150 EXPERIMENTAL FARMS. 



Experiment B. 



A second series of composite tests was made during the week, May 15th to 22nd. 

 The mathematical average of the daily tests of the samples going to make up the 

 composites A and B was 3'05 per cent fat. 



The series was carried on in duplicate : — 



Composite sample A.= 600 c.c. milk -{- 1"2 grms. potassium bichromate. 

 " " B.= 600 c.c. " 4- 2-4 " " " 



(This is equal to 18 grains and 36 grains to the pint, respectively.) 



The fat readings are as follows: — 



A. B. 



Date. Fat, Fat, 



per cent. per cent. 



May 2Yth 3-0 295 



June 3rd 3-1 3-00 



" 10th 3-1 3-00 



" 17th 3-0 300 



" 24th 3-0 2-95 



July7th 30 2-90 



After this Inst date the fat in both samples became slightly curdy and the percent- 

 age gradually lower. The samples were shaken daily and tests made every week until 

 September 26th, when the respective readings were A. 2*20 per cent fat; B. 2'50 

 per cent fat. Both samples had curdled and were measured with diflSculty. 



For more than seven weeks the milk had retained its fluidity and yielded a cor- 

 rect percentage of fat, though the preservative had been increased to 36 grains to 

 the pint. 



Experiment C. 



A further series was then comn^enced in which composite sample A. had 36 

 grains potassium bichromate, and B. 72 grains potassium bichromate to the pint. 

 The fat readings were made at intervals of one week from June 3rd to September 

 30lh. The mathematical average of the peicentages of fat, obtained from the daily 

 tests of the samples making the composites, was 3'51. 



The results were in accord with those obtained in Experiment B., viz., the per- 

 centage of fat after a time became gradually less. On September 9th, the fat in 

 both samples read 3*1 per cent. It was further observed that, as the quantity of 

 bichromate was increased, the volume of acid had to be slightly decreased in order to 

 get clear readings. Towards the close of the test period, as the bichromate becomes 

 reduced, however, the volume of acid must again be increased. 



Experiment D. 



Date of composite sample. May 29th — June 3rd. Potassium bichromate to the 

 amount of 230 grains to the pint was added. The percentage of fat, as obtained from 

 the mathematical average of the daily tests, was 3*35. 



The normal quantitj^ of acid (17*5 c.c.) charred the fat so that it could not be 

 read. The amount was gradually reduced till the readings became distinct, and as a 

 result it was found that 11*5 c.c gave clear readings and the correct percentage of 

 fat. On June 24lh the milk with this quantity of acid still showed 3*3 per cent fat. 

 After this date the milk became lumpy and the fat adhered to the sides of the bottle, 

 so that a representative sample could not be taken up in a pipette. 



Conclusions. — A consideration of all these results will show that the exact amount 

 (i.e., within certain limits) of potassium bichromate to be added is of no moment. For 

 ordinary work from 3 to 7 grains (measured roughly on the point of a knife or in a small 



