212 Professors Redtenbacher and Liebig 



ought we not first to compare these with other experiments, 

 in which this source of error is entirely avoided ? 



A question also arises, whether naphthalin, a substance the 

 atomic weight of which cannot be determined with certainty, 

 as it enters undecomposed into no combination, is a proper 

 substance to select as the means of determining the atomic 

 weight of carbon or of hydrogen? That body must indeed 

 be rejected on this account, for it is not in our power to con- 

 trol our analytical results from a knowledge of the weight of 

 its atom, that is, the sum of the atomic weights of the elements 

 composing it. 



When we also consider, that the naphthalin, which in the 

 above experiments was submitted to combustion in a glass 

 tube with oxide of copper, is a volatile body, that it cannot 

 be introduced into the combustion tube, with oxide of copper 

 that is absolutely free from moisture; and bear in mind also, 

 that, owing to the volatility of the substance, this moisture 

 cannot previously to combustion be removed by means of 

 exhaustion, we cannot doubt the existence of a source of error, 

 which must increase the per-centage of hydrogen beyond that 

 which actually existed in the substance ; for, however small 

 the quantity of this hygroscopic moisture may be, it is never- 

 theless always present; it is weighed with the chloride of 

 calcium tube, and its hydrogen added to that contained in 

 the substance. 



In all analyses hitherto conducted, even those in which 

 the whole of the hygroscopic water had been removed as 

 nearly as possible before combustion, by means of exhaustion, 

 it is observed that the experiment invariably gives rather 

 more hydrogen than is indicated by calculation. This excess 

 amounts in good analyses to from 0*1 to 0'2 per cent. It is 

 found, however, that this, in reference to the quantity of sub- 

 stance employed in analysis, is not sufficient to affect the pro- 

 portion of the elements, to the extent observed in the analyses 

 of naphthalin ; the excess in those analyses is however dimi- 

 nished, when allowance is made in the calculation for this 

 error. 



There exists therefore some other cause affecting the deter- 

 mination of the equal quantity of the elementary constituents 

 of an organic substance, in such a manner, that one of them, 

 namely, the carbon, when calculated from the quantity of car- 

 bonic acid obtained by combustion, amounts to more than the 

 weight of the carbon which is contained in the matter analysed. 

 On this account a new determination of the atomic weight of 

 carbon appears to be indispensable, and we have united, in 



