S. T other — Growth of Crystals. 541 



solubility than the other parts of the crystal. Since a crystal 

 is enlarged by deposition only at those points where it is in 

 contact with a supersaturated solution, the degree of concen- 

 tration necessary for crystal growth is greater at the base of 

 the crystal than elsewhere. This means that a crystal is able 

 to raise itself by downward growth, as in the experiments out- 

 lined above, only when its base is in contact with a solution 

 supersaturated with respect to it, and that the degree of con- 

 centration necessary for growth increases with any increase in 

 the weight supported per unit area of the supporting surface. 



When supersaturation is brought about either by evaporation 

 or by cooling, the denser and therefore more concentrated por- 

 tions of the solution tend to settle to the bottom, and in the 

 latter case the bottom of the crystallizing dish is itself a cooling 

 surface. It is possible that adsorption in some slight degree 

 helps to increase the concentration of the layer of solution in 

 contact with the bottom of the dish, but adsorption can not be 

 the direct cause of the upward pressure developed by crystals 

 in the foregoing experiments, for otherwise the presence of 

 unweighted crystals in the same dish with weighted crystals 

 would not prevent the upward growth of the latter. The 

 weighted crystals fail to grow in height because the unweighted 

 crystals, with their lesser solubility, prevent the necessary 

 degree of supersaturation in the layer of solution resting on 

 the bottom of the dish. Another reason for believing that 

 adsorption is of no importance in the formation of cavities and 

 upward growth of crystals is the absence of appreciable hollows 

 on the upper surface of crystals grown under heavy weights or 

 even under glass plates weighing only 0*17 gram. There are 

 no grounds for assuming that adsorption is any more efficacious 

 in producing hollows on the under than on the upper surface 

 of a crystal when both surfaces are in contact with similar 

 material. In the few cases where insignificant shallow hollows 

 were observed on the upper surface of crystals they were 

 probably due to the cover plate acting as a cooling surface. 



The objections raised against the adsorption hypothesis apply 

 with equal force to the theory that the pressure accompanying 

 crystal growth, in the experiments described above, is in any 

 considerable degree due to those forces which are peculiar to 

 the growth of crystals and which determine the development 

 of crystal faces. The thin layer of solution due to capillarity 

 and adsorption is always found between the upper surface of a 

 crystal and the cover glass, as well as between the lower sur- 

 face and its support; and the concentration of the solution in 

 contact with the upper surface can be increased by diffusion 

 from a surrounding solution of higher concentration just as 

 easily as the solution under the base of the ciwstal. The upper 



