HOSTETTER: APPARATUvS for growing CRYSTAIvS 93 



form flattened octahedra, diagrams of which are shown in position 

 in the sketches of the crystalHzer. Ordinarily such crystals 

 are nearly free from any of the related faces such as the cube 

 and the rhombic dodecahedron; the cube face may be observed 

 but it is always small and the rhombic dodecahedral face is 

 even less developed. Such is the case when the crystals are 

 grown continuously without periods of solution intervening. 

 If, for any reason, the growth of the crystal is stopped and solu- 

 tion of the crystal takes place the edges of the crystal are rounded 

 in the initial stages of dissolution: when conditions are changed 

 so that growth recommences the rounded portion of the crystal 

 flattens during growth and in so doing a rhombic dodecahedral 

 face is developed. During the future growth of the crystal 

 this new face is very prominent and in none of the cases observed 

 here has it ever filled out to form the sharp edge of the original 

 octahedron. A similar development of the rhombic dodeca- 

 hedral face can be induced if the edge of the octahedron is re- 

 moved by other means and the crystal then allowed to grow. 

 The "repair" of crystals during growth is thus seen to follow 

 along the lines of least resistance. 



It should be emphasized here that the mere fact that all es- 

 sential conditions are under control in the crystal-growing ap- 

 paratus described above, is not, in itself, a guarantee that any 

 salt can be made to form large crystals under the conditions 

 obtaining therein. Some salts may be readily enough crystal- 

 lized in large well-formed crystals — other salts under the same 

 conditions will yield a multitude of small crystals rather than 

 a few large ones. Potassium alum and sodium chlorate were 

 grown very successfully in this apparatus but experiments with 

 ammonium chloride yielded only a mass of fine, fernlike crystals 

 instead of growth on certain crystals which had been introduced 

 as nuclei. In this case the effect was not caused by incorrect 

 adjustment of conditions for these fine crystals appeared and 

 increased in size in the crystallizer, thus showing that con- 

 ditions were optimum. On several occasions all crystals except 

 one were carefully removed from the crystallizing chamber 



