PRESERVATIVES. 223 



the fat has already been extracted, and the amount of ash is 

 so small that it may be neglected without great error. 



Ash. In place of an estimation of the salt, an ash deter- 

 mination is often made, and the ash taken as salt. The results 

 are, however, always slightly above those obtained by titration, 

 as butter itself, to which no salt has been added, gives a small 

 ash ; preservatives, such as borax, will also swell the weight of 

 the ash. 



Preservatives. The preservatives most largely used in 

 butter consist of sodium borates ; sulphites and nitrates have 

 also been used, usually in conjunction with borates ; fluorides 

 are also employed ; formalin has been recommended, but appears 

 to be rarely used. These should be tested for in the aqueous 

 portion which sinks to the bottom on melting the butter at a 

 low temperature. The reaction with turmeric paper applied 

 to the liquid direct will show the presence of free boric acid. 

 If no reaction or a feeble one be obtained, a little of the liquid 

 may be acidified with very dilute hydrochloric acid, and tested 

 with turmeric paper. A pinkish -brown coloration, turned 

 greenish-black by dilute alkali, will show the presence of boric 

 acid in combination. It will usually be found, if the butter 

 is preserved in this way, that a reaction is obtained from the 

 liquid itself, and a much stronger one after acidifying. The 

 presence of sulphites may almost always be detected by the 

 smell of sulphurous acid developed on acidifying. Nitrates may 

 be found by the diphenylamine test. Monier-Williams tests 

 for fluorine by shaking 10 grammes, after melting, with ether 

 and 1 or 2 c.c. of water in a separating funnel. The aqueous 

 layer is run off into a test-tube, a few drops of hydrogen peroxide 

 added, and 1 c.c. of a solution containing 2 per cent, of titanium 

 sulphate in 10 per cent, sulphuric acid. The colour is compared 

 with a similar test made on pure butter (or margarine). In the 

 presence of fluorides the orange-yellow colour of the peroxide 

 will be partially or wholly discharged. While not quite character- 

 istic, it is a useful sorting test, and the presence of fluorides may 

 be proved by testing as in milk (p. 208). 



For the quantitative estimation of preservatives 50 grammes 

 of butter should be placed in a stoppered cylinder, 50 c.c. of 

 chloroform added, and the mixture warmed gently till perfect 

 mixture takes place. A quantity of water, which will, with that 

 present in the butter, make up 50 grammes, is added, and, after 

 shaking, the cylinder is put aside to allow the aqueous portion 

 to separate. Each cubic centimetre of the solution will contain 

 the preservative in 1 gramme of butter. 



For the estimation of boric acid Thompson's method is con- 

 venient (p. 110). As butter is practically free from phosphates, 



