295 
by taking averages, as e.g. Desiye did in his paper on the VAN DER 
Waats’ cohesion forces *). 
In our first paper (see p. 1188) we assumed for the attractive 
force = f X Aw, in which «a represents the distance of the moving 
molecule from the neutral point; and for the repulsive force on 
collision 2e > y, in which y represents the compression of the 
molecule. 
Later on (p. 897) I substituted for the two separate laws of force 
one single law of the form (cf. the cited paper for the meaning of 
the different letters) 
Py oe |: rd 
(ls)? — 2? 
which rendered the solution of the problem raised tlrere still just 
possible by the aid of elliptical functions. 
Though these two laws of attraction by no means represent 
reality quantitatively accurately, yet at low temperatures we found 
a relation between ME and 7, which is analogous with PLANck’s 
well-known expression — which certainly proves that the essential 
part of our considerations (viz. our observance of the t/me-average) rests 
on solid foundations. I'he exact form of the law of attraction seems 
here to a certain extent to be of minor importance, and according 
to the results of the two papers to have influence only on some 
numerical coefficients. 
1) Phys. Zeitschr. 21, 178 (1920). In 1908 VAN DER Waars Jr. already treated 
a simular problem, but he still considered the atoms (molecules) as electric 
double points which, like Depije, he besides supposed far enough apart to 
simplify the problem. He found that the force decreased more rapidly 
than 77. 
Later also KEESOM wrote a paper in connection with the said paper by 
Dezije (These Proc. Vol. XXIII, N°. 6, p. 939 and 943; also Phys. Zeitschr. 
22, 129 (1921) and Mededeelingen Utrecht N°. 6) on the question of the 
forces of attraction. There he demonstrates that for H,, O, and N, DEBIJE’s 
quadrupoles yield a too large value; further that — atleast for the said gases 
— the quadrupole-attraction has considerably more influence on the second 
virial-coefficient B than the so-called “induced” attraction (unless the tempe- 
rature is very high), and that the VAN DER WAALS cohesion-forces can chiefly 
be attributed to forces which the molecules exert on each other in virtue of 
their quadrupole-momenta. 
BurGers (Dissertatie, Leiden 1918, p. 186) calculated the quadrupole- 
momentum of the H,-molecule according to the model constructed by BoHr 
and DeBĲE (to which, however, there are several objections), and found a 
remarkable agreement with the value 2,03.10-% (electrostatic units  cm”.; 
uncorrected for the polarisation of the molecules in each others’ electric 
field), derived by KrEEsSom (Comm. Leiden, Suppl. 39a, p. 15) from the 24 
virial-coefficient, viz. 2,05.10-—26, 
