ATOMIC WEIGHTS 179 



carbon dioxide and water under pressure. B}' heating in a silver dish 

 the bicarbonate was converted into the normal salt, which was weighed, 

 and then ignited at a temperature of 1300°-1400°. The data are as 



follows : 



3.9772 grm. CaCOj gave 2.2268 grm. CaO. 55.989 per cent. 



2.3614 " 1.3218 " 55.975 



3.2966 " 1.8456 " 55.985 



Mean, 55.983, ± .0028 



Hence Ca=: 39.966. 



Hinrichsen,^ in his two separate communications, gives analyses of spar 

 from two distinct localities, namely, Iceland and tlie Crimea. In each 

 case very small quantities of impurity were present, which were carefully 

 determined and corrected for. The spar, previously freed from all traces 

 of moisture, "w^as ignited in an electric furnace, at a temperature between 

 1200° and 1400°. The results obtained, with all corrections applied, 

 and vacuum weights, are subjoined : 



First Series. 



30.72157 grm. CaCOa gave 17.22354 grm. CaO. 56.0633 per cent. 



32.77791 " 18.375587 " 56.0617 



34.45625 " 19.31698 " 56.0623 



33.36885 " 18.70723 " 56.0620 



Second Series. 

 31.20702 grm. CaCOa gave 17.49526 grm. CaO. 59.0608 per cent. 

 22.00588 " 12.33642 " 56.0602 



Mean of both series as one, 56.0617, ± .0003 



Hence Ca = 40.145. 



Combining all these determinations, we have for the percentage of 

 CaO from CaCO, :' 



Dumas 56.073, ± .016 



Erdmann and Marchand, 1 56.000, ± .007 



Erdmann and Marchand, 2 56.028, ± .0047 



Erdmann and Marchand, 3 .56.006, ± .0043 



Herzfeld 55.983, ± .0028 • 



Hinrichsen 56.0617, rt: .0003 



General mean 56.0603, ±: .0003 



The effect of this combination is practically to discard all of the de- 

 terminations except that of Hinrichsen. Herzfeld's figures are cer- 

 tainly too low, and probably because of undetermined impurity in liis 



' Zeitschr. physikal. Chem., 39, 311, 1001; and 40, 747, 1902. 



