MAGIC LANTERN. 157 



by inserting glass plates containing suitable figures in a 

 trough having two of its sides parallel, and made of plate 

 glass. The trough must be introduced at E F, so that 

 the figure on the glass is at the proper distance from the 

 object lens D. When the trough is filled with water or 

 with any transparent fluid, the picture at P Q will be 

 seen with the same distinctness as if the figure had been 

 introduced by itself into the groove E F ; but if any 

 transparent fluid of a different density from water is 

 mixed with it, so as to combine with it quickly or slowly, 

 the appearance of the figure displayed at P Q will under- 

 go singular changes. If spirits of wine, or any ardent 

 spirit, are mixed with the water, so as to produce through- 

 out its mass partial variations of density, th figure at 

 P Q will be as it were broken down into a thousand 

 parts, and will recover its continuity and distinctness 

 when the two fluids have combined. If a fluid of less 

 density than water is laid gently upon the water, so as to 

 mix with it gradually, and produce a regular diminution 

 of density downwards ; or if saline substances soluble in 

 water are laid at the bottom of the trough, the density 

 will diminish upwards, and the figure will undergo the 

 most curious elongations and contractions. Analogous 

 effects may be produced by the application of heat to the 

 surface or sides of the trough, so that we may effect at 

 the same time both an increase and a diminution in the 

 density of the water, in consequence of which the magni- 

 fied images will undergo the most remarkable transforma- 

 tions. It is not necessary to place the glass plate which 

 contains the figure within the trough. It may be placed 

 in front of it, and by thus creating as it were an atmo- 

 sphere with local variations of density we may exhibit the 

 phenomena of the mirage and of looming, in which the in- 

 verted images of ships and other objects are seen in the 

 air, as described in another letter. 



