SORBY — STRUCTURE OF CRYSTALS. 461 



form of the crystal and that of the cavities. For example, in the 

 cubic crystals of the chlorides of sodium and potassium they are 

 rectangular, as seen in figs. 4, 17, & 19; and in the octahedral 

 crystals of alum they are often equilateral triangles, like figs. 28 & 

 30 : but, from forms thus essentially related to the planes of the 

 crystals, they pass into all kinds of irregular shapes only slightly 

 related, as shown in a remarkable manner by the cavities in chloride 

 of potassium formed at 100° C. (figs. 8 & 10). 



The arrangement of the cavities is also sometimes related to the 

 form of the crystal, as shown in fig. 1 ; and bands parallel to the 

 bounding edges are very common, being in fact lines of growth, pro- 

 bably indicating variations in the rate of deposition. In some cases 

 they occur as bands of single cavities, twisting and curving about 

 without any very definite connexion, as shown by fig. 32 ; or they 

 are scattered promiscuously through the entire crystal. 



c. Expansion of fluids hy heat. 



By the experiments described above, I have shown that, at the 

 temperature at which they are formed, the fluid-cavities in crystals 

 are full of the fluid, and that, if they be examined at a lower tempe- 

 rature, they contain vacuities, owing to the contraction of the fluid 

 on cooling. Hence I think it is only reasonable to conclude that, 

 provided the temperature were not known, it might be ascertained 

 approximately by determining what increase of heat would be re- 

 quired to expand the fluid so as to fill the cavities. In some cases 

 this can be learned by direct experiment ; but generally it cannot, and 

 we must have recourse to calculation, From the nature of the case 

 the temperature is that required to cause the liquid to expand so 

 much that the increase in volume is equal to the size of the vacuity. 

 Taking, then, the volume of liquid for unity, and representing the 

 relative sizes of the vacuity by V, it is easy to perceive that, if the 

 law of the expansion of the liquid were known, and the value of V 

 had been ascertained by observation, the temperature could be cal- 

 culated. 



In order to be able to do this, I have (June 1858) made an ex- 

 tensive series of experiments to ascertain the law of the expansion of 

 water and saline solutions up to a temperature of 200° C. (392° F.). 

 This I did by hermetically enclosing the liquid in strong glass tubes 

 about 2 inches long and ^th of an inch internal diameter, and 

 heating them in a bath of paraffine, with such arrangements that the 

 increase in volume could be measured, by means of a micrometer 

 microscope, to within the y^^th of an inch. This, however, is not 

 the place for anything but the general conclusions. By appropriate 

 experiments and calculations, I find that the increase in volume 

 may be represented very accurately by an expression of the form 

 Y=Bt + Ct 2 , where t is the temperature in degrees Centigrade, and 

 B and C constants, the value of which depends upon the nature of the 

 aqueous solution. Perhaps, indeed, in reality there may be terms 

 involving higher powers of t than t 2 ; but if so, they are so nearly 



2h2 



