The error (owing to imperfect removal) only amounts to one or two tenths per cent. of the butter taken. As the salt is accurately estimated afterwards, the loss falls on the curd. The salt may be determined by careful ignition of the filter and residue, the incombustible matter consisting almost wholly of common salt, while the curd is ascertained by loss of weight. This method is not to be recommended; for without great care some of the salt will be volatilised and lost, the error causing the amount of curd to appear excessive.
Ignition also renders any further examination of the curd an impossibility. A far preferable plan is to return the weighed curd and salt to the filter, and to wash them with cold water. The filtrate is made up to 100 c. c., and the salt is estimated in a half of it by titrating with decinormal nitrate of silver. The remaining portion of the solution can be employed for the estimation of sugar, if desired. This is effected by inverse titration with Fehling’s copper solution, in the same way as grape sugar. The estimation of sugar may sometimes be of interest, as a means of ascertaining whether the aqueous portion of the butter consisted of mere water or of serum of milk. In other words, the estimation of the sugar may furnish a means of ascertaining whether an excess of water in the butter is due to insufficient removal of the butter-milk, or to subsequent incorporation of water. Every 0·001 gram of milk sugar represents about 0·022 of average milk serum.
The residue insoluble in cold water usually consists almost wholly of casein. If, however, the butter has been adulterated with mashed potatoes, flour, or other starchy matters—said to be occasionally employed—they will be found here. The presence of starch in the residue will, of course, be readily indicated by treating it with hot water, and testing the cooled liquid with solution of iodine. By pressing out a small portion of the butter between two slips of glass, so as to obtain a thin film, and observing it under the microscope (or by observing the caseous residue after treatment with cold water), the nature of the starch may be ascertained.
The solution of the fatty matter in benzoline is evaporated at 100° C. till it no longer
decreases in weight. The average proportion of fatty matter in butter is about 85 per cent. If less than 80 per cent. the butter must be considered adulterated. It is evident that a careful estimation of the per-centage of fatty matter would often render separate estimations of the water, curd, and salt unnecessary; for unless the sum of the three latter constituents exceeded 20 per cent. the butter could not be considered as adulterated, unless by an admixture of other fats.
An easy and rapid method of estimating the fat in the undried butter is therefore a great desideratum; but unfortunately no satisfactory method is at present known. The indirect estimation of the fat, by subtracting the sum of the per-centages of water, curd, and salt from 100·00, ought to agree with the direct estimation of fat within 0·5 per cent., and the variation is often much less.
2. Dr Dupré adopts the following method: About 5 grams of the dry filtered butter fat are weighed in a small strong flask; 25 c. c. of a normal alcoholic soda solution are added; the flask is closed by means of a well-fitting cauotchouc stopper, firmly secured by a piece of canvas and string, and heated in a water-bath for about an hour. When cool the flask is opened, the contents—which are semi-solid—carefully liquefied by heat, and washed into a flask with hot water. This flask is now heated for some time on a water-bath to expel the alcohol, some more hot water is added, and 25 c. c. of diluted sulphuric acid somewhat stronger than the alkali used, are run in. The contents are allowed to cool, and the acid aqueous solution below the cake of fatty acids is passed through a filter. The fatty acids in the flask are washed by hot water in the manner recommended by Dr Muter, i.e., each time allowed to cool; all the washings are passed through a filter.
The author uses no cambric, but passes everything through paper. With care scarcely any of the fatty acid will find its way into the filter. After the washing with water is completed and the flask drained, he washes any fatty acid that may be on the filter into the flask by means of a mixture of alcohol and ether on a water bath, and finally dries the fatty acids in the flask at a temperature of 105° C. The drying can be done readily if the melted fat is now and then shaken briskly, so as to subdivide the water as much as possible. In this way the acids when once in the flask are not taken out until their weight has been taken, thus reducing the risk of loss to a minimum. Meanwhile the acidity of the aqueous filtrate and washings is estimated by decinormal soda solution. Subtracting from the amount required to the proportion necessary to neutralise the excess of acid added in decomposing the soap, the rest represents the soluble fatty acids contained in the butter taken, and on the assumption of its being butyric acid, we can, of course, calculate the amount of this acid present. When once the equivalent of the soluble acids present in butter is fairly determined, this, of course, will have to be substituted for that of butyric acid. The results thus obtained are very accurate, and the process is very simple in execution.
The author has satisfied himself by repeated experiments that the alkalinity of the alcoholic soda solution by itself is not altered by the process. The author places no reliance on the specific gravity test, as he finds that mutton dripping, and other fats likely to be used as adulterants of butter, may acquire a specific gravity above ·911 by being strongly and repeatedly heated. He thinks, however, that any sample of butter below ·911 may safely be pronounced adulterated.
In a subsequent note Dr Dupré states that he has effected the saponification, decomposition of the soap, and the washing and drying of the fatty acids at ordinary temperature, thus still further reducing the risk of breaking up the higher into lower acids. The saponification is readily effected by using a sufficiency of alcoholic soda. Between four and five grams of the dry butter fat were shaken up for several minutes with 100 c. c. of normal alcoholic soda. The butter soon dissolves, but after a time the solution gelatinises to a clear transparent mass. (The temperature of the laboratory at the time of these experiments ranged between 22° and 50°.) This jelly is now allowed to stand over night, during which time the smell of butyric ether, very strong at first, entirely disappears. In one of the experiments the alcohol was allowed to evaporate spontaneously before the acid was added; in the other (made with a different sample of butter) the soap was dissolved in about half a litre of water, and at once decomposed by the addition of hydrochloric acid.