104 M. E. Pringsheim on the Radiometer, 



elastic glass tube. The cocks and stopper were all well 

 greased, which enabled me, with a little attention, to keep the 

 apparatus completely exhausted for days together without any 

 perceptible increase of the air-pressure. 



I. Action of the Glass Case. 



On the direct action of the glass case of the radiometer no 

 prior experiments are known to me. It is true that experi- 

 ments were made by Crookes * for the purpose of ascertaining 

 what influence is exerted by a greater or less distance of the 

 vanes from the glass sides, but only in order to show that the 

 motion depends on a reciprocal action between the vanes and 

 the case. 



First, the question presents itself, Is the glass case, in itself, 

 an indispensably necessary condition for the generation of the 

 motion of the radiometer ? — that is, would the motion cease, or 

 be different in quality, if the movable part of the apparatus 

 were suspended in an unlimited space of rarefied air ? Crookes 

 once affirms this f , emphatically stating that in the radiometer, 

 in contrariety to the otheoscope, the glass case is necessary 

 for the motion, but without mentioning his reasons either in 

 the place cited or, to my knowledge, anywhere else. He was 

 perhaps led to this conclusion by theoretical considerations, 

 which he indicates in another place \ y where he adopts Stoney's 

 theory. 



To decide this question directly by a simple experiment is, 

 of course, impossible, since we cannot produce even a rough 

 approximation to an infinite space of air so highly rarefied. 

 Hence the decision of this point is only possible by approach- 

 ing it indirectly, and probably not without theoretical dis- 

 cussions ; therefore we have not yet come to the place for 

 entering more particularly into it. 



A supposition, however, which certainly at first obtrudes 

 itself respecting the behaviour of the glass case, is very easily 

 refuted. For a very simple observation shows that in the 

 ordinary radiometers, consisting of thin mica vanes blackened 

 on one side with soot, irradiation of the blackened side 

 alone with sunlight or an ordinary artificial source of light 

 produces rotation in the same direction as irradiation of 

 the bright side only, so that in both cases the black sur- 

 face recedes. Now, on the assumption that the rotation is 

 produced by the difference of temperature which actually 



* Comptes Rendus, lxxxiii. p, 1233 (1876) ; Proc. Roy. Soc. xxv. 

 pp. 308, 309 (1876). 



t Comptes Rendus, lxxxiv. p. 1081 (1877). 



% Proc. Roy. Soc. xxv. p. 308 (1876) ; Nature, xv. p. 224 (1877). 



