MAGNESIUM, 135 



Hence the subjoined values for glucina: 



From (i) GIO = 25. 130, ± .0269 



From (2) " =25.151,^.1783 



From (3) " = 24.891, rb .OO25 



General mean GIO = 24.893, zfc .0025 



And Gl = 9.014, zfc .0025 



If = 16, Gl = 9.083. 



All the values but that derived from the third ratio might obviously 

 be rejected. Their influence upon the final mean is altogether trivial. 



MAGNESIUM. 



There is perhaps no common metal of which the atomic weight has 

 been subjected to closer scrutiny than that of magnesium. The value 

 is low, and its determination should, therefore, be relatively free from 

 many of the ordinary sources of error ; it is extensively applied in chem- 

 ical analysis, and ought consequently to be accuratelj^ ascertained. 

 Strange discrepancies, however, exist between the results obtained by 

 different investigators ; so that the generally accepted figure cannot be 

 regarded as absolutely free from doubt. 



The early determinations made by Berzelius, Longchamp, and Gay- 

 Lussac need not be considered here, as they have only antiquarian value. 

 The investigations which demand attention are those of Scheerer, Svan- 

 berg and Nordenfeldt, Jacquelain, Macdonnell, Bahr, Marchand and 

 Scheerer, Dumas, Marignac, Burton and Vorce, and Richards and Parker. 



Scheerer's method of investigation was exceedingly simple.^ He 

 merely estimated the sulphuric acid in anhydrous magnesium sulphate, 

 employing the usual process of precipitation as barium sulphate. 'He 

 gives no weighings, but reports the percentages of SO3 thus found. In 

 his calculations, = 100, SO3 = 500.75, and BaO = 955.29. It is easy, 

 therefore, to recalculate the figures which he gives, so as to establish 

 what his method really represents, viz., the ratio between the sulphates 

 of barium and magnesium. 



Thus revised, his four analyses show that 100 parts of MgSO^ yield 

 the following quantities of BaSO^ : 



Per cent. SO.^. 



193.575 66.573 



193.677 66.608 



193.767 66.639 



193.631 66.592 



Mean, 193.6625, ± .0274 



*Poggend. Annalen, 69,535. 1846. 



