118 M. E . Pringsheim on the Radiometer. 



If the bright side was irradiated by the flame of a Bunsen 

 burner, a positive deflection again occurred, which varied 

 between 100 and 200 scale-divisions, according to the dis- 

 tance and size of the flame. 



On irradiating the smoked side, the Dutch metal behaved 

 (as might have been foreseen) quite similarly to the mica ; 

 for both underwent a strong positive deflection. 



The motion of the radiometer- vane can be produced not 

 only by a difference in the substance and temperature of its two 

 faces, but also by a difference in the shape, and, lastly, by a dif- 

 ference in the position with respect to the glass case. Many 

 experiments have been made with curved vanes, especially 

 by Crookes* and Zollnerf, which, however, do not completely 

 agree with each other. Yet from all of these experiments it 

 follows that the convex face corresponds to the blackened side in 

 ordinary radiometers, the concave face to the bright side. Very 

 peculiar were the results found by Crookes with aluminium 

 vanes lampblacked, some on one and others on both sides; for 

 whenever in those experiments any motion was generated, on 

 employing a luminous source of heat the concave side went 

 before, even when both sides were lampblacked but only the 

 concave surface was irradiated. The perfect equality of tem- 

 perature of both sides, assumed by Crookes in order to account 

 for this phenomenon, appears to me, after the above-commu- 

 nicated experiments on the influence of the heat-conduction, 

 very unlikely. Crookes, moreover, on Stokes's proposal, has 

 made an experiment with a radiometer having peculiarly curved 

 vanes, which shows it to be very probable that the rotation is 

 not a result of the difference of form of the two sides by 

 itself, but much rather arises from their different position to 

 the glass case. The influence of the position of the vane with 

 respect to the glass appears considerable in radiometers with 

 transparent vanes placed obliquely ; and in these it is found 

 that the side which is nearest to the glass is repelled the most 

 strongly. To this cause is certainly to be referred also the 

 phenomenon remarked by Crookes t, that a radiometer-cross 

 rotates more rapidly in a small than in a large vessel. 



The Laws of Kadiometek-motion. 



The results of the experiments made by the various obser- 

 vers with radiometric apparatus cannot at once be comprised 

 in general propositions from which it would be possible to 



* Proc. Roy. Soc. xxv. pp. 312-314 ; Comptes Mendus, lxxxiii. pp. 1289- 

 1291 : Nature, xix. p. 88 ; Phil. Trans. 1878, pp. 294-302. 

 t Pogg. Ann. clx. pp. 160, 164, & 165. 

 \ Comptes Rendus, lxxxiii. p. 1233 ; Proc. Eoy. Soc. xxv. pp. 308, 309. 



