J. J. BERZELIUS: 6.533 (0=16); 40.83 (0 = = 100). Berzelius found that 1.874 lithium sulphate gave 3.9985 barium sulphate, and calculated this relation for S 200.75; Ba855.29. He also found 4.4545 melted carbonate = 6.653 sulphate, but rejected the analysis. (Lehrbuch, 3, 1229, and Jahresbericht, 10, 1830, 96.)

R. HAGEN: 6.57 (0 = 16).

Hagen precipitated lithium sulphate with barium chloride, and found that 0.852 dry lithium sulphate gave 1.8195 barium sulphate whence he calculates Li= 6.493. [If Ba= 137.08; S = 32; this relation gives Li= 6.57.] (Poggend. Ann., 48, 1839, 363.)

J. W. MALLET: 6.95 (016); 86.89 (0 = 100).

In two experiments a known weight of lithium chloride was precipitated by argentic nitrate, and the argentic chloride weighed. In one experiment lithium chloride was titrated with argentic nitrate by Pelouze's method. The number is the mean; the extreme difference is 0.18 for O 100. Mallet takes Ag=1349.66; Cl=443.28. The alkalis were separated from the lithium salt by repeated treatment with ether and alcohol. The salt was examined for impurities, and was fused with a little ammonium chloride to prevent the formation of oxy-chloride. (Silliman's Amer. Journ., (2,) 22, 1856, 349.)

=

L. TROOST: 6.5 (016).

Troost found this number from analysis of the carbonate which had been crystallized from water containing carbon di-oxide and dried at 200°, but does not regard it as definitive. (Annal. de Chim. et de Phys., (3,) 51, 1857, 111.)

J. W. MALLET: 7 (0 = 16).

Troost having objected to Mallet's former method of determination, he redetermined it by precipitating the sulphate with a standard solution of barium chloride, the precipitating power of which had been tested on the sulphates of magnesium and sodium. This method was adopted to avoid the well-known imperfections of the sulphur determination. Compared with sodium sulphate the atomic weight of Li was found = 6.92 and 6.95. Compared with magnesium sulphate it was found 7.07 and 7.09.

=

Mg=24; Na 23. The sulphate was prepared from carbonate, and dried somewhat below a red heat. (Silliman's Amer. Journ., (2,) 28, 1859, 349.)

K. DIEHL: 7.026 (0 = 16).

Determined by analysis of lithium carbonate with Bunsen's apparatus and in his laboratory. Four experiments; extreme difference, 0.024. C=12. The salt was purified from alkalis by precipitation as carbonate, resolution in acid and reprecipitation, repeated until the sodium line was no longer visible. Diehl found that precipitation of the sulphate with barium chloride gave a nearly constant error on account of the retention of lithium in the precipitate, and led to nearly the same results as Berzelius got. (Liebig's Ann., 121, 1862, 93.)

L. TROOST: 7 (0 = 16).

Troost found 1.309 grammes lithium chloride = 4.42 argentic chloride, and 2.75 lithium chloride = 9.3 argentic chloride. From these analyses he deduces the values 7.03 and 6.99. By decomposing the carbonate, dried at 100°, with silicic acid, he found 0.97 carbonate 0.577 carbon di-oxide and 1.782 carbonate 1.059 di-oxide, and infers for Li 7 and 7.02. In one experiment the carbonate was converted into sulphate. 1.217 carbonate gave 1.808 sulphate. Troost calculates Li=7.06. [If CI= 35.457; Ag =107.93; C=12; S= 32; these determinations give, in the same order as above, 7.01; 6.94; 6.98; 7.02; 7.07.] The carbonate was purified by solution in water containing carbon di-oxide, and reprecipitation by boiling, the operation being repeated until the salt was spectroscopically pure. (Paris Comptes Rend., 54, 1862, 366.)

J. S. STAS: 7.022 (0 = 16).

According to the mean of three determinations, 100 parts of silver 39.358 lithium chloride; extreme difference, 0.005. If Ag=107.93; Cl = 35.457; this ratio gives Li =7.022. This value is confirmed by experiments on the conversion of the chloride into the nitrate, the results of which give Li=7.018. The comparison with silver was made according to Pelouze's modification of the silver titration method. The chloride was purified from alkalis, after preliminary treatment with ether and alcohol, by pouring the dissolved salt into a boiling solution of ammonium car

bonate containing ammonia in excess. All weighings reduced to vacuum. (Stas, Untersuch. über Chem. Prop., Leipzig, 1867.)

MAGNESIUM.

Regnault and Kopp have each determined the specific heat of this metal. It answers to an atomic weight of about 24. (Gmelin-Kraut, l. c.)

J. J. BERZELIUS: 25.3 (0 = 16); 158.139 (O = 100). Determined by dissolving magnesium oxide in dilute sulphuric acid, evaporating and heating to incipient redness. 100 oxide were found 293.985 sulphate. The sulphate was perfectly soluble in water and had therefore lost none of its acid. The oxide was purified by solution in an aqueous solution of carbon di-oxide and reprecipitated by boiling. S200.75. (Poggend. Ann., 8, 1826, 188; and Lehrbuch, 3, 1227.)

Marchand and Scheerer recalculated this analysis for S = 200 and reached the value 157.74. They assert that the oxide may have contained alkalis and that the sulphuric acid carries off magnesium sulphate in volatilizing. (Erdmann's Journ. für Prak. Chem., 50, 1850, 392.)

W. HENRY: F. H. WOLLASTON: 23.36 (0 =
(0 = 100).

16); 146

Henry found that magnesium sulphate contained 33 per cent. magnesium oxide. If S 200 the value follows.

(Phil. Trans., 104, 1814, 21.)

-LONGCHAMP: 15.35 (0 = 16).

In two experiments, Longchamp found that five parts of crystallized magnesium sulphate are equivalent to 4.91 barium sulphate. [If Ba=137.08; S32, the number follows.] Longchamp found 53 per cent. water which is much too high. According to Marchand and Scheerer, the data for the anhydrous salt give Mg= 97.37, for S = 200; Ba = 856.8, [or 15.74.] (Annal. de Chim. et de Phys., 12, 1819, 265.)

L. J. GAY-LUSSAC: 23.62 (0 = 16).

16.205 grammes crystallized sulphate were found equal to 15.345 barium sulphate, and 19.395 magnesium sulphate

to 18.3455 barium sulphate. Calculating from the anhydrous salt Gay-Lussac found from these experiments respectively Mg=147.23 and Mg= 148.09 for Ba= 856.8; S=200. The salt was found to contain 51.43 water. [Calculated from the anhydrous salt these data give Mg =23.55 and 23.68. Calculated from the hydrous salt (7 molecules water) the numbers give 24.14 and 24.41, if S = 32; Ba= 137.08.] Gay-Lussac remarks that the sulphate is partially decomposed at a red heat. (Annal. de Chim. et de Phys., 13, 1820, 308.)

T. SCHEERER: 24.16 (0 = 16); 150.97 (O= 100). Mean of six experiments on the precipitation of the sulphate with barium chloride. Extreme difference, 0.79. S =200.75; Ba=855.29. After weighing, the barium sulphate was treated with dilute HCl and the chloride thus extracted allowed for. (Poggend. Ann., 69, 1846, 535.)

T. SCHEERER: 24.21 (0 = 16); 151.33 (0=100). Barium sulphate formed as in the last determination was fused with soda, the barium carbonate dissolved in HCl, and reprecipitated as sulphate. In the filtrate additional magnesia was found. If the error in the former determination was the same, its corrected value would be as above. (Poggend. Ann., 70, 1847, 407.)

SVANBERG and NORDENFELDT: 24.72 (0 = 16);

154.504 (O = 100).

Four experiments were made on the calcination of the oxalate, and three on the conversion of the magnesia so obtained into sulphate. The oxalate was dried at from 100° to 105° and heated to redness until the weight was constant. The oxide was dissolved in sulphuric acid, evaporated and the excess driven off by heat. The oxalate was prepared from the sulphate by precipitation with sodium carbonate and digestion with oxalic acid. The number is the mean of all experiments; extreme difference, 0.514. = 200.75; C 75.12; H=12.48. (Erdmann's Journ. für Prak. Chem., 45, 1848, 473.)

S

=

According to Marchand and Scheerer, the data give Mg =154.27 for S = 200; H=12.5; C = 75.

MARCHAND and SCHEERER: 24.03 (0 = 16); 150.19 (0 = 100).

Eleven experiments were made in calcining massive magnesium carbonate from Frankenstein, and weighing the

caustic magnesia formed. The carbonate was dried at 300°, and the carbon di-oxide, which passes off above 230°, was caught by caustic baryta solution and determined. The traces of carbon di-oxide not expelled by a yellow heat were set free by solution in chlorhydric acid and also determined as barium carbonate. The silicic acid, etc., were also determined. The mean in air is 150.34; in vacuo as above. Extreme difference, 0.57. C-75. Eleven other experiments were made with comparatively impure material and less precaution, tending to confirm the above. (Erdmann's Journ. für Prak. Chem., 50, 1850, 409.)

T. SCHEERER: 24 (0 = 16); 150 (0 = 100).

By separating the neutral sulphates of magnesium and calcium by means of alcohol, Scheerer found that the magnesites used by Marchand and himself contained from onefourth to one-half per cent. calcium oxide. This correction would make their determination almost exactly 250 or 24. (Liebig's Ann., 110, 1858, 236.)

V. A. JACQUELIN: 24.408 (0 = 16); 152.55 (0 .100). Anhydrous, neutral magnesium sulphate, obtained by solution of the oxide in sulphuric acid and heating to redness, gave 33.56 per cent. pure oxide. The method adopted is not described. This oxide by treatment with sulphuric acid gave the original amount of sulphate. If S 200, the number follows. (Annal. de Chim. et de Phys., (3,) 32, 1851, 195.)

A. MACDONNELL: 23.9 (0 =16).

=

Determined from analyses of anhydrous and of crystaltized magnesium sulphate. (Brit. Assoc. Rep., 1852, part 2, 36; and Kopp's Jahresbericht, 5, 364.)

J. F. BAHR: 24.77 (016); 154.842 (0 = 100). A known weight of purified magnesium oxide was dissolved in sulphuric acid, evaporated and heated to redness till the weight was constant. The number is the mean of three experiments; extreme difference, 0.515. The oxide was obtained from meteoric olivin. After removal of the heavy metals, the solution was evaporated to dryness with soda, washed and heated to redness. The oxide so obtained was dissolved in acetic acid, oxalic acid was added, the