1885.] NEW-YORK MICROSCOPICAL SOCIETY, 141 



ena of dilatation produced by a gentle heat. Other cavities con- 

 tain two liquids, — the heavier saline solution next the wall and 

 enveloping the salt-crystals, and the lighter carbon dioxide 

 within. In this case three outlines are usually exhibited in each 

 cavity, — that of the cavity itself on the outside, then the limiting 

 line between the layers of saline solution and carbon dioxide, 

 and, within, the circular outline of the gas-bubble. In some 

 cavities the volume of the gas-bubble is very large in proportion 

 to the amount of liquid carbon dioxide. On heating, there 

 ensues, in this" case, a rapid evaporation of the liquid into the 

 inner space. Thilorier has stated, in the record of his laboratory 

 experiments on tubes partly filled with this volatile liquid, that 

 when half full, such a tube acted as a retrograde thermometer, 

 the increase of temperature being marked by a diminution of the 

 volume of the liquid on account of its vaporization ; but, when 

 two-thirds full, it served as a normal thermometer of great sen- 

 sitiveness, the volume of. the liquid expanding with the rise of 

 temperature. So, in these cavities with large bubbles, the 

 amount of liquid carbon dioxide present is so small that, at the 

 temperature then prevailing, it is reduced to so thin a layer as to 

 be insufficient to wet any longer the inner surface of the saline 

 layer. Then the saline solution lies as usual next the wall, with 

 a huge gas-bubble immediately in contact ; but within the latter 

 lies the limpid liquid carbon dioxide gathered into a ball, the 

 liquid and gas having interchanged positions. All these varieties 

 of condition may be distinguished after a little study, especially 

 by noting the depth of the dark rim or shadow of the outline of 

 the lighter liquid next the denser, or of the gas-bubble next 

 either liquid — the depth of the shadow being in each case pro- 

 portionate to the difference in density of the two fluids. 



In some cases the bubble adheres firmly to the wall or corner 

 of a cavity and cannot be dislodged. Commonly it runs freely 

 around the wall of the cavity, as the thin section is rotated upon 

 the stage, always, of course, adhering to the uppermost side of 

 the cavity as in a natural spirit-level. Thus it mi^ht be fancied 

 that any sagacious gnome, having lost his way in the subterranean 

 recesses, might determine his reckoning, as to up and down, by 

 the position of the bubbles of the liquid-cavities in the rocky 

 walls surrounding him. 



An interesting feature of the smallest cavities, usually less than 



