^\R. J. LARMOR ON A DYNAMICAL THEORY OF 
w -J- 
tliat it would uot go up to iufiuity nor go down to zero : but there is nothing in its 
genei’cal aspect, at any rate for g positive, as for example given by von Helmholtz, 
and verified by Peluger for anomalously dispersive solid dye stuffs, which specially 
favours any one theory of dispersion. 
30. The medium has hitherto been taken as absolutely transparent, that is, no 
degradation of energy occurs in it, the absorption bands above so called being really 
produced by total and nearly total reflexion of the radiation, which thus is not 
absorbed by the medium, but simply cannot get into it. Suppose that there is 
present a conduction current, which may be considered to include all causes which 
put 13^ out of ] 3 hase with 13 and so lead to regular absorption of the energy of the 
waves: it may be represented, as in Part II., § 11, by the formula 
z= Ic {in' djdt + o-') 33 ; 
we now have 
K' = 1 + + 
thus simply adding to the formula for the square of the index of refraction the terms 
— {kill + kp~'^<r'L)l{m'^iy^ + cr'^.), which satisfactorily represent the general features 
of metallic propagation as was shown in Part II., § 11.'" 
It is noteworthy that as the period becomes very rapid the effective index ot 
refraction always approximates to unity ; so that very short waves will not suffer 
sensible reflexion, refraction, or diffraction even while their length may include many 
molecules of the material medium. In fact when the period of the radiation is 
sufficiently high, the free periods of the polar molecules are not quick enough to 
enable them to respond,! while the comparatively free ions are prevented by their 
[* For example, it explains why the real part of is negative for metals. In Part II. the generally 
received contention (based on a narrower theory) that jj? cannot be negative for purely elastic, that is 
dielectric, media was admitted without examination : but it is obviously inconsistent with the discussion 
above. As a concrete illustration, for a stretched thread weighted with equidistant particles, the 
squai'e of the velocity of propagation of transverse waves of sufficiently short periods is negative: yet 
no inference follows as regards instability.] 
t Foi' a similar reason, the periods of luminous radiation are already too high to allow magnetic 
polai'ization to play any part in its propagation. 
The statement in the text involves the currently receivetl exjilanation of the Roiitgen radiation. 
The different view has been recently advanced by Sir George Stokes that it may consist of sudden 
shocks transmitted through the fether from impacts by the molecules of the cathode streams. The mole¬ 
cules of matter lying in the track of the rays would not have time to be sensibly polarized by a sudden 
pulse which is over in a small fraction of their natural periods, and thus the pulse would pass across in 
the spaces between them, like sound through a grove of trees, without sensible refraction or diffraction . 
on the other hand the disruptive effect would resemble that of an explosive wave. Such pulses could 
hardly be other than the irregular beginnings of regular v/ave-trains sent out by the individual \ibia- 
tiim molecules; and as all radiation consists of such intermittent trains each with its irregular begin- 
ning, it would be assumed that the initial pulse is very much more intense in the electric bulb than in 
ordinary light, though still perhaps representing but a small portion of the total energy of the radia¬ 
tion. That the bombardment by the cathode streams is of a very disruptive, so to speak explo6i^e, 
