=

oxide. If C75.4, and S 200, the value follows. (Phil. Trans., 104, 1814, 20.)

100 parts of barium chloride gave 138.08 and 138.06 parts argentic chloride. [If Ag = 107.93, and Cl 35.457, the above value follows.] Berzelius also determined barium from the sulphate; 100 parts barium chloride gave 112.17 and 112.18 parts sulphate. Calculated for S200.75 this determination is almost identical with the other; Berzelius, however, expressly adopts the former. [Calculated for S 32.0742, it gives 135.74.] (Poggend. Annal., 8, 1826, 189, and Lehrbuch der Chemie, 5th ed., 3, 1229.)

E. TURNER: 137.4 (0 = 16).

=

Turner determined the chlorine contents of barium chloride at 34.016 per cent. by precipitation with silver. This number was the mean of the best two experiments made, and the value follows from it on the assumption that Cl= 35.42. The barium chloride was prepared from native carbonate by solution in chlorhydric acid, precipitation of impurities with barium oxide, ignition of the chloride, treatment with alcohol, and recrystallization. (Phil. Trans., 119, 1829, 291.*)

=

100).

T. THOMSON: 136 (0 = 16); 850 (O Thomson had formerly determined this atomic weight at 875 by mixing potassic sulphate with barium chloride in such proportions that the supernatant liquid contained no sensible amount of either sulphuric acid or barium. Turner having shown the fallacy of this method, Thomson substituted ammonium sulphate, and also sulphuric acid for the potassium salt, and found 9.5006 barium oxide equivalent to 5.00 sulphuric anhydride. He also analyzed the chloride with argentic nitrate, assuming silver 1375, and chlorine = 450, and reached the same conclusion with reference to barium. (Thomson's System of Chemistry, 7th ed., 1, 1831, 426.)

=

*Turner made the discovery in the course of this investigation that barium sulphate carries down other salts, such as potassic sulphate, which cannot be extracted from the precipitate by any degree of washing, and that determinations, with barium sulphate, are consequently unreliable. Although Berzelius drew attention to the importance of the observation, and Thomson was obliged to acknowledge errors in his work from this cause, the fact was for a long time nearly forgotten, as can readily be proved from the contents of this digest.

SALVETAT: 136 (0 = 16); 850 (0 = 100).

Determined from the loss of weight ensuing on the decomposition of barium carbonate by sulphuric acid. Details not given. (Paris Comptes Rendus, 17, 1843, 318.)

J. PELOUZE: 137.28 (0 = 16); 858.01 (0 = 100). Into a nitric acid solution of a known weight of perfectly pure silver, a known and slightly more than equivalent weight of barium chloride was introduced. The excess was titrated with decimal silver solution. The value is the mean result of three experiments, which give an extreme difference of 0.22 for O= 100. The barium chloride was purified by recrystallizations continued till determinations gave a constant result, and was dessicated in part at 200°, and in part at a temperature just below redness. Pelouze took Ag=1349.01, and Cl=443.2. (Paris Comptes Rendus, 20, 1845, 1047.)

C. MARIGNAC: 137.08 (016); 856.77 (O = 100).

Determined by six experiments on the equivalence of silver and barium chloride performed by Pelouze's method, . (vide supra.) 100 silver were found equivalent to 96.365 barium chloride; extreme difference, 0.038; hence the value taken. Marignac takes Ag= 1349.01, and Cl=443.2. The barium chloride was purified as follows: Commercial chloride was crystallized from boiling aqueous solution; the crystals were heated to redness, dissolved in boiling water,. treated with carbon di-oxide, filtered and crystallized, and these crystals were washed with alcohol and again recrystallized. Determinations were made at each stage and the purification was continued until constant results were obtained. (Liebig, Annal., 68, 1848, 214; Bibl. Univ., Arch. des Sciences, 8, 265.)

H. STRUVE: 136.26 (0 = 16).

100 parts of barium chloride gave 112.0938 parts of sulphate as a mean of two experiments; extreme difference, 0.005. S = 32; Cl=35.4624. (Liebig, Annal., 80, 1851, 204; Oefversigt af Kongl. vet. Acad. Foehr., 6, 165.)

T. ANDREWS: 137.578 (0 = 16).

Andrews obtained this number from two nearly coincident experiments of which he gives no details. (Brit. Assoc. Rep., 1852, pt. 2, 33.)

C. MARIGNAC: 137.16 (016).

Three experiments were made on the titration of air-dried barium chloride in crystals by Pelouze's method, (vide supra.) Five grammes of the salt required for precipitation (1) 4.4205; (2) 4.4195; (3) 4.4210 grammes silver. Three experiments were made on the conversion of the same barium chloride into sulphate. Ten grammes of the salt gave (1) 9.543; (2) 9.544; (3) 9.542 grammes sulphate. In each of the latter experiments the water was determined, and was found to vary no more than 0.0005 grammes. Comparison of the two series gives for Ag=108, S16, and 0 = 8; barium equal to (1) 68.57; (2) 68.61; (3) 68.55; in mean 68.58, or one-half of 137.16. This result is independent of the possible trace of water the chloride might have contained. In another series of three experiments the water was driven off at a low red heat and determined, and the salt analyzed by Pelouze's method. It was proved that barium chloride is not decomposed at the temperature employed. (1) gave 68.61; (2) 58.59; and (3) 68.55, or a mean of 68.583. The salt for the experiments marked (1) was prepared by recrystallization and precipitation with alcohol; that for (2) by a repetition of the same process, and for (3) by resolution of (2) and precipitation with chlorhydric acid gas. Marignac proved that the precipitated argentic chloride contained entirely insignificant traces of barium salt. Cl Cl35.5. (Bibl. Univ., Archives des Sciences, Nouv. Série., 1, 1858, 209.)

J. DUMAS: 137 (0 = 16).

Determined by fifteen experiments on the titration of barium chloride with argentic nitrate, which give a general average of 68.516 with an extreme difference of 0.11. The barium chloride was prepared from pure nitrate and pure carbonate, and from commercially pure chloride after it had been freed from lead by precipitation with barium sulphide. The chloride was precipitated from solution by chlorhydric acid gas and melted in a current of chlorine to prevent oxidation. Ag=108; Cl = 35.5. (Annales de Chimie et de Physique, (3,) 55, 1859, 137.)

BERYLLIUM.

The atomic heat of beryllium has been determined by J. Emerson-Reynolds by direct comparison with that of silver.

In a calorimetric apparatus constructed for the purpose, the amount of heat given off during cooling by 108 parts of silver heated to 100° was found to be equal to that communicated by a little more than 9.2 parts of beryllium under the same conditions. Assuming the atomic weight of the metal to be 9.2, the atomic heat found would be 5.91. The smallness of this number the observer accounts for by supposing that there was a trace of platinum present introduced by the use of platinum vessels in the course of reduction. (Phil. Mag., (5,) 3, 1877, 38.)

J. J. BERZELIUS: 14.5 (0 = 16).

0=

3

Berzelius analysed the salt formed by saturating dilute sulphuric acid with beryllium oxide. From the amount of barium sulphate obtained he inferred that the atomic weight of beryllium was 331.261 on the supposition that the oxide was Be, + 0, and that the salt was neutral. Berzelius took 0 = 100; S=200.75, and Ba= 855.29. [Awdejew having discovered that this salt is basic, this value is reduced to 90.63; or, for 016, to 14.5.] Berzelius accepted Awdejew's determination in preference to his own. (Poggend. Annal., 8, 1826, 187; and Lehrbuch der Chemie, 5th ed., 3, 1225.)

T. THOMSON: 36 (0 = 16).

Experiments not given. The value is four times nine, and may have arisen from a mistake as to saturation. (System of Chem. 7 ed., 1, 1831, 459.)

AWDEJEW: 13.85 (0=16); 86.58 (0

=

= 100). Beryllium sulphate, in chlorhydric acid solution, was decomposed with barium chloride. In the filtrate the excess of barium chloride was precipitated with sulphuric acid, and the beryllium oxide thrown down with ammonia, dried, heated, and weighed. The beryllium sulphate was prepared from pure carbonate by treatment with sulphuric acid and precipitation with alcohol. It was purified by recrystallization. Four experiments were made, the mean of which calculated for S201.165, gave Be 58.084 with an extreme difference of 1.955. (Poggend. Annal., 56, 1842, 106.) Weeren recalculated these analyses for S=200 and got 57.72, [or of 86.58.] (Poggend. Annal., 92, 1824, 124.)

=

J. WEEREN: 13.83 (0 = 16); 86.46 (0 = 100). Weeren followed the same method as Awdejew, except that he precipitated the beryllium with ammonium sulphide, the oxide being soluble in excess of ammonia. The mean of four experiments gave 57.64, the extreme difference being 1.52 for O=100, [57.64 is of 86.46.] took S=200. (Poggend. Annal., 92, 1854, 124.)

G. KLATZO: 13.89 (0 = 16).

Weeren

Klatzo made five analyses of the sulphates containing seven and four molecules of water, precipitating the sulphuric acid as barium sulphate, and the beryllium as oxide by means of ammonia. From a comparison of the sum of the oxide found in all the analyses with the total amount of barium sulphate found, Klatzo deduces Be=9.227, for Ba= 137, and S32. [If Ba is taken equal to 137.16, and S= 32.07, and if each of the analyses is calculated for itself, Be 13.89. The extreme difference is 0.45.] The sulphates were purified by recrystallization, and treatment with alcohol. (Erdmann's Journ. für Prak. Chemie, 106, 1868, 227; Klatzo, Ueber die Constitution der Beryllerde, Dorpat, 1868.)

L. F. Nilson and O. Pettersson have redetermined the specific heat of beryllium within a few weeks. They find the specific heat 0.4079, corresponding to a trivalent metal and a sesqui-oxide. The investigation seems to have been made with great care, while that of Emerson-Reynolds was merely preliminary. (Berlin, Bericht der chem. Ges., 11, 1878, 386.)

BISMUTH.

Dulong and Petit, Regnault, and Kopp, have determined the specific heat of Bismuth. It corresponds to an atomic weight of about 210. (Gmelin-Kraut, l. c.)

=

P. LAGERHJELM: 212.86 (0= 16); 1330.377 (0 = 100). Metallic Bismuth was oxidized in a weighed vessel by nitric acid, and the nitric acid expelled by heat. 10 grammes of bismuth gave 11.1275 oxide. (Berzelius Lehrbuch der Chemie, 5th ed., 3, 1216; Stockholm, Akad. Handl., 34, 1813, 219.)