162 TRANSACTIONS OF THE [APR. l, 
would have been caused by the passage of the beam through 
the glass itself. The tube being exhausted until the gauge 
showed a pressure of only four millimeters, the current was 
turned on and the glowing strip examined by means of the bi- 
quartz polariscope. The emitted light was still found to be 
polarized for oblique angles of emergence, and did not appear to 
have undergone any change in intensity. In order, however, to 
ascertain whether or not the effect of the air was altogether 
negligible, more delicate experiments were necessary. These 
will be hereafter described. 
It may also be added at this point that experiments were 
made upon the glowing platinum by means of a Jamin compen- 
sator, in order to make certain that the light under consideration 
was simple partially polarized light, and not elliptically polar- 
ized. From our knowledge of the way in which elliptically po- 
larized light is produced, viz., by causing retardation or acceler- 
ation in one of the components of plane polarized light, it would 
immediately be inferred that the light emitted by the platinum 
is not elliptically polarized, since there is no reason to suppose 
that it was at first plane polarized. The correctness of this con- 
clusion was immediately shown by the use of the compensator. 
For, the compensator having been set at zero by the use of plane 
polarized light, when the beam of light from the platinum was 
made to fall upon the instrument, the dark central line showed 
no displacement whatever from the zero position. Had the light 
been elliptically polarized the black central line would have been 
of necessity somewhat displaced, no matter what angles the com- 
pensator planes made with the rectangular axes of the ellipse. 
Furthermore, for elliptically polarized light there is always some 
position of the analyser which gives an absolutely black line, 
whereas, if the light is partially polarized, the line can never be - 
any darker than the field which is illuminated by that quantity 
of natural light which is contained in the partially polarized 
beam. Applying these considerations to the light emitted by 
the platinum it was shown without question that the light is 
partially and not elliptically polarized. 
2. Having thus examined the nature of the light. and proved 
that the phenomenon is inherent in the body itself, experiments 
were made upon the following substances with. results as indi- 
cated. 
SoLips.—WMetallic. 
Platinum (polished).—Polarization very strong near grazing emergence, 
but falling off rapidly as the angle diminishes. Scarcely perceptible 
at ten degrees. 
. 
