September 30, 1910] 



SCIENCE 



447 



Table XIII. contains the values of e obtained 

 from all of the observations recorded in table 

 XII. except the first four and the last six. 

 These are omitted not because their introduc- 

 tion would change the final value of e, for as 

 a matter of fact this is not appreciably altered 

 by including them, but solely because of the 

 experimental errors involved in work upon 

 either exceedingly slow or exceedingly fast 

 drops. When the velocities are exceedingly 

 slow residual convexion currents introduce 

 errors, and when they are exceedingly fast 

 the time determination becomes uncertain. 



The final mean value of e is 4.9016 X 10"^". 

 The probable error computed from the number 

 of observations shown in the last column and 

 their average divergence should be about one 

 tenth of one per cent. Since, however, the co- 

 efficient of viscosity of air is involved in the 

 formula the accuracy with which e is known 

 is limited by that which has been obtained in 



TABLE XII 



the measurement of this constant. Aiter a 

 prolonged and very careful study of all the 

 data available on the viscosity of air I have 

 chosen as the most probable value of fx at 

 15°.00017S5. For reasons which will be de- 

 tailed elsewhere it is thought that the error 

 in this value is less than one haK of one per 

 cent. 



It is most interesting that the agreement 

 between Cunningham's rational formula and 

 our experimental results is so perfect. How 

 perfect it is may be seen graphically from 

 Fig. 2, in which the curve is computed from 

 7 under the assumption of e^ 4.9016 and our 

 experimentally determined values of e are 

 plotted about this curve, every observation 

 contained in Table XII. being shown in the 

 figTire. Nevertheless, it is to be particularly 

 emphasized that the correctness of our final 

 value of the elementary electrical charge is 

 completely independent of the correctness of 

 any theory whatever as to the cause 

 of the failure of Stokes's law for small 

 drops. It is entirely possible that a series of 

 experiments of this kind upon substances 

 other than oil may lead to other values of A, 



tabm: xni 



Final mean e = 1.9016 



