of PROUST, is to the chlorine as 7.5 to 34.4; and the arsenic being as 21.9, the oxygene, from the analysis of the same chemist, is to the chlorine as 7.3 to 33.6. The analyses of the oxides of the other metals being at variance with those of the chlorine combinations, I was induced to make the following experiments, with the hope of discovering the cause of the difference.

100 grains of lead, which had been precipitated from the nitrat of lead by zinc, were dissolved in nitric acid and thrown down by carbonat of potash. This precipitate of carbonat of lead was well washed and dried and heated to dull redness for a quarter of an hour in a platina crucible; by this treatment all the carbonic acid was expelled; the remaining yellow oxide weighed 107.7 grains, and it dissolved in muriatic acid without effervescing, and without affording any residue of brown oxide. Hence the yellow oxide of lead appears to contain 7.15 per cent. of oxygene. And this proportion of oxygene in the oxide compared with that of chlorine in plumbane, lead being as 97.2 appears to be in the ratio of 7.5 to 33.8, instead of that of 15.6 the estimate of KLAPROTH, or of 11.2 the estimate of Dr. THOMPSON to 33.8. KLAPROTH might have been misled by considering the hydrated oxide as a true white oxide free from water.

According to M. PROUST the peroxide of antimony contains 23 per cent. of oxygene, and the protoxide 18.* I have repeated this chemist's experiments; my results, in which the peroxide is concerned, agree with his; but there is not the same concordance in those relating to the protoxide. The protoxide I used was either prepared by the decomposition of * Journal de Physique, Tom. LV.

the butter of antimony, or of the sulphat, by a boiling solution of carbonat of potash. This oxide, in its purest state, I have always found as M. PROUST describes it, of a light fawn colour before fusion, and afterwards in mass of a gray colour, and of a radiated crystalline texture. 100 grains of it that had been fused were heated in the state of powder with strong test nitric acid in a platina crucible, when nitrous gas ceased to be produced, the excess of nitric acid was expelled by a gentle heat, and the oxide was heated to dull redness, the increase of weight after this, was equal to 10.4 grains; nitric acid was again added and the process repeated, but without any alteration of weight being produced. Hence as the peroxide contains 23 per cent. the protoxide seems to contain 15 per cent.; which proportion of oxygene very nearly agrees with that of chlorine in the butter of antimony, for antimony being as 42.5, the former is to the latter as 7.5 to 34.6, instead of 33.6. I put some confidence in this estimate of the proportion of охуgene in the protoxide, not only on account of its agreement with the analysis of the butter of antimony, but because it was confirmed on the repetition of the experiment.

KLAPROTH Concludes from his experiments, that the oxide of bismuth, prepared by means of nitric acid, contains 17.7 per cent. of oxygene, and in consequence this oxide has been considered distinct from that which is formed by direct calcination of the metal, and which contains a much smaller proportion. But there is reason to believe that this difference does not really exist, and that there is only one known oxide of bismuth, and that KLAPROTH'S oxide was an hydrated oxide; for I have found that 100 grains of bismuth, converted by nitric acid into oxide, precisely in the same manner as the

protoxide of antimony was more highly oxidated, gained only 11.1 grains. KLAPROTH did not heat his oxide to redness, and hence apparently the discordance. From the above result, which I have confirmed by repetition of the experiment, oxide of bismuth seems to contain 10 per cent. of oxygene and bismuth being as 67.5, the oxygene in the oxide is to the chlorine in the butter of bismuth, as 7.5 to 34.2.

6. On the Relation between the Proportion of Sulphur in the Sulphurets, and the Proportion of Chlorine in some of the Combinations of Chlorine and the Metals.

The last section afforded proofs of the useful application of the general analogy of definite proportions in correcting the results of chemical analyses. In the present section, it is my intention to pursue a little further, the plan that I have adopted in the preceding, and to apply another test to the analyses of the combinations of the metals and chlorine, by comparing some of them with the combinations of the same metals and sulphur.

I was first led to examine the sulphurets of tin on a different account. Aurum musivum, it has been observed, is formed when stannane is heated with sulphur. According to M. PROUST, this substance is a sulphuretted oxide of tin. Were this opinion correct, an argument might evidently be deduced from it, in favour of the existence of oxygene in chlorine. To satisfy myself respecting this, I endeavoured to ascertain whether any sulphureous acid gas is produced by the decomposition of aurum musivum by heat, as it is commonly asserted. I heated to redness in a bent luted green glass tube connected with a pneumatic mercurial apparatus about 20 grains of

aurum musivum, prepared by the decomposition of stannane with sulphur, no more gas was produced than the expansion by heat occasioned, sulphur sublimed, and a gray sulphuret of tin remained. These results I have several times obtained, and not only with aurum musivum prepared as the preceding, but with some also made according to WOULFE's process. As no sulphureous acid gas was produced, and as sulphur sublimed, it may be concluded that aurum musivum differs merely from the gray sulphuret in containing a larger quantity of sulphur. My next object was to ascertain the exact proportion of sulphur in both these sulphurets, for the sake of comparison with the combinations of tin and chlorine.

100 grains of tin in a finely divided state, as precipitated from the muriat of this metal by zinc, were heated in a glass tube intimately mixed with sulphur, the combination of the two was accompanied with vivid ignition, the sulphuret formed weighed 127.3 grains, and broken, it appeared perfectly homogeneous; it was pounded, and again heated with sulphur; but the excess of sulphur being expelled, the fused sulphuret had not increased in weight. The second time I made this experiment, I obtained the same result.

50 grains of aurum musivum, purified from mixed sulphur by exposure in a close vessel to a dull red heat, were decomposed by a bright red heat in a small green glass tube nicely weighed, and having only a very small orifice; the loss of sulphur, by conversion into the gray sulphuret, was equal to 9.3 grains. Hence, as 40.7 grains of gray sulphuret contain 8.72 grains of sulphur, 50 grains of aurum musivum appear to contain 18.02 grains.

The ratio in which sulphur combines with bodies is to that

in which oxygene and in which chlorine combines, as 15 to 7.5 and 33.6. This appears from the proportions of the constituent parts of sulphuretted hydrogene and sulphureous acid gas, for I have found 100 cubic inches of the former to weigh 36.64 grains, and 100 of the latter 68.44 grains. In the comparison, therefore, between the sulphurets of tin and the combinations of this metal and chlorine, 15 by weight of sulphur are equivalent to 33.6 of chlorine. And the tin being as 55, it appears from the analysis of the gray sulphuret and stannane, that the sulphur is to, the chlorine as 15 exactly to 33.4; and from the analysis of the other two compounds, aurum musivum and the liquor of Libavius, as 15.5 to 33.5, or as 31 to 67.

The proportions of sulphur in the two sulphurets of iron, do not accord with the proportions of oxygene in the oxides, or of chlorine in the chlorine combinations; but I am yet ignorant of the cause of this difference.

100 grains of lead, heated with sulphur in a glass tube, afforded, in two trials, 115.5 grains of fused sulphuret. Hence lead being as 97.2, the sulphur is to the chlorine in the respective combinations as 15.09 to 33.8.

Sulphuret of antimony contains 25.9 per cent. of sulphur. Hence antimony being as 42.5, the sulphur in the sulphuret is to the chlorine in the butter of antimony, as 14.86 to 34.6.

100 grains of bismuth heated with sulphur afforded 122.3 grains of sulphuret. Hence bismuth being as 67.5, the sulphur is to the chlorine as 15.08 to 34.2.

In the following table, the proportions are collected in which chlorine, sulphur, and oxygene combine with several metals; the numbers representing the metals are kept constantly the same, for the greater facility of comparison.