456 PROCEEDINGS OF THE GEOLOGICAL SOCIETY. 



of the crystal, more salt were to be deposited at the upper part of 

 the notch than at the bottom, as shown by the dotted line, a portion 

 of the liquid would be enclosed in the crystal, and form a fluid- 

 cavity like that seen on the right-hand side ; whereas, if the crystal 

 grew so as to enclose the projecting portion of fluid by a plane 

 parallel to the face of the crystal, a fluid-cavity like that on the 

 left side would be produced. These two cavities, as well as 

 figs. 3, 4, and 5, illustrate the usual character of those in common 

 salt or chloride of potassium. When deep, they have a broad dark 

 outline, like fig. 4 ; but when flat and very shallow, they are bounded 

 by a narrow black line, as in fig. 5. Often they are much longer 

 in one direction than in any other, so as to be tubes, like fig. 3 ; 

 and this is especially the case in the long prismatic crystals of chlo- 

 ride of potassium sometimes formed on the cooling of a warm 

 solution. 



The manner in which the cavities are produced necessarily causes 

 them to be full of fluid at the time of their formation. If this takes 

 place at the ordinary temperature, and they are kept in fluid, they 

 remain quite full, and none of the cavities contain bubbles. When, 

 however, they are formed at a higher temperature, the fluid of course 

 contracts on cooling, and, just as happens on the cooling of a glass- 

 tube filled with hot water, a vacuity like a bubble makes its appear- 

 ance, as shown in fig. 7. In order to obtain crystals slowly deposited 

 at the heat of boiling water, I employ a flat-bottomed flask, over the 

 mouth of which I tie a piece of blotting-paper, and keep it in a bath 

 of boiling water, with the long neck projecting through a hole in a 

 metal plate covering the bath. 



For these experiments no substance is more convenient than com- 

 mon salt, because it is so very little more soluble in hot than in cold 

 water. Having, then, a concentrated solution in the flask, it evapo- 

 rates through the blotting-paper, and crystals are slowly deposited at 

 the heat of boiling water. When a number of about -^jth of an inch 

 in diameter are formed, the hot solution is quickly poured off, and a 

 small quantity of a concentrated cold solution added, so as to prevent 

 the deposition of salt when the remaining solution becomes cold ; and 

 then, separating the larger and very small by means of sieves, the 

 crystals of about ^frth t° sM-h °f an mcn m diameter are mounted 

 in glass cells in a concentrated solution of the salt. On examining 

 them with the microscope, it is seen that the greater part of the ca- 

 vities contain small bubbles, produced by the contraction of the fluid. 

 Adopting the same method with chloride of potassium, in some cases 

 most excellent thin flat crystals are formed, containing many very 

 interesting cavities. A portion of one of these crystals is drawn in 

 fig. 6, which serves to show the great number and peculiar form of 

 the cavities, and how they occur in bands parallel to the edges of 

 the crystal. Unlike when formed at the ordinary temperature, 

 manv of the cavities are of very complicated forms, as shown in 

 figs."8 and 10. 



In order to ascertain the relative size of the fluid and bubbles in 

 the cavities, it is best to measure with the micrometer such long re- 



