On Solidification and Viscosity of Supercooled Liq uids. 239 



face of separation of the liquid and solid in which the fall 

 of pressure takes place, O the melting-point on the abso- 

 lute scale, and the actual temperature at the surface of 

 separation. 



If we regard F and a as constants and A as being propor- 

 tional to the viscosity of the liquid, we can write the above 

 formula thus : 



» = C~ 



where is a constant, s = o —0 the actual supercooling, and 

 V is the viscosity of the liquid. 



Since V increases as the temperature falls, v may attain a 

 maximum value and then fall when s is increased. 



The experiments described in this paper were undertaken 



with the object of testing the formula v = C^; and they show 



that the main features at any rate of the observed relation 

 between the velocity and temperature can be deduced from 

 the results for the viscosity by means of this formula. In 

 fact, when the constant C is so chosen as to make the maximum 



value of C=y- equal to the maximum value of the velocity, then 



s 

 the values of C=r^ agree fairly well with the observed values 



of the velocity. 



To find the extent to which the temperature of the soli- 

 difying surface, or rather that of the tips of the growing- 

 rays of solid, is raised above the original temperature of the 

 liquid, experiments were made in which a thermocouple was 

 immersed in the liquid, and the change in its temperature 

 observed as the boundary between the solid and liquid passed 

 it. The substance first used was salol, which was chosen 

 because it solidifies very slowly (maximum velocity about 

 4 millimetres per minute), so that the thermocouple has a 

 chance of taking up the temperature of the surface of 

 separation. 



The salol was contained in a glass U-tube which was im- 

 mersed in a water-bath at constant temperature, and the 

 thermocouple was put down one of the limbs of the U-tube. 

 The couple used was made from iron and nickel wires each 

 0"2 millim. in diameter, which were twisted together for 

 about 2 millims. and kept apart by a thin capillary of glass 

 round the iron wire. The iron and nickel wires dipped into 

 small mercury cups kept in the same water-bath, from which 



