Heat radiated by Rough and by Bright Surfaces. 83 



centre was exactly in a straight line with the axis of the pile. 

 The distance of the platinum plate from the pile was 280 mil- 

 lims. Between them was placed a large screen with a diaphragm, 

 through which the rays arrived at the thermo-pile. The open- 

 ing in the screen had the same size as the radiating disk. In 

 order that none but perpendicular rays might fall upon the pile, 

 a second screen was placed at a short distance from it, with an 

 opening of the same size as the surface of the pile. With this 

 arrangement, the conical reflector with which thermo-piles are 

 usually provided, would not have caused any increase in the 

 effect. Indeed this was not necessary, in consequence of the 

 extraordinary sensitiveness of the galvanometer employed. The 

 pile itself was placed in a large cardboard box, which had no 

 opening except that of the diaphragm just mentioned. Its front 

 surface formed a square, the side of which, 13 millims. long, 

 was but little greater than the diameter of the disk. The back 

 surface of the pile remained covered with its brass cap ; and in 

 order to protect it still further from changes of temperature, it 

 was surrounded by a thick layer of loosely-packed cotton-wool. 

 In order that the temperature of the room might remain as con- 

 stant as possible, it was heated in the winter late in the evening. 

 The currents of air, which the warm stove would otherwise have 

 produced, did not then exist on the following morning. 



Instead of the galvanometer with two nearly astatic steel 

 mirrors which I have previously described*, I have employed 

 one of a different construction with two magnetic needles, and a 

 silvered- glass mirror supported below the needles by a prolonga- 

 tion of the small bar which connects them. This system of 

 needles and mirror has a very much smaller moment of inertia 

 than two steel mirrors, however thin they may be made : conse- 

 quently this galvanometer gave deflections very much greater than 

 those of one with two steel mirrors which I had previously usedf. 



In the first place, the two platinum disks were tried with 

 bright unplatinized surfaces, in order to ascertain whether under 

 these conditions they radiated equal quantities of heat to the 

 pile. When this was the case, it was certain that one occupied 

 exactly the same position as the other, and consequently that 

 they were properly attached, as well as that the flame remained 

 constant. A uniform coating of spongy platinum was then ap- 

 plied to one disk, and the radiating powers of both were afterwards 

 compared J. The sponge, as has been already stated, raises the 



* Poggendorif' s Annalen, vol. cxviii. [Phil. Mag. S. 4. vol. xxvi. p. 23.] 

 f [A foot-note of considerable length occurs at this point of the original, 



giving a description of the galvanometer used for these experiments. — 



Transl.] 



% The platinum-sponge is conveniently applied in the following man- 



G2 



