992 
null-method Compron and Roeniry') have actually established the 
absence of any such effect for the (ferro-magnetic) crystals of magnetite. 
2. In the theory of Lanerevin-Wuiss the rotational movement of 
the elementary magnets gives its own contribution to the kinetic 
energy and therefore to the specific heat, whereas in our theory 
the corresponding term does not occur. At first sight this appears 
strange, since the form of equation (2) seems to point to equiparti- 
tion. But a similar result will always be obtained, in cases, where 
the lowest quantum-motions which are possible, possess a very small 
difference of energy As with respect to each other (in our case 
2 uHleos 9 with a field H). In those cases there is always a range 
of temperatures 7, where 7’ is small enough for the higher quan- 
tum-orbits to be disregarded, and at the same time large enough 
with respect to Ae. 
3. Although the transition between the right- and left-hand 
motions requires an amount of energy small as compared to 77’, 
still it may require the coincidence of favourable circumstances to 
bring about the corresponding reversal of the motion (moment of 
momentum). Since we are dealing with a quantic process, it is 
probably difficult to treat this question quantitatively. In general we 
may expect, that the corresponding retardations in the establishment 
of the magnetisation would show themselves most easily at very low 
temperatures and rapidly alternating fields’). (For light-vibrations 
xy is always=0O). They would for instance give rise to a kind of 
hysteresis and a corresponding development of heat, when gadolinium- 
sulphate. is periodically magnetized in opposite directions. 
Leiden. Physical Department. 
1) A. H. Compton and O. Roanuey. Is the atom the ultimate magnetic particle? 
Phys. Review. 16 (1920) 464. 
2) The possibility of this retardation was pointed out by Lenz in his address 
at Nauheim (lc. p. 615) from the point of view of the sudden reversals of the 
magnetic atoms. Previously to this in the beginning of July 1920 the question was 
discussed by Prof. KAMERLINGH ONNES and me, both from the point of view of 
Weiss’ theory and of the assumptions ol this paper, together with the possibility 
of testing it experimentally. 
