THE ELECTRIC AND LUMINIFEROUS 2*[ED1UM. 
795 
its steady equilibrium state with excessively great effective elasticity, while the 
tractions necessary to equilibrate a free boundary are non-existent. Such a hypo¬ 
thesis looks like explaining one mther by means of a new one, but it is perhaps not 
really more complicated than the facts ; on our present principle of iuterpretation, the 
change of gravitation in the field due to a disturbance at any point must have been 
propagated somehow, while in the machinery that ti’ansmits electric and luminiferous 
disturbances no elasticity has yet been recognized anywhere near intense enough to 
take part in such a propagation. 
We may not surmount the difficulty by the assumption that, in addition to the 
finite resistance to rotation which is the cause of the propagation of the radiation, the 
medium also possesses an enormously greater static resistance to rotations of some 
more fine-grained structure, and that the surface integral of the rotation over any 
surface enclosing a vortex-atom is a positive constant, of course definite and unchange¬ 
able in value for each atom ; for this would lead to gravitational repulsion instead of 
attraction. The term must be in the kinetic energy, not in the potential energy of 
the medium. 
104. In a representation of a magnetic or other medium,'^ imagined to be composed 
of gyrostatic elements spinning indifferently in all directions, and linked into a 
system by an arrangement like idle-wheels between them, in fact by an ideal system 
of universal ball-bearings, the kinetic energy lunction would have a rotatory part 
T = 1 C f 
‘Ml , ‘V , ‘Ml\ ,lr 
dt- ^ ^ dt^' ’ 
where {f, g, k) is the absolute rotation of an element, which is supposed from the 
connecting mechanism to be a continuous function of position in the system. 
We would have therefore 
“ ^ ) 1 dt \ly dz ) dt dt \ dz ~ dx) It dt \dx ~ dy)\ ^ 
n^/dh 
J 
Thus the kinetic forclve which is the ecjuivalent of the acf-ual applied forcive in the 
medium per unit volume, arising from its potential energy and such extraneous forces 
as act on it, is 
d^ 
C curl (/, g, h), or - C — (^, p, Q 
dl^ 
If we suppose the displacement (^, p, to be originally derived from a potential 
* Cf. Maxwell’s “Hypothesis of Molecular Vortices,” ‘Treatise,’ §§ 822-7, 
