131 



The growth velocity coefficients foi' the different faces 

 of a crystal do not vary at the same rate when the tem- 

 perature is changed, as we indicated above; hence, some 

 faces become prominent at certain temperatures and the 

 crystal develops more in the direction of these faces. The 

 temperature of crystallization determines, therefore, the 

 shape of the crystals. In general when the velocity of 

 crystallization is very low for one face it is low for all the 

 faces and the crystal has the shape of a spherulite. 



Walton and Judd (1914) studied the rate of growth of 

 ice in a long glass tube previously filled with distilled 

 water and subcooled to various temperatures. They 

 initiated crystallization by ice-seeding at one end of the 

 tube and measured the velocity of the congelation-wave as 

 it moved toward the other end. They found a rate of 65 

 mm. per second at - 8°. 



Hartmann (1914), in experiments of the same kind, ob- 

 tained 46.6 mm. per second at - 7°. 



Tammann and Biichner (1935a), who succeeded in keep- 

 ing w^ater subcooled to - 13.4°, found 41.3 mm. per second 

 at - 7.2° and 96.8 mm. at - 13.4°. They also measured the 

 crystallization velocity when ice was formed from heavy 

 water and obtained almost the same values as with ordinary 

 water, at the same degrees of subcooling (that is, for tem- 

 peratures calculated from + 3.8° as the freezing point 

 of heavy water and from 0° as the freezing point of ordi- 

 nary water). 



The same authors (1935b) determined the lowering of 

 the crystallization velocity of water when substances such 

 as sodium chloride, sulphuric acid, glycerine, alcohol and 

 sugar, were dissolved in it. They found that, at 5 degrees 

 below the freezing point, 0.85 mole of sodium chloride in 

 1000 gr. of water slows the rate of congelation to about 6 

 mm. per second; while 1.1 and 2.9 moles of sugar, in the 

 same conditions, slow it to about 0.3 and 0.02 mm. respec- 

 tively. Sugar, it was remarked, has a particularly strong 

 retarding effect. It might be of interest to notice that in 

 very dilute concentrations (about 0.01 mole per liter) the 



