Polarization hy Rejlection. 555 



on AB at an angle* of 34° with the surface, is reflected in the 

 direction EF, which, by the construction, is a vertical line. 



From the surface CD, on which, at incidence, it also falls at 

 an angle of 34°, it is reflected to the pile, whose opening in- 

 clines downwards at an angle of 22°, so as to receive the rays 

 directly. From this it is clear that the whole apparatus con- 

 nected with the first plate AB may revolve round the vertical 

 line EF as an axis, until the plane of section be perpendicular 

 to the plane of the paper, and that yet the heat shall be cor- 

 rectly reflected to the pile. In this case it is clear that the 

 planes of reflection becoming perpendicular, a minimum of 

 heat will be reflected if polarization take place*. 



Such appears to be the case with all the kinds of heat that 

 J have tried. The disturbing influence of conduction is here 

 more difficultly avoided, and serves to diminish the apparent 

 effect. The quantities of heat reaching the pile from any 

 non-luminous source are always small. The results, how- 

 ever, are well marked, and seem decidedly to indicate that 

 under the particular circumstances of the observation, dark 

 heat is more completely polarized than the more reflexible 



* Square tubes of wood serve to inclose the apparatus and facilitate its 

 adjustment. Other means not represented in the figure were also used for 

 preventing direct heat from reaching the pile in any position. 



