SORBY STRUCTURE OF CRYSTALS. 457 



gular tubes as figs. 9 and 50. It may also be determined with suffi- 

 cient accuracy from cavities like figs. 101 and 113, which are equally 

 deep throughout, and have a flattened vacuity, so that the propor- 

 tion between their areas is that between their volumes. If, however, 

 neither of these kinds can be found, the best approximation that can 

 be made is to be derived from such cavities as are nearly equiaxed, so 

 that the relative magnitude of the cavity and vacuity equals the cube 

 of the ratio of their diameters. 



In the very excellent tubular cavity (fig. 9) in chloride of potassium 

 formed at a heat not much below that of boiling water, the vacuity 

 is about *025 of the fluid ; but, if the cohesion of the sides produced 

 no effect, it should have been about *030. Hence I think we must 

 conclude that the cohesion of the liquid to the sides of so small a 

 cavity slightly diminishes the size of the vacuity, either by stretching 

 the fluid or the substance of the crystal. That very minute cavities 

 do produce such an effect, is proved by the fact that they often con- 

 tain no bubbles, as though the cohesive force entirely counteracted 

 the contraction of the fluid, and operated like the reverse of a very 

 great pressure ; and M. Berthelot has also shown (Annales de 

 Chimie, 3 e ser. t. xxx. p. 232) that this occurs to a certain extent, 

 even in glass- tubes. 



When fluid-cavities are large, the bubbles move about, if the cry- 

 stal be turned, like those in spirit-levels ; but when small, this test 

 is not easily applied. We can readily see, with a high magnifying 

 power, that the constant tremor of the ground causes the bubble in a 

 level to be in constant motion ; but no such movement can be seen 

 in the larger fluid-cavities. In the very small, however, the bubbles 

 move in a most striking manner, as if they were minute animalcules 

 swimming about and exploring every part of the cavities. The true 

 physical cause of this movement still remains to be determined ; for 

 hitherto I have not been able to form any theory that was not appa- 

 rently upset by facts subsequently discovered. Whatever be the 

 cause, the fact of the movement is of very great value in these in- 

 quiries, since, when a bubble moves about, the substance in the 

 cavity must of course be a liquid. 



The peculiarities of fluid-cavities formed when crystals are deposited 

 from a solution containing some other salt, can be studied to great 

 advantage by crystallizing common salt from a solution of bichromate 

 of potash. If the solution be saturated with the bichromate at the 

 ordinary temperature, it is of a deep yellow colour, and the fluid- 

 cavities in the crystals of common salt deposited from it are seen to 

 be filled with this yellow liquid ; and the crystals are thus rendered 

 yellow, and remain so when mounted in a colourless solution of com- 

 mon salt. If, however, the crystals be formed at the temperature 

 of boiling water in a solution containing more of the bichromate than 

 can remain dissolved at the ordinary temperature, small crystals of 

 the deep red colour of that salt are deposited inside the fluid-cavities, 

 as shown by figs. 1 1 and 1 2. When the solution is saturated at 

 100° C. with both common salt and the bichromate, so that, on slow 

 evaporation at that temperature, crystals of both salts are deposited, 



