Infiuevce of the Separator on Bacteria. 63 



Table I. 



A comparison of -this table with triose following demon- 

 strates that the amount of contamination in every case is en- 

 tirely negligible, being in fact much less than the experimental 

 error. 



The milk for the first twenty-three tests was good market 

 milk, and for the last fifty-five tests was milk hauled in from 

 distant farms and fairly representative of such milk. Table II 

 gives an average of the first twenty-three tests. 



Table II. 



This table shows, first, that a somewhat larger proportion of 

 bacteria go with the skim milk than with the cream; second, 

 that the number of bacteria in the mixed milk (after separa- 

 tion) is nearly double the number found in unseparated milk; 

 and third, that the bowl sediment contains nearly five times as 

 many bacteria per gram as does the unseparated milk per cc. 

 This indicates that the organisms grow in masses in the whole 

 milk, and that the usual method of estimation does not give the 

 actual number present. The enormous centrifugal force may 

 kill large numbers, but the counts show an actual increase. 

 The average of fifty-five tests on a diff'erent grade of milk is 

 given in Table III. In this series tests were not made on the 

 bowl sediment. 



