Reaction before Complete Equilibrium. 65 



of the reaction (the ice was also quickly weighed for the same 

 purpose) ; the surface at the time t„ t 2 , &c. was calculated 

 from the temperature of the liquid at the respective times. 



The velocity of ice melting cannot be investigated with the 

 same accuracy as that of ice separation. It is absolutely im- 

 possible to get an equation for the velocity of ice melting, if ice 

 separated from overcooled liquids (in small crystals) be used for 

 the experiment : there is no possibility of warming the liquid 

 by any means, without having a very important part of the 

 separated ice melted at the same time, and so also we cannot 

 transfer without loss separated ice to another liquid which is at 

 a temperature above 0° C. Under such conditions we cannot 

 know the initial temperature of the liquid t J nor the surface 

 of the solid ice in contact with the liquid t — t ov . Further, 

 every individual crystal is partly melting away, and we have 

 no possibility of getting any knowledge as to the variation of 

 the surface of the ice-crystals with the change of temperature 

 of the liquid. It will be seen later on that even small errors 

 in the determination of the values of t ov are quite sufficient to 

 make the value of the constant C incorrect by 100 per cent. ; 

 speaking more exactly, the constant cannot be evaluated. 



The cubes of ice were first cut out of an ice block and then 

 melted away and polished on heated plates of thick iron with 

 a perfectly smooth surface, and this was continued until they 

 received the desired accurate form of a cube. After having 

 been quickly weighed they were placed in an iron cage (the 

 rods of the cage were covered with thick indiarubber tubes to 

 prevent the formation of any grooves in the ice), which was 

 fixed to the stirrer; this prevented the cube of ice from 

 rotating, and always kept it below the surface of the liquid. 

 It was found that in this way a considerable portion of the 

 ice can be melted away without any essential change in the 

 /form of the cube being produced by it. 



5. The Velocity of Crystallization of supersaturated solutions. 

 (Equilibrium between the sepa rated salt and the salt solu- 

 tion.) 



I find that the equation 



log («i - f J ~ log fa - O + log (*. - t t ) - log (t - 1 2 ) 



= C (to - T t ) (t - t ov ) , especially = k(r 2 - rj), 



holds good, where t is the temperature of equilibrium, t 00 is 

 the temperature at which the crystallization is started, t l , t 2 

 are the temperatures of the liquid at the times r l} t 2 . The 

 method employed is the same as that of § 3. These reactions 

 Phil. Mag. S. 6. Vol. 2. No. 7. July 1901. F 



