CHEMISTRY OF SOLIDS DESCH 237 



be this variation, when the chemical composition of the distinct 

 varieties may be identical, so far as analysis is able to give infor- 

 mation ? 



Again, the crystalline S3'stem will not account for the differences 

 in the building up of individuals to form aggregates. Rock salt 

 and cuprite crystallize in cubes, and the space lattice has a very 

 similar form in the two minerals ; but when the salt forms multiple 

 growths, the cubes arrange themselves in characteristic stepped 

 pyramids, whilst the red oxide of copper may form the most beauti- 

 ful hair-like threads, a tissue of scarlet silk, as liuskin calls it. 

 Neither mineral ever assumes a form which is characteristic of the 

 other, the simple cube being once departed from. Why should this 

 be? It is Imown that the presence of traces of foreign matter may 

 cause differences of habit, the most famous example being that of 

 the crj^stallization of common salt in octahedra instead of cubes when 

 a small quantity of urea is added to the solution, but the explanation 

 of these facts is still imperfect. 



The work of John sen and of Gross has shown that the appearance 

 of a face on a crj^stal placed in a supersaturated solution is really 

 determined by the velocity of growth in a direction normal to that 

 face, those faces being produced which have a minimum velocity 

 of growth. Some light is thrown on the subject by a study of the 

 growth of a crystal when solvent is completely excluded, the sub- 

 stance used being sublimed in a vacuum. This has been undertaken 

 by Volmer, who finds that cadmium, zinc, and mercury crystals 

 grow in this way in a high vacuum. When small nuclei are present, 

 those grow which have the face with the smallest velocity of growth 

 perpendicular to the stream of impinging molecules. The differences 

 between different faces are large, so that under these conditions 

 either flat tables or long prisms are usually formed, according to 

 the direction of the original nucleus. The crystal grows by the addi- 

 tion of thin laminse, probably only one molecule thick, which spread 

 over the surface. This is likely to be the process when tlie crystal 

 is growing in a solution or in a molten mass, as well as in the vapor; 

 and, in fact, when cadmium or tin is being deposited electrolytically 

 at a cathode, or vvdien lead iodide is being formed from a solution of 

 a lead salt and an iodide, the growth of the crystal may be watched 

 under the microscope, when a thin film begins to form at some point 

 on a face, and extends over the face, maintaining a uniform thicloiess 

 throughout. It is realized that in the presence of a foreign sub- 

 stance, either molecules or ions may attach themselves to such a sur- 

 face by their residual affinity, and this will necessarily affect the 

 addition of further layers of the original substance. In other words, 

 the velocity of crystallization in a direction normal to that face will 

 be changed. As the residual affinity of different faces of a crystal 



