PHYSICS: R. A. MILLIKAN 
233 
directly the time of pushing a key, in terms of the indications of a 
standard clock, with an error which is never more than 1/100 second. 
The times of fall and rise were from 14 to 60 seconds. This makes the 
error in the mean of a series of time determinations a wholly negligible 
quantity. The electric field strengths were measured to 1 part in 3000, 
by a 750 volt Weston standard laboratory voltmeter calibrated at fre- 
quent intervals during the experiment against 3 standard Weston cells. 
The coefficient of viscosity of air was redetermined with extraordinary 
precision by Dr. E. L. Harrington who, using the constant deflection 
apparatus designed by Dr. Gilchrist and myself, succeeded in introducing 
such improvements in conditions and perfections in detail as to make 
his final value altogether unique in its reliability and precision.^ In 
view of this and other work now in progress with the same apparatus it 
is hardly possible that the correct value of t] for dry air at 23°C. can be 
more than 1 part in 2000 removed from the value 0.00018227. This 
is within less than 0.1% of my former value, viz., 0.0001824.^ All the 
other elements of the problem have been looked to with a care which 
is the outgrowth of six years of experience with measurements of this 
kind. 
That portion of the investigation which has had to do with the test- 
ing of the general validity of the method has been reported in detail 
elsewhere.^ Suffice it to say here that I find no indications whatever 
that, when properly used, it ever fails, or that it ever even remotely 
suggests the existence of a subelectron. 
The precision of the method is sufiiciently attested by the consistency 
of the results on different drops, provided no constant error inheres in 
the measurement of the dimensions of the condenser, the volts, the time, 
or the viscosity of air. The extent of this consistency is shown in the 
figure and the table which present the observations on 25 consecutive 
drops taken with all possible precautions during a period of several 
months. The data on these drops are treated precisely in the manner 
adopted in the 1913 article. A more detailed presentation of this work 
will be pubhshed elsewhere. It will be seen from the table that the final 
mean value of is 61.126 X 10"^ There is hut one drop in the table 
which yields a value of e^ differing from this by as much as one-third of 
one per cent and the probable error of the mean computed by least squares 
is one part in 4000. This value is 0.07% higher than the value 61.086, 
which I pubhshed in 1913. Both values however are computed in terms 
of 0.0001824 as the coefficient of viscosity of air. The new value 
0.00018227 is more reliable than the old and is 0.07% lower so that 
the new value of e computed solely from the new data obtained in this re- 
