INTERFERENCE OF LIGHT 331 



second method described. In this case a fresh film should 

 always be taken up after all arrangements are made, as the 

 bands are most numerous when the film is thickest. They 

 will be seen to shift as the film thins, but care must be taken 

 to avoid a current of air on the film as far as possible. 



The colours of a film of condensed vapour are readily pro- 

 jected according to Eeade's method. A plate of glass is 

 rubbed with soap, and then nearly cleaned off with a chamois 

 leather; the plate is then adjusted in the Bunsen holder. 

 Then taking a piece of rubber tube about a foot long and 

 | -inch in bore, this is gently breathed or blown through, with 

 the othor end directed against the centre of the soaped side, 

 which is projected 

 like the soap film. 

 Coloured rings will 

 appear on the plate. 

 One precaution is 

 necessary, however: 

 as the experiment 

 depends upon the 



condensation of the pio 1S2 ._ AnaIysIsotEiDg3 



breath by cold, an 



alum trough must be placed to intercept the heat of the 

 lantern, and the plate must be cold when placed in position. 



Any piece of tolerably flat iridescent glass may also be 

 projected. 



That light is destroyed by interference, even with films 

 too thick to exhibit colour, may be shown by analysing the 

 light from a slit, as in the analysis of the soap film, from the 

 film of air between two pieces of plate glass pressed but not 

 rubbed together ; or by carefully splitting a very thin film of 

 mica, and bending it round a blackened convex surface ; the 

 reflection from this will itself give a slit, or rather a line of 

 light, if placed in the parallel beam. The spectrum will be 

 crossed by a number of dark parallel lines ; more and thinner. 



