178 TRANSACTIONS OF THE [aPr. 1, 
upon which experiments had been made. These calculations 
were made as follows : 
Quincke’s values for the D line are 
ATTCE, wormule (zie jogs = ae 
Then for a = 80° formule (3) (2) and (1) give 
K? = .9348 K? == 4013 
Assuming now the incident beam to have had an intensity 
unity, the emitted beam polarized in the plane of emergence 
would have an intensity 
1 — K” = .0652 
And the beam polarized in the plane perpendicular to the plane 
of emergence an intensity 
1 — K”? = .5987 
.5987—.0652 
Therefore the degree of polarization = p = 3.7 gs5 —-834. 
The complete results of the calculations for the platinum 
are as follows : 
CO pe RE ee ee ARS Oe eel Kee ee Kee eee 
SOR ph e9Ot8 veer cao & een Oboe ie )sn OO Ri) Seeman Bee 
DOs oe Se Osi cs ehUSE ee ae OAs cn ee ee OO ke amen Coles 
GOL Me 82848 tO gee olsOGy. : OLBin~ es - eAge 
DOS Pent STB 2 Fh BAS SU e YRS stare ety eee oan 
AQe RS oi 40982 5 SSO98IL Se seo Le 401 See Selo 
BO 4 eae LID Do 68307 2 eee eGo) sae OOO een eel eeg 
Considering the number of assumptions which have been 
made, the correspondence between these quantities and those 
given by experiment is altogether remarkable, and points with 
as much certainty as the work upon uranium glass to the con- 
clusion that the phenomenon is simply one of refraction. 
XIII. 
DIFFERENCE IN CoLoR OF IMAGES. 
In the course of these observations upon platinum, another at 
first unaccountable phenomenon was noticed. At large angles 
of emergence the color in the two images was notably different. 
The feeble image, 7. e., the one corresponding to vibration per- 
pendicular to the plane of emergence seemed for large angles to 
