364 DEDUCTIONS FROM ANALYSIS. 



A high percentage of water does not appear to have any effect 

 on the keeping qualities of the butter ; a large percentage of 

 solids not fat or curd seems to be distinctly inimical to its good 

 preservation. 



Speaking broadly, butters containing about 13| per cent, of 

 water have the best flavour. When the limits of 12 per cent, 

 on one hand, and 15 per cent, on the other, are passed, a distinct 

 falling-off in quality is usually found. To this rule, however, 

 exceptions are numerous. 



During very hot weather, if the butter is very soft when taken 

 out of the churn, there is a difficulty in working the water out 

 to a sufficient extent ; during very cold weather the butter may 

 be so hard that it cannot be efficiently worked. In both these 

 cases the water may somewhat exceed 16 per cent. An organism 

 has been described which produces changes in the cream which 

 prevent the water from being worked out, but it is fortunately 

 not frequently met with. 



Detection of Adulteration of Butter. The most useful and 

 rapid preliminary test is examination with the butyro-refracto- 

 meter. Any sample showing a refractive index of less than 

 46 at 35 C. is most probably genuine, but may, however, 

 contain coconut oil. The Reichert-Wollny process should next 

 be applied. Any sample requiring less than 20 c.c. for 5 grammes 

 may be taken as adulterated ; samples requiring more than 

 28 c.c. may be passed as genuine, though they cannot absolutely 

 be certified as free from adulteration. Any sample taking a 



N 

 volume of alkali between the limits given above must be 



examined further. The Polenske and Ave-Lallemant methods 

 should be employed, and if the results are suspicious the phyto- 

 steryl acetate test should be used. Baudouin's, Becchi's, and 

 Halphen's tests should be applied. A well-marked reaction 

 with any or all of them will furnish strong presumptive 

 evidence of the presence of margarine containing vegetable oils. 

 The soluble and insoluble fatty acids, saponification equivalent, 

 and especially the mean molecular weight of the insoluble fatty 

 acids should be determined. 



Coconut oil can be detected readily by the figures thus 

 obtained. A high Polenske figure indicates this adulterant, and 

 the amount can be calculated from the formula on p. 249. 

 The ratio between the Reichert-Wollny figure and the differ- 

 ence between the insoluble fatty acids and 95-5 is much 



K-W 



depressed ; in butter the ratio is about T - =3-5 (R-W = 



95-5 1 



Reichert-Wollny figure, and I = Insoluble fatty acids) ; while 

 coconut oil gives a value of approximately 0-75. The mean 



