158 TRANSACTIONS OF THE [APR. 1, 
to a certain depth; and the rays emitted by the interior mole- 
cules undergo refraction at the surface.” Since the time of 
Verdet, I believe no one has made any experiments upon the 
subject except Violle, who has a brief note in the Compte 
Rendu of 1887, Vol. 105, p. 111, in which he states that, while 
making some other experiments upon molten silver, he took oc- 
casion to measure the percentages of polarization in the light 
emitted by that substance at various incidences. He plotted 
the curve of these percentages and found that it was very well 
represented by the empyrical formula pe = (1 — cos i) (1 + 
" cos 75° + 4), where pe represents the ratio of polarized light to 
the whole light in the emitted beam and i the angle of incidence. 
Assuming, then, the phenomenon to be due to refraction, he 
argues that the equality of the amounts of polarization in the 
reflected and refracted beams would require that e pe =r pr, 
where e is the proportion of the whole light emitted, r the pro- 
portion reflected, pe the proportion of polarized light in the 
emitted beam, and p; the proportion of polarized light in the re- 
flected beam. Then, since the whole light is either emitted or 
reflected, e + r = 1,and the formula r = per immediately fol- 
lows. Taking the experimental values which have been deter- 
mined for p; by reflection at ordinary temperatures, he finds 
that his own results for pe, when substituted in this formula, give 
a uniformly high reflecting power for molten silver; a result 
which agrees with the known properties of ordinary polished 
silver. This forms the nearest approach to a verification of 
Arago’s assumption which has yet been given. 
Such is the extent of the work which has thus far been done 
upon polarization by emission. 
III. 
Discussion oF ARAGO’S EXPLANATION. 
The explanation of Arago and Verdet is as yet the only one 
which has been offered to account for the phenomenon. This ex- 
planation does not rest upon careful experimental proof, and, 
furthermore, there seems to be considerable reason for doubting 
its correctness. According to that explanation the light which 
comes to the eye from the surface particles is natural light ; 
but mixed with this unpolarized light is a quantity of light 
which has worked its way up from uncertain depths, has under- 
gone reflection and refraction at the surface, and is consequently 
polarized upon emergence. Aside from the intrinsic difficulty of 
this conception, the first experiments which were made in this 
