130 THE ANALYST. TESTING SKIM-MILK BY THE LACTOCRITE. BY HARALD FABEB. Read at the Meeting, April, 1887. IN a previous paper on the Lactocrite I tried to prove how very exact results may be obtained when determining the amount of fat in whole milk by means of this apparatus, and how easilyit is worked. I mentioned that the results obtained, when testing skim- milk, would fall too low, and by two examples I showed that the difference is about 0.1 per cent., when the skim-milk contains about 1.0 per cent. of fat; I shall now give the results of some experiments on milk containing still less fat, viz., skim-milk from very close hand skimming and from cream separators. I mentioned that the differences between the Lactocrite and analysis are found to be about 0.2 per cent. for separated milk, the result of a series of experiments carried out in Sweden being :- Chem.Analysis. Lactocrite Difference. 0.50 0.30 0-20 0.50 0.29 0.21 0.43 0.20 0.23 0.36 0.10 0.26 0.34 0.13 0.22 0.26 0.05 0.21 0.1 9 0.00 0.1 9 0.19 0.00 0.19 My own experience showed me that sometimes the difference was still larger, some- times also less, making the use of the Lactocrite for testingseparated milk rather doubt- ful. Besides, when the Lactocrite shows no fat, it is impossible to know how small an amount has been left by the skimming. In one case I found, in testing a milk contain- ing 0.31 per cent. of fat, that the Lactocrite showed only a trace which could not beTHE ANALYST. 131 measured, so that separated milk with less than 0.3 per cent. of fat can not be tested at all.The reason why the very small amounts of fat, so to speak, disappear, I cannot give. It might be imagined that the boiling with acid would attack the small globules, and in separated milk only the very smallest globules are left; but if that was the case it would seem very strange that so exact results are obtained with whole milk, as the size of the globules varies much in different samples of milk, and especially in milk from Merent breeds. Six sclmples of the same separated milk, containing 0.49 per cent. of fat, were tested in the Lactocrite, three of them being boiled with acid for only four minutes, the other three for eight minutes. The following experiment seems also to controvert this theory :- The results obtained were as follows :- Boiled 4 min S min.Per cent. of fat found 0.2 0.2 19 0.2 0.25 ,,*a 0.25 0.25 If the acid did attack the fat, less fat, ought to be left after the longer boiling with acid, which was not the case. It may seem of little interest to be able to estimate such small amounts of fat with very great m u w , but in butter dairies it will be found of the greatest importance. Farmers look for help and information from agricultural chemists, and chemical analysis can hardly benefit the dairy farmers more than by teaching them how t o skim the milk thoroughly with a separator, if they go in for buttermaking; as by a well-worked good separator the milk may be made to produce 20 per cent. more butter than by the old system, in some cases even considerably more. But this is only done when the sepamtar is made to skim the milk thoroughly, which alone can be controlled by a test of the amount of fat left in the skim-milk.Seeing how great a service the Lactocrite might yield the buttermaker, if it could be made to test the sepamted milk with sufficient exactness and in a simple way, I tried to alter the method of working, and finally found that by using it in the same way as the (( Control Centrifuge ” of Professor Fjord (see ANALYST of January) very satisfac- tory results could be obtained. The method I use is still easier than the ordinary working of the Lactocrite. All t h t is required is to fill the test-boxes with the skim-milk until it reaches about half- way up on the graduated glass, and then place them in the warm disc and make this revolve.It take a longer time for these very small globules to rise in milk not treated with acid, so that separating for twenty or thirty minutes will be necessary. After that time a column of crmm is found in the graduated glass, which it is very easy to read off. This cream, however, does not directly give any information of the amount of butterfat in the skim-milk; a calculation is necessary. I have found that multiplying the degrees of cream [each degree corresponding to 0.1 per cent. of fat when testing whole milk] by the factor 0.03 gives the per centage of fat in the skim- milk with remarkable correctness. Below are the results of fourteen samples, resulting from eleven different sepmtings of different milks, separated by dXerent separators :-132 THE ANALYST.Per cent. of fat by Analysis. 0.08 1.01 0.13 0.22 0.24 0.25 0.28 0.3 1 0.31 0.32 0.35 0.51 0.55 0.76 Degree of cream by Lactocrite. 21 4 7 8 9 11 13 15 168 l2 79 9$ IS$ 22h Degree of cream x 0.03. Difference. 0.075 0.005 0.03 0.08 0.1 2 0.01 0.21 0.01 0.225 0.015 0.24 0.01 0.27 0.01 0.285 0.025 0.33 0.02 0.39 0.07 0.45 0.10 0.495 0.045 0.555 0.005 0.675 0.055 Average 0.007 Numeric average 0.031, Only in four cases is the difference larger than 0.05, which I consider the difference to bs guaranteed by chemical analysis, if this shall claim to be exact. It is, perhaps, not so very strange that tolerably uniform results should in this way be obtained by a test, based on the ‘‘ cream,” which is otherwise not a very uniform product. For, in the first place, the centrifugal force, when, acting strongly for half an hour, is well able to cause all the globules to rise, and pack them closely, and secondly, these globules, including (in separated milk) only the very smallest, are of much more uniform size than the globules in whole milk. At any rate the number of requirements seems sufficient to prove that, worked in this way, the Lactocrite may be used as a practically reliable test for skim-milk. By means of the Lactocrite, therefore, a dairy farmer may examine the working of a separator before buying it, and he may at any time test whether his milk is skimmed so closely as to give him the largest possible yield of butter, since he will be able to perform in a short t h e a sufficiently correct analysis of his skim-milk at very little cost. By this machine, therefore, may be established a control which hitherto h&s been largely neglected. Conclusion of the Society’s Proceedinp.