124 Dr. M. Wildermann on an Experimental 



In the above tables under a and /3 two values are given. 

 In the case of cane-sugar, alcohol, and urea (with the excep- 

 tion of the solution of urea which contains 0*001813 molecules 

 per litre) there were, with all concentrations, fine needles of 

 ice in the solution, but no ice-cap on the thermometer. In 

 square brackets there are given under a and /3 the values 

 which have been obtained when the freezing-point of water, as 

 read off when there was a well-closed ice-cap round the bulb 

 of the big (and the little) thermometer, is taken as the true 

 zero ; the other figures (without brackets) under a and /3 are 

 obtained when the freezing-point of water, as read off when 

 there is no ice-cap round the bulb, is taken as the true zero. 



In the latter case (of a well-closed ice-cap) the freezing- 

 point of water has been found to be under my experimental 

 conditions on an average O- 0016 higher *, and may err, like 

 all determinations in very dilute solutions, by about 1, 2, or 

 less often 3 ten-thousandths of a decree. In the above tables 

 the figures and results are given exactly as they were obtained 

 from the observations. In these tables, in the cases of cane- 

 sugar and urea, Van't HofFs constant appears very clearly if 

 the completely closed ice-cap be avoided in the determination 

 of the freezing-point of water, and would appear still more 

 clearly if one were to make the allowable correction of from 

 1 to 2 or 3 ten-thousandths of a degree : in short, so far 

 as the degree of exactitude of the method itself allows, Van't 

 HofFs constant appears very clearly. In the cases of alcohol 

 and cane-sugar the values of /3 fall in the more dilute solutions 

 short of 1*87 by about 1^ per cent. 



The importance of the question on which we are engaged 

 induces me to point out that Van't HofFs constant can also be 

 demonstrated most distinctly, independently of the freezing- 

 point of water. Since cane-sugar, urea, and alcohol can only 

 be very little dissociated, we can determine the several de- 

 pressions and concentrations, starting from a given solution 

 which does not cause the formation of an ice-cap round the 

 bulb of the thermometer (instead of taking water as the 

 starting-point), and state the results in relation to this ; by 

 this means any error which attaches to the determination of 

 the freezing-point of water is eliminated. Starting, then, 

 from the most dilute solution, in which no ice-cap was 

 observed, the following results are obtained : — 



* The convergence temperature was here above the freezing-point 

 temperature, 



