48 THE ERRORS OF PRECISION. 



The starting material must be an element or simple 

 compound, strictly pure and accurately weighable. 



Our Standard of Matter. 



The diamond is pre-eminently such a material. It has 

 been selected by us as the standard of matter for all atomic 

 weight determinations. Comptes Rendus, T. 117, pp. 1075- 

 1078; 1893. True Atomic Weights, p. 174; 1894. 



The properties peculiar to the diamond which make it 

 almost the only substance fit to be taken as standard of 

 matter are its absolute resistance to all ordinary physical and 

 chemical agencies making it weighable with highest pre- 

 cision; and its practically absolute chemical purity, the 

 foreign matter present being incombustible and remaining 

 as a perfectly fixed, exactly weighable ash at the close of the 

 combustion. See p. 39. 



The diamond occurs native in perfectly suitable speci- 

 mens, of not an excessive cost, considering its almost ideal 

 qualities as the fundamental standard of matter for chemical 

 science. 



As the diamond will stand the action of even aqua regia 

 and quite a considerable degree of temperature without 

 change, all accessory accretions may be completely removed. 



The atomic weight of the carbon of the pure diamond 

 we take as 12 exactly. 



It is most essential to state plainly already at this point, 

 that no form of carbon can be used for this purpose of a 

 standard of matter, other than the diamond. Even graphite 

 cannot be employed, and artificial forms of so-called carbon 

 are entirely out of the question. These forms of carbon all 

 lack one or the other of the properties of the diamond. 

 Already Dumas noticed that they cannot be weighed with 

 accuracy. We shall come back to this subject under carbon. 



Pure oxygen we can obtain by proper washing and drying. 



The combustion of the pure diamond in pure oxygen 

 gives only carbon dioxide gas which is completely absorbable 

 and accurately weighable, as was first practiced by Dumas in 

 1840. True Atomic Weights, pp. 19-22; 1894. Also p. 39, 

 supra. 



