Movement of the Glass Case of a Radiometer. 393 



being nearly in a condition of equilibrium, though an unstable 

 one, remains a short time without much change of place. Then 

 it moves rapidly to its position of equilibrium, the surface which 

 is its locus forming the sheet. Then it remains in its position of 

 equilibrium during the greater part of the discharge, approaching 

 the axis again as the discharge falls, so that its equilibrium position 

 is not so far from the axis. Thus we see two bright curves cor- 

 responding to the two positions of approximate rest united by 

 a less bright sheet, the first curve being nearly a straight line, 

 and the second nearly a helix traced on a cylinder of which the 

 former line is a generating line. 



" It was noticed that the sheet projected a little beyond the 

 helix. This may be explained by cousidering that at first the dis- 

 charge is more powerful than can be maintained, so that the curve 

 reaches a little beyond the distance that can be maintained." 



The appearance of the discharge when viewed in a revolving 

 mirror (except the projection beyond the sheet, the illumination 

 of which was too feeble to be observed) confirmed the above 

 remarks. 



" On the Movement of the Glass Case of a Radiometer." By 

 "William Crookes, F.R.S. &c. 



During the discussion which followed the reading of Prof. 

 Reynolds's and Dr. Schuster's papers at the last meeting of the 

 Royal Society I mentioned an experiment bearing on the observa- 

 tions of Dr. Schuster. I have since tried this in a modified form; 

 and as the results are very decided and appear calculated to throw 

 light on many disputed points in the theory of these obscure 

 actions, I venture to bring a description of the experiment, and to 

 show the apparatus at work, before the Society. 



I made use of a radiometer described in a paper communicated 

 to the Society in January last. I quote the description from 

 paragraph 184. "A large radiometer in a 4-inch bulb was made 

 with ten arms, eight of them being of brass and the other two 

 being a long watch-spring magnet. The disks were of pith, 

 blackened on one side. The power of the earth on the magnet is 

 too great to allow the arms to be set in rotation unless a candle is 

 brought near ; but once started it will continue to revolve with 

 the Jight some distance off." 



This radiometer was floated in a vessel of water ; and four 

 candles were placed round it so as to set the arms in rotation. 

 A mark was put on the glass envelope so as to enable a slight 

 movement of rotation to be seen. The envelope turned very 

 slowly a few degrees in one direction, then stopped and turned 

 a few degrees the opposite way ; finally it took up a uniform 

 but excessively slow movement in the direction of the arms, but 

 so slow that more than an hour would be occupied in one revolu- 

 tion. 



A powerful magnet was now brought near the moving arms. 

 They immediately stopped, and at the same time the glass envelope 



