The Value of e/m

Abstract

IT is quite customary, at the present time, to use as the most probable value of e/m that derived by Paschen from spectroscopic data, as given by equations (12) and (13), page 272, and (15) and (16), page 275, of Sommerfeld's “Atombau,” third edition. Taking Paschen's own estimate of the error in R_He and R_H, we have a probable error in e/m of about 0.2 per cent. But I have shown by a more detailed consideration of all available data (Physical Review, 17, 589, 1921) that Paschen's estimate of error for RH (±0.06) is certainly too small, and that the true probable error is nearer ±0.2. The latter figure leads to an error of 0.5 per cent in e/m, and the Paschen data, combined with the best data on atomic weights (4.002 and 1.0077 for He and H) result in e/m =1.768±0.009, where the error in RH has alone been considered.