bonate of lead in nitric acid for the final lead determination, was in each case incinerated, and then fused with carbonate of soda before the blowpipe. A distinct button of metallic gold was obtained in every instance. The amount of gold obtained varied, however, to a very considerable extent: it was abundant in the American and Spanish coin, less was observed in the English and Mexican, and but little in the French coin or the American fine silver.*

The silver of the French coin was undoubtedly prepared by the sulphuric-acid process,† in which the granulated alloy of silver, gold,

* Besides the quantitative analyses recorded above, which have been made by this method, it should be mentioned, that in several qualitative experiments upon American coins, by the method as described or slightly modified, indications of lead were obtained in every instance. The same remark is true of two qualitative analyses, one of a dime and another of a quarter-dollar, made as follows: the coin having been dissolved in nitric acid, and the insoluble black residue of sulphide of silver and a little gold separated by filtering, the solution was mixed with pure acetate of soda (see Fresenius, op. cit., p. 428, § 163, 1. y) to prevent precipitation of chloride of lead in subsequent operations, and the whole heated over the water-bath; the silver was then precipitated with dilute chlorhydric acid, and the chloride of silver collected upon a filter and washed. Through the filtrate sulphuretted hydrogen was passed, the mixed sulphide of copper, lead, and gold collected upon a filter, treated with nitric acid, the acid solution evaporated to dryness, the residue taken up with water, and the solution filtered; in the filtrate lead was detected, while the residue contained gold.

With regard to this residue of gold, which has also been alluded to in the text, it should be remarked that it has been entirely neglected by many analysts who have published determinations of the amounts of gold in various silver coins. Indeed, it would appear as if the fact were not generally known to chemists, that nitric acid can dissolve a little gold (as it can platinum), when an alloy containing a small quantity of gold is treated with this acid (compare Mulder, op. cit., p. 159). Another common error of analysts has been, to class as "gold" the dark-colored residue which is usually left when silver coin is dissolved in dilute nitric acid. This precipitate does, in fact, generally contain a little gold, though in our own experiments by far the larger portion of this metal passed into solution in the nitric acid, and was separated at a subsequent stage of the analysis, as has been described; Gay Lussac (Ann. Ch. et Phys., 1836, [2.] LXIII. 334) long ago showed that it is principally composed of sulphide of silver; to which fact attention has also been called by Mulder, in his excellent little treatise, just cited, p. 178.

† See D'Arcet in Dingler's Polyt. Journ., 1828, XXVIII. 3; from Annales Mensuelles, Mai, 1827, p. 131. Also, Ure's Dictionary of Arts, &c., 4th edit., (Boston, 1853,) II. 543. Dumas's Traité de Chimie Appliquée aux Arts, (Paris, 1833,) IV. 464.

copper, &c. is dissolved in hot concentrated sulphuric acid, and the silver subsequently precipitated, by means of metallic copper, from a somewhat diluted hot solution of the sulphate of silver thus obtained. Since ordinary commercial sulphuric acid is used in this operation, it is not strange that a portion of the lead with which it is contaminated should be transferred to the silver. The lead salt is perhaps reduced by the metallic copper in the manner so well described by Odling, in his memoir "On the Reciprocal Precipitations of the Metals." It is, moreover, not impossible that some sulphate of lead may fall when the acid liquor is diluted, and become mixed with the precipitated silver. It is not credible, however, that the lead in the coin can have been derived from the copper used to form the standard alloy, for this supposition would imply that copper of very inferior quality had been employed. According to Karsten, copper which contains but one per cent of lead is utterly unfit for manufacturing purposes, since it cannot be worked at any temperature. Now the least amount of lead which we found in fifty cents' worth of American silver is 0.0253 gram. in 10 half-dimes of 1853. These 10 half-dimes contain 1.243 gram. of copper, and if this metal had been the source of the lead, it must have contained more than two per cent of that impurity.

We are ignorant of the process of preparing silver which was in use at the British Mint in 1816. Perhaps the silver in the coin of that date was obtained by cupellation, and it is well known that silver so obtained almost always contains lead. §

*D'Arcet, Journ. für tech. u. œkon. Ch., 1829, IV. 420.

† Quar. Journ. Chem. Soc. of London, 1857, IX. 289.

In his System der Metallurgie, (Berlin, 1832,) V. 245.

§ Authorities differ as to the amount of lead contained in crude cupelled silver ("lightened silver,” argent éclairé, Blicksilber). According to Berthier (Essais par la Voie Seche, (Paris, 1848,) II. 724), it contains, on an average, only 1 per cent of lead. Kerl (Handbuch der metallurgischen Hüttenkunde, (Freiberg, 1855,) III. 152), on the other hand, says that it contains from 5 to 10 per cent of impurity; and according to Karsten (System der Metallurgie, V. pp. 200, 201) Blicksilber contains at least 12 per cent of lead, and often more, the proportion of lead to silver being entirely dependent upon the temperature of the cupelling furnace. Since such silver is unfit for use, it is refined by small portions, either by a second cupellation at a high heat, or by melting it in crucibles with saltpetre and borax or some other flux (Kerl, op. cit., III. pp. 181 to 198); but it appears to be impracticable, or at least not advantageous in practice, to remove the last traces of lead from the silver by either of these processes. (See Kerl, op. cit., I. 224; or Plattner, Probirkunst mit dem Löthrohre, (Leipzig, 1853,) p. 403).

The silver of the Spanish and Mexican coins was doubtless prepared by the American system of amalgamation.* In this process, speaking in general terms, it would appear that silver is reduced by means of metallic mercury from a solution of chloride of silver in chloride of sodium, the amalgam which is formed being exposed meanwhile to the action of chloride of copper and perchloride of iron.

It is to be inferred from the investigations which we have cited, that under these circumstances the amalgam should be almost entirely free from any contamination with the more strongly electropositive metals.† It would seem, indeed, that the American system of amalgamation furnishes purer silver than is obtained by any of the other processes which are employed upon the large scale.

So far as concerns the occurrence of lead in the silver coin of our own country, it could probably be mainly, if not altogether, avoided, by employing zinc free from lead, such as is manufactured in Pennsylvania, as we have already described in our Memoir.

It is interesting to observe, that the occurrence of lead in some of the silver coins of the ancients has been thought to indicate "that the process of separating lead and silver was less perfectly executed in the ages of antiquity than is at present the case." § Yet, in none

* Described by Humboldt in his Essai Politique sur le Royaume de la Nouvelle Espagne, (Paris, 1811,) II. 558. See also Karsten, Abhandlungen der phys. Klasse der Akad. der Wissenschaften zu Berlin, 1828, p. 1; and Karsten u. Dechen's Archiv für Mineralogie, etc., 1829, I. 161; and again, ibid. XXV. 178, and in Dingler's Polyt. Journ., 1852, CXXVI. 357. Compare the subsequent statements of Boussingault, Ann. Ch. et Phys., 1832, [2.] LI. 350.

In the European system of amalgamation, as practised at Freiberg in Saxony, where the silver is reduced from its chloride by metallic iron instead of quicksilver, a similar degree of purity in the silver is not to be expected. (For analyses of such silver, "Tellersilber," see Kerl, op. cit., I. 234.)

Walchner, Schweigger's Journal für Ch. u. Phys., 1827, LI. pp. 204, 205. J. W. Draper, Silliman's Am. J. Sci., 1836, [1.] XXIX. 160. Sarzeau, Journal de Pharmacie, 1839, XXV. 503. Brüel and Hausmann, Karsten u. Dechen's Archiv für Mineralogie, etc, 1844, XVIII. 505; also in J. pr. Chem., XXX. 334. J. W. Mallet, Trans. Royal Irish Acad., 1853, XXII. 319. Brüel and Hausmann refer also to earlier determinations of lead in antique coins by Klaproth (probably in his Beiträge zur chemischen Kenntniss der Mineralkörper, 1795 – 1780, B. VI.), and by Goebel (doubtless in his brochure, Ueber den Einfluss der Chemie auf die Ermittelung der Völker der Vorzeit, etc., Erlangen, 1842); but they do not indicate precisely where these analyses are to be found, nor have we any means of ascertaining this point.

§ Brüel and Hausmann, loc. cit., Archiv. p 509; J. pr. Chem., p. 338.

of the recorded analyses of ancient silver coins to which we have had access have we been able to find that any one has detected so large an amount of lead in these coins, as we have shown to occur in American fine silver of the year 1860, if we except a single analysis by Professor Draper, who found nearly 3 per cent of lead in a silver coin of Hadrian. The greatest percentage of lead observed by Brüel (the author of the remark just quoted) was only 0.12. It is probable, however, that the methods of analysis none of which are recorded employed for separating lead from silver by the chemists to whose labors we have referred, were less delicate than the process which we have ourselves made use of.

It is interesting also in this connection to observe, that the American system of amalgamation, which at one time—before its peculiar fitness for the circumstances of the case in which it is employed had been recognized was so frequently criticised by European metallurgists, affords silver which is less strongly contaminated with lead, and is probably purer in other respects, than is produced by any other process of manufacture.

II. On the extreme Difficulty of Removing the last Traces of Carbonic Acid from large Quantities of Air.

In the course of our research upon the impurities of zinc, we instituted a series of experiments, in order to ascertain whether the disagreeable odor of hydrogen gas, as generated from common zinc by means of sulphuric acid, could be attributed to the presence of any gaseous compound of carbon,- an apocryphal doctrine which seems to be quite generally credited.

The results of these experiments were entirely negative, in so far as they related to the point in question, and we should not have thought of publishing them, had it not been proved to us, by a memoir recently printed in Poggendorff's Annalen, that one of the phenomena which we then observed had not been sufficiently dwelt upon by chemists.

In our experiments above referred to, a jet of hydrogen was burned in a glass globe thirty centimetres in diameter, through which was drawn, by means of an aspirator, a steady current of air. Before entering the globe, this air had passed through an apparatus, described in full below, which was intended to deprive it of all its carbonic acid, and on leaving the globe it was drawn through a bottle containing

lime-water, carrying with it, of course, the products of the combustion of the hydrogen. We first used, in order to absorb the carbonic acid completely from the air in which the hydrogen burned, two cylinders twenty-seven centimetres high and five centimetres in diameter, filled with fragments of pumice-stone moistened with caustic potash; not satisfied with this large absorption surface, we next added to the cylinders two Wolfe bottles containing a concentrated solution of caustic potash, and finally substituted for the Wolfe bottles three five-bulb potash-tubes. We repeatedly maintained a burning jet of hydrogen in the globe for periods varying from four to six hours, with air purified by passing through the three potash-tubes and two cylinders described, and always obtained the same result; viz. there was never any perceptible cloud of carbonate of lime in the bottle containing lime-water during the progress of the experiment, but after standing twelve hours, an unmistakable deposit of crystalline carbonate of lime was invariably found at the bottom of the lime-water. We might have regarded this as sufficient evidence of the presence of an infinitesimal amount of carbon in the hydrogen, had we not found by repeated trials, that the burning of the hydrogen had no influence whatever on the formation of this crystalline deposit. 156,000 c. c. of air (the contents of our aspirator), passed through the purifiers we have described, still retained sufficient carbonic acid to produce a deposit of crystalline carbonate of lime, when allowed time to separate from the lime-water by crystallization.* In order to render the ex

The fact that carbonate of lime may at first be dissolved by lime-water, has been clearly shown by Vogel. (Schweigger's Journ. f. Ch. u. Phys., 1821, XXXIII. 207.) It is moreover very distinctly affirmed in the following passage from an article, which has recently fallen under our notice, by Berthollet (Annales de Chimie, 1789, III. 68): — "I am indebted to M. Welter for an observation upon the use of lime-water, which may be useful in cases where one wishes to detect small quantities of carbonic acid. Lime-water has the property of dissolving a little carbonate of lime, of which one can assure himself by blowing into it with a tube; the air expired produces a cloud which redissolves entirely, until the limewater is saturated with the carbonate of lime which has been formed. If, therefore, it is desired to detect small quantities of carbonit acid by means of limewater, it is necessary to agitate some carbonate of lime with the latter, in order that it may be saturated, before filtering it." That the carbonate of lime separates after a time from the lime-water which had previously held it in solution, has also been shown by one of us (Am. J. Sci., 1858, [2.] XXV. 42), — at that time entirely ignorant of the experiments of Vogel and Welter.