ANALYTIC PROCESSES 29 
tion of casein are mixed, the albumin precipitated 
by Almén’s tannin reagent, filtered, and the 
nitrogen in the precipitate determined as above. 
The same factor is used. 
Almén’s reagent is prepared by dissolving 4 
grams of tannin in 190 c.c. of 50% alcohol and 
adding 8 c.c. of acetic acid of 25%. 
In a mixture of milk and whey (prepared with 
rennet) in about equal parts, Richmond and 
Boseley found about 0.3% of albumoses not pre- 
cipitated by the copper sulfate nor by magnesium 
sulfate, but precipitable, along with the albumin, 
by a solution of tannin. The separation may be 
effected by diluting the filtrate from the magne- 
sium sulfate precipitation, acidifying slightly with 
acetic acid, and boiling, when the albumin will be 
coagulated and precipitated. The albumoses 
may be separated by filtering the solution and 
precipitating with tannin solution. The pre- 
cipitated proteins are best estimated by de- 
termining the nitrogen in the moist precipitate. 
The separation of the proteins may be effected, 
though less accurately, but the use of acetic acid, 
as recommended by Hoppe-Seyler and Ritt- 
hausen. 
Leffmann and Beam have modified the process 
to avoid the delay and trouble of washing the 
precipitate, as follows: 10 c.c. of the milk are 
mixed with saturated magnesium sulfate solu- 
