1895.] NEW YORK ACADEMY OF SCIENCES. 173 
to each other, the amount of polarization in a beam of natural 
light which has undergone single refraction is 
1 
Bry — 371’, 13! — 71’ __ cos? (a—P) : fe ed 
US ee at Ca ie naa aa ; 1+cos? (a + 8) fae 
1. 
cos? (a +B) 7 
The only unknown quantity in this formula is the angle /. 
In order to determine 3 for any given angle it was only neces- 
sary to determine the index of refraction of the 
uranium glass. 
The glass being of considerable thickness the 
microscope method was the one best adapted 
to this determination. d being the thickness of 
the glass,and a the change in focus due to 
the introduction of the glass between the ob- 
ject o and the objective, the index wu is given 
by the formula (see Kohlrausch Praktische 
Physik, p. 151). 
d 326 mm iu sin a 
u = d—a— 326mm—110mm — Lol= sin 3 
The substitution of the various values of 3, thus found, in the 
formula for p gave the following values : 
a B Pp 
oh SEE 1 aa eee ae S34 Se ot 351 
3 el ne ie A Garett (7a UA 315 
BUS wen is saree EAA A Se ee. tees Sf 251 
lee eet ha: ahs Bae AGRE ee oe ws an 206 
i uhee ga CERT: gl |) Aner cee eee 153 
SON agg oe Pea Ege Sh Feb een Oa eh coe Sn 125 
HU Ware ieee eels hee hs ek ae 058 
The correspondence between the results given by experiment 
and those given by this calculation from Fresnel’s formulz was 
unexpected. The experiments were completed more than a 
month before any calculations were made, so that I had no idea 
at the time of making the experiments what would be the na- 
ture of the results given by calculation. 
The differences between the two sets of values are hardly 
greater than the possible errors of observation. The differences 
at 65° and 50° are quite large, but might have been due to the 
lack of perfect uniformity in the luminous surfaces. On the 
whole, the agreement between the two sets of results indicates 
