104 ANALYSIS OF MILK. 



a briskly boiling water-bath, and kept there for exactly seven 

 minutes. The flask and its contents are then cooled, reweighed, 

 and the amount of water lost by evaporation replaced, and the 

 whole thoroughly mixed, filtered, and examined polarimetrically, 

 the temperature {t) at the time of polarisation being noted. 



The reading after inversion is subtracted from that before 



,,..,,, 142-66 -0-5i° ,, , ^ ,. 



mversion and divided by r-r- (t = degrees Centi- 

 grade), the result will be the reading due to cane sugar (C), and 

 this, subtracted from the reading before inversion, will be the 

 reading due to milk sugar (M). 



From these readings the percentages of cane sugar and milk- 

 sugar are calculated as follows : — 



Let W = number of grammes of condensed milk per 100 c.c., 

 F = percentage of fat, 

 P = percentage of proteins, 



then percentage of cane sugar is — 



/ W X F X 1-08 -f- W X P X 0-84 \ 



C X V 100 ; X ^- X ^-50 



^ ^ 100 1-33 ^ W • 



and the percentage of milk-sugar is — 



/W X F X 1-08 -f W X P X 0-84 



100 



«) ^ 



-^ s 100 7 1 100 



^ ' 100 ^ 1-106 ^ W • 



He also, at the author's suggestion, has worked out the following 

 method, which avoids the calculation ; add to 100 c.c. of diluted 

 condensed milk — 



(a) 10-25 c.c. of water, 



(6) a number of c.c. equal to W X F X 1-08, 



(c) 3 c.c. of acid mercuric nitrate ; 



polarise, and calculate the reading due to cane sugar as before. 

 Then the difierence between the reading before inversion 



and that due to cane sugar multiplied by -^ will give the per- 

 centage of milk-sugar, and the reading due to cane sugar multi- 

 plied by ^ will give the percentage of cane sugar. 



The fat may be estimated by a rapid method, as the deviation 

 from the truth due to an error in the fat estimation of 0"5 per 

 cent, is only O'l per cent, in the cane sugar ; the possible error 

 due to the use of the simplified method does not exceed i 0'2 

 per cent. Richardson and JafEe have proposed a somewhat 



