i MILK 107 



of the hand. In this way the large pieces of ice soon begin 

 to melt. When the temperature has risen five- or six- 

 hundredths of a degree the cylinder is returned to the freezing 

 mixture before all the frozen milk has melted. The fall of the 

 mercury must now be carefully watched, and the stirrer raised 

 arid sunk again twice each second. After a few seconds the 

 column of mercury stands still, so if there is no movement 

 of it after the stirrer has been raised twice or three times , 

 the temperature may be read off. After one or two minutes 

 another reading should be taken, and the thermometer should 

 then show the same temperature as before. If this is the case, 

 then the reading may be taken as correct. . Sometimes, but 

 not often, the temperature rises half- to one-hundredth part 

 of a degree in this time, and if so this latter reading is the 

 correct one. Should the temperature, however, fall during 

 this pause of a few minutes, then the determination must 

 be made afresh. 



In each series of cryoscopic tests a correction of the zero 

 has, moreover, to be made. This is because each thermometer 

 alters slightly for two or three years after being made. The 

 glass shrinks as it were, and so the bulk of the thermometer 

 becomes smaller, and the zero point is consequently raised. 

 Where the measurements have to be so carefully made as in 

 the present case, the change cannot be neglected. 



The correction is made simply by determining the freezing- 

 point of distilled water, and the position of this on the scale 

 is taken as the zero. 



Winter 1 has made use of the fact that a lowering of the 

 freezing-point, A, is almost proportional to the addition of 

 water, and has constructed the following formula : 



a 



where E is the volume of water which has been added to 

 the milk, V the total volume of water and milk, a the normal 

 lowering of the freezing- point of milk (0'55 C.), and A the 

 observed lowering of the freezing-point. 



The following table, which has been drawn up by Winter, 

 permits of the percentage of added water being directly 

 ascertained from the lowering of the freezing-point : 



1 Bull de la Soctite Chim., 1895, Vols. XIII XIV, No. 24, p. 1,102. 



