116 EF. C. Pickering—Measurements of the 
tween the telescopes is then altered so as to make 7 successivel 
10°, 20°, 30°, &c., and the observation repeated. Various ad- 
justments must be made to eliminate constant errors, but they 
need not be detailed here. One series was made with a glass 
prism having an index of 1°52, a second with a plate of colored 
glass, a third with a sheet of plate glass and others with four and 
ten microscope slides. The latter were used, as the thickness of 
the plate-glass was such that, when a number of plates were 
placed between the telescopes, a portion of the internal reflec- 
tion would be lost. 
To measure the polarization by refraction, two similar meth- 
ods were employed. The plates were placed vertically over the 
center of a graduated circle, and a piece of ground-glass was 
viewed through them by the polarimeter. The plates were then 
set at various angles, and the polarization measured in each case. 
All these observations were made in cloudy weather, to elimin- 
ate the effect of sky polarization. Several series were made 
with the optical circle and Babinet’s wedges, placing the tele- 
scopes opposite each other, and recording the angles of the 
Nicol for various positions of the plates. Still a third method 
was employed, already described in connection with the Arago’s 
polarimeter. 
It will be noticed that no observations are given of the polar- 
ization of a beam transmitted by one surface of glass. ‘There 
seemed to be no easy method of measuring this quantity. It 
might be done by making aseriesof prisms of such angles that 
when the light was incident on one face at 10°, 20°, 30°, &e., 
the refracted ray would strike normally on the second face. 
The effect of the latter would then be nothing, so that the po- 
larization would in this case be entirely due to the first surface. 
o determine the polarization of a single surface of glass & 
plate was covered on one side with lamp-black, but it was found 
to give the same results as when laid on velvet. A plate of col- 
ored glass was therefore used instead. Fig. 6 gives the results 
of the observations of the polarization effected by one surface 
of glass, fig. 7, that of two surfaces, fig. 8, of four, fig. 9, the 
reflected beam for twenty surfaces, and fig. 10 the corresponding 
refracted beam. The smooth curves give the theoretical polar- 
ization, including the internal reflection, and the dotted lines 
omitting it. The observations of the reflected beam agree very 
well with theory, especially when the number of surfaces 18 
sm e refracted beams, on the other hand, show a devia- 
tion from theory which becomes perceptible for one plate above 
80°, for four plates above 65°, and for ten plates above 20°. 
The errors most likely to occur, which tan 3 be Common to all 
the observations on the refracted beams, are, first, stray Fats 
t 
or light entering the instrument without passing through ¢ 
