•248 
MR. J. LARMOR ON A DYNAMICAL THEORY OF 
or fji' = 771 — Ijm . m'a, where iii = + ?7i“) K' — Thus the 
emergent vibration-vector is represented by 
that is 
B' exp t [{I oLin) ^ + (m — Ijm. am') -q — pt], 
B' exp — ah . 77) exp t {il + ^y") ^ ~ • “i) V ~ P^)- 
It therefore emerges at an angle of refraction xfj, away from the edge, given by 
cot xfj = {m — Ijm . aj) j {I aj) = mjl . [ 1 — ocjl + m“^) j, 
the angle of incidence being ^ where cot = mil. Thus ^ ^ the devia¬ 
tion IS \}i — (f) — a that is a (/ sec ^ — 1), where / is the real part of (K' — siir (j)f. 
When the angle of incidence (f) is small, the deviation is thus {n — 1) a, where n is the 
real part of the complex refractive index K'K Thus the experiments of Kuxdt on 
metallic prisms, and of Pfluger on anomalously refracting media,determine the 
march of n. Although sin^ ^ is not usually very considerable compared with K', and 
thus oblique incidence on the prism does not very greatly affect the deviation,! yet it 
would seem desirable to have observations at oblique incidence, as they would give 
data for determining the imaginary part of the index also by this uniform method, 
and thus its complete value. If this were known for the neighbourhood of an absorp¬ 
tion band, we should possess all the data requisite to guide and correct theory in the 
matter of optical dispersion ; but a knowledge of n by itself is not of much service in 
this respect. The value of this method of prismatic deviation lies in the fact that 
the complex index is determined without the intervention of any considerations as to 
dynamical theory or the effect of surface contamination on polarization, which must 
enter into the interpretation of experiments on reflexion. 
Tlic Mechanical Tractions on Dielectric Interfaces: and the Mechanical Bodily 
Forcive. 
35. When the local part ot the forcive on the polarized molecules of the medium, 
arising from their interaction with the neighbouring polarised molecules, is left out 
of account, the remainder, which is the mechanical force on the element of volume, is 
derived from the energy function — {f 'P + d'Q d" hJB ); this would be also a 
potential function of the forces were it not that in it only the electric force (P, Q, B.) 
is to be varied. When however the dielectric is homogeneous, the negation of 
perpetual motions requires thaty’h/P + y dQ + liclM shall be a complete differential; 
thus Avhen the law of induced polarization is linear, the force will be derived from a 
])otential function — h{f'P + y'Q + /fB), and so will be balanced, as regards the 
interior of the medium and as regards the translatory part, by a hydrostatic pressure 
* A. PrLUGER, ‘ Wied. Aim.,’ 06 , 1895, p. 412. 
t Cf. the measures of Shea, ‘ Wied. Ann.,’ 56. 
