CRYSTALLIZATION. 279 



energy will be most easily produced inside as well as outside, and in a 

 crystalline mass of course they will be parallel to the external faces of 

 the crystal. We see the same thing in the action of solvents. Most 

 metals assume a crystalline texture on cooling- from fusion, and when 

 slowly acted on by dilute acids the surfaces of greater energy are most 

 easily attacked, in accordance with the laws of energy, and the uiulis- 

 solved metal is left with surfa(;es of least energy which are the faces of 

 crystals. This is easily seen on treating a piece of tin plate or of gal- 

 vanized iron with very dilute a(iua regia. In fact, solution is closely 

 connected with surface energy. It is probably the low surface energy 

 of one form of crystals of sulphur which makes them insoluble in car- 

 bon disulphide, and this low" surface energy may be an electrical effect. 

 I pointed out that the develoi)m('nt of all the faces of a form and the 

 similar moditicationof all corresponding edges and angles of a crystal are 

 in general necessary in order to juoducc equilibrium under the surface 

 tensions. But avc sometimes lind crystals with only half the modilica- 

 tions required for symmetry. In such cases the surface tensions nnist 

 l)roduce a stress in the interior tending to deform the molecules. When 

 the ciystal was growing there must have been equilibrium, and there- 

 fore a pressure equal and op])osite to this effect of the surface tension. 

 There are various ways iu which we nuiy sui)pose that such a force 

 would arise. The electric field might give rise to a stress in op})osition 

 to the aggregati<m of the molecules in the closest possible way, and 

 then the crystal would grow such faces as would produce an equal and 

 opi)osite stress. Inequalities of temperature or the presence of mole- 

 cules of other kinds amongst those of the crystal might produce similar 

 results. When the stress due to electricity or to temperature was 

 removed by change of circumstances, that due to the surface tensions 

 would pL'rsist, and the crystal would be left with an internal strain. 

 Crystals of this .sort, with unsymmetric faces, generally betray the 

 internal strain either by developing electricity of oi)posite kinds at the 

 two ends when heated or cooled, or they affect polarized light, rotat- 

 ing the plane of polarization. That these effects are due to the internal 

 strain is shown l>v the fact that tourmalines and other ciystals which 

 are pyro-electric when unsymmetrical sbownosu(;h i)roperty when sym- 

 metrically grown. Also sodium chlorafe in solution, (piartz w hen fused, 

 and so on, lose theii' rotatory power. Substances which in solution show 

 rotatory power as a rule de\'clop unsymmetric crystals. This is well 

 seen in the tartrates. The constitution of the molecules must l)e such 

 that they will not with(»ut some strain form crystals; and e(iuilil)rium 

 when t])e crystal is growing is attained by means of theoi)])osing stress 

 due to want of syinmetiy in the surf;;ce tensions. In all such crystals 

 the rotatory i)ower of the solution <lisai)])ears in whole or in part. We 

 can not test this in biaxial crystals, l)ut, according to Dcs (Moiseaux, 

 suli)hate of strychnine is the onl.\' substance which shows rotation both 

 iu the solution and in the crystalline form, and in it the rotatory power 



