quite so high as 8 grains. In one experiment in which the heat was applied with much caution, the loss of weight was 7*8 grains, and the salt was completely soluble in water. It is obvious from this, that the water of crystallization in 17-125 grains of sulphate of nickel lies between 7-8 and 8 grains. Now, 7-875 grains of water are equivalent to 7 atoms of that liquid : and as the quantity of acid in this weight of salt is just 5 grains, equivalent to an atom, we can have no doubt that the true quantiy of water is likewise equivalent to 7 atoms, or 7*875 grains. Five acid and 7-875 water, make 12-875, which ot j de being subtracted from 17*125 leaves 4<25, which rf "*ei- must be the weight of protoxide of nickel, and z 4 360 IRON, must likewise (since the salt is neutral) indicate the atomic weight of protoxide of nickel. Thus it appears, that sulphate of nickel is composed of 1 atom sulphuric acid 5 1 atofli protoxide of nickel 4-25 7 atoms water . 7*875 17-125 There is no reason to doubt that this oxide is a compound of 1 atom nickel and 1 atom oxygen. Atom of An atom of nickel, therefore, must weigh 3*25. 3. Besides the protoxide of nickel there is an- other oxide, having a black colour, which was first pointed out by Thenard, and which he ob- tained by passing a current of chlorine gas through water, having protoxide of nickel sus- pended in it. A portion of the oxide dissolves in the muriatic acid evolved, while another por- tion assumes a black colour and is converted into peroxide. It will be obvious, from the account given below of the constitution of this oxide, that id of the protoxide of nickel must be dis- solved and converted into muriate of nickel, while |ds become peroxide of nickel and remain undissolved. Atom of When peroxide of nickel is dissolved in am- Ukkei. e monia, an effervescence takes place, owing to the escape of azotic gas: for the peroxide is con- verted into protoxide at the expense of the hy- drogen of the ammonia. An effervescence also 16 COBALT. 361 takes place when this oxide is digested in acids : the excess of oxygen flies off in the state of gas, and the protoxide thus formed unites with the acid. From some experiments of Rothoff, quoted by Berzelius,* it would appear that the oxygen in peroxide of nickel is to that in the protox- ide as 3 to 2. If this be the case, the atomic weight of peroxide of nickel will be 4'75, and it will be a compound of 1 atom nickel . . 3-25 1 atom oxygen . . 1*5 4-75 If this supposition be well founded, 47*5 grains of peroxide of nickel, when dissolved in sulphu- ric acid, will lose 5 grains, or 14754 cubic inches of oxygen gas. SECT. III. OF THE ATOMIC WEIGHT OF COBALT. THE sulphate of cobalt, when pure, crystallizes in small red rhombic prisms, very similar to the crystalline form of protosulphate of iron. It may be exposed to an incipient red heat, and * Annals of Philosophy, III. 355. 362 IRON, NICKEL, COBALT, &c. deprived of the whole of its water of crystalliza- tion without losing any of its acid. Sulphate of i. 9-25 grains of anhydrous sulphate of cobalt cobalt ana- L were dissolved in water, and the liquid mixed with a solution gf I3*%5 grains of chloride of ba- rium. A double decomposition took place ; sul- phate of barytes precipitated, and the clear re- sidual liquid being examined, was not affected by sulphate of soda or muriate of barytes show- ing that it contained neither barytes nor sulphu- ric acid. From this experiment it is obvious, that 9*25 grains of anhydrous sulphate of cobalt contain exactly 5 grains of sulphuric acid : the remaining 4-5 grains must be protoxide of co- balt. Consequently, this salt is composed of Sulphuric acid . . 5 Protoxide of cobalt . 4-25 The salt being neutral, and 5 being the atomic Atom of weight of sulphuric acid, 4'&5 must be the atomic protoxide . , ,, of cobalt weight ot protoxide of cobalt. 2. When 17'125 grains of crystallized sulphate of cobalt are exposed to an incipient red heat in a platinum crucible, they lose 7*875 grains of their weight. Now, this is equivalent to 7 atoms of water. It is obvious from this experiment, that sulphate of cobalt is composed precisely as sulphate of nickel, or of COBALT. 363 1 atom sulphuric acid 5 1 atom protoxide of cobalt 4-25 7 atoms water . . 7'875 17-125 3. There seems no reason to hesitate about considering this oxide as a compound of 1 atom oxygen and 1 atom cobalt. On that supposition, the atomic weight of cobalt will be 3*25, the Atom of same as the atomic weight of nickel. 4. If protoxide of cobalt be a compound of 1 atom cobalt and 1 atom oxygen, it follows, that 100 parts of this oxide are composed of Cobalt . . . 76-47 Oxygen . . . 23-53 100-00 Now, when the protoxide of cobalt is precipitated from an acid, edulcorated and dried in the open air, it absorbs oxygen during the drying, and is converted into a black powder, which is the per- oxide of cobalt. This peroxide dissolves with effervescence in muriatic acid, while chlorine gas is disengaged. When heated, it gives out oxygen gas 5 and, from the experiments of Rothoff, as stated by Berzelius,* it appears, that when 111 grains of peroxide of cobalt are heated, they are converted into 100 grains of protoxide : so that * Annals of Philosophy, III. 356. 364 IRON, NICKEL, COBALT, c. 100 grains of protoxide become peroxide by uniting with 11 grains of oxygen. But 11 2 = %%, which approaches very near the oxygen in 100 grains of protoxide. Atom of From this we see that the oxygen in the prot- cobah. e oxide is to that in the peroxide as 2 to 3. Con- sequently, the atomic weight of peroxide of co- balt must be 4-75, and it must be a compound of 1 atom cobalt 3-25 1 atom oxygen 1-5 4-75 Thus its composition and weight agree with those of peroxide of nickel ; like that peroxide, it does not appear capable of uniting with acids, nor of entering into any permanent combina- tions like the peroxide of iron. SECT. IV. OF THE ATOMIC WEIGHT OF MANGANESE. THE atomic weight of the protoxide of this metal may be made out from the protosulphate in the same way as the protoxide of iron, nickel, and cobalt. Protosulphate of manganese, when pure, forms large transparent crystals, having a beautiful flesh colour, and the shape of a very oblique rhombic prism, the faces of which meet MANGANESE. 365 at angles of about 148 and 32. Its taste is somewhat similar to that of Glauber's salt, but rather more disagreeable ; and like that salt, it acts as a cathartic when taken to the extent of about half an ounce. It is very soluble in water, but is not altered by exposure to the air, whether the atmosphere be dry or moist. 1. 15-125 grains of these crystals were dissol- Sulphate of manganese ved in water, and the liquid mixed with a solu- analyzed, tion of 13-25 grains of chloride of barium. A double decomposition took place, and when the sulphate of barytes had subsided, the residual liquid was neither affected by sulphate of soda nor muriate of barytes. From this experiment it follows, that 15-125 grains of protosulphate of manganese contain exactly 5 grains of sulphuric acid. 2. When 15-125 grains of protosulphate of manganese are heated in a platinum crucible, they melt, and the water of crystallization is gra- dually dissipated. But it is scarce possible, by this process, to drive off the whole water of crys- tallization without allowing some of the acid at the same time to escape. Dr. John, from 15*125 grains of this salt, drove off 5 -345 grains of wa- ter. I repeated this experiment with great care, and found that I could disengage 5 '48 grains of water from 15-125 grains of the salt. But if the heat was raised high enough to produce a greater diminution of weight, a portion of the salt was 366 IRON, NICKEL, COBALT, C . altered in its appearance, and rendered inso- luble. Five atoms of water weigh 5 '625, which weight exceeds 5-48 by only 0*145, or about the eighth part of an atom fc There is no reason to doubt that the aqueous portion of this salt constitutes a determinate number of atoms; and since 15*125 of the salt contain exactly one atom of sul- phuric acid, the atomic weight of the water must exist also in the same quantity. Now, the results just stated point out unequivocally that five atoms of water are united in it with an inte- grant particle of anhydrous protosulphate. Thus it appears, that 15*125 grains of protosulphate of manganese contain 5*625 grains of water. 3. The five grains of sulphuric acid and the 5*625 grains of water make together 10*625 grains. This quantity, subtracted from the ori- ginal weight, leaves 4*5, which must denote the quantity of protoxide of manganese present. Thus it appears, that protosulphate of manganese is composed of Sulphuric acid 5 Protoxide of manganese 4*5 Water . . . 5-625, 15-125 Atom of ^ S ^ e sa ^ * s neutra ^ an ^ as 5 is the weight of protoxide an atom of sulphuric acid, 4*5 must be the of manga- . . * . nese. atomic weight of protoxide or manganese. MANGANESE. 367 Thus it appears, that protoxide of manganese and protoxide of iron have the same atomic weight. If protoxide of manganese be a com- pound of one atom manganese and one atom oxygen, (as there is the strongest reason for be- lieving) then an atom of manganese weighs 3*5, or the same as an atom of iron. 4. Manganese occurs most commonly in the Native . 1 . . , . black oxide. state or a black coloured oxide, which has been long employed to deprive glass of its colour, and is at present consumed in great quantities by the bleachers and the preparers of the bleaching powder. This black oxide is most commonly impure ; but it occurs occasionally crystallized in small needles, which have the form of rhom- bic prisms, and the lustre and colour of iron, though rather darker. I have met with speci- mens of this kind perfectly pure. When 11 grains of this pure oxide are kept for some time in a low red heat in a platinum crucible, they lose exactly one grain of their weight, which consists of pure oxygen gas. By this exposure to heat, the metallic lustre disappears, and the mass assumes a brownish black colour. The powder thus formed is a peculiar oxide of man- ganese; for it dissolves without effervescence in sulphuric and in muriatic acid, and forms dark red coloured solutions. This sulphate and muriate do not crystallize ; indeed, when the muriate is heated, chlorine gas is evolved, and 368 IRON, NICKEL, COBALT, &c. the common colourless protomuriate of manga- nese obtained. But the red sulphate of manga- nese may be concentrated without losing its co- lour. Sulphurous acid, or nitrous acid, instantly renders it colourless, by depriving it of the ex- cess of oxygen which it contains above the pro- tosulphate. ^ us> besides the protoxide of manganese, there are two other oxides, both of which are black. I determined the composition of the brownish black oxide in the following way : A quantity of carbonate of manganese, which I knew to con- tain exactly 45 grains of protoxide, was dissol- ved in nitric acid in a platinum crucible ; the solution was evaporated to dryness, and the cru- cible with its contents was then exposed to a red heat : by this means, the nitric acid was decom- posed, and driven off, and a black shining pow- der obtained, which weighed exactly 50 grains. This black powder possesses the same properties with the brownish black powder obtained by ex- posing native black oxide of manganese to a low red heat for it forms the same red coloured solutions with sulphuric and muriatic acids, which are rendered colourless by sulphurous acid, or when heated with sugar.* * When carbonate of manganese is exposed to a red heat, it assumes a brown colour ; but is not converted into deutoxide, unless it be previously mixed with nitric acid. A quantity of carbonate containing 4<-5 grains of protoxide of manganese, when thus treated, instead of weighing 5 grains as it would do if it were converted into deutoxide, only weighs 4-8 or MANGANESE. 369 Thus it appears, that 45 protoxide of manga- nese become 50 deutoxide ; or, 4*5 protoxide become 5 deutoxide : but 4 *5 is the atomic weight of protoxide of manganese, and it con- tains 1 atom of oxygen. Five, therefore, must be the atomic weight of deutoxide ; and it is ob- viously a compound of 1 atom manganese 3-5 Ji atom oxygen 1*5 5-0 It is, therefore, similar in its composition to the peroxide of iron. As its colour is either a brown- ish black, or a shining black, according to the mode of preparing it, we cannot with propriety distinguish it by its colour ; and shall, therefore, henceforth give it the name of deutoxide of man- ganese. 5. The composition of deutoxide of manga- perhaps 4-83; showing that it is still partly in the state of protoxide. This is the oxydum mcmganoso-manganicum of Arfwedson, which he considers as composed of 1 atom protoxide 4-5 :2 atoms deutoxide 10 3)14-5 4-8333 = atomic weight of the compound. When the black oxide is exposed to too high a temperature, it becomes analogous to this in its composition, showing that it is partly converted into protoxide. VOL. I. A a 370 IRON, NICKEL, COBALT, &c. nese being known, it is easy, from what has been already said, to determine the composition of the native black oxide employed by manufac- turers : 11 black oxide, when heated to redness, Atomic i ose i of oxygen, and become 10 deutoxide : or, weight of J ~ which is the same thing, 5 '5 black oxide, when deprived of 0*5 oxygen, become 5 deutoxide ; consequently, the atomic weight of native black oxide is 5 %5, and it is obviously composed of 1 atom manganese 3-5 2 atoms oxygen 2 5-5 This is an oxide not analogous to any of the known oxides of iron, nickel, or cobalt. It has been hitherto distinguished by the name of per- oxide of manganese ; but, as there is reason to believe that manganese forms at least one, if not two additional oxides, it will be better hence- forth to call it tritoxide of manganese. *J O ^" Scheele, m &ny years ago, made a curious experiment with the tritoxide of manganese. He mixed it with caustic potash, and exposed it to a heat sufficiently strong to melt the alkali. The fusion being kept up for some time, a dark green, or blackish mass was formed, which was soluble in water. By degrees, a yellow powder precipitates from this solution, and the liquid assumes a blue colour ; some time after, the co- MANGANESE. 371 lour becomes violet, and then red ; and, finally, black oxide of manganese precipitates, and the liquid loses all colour. In consequence of this curious succession of colours, the liquid thus ob- tained used to be distinguished by the name of mineral chameleon. We may employ saltpetre instead of potash, and the process will succeed equally well. When the liquid had assumed a red colour, Chevillot and Edwards* concentrat- ed it by evaporation, and obtained a red colour- ed salt in needles. This salt is neutral, and is composed of potash, united to an oxide of man- ganese, which possesses acid properties, and which, consequently,, is different from the three oxides, which have been already noticed. This red salt has received the name of man- of potash. ganate of potash. It possesses some singular properties. It crystallizes in four-sided rectan- gular prisms, has the diamond lustre, a purple colour, and a taste at first sweet, and similar to that of walnuts, but which leaves a very disa- greeable impression in the mouth. When these crystals are heated, in contact with hydrogen gas, they set it on fire. They detonate violently with phosphorus, set fire to sulphur, antimony, and arsenic, and rival chlorate of potash itself in their power of supporting combustion. The acid cannot be separated from this salt without de- * Ann. de Chhn. et de Phys. VI J I. 337. A a2 372 IRON, NICKEL, COBALT, &c. composition. Sulphuric acid disengages it from the potash ; but when we endeavour to separate the sulphate of potash by evaporation, the red colour disappears, and tritoxide of manganese precipitates. B u t Dr. Forchhammer succeeded by a very mer s mode of obtaining ingenious process in obtaining manganesic acid manganesic , . . , , acid. in a separate state, or at least combined merely with water. He mixed together tritoxide of manganese and crystallized nitrate of barytes, and exposed the mixture to a heat just sufficient to produce the fusion of the salt. Great care, he says, must be taken not to drive oft' the water of crystallization, otherwise the process does not succeed.* The nitric acid is gradually decompos- ed. The tritoxide of manganese unites with more oxygen, is converted into an acid, and combines with the barytes, as that alkali is disengaged. After the process is finished, the mass must be digested in water, which dissolves any pure ba- rytes, or nitrate of barytes, that may be present; but does not touch the compound of barytes and acid of manganese, \vhich is insoluble in that liquid. To this insoluble residuum Dr. Forch- liammer added the quantity of sulphuric acid just capable of saturating all the barytes which it contained. The acid of manganese in the * I have not attempted to repeat this process, because I have never met \vith any nitrate of barytes that contained any water of crystallization. As far as my observations go, the salt is always anhydrous. MANGANESE. 373 mass is immediately resolved into two substances, tritoxide of manganese, and manganesic acid. The former remains mixed with the sulphate of barytes ; but the latter dissolves in water, and gives that liquid a beautiful red colour. 7 When tritoxide of manganese and nitre are heated, the solution is at first green, and it gra- dually passes into red, while, at the same time, a quantity of deutoxide of manganese precipi- tates. When an acid is poured into the green coloured solution, it becomes red ; an alkali, on the other hand, renders the red coloured solution green. When alcohol, or any animal or vege- table substance, is added to the red solution, deutoxide of manganese precipitates ; while, at the same time, the red colour is destroyed. Dr. Forchhammer is of opinion, from these and similar phenomena, that there are two acids of manganese, which he calls manganous and manganesic acids the former containing the least, and the latter the most oxygen. The manganous acid exists in the green coloured liquid, which is a compound of that acid arid potash ; while the red salt is a compound of manganesic acid and potash. 8. To determine the composition of manga- His method i TX -n i i i .of analyz- nous acid, Dr. Jborchhammer prepared a quantity ing ma n- of the manganite of potash, and decomposed it by the addition of a solution of nitrate of lead. A dark coloured precipitate fell, consisting of a A a 3 374" IRON, NICKEL, COBALT, &c. combination of peroxide of lead and deutoxide of manganese ; and this precipitate he consider- ed as containing all the oxygen and all the man- ganese which existed in the manganous acid. This precipitate, being washed and dried, was mixed with bisulphate of potash, and exposed to a heat capable of fusing it in a retort. The oxy- gen gas given out was collected and measured ; its weight was O1303 parts. The salt in the re- tort was dissolved in water (the lead remaining behind in the state of sulphate) ; the manganese was precipitated by carbonate of soda, and ex- posed to a red heat. The deutoxide obtained by this process weighed O282 parts. Now, deutoxide of manganese is (as we have seen) a compound of 3 '5 manganese +1-5 oxy- gen ; consequently, 0-282 parts of this oxide contain Manganese . 0-1974 Oxygen . 0-0846 0-2820 If to this portion of oxygen we add the O1303 parts driven off by heat, we shall have manga- nous acid composed of Manganese 0-1974 or 3-5 Oxygen 0-2149 - 3-8103 0-4123 MANGANESE. 37 5 Now, 3*8103 approaches nearly to 4 atoms of oxygen from which it would appear, that man- ganotis acid is a compound of 1 atom manganese and 4 atoms oxygen ; and that its atomic weight is 7*5. But Dr. Forchhammer concludes (though I do not see the force of his reasoning), that, in reality, manganous acid contains only 3 atoms of oxygen, united to 1 atom of manganese. Its atomic weight, on that supposition, will be 6%5. Dr. Forchhammer's process of experimenting And man- and reasoning, by which he made out the com- position of manganesic acid, is very ingenious, but rather abstruse. I shall endeavour to lay a simple view of it before the reader : Through a solution of manganite of potash, he passed a current of carbonic acid gas. By this process the manganite was converted into manganate of potash ; while, at the same time, 0-136 parts of deutoxide of manganese were thrown down. The manganate was then decomposed* by means of alcohol ; 0*214 parts of deutoxide of manganese were 'precipitated. Now, by the first process, the manganesic acid formed con- tained all the oxygen in the manganous acid, except what remained in the portion of the deu- toxide separated. The whole deutoxide separat- ed was 0-136 + 0-214 = 0-350 ; and this was the quantity of deutoxide contained in the manga- nous acid. The manganesic acid contained only 0-214 parts ; and with these 0-214 parts A a I 376 IRON, NICKEL, COBALT, &c. there was combined just as much oxygen as with the 0*350 parts in the manganous acid. But manganous acid is composed of 5 deu- toxide of manganese + 1 *5 oxygen ; consequent- ly, 0-350 are composed of 0*245 deutoxide + 0*105 oxygen ; and manganesic acid is a com- pound of Deutoxide of manganese 0-214 Oxygen . . 0-105 Deutoxide of manganese is a compound of 3*5 manganese + 1*5 oxygen; consequently, 0-214 deutoxide consists of 0*1498 manganese + 0-0642 oxygen. By adding together the 0*105 oxygen and the 0*0642, we obtain 0*1692 for the whole oxygen in manganesic acid, combined with 0*1498 of manganese. Thus it appears, that manganesic acid is a compound of Manganese 0-1498 or 3-5 Oxygen 0-1692 - 4-620 0-3190 If we make the same allowance for excess of oxygen that Dr. Forchhammer made in the case of manganous acid, manganesic acid may be considered as a compound of 1 atom manganese and 4 atoms oxygen, which would make its atomic weight 7 '5. It would appear from the preceding investiga- tion, that oxygen and manganese combine in no MANGANESE. fewer than 5 proportions, and form 5 oxides, the composition of which is as follows : manganese. Oxides of 1 Protoxide . 3-5+1 2 Deutoxide . . . 3-5 + 1-5 3 Tritoxide . . . 3-5 + 2 4 Manganous acid . . 3-5 + 3 5 Manganesic acid . . 3-5 +4 9. Berzelius describes an oxide of manganese containing less oxygen than any of the preced- ing. It was obtained by leaving metallic man- ganese for more than a year in a phial not very closely corked. It fell into a dark coloured powder, which, according to Berzelius' analysis, or rather calculation, was a compound of Manganese . . 100 Oxygen . . 10-666 I have had no opportunity of making any ex- periments on this supposed oxide : but, if it exist, it must be a compound of 3 atoms man- ganese with 1 atom oxygen. This would con- sist of Manganese 10-5 or 100 Oxygen 1-0 - 9'5238 which differs but little from the estimate of Ber- zelius. 10. Dr. John found, that when metallic man- ganese is put into water, that liquid was slowly m 378 IRON, NICKEL, COBALT, &c. decomposed ; hydrogen gas being evolved, while the manganese lost its metallic lustre and assumed a green colour : 80 parts of manganese thus treated, became 92 parts of the new oxide. Now, 80 : 12 i : 3-5 : 0-525. Hence it would appear that this oxide, which I have not had an opportunity of examining, is a compound of 1 atom manganese and half an atom of oxygen ; or of 2 atoms manganese . . 7 1 atom oxygen ... 1 This oxide is incapable of uniting with acids ; but Dr. Forchhammer has made a set of experi- ments to show that it combines with sulphur, and that it constitutes the mineral found in Transyl- vania and Cornwall, and considered at present as a sulphuretted oxide of manganese. It seems, however, proved that the Transylvanian ore is a sulphuret of manganese. I think the existence of the supposed protox- ide of manganese of Berzelius exceedingly doubtful. But I see no reason to doubt the ex- istence of the suboxide of John though that chemist was mistaken in considering it as the basis of the salts of manganese. The real basis is the substance which I have here distinguished by the name of protoxide, and which contains just twice as much oxygen as the suboxide of John. CERIUM. 379 Of all the substances hitherto examined, man- ganese is capable of uniting with the greatest number of doses of oxygen. SECT. V. OF THE ATOMIC WEIGHT OF CERIUM. CERIUM has not hitherto been obtained in the metallic state ; but it has been ascertained that it combines with two proportions of oxygen, and forms two oxides ; both of which possess Alka- line properties, and are capable, of consequence, of combining with acids, and of constituting the bases of a variety of salts. The only person who has hitherto attempted Muriate of * * ri to determine the atomic weight of this metal and its oxides, is M. Hisinger. His deductions were made from an analysis of the protomuriate of cerium. A quantity of this salt was dissolved in water, and the muriatic acid thrown down by nitrate of silver. The chloride of silver, after fusion, weighed 1*819, equivalent to 0-4609 parts of muriatic acid. The solution having been freed by muriatic acid from the excess of silver which it contained, was mixed with carbonate of am- monia to throw down the protoxide of cerium. This oxide, after being heated to redness in a close vessel, weighed 0-684 parts.* From this * Annals of Philosophy, IV. 355. cerium ana. ty*** by 380 IRON, NICKEL, COBALT, c. analysis it follows, that protomuriate of cerium is composed of Muriatic acid 0-4609 or 4-625 Protoxide of cerium 0-684 - 6-862 As the salt was. neutral, the conclusion to be drawn from this analysis is, that the atomic weight of protoxide of cerium is 6 -862. But I think there is reason to suspect that the muriatic acid in this analysis is somewhat over- rated, and the protoxide of cerium somewhat un- derrated : at least, I generally find, that when nitrate of silver is poured into the solution of a metalline salt, the chloride obtained contains traces of the metal previously in combination with the muriatic acid. In order to obtain cor- rect results from the analysis of a metalline mu- riate, the metallic oxide should be thrown down in the first place, by means of an alkaline car- bonate or pure alkali ; and after the metallic ox- ide has been separated by the filter, the liquid is to be neutralized by nitric acid, and the muriatic acid thrown down by nitrate of silver. These considerations induced me to make the following experiments in order to determine the atomic weight of protoxide of cerium. i A quantity of oxalate of cerium was ex- ments of 1 J the author, posed to a red heat, in order to destroy the acid which it contained. The peroxide of cerium thus obtained, was dissolved in muriatic acid, and the liquid was repeatedly evaporated to dry- CERIUM. 381 ness with an excess of acid till it became quite colourless. Carbonate of ammonia was then mixed with it in sufficient quantity to throw down the protoxide of cerium in the state of a carbonate. The carbonate of cerium, when al- lowed to dry in the open air, was a very light, bulky, white, tasteless substance, in small scales or plates, somewhat resembling fhe appearance of earthy talc. It was dissolved in dilute sul- phuric acid till the acid refused to take up any more. The solution was colourless, and on being sufficiently concentrated deposited protosulphate of cerium in small crystals, which had the form of six-sided prisms, and some of them seemed to be three-sided prisms ; but they were too small to admit of any thing like an accurate measure- ment of their angles. This salt has a flesh colour, a sweet taste, with a slight degree of astringency, and the property of reddening vegetable blues. It is pretty solu- ble in water ; though the precise degree was not determined. 2. 15'6&5 grains of these crystals were dis solved in distilled water, and the liquid mixed rium ana. Ivzed with a similar solution of 13 '25 grains of chlo- ride of barium. A double decomposition took place, sulphate of barytes precipitated, and the residual liquid was not affected by sulphate of soda nor muriate of barytes. Hence, it con- tained no sensible quantity of sulphuric acid, or 382 IRON, NICKEL, COBALT, c . barytes. From this experiment it follows, that 15*625 grains of protosulphate of cerium con- tain 5 grains of sulphuric acid. 3. 15-625 grains of the crystals of protosul- phate of cerium, were exposed for an hour on the sand bath to a heat of about 500 or rather more. The loss of weight was 3*125 grains ; and this loss was not increased though the heat was continued for twenty four hours. The salt thus treated, had become white. It still red- dened vegetable blues, and was completely solu- ble in water. Hence, the loss of weight must be ascribed to the escape of water. Now, the weight of three atoms of water is equivalent to 3'375 grains. This exceeds the loss of weight, actually sustained by the salt, by 0-25 grains, equivalent to Jzrth of an atom. The salt being as neutral as possible, and 15*625 grains of it containing 5 grains of sulphuric acid, which is equivalent to an atom, the same weight of salt must contain a determinate number of atoms of water ; and it is obvious, that the true quanti- ty of water must be 3 atoms, or 3-375 grains. I found that the salt could not be rendered per- fectly anhydrous without at the same time losing a portion of its acid. 4. Five sulphuric acid, and 3*375 water added together make 8*375, which being subtracted from 15*625, the total quantity of salt used, leaves 7*25, which must represent the weight of CERIUM. 383 protoxide of cerium in the salt, and must also be the atomic weight of that protoxide. From the preceding analysis the reader will perceive, that protosulphate of cerium is com- posed of 1 atom sulphuric acid . 5 1 atom protoxide of cerium 7*25 3 atoms water . . 3-375 15-625 7 -25, then, is the true atomic weight of protoxide Atom of f, . ,, 111-11 i protoxide or cerium, which is a little higher than the num- of cerium, ber deduced from Hisinger's analysis of the muriate of cerium. The probable reason for this difference has been already hinted at. 5. According to Hisinger's experiments the oxygen in the protoxide of cerium is to that in the peroxide as 2 to 3. My stock of cerium was not sufficient to enable me to make a simi- lar set of experiments on peroxide of cerium, of per - But there is every reason to confide in the ratios cerium. which Hisinger has deduced from his analyses ; all of which, having been made in the same way, would be subject to the same errors, which would not affect the ratios between the two oxides, though it had an effect upon their abso- lute weights. It is easy from these ratios to de- duce the atomic weights and composition of cerium and its oxides, as follows : 384 IRON, NICKEL, COBALT, &c. ATOMIC WEIGHT. CERIUM. OXYCEtf. Cerium 6-25 Protoxide 7'25 composed of 1 atom 4. 1 atom Peroxide 7 '7 5 1 + 11 Action of 6. The following experiments to determine the reagents. t m c effects of reagents on protomuriate of cerium, though they contain little novelty, are perhaps worth recording ; because the salts and all the reagents used were exceedingly pure. 1 . Prussiate of potash A chalky white precipitate. 2. Gallic acid and infusion of galls O 3. Oxalic acid A copious white precipitate. 4. Oxalate of Ammonia Ditto, redissolved by agitation. 5. Tartaric acid O- 6. Tartrate of potash A copious white precipitate. 7. Bitartrate of potash O. 8. Phosphoric acid A white precipitate. 9. Phosphate of soda Abundance of white flocks. 30. Carbonate of ammonia A milk white flocky precipitate. CHAP. XL OF THE ATOMIC WEIGHTS OF ZINC, CADMIUM, LEAD, TIN, BISMUTH, COPPER, MERCURY, AND SILVER. THESE eight metals possess several properties in common by which they are distinguished from all the other metals. 1. They all possess a cer- characters tain degree of malleability ; for bismuth, by far tais treated the most brittle of them all, may be dimpled by chapter? 8 a blow of the hammer ; and zinc, the next in order, is very malleable at a temperature some- what above that of boiling water. 2. They all melt at rather low temperatures, when compared with the melting point of most of the other metals ; for copper, which requires the greatest heat to melt it, may be fused in a common fire urged by bellows. It requires some care to keep a silver crucible red hot in a common fire with- out fusing it. The other six melt in a heat be- low redness. Indeed, mercury requires so very little heat to melt it, that it remains always fluid even in our coldest winter weather. 3. Their oxides are easily decomposed by heat and char- coal, so that there is no difficulty in obtaining VOL, I. B b 386 ZINC, CADMIUM, LEAD, &c. these bodies in the metallic state, and, compara- tively speaking, in a state of considerable purity. 4. When the last seven of these metals are held in solution by an acid (particularly sulphuric, or muriatic, when these acids form soluble salts with the metallic oxide) if we introduce into the solution a plate of zinc, the metal previously in solution is thrown down in the metallic state, while a corresponding portion of the zinc is dis- solved in its place. A plate of cadmium intro- duced into a solution of any of the six metals placed after it in the title of this chapter, will also throw it down in the metallic state, while a corresponding portion of the cadmium enters into solution. In like manner a plate of lead will throw down any of the five metals after it in the metallic state. A plate of tin will throw down any of the four metals after it ; a plate of bismuth, any of the three metals after it ; a plate of copper either of the two metals after it 5 and mercury will throw down silver in the metallic state. But silver is incapable of throwing down any of the other metals mentioned in the title of this chapter, though it is itself thrown down by them all. Mercury throws down silver, but none of the others ; copper throws down mer- cury and silver, but none of those that precede it. The same remark applies to every particular metal ; it throws down all those which are placed CADMIUM. 387 after it in the title of the chapter, but none of those that are placed before it. Iron has likewise the property of throwing down all the metals belonging to this chapter, except zinc, in the metallic state. When a plate of zinc is put into a solution of iron, the iron falls very slowly in the state of peroxide. Cop. per is a good precipitant of mercury and silver ; zinc precipitates cadmium, tin, and copper very well from muriatic acid, and lead very well from acetic acid. The reader will recollect that the atomic weight of zinc was established in the first sec- tion of the first chapter of this work. The know- ledge of it was necessary in order to enable us to determine the specific gravities of oxygen and hydrogen gases, upon which the whole doc- trine of atomic proportions ultimately depend. The weight of an atom of zinc is 4-'&5 ; it com- 4 tom of bines with only one proportion of oxygen, and forms one oxide, the atomic weight of which is 5-25. SECT. I. OF THE ATOMIC WEIGHT OF CADMIUM. CADMIUM is still a very scarce metal ; but by the kindness of Mr. Harepath, I was favoured Bb2 388 ZINC, CADMIUM, LEAD, &C. with a quantity of the matter which adheres to the roof of the room in which zinc is smelted in Bristol. This sublimate contains about ten per cent, of cerium ; by extracting it, I have had it in my power to examine the constitution of its different salts. Stromeyer, to whom we are indebted for our knowledge of this metal, has analyzed several of its salts, and ascertained how much the metal increases in weight when it is converted into an oxide : and from the well known precision of this excellent chemist, we have every reason to expect that the results of his analyses will give us the atomic weight of this metal, at least exceedingly near the truth. It will be proper, therefore, in the first place to state his experiments. M. Stromeyer analyzed the carbonate, sul- phate, nitrate, muriate, phospate, and borate of cadmium. He has not yet published the details of his analyses ; or at least, the second volume of his Untersuchungen, in which they will probably be found, has not yet reached Great Britain. This puts it out of our power to judge of the accuracy of the data which he employed in his calculations. Under this uncertainty, I shall make choice of the sulphate to determine the atomic weight of cadmium, because I know from other analyses which he has published in detail, that he employs the constituents of sul- phate of barytes as given by Berzelius, which, CADMIUM. 389 though not rigidly exact, approach exceedingly near the truth. According to Stromeyer, sulphate of C3L ^" mium has a good deal of resemblance to sul- cadmium by Stro- phate of zinc in its appearance. Its crystals are meyer. four-sided rectangular prisms, which are very soluble in water, and effloresce rapidly when exposed to the air. Heat easily drives off the water of crystallization ; but the salt may be exposed to a red heat without losing any of its acid ; this makes it easy to determine its water of crystallization. The constituents of this salt as determined by Stromeyer are Sulphuric acid . 28-5230 Oxide of cadmium 45*9564 Water . . . 25-5206 100* Now, these numbers bear to each other the same ratios as the following : Sulphuric acid . 5 Oxide of cadmium 8-056? Water . . 4-454 It will be obvious to the reader that 5 is the number which denotes the atomic weight of sul- phuric acid ; 4*454 is so near 4%5, which repre- sents the weight of four atoms of water, that we Gilbert's Annalen, LX. 202. Bb3 390 ZINC, CADMIUM, LEAD, &c. can have no doubt that four atoms is the exact quantity of water in an integrant particle of the salt. It is more than probable that the 0*046, which is wanting to complete the weight of the four atoms of \\toter, has been given to the oxide of cadmium. Hence, it is likely that the constituents of the salt should have been stated as follows : 1 atom sulphuric acid 5 1 atom oxide of cadmium = 8*0107 4) atoms water 4-5 8*0107 denote the weight of an atom of oxide of cadmium, according to this analysis a number approaching 8 so nearly, that we can have no hesitation in making choice of 8 as the true ato- mic weight of this oxide. For it is obvious from the great number of atoms already investi- gated, as well as to be investigated, that all the atomic weights of simple bodies are multiples of 0-&5. Now, 8 is such a multiple ; while neither 8-0567 nor 8-0107 is so. To determine the composition of sulphate of cadmium with the requisite exactness, I pre- pared a quantity of very pure crystals, by crys- tallizing a rather concentrated solution of the salt under the exhausted receiver of the air- pump over sulphuric acid. The crystals were large, flat, four-sided prisms, apparently rectangu- lar, but from their efflorescing property, I was not able to measure their angles by the reflecting goniometer. In most of them, the edges were CADMIUM. 391 truncated, rendering the prism hexagonal, and it was terminated at both extremities by an elon- gated hexagon. These crystals, after four suc- cessive crystallizations, still reddened vegetable blues. Hence, I conclude that oxide of cad- mium, like oxide of zinc, js incapable of com- pletely neutralizing sulphuric acid. The taste of the salt has some resemblance to that of sul- phate of zinc ; but it is rather more acrid and disagreeable. Guided by the analysis of Stromeyer given above, I dissolved 17*5 grains of these crystals, dry, but quite free from efflorescence, in water, and mixed the solution with one of 13'25 grains of chloride of barium. After the sulphate of barytes had subsided, the supernatant liquid was tested by sulphate of soda and muriate of ba- rytes, but it was not affected by either. From this we see that 17'5 grains of crystallized sul- phate of cadmium contain just 5 grains of sul- phuric acid. 17*5 grains of the same crystals, heated very cautiously to redness in a platinum crucible, lost just 4-5 grains, which is equivalent to 4 atoms of water. The acid and water in 17*5 grains of sulphate of cadmium amounting to 9*5 grains, it is obvi- ous, that the 8 grains wanting to make up the whole weight must be oxide of cadmium ; and that the true composition of the salt is 392 ZINC, CADMIUM, LEAD, &c. 1 atom sulphuric acid 5 1 atom oxide of cadmium 8 4 atoms water . 4-5 17'5 I placed an exceedingly concentrated solution of sulphate of cadmium, along with sulphuric acid, under the exhausted receiver of an air pump. Instead of crystallizing, the liquid de- posited a white, hard, shapeless crust, similar in appearance to enamel. When a solution of 15-25 grains of this crust was mixed with a solution of 13'25 grains of chloride of barium, a double de- composition took place; and after the sulphate of barytes had subsided, the supernatant liquid was tested with sulphate of soda and muriate of barytes, but was not affected by either. Hence the constituents of this crust are 1 atom sulphuric acid 5 1 atom oxide of cadmium 8 2 atoms water . 2 '2 5 15-25 Sulphate of zinc has the property of being de- posited in the same way, as was shown in the first section of the third chapter of this work. Several of the other salts of cadmium were likewise analyzed. The results, which will be seen in a subsequent chapter, all coincide to CADMIUM. 393 show that 8 is the atomic weight of oxide . oxide of Cadmium. cadmium. As cadmium unites with only one dose of oxygen, it is obvious that the oxide is a com- pound of one atom cadmium and one atom oxy- gen j consequently, the atomic weight of cad- At <>m of mium must be 7- This is corroborated by Strom- eyer's experiments. He found that 100 cad- mium, when oxydized, were converted into 114*352 of oxide of cadmium. Therefore, oxide of cadmium is composed of Cadmium 100 or 6'96?7 Oxygen 14-352 - 1 6*9677> the atomic weight of cadmium, deduced from this experiment, approaches very nearly to 7, the number deduced from the sulphate. That it should be rather less than 7 is not surprizing 5 for cadmium is so volatile a metal, that it is not easy to expose it to heat, without some loss. Stromeyer likewise formed and analyzed some alloys of cadmium. They deserve our attention, because they serve to corroborate the accuracy of the atomic weight which has been assigned to cadmium. The alloy of cadmium and platinum is a com- Alloy of ' cadmium Of and plati- num. Platinum 46-02 or 12 Cadmium 53-98 - 14-07 100-00* * Gilbert, LX. 209. 394 ZINC, CADMIUM, LEAD, &c. It will appear hereafter, that the 12 represents the atomic weight of platinum. 14'07 must de- note two atoms of cadmium 5 for the alloy con- sists obviously of 1 atom of platinum, united to 2 atoms of cadjnium. The atom of cadmium, deduced from this alloy, is 7'035. Amalgam The amalgam of cadmium is composed of of cadmi- um. Cadmium 21-7391 or Mercury 78-2609 - 25 100 25 is the atomic weight of mercury, so that the amalgam is a compound of 1 atom mercury, and 1 atom cadmium. The atomic weight of cad- mium, deduced from this amalgam, is 6'944. The atom of cadmium, deduced from a mean of these two alloys, is 6*9895 ; and if we deduce the atomic weight of this metal from a mean of these four sets of experiments of Stromeyer, which is the most likely way to come at the truth, we obtain 7*00085. Now, this approaches so very near 7, that the most scrupulous chemist can- not hesitate about adopting 7 as the true atomic weight of this metal ; so that the experiments of Stromeyer, even without mine, are sufficient to establish that important point. Thus it appears that an atom of cadmium weighs 7> and an atom of oxide of cadmium 8. Sulphuret of cadmium, when digested in nitric LEAD. 395 acid, is converted into sulphate of cadmium. It is, therefore, a compound of one atom sulphur and one atom cadmium, and its atomic weight is 9. SECT. II. OF THE ATOMIC WEIGHT OF LEAD. I HAVE already in a former section of this work described an experiment, from which the atomic weight of protoxide of lead is easily deduced. Nitrate of lead crystallizes in white opaque oc- tahedrons, and is quite anhydrous. 2O75 grains of it are just decomposed by 11 grains of anhy- drous sulphate of potash ; and the sulphate of lead obtained weighs exactly 19 grains. Now, these 19 grains contain 5 grains of sulphuric acid ; so that the salt must be a compound of Sulphuric acid . 5 Protoxide of lead 14 19 The salt being neutral, and 5 representing the A / o t xi f eof weight of an atom of sulphuric acid, 14 must lead. represent the weight of an atom of protoxide of lead. If it be a compound of 1 atom lead and 1 Atom of atom oxygen, then an atom of lead weighs 13 ; e and that 13 is the true weight of an atom of lead 396 ZINC, CADMIUM, LEAD, c. will appear from the following experiment : 15 grains of pure galena or sulphuret of lead were reduced to powder, and digested with nitric acid in a small retort, the weight of which had been previously ascertained, and marked upon it with a diamond. By this process the galena was gra- dually converted into sulphate of lead. When the conversion was completed, the nitric acid was distilled off, and the retort was kept in a heat of above 600, till it became quite dry in the inside ; it was then allowed to cool, and the weight of the sulphate of lead thus formed was exactly 19 grains. Now, 5 of these grains are sulphuric acid, containing 2 grains of sulphur ; consequently, 2 represents the quantity of sul- phur in 15 grains of sulphuret of lead. The re- mainder of the weight being lead, it is obvious that sulphuret of lead is composed of Lead . . 13 Sulphur . . 2 15 2 represents the weight of an atom of sulphur, and 13 is obviously the atomic weight of lead. 2. Besides the protoxide, or yellow oxide of lead, there are two other oxides of this metal ; namely, the red oxide, or deutoxide, and the Com osi b rown oxide, or peroxide. tionofred The beautiful pigment called red lead, is a LEAD. 397 mixture of deutoxide and protoxide in variable proportions. In some very beautiful red lead, I have found the proportion of protoxide amount to nearly one-half of the whole weight. If we digest red lead in acetic acid, we dissolve the protoxide, and leave the deutoxide untouched ; it has a dark red colour, and is destitute of the beauty which characterizes red lead. Various experiments have been made to determine the additional quantity of oxygen necessary to con- vert the protoxide of lead into deutoxide. I made a set of experiments on the subject about twenty years ago ;* but as I was not aware at that time, of the mixture of protoxide present in the pigment, I did not, of course, attempt to separate it, and consequently underrated the proportion of oxygen which it contained. Even Berzelius, though aware of the existence of the protoxide of lead in red lead, does not seem to have succeeded in removing the whole of it. But the proportion of oxygen, which he assigns to this oxide, though below the true quantity, affords a sufficient ap- proximation to enable us to determine the com- position of the deutoxide of lead. 9'835 parts of deutoxide, when heated to red- ness, gave out 0*29 parts of oxygen, and left 9 -545 parts of protoxide of lead.t According to * Nicholson's Jour. VIII. 289. f Afhandlingar, III. 168, or Ann. de Chim. LXXV.TII. 14. But there is a typographical error in the French translation. 398 ZINC, CADMIUM, LEAD, &c. this experiment deutoxide of lead is composed of Protoxide of lead 9-545 or 14 Oxygen . 0-290 - 0-43 T^ foUwing.experiment, which I made with great care, will show that the true composition of deutoxide of lead is 14 protoxide 4- 0-5 oxy- gen. 29 grains of deutoxide of lead were di- gested in dilute, but pure nitric acid, till all ac- tion was at an end ; the whole was then thrown on a double filter, and the peroxide of lead remaining undissolved was well washed and thoroughly dried in a temperature of nearly 500 it weighed 15 grains. Hence it follows, that the protoxide dissolved by the acid weighed 14 grains. Thus, 29 grains of deutoxide of lead were resolved into Protoxide of lead . . 14 Peroxide of lead . . 15 29 But it will be shown immediately that 15 repre- sents the atomic weight of peroxide, and that it is a compound of 1 atom lead and 2 atoms oxy- gen. Protoxide of lead is a compound of 1 atom lead and 1 atom oxygen : hence it follows, that 29 deutoxide of lead are composed of 2 atoms lead =26 or 1 atom lead 13 3 atoms oxgen = 3 - ] atom oxygen 1*5 29 14-5 LEAD. 399 Deutoxide of lead does not seem capable of entering into any chemical combination ; it is, therefore, in the present state of our knowledge, comparatively insignificant as a chemical sub- stance. 3. The peroxide of lead is a flea brown Atom of coloured powder, which was discovered by lead. Scheele. It may be obtained by digesting ni- tric acid on red lead. A portion of the deutox- ide is reduced to the state of protoxide, and dis- solves in the acid, while the oxygen given off by it uniting to another portion converts it into peroxide. If the deutoxide be pure, it is evident from the experiment stated above, that just one half of the lead is dissolved and the other half converted into peroxide. I found long ago, that 100 grains of this ox- ide, when exposed to a red heat, left 91 grains of protoxide.* Now, if the 9 grains of loss be considered as oxygen, it will follow from this ex- periment, that peroxide of lead is composed of 100 lead + 18*34 oxygen, or of lead 13 + oxygen ^384. But when I made these experiments I was not aware that metallic oxides are apt to re- tain water : the peroxide which I employed had been merely dried in the open air, and conse- quently, was not quite anhydrous. The loss of weight was partly water and partly oxygen ; * Nicholson's Journal, VIII. 288. 400 ZINC, CADMIUM, LEAD, &C. consequently, my estimate of the quantity of oxygen in this oxide was too high. Berzelius dried the peroxide of lead in a tem- perature at which tin melted. He found, that 5 parts of the Qxide thus dried, lost, when ex- posed to a red heat, 0-325 parts ; the residue, which weighed 4 '675 parts, being digested in acetic acid, left 0-13 parts of sulphate of lead and silica.* It follows from this experiment, that peroxide of lead is composed of Protoxide of lead 4'545 or 14 Oxygen . 0-325 - 0-973 4-87 No doubt the true composition is Protoxide of lead . . 14 Oxygen .... i 15 The atomic weight is 15, and the oxide is a compound of 1 atom lead =13, and 2 atoms oxygen = 2. Thus, the constituents of the three oxides of lead are as follows : LEAD. OXYGEN. ATOMIC WEIGHT. Protoxide 1 atom + 1 atom . . 14 Deutoxide 1 + 1J . . 14-5 Peroxide 1 + 2 ... 15 * Afhandlingar, III, 170., or Ann. de Chim. LXXVIII, 16. The French translation is unintelligible, because a sentence is left out. TIN. 401 4. When thin plates of pure lead are intro- chloride of duced into dry chlorine gas, the gas is gradually absorbed and the lead converted into a white matter, which is a chloride of lead. This chlo- ride is easily obtained by mixing together solu- tions of nitrate of lead and of common salt : small brilliant crystals precipitate, which, when heated in a close vessel, melt and assume the form of a greyish matter, formerly distinguished by the name of plumbum corneum. I find, by a careful experiment, that 13 grains of lead, when converted into chloride, weigh 17*5 grains. Hence, it is obvious that this chloride is com- posed of 1 atom lead . . 13 1 atom chlorine . 4-5 17-5 SECT. III. OF THE ATOMIC WEIGHT OF TIN. TIN is a metal which forms two different oxides, Oxides of when it combines with oxygen. The protoxide ir may be obtained by dissolving tin in strong mu- riatic acid (taking care to exclude all access of atmospheric air to the inside of the small retort in which the process is going on). When the muri- atic acid is saturated with the tin, the liquid must Voi. I. Cc 402 ZINC, CADMIUM, LEAD, c. be poured at once into a glass vessel containing as much caustic potash or soda as will saturate the whole muriatic acid in combination with the protoxide. A black powder falls, which is prot- oxide of tin ; it may be separated by a filter, edul- corated and dried. When in combination with acids, this oxide has a great tendency to combine with more oxygen, and to be converted into peroxide ; but when in a separate state, and dry, I have kept it for years in a common phial, without observing any alteration in its colour. The peroxide of tin has a yellowish colour. It may be obtained by digesting tin in nitric acid, and gradually raising the temperature to redness. The acid is decomposed and driven off, while the peroxide of tin remains behind in a state of purity. Both of these oxides form salts with muriatic acid ; but my attempts to determine the atomic weight of tin, by a rigid analysis of these salts, did not succeed. The following experiment, however, which I have several times repeated, gives the composition of peroxide of tin, from which it is easy to deduce the atomic weight of this metal. Atom of 7*25 grains of pure tin were put into a plati- num crucible, and dissolved in a sufficient quan- tity of very dilute nitric acid. The solution was slowly evaporated to dryness, and the crucible gradually raised to a red heat. The peroxide of TIN. 403 tin formed in this way weighed 9*25 grains ; hence it is evidently a compound of Tin . 7-25 Oxygen . 2 9-25 But 2 denotes 2 atoms of oxygen ; consequent- ly, 7*^5 must be the weight of an atom of tin. Peroxide of tin obviously weighs 9 '25, and is a compound of 1 atom of tin and 2 atoms of oxy- gen. 2. The experiments of Berzelius* leave no Atom . <> f doubt that the protoxide of tin contains just half the oxygen in the peroxide. It is, therefore, a compound of 1 atom tin and 1 atom oxygen, and its atomic weight is 8 '25. 3. Tin combines with two proportions of chlo- Chlorides rine, as well as of oxygen, and forms two chlo- rides. The protochloride is most easily formed pure, by amalgamating together tin and mer- cury, mixing the amalgam with a sufficient quan- tity of calomel ; and exposing it in a glass tube to a heat at first low, but gradually raised high enough to expel all the mercury and the calo- mel. J'%5 grains of tin treated in this way, left 11'75 grains of protochloride. Hence this chlo- ride is obviously composed of Nicholson's Journal, XXXV. 122. Cc2 404 ZINC, CADMIUM, LEAD, & C . 1 atom tin . . 7'25 1 atom chlorine . 4-5 11-75 This chloride is a grey matter, having a resinous lustre and fracture. When put into chlorine gas, it catches fire, and is converted into per- chloride. Fuming H- 4. The perchloride of tin has been long known '" under the name of fuming liquor of Libavius. It may be obtained by mixing together amalgam of tin and corrosive sublimate, and distilling with a very moderate heat. A colourless liquid comes over at first, consisting chiefly of water ; then the fuming liquor rushes all at once with such rapi- dity, that I have more than once lost the whole product of the distillation. Dr. John Davy found that this perchloride may be obtained likewise by mixing together concentrated permuriate of tin and sulphuric acid, and distilling with a gentle heat. This perchloride gives out very copious and dense fumes when exposed to the air, in conse- quence of its great avidity for moisture. Adet found, that when three parts of it are mixed with one part of water, the whole condenses into a solid state. I see no reason to doubt that this perchloride, when in a state of purity, is a com- pound of 1 atom tin and 2 atoms chlorine ; and that its atomic weight is 16-&5 ; but I have not TIN. 405 been able to verify this opinion by direct analy- sis. I made three experiments upon fuming li- quor of Libavius, prepared very carefully in my laboratory. The quantity of chlorine found in each did not exactly correspond ; but the mean of all the three gave for the constituents of the chloride 1 atom tin . . 7'25 2 J atoms chlorine 9*112 16-362 The analysis of this compound by Dr. John Davy* gave very nearly the same result ; or, rather, he found a somewhat greater proportion of chlorine, viz. 9*18 parts united to 7*%5 of tin. It would seem from these analyses, that fuming liquor of Libavius, when prepared by means of amalgam of tin and corrosive sublimate, is apt to contain an excess of chlorine. I find that permuriate of tin in crystals is composed of 1 atom peroxide of tin . 9-25 2 atoms muriatic acid . 9-25 3 atoms water . . 3-375 21-875 When this salt is mixed with sulphuric acid, and * Phil, Trans. 1812, p. 177. Cc3 406 ZINC, CADMIUM, LEAD, &c. distilled by a gentle heat, the perchloride ob- tained would probably be neutral ; but I have not tried to prepare it in this way. of u !fn urets 5 ' The sul P nurets *' tin > which are two in number, likewise furnish decisive evidence of the atomic weight of tin. The protosulphuret of tin may be obtained by melting together tin and sulphur in a glass tube, and keeping up the heat till the excess of sul- phur is sublimed. When 7*25 grains of pure tin are thus converted into sulphuret, the weight of the new compound is 9*25. Hence it is a compound of 1 atom tin . . 7-25 1 atom sulphur . 2 9-25 The persulphuret has been long known by the name of Mosaic gold. It is under the form of light scales, of a brownish yellow colour, and great beauty and lightness. It is soluble in ni- tromuriatic acid, and likewise in potash ley. The analysis of this sulphuret of tin by Dr. John Davy is sufficient to prove that it is a compound of r atom tin . . 7-25 2 atoms sulphur . 4 11-25 BISMUTH. 407 for he found the constituents by analysis to be* Tin . 100 or 7'25 Sulphur . 56-25 - 4-078 SECT. IV. OF THE ATOMIC WEIGHT OF BISMUTH. IN order to determine the atomic weight of bis- muth with accuracy, it is requisite to employ that metal in a state of purity. The bismuth of commerce is usually contaminated with iron, and almost always contains some sulphur. To obtain pure bismuth, I dissolved a quantity of the bis- muth of commerce in nitric acid. The crystals of nitrate of bismuth being thrown into water, were decomposed, and a white powder, consist- ing chiefly of hydrated oxide of bismuth, fell to the bottom. This hydrate was well washed, dried upon the filter, and reduced to the me- tallic state, by heating it in a covered crucible with black flux. 1. Nine grains of bismuth purified in this way, Atom of were put into a platinum crucible, and dissolved bismuth, in nitric acid. The solution was evaporated to dryness by a gentle heat, which was afterwards gradually raised till the crucible became percep- * Phil. Trans. 1812, p. 199. 408 ZINC, CADMIUM, LEAD, &c. tibly red hot. By this means, the acid was driven off, leaving behind it the bismuth reduced to the state of an oxide. It weighed exactly 10 grains, and consequently, the metal, when oxy- dized, had combined with one-ninth of its weight of oxygen. Hence, oxide of bismuth is com- posed of Bismuth 9 Oxygen 1 10 But 1 being the weight of an atom of oxygen, it is obvious that 9 must be the atomic weight of bismuth ; and the atomic weight of oxide of bismuth is 10. This metal combines with only 1 atom of ox- ygen, and forms only one oxide. chloride of 2. The chloride of bismuth may be formed by mixing together bismuth in powder, with corrosive sublimate in a glass tube, and exposing the mixture to a heat, at first low, but gradually raised to the boiling point of mercury, or a little higher. The excess of corrosive sublimate and the reduced mercury, are gradually volatilized, leaving the chloride of bismuth in a state of purity. It is a solid substance, having a granu- lar texture, a greyish white colour, and a silky lustre. I mixed 9 grains of pure bismuth in powder, with 40 grains of corrosive sublimate in a test tube. This test tube, with a cork fitted BISMUTH. 409 to it, was left for three or four days on the sand bath ; at first the heat applied did not exceed that of boiling water, but the tube was brought gradually towards the hottest part of the sand bath, till the temperature was at last between 600 and 700. Finally, it was raised high enough to drive off all the mercury and the cor- rosive sublimate which had escaped decomposi- tion ; the weight of the chloride of bismuth was within Tsth of a grain of 13'5 grains. The loss of roth of a grain, I ascribe to the volatilization of a small portion of the chloride, which it is very difficult to avoid. If we compare this ex- periment with the preceding, it will be obvious that the chloride of bismuth is a compound of 1 atom chlorine 4*5 1 atom bismuth 9 13-5 3. The atomic weight of bismuth deduced from the oxydation of the metal, is still farther confirmed by the sulphuret. This sulphuret occurs native, and may be easily formed artifi- cially, by heating together bismuth and sulphur, and keeping up the heat till the excess of the sulphur is driven off. Nine grains of pure bis- muth converted into a sulphuret in this way, in a green glass tube, became eleven grains of sulphuret, showing that the sulphuret is a com- pound of 410 ZINC, CADMIUM, LEAD, c. 1 atom bismuth 9 1 atom sulphur 2 11 The previous analyses of sulphuret of bismuth by Dr. John Davy and Mr. Lagerhjelm almost coincide with this experiment of mine. Dr. John Davy * found it a compound of Bismuth 9 Sulphur 2-007 M. Lagerhjelm obtained t Bismuth 9 Sulphur 2-0268 Dr. John Davy's experiment was conducted nearly in the same way as my own : he em- ployed 100 grains of bismuth I presume the bismuth of commerce. Now, the presence of a very small portion of iron in this bismuth, would account for the slight excess of sulphur, which scarcely exceeds one third per cent. La- gerhjelm's analysis is not quite so near the truth as Dr. Davy's. 4f. Vauquelin, in a set of experiments, pub- lished a number of years ago, upon the quantity of sulphur with which the metals combine when heated or fused with that substance, states that he found sulphuret of bismuth composed oft * Phil. Trans. 1812. p. 200. f Annals of Philosophy, IV. 357. Ann. de Mus. d'Hist. Nat. XV. 9. COPPER. 411 Bismuth 68-25 or 9 Sulphur 31-75 - 4-186 100-00 If this sulphuret of Vauquelin exist, (for I have not succeeded in my attempts to form it artifi- cially,) it is evidently a bisulphuret, or a com- pound of two atoms sulphur with one atom of bismuth. SECT. V. OF THE ATOMIC WEIGHT OF COPPER. COPPER forms two oxides ; the red oxide, which Oxides of occurs native, and does not seem capable of combining with acids; and the black, which constitutes the basis of all the cupreous salts. It was shown long ago by Chenevix, and his experiments have been amply confirmed by succeeding researches, that the black oxide con- tains exactly twice as much oxygen as the red ; and I shall show immediately, that if we reckon the weight of an atom of copper to be 4, then the composition of these oxides is as follows : COPPER. OXYGEN. ', 1. Red oxide 8 + 1 or 2 atoms copper + 1 atom oxygen. 2. Black oxide 4 + 1 or 1 atom copper + 1 atom oxygen. I shall therefore distinguish the black oxide by 412 ZINC, CADMIUM, LEAD, &c. the name of oxide of copper ; while to the red oxide we shall apply the term suboxide, indicat- ing that it is a compound of two atoms copper and one atom oxygen. Such suboxides in gen- eral are incapable of constituting permanent salts with acids. As the black oxide of copper constitutes the basis of all the cupreous salts, the knowledge of its atomic weight is of most importance in a chemical point of view. We may employ the common sulphate of copper, as a good instru- ment for determining the atomic weight of this oxide ; because being manufactured to a con- siderable extent, and employed both in medi- cine and the arts, it may be easily procured in a state of purity. It has a fine blue colour, an ex- ceedingly disagreeable cupreous taste, reddens vegetable blues, and crystallizes in oblique four- sided prisms, the faces of which are inclined to each other at angles of 124 and 56. Analysis of 1. When a solution of 15*625 grains of these crystals in water, is mixed with a solution of 13*25 grains of chloride of barium, a double de- composition takes place, sulphate of barytes pre- cipitates, and the supernatant liquid is neither affected by sulphate of soda nor muriate of ba- rytes. Hence, it follows, that 15*625 grains of this salt contain just 5 grains of sulphuric acid. 2. If the whole oxide of copper be precipita- ted from 15*625 grains of blue vitriol, and col- COPPER. 413 lected without loss, it will weigh exactly five grains. This experiment is not easily made, be- cause we have no good precipitant of oxide of copper. But I have accomplished it in various ways. If we throw down the oxide of copper by caustic potash, we obtain very nearly five grains, though I have never got the exact amount in this way. In general, the oxide of copper thus procured weighs about 4-9 or 4'95 grains. When the copper is thrown down by sulphuret- ted hydrogen, or by hydrosulphuret of ammonia, the precipitate, when collected without loss, and sufficiently dried, weighs 8 grains, and when dis- solved in aqua regia, and precipitated by muriate of barytes, the sulphate of barytes obtained weighs 29*5 grains, indicating 10 grains of sul- phuric acid, equivalent to 4 grains of sulphur ; consequently, the copper in 15 '625 grains of blue vitriol weighs exactly 4 grains.* This, coupled with the amount of the precipitate by means of caustic potash, leaves no doubt that the true weight of the oxide of copper in 15*625 grains of blue vitriol is 5 grains ; and that it is a compound of * If we precipitate the copper from 15-625 grains of sulphate of copper by means of a plate of zinc, taking care in the first place, to acidulate the solution of the salt with a little sulphuric acid, to prevent the copper from adhering to the zinc ; the precipitated copper, if it be dried without allow- ing it to absorb oxygen, will weigh exactly four grains. This experiment is not easily made, because the copper while drying has a tendency to absorb oxygen. But it is easy in this way to obtain a very near approxi- mation. 9 414 ZINC, CADMIUM, LEAD, &c. Copper 4 or 1 atom Oxygen 1-1 atom 3. The only other constituent of blue vitriol being water, it is obvious that the salt is a com- pound of 1 atom sulphuric acid 5 1 atom oxide of copper 5 5 atoms water 5-625 15-625 Disuiphate 4. There exists another sulphate of copper, which I formed by the following process: 31*25 grains of blue vitriol were dissolved in water in a retort, and the solution was mixed with 10 grains of peroxide of copper. The retort was put upon the sand bath in such a position that it was kept boiling several hours every day, taking care to add fresh water in proportion as that liquid was dissipated ; and this was continued till all the oxide of copper had totally disap- peared. This happened in about a couple of months ; at the same time the liquid lost its blue colour, while a green coloured powder was de- posited. The liquid being tested for copper by ammonia, was found to contain none. It is ob- vious, therefore, that the green powder contained all the acid and oxide of copper in 31 '25 grains of blue vitriol, together with the additional 10 grains of oxide added. The green powder, there- fore, was a compound of COPPER. 415 Sulphuric acid 10 or 5 Oxide of copper 20 -JO This powder was tasteless, insoluble in water, and produced no effect on vegetable blues. It was obviously a compound of 1 atom acid, and 2 atoms oxide, or it was a disulphate of copper. 5. Sulphuret of copper is easily formed by heating together copper filings and sulphur, per. The combination takes place suddenly when the mixture has reached a certain temperature, and the sulphur from being in a liquid state becomes suddenly solidified. This explains the reason of the ignition which takes place, and which at- tracted so much of the attention of the Dutch chemists. The first person who attempted to determine the composition of this sulphuret with the requisite precision was Berzelius. He found that 10 parts of copper, when converted into sul- phuret, became 12*56 parts.* Now, 10 : 2-56 : : 8 : 2-048 ; so that according to this experiment, sulphuret of copper is composed of Copper 8 Sulphur 2-048 The copper of commerce, which Berzelius pro- bably employed in his experiment, is not quite pure. I believe it is often contaminated with some protoxide of copper, and always with some other Af handliugar, III. 187. J6 416 ZINC, CADMIUM, LEAD, &c. metals. I prepared some very pure copper on purpose, and found that 8 grains of it, when con- verted into sulphuret in a green glass tube, be- came exactly 10 grains. Hence, it follows that this sulphuret i a compound of Copper 8 Sulphur 2 10 2 is an atom of sulphur and 8 is two atoms copper. So that this compound is a disulphuret of copper. Besides these two sulphurets of copper, a of copper. . A f A J third seems to exist in combination with proto- sulphuret of iron ; for M. Henry Rose found copper pyrites to be a compound of Copper 34-4 Iron 30-47 Sulphur 35-8? Silica 101-01* The excess of 1 per cent, was probably in the sulphur. But let us suppose the result to be exact, the above numbers are equivalent to * Ann. de Chim. et de Phys. XXV. 197. COPPER. 417 Copper 4-000 or 1 atom Iron 3'543 - 1 atom Sulphur 4-170 - 2 atoms It is obvious, that copper pyrites is a compound of 1 atom sulphuret of copper 6 1 atom sulphuret of iron 5-5 Thus, there are three sulphurets of copper ; namely, COPPER. SULPHl'R. 1. Disulphuret composed of 2 atoms + 1 atom. 2. Sulphuret 1 atom + I atom. 3. Bisulphuret 1 atom -f- 2 atoms. 6. If 100 grains of peroxide of copper be Atom of dissolved in muriatic acid, and 80 grains of cop- copper. per filings be added to the solution, and the whole agitated in a well stopped phial nearly filled by the mixture, the copper filings are in- stantly dissolved without the evolution of any gas, and the liquid, which has become black and opaque, is found to contain a solution of sub- oxide of copper in muriatic acid. This experi- ment, for which we are indebted to Mr. Chenevix, but which I have frequently repeated, demonstrates the composition of suboxide of copper : for 1 00 grains of oxide of copper are composed of Vor,. I. D cl 418 ZINC, CADMIUM, LEAD, &c. Copper 80 Oxygen 20 100 So that by the addition of 80 grains, the copper in the liquid is exactly doubled, while the oxy- gen with which it is combined remains the same as at first. Hence, suboxide of copper is ob- viously composed of Copper 160 or 16 or 8 Oxygen 20 - 2-1 It is a compound of 2 atoms copper and 1 atom oxygen, and its atomic weight is 9 Chloride of 7. Chlorine and copper combine in two pro- portions, and form two chlorides analogous in their composition to the two oxides of copper. The perchloride is easily obtained by dis- solving peroxide of copper in muriatic acid, and concentrating the green coloured solution. Permuriate of copper is formed, and may be ob- tained in long four-sided rectangular prisms with square bases. This salt I find, by a careful ana- lysis, is composed of 1 atom muriatic acid 4 '62 5 1 atom oxide of copper 5 2 atoms water . 2-25 11-875 COPPER. 149 When 11*875 grains of this salt are exposed to a temperature of between 400 and 500, or in- deed to a lower temperature, there is a diminu- tion of weight amounting to exactly S'SJS grains, the green colour disappears, and an orange brown matter remains, weighing exactly 8*5 grains. Now, the 3-3J5 grains of matter driven off consist of pure water, as may be ascertained by performing the experiment in a retort, to which a receiver is luted. The water will come over into the receiver quite pure, unless too much heat has been applied. In that case some chlorine is apt to pass off, owing I presume to the agency of the common air in the retort. 3*375 grains of water is the equivalent for 3 atoms of that liquid ; yet the salt contained only 2 atoms. The other atom of water was formed by the union of the atom of oxygen in the oxide with the atom of hydrogen in the muriatic acid. Thus the salt is changed into chloride of copper, which, as is obvious from the facts just stated, is a compound of 1 atom copper . 4 1 atom chlorine 4'5 8-5 8. The subchloride of copper was noticed by . 11 -i-ii ride of cop- Boyle ; and it has been minutely described by per. Prout, Chenevix, and Dr. John Davy. I have Dd2 420 ZINC, CADMIUM, LEAD, &C. not made any attempts to analyze it, because the previous analysis of Dr. Davy is perfectly accu- rate. He found it composed of atoms copper atom chlorine 8 4-5 12-5 The constituents of the chlorides and sulphuret of copper, compared with those of the oxides, leave no room for doubt that the true atomic weight of this metal is 4.* * It is hardly necessary to remark, that the atom of copper might be represented by 8, as well as by 4-. In that case, red oxide of copper would be composed of 1 atom copper = 8 1 atom oxygen = 1 ind black oxide of The chlorides would be : protochloride I atom copper = 8 2 atoms oxygen = 2 10 1 atom copper = 8 1 atom chlorine = 4*5 12-5 perchloride 1 atom copper = 8 2 atoms copper = 9 17 MERCURY, SECT. VI. OF THE ATOMIC WEIGHT OF MERCURY. IT has been ascertained by unexceptionable ex- periments, that mercury combines with two doses of oxygen, and forms two oxides. The oxides of protoxide is black, and the peroxide red. The mercury volatility of mercury and its compounds presents and the sulphurets : protobulphuret of bisulphuret of pcrsulphuret of I atom copper 1 atom sulphur 1 atom copper 2 atoms sulphur 8 8 10 8 i 12 1 atom copper = 8 4 atoms sulphur = 8 16 This was the idea I originally formed of these compounds, and it seems to be the one generally adopted by chemists. It was the consideration of the salts of copper that induced me to adopt 4< for the atomic weight of copper ; for, if we represent the atom of copper by 8, all the salts of copper, without exception, will be bisalts, or will contain 2 atoms of acid, united to 1 atom of oxide. It was an object of some importance to get rid of this anomaly, which involved the salts of copper in a good deal of confusion. Tlxis was accomplished by adopting 4 for the atomic weight of copper, as had been already done by Dr. Wollaston and Dr. Prout. D d 3 ZINC, CADMIUM, LEAD, &c. a formidable obstacle to the rigid determination of their atomic weights. Deter- 1. Fourcroy, long ago, stated, that the pro- toxide of mercury is a compound of 100 metal and 4 oxygen ; but, from the loose way in which this chemist was accustomed to express himself, we can consider this determination in no other light but that of a fortunate conjecture. Mr. Donovan, Donovan, from a number of experiments on both oxides, has concluded that protoxide of mercury is Composed of Mercury . 100 or 2 4-2 71 Oxygen . 4-12 - I* And peroxide of Mercury 100 or 25-575 Oxygen . 7'82 - 2 The atomic weight of mercury, deduced from these two oxides, not being exactly the same, it is obvious that the experiments from which the constitution of these oxides has been deduced, are not quite exact. The mean of both gives us 24 '923 for the atomic weight, which differs less than ^joth from 25, the atomic weight of mer- cury, deduced from the composition of protoxide of mercury, as stated by Fourcroy. M. Sefstrom made a set of experiments in Ber- and Sef- strom. * Annals of Philosophy, XIV. MERCURY. 423 zelius 5 laboratory on the composition of the oxides of mercury, with the results of which only I am acquainted.* He found the constitu- ents of protoxide of mercury Mercury . 100 or 25-063 Oxygen . 3-99 - 1 and those of red oxide Mercury . 100 or 25-031 Oxygen . 7'99 - 2 If we take the mean of these two analyses, which do not exactly agree with each other, though they come much nearer than those of Donovan, we obtain 25-047 for the atomic weight of mer- cury, which exceeds 25 by less than jioth part. Now, as the atomic weight of mercury, as de- duced from Donovan's experiments, is only a very little less than 25, while the same weight, deduced from the experiments of Sefstrom, is a very little more than 25, the inference from these experiments seems to be, that an atom of mercury weighs exactly 25. 2. The experiments of Donovan and Sefstrom leave no doubt that the oxygen in the peroxide of mercury is just double that in the protoxide. It is, therefore, composed of 1 atom mercury and 2 atoms oxygen ; and its atomic weight is Annals of Philosophy, III. 355. Dd4- ZINC, CADMIUM, LEAD, &c. 27 j it' we reckon the atomic weight of mercury 25. The truth of this is sufficiently established by the following experiment. menteof ^ grain s of pure peroxide of mercury, pre- the author, pared on purpose for the experiment, were put into an eight-ounce phial, and as much muriatic acid poured upon it as dissolved it completely. The phial was then placed upon the sand bath (having a charcoal stopper fitted into its mouth), and left till the liquid was evaporated to dryness. The phial was then covered about two-thirds up with sand, and the temperature was gradually raised sufficiently high to sublime the corrosive sublimate, into which the red oxide had been converted, as high as the middle of the phial. The phial, after being allowed to cool gradually, was weighed, and the weight of the corrosive sublimate was exactly 34 grains. From this experiment it appears, that 27 per- oxide of mercury form exactly 34 corrosive sub- limate. To understand this experiment it is necessary to know, that mercury combines with two pro- portions of chlorine, and forms two chlorides. The protochloride, usually denominated calomel, is a compound of 1 atom mercury and 1 atom chlorine ; while the perchloride, or corrosive sub- limate, is a compound of 1 atom mercury and 2 atoms chlorine. With the knowledge of these facts, we may MERCURY. 425 now analyze the conversion of 27 grains of per- oxide of mercury into 34 grains of corrosive sub- limate. 27 peroxide are composed of Mercury . 25 Oxygen . 2 27 When muriatic acid is poured upon the 27 per- oxide, two atoms of the acid, equivalent to 9 '25 grains, unite with the oxide, forming a com- pound, which will amount to 36'25 grains, and be a compound of 1 atom peroxide 27 2 atoms muriatic acid 9*25 36-25 When the solution is evaporated to dryness, the acid and peroxide mutually act upon each other : the oxygen of the oxide combines with the hy- drogen of the acid, constituting 2 atoms of wa- ter, which are expelled by the heat ; while the 2 atoms of chlorine combine with the 1 atom of mercury, and form a perchloride. Thus by the heat, the 36-25 of permuriate of mercury is re- solved into Oxygen 2 \ > constituting two atoms water. Hydrogen 0-25 j 2-25 426 ZINC, CADMIUM, LEAD, &c. Mercury 251 ,. ., > constituting an atom perchloride. Chlorine 9 / 34 It is clear, therefore, that 27 peroxide of mer- cury could not oecome 34 corrosive sublimate, Atom of unless the atomic weight of mercury were 25, mercury. and unless peroxide of mercury were composed of 1 atom mercury and 2 atoms oxygen. 3. The analysis of calomel by Chenevix is so conclusive, that I did not consider it at first as worth while to make any researches on the sub- ject. But it may be satisfactory to mention the Atom of following experiment, which was made with an- other object in view. A quantity of crystallized protonitrate of mercury, which I knew to contain precisely 25 grains of mercury, was dissolved in water, and the solution mixed with a solution of 7*5 grains of decrepitated common salt : calo- mel precipitated, which, when collected on a double filter, washed and dried, weighed 29*5 grains j and the residual liquid being examined for mercury and muriatic acid, no traces of either could be detected. From this experiment it is obvious, that calomel is a compound of 1 atom mercury . 25 1 atom chlorine . 4-5 of caref ul analysis of cinnabar gave evi- MERCURY. 427 dence equally satisfactory, that the atomic weight of mercury is 25. It will be proper to premise, in the first place, that the cinnabar which I em- ployed was not quite free from traces of silica and peroxide of iron : but I made allowance for these bodies, and have left them out of the re- sults to avoid complicated statements as much as possible. Nitric acid has not the property of decompos- ing cinnabar, but nitromuriatic acid dissolves it with rapidity, when assisted by heat. 29 grains of cinnabar in powder were put into a retort, along with a sufficient quantity of nitromuriatic acid : the retort being placed on the sand bath, a pretty strong effervescence began, and the cin- nabar disappeared in a few minutes, with the ex- ception of a yellow matter, which was separated. This yellow matter was soft and ductile. When first collected, it weighed 3*46 grains ; but being left a few days exposed to the air it became hard and brittle, and the weight was reduced to 3*2 grains. It was pure sulphur, burning all away with a blue flame without leaving any resi- due. The nitromuriatic solution was supersaturated with carbonate of soda without the appearance of any precipitate. It was, therefore, evaporated nearly to dry ness, redissolved in water, and mixed with a solution of caustic soda : the mercury precipitated in the state of brownish red flocks, 428 ZINC, CADMIUM, LEAD, c. and being carefully collected and weighed, amounted to 27 grains very nearly. The liquid, thus freed from mercury, was neu- tralized by nitric acid, and being mixed with ni- trate of barytes, a precipitate of sulphate of barytes fell, which, after being washed, dried, and heated to redness, weighed 6 grains, equiva- lent to 0-81 grains of sulphur. Thus, we have obtained from 29 grains of cinnabar Peroxide of mercury . 27 Sulphur . . . 4-01 31-01 This exceeds the quantity analyzed by 2 grains. The reason is, that in the cinnabar the mercury is in the metallic state, while in the analysis it was obtained in the state of peroxide : but 27 peroxide are equivalent to 25 in the metallic state. Hence, cinnabar is a compound of Mercury . . 25 Sulphur . . 4-01 29-01 * The O01 grain of excess is, doubtless, owing to an error in the analysis ; which, with all my care, I was not able to avoid. But I consider the analysis as proving that cinnabar is a compound of SILVER. 429 1 atom mercury 2 atoms sulphur 25 4 29 showing that &5 is the atomic weight of mer- cury. The experiments related in this section, con- stitute the best of a very great number which I have made upon this difficult metal. They leave Atomic no doubt on my mind that the atomic weights of mercury and its compounds are as follows : ATOMIC WEIGHT. Mercury 25 MERCURY. OXYGEN. Protoxide 26 1 + 1 atom Peroxide 27 1 atom + 2 CHLORINE. Protochloride 29-5 1 4- i Perchloride 34 1 + 2 IOOI.VE. Protiodide 40-5 1 + 1 Periodide 56 1 + 2 At the same time, it is proper to state, that I have made no attempts to analyze the iodides of this metal. SECT. VII. OF THE ATOMIC WEIGHT OF SILVER. I HAVE already stated an experiment in page 105 of this work, from which the atomic weight of 16 430 ZINC, CADMIUM, LEAD, c. silver may be deduced. It was there mentioned, 5 that when solutions of 21 %5 grains of nitrate of silver and 7%5 grains of decrepitated common salt are mixed together, a curdy precipitate falls, and the residual liquid is neither affected by ni- trate of silver nor common salt. It is evident from this experiment, that the silver in 21*5 grains of nitrate of silver, and the chlorine in 7%5 grains of chloride of sodium, just saturate each other. Now, 7-5 grains of chloride of sodium con- tain exactly 4%5 grains of chlorine. And, if 21 %5 grains of nitrate of silver be cautiously heated to redness, the salt will be decomposed, the acid driven off, and the oxide reduced ; the weight of silver remaining will be exactly (or at least very nearly) 13 '75 grains. It is obvious from this, that the chloride of silver, formed by the mutual decomposition of 21/5 nitrate of sil- ver and 7*5 chloride of sodium, must consist of Silver . . . 13-?5 Chlorine . 4-5 18-25 Accordingly, when the precipitate is collected without loss, and after being well washed and dried is exposed to a heat sufficiently high to fuse it, the weight is exactly 18 '5 grains. SILVER. 431 Now, as 4 '5 is the weight of an atom of chlo- rine, and as the chloride is neutral, it is obvious that 13-75 must represent the weight of an atom of silver. We have various analyses of the chloride of silver by preceding chemists ; some of which it will be worth our while to state. According to Dr. Marcet,* it is composed of Chlorine 4-5 Silver 13-844 According to M. Gay-Lussac,* its constituents are Chlorine 4-5 Silver 13-683 According to Berzelius,t it consists of Chlorine 4-5 Silver 13-758 This determination of JBerzelius differs from mine by only rrVgth part, a quantity so small, Difficulties that in almost every case of chemical analysis it lysis, may be neglected. I am almost ashamed to acknowledge the great number of times that I repeated this experiment before I could obtain what I considered as perfect accuracy. I began by employing the lunar caustic of commerce ; * Annals of Philosophy, XII. 347. f Afhandlingar, III. 192, 4>32 ZINC, CADMIUM, LEAD, Sec. but was constantly foiled. I then prepared ni- trate of silver in platinum vessels, and found that it was anhydrous, and that 21 %5 grains of it were just precipitated by J-5 grains of chloride of so- dium ; but the jveight of chloride obtained did not tally with the calculated weight. After many trials to discover the cause of the discor- dancy (for the chloride was always slightly de- ficient) I began to suspect that the distilled water which I employed might perhaps contain some ammonia, sufficient to dissolve a small portion of the chloride, and thus to occasion a deficiency, which became perceptible in conse- quence of the great quantity of that liquid with which it had been washed. This induced me to acidulate the water with a few drops of nitric acid in my next experiment. By this addition I found that the chloride of silver (as far as I could weigh it) amounted to just 18-25. The slight excess of chloride in the experiment of Berzelius, and the still greater excess in that of Marcet, the former amounting to O008 grain, and the latter to 0'094 grain, I ascribe to a minute portion of some muriate in the filter em- ployed by these chemists. This could have no effect in my trials ; because I employed just the quantity of nitrate of silver that was decomposed by 7%5 grains of common salt. Hence, my chlo- ride was always deficient ; and I had more than once given up the experiment as desperate be- SILVKLJ. 433 fore I thought of acidulating the water with nitric acid. 2. The preceding experiment gives us not only the atomic weight of silver, but enables us likewise to determine the atomic weight of oxide of silver ; for it is obvious from it, that nitrate of silver is anhydrous, and that 21-5 grains of it contain exactly 13*75 grains of silver. Now, as the salt is neutral, and as 13*75 is equivalent to an atom of silver, it is obvious, that the 21*5 of the salt must contain an atom of nitric acid ; and that the difference between 21 '5 and 20*5, (viz. 18-75 _j_ 6*75) must be the quantity of oxy- gen with which the silver is united. Hence, ni- trate of silver is composed of (of 1 atom silver 13-75) i 1 atom oxide of silver -J > 14-75 ( 1 atom oxygen 1 ) I atom nitric acid . . . . .6-75 21-5 It is obvious, that the oxide of silver is a com- Atom of , oxide of sil- pound of 1 atom silver + 1 atom oxygen, and ver . that its atomic weight is 14*75. 3. When oxide of silver is dissolved in ammo- Suboxide nia, and the solution left exposed to the air, a brilliant black pellicle collects on the surface. This pellicle was first examined by Mr. Faraday, from whose experiments it appears, that it con- tains two thirds of the oxygen in the common oxide of silver. My experiments on this pellicle, VOL. I. E c 434 ZINC, CADMIUM, LEAD, c. which were made soon after Mr. Faraday had published his observations, were upon too small a scale. My object was merely to verify Mr. Faraday's statement ; but as far as I went I ob- tained just the s but his estimate was founded on the specific gra- vity of oxygen gas as determined by Berzelius and Dulong, which I have shown in a former chapter to be too low. When we correct his es- timate, the proportion of oxygen united to 100 gold turns out to be 12-0008, a number which I am persuaded is almost exact.* SECT. II. OF THE ATOMIC WEIGHT OF PLATINUM. PLATINUM dissolves, as every body knows, in strong nitromuriatic acid with effervescence. The solution has a dark brown colour, and is very as- tringent and corrosive. When concentrated by evaporation it does not crystallize, but forms a dark coloured dirty looking matter, not easily * 1 have some suspicion that the true atomic weight of gold may ulti- mately turn out only 12-5 instead of 25. If that be the case, the triple salt analyzed in this section would be a compound of 2 atoms chloride of gold and 1 atom chloride of sodium. We have not data, at present, to deter- mine this point with certainty. Vor, I. F f 450 GOLD, PLATINUM, PALLADIUM, &C. analyzed with accuracy. From the nitromuri- atic solution of platinum it is easy to form tri- ple salts by adding solutions of muriate of potash, muriate of soda, or muriate of ammonia. The triple salts with ^muriates of potash and ammonia are insoluble powders of a fine yellow colour : the triple salt with muriate of soda is soluble, and crystallizes in beautiful rhombic prisms. These salts furnish a method for determining the atomic weight of platinum, to which I had re- course, after a few unsuccessful trials with muri- ate of platinum. Analysis of ]_. Triple muriate of platinum and potash is bichloride r of platinum an insoluble yellow tasteless powder, and when num. properly dried contains no water of crystalliza- tion. After a number of preliminary trials to enable me to judge how I ought to proceed, and which it is unnecessary to detail, I analyzed the salt in the following manner : 30*5 grains of it were put into a platinum crucible, and subjected to a heat which was slowly and cautiously raised till the crucible be- came red hot : by this process the platinum was reduced to the metallic state, and the loss of weight sustained amounted to 9 grains, it was owing to the escape of chlorine in the form of gas. If the salt has been overheated, (which happened in some of my first trials) the loss of weight will exceed 9 grains ; but when this hap- pens, the weight of the chloride of potassium PLATINUM. 4,51 remaining mixed with the platinum will be found just as much less than 9*5 grains, as that of the loss exceeds 9 grains. This gives us the cause of the excessive loss ; it is owing to the volatiliza- tion of a portion of that salt. . The matter remaining in the crucible, when the heat had been properly applied, weighed 21 '5 grains : water was poured upon it repeated- ly, and digested on it for some time, to dissolve off all the chloride of potassium. The platinum being then dried, was weighed, and its weight was found to amount to exactly 1 2 grains ; con- sequently, the weight of the chloride of potassi- um was 9*5 grains. From this analysis it appears, that the triple salt is composed of Platinum . . . 12 Chlorine ... 9 Chloride of potassium . 9*5 30-5 But the salt is neutral 9 indicates the weight of % atoms of chlorine, and 9*5 of an integrant par- ticle of chloride of potassium : 1 2, therefore, must either be the weight of an atom of plati- num, or at least, a multiple of it. From experi- ments to be stated immediately, it will be seen, that 12 represents the atomic weight of platinum. In the salt we find 1 atom of platinum combined with 2 atoms of chlorine, constituting a bichlo- FfS GOLD, PLATINUM, PALLADIUM, c. ride. The constituents of the triple salt are ob- viously 1 atom bichloride of platinum =21 1 atom chloride of potassium = 9*5 30-5 And the atomic weight of an integrant particle of the salt is 30-5. Berzelius has published an analysis of this powder, which differs but little from mine. He found the constituents Platinum . 40 or 12-2 Chlorine . . 292 - 8-906 Chloride of potassium 30-8 - 9'394 100 30-5 The atomic weight of the salt is the same as that deduced from my own experiments. Not having seen the particulars of Berzelius' expe- riments, but merely the result, I cannot judge of the cause of the slight discrepancies between the proportions of the ingredient, when his analysis is compared with mine : they are probably, in a great measure, owing to the difference of the data, which he employs for his calculations. But, notwithstanding these slight differences, the analysis of Berzelius is a most important corro- boration of the very near approach to accuracy which I have attained in mine. PLATINUM. 453 2. The triple muriate of platinum and soda Analysis of , . , , ! i i i bichloride has not hitherto been described in any chemical of platinum book which I have seen though it must have an often been prepared by practical chemists. It may be worth while, therefore, to mention its most striking properties before proceeding to its analysis. Its colour is a fine orange red it crystallizes in flat, and very oblique four-sided prisms its taste is cooling, bitter, and disagreeable it has no smell its specific gravity is 1 *980 at the temperature of 62 '5, 100 parts of water dissolve 107'2 parts of the crystals of this salt it is so- luble, likewise, in alcohol. I analyzed it in the following manner : 37'5 grains of the crystals being cautiously heated on the sand bath for about an hour, were found to sustain a loss of weight amounting, at an average, to 9 grains. Sometimes, when the heat happened to be too low, the loss of weight was less than 9 grains ; and at other times when it was too high, it exceeded that amount. The salt, thus deprived of its water of crystallization, was heated to redness in a platinum crucible, and the heat was continued till the platinum in the salt was reduced to the metallic state : the addi- tional loss of weight sustained was 9 grains this loss was owing to the disengagement of chlorine. The matter remaining in the crucible was di- gested in distilled water till every thing soluble F f 3 454 GOLD, PLATINUM, PALLADIUM, & c . was taken up. The platinum now remaining in a state of purity, after being dried, weighed 12 grains ; consequently, the weight of the chloride of sodium removed by the distilled water was just 7\5 grains* From this analysis it follows, that the triple muriate of platinum and soda is composed as follows : Platinum . . . 12 Chlorine ... 9 Chloride of sodium ,. 7-5 Water ... 9 37-5 In one portion of this salt which I examined, the water, instead of 9 grains, amounted only to 8*31 grains, which would indicate only 7% atoms of water instead of 8 atoms. The cause of this dif- ference I am unable to explain : perhaps the salt had been accidentally deprived of a portion of its water. The constituents of this salt give us 12 for the atomic weight of platinum, precisely like that of the preceding salt. We may either consider it as composed of 1 atom bichloride of platinum =21 1 atom chloride of sodium . = 7-5 8 atoms water . . . =9 PLATINUM. 455 Or, its constituents in the crystallized state may be 1 atom bipermuriate of platinum = 23-25 1 atom muriate of soda .' . = 8-625 5 atoms water = 5'625 S7'5 The constituents, which ever supposition we adopt, are the same ; the only difference consists in the way in which they are arranged. If the salt be a permuriate, 3 atoms of the water must go to convert the platinum and sodium into oxides, and the 3 atoms of chlorine into muriatic acid. 3. Ammonio-muriate of platinum is formed by Analysis of . ., , ammonio- mixing sal-ammoniac with a concentrated sol u- muriate of tion of muriate of platinum. This salt is soluble pla1 in water, though very slightly, and it crystallizes in small grains seemingly octohedrons. Its co- lour is a rich yellow, and it contains no water, but is composed of 2 atoms chlorine 9 1 1 atom platinum 12 J 1 atom sal-ammoniac . . . 6 %> 75 27-75 The preceding analyses are sufficient to show that 12 either denotes the atomic weight of pla- tinum, or that it is a multiple of that weight. Some other compound of platinum is still want- F f 4 456 GOLD, PLATINUM, PALLADIUM, &c. ing to enable us to decide which of these two suppositions is the true. After some unsuccess- ful attempts, I at last found that the sulphate of platinum put it in my power to solve the problem. 4. I formed a quantity of sulphate of platinum by the process pointed out by Mr. Edmund Davy. f f 19 grains of this dry sulphate were put into a platinum, small green glass retort, and exposed to a red heat till the platinum was reduced to the metal- lic state; it weighed exactly 12 grains. This experiment was repeated several times, in order to collect the oxygen gas evolved by the reduc- tion of the platinum, and to determine the quantity of sulphuric acid separated; for both these products could not be conveniently collect- ed at the same time. The oxygen gas evolved was rather less than three cubic inches, and amounted, therefore, very nearly to 2 grains. The sulphuric acid was a very little less than 5 grains ; the reason was, that a little sulphurous acid was disengaged, which I found it difficult to obtain without some loss. From these experiments it is obvious enough, that the constituents of sul- phate of platinum are Platinum 12 Oxygen 2 Sulphuric acid 5 19 Now the salt was neutral, and 5 being equiva- PLATINUM. 457 lent to an atom of sulphuric acid, it follows that 12 must denote an atom of platinum : 2 being A 1 to . m of * f platinum. equivalent to 2 atoms of oxygen, it is obvious that the oxide of platinum in the sulphate, which is the peroxide, is a compound of 1 atom plati- num and C 2 atoms oxygen. Consequently, its Atom . of atomic weight is 14. The knowledge of the atomic weight of pla- tinum deduced from this salt, enables us to de- cide, that the triple chlorides of platinum and potash, and platinum and soda, contain 1 atom of platinum united to 2 atoms of chlorine. Mr. Edmund Davy made an analysis of the sulphate of platinum, a good many years ago, and his results, though not quite accurate, con- stitute near approximations to the truth. It was his analysis that pointed out to me the mode which I had to follow. My method being nearly the same with his, I considered it unnecessary to describe it minutely. Mr. E. Davy's results are as follows :* Platinum 12-4697 Oxygen 1-5352 Sulphuric acid 4-9951 19-0000 The weight of an integrant particle of the salt deduced from this analysis is the same as from my own, namely 19. The proportion of sulphuric * Phil Mag. XL. 354. 458 GOLD, PLATINUM, PALLADIUM, & c . acid is very nearly correct ; but the platinum is too much, while the oxygen is as much too little, as if the metal had not been completely reduced to the metallic state. 5 * * tr * ec * to P re P are a quantity of peroxide of platinum by the process pointed out by Berze- lius. Finding it to contain both muriatic acid and an alkali, I reduced it to a fine powder, and boiled it for a week in a flask kept nearly full of distilled water. After this treatment, the pro- portion of salt which it contained was so exceed- ingly small, that I am disposed to consider the characters which my powder possessed, as very nearly those of pure peroxide of platinum. Its colour was a brownish black ; it was tasteless, insoluble in nitric, sulphuric, oxalic, and acetic acids ; but it dissolved, though slowly, and 'not without considerable boiling, in muriatic and phosphoric acids. I made no attempt to analyze it, both because it was not absolutely pure, and because I was not sure whether it was perfectly anhydrous. Amalgam g. Conceiving that the amalgam of platinum was num. likely to throw some light on its atomic weight, I made a quantity of it by triturating platinum pow- der and mercury in a mortar ; beginning with a few grains of each, and adding additional portions of both metals as the amalgamation proceeded. The fluid amalgam thus formed was put into a leathern bag, and subjected to strong pressure in PLATINUM. 459 order to squeeze out all the uncombined mer- cury. I obtained a solid brittle amalgam, having the colour of mercury, a granular texture, and a brittle consistency. 61 grains of this amalgam were exposed to a heat at first moderate, but gradually raised as much as possible in a wind furnace, and kept up for half an hour. The loss of weight sustained was 24'9 grains, and there remained 36-1 grains of platinum appa- rently pure. Now, 4*9 is very nearly 25, which denotes the atomic weight of mercury ; and 36*1 is also exceedingly near 36 = 3 atoms of pla- tinum. It is obvious from this, that the amal- gam was a compound of 3 atoms platinum 36 1 atom mercury 25 61 Conceiving that by allowing the fluid amalgam of platinum to crystallize, I should probably ob- tain a compound of the two metals in a different proportion, I placed a quantity of the fluid amal- gam on an inclined plane, and when it had be- come solid by the running off of the superfluous mercury, I analyzed it in the way just described, and found its constituents to be Platinum 12 Mercury 35*3 460 GOLD, PLATINUM, PALLADIUM, &c. It is obvious that the crystals had retained a portion of the mercury mechanically lodged be- tween their particles. In fact, these crystals resembled crystals of a salt taken out of the mother liquor before being dried ; yet I think it not unlikely from this experiment, that the crys- tals of amalgam of mercury, if they could be obtained in a state of purity, would consist of 1 atom platinum 12, 1 atom mercury 25 37 7. From the preceding experiments there seems no doubt that the atom of platinum weighs 12; and that it forms two oxides, only one of which, the peroxide, has been yet obtained in a separate state. The atomic weight of the per- oxide is 14. The protoxide must be a com- pound of 1 atom platinum and 1 atom oxygen, and its atomic weight must be 13. SECT. III. OF THE ATOMIC WEIGHT OF RHODIUM. THE soda-muriate of rhodium, a triple salt dis- covered by Dr. Wollaston, and with a quantity of which he was so obliging as to supply me a RHODIUM. 4-61 good many years ago, enabled me to determine the atomic weight of rhodium and its oxides, with considerable precision. This salt forms rhomboidal crystals of an intensely deep red colour, and not altered by exposure to the air. It is insoluble in alcohol, but sufficiently so in water. Carbonate of ammonia produces no ef- fect upon the solution of this salt ; but caustic ammonia, as well as the caustic fixed alkalies precipitate the peroxide of rhodium in yellow coloured flocks. An excess of the fixed alkalies redissolves the whole of the precipitate. Though ammonia does not act so powerfully as a solvent, yet it is not without its effect ; for the whole oxide of rhodium is never obtained if an excess of ammonia be added to the solution. 1. A hundred grains of crystals of soda mu- Analysis of r TIT- j soda-mu- riate of rhodium were dissolved in water, anciriateofrho- the peroxide was precipitated by ammonia, tak- ing care not to add that alkali in excess. The oxide of rhodium, when collected, washed, and dried, was found to weigh 26*07 grains. Into a solution of another 100 grains of the same salt a polished plate of zinc was put, and left in the solution till the whole of the rhodium which it contained had been precipitated in the metallic state. To prevent the rhodium from adhering to the zinc, which would have made it difficult to weigh it correctly, the liquor had been acidulated with some muriatic acid. The 462 GOLD, PLATINUM, PALLADIUM, &c. weight of the metallic rhodium obtained in this experiment was 19'1 grains. From this preliminary experiment it appears, that 26*07 grains of yellow oxide of rhodium contain 19'1 grains of metallic rhodium, and consequently, 6 '97 grains of oxygen : therefore, yellow oxide of rhodium is a compound of Rhodium 1 9-1 or 5-486 Oxygen 6*97 - 2 I was induced to infer from this experiment, that the atomic weight of rhodium is 5'5 ; and that the yellow oxide is a compound of 1 atom of rhodium, and 2 atoms of oxygen ; and con- sequently weighs 7*5. 2. 28'625 grains of the crystals of soda-mu- riate of rhodium were put into a small green glass retort, and exposed to a heat which was gradually raised to an incipient redness, and kept up a considerable time in order to dissipate completely the aqueous vapours which are dis- engaged from the salt by the heat. This ex- periment cannot be made in a metallic vessel without disengaging some muriatic acid, and re- ducing a portion of the rhodium to the metallic state ; but I have kept the salt in a low red heat for upwards of an hour in a green glass retort, and found it still perfectly soluble in water with the usual dark red colour, and consequently, still retaining the whole of its acid. The loss of RHODIUM. 463 weight in this experiment was 9 grains ; which is, therefore, the water of crystallization in 28-625 grains of soda-muriate of rhodium. 3. 28 '625 grains of soda-muriate of rhodium were dissolved in water, and precipitated by am- monia added very cautiously, and all excess avoided. The yellow oxide obtained weighed 7%5 grains. This step of the analysis cost me a great deal of trouble, and was repeated about ten times before I succeeded in separating the whole of the oxide, and collecting it without loss. 4. The residual liquid, from which the oxide of rhodium had been thrown down, was evapo- rated with great caution to dryness, and the re- sidual salt weighed. The weight did not exactly agree in the different trials ; doubtless, because I applied the heat so cautiously that the whole aqueous portion was not driven off; but the average weight of this saline residual was 14 '5 grains. It was a mixture of common salt and sal-ammoniac. Being exposed to a heat suffi- cient to drive off the sal-ammoniac, the residual common salt weighed 7 ''5 grains. Consequently, the loss of weight was 7 grains ; but the weight of an integrant particle of sal-ammoniac is 6*75 grains ; for it is composed of 1 atom muriatic acid 4- '62 5 1 atom ammonia 2.125 6-75 464 GOLD, PLATINUM, PALLADIUM, &C. It is obvious that the O25 grain of loss, over and above this integrant particle of sal-ammoniac, must have been water. Now, the 4*625 grains of muriatic acid, in the 6 '75 grains of sal-ammo- niac volatilized, must have existed in the 28 '625 grains of the soda-muriate of rhodium, and must have been in combination with the oxide of rhodium. From the preceding analysis it appears, that soda-muriate of rhodium is a compound of Yellow oxide of rhodium 7-5 Muriatic acid . . 4-625 Common salt . 7-5 Water 9-0 28-625 Atom of Now, this salt is perfectly neutral : 9 re- peroxide of J rhodium, presents eight atoms or water ; 7'5 is an inte- grant particle of chloride of sodium ; and 4-625 an integrant particle of muriatic acid. 7%5 there- fore, must represent the weight of an integrant particle of yellow oxide of rhodium. The con- stituents of the soda-muriate of rhodium may therefore be represented as follows : 1 atom permuriate of rhodium 12'125 1 atom muriate of soda . 8 '62 5 7 atoms water . . 7*875 28-625 For one of the atoms of water separated by heat, RHODIUM. 465 was ill all probability formed by the change of the integrant particle of muriate of soda into chloride of sodium. 5. The preceding analysis, when compared with the preliminary experiment, leaves no doubt that the weight of an atom of rhodium is Atom of . - , _ rhodium. 5%5, and that the yellow oxide is a compound of 1 atom rhodium and 2 atoms oxygen. This oxide of rhodium is described by Ber- oxides of zelius as brown ; but when obtained by precipi- r tation from the soda-muriate of rhodium, it is always yellow. When exposed to heat it be- comes slowly brown, and at last black. During this change of colour, I could detect no escape of moisture, to induce me to suppose that the yellow coloured powder is a hydrate, and the brown the anhydrous oxide. Whether the brown oxide be the same, or different from the yellow, I have not ascertained ; most probably it is a mixture of the yellow and the black oxides. The black oxide is undoubtedly the protoxide of rhodium, and a compound of 1 atom rhodium and 1 atom oxygen ; so that its atomic weight is 6-5. If the brown powder should appear, upon far- ther examination^ to be a peculiar oxide, it is obvious that it must be a compound of 1 atom rhodium and li atom oxygen ; and that its ato- mic weight must be 7 I did not succeed in my attempts to dissolve it in acids ; but this insolu- Vot. I. G g 466 GOLD, PLATINUM, PALLADIUM, c. bility may have been owing to the heat to which it had been exposed. SECT. IV. OF THE ATOMIC WEIGHT OF PALLADIUM. MY stock of palladium was so small, that I was under the necessity of making my experiments upon a very minute scale. This precluded the possibility of attaining the mathematical accu- racy necessary for determining the atomic weight with certainty ; but as my experiments, so far as they go, confirm those already made by Berze- lius on this metal, I flatter myself that the num- ber which I have deduced from them is a very near approximation to the truth. From the great scarcity of this metal at present, it is unlikely that quantities sufficiently large to decide the point with certainty, will speedily be at the dis- posal of any chemist. 1. Palladium, as is known, dissolves in all the acids ; but by far the best solvent of it is nitro- muriatic acid : the solution has a very deep red Properties colour. When we evaporate the solution to of paiiadi. dryness, and re-dissolve the residual salt in wa- ter, the colour of the liquid is a reddish brown ; Hs taste is intensely astringent, without any ac- companying flavour of a disagreeable nature. PALLADIUM. 467 i This solution is precipitated by potash in dark red flocks ; by ammonia in buff coloured flocks. The precipitate by potash readily dissolves in nitric acid ; but the ammoniacal precipitate does not dissolve in that acid till the action of the acid is assisted by heat. Gallic acid gives the liquid a greenish brown colour, but occasions no immediate precipitate ; prussiate of potash throws down reddish brown flocks, but the solution is not rendered colourless. Muriate of palladium has a reddish brown co- lour. It does not crystallize ; its taste is in- tensely astringent. When the dry salt is expos- ed to the air, it becomes moist on the surface, but does not flow into a liquid. This salt dis- solves readily in water. 2. I dissolved 7 grains of palladium in muria- Potash mu- tic acid (assisted by a little nitric acid). To the iadium. P solution I added a quantity of carbonate of pot- ash, which I knew to contain exactly 6 grains of potash. This solution being evaporated, yielded crystals of triple muriate of palladium and pot- ash to the very last drop. This experiment shows, that in the triple salt the palladium bears to the potash the ratio of 7 to 6. 3. Berzelius has given us the result of two * id f. * m . palladium. experiments which he made with palladium. He oxydized 100 parts of palladium, by heating them in a platinum crucible, with some potash and nitre ; the oxide weighed 114'209 parts. G 2 468 GOLD, PLATINUM, PALLADIUM, &C. As palladium forms only one oxide, we may consider it as a compound of 1 atom metal and 1 atom oxygen. Now, 14-209 (the oxygen which united with the metal) : 100 (the weight of metal) : : 1 (atomic weight of oxygen) : 7*038 = atomic weight of palladium. S , ulp ^ ur , e . t He informs us likewise, that 100 parts of pal- of palladi- um, ladium, when converted into sulphuret, combine with 28*15 parts of sulphur. If we consider the sulphuret thus formed as a compound of 1 atom metal and 1 atom sulphur, we obtain 7' 104 for the atomic weight of palladium ; for 28-15 : 100 : : 2 (atom of sulphuret) : 7'104. * These two separate estimates of the atomic weight of palladium approach so nearly to 7, that I was led from them to consider 7 as the true atomic weight of this metal. It was this that induced me to make choice of 7 grains of palladium, and to add to the muriatic solution 6 grains of potash, both of these quantities being equivalent to the atomic weight of these bodies. The result of the experiment shows that the opinion was correct, or very nearly so, and au- thorizes us to make choice of 7 as the true ato- mic weight of palladium. It would have been desirable to have made the experiment upon a larger scale to have employed, for example, * Essai sur la Tlieorie des Proportions Chimiques, p. 137. PALLADIUM. 469 70 grains of palladium and 60 grains of potash. Such quantities would have shown more decisive- ly whether there was an excess of either of the ingredients ; but, unfortunately, my stock of palladium was inadequate to any such experi- ment. 4. The triple muriate of palladium and potash J crystallized in very small needles, of a mostiadium. beautiful brown colour. Dr. Wollaston describes them as four-sided prisms j and says, that when viewed transversely, they are of a bright green ; but when seen in the direction of their axes, they are dark red. In my crystals, from their extreme minuteness, it was impossible to ob- serve these appearances. Was the brown colour owing to a mixture of the two colours described by Wollaston ? It has an intensely astringent taste, accom- panied with a perceptible sweetness. It is in- soluble in alcohol, but pretty soluble in water ; and the solution has a dark red colour, like that of muriate of palladium. It contains very little water ; for 10 grains of it, when kept 24 hours upon the sand bath, in a temperature of nearly 600, lost only 0-12 grain of weight. When heated to redness the salt melted, and I kept 10 grains in a state of fusion for half an hour ; the loss of weight was 0*8 grain. When exposed to the intense heat of a wind furnace, the palladium is reduced to the metallic state ; but the tem- GgS 470 GOLD, PLATINUM, PALLADIUM, c. perature requisite for this reduction is so nign, that the greatest part of the muriate of potash is volatilized. The weight of palladium obtained from 10 grains of the triple salt was 3*08 grains. Now, it was shown by a preliminary experi- ment, that the potash in this salt bears to the palladium the ratio of 6 to 7 5 consequently, 10 grains of the salt must contain 2*64 grains of potash. The palladium in the salt is doubtless in the state of oxide. If the atomic weight of palladi- um be 7> and that of oxide of palladium 8, then (since 7:8:: 3-08 : 3-52,) 3-52 grains must be the weight of oxide of palladium in 10 grains of the triple salt. Both of these bases in the triple salt are com- bined with muriatic acid. Now, the atomic weight of muriatic acid being 4*625, it is evi- dent that 8 grains of oxide of palladium will be just saturated by 4*625 grains of muriatic acid ; and 3 '52 grains of oxide will require for satura- tion 2-035 grains of muriatic acid. And, since 6 potash require 4*625 of muriatic acid, 2*64 grains of potash must require for sa- turation 1 -526 grain of muriatic acid. From this it appears, that 2*035 grains of mu- riatic acid are in combination with the oxide of palladium, and 1*526 grain with the potash. These calculations enable us to state the con- PALLADIUM. 471 stituents of the triple muriate of palladium as follows : Oxide of palladium 3-52 1 ^ ,, . .- > 5-555 muriate of palladium, Muriatic acid 2-035 j Potash 2-64 \ A . _ > 4- 166 muriate of potash. Muriatic acid 1 -526 J Water . . . 0-279 10-000 This estimate would give us the constitution of the salt in atomic numbers as follows : 1 atom muriate of palladium = 12-625 1 atom muriate of potash = 10-625 f ths of an atom water . = 0-675 23-925 Had the quantity of water amounted to 1 atom instead of fths of an atom, the analysis would have been more satisfactory. The reader will observe that with respect to palladium, I was obliged to rest satisfied witli the preliminary experiments which I made on the other metals, before attempting the analysis from which the atomic weights are deduced. These preliminary trials I have thought it need- less to state in the other sections ; but my stock of palladium having been exhausted by them, I was under the necessity of drawing my conclu- sions respecting this metal from them entirely. The above analysis, (if analysis it can be called) Gg 4 GOLD, PLATINUM, PALLADIUM, &c. was thrice repeated with very nearly the same result. chloride of 5^ j ma d e three attempts to ascertain the ato- palladium. mic weight of palladium, by forming a chloride. My method was* to dissolve 7 grains of palla- dium in muriatic acid, to evaporate the solution to dryness, and to expose the muriate to heat in a covered platinum crucible. The first trial gave 11 -9 10 grains of chloride The second . 12-133 The third . 12-231 Mean . . 12-091 These weights are too high, if 7 be the atomic weight of palladium j for if we subtract 7 from 12-091, the remainder is 5*091, which some- what exceeds the weight of an atom of chlorine. Perhaps the salt was not exposed to a heat suffi- ciently high to decompose the muriate com- pletely. The experiment is scarcely susceptible of accuracy ; because at the heat necessary to decompose the muriates, a portion of the chlo- rine is apt to escape. The preceding experiments, imperfect as they are, lead to the conclusion, that the atomic weight of palladium is 7, and that of its oxide 8. These numbers are certainly very near the truth. IRiDIUM. 473 SECT. V. OF THE ATOMIC WEIGHT OF IRIDIUM. I PREPARED a quantity of muriate of indium from the black matter which remains when crude platina is dissolved in nitro-muriatic acj$. The process which I followed was nearly that pointed out by Vauquelin in his paper on os- mium and iridium. This black matter was found to contain a very considerable proportion of iron and some titanium ; the quantity of in- dium which it yielded was much less than I had anticipated. 1. Muriate of iridium crystallizes in tetrahe- Properties drons ; it has a great deal of lustre, and a brown colour so deep that the salt appears black. Its taste is intensely bitter and astringent ; its solu- tion in water has a red colour, so deep that the liquid continues opaque even after it has been a good deal diluted with water. When so much diluted as to be transparent, the colour appears yellow. The aqueous solution of muriate of iridium is rendered colourless by the addition of a few drops of any of the following bodies : Gallic acid, Prussiate of potash, 474 GOLD, PLATINUM, PALLADIUM, & c . Ammonia, Potash, Soda, Lime^Water, JSarytes water, Sulphuretted hydrogen, Protosulphate of iron, Sulphate of zinc. Zinc or tin, when left in the liquid, renders it colourless in twenty-four hours. The following bodies produce no alteration in the colour and transparency of this liquid. Blue vitriol, Persulphate of iron, Muriate of manganese, Nitric acid, Oxalic, tartaric, citric, acetic acids, Sulphuric, nitric, muriatic acids. When nitrate of mercury is dropped into the solution of muriate of iridium, an abundant precipitate of orange flocks falls down, and the liquid becomes colourless. Muriate of tin pro- duces a slight yellow precipitate, and renders the liquid colourless. Such are the effects of reagents upon this sin- gular substance. It is not precipitated either by alkalies or by acids so that no good method of analyzing it readily presents itself. Yet, upon making the trial, I found the analysis of it both easy and satisfactory. IRIDIUM. 4-7^ 2. A hundred grains of the crystals were put Analysis or . A , , chloride of into a small green glass retort, and kept red hot indium. for two hours by means of a charcoal fire. The crystals did not alter their appearance in the least; yet they emitted a perceptible smell of chlorine, and the salt sustained a loss of weight amounting to 2'6&5 grains. The crystals thus treated were completely soluble in water ; they were therefore in a state of a pure anhydrous chloride of iridium. 3. 82%5 grains of this chloride dried in a red heat were put into a platinum crucible, and ex- posed to the strongest heat which I could raise in a wind furnace. When the crucible was cold, I found that the chloride had been decomposed, and that there remained 37'5 grains of metallic iridium. The loss of weight sustained was there- fore 45 grains, and this loss was obviously occa- sioned by the dissipation of the chlorine. This experiment will be more easily under stood if we divide the quantity of chloride em- ployed by ten. 8*25 grains of chloride of iri- dium, when decomposed by heat, lose 4'5 grains of chlorine, and the remaining 3*75 grains consist of metallic iridium. Thus we see that the chlo- ride is composed of Iridium . . . 3*75 Chlorine 4-5 8.25 476 GOLD, PLATINUM, PALLADIUM, & c . Now, as 4-5 is equivalent to an atom of chlorine, Atom of 3.75 mu st represent the atomic weight of iri- iridium. dium. I repeated this experiment twice, employing each time 8*25 'grains of chloride .of iridium ; because my stock of salt had been too much ex- hausted by the preceding trials to allow me to employ a greater quantity. These two experi- ments did not quite agree with each other ; the first leaving a greater quantity of iridium than 3*75 grains, and the second a smaller quantity : the mean of the two made the quantity of iridi- um amount to 3*88 grains. This approaches as near to 3*75, the atomic weight deduced from the first experiment, as could be expected, when we take into view the very small scale on which the experiments were made. For when we em- ploy in our trials very small quantities of matter, exact results are not to be looked for ; because the unavoidable inaccuracies arising from weigh- ing, or from want of absolute correctness in the weights which we employ, bear too great a pro- portion to the whole quantity used. But the first experiment, which was made with great care upon a quantity sufficiently great for accuracy, gives us, I conceive, the true atomic weight of iridium. I believe the crystals usually called muriate of iridium should be called chloride of iridium; for the loss of weight being only -625 per cent., OSMIUM. 477 does not amount to quite so much as the fifth of an atom of water. This explains the reason why the iridium is not precipitated by alkalies, as is usually the case with metallic oxides. When the solution of this salt in water is made without heat, it has a fine green colour ; but if we boil the green liquid, it gradually loses its original colour and becomes red. This change of colour has been generally ascribed to the ad- dition of oxygen to the oxide of iridium. It has been concluded from it that there are two oxides of iridium, the protoxide is green and the per- Oxides of oxide red. Whether this explanation be the true m one I have not been able to determine, as all my attempts to obtain the oxides of iridium in a separate state have ended in disappointment. If there be two oxides of iridium, the protoxide is doubtless composed of 1 atom iridium and 1 atom oxygen, and its atomic weight must be 4'75. Peroxide of iridium must be a compound of 1 atom iridium and 2 atoms oxygen, and its atomic weight must be 5^5. SECT. VI. OF THE ATOMIC WEIGHT OF OSMIUM. I AM still unacquainted with the atomic weight of osmium the quantity of that anomalous sub- 478 GOLD, PLATINUM, PALLADIUM, c. stance at my disposal being too small to enable me to prosecute the subject to any extent. 1. I mixed together equal weights of zinc filings and osmium, and putting the mixture into a glass tube benf like a retort, I poured on it a quantity of dilute sulphuric acid, to see whether the hydrogen gas evolved contained any osmium in solution : but I could obtain no evidence that it contained any. The gas had no sensible smell nor taste ; it was not absorbed by water, and did not produce any sensible effect upon the solu- tions of lead. It burned just like common hy- drogen gas, and required, as usual, half its volume of oxygen gas for complete combustion. The sulphuric acid solution of the zinc was black and opaque, and contained, obviously, a portion of the osmium in a state of solution. But upon attempting to obtain this triple solution in a 'dry state, the colour gradually became lighter as the evaporation advanced, and when the whole had been reduced to dryness, all traces of the presence of osmium had vanished. Osmium, in the metallic state, was digested in caustic soda ley : a yellow coloured liquid was obtained, but the greatest part of the osmium remained undissolved. END OF VOLUME FIRST. USE RETURN TO DESK FROM WHICH BORROWED LOAN DEPT. This book is due on the last date stamped below, or on the date to which renewed. Renewed books are subject to immediate recall. JAN 19 1975 LD 21A-50m-3,'62 (C7097slO)476B General Library University of California Berkeley