346 On the Forms and States of Fluids [1831. 



focal length due to the curvature at different parts of the sur- 

 face of fluid, so that by observing the luminous figure produced 

 and its transitions as the screen was moved nearer or further, 

 the general form of the surface could be deduced. Each heap 

 with a certain distance of screen gave a star 

 of light 0, fig. 24, which twinkled, i. e. ap- 

 peared and disappeared alternately, as the 

 heap rose and fell. At the corners x equi- 

 distant from these, fainter starred lights ap- 

 peared ; and by putting the screen nearer to --- 

 or further from the surface, lines of light, in 

 two or even four directions, appeared inter- 

 secting the luminous centres and apparently permanent, whilst 

 circumstances remained unchanged. These effects could be 

 magnified to almost any scale (72). 



97. When heaps of similar magnitude were produced, with 

 diluted ink on glass (75), and white paper or an illuminated 

 screen looked at through them, a chequered appearance was 

 observed. In one position, lines of a certain intensity separated 

 the heaps from each other, but the square places representing 

 the heaps looked generally lighter. In another position, when 

 but little reflected light came from the surface of the heaps, 

 their places could be perceived as dark, from the greater depth 

 of ink there. By care, another position could be found in which 

 the whole surface looked like an alternate arrangement of light 

 and dark chequers, fig. 25, not steady, but with F - 2 5 



a quivering motion, which further attention could 

 trace as due to a rapid alternation in which the 

 light spaces became dark and the dark light, 

 simultaneously. When, instead of glass, a bright 

 tin plate was used under the diluted ink, the 

 chequered spaces and their alternations could 

 be seen still more beautifully. 



98. It was in consequence of these effects that very large 

 arrangements were made (72), giving heaps that w T ere two inches 

 and a half wide each*; and now it was evident, by ordinary 



* This estimate is given in accordance with the mode of estimating the former 

 and smaller heaps, as if the heaps were formed simultaneously ; but it is evident 

 that if only half the number exist at once, each heap will have twice the width 

 or four times the area of those which can be formed if all exist together. 



