1895. } NEW YORK ACADEMY OF SCIENCES. 175 
In all of the following experiments, sheets of rolled platinum 
-06 mm. in thickness were heated to a white heat by means of a 
Bunsen burner, care being taken to prevent light from any other 
sources from vitiating the results. The observations are here 
given in full. 
80° 70° 60° 
Left Right Left Right Left Right. 
23.1 12.0 30.2 207 36.9 24.5 
21.4 eS S22 21.0 36.8 26.1 
22.5 TES ae Hil 21.2 35.0 25.0 
22.3 10.9 31.0 20.5 37.0 25.4 
23.2 11.4 aPhal Ales: 39.6 O55 . 
30.0 25.4 
22.5 11.49 31.52 20.94 36.05 25) Be 
oe —[—— wee 
2a 2 w = 52.4 2in—— Olea 
ip — -829 ‘p —-.610 p= .481 
50° 40° 30° 
Left. Right. Left. Right. Left. Right. 
40.0 29.7 45.6 35.0 47.6 38.0 
39.1 29.0 45.5 33.3 47.5 36.3 
41.2 29.3 44.4 34.2 47.2 36.2 ° 
39.4 28.5 43.8 34.6 47.8 Sued 
41.0 29.6 44.4 34.5 48.0 36.0 
40.14 29.2 44,74 34.3 47.62 36.72 
See — ae 
2 w = 69.34° 2w = 79° 2 w = 84.34 
— ond elon eos 
XT 
APPLICATION OF CaAucHy’s FoRMUL2 FOR METALLIC REFLEC- 
TION TO THE CASE OF PLATINUM. 
Fresnels formulz for reflection: rest upon the hypothesis that 
the time required in the process of reflection is infinitesimal in 
comparison with a wave period, and hence that the phase of vi- 
bration of the reflected ray is either the same as that of the in- 
cident ray,or else differs from it by the quantity +. It follows 
from this assumption that the reflected ray is plane polarized, if 
the incident ray is plane polarized. 
When the reflection takes place at the boundary surface be- 
tween air and a metal, experiment shows this assumption to be 
incorrect, and hence Fresnel’s formulz become inapplicable. 
If the phenomenon here considered be due to reflection, the 
