on the Atomic Weight of Carbon. 227 



If we' compare our experiments separately with jone another, 

 that is to say, the atomic weight of carbon as deduced from 

 the one salt, with that from the others, we observe that the 

 numbers obtained in the analyses of the same salt, agree more 

 perfectly with one another, than do the mean numbers result- 

 ing from the analyses of the different silver salts. 



If the experiments themselves be assumed to be perfectly 

 exact, this discrepancy must, from the very nature of things, 

 result from some cause. This cause can be no other than the 

 difference between the specific gravities of the different salts. 

 They were weighed, not in vacuo, but in air, and in the pro- 

 portion of their unequal densities, they must displace unequal 

 volumes of air. All these salts in the above weighings must 

 lose weight, and those which are specifically lighter more 

 than the specifically heavier salts. Dr. Clark, at the meeting 

 of the British Association held at Birmingham in 1839, called 

 attention to the influence which these relations exert on deter- 

 minations by weight, and corrected the experiments of Berze- 

 lius accordingly. It is clear, that in weighing of from one 

 to two grammes of substance, these differences will not affect 

 the relation of numbers, but when 20 or more grammes are 

 weighed, the correction ought not to be neglected. 



Such a correction must result from a knowledge of the spe- 

 cific gravities of salts. We determined those of the four 

 salts analysed by us, by taking a known weight of the salt in 

 a saturated solution, and comparing the specific gravity of 

 this liquid with that of pure water. 



3*1281 gr. acetate of silver displaced 1 gr. = 1 c. c. water 

 at 15° C. But 1 c. c. of air at that temperature weighs 

 0-00123 gr. (log. 0-0905 137 — 2); 28-800 gr. acetate of silver 

 displaces therefore 0*0113 gr. atmospheric air. These 28*803 

 gr. of salt weigh consequently 28-814 gr., but the weight was 

 estimated by means of a brass weight of 28*803 gr., which 

 having the sp. gr. 7*8, displaced 0*00455 gr. of air; that is to 

 say, it lost that weight in air. 



The above 28*814 gr. of acetate of silver, weighed therefore 

 28*8098 gr., and the 18*612 gr. of metallic silver which remains 

 behind after the calcination weighed only 18*6113 grammes. 



According to these corrections, 28*8098 gr. of the silver 

 salt, yield 18*6113 gr. of metallic silver. 



If we correct in this manner the atomic weights of the tar- 

 trate, racemate, and malate of silver, the specific gravities of 

 which are in vacuo, respectively = 3*4321, 3*7752, and 4*0016, 

 we obtain the following results : — • 



Q2 



