298 
F 1 l—zt«e 1 l—ez 
Dan (l—z)}? hel (U —z a) —a')h br (l—z +a)? a ((l—z)*—a?)*/s jaN 
ie 1 4 Iz en 1 ne lt+z | 
(+z)?  ((+2=a)?—a’)'2” (lee)?  ((LH2)—a!)h| 
and this will be the law of action between M and the two adja- 
cent atoms P and Q. The first four terms refer to P, towards which 
M e.g. moves; the last four to Q, from which M then moves away. 
The expression (1) yields in the equation of motion perfectly 
unintegrable forms, and we shall try to find for them an approxi- 
mate expression, when z aud « are not too great with respect to /. 
At any rate it appears at once that #’ contains the factor «x, so that 
the law of action becomes a purely periodical one. 
Whenever «& becomes =O, ie, the electron (fictitiously) moves 
through the nucleus, / will also become = 0, and the total 
force change from positive (at wv positive), i.e. directed towards the 
right, into negative (at # negative), i.e. directed towards the left, 
and vice versa. In reality for «=O both the first part ot the second 
member of (1) becomes = 0, and the second part. 
When in his cited article (p. 179 righthand side) Desi states that the 
potential of a sphere with charge + e in its centre and —e on its 
circumference is on an average —0, he is, of course, right. But in the first 
place I object to this view of the problem, since it will depend on the 
mutual position of the electrons in their orbits round the two centres 
and on their phace-difference, what action will result; which also 
renders it doubtful whether all orientations of the two electrons 
on the two sphere-surfaces will, indeed, be equivalent — even on an 
“average”. And in the second place it seems to me that his method 
— in order to find still a positive valne (i.e. attractive action) for 
the resulting force — of taking the action into account which one 
atom exerts on the electric moment of the other, is open to doubt. 
For according to DerBisr himself the electric field of this one atom 
will be on an average = O (see a few lines lower). How can then 
this field, which is=-0 on an average, exert an appreciable polari- 
zing action ') on the other atom? 
That the attractive action with the periodicity found by us, is 
1) Even apart from the fact that the polarization will certainly always be 
very small, because in my opinion the exceedingly great velocity of the 
electrons in their orbits excludes an appreciable deformation. DEBIJE finds 
finally for the attractive action proportionality as r—9 (for so-called dipole 
gases on the other hand 7-7, cf. note 1 on page 183), as against VAN DER 
Waars Jr. as r—(7+*), 
