on the Atomic Weight of Carbon. 



219 



'* These results, which range about between 1293 and 1296, 

 appear to prove that the atomic weight of lead lies between 

 those two numbers. The mean of these experiments differs 

 so little from the number arrived at in my former ones, 

 namely, 1294'489, that I consider it unnecessary to alter the 

 latter. 



" If the atomic weight of hydrogen is 12*5, then the atomic 

 weight of lead, supposing it to be a multiple of that number, 

 is exactly either 1287*5, or 1300; and if either of these num- 

 bers be the true one, it appears to me that my results must 

 have oscillated about one of them; instead of which, they, as 

 we see, oscillate about a number which lies precisely midway 

 between the above-named." 



It may therefore be concluded, that the fact of an atomic 

 number being a multiple of the equivalent of hydrogen is no 

 proof of its exactness. 



There are other means of controlling and estimating anew 

 the atomic weight of carbon. The direct method, viz. that 

 of burning a known quantity of pure carbon, and ascertaining 

 the quantity of carbonic acid formed, is but little fitted for 

 the solution of the point in question, since a complex appara- 

 tus is required for collecting the gas, a circumstance which 

 lessens the dependence to be placed in such determinations. 

 If, indeed, it be remembered, that even in operating on several 

 grammes of carbon, the variations generally amount to from 

 8 to 10 millegrammes, it will be evident that a complex appa- 

 ratus offers no security for absolute accuracy. We have 

 therefore selected a different method ; an ordinary one, it is 

 true, but which has been hitherto generally acknowledged to 

 be the most certain and the most free from error in analytical 

 chemistry. 



We can, for instance, estimate with great certainty the 

 atomic weights of many organic compounds, namely, nu- 

 merous organic acids, by determining the proportions in which 



