CALCIUM. 27 



C. Marignac: Jf^O.^OS (0 = 16); 251.3 (O = 100). 



Determined by precipitating calcium chloride with argen- 

 tic nitrate ; Ag = 1349.01 ; CI = 443.2. Marignac laid no 

 weight on this determination finding it impossible to pre- 

 pare calcium chloride which did not show an alkaline re- 

 action. The presence of caustic lime would make the result 

 erroneously high ; no doubt Berzelius' early analysis was 

 defective from the same cause. {Berzelius' Jahresbericht, S4-> 

 1844, 103; Bibl. Univ., ^6, 1843, 367.) 



Erdmann and Marchand : 40.007 (O = 100). 



Four experiments were made on the calcination of cal- 

 cium carbonate enclosed in a double platinum crucible in a 

 wind-furnace, till the weight was constant. A mean of 56 

 per cent, calcium oxide was found with an extreme dif- 

 ference of 0.05. This gives Ca = 40 for C = 12. Two ex- 

 periments were made by decomposing calcium carbonate 

 by sulphuric acid. These gave a mean of 43.99 carbonic 

 acid; diflerence, 0.02. The value taken is the mean of all 

 experiments. The carbonate was prepared by precipitating 

 calcic chloride with ammonium carbonate, and drying at 

 160° to 180°. Confirmatory experiments were made on ice- 

 land spar. The weighings are reduced to vacuum. {Erd- 

 mami's Journ. fllr Prak. Chem., 26, 1842, 472.) 



Berzelius maintained that Erdmann and Marchand em- 

 ployed material containing water, chlorine and magnesium. 

 Erdmann and Marchand answered that there could be no 

 magnesium and was no chlorine but that they had convinced 

 themselves that spar is the only compound of certain and 

 constant composition. Berzelius replied that the}' then 

 admitted that their carbonate contained water. Erdmann 

 and Marchand appealed to their experiments on spar, upon 

 which Berzelius made experiments showing that s[)ar, too, 

 retains water at 200°. This Erdmann and Marchand de- 

 nied and finally assert that all the carbonic acid is not driven 

 off at any attainable temperature, and that their results 

 were therefore too his^h instead of being too low. The 

 error they estimate to exactly cover the difterence between 

 their averages and 40. (Erdmann' s Journ. fiir Prak. Chem., 

 31, 1844, 257; 37, 1846, 75; 50, 1850, 237.) 



Erdmann and Marchand: 4,0.062 {0 = 1Q)', 250.30 

 (O = 100). 



The spar experiments referred to above. Six analyses were 

 made as before, giving a mean of 56.028 oxide ; extreme 



