in the Cavities of Minerals. 17 



two fluids, by means of transmitted light. The sides of the ca- 

 vity being inclined to one another, like those of a prism, it is 

 manifest, that if a b c is the boundary of two fluids of equal re- 

 fractive power, the image of a luminous object will have the 

 same deviation, by the refraction of both. As the cavity, how- 

 ever, is too minute to permit an image to be distinctly seen 

 through it, it becomes necessary to look with a microscope at 

 the illumination of the surface of the cavity, and if the two re- 

 fractive powers are equal, the portion above a b c will be dark, 

 when the portion below it is dark, and vice versa. I found, how- 

 ever, that the portion of fluid B a b c was often dark, when the 

 second fluid below a be was light, and I therefore concluded that 

 this arose from their unequal refraction. To this conclusion it 

 may be objected, that the inclination of the refracting faces 

 might accidentally be different behind B a b c, although it is not 

 likely that the portion possessing this difference of inclination 

 would be bounded by a curve line a b c. I therefore applied heat 

 to the specimen, and, by expanding the new fluid at B and C, 

 the bounding lines were made to move from a be, def, to mnn'o, 

 and I remarked, that, during this change of position, the boun- 

 dary of the two fluids was always the boundary of the unequal 

 shades produced by unequal refraction. 



As the arrangement of the fluids which enabled me to make 

 these experiments, possesses a peculiar interest, I have carefuUy 

 looked for similar cavities, but I have not succeeded in finding 

 more than a few examples, one of which is represented in 

 Fig. 12., as it appears at the temperature of 32. This cavity 

 consists of two wide portions, separated by a narrow channel. 

 The new fluid occupies the portion between cc, dd, and also that 

 between a a and b b, these two portions being separated by the 

 second fluid dd, a a. The whole vacuity exists at V. If we now 

 apply heat, the new fluid at N and N expands, and the bounda- 

 ries dd, a a and b b, advance towards B. The vacuity V becomes 

 VOJL. x. P. i. c 



