u8 GUIDE TO DAIRYING IN SOUTH AFRICA 



various percentages of butter fat obtained in the 

 second speed trial, and should prove very conclu- 

 sively how easily the butter-fat content of cream 

 can vary very materially even from the same. milk, 

 if the correct speed is not maintained. 



In making the tests as shown in Fig. 13, 

 9 grammes of cream by weight were taken, so the 

 percentage of butter fat as depicted in the necks of 

 the bottles must in each instance be doubled. 



Turning now to the third set of experiments, 

 viz. the effect of separating under identical con- 

 ditions naturally rich and naturally poor milks, we 

 again arrive directly at another reason why cream 

 tests vary, even though the separator be turned at 

 the correct speed. Table III. gives the complete 

 figures obtained in this trial, and Fig. 14 the per- 

 centages of butter fat contained in the cream from 

 the respective milks. It will be noticed that the 

 percentage of butter fat contained in the cream 

 when using a 4 '8 % milk was 49 %, whereas when 

 a 2 '9 % milk was put through the machine under 

 exactly similar conditions only a cream containing 

 38 % butter fat was obtained. 



This to my mind proves very effectively that 

 even when working a machine at correct speed 

 there still may be variations in the cream tests 

 owing to the fluctuations from time to time which 

 will occur in the butter-fat content of milk used for 

 the production of cream. This is still more than 



