ON CRYSTALLOGAPHY. 143 



perceived, the liquid must bo pourod off, and a portion of new 

 liquid put <n its place ; otherwise the crystal is infallibly der- 

 troyed. Mr. Leblanc has observed, that this singular change be. 

 gins first at the surface of the liquid, and extends gradually to the 

 bottom ; so that a crystal, if large, may be often perceived in a 

 state of increase at its lower end, while it is disappearing at its up- 

 per extremity. Mr. Leblanc even affirms that saline solutions al- 

 most always increase in density according to their depth from the 

 surface. 



4. The phenomena of crystallization seem to have attracted but 

 little of the attention of the ancient philosophers. Their theory, 

 indeed, that the elements of bodies possess certain regular geome- 

 trical figures, may have been suggested by these phenomena ; but 

 we are ignorant of their having made any regular attempt to ex- 

 plain them. The schoolmen ascribed the regular figure of crys. 

 tals to their substantial forms ; without giving themselves much 

 trouble about explaining the meaning of the term. This notion 

 was attacked by Boyle ; who proved, that crystals are formed by 

 the mere aggregation of particles*. But it still remained to ex. 

 plain, why that aggregation took place ; and why the particles 

 united in such a manner as to form regular figures? 



The aggregation is evidently the consequence of that attractive 

 force which has been examined in the last section. But to explain 

 the cause of regular figures is a more difficult task. Newton has re. 

 marked, that the particles of bodies, while in a state of solution, arc 

 arranged in the solvent in regular order and at regular distances; 

 the consequence of which must be, that when the force of cohesion, 

 becomes sufficiently strong to separate them from the solvent, 

 they will naturally combine in groups, composed of those particles 

 which are nearest each other. Now all the particles of the same 

 body must be supposed to have the same figure ; and the combi- 

 nation of a determinate number of similar bodies must produce si- 

 milar figures. Hauy has made it exceedingly probable that these 

 integrant particles always combine in the same body in the same 

 way ; that is to say, that the same faces, or the same edges, al. 

 ways attach themselves together ; but that these differ in different 

 crystals. This can scarcely be accounted for, without supposing 

 that the particles of bodies are endowed with a certain polarity 



* Treatise on the origin of form* nnd qualities 

 TOL. TI. I. 



