POLARISED LIGHT 385 



flat ends being supported horizontally in front of the polariser, 

 filled with suitable vapour, and the parallel polarised beam 

 sent through it. The particles in the tube then act as 

 analyser, so that the phenomena can be observed by all 

 spectators in a direction approximately at right angles to the 

 tube. In one position of the polariser, light is copiously 

 reflected to such spectators, while it is invisible to anyone 

 looking down upon the tube. When the polariser is rotated 

 90 the light is extinguished horizontally, but reflected above 

 and below. A better method, however, is to interpose between 

 the polariser and the tube a large plate of selenite or mica, 

 when the tube will glow with the usual colours. 



Such a tube should not be less than 2^ or 3 inches 

 diameter, and 18 inches long. Professor Tyndall generally 

 preferred, having exhausted the tube, to introduce sufficient 

 nitrite of butyl vapour in air to depress the mercury-gauge -$ 

 inch, and to add sufficient hydrochloric acid vapour in air to 

 depress the gauge a further J inch. He also employed vapour 

 of carbon disulphide, amyl nitrite, and other compounds. But 

 it is very much easier, and saves the trouble of exhausting 

 and a great deal of rather delicate manipulation with air- 

 tight and expensive apparatus, to employ a whiff of tobacco 

 smoke, which is perfectly effectual. Instead of using the 

 parallel beam from the polariser, very fair results can be had 

 from the Nicol analyser in the nozzle of the optical front, 

 but the polarised beam does not then, of course, fill the tube 

 BO completely. 



It is however easier and more convenient in every way, to 

 show the phenomena in liquids, which can be conveniently 

 done in a large rectangular glass trough or cell. Far the 

 best method, however, not only for economy and convenience, 

 but for its truly magnificent effect, is one for which I was 

 originally indebted to Mr. John Thomson of Dundee. For 

 this beautiful experiment we require a glass jar with foot, 

 which should be about 2^ inches diameter, and is shown at 



