in the Cavities of Minerals. 21 



ness of the cavity, and not a portion of that thickness, as in 



Fig. 18 *. 



With the view of confirming this explanation, I took a cavity 

 AB, Fig. 22., in which the new fluid N occupied the whole 01 

 one side of the cavity, and the second fluid W the whole of the 

 other side. Having made the vacuity vanish, and increased the 

 heat to about 200, the effect of this was to expand N, and make 

 the boundary a b move very slowly towards A ; but in a short 

 time, a portion of the fluid W, which had thus been pressed out 

 along the bottom of the cavity, made its appearance at the end 

 B, and gradually increased in quantity as a b moved towards A. 

 The new fluid then occupied the space between the dotted lines 

 c d and ef, which contained a greater area than the space be- 

 tween a b and B. The portion efB of the second fluid remained 

 for two hours in the position shewn in the figure ; but being 

 connected below N with the other portion cdA, it was drawn 

 over to the other side, and occupied its original position, as 

 shewn by A a b. 



In one of the Quebec crystals of Quartz, where the cavities 

 are filled with a slightly yellowish fluid, I observed a very deep 

 cavity, such as that shewn in Fig. 23., where the globule N ex- 

 panded very considerably to the width of n n by a considerable 

 heat. I sought in vain for a vacuity, which, however, might 

 have been concealed in a cavity of such a depth, and of such ir- 

 regularity of surface ; but, upon plunging the crystal in hot wa- 

 ter, and applying the microscope, I observed two very minute 

 globules, either of vacuity or something else, floating within N, 

 which gradually diminished and disappeared. During another 



* In Figs. 20. and 21. there are small squares, such as S, S, within the cavities, 

 which seem to be filled up with crystallized matter. These squares being sometimes 

 united only by contact with the surface of the cavity, exhibit very brilliantly the 

 colours of thin plates. 



