1026 Proceedings of Boy al Society of Edinburgh. [sess. 
of that system most desirable. Ewing, in his treatment, assumed 
that the molecular magnets were arranged in consecutive and 
adjacent groups of four, placed at the corners of a square, and 
that the distance between nearest centres only slightly exceeded 
the dimensions of the molecule. Thus, in the magnetised condi- 
tion of the medium with no external field, a pole of one molecule 
was regarded as being practically under the control of the opposite 
pole of the consecutive molecule alone ; and a line of such consecu- 
tive molecules realised one of Faraday’s lines of induction. 
Ewing’s aim was mainly to account for the characteristic form 
of the induction curve in strongly magnetic materials and for 
the observed magnitude of residual magnetisation. The primary 
object in the present investigation is to determine whether or not 
a satisfactory explanation can be given of the characteristic 
phenomena manifested in the magnetisation of crystals of the 
cubic system. A secondary purpose is to extend the Boscovichian 
theory of the constitution of matter so as to include the subject 
of magnetism. 
2. The assumptions made are to some extent different from 
those made by Ewing. As in his work, it is assumed that the 
molecule possesses the qualities of an ideal magnet. Should it be 
the case that the effective magnetic moment of the molecule is not 
actually independent of its distance or orientation, full account of 
the fact can be taken, in the expressions subsequently found 
(§§ 9, 11), by the alteration of constants in a manner which would 
be perfectly definite when the actual conditions were given. It 
is assumed further that the space distribution of the centres of 
molecules is that of the closest-packed homogeneous assemblage 
of Boscovichian points. Other arrangements are not excluded, 
but their treatment is not taken up in the present paper. It is 
conceivably quite possible that a comparison of results obtained 
with different arrangements may lead to magnetic tests discrimi- 
nating the actual arrangement of the magnetic constituents of 
molecules from amongst those which are theoretically possible on 
the cubic system. It is assumed, at first, that the medium is of 
practically infinite extent, so that demagnetising force is not 
evident; but the effect at boundaries is specially discussed (§ 19). 
