316 Report of the Department of Chemistry of the 



been deposited on the walls of the bowl during the whirling. The 

 results of analysis of the " separator slime " deposited after each 

 two hours of whirling are given in the table following: 



These results show that two-thirds of the insoluble inorganic 

 phosphorus was removed during the first two hours of whirling, again 

 indicating that the phosphates are heavier than the casein. The 

 ratio of casein to phosphates is here also shown to be wholly irregular, 

 indicating no definite combination. 



Expressing the data in Table VII in the form of gram equivalents, 

 we have the figures contained in the following table: 



Table VIII. — Amounts of Acids and Bases Expressed as Gram Equivalents. 



Casein as gram 



equivalents 



of acid. 



Phosphates as 



gram 



equivalents 



of di-basic 



acid. 



Sum of gram 



equivalents of 



casein and 



phosphates. 



Gram 

 equivalents 

 of calcium. 



1st deposit. . 

 2nd deposit. 



81. 6 x 10_ 

 82.0 x 10 



81 . 1 x 10 

 50.4 x 10 



162.7 x 10 

 132.4 x 10 



154.5x10 , 

 134.6x10 



An examination of these figures shows that there is the same balance 

 between the acids (casein and phosphoric acid) and the bases (cal- 

 cium and magnesium) in the two separate deposits, even when the 

 inorganic phosphorus is as unevenly distributed between them, which 

 furnishes proof for two points: (1st) The inorganic phosphorus 

 must be in the form of neutral calcium phosphate (CaHPCU), for 

 otherwise the balance between bases and acids would be altered, 

 acid calcium phosphate (CaH 4 P 2 0s) giving an excess of acid and 

 tri-calcium phosphate (Ca 3 P20 8 ) an excess of base in the " slime " 

 deposited in the first whirling. (2nd) If the phosphates were in 

 combination with the casein, we should expect to find the ratio 

 between the organic phosphorus and the inorganic phosphorus the 

 same in both deposits, but, instead of uniformity, we find the ratio 

 showing so wide a variation as 1:1.92 and 1:1.19 in the two cases. 



