in the Cavities of Minerals. 501 



menon, seen frequently, and under more favourable circum- 

 stances, not only from its intrinsic interest, but because a 

 distinguished philosopher had treated with an air of incredu- 

 lity an observation which I had made of a similar kind. 

 There can be no higher testimony to the novelty and import- 

 ance of a scientific fact, than when a competent judge raises it 

 to the supernatural. 



I come now to describe a property of the dense fluid, so 

 new and remarkable that it cannot fail to excite the attention 

 of chemists. This fluid occupies the whole of a large cavity 

 ABCDE, fig. 2, with the exception of a bubble at A, which 

 must be either a vacuum, as it is in all cavities containing only 

 this fluid, or a bubble of the expansible fluid, or the vapour 

 of the dense fluid, or some gaseous body. It cannot be a 

 vacuum, because it expands with heat, in place of being filled 

 up by the expansion of the fluid. It cannot be the expansible 

 fluid, because cold would contract it, and produce a vacuity. 

 It cannot be the vapour of the expansible fluid, because there 

 is no expansible fluid to throw it oif, and it has not the optical 

 properties of its vapour. It cannot be the vapour of the fluid 

 in the cavity, for it does not disappear by the application of 

 cold, and does not become a vacuit}', which fills up by the 

 expansion of the fluid. It is therefore an independent gas, 

 which exhibits the following phaenomena. 



When heat is applied, the bubble A expands, not by the 

 degradation of its circular margin passing into vapour, as in 

 the vapour cavities described in a former paper, but by the 

 rapid enlargement of its area. When it attains a certain size, 

 it throws off a secondary bubble B, which passes over a sort 

 of ridge or weir inno^ in the bottom of the cavity, and settles 

 at B. If the heat is continued, these two bubbles increase in 

 size ; but it was instantly withdrawn when B had begun to 

 swell. As the topaz began to cool, both the bubbles A and 

 B quickly contracted. The primary bubble A returned gra- 

 dually to its original condition, and B, when reduced to a 

 single speck, would have disappeared, had the cooling not 

 been stopped. This speck swelled again by the application of 

 heat, and so did the bubble A. When the speck at B was 

 allowed to vanish, which it did on the spot which the bubble 

 occupied, the fresh application of heat did not revive it at that 

 spot, but merely expanded the primary bubble A, which 

 again threw off" a secondary bubble B, which exhibited by 

 heat and cold the same phaenomena as before. These ex- 

 periments I repeated many times with the same result. It will 

 naturally be asked, what was the condition of the fluid itself 

 wliich has the properly of expanding by heat ; and what be- 



