294 Prof. Magnus on the Polarization of Heat. 



experiments were therefore confined to mercury, fatty oils, gly- 

 cerine, melted resins, paraffine, and such like. 



Liquids present also an especial difficulty in consequence of 

 their surfaces, when the liquids are warmed, not being so flat 

 as they are at ordinary temperatures. In all cases when we 

 examine a liquid which is of a somewhat higher temperature 

 than the surrounding air, we find the image reflected from 

 it to be no longer completely well defined. Heat, in fact, is 

 continually given off from the surface, the cooler portions sink 

 to make room for the warmer portions which rise ; and hence 

 are produced movements, small indeed, but which, in conse- 

 quence of their taking place to a greater extent at one part than 

 at another, give rise to different motions of the surface at differ- 

 ent places, and do not allow it to retain its smoothness. 



Consequently too much reliance must not be placed upon the 

 numbers. Nevertheless it was found that a portion of the heat 

 radiated from all the liquids examined was polarized. Mercury 

 showed a polarization of 32*0 and 32*2 per cent., a value at 

 least as great as that found for solid metals*. Other liquids, 

 on the contrary, gave less values. 



Rape-oil . . .{Jii PW,enfc 



Colophony 



. 7-26 



White wax 



. 73 



Glycerine . 



. 5-61 



Paraffine . . 



. 5-0 



In order to see if the diathermancy of the liquid had influ- 

 enced the results, the bottom of the tray containing the liquid 

 was covered with a bright metal plate, with a blackened metal 

 plate, and with lampblack, in succession, before pouring in the 

 liquid. Under each of these conditions the same values were 

 found for the same liquids. It hence appears that the radiating* 

 power of the bottom of the tray has no influence in these expe- 

 riments, and that consequently the liquids examined are ather- 

 manous for heat of the low temperature here employed. 



The following is a Table of the values for all the substances 

 hitherto examined : — 



* Phil. Mag. loc. cit. 



