on Artificial Mirages and Tornadoes. 351 



The three layers may now be cautiously stirred to hasten 

 the diffusion, after which they should be allowed to rest a few 

 minutes until the strias have disappeared. 



If now a beam of light from an arc-lamp, made parallel by 

 means of a condensing-lens, be thrown obliquely into one end 

 of the trough, it will be seen to traverse the liquid in the 

 form of a most beautiful blue wave, the curvature of which 

 varies with the angle at which the ray enters. Rays of 

 light travelling in sine curves are shown in figs. 2 and 3, 

 which were photographed directly from the trough. 



Prof. Everett showed in his paper that a parallel or slightly 

 divergent ray entering a medium of this description would, 

 converge to a linear focus, and then successively diverge 

 and converge to conjugate foci. This phenomenon is shown 

 in fig. 4, which is a photograph of the trough with a rather 

 wide beam of horizontal parallel light entering the end. 

 This experiment I have never seen described, though Exner 

 has shown that the eyes of some insects operate in a similar 

 manner, the visual organ consisting of a transparent cylin- 

 drical body, the axis of which has a high refractive index, 

 while as we approach the surface the optical density decreases 

 continuously. 



The beautiful miniature desert-mirages that I have witnessed 

 on the level city pavements of San Francisco (see letter and 

 photograph in ' Nature ' for Oct. 20, 1898), suggested to me 

 the idea of reproducing this phenomenon on a small scale in 

 the class-room. 



Although 1 have already described very briefly an expe- 

 riment of this nature, I will repeat the description now some- 

 what more in detail. Three or four perfectly flat metal 

 plates, each one about a metre long and 30 cms. wide, should 

 be mounted end to end on iron tripods and accurately levelled. 

 The plates should be thick enough not to buckle when heated, 

 say 0'5 cm. I have used plaster plates, made by casting 

 plaster of Paris on plate glass, with some success, though they 

 are fragile and not very durable. 



Probably plates of slate would serve admirably, since they 

 will stand a fair amount of heating, and can be obtained very 

 flat and smooth. 



The plates must be thickly sprinkled with sand to destroy 

 all traces of reflexion at grazing incidence, and the sand sur- 

 face should appear perfectly level when looked along from a 

 point just above its plane. On the absolute flatness of the 

 desert depends the successful working of the experiment; 

 therefore too much care cannot be given to the adjustment of 

 the plates. An artificial sky must be formed at one end of the 

 desert. If the experiment is being performed at night, a 



2 B2 



