142 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 54 



From ratios 47 and Ag C = 11.9320, ± .0077 



" 1. 43, and Ag 11.9957, ± .0004 



" 2 12.0000, ± .00029 



1, 4G, and Ag 12.0007, ± .0007 



" 1, 50, N, Ag, and Br 12.0138, =t .0044 



" 1, 49, N, Ag, and Br 12.0175, ± .00052 



" 48, N, and Ag 12.0197, ± .0014 



" 55, Na, and S 12.0270, ± .0044 



" 1, 44, and Ag 12.0475, ±: .0010 



" 1, 45, and Ag 12.0499, ± .0098 



General mean, C = 12.0038, ± .0002 



That this mean is higher than the atomic weight given in ratio (2) 

 does not prove it to be in error. Scott's recent determinations, the fifth 

 and sixth given above, are even higher, and the cause of the discrepancy 

 is undetermined. The general mean of all determinations agrees well with 

 the results obtained by modern physical methods, and may, therefore, 

 stand, until it is superseded by something of less uncertainty. 



As for hydrogen, new values for its atomic weight can be deduced from 

 eleven of the fundamental ratios. The computation has been roughly 

 made, and found to be withoTit significance. The combined values, so 

 obtained, are of such small weight in comparison with ratio (1) that 

 they only modify it in the sixth decimal place, a change which is not 

 worth considering. 



To sum up: The subjoined values, referred to = 16 as the standard, 

 have been computed from all the ratios, old and new, good, bad and 

 indifferent : 



That these values are final, is not to be supposed. That they are, 

 in the strict mathematical sense, the most probable values deducible from 

 the experimental data, is also questionable. But that they are highly 

 probable values, in harmony with the best modern evidence, can safely 

 be asserted. The inferior determinations, low in weight, have practically 

 vanished, one might almost say self -rejected, but not thi-own out arbi- 

 trarily. The good measurements overwhelm the doubtful ones, whose 

 influence upon the final computations is almost negligible. The nine 

 values as given above, will be used in calculating the atomic weights of 

 all the other elements. 



