COMBINING NUMBERS OF THE ELEMENTS. 917 composed by heat, left 60.839 of chloride of potassium (KCl); and 22.032 of pure silver required 15 216 of chloride of potassium for its complete precipitation. 14'427 of chloride of potassium gave 27.749 of chloride of silver (AgCl). Berzelius calculates from these results that the equivalent of chlorine is 35'46; and this coincides exactly with the recent elaborate experiments of Stas. Maumené, by heating chloride of silver in a current of hydrogen, found that 100 parts of silver were united with 32.856 of chlorine. The same chemist obtained from 100 parts of chlorate of potassium 60.791 of chloride of potassium: and from 100 parts of chloride of potassium, he obtained by precipitation 19275 of chloride of silver. These experiments furnish data from which the atomic weights of potassium and silver may be determined, as well as that of chlorine, in the manner following:- The composition of chlorate of potassium is represented by the formula KCl03; when heated it gives off the whole of its 3 atoms of oxygen. The atom of chloride of potassium therefore will be the quantity which is combined with 48 parts (or 3 atoms) of oxygen. Now, taking Maumené's result that 39.209 parts of oxygen are combined in chlorate of potassium with 60791 of chloride of potassium, we have 39 209:48:: 60'791: x (=74°4208, 1 at. of KCl). If 100 parts of chloride of potassium produce 192.75 of chloride of silver, I atom, or 74'4208, of chloride of potassium will furnish 1 atom of chloride of silver. For 100: 192'75:: 74°4208 : ≈ (=143°446, 1 at. of AgCl) ; and 132.856 of chloride of silver contain 32.856 of chlorine; consequently (I atom of chloride of silver containing 1 atom of chlorine), we find the atomic weight of chlorine as follows: 132.856:32.856:: 143'446: x (=35'476): but the atomic weight of chloride of silver being that of silver is found by deducting the at. wt. of Cl leaving the atomic weight of silver. and the atomic weight of chloride of potassium being deduct from it the atomic weight of chlorine . =143'445 = 35'476 =107'970 =74°4208 =35'476 918 DATA FOR THE DETERMINATION OF THE No material error can therefore arise if the atomic weight of chlorine be taken as = 35'5 =108.0 = 39°0 Dumas has also checked these numbers by burning finely divided silver in a current of perfectly dry chlorine. He thus found that 108 parts of silver combined with 35'505 of chlorine as a mean of two experiments. 12. Chromium.-The atomic weight of chromium was determined by Berlin, by converting chromate of silver (Ag2¤rÐ ̧) into the chloride; the number calculated from this result is 52.694; and that deduced from the reduction of the chromic anhydride (Єre) to the sesquioxide of chromium, in the same series of experiments, is 52'54. Péligot's experiments on the acetate would make it between 52 and 536, but he considers 52:48 as nearest the truth. 13. Cobalt.-Rothoff found that 269 2 parts of the protoxide of cobalt (Co), when converted into chloride (CoC) by means of hydrochloric acid and precipitated by means of nitrate of silver, gave 10299 of chloride of silver: hence the atomic weight of cobalt is 58.98. The number 59'08 is the mean of 5 experiments made in a similar way by Dumas. Schneider, however, subsequently, from the decomposition of the oxalate, obtained results corresponding to the number 60; but the more recent experiments of Russell confirm the former number, and show that the atomic weight is identical with that of nickel. 14. Copper.-Berzelius obtained from 7.68075 grammes of oxide of copper (Cu), which were reduced in a current of hydrogen, 6·13075 of metallic copper; hence the atomic weight of the metal is 635. Erdmann and Marchand, by a similar method, obtained numbers which would make it 63:52. 15. Fluorine.-Berzelius found that 100 parts of fluor-spar (CaF) when heated with an excess of sulphuric acid, yielded 175 of sulphate of calcium (CaSO4). Louyet, on repeating this experi ment, obtained 174.361 parts of sulphate of calcium. The equivalent of fluorine deduced from this latter result is 19. (See p. 163.) And these results have been confirmed by Dumas, who made similar experiments upon the fluorides of calcium, potassium, and sodium. 16. Glucinum.-Awdejew found that 100 parts of the chloride of this metal contained 88.42 of chlorine; hence, if the chloride be represented as Gl,Cl, the combining number of the metal will COMBINING NUMBERS OF THE ELEMENTS. 919 be 6·97; but if the chloride be regarded as GCl2, the atomic weight of the metal is 9.30. 17. Gold.-Berzelius, by reducing the double chloride of gold and potassium in a current of hydrogen, determined the combining number of this metal at 196 66. By an earlier series of experiments he found that 142'9 of metallic mercury precipitated 93'55 of gold from the terchloride (AuCl2); 3 atoms of mercury causing the precipitation of 2 atoms of gold: and assuming the atomic weight of mercury to be 200, this would make the number for gold 196:44. 18. Hydrogen.—The equivalent of hydrogen was determined with great care by Dumas, by the method already described at p. 52. He ascertained, as a mean of nineteen experiments, that 8 parts of oxygen combined with 10012 of hydrogen to form water; the lowest quantity which these experiments gave being o'9984, the highest 10045. The quantity of water collected in each of these experiments was considerable, varying from 230 to 1100 grains. Erdmann and Marchand repeated these experiments with similar results. Berzelius and Dulong concluded, from researches performed long previously upon a similar principle, though on a smaller scale, that the quantity of hydrogen united with 8 parts of oxygen was o'9984, which coincides with the lowest number obtained by Dumas. It is obvious that no appreciable error can be committed by assuming hydrogen to possess an atomic weight of 1, that of oxygen being 16, if water be taken as H2O. 2 19. Iodine.-De Marignac determined the number for iodine by a process analogous to that which he employed for chlorine. The atomic weight of iodide of potassium (KI) he fixed at 165'951; deducting from this 38.95, Maumené's number for potassium, we obtain 127 as the combining number of iodine. Dumas, by two experiments upon the iodide of silver, which he converted into chloride by heating it in a current of dry chlorine, obtained the same result. 20. Iridium.-Berzelius deduced the number for iridium from an analysis of the chloride of iridium and potassium (2 KCl,FrCl); his results would make it 197 12, which is identical with the number obtained for platinum. 21. Iron.-Berzelius found that 1586 grammes of pure iron converted first into nitrate, and then into sesquioxide by ignition, gave 2 265 of sesquioxide (Fe,3); and Svanberg and Norlin, by reducing sesquioxide of iron in a current of hydrogen, obtained from 35 783 of sesquioxide, 25'059 of metallic iron; making the 920 DATA FOR THE DETERMINATION OF THE atomic weight of iron 56:08. Erdmann and Marchand, by the method last named, fixed the atomic weight at 56·002, and Maumené has also arrived at a similar result by dissolving iron in aqua regia, and precipitating the oxide by means of ammonia. Still more recently, Dumas has corroborated the same number by decomposing the chloride of iron by means of nitrate of silver-the mean of his four experiments giving 56'14. 22. Lead.-21.9425 grammes of oxide of lead (Pb) were reduced by Berzelius in a current of hydrogen, and gave 20'3695 of metallic lead; from the mean of his five experiments, the atomic weight of the metal would be 207.14. This result has been confirmed by De Marignac, who obtained, by the precipitation of 5 grammes of chloride of lead (PbCl2) 3·8835 of chloride of silver; similar experiments by Dumas would make the number for lead 207'1, whilst from those of Stas it would be 206'912. 23. Lithium. The number obtained by Berzelius for this metal, by neutralizing fused carbonate of lithium (Li,,) with sulphuric acid, is probably inaccurate, as the carbonate of lithium has subsequently been found to lose a little of its acid when melted. Mallet estimates the atomic weight of lithium at 6'97. Diehl, from an analysis of the carbonate, obtained the number 7'026; and Troost, from the chloride, obtained the number 701, as well as by decomposition of the carbonate by means of silica. therefore the number adopted. 7:00 is 24. Magnesium.-Berzelius found that 100 parts of magnesia, dissolved in pure sulphuric acid and ignited, gave 293.985 of sulphate of magnesium (MgSO); hence the equivalent of magnesium would be 25'3; but this result is probably a little too high. Scheerer, by ascertaining the quantity of sulphate of barium produced by a given weight of sulphate of magnesium, determined the number for magnesium at 24:22. The results of Svanberg and Nordenfeldt, by the decomposition of the oxalate of magnesium (MgC,,,2H,O) by heat, made it 24'7; those by converting a known weight of magnesia into sulphate, gave it as 24'74 whilst those of Marchand and Scheerer, by ignition of the native carbonate, indicated the number 24°04. Dumas found extraordinary difficulty in procuring chloride of magnesium quite free from magnesia. The mean of II experiments upon the precipitation of the chloride (MgCl) by nitrate of silver gave 246 as the combining number of magnesium. The mean of these results is 24:32. 25. Manganese.—4°20775 of chloride of manganese (MnCl,) gave Berzelius 9'575 of chloride of silver; the atomic weight of COMBINING NUMBERS OF THE ELEMENTS. 921 the metal, from a mean of two such experiments, is 55'14. Dumas, as a mean of five such experiments conducted in a similar manner, obtained the number 54'96. 26. Mercury.-Erdmann and Marchand obtained from 118.3938 grammes of red oxide of mercury (Hg) 109.6308 of mercury; a mean of five experiments gave 2002 as the equivalent of the metal. It may be safely estimated at 200. 27. Molybdenum.-Berzelius regarded the number originally given by himself for this metal, only as an approximation. Svanberg and Struve, from an extensive series of experiments, considered that the most accurate results were obtained by roasting the bisulphide of molybdenum in air, and they conclude that 100 parts of the bisulphide (MOS) yield 89.732 of molybdic anhydride (Mo); hence, if the equivalent of sulphur be taken at 32, that of molybdenum will be 92.12. Berlin, from the quantity of molybdic anhydride left by the salt (HAN),0,5 Mo ̧,3 H2, found the number for the metal (from a mean of four experiments) to be 91.96. Dumas, however, by reducing crystallized molybdic anhydride in a current of hydrogen, found, as a mean of 5 concordant experiments, 96 as the number for this metal. 28. Nickel.-Rothoff converted 188 parts of oxide of nickel into chloride (NiCl2), and obtained from it 718.2 of chloride of silver; the number for nickel hence deduced is 59'08; and Dumas, as a mean of 5 experiments on the same plan, obtained the number 59'02. Russell found it by reducing the oxide in a current of hydrogen 58'738. 29. Nitrogen.-De Marignac, by converting 200 grammes of silver into nitrate (AgNO3), obtained 314.894 of the salt; 14'110 of nitrate of silver required for precipitation 6·191 of chloride of potassium; 10'339 of silver converted into nitrate required 5120 of chloride of ammonium for complete precipitation; the mean result, as calculated by Berzelius from several experiments performed in this manner, gives 14'004 as the number for nitrogen. Stas, by synthetic experiments upon nitrate of silver, fixed it at 14'041. Anderson, by the decomposition of nitrate of silver by heat, concluded that the atomic weight of nitrogen was 13'95; and Svanberg, from the analysis of nitrate of lead, obtained the same result: the number for nitrogen may therefore be taken as 14. 30. Osmium. The atomic weight of this metal was calculated by Berzelius from the result obtained by heating the chloride of osmium and potassium (2KCl,OsCl) in a current of hydrogen, 13165 grammes of the salt leaving o'401 gramme of KCl and |