Vol. 7, 1921 
PHYSICS: L. B. LOEB 
9 
The dotted curves are the observed and the full curves are the computed 
ones. The experimental data are given in the legend. The shapes of 
the curves at atmospheric pressures or at low values of N, where the carriers 
are all ions, should not agree. For in deducing the theory we assumed 
that at V 0 all carriers succeeded in reaching E, which is not the case in 
fact. Also at the lower pressures close agreement is not to be expected, 
for the velocity acquired in the field becomes commensurable with the 
velocity of agitation W. Barring these points the general shapes of the 
theoretical and observed curves, and the changes in shape of these curves 
with p and N, are quite similar. It is also evident that the points of in- 
l^iihfel Mt lt k \ ■■■■\, -.\ : \ hi. ,- 
§8 WlOMBUOH (MOTIt A»D CniiPOTED HPBH.ITT 
1 - — r* 
m 
3 
% 
r 
tej 
'■■ 
0 10 20 30 40 50 60 70 80 90 100 
:'V c 
[■ 
o: pari 
m 
son or 
» 1.6 
1 
0BEKR 
9 
no 
TC 
t - 
v 
.45 
m 
1 
CD MM 
ILI1 
■. 
t nmva 
s 
.... 
■]- 
I 
- ' " ; 
6 
A' 
/zs. 
.. h 
i 
~b 
- 
— a 
_.„.. 
: | 
::■ 
: 
0 10 20 30 40 50 60 70 80 90 100 
flection, or the asymptotic feet, of the corresponding observed and com- 
puted curves fall on values of V which lie close together. This means 
that the values of the hybrid, or abnormal, mobilities estimated from the 
two sets of curves are nearly the same. There is, therefore, sufficient 
similarity in the two sets of curves to permit one to assert that the curves 
are represented by the same type of equation. We may then conclude 
that the Thomson theory is in good qualitative agreement with the results 
observed. 
With the evidence before us I believe we are justified in concluding that 
the mechanism of negative ion formation in air consists in the electron 
attaching itself to a molecule to form a negative ion on the average in one 
out of 250,000** molecular impacts. Now it has been shown that in pure 
nitrogen 3,4 the electron does not attach to form an ion to any appreciable 
extent. We must, therefore, assume that the electron attaches to the 
oxygen molecule in air. Measurements made in pure oxygen give n 
as about 50,000. As there are four molecules of nitrogen in air to one of 
oxygen this agrees quite well with our conclusion. What the significance 
of n is, whether it depends on the electronic ring in the molecule struck, 
whether it depends on some particular state of the molecule struck, or 
whether it depends on the energy conditions of the impact remains for 
the future to say. 
