1889.] on the Discharge of a Leyden Jar. 423 



but with a slight rattle in addition, indicating intermittence as well 

 as alternation. Rotate the mirror, and everyone should see one or 

 other of the serrated bands of light at nearly every break of the 

 primary current of the coil. [Rotating mirror to analyse sparks.] 



The musical sparks which I have now shown you were obtained 

 by me during a sj)ecial digression * which I made while examining 

 the effect of discharging a Leyden jar round heavy glass or bisulphide 

 of carbon. The rotation of the plane of polarisation of light by a 

 steady current, or by a magnetic field of any kind properly disposed 

 with respect to the rays of light, is a very familiar one in this place. 

 Perhaps it is known also that it can be done by a Leyden jar current. 

 But 1 do not think it is ; and the fact seems to me very interesting. 

 It is not exactly new — in fact, as things go now it may be almost 

 called old, for it was investigated six or seven years ago by two most 

 highly skilled French experimenters, Messrs. Bichat and Blondlot. 



But it is exceedingly interesting as showing how short a time, 

 how absolutely no time, is needed by heavy glass to throw itself into 

 the suitable rotatory condition. Some observers have thought they 

 had proved that heavy glass requires time to develop the etitect, by 

 spinning it between the poles of a magnet and seeing the effect 

 decrease ; but their conclusions cannot be right, for the polarised 

 light follows every oscillation in a discharge, the plane of polarisa- 

 tion being waved to and fro as often as 70,000 times a second in my 

 own observation. 



Very few persons in the world have seen the effect. In fact, I 

 doubt if anyone had seen it a month ago except Messrs. Bichat and 

 Blondlot. But I hope to make it visible to most persons here, though 

 I hardly hope to make it visible to all. 



Returning to the Wimshurst machine as exciter, I pass a dis- 

 charge round the spiral of wire inclosing this long tube of CS2, and 

 the analysing Nicol being turned to darkness, there may be seen a 

 faint — by those close to not so faint, but a very momentary — restora- 

 tion of light on the screen at every spark. (CS2 tube experiment on 

 screen.) 



Now I say that this light restoration is also oscillatory. One way 

 of proving this fact is to insert a biquartz between the Nicols. With 

 a steady current it constitutes a sensitive detector of rotation, its 

 sensitive tint turning green on one side and red on the other. But 

 with this oscillatory current a biquartz does absolutely nothing. 

 (Biquartz.) 



That is one proof. Another is that rotating the analyser either 

 way weakens the extra brightening of the field, and weakens it 

 equally either way. 



But the most convincing proof is to reflect the light coming 



* Most likely it was a conversation wMch I had with Sir Wm. Thomson, at 

 Christmas, which caused me to see the interest of getting slow oscillations. My 

 attention has mainly been directed to getting them quick. 



