September 23, 1887.] 



SCIENCE. 



155 



In the case of sample No. 33, which curdled even when slightly 

 warmed with water, and would not mix well with cold water, the 

 portions needed for each determination were separately weighed 

 out. 



Total Solids. — Of the diluted milk a measured volume was 

 diluted again with an equal volume of water, so that 5 cubic cen- 

 timetres corresponded to I gram of the condensed milk, and then 

 5 cubic centimetres were dried in a flat-bottomed platinum dish (40 

 millimetres in diameter at the bottom), at first on the water-bath, 

 then in an air-bath at 100° to 105° C, until the loss of weight 

 after drjing half an hour was less than 2 milligrams. Compara- 

 tive e.xperiments showed that under the above conditions the dry- 

 ing was as thorough as if the milk had been first coagulated with 

 acetic acid, jvhile the method was more convenient. At first 2 

 grams of milk were used, but the result was the same, whila the 

 drying was far more tedious. 



Duplicate determinations were made. The greatest difference 

 was 0.29 per cent ; usually it was much less ; occasionally the re- 

 sults were identical. 



Ash. — The dried milk was ignited in the same dish at a scarcely 

 visible red heat, until no black carbonaceous portions were left. 

 In one case the chlorine in the ash was determined by Volhard's 

 volumetric method, and found to be 9.52 per cent. 



Fat. — Rather thick white filter-paper was thoroughly extracted 

 with ether in a Soxhlet apparatus, and 5 cubic centimetres of the 

 diluted milk (equal to 2 grams of the condensed milk) dropped on 

 a nearly square strip of this paper large enough to conveniently soak 

 up the milk. To avoid the formation of candied spots, the milk was 

 uniformly spread over the paper by brushing with a small, narrow 

 strip of the same sort of paper. After drying in the air, the paper 

 was rolled to a loose cylinder, and dried in an air-bath at 100*^ C. 

 for about an hour and a half. The fat was then extracted with 

 ether for two hours in a Soxhlet apparatus ; and a second extrac- 

 tion was made, lasting from an hour to an hour and a half longer. 

 The second extraction usually yielded less than 4 milligrams more 

 of fat, and often none at all. The fat determination was made in 

 duplicate. The greatest difference was 0.2 per cent ; usually only 

 a few hundredths of a per cent. 



At first, extraction after drying with sand was employed, but 

 comparative tests showed that the paper method yielded better re- 

 sults in less than half the time. 



It is a very difficult matter to extract all of the fat from a dried 

 condensed-milk residue in any other way than by using paper, es- 

 sentially Adams's method. Blotting-paper would not be as good 

 as the thinner filter-paper, because there is so much cane-sugar 

 present in some of the milks. 



Caserne and Albumen. — Ritthausen's method was followed, es- 

 sentially as described by Dietzsch (Nahruiigsmittel iind Getrdnke, 

 Zurich, 1884); 5 cubic centimetres of the diluted milk, equal to 2 

 grams of the condensed milk, being further diluted with water to 

 40 cubic centimetres, and then treated with enough of a solution of 

 copper sulphate {6,35 grams in 100 cubic centimetres of water) to 

 insure quick separation of the coagulated albumen after stirring, 1 5 

 drops being added in almost every case. Then enough of a 5-per- 

 cent solution of caustic potash was added to render the mixture 

 nearly neutral to blue litmus-paper ; an excess of the potash being 

 avoided, as this would hold some of the caseine in solution and 

 render the filtrate turbid. In most cases 5 drops was found to be 

 a proper quantity. After settling clear, the fluid was decanted into 

 a weighed filter 11 centimetres in diameter, previously dried at 100° 

 C. The precipitate remaining in the beaker was stirred up with 

 20 to 30 cubic centimetres more of water, and finally the whole of 

 the precipitate was brought on the filter, the washing being con- 

 tinued until 100 cubic centimetres of liquid had passed through the 

 filter. This filtrate was preserved for the milk-sugar determina- 

 tion. The precipitate and filter were weighed together, after drj'- 

 ing at 100" C, until the loss of weight after drying half an hour did 

 not exceed i milligram. The filter and precipitate were next in- 

 cinerated in a porcelain crucible, and the weight of the residue de- 

 ducted from the weight of the dry precipitate : the difference was 

 the weight of the albumen (including caseine) and fat ; and after 

 deducting the weight of the fat the percentage of the albumen 

 (caseine) was calculated. S. W. Parr (Amer. Chem. Joiirn., vii. 



p. 246) has shown that the results by Ritthausen's method are 

 " neariy, if not quite correct." It is probably the best method for 

 condensed-milk analysis. 



Milk-Sugar. — This was determined in most cases by treating 

 25 cubic centimetres of the filtrate just mentioned with 15 cubic 

 centimetres of Fehling's solution (68 grams of caustic soda and 187 

 grams of tartrate of potassium and sodium in 500 cubic centime- 

 tres of water ; 34.64 grams of copper sulphate in 500 of water ; 

 the two solutions being mixed at the time of using them combined 

 as Fehling's solution) in a porcelain dish resting on wire gauze over 

 a Bunsen burner. The contents of the dish were rapidly brought 

 to boiling, and then boiled for four minutes, after which the liquid 

 was filtered through a filter 6 centimetres in diameter, and the pre- 

 cipitated suboxide of copper was washed, chiefly by decantation in 

 the dish, with about 40 cubic centimetres of water, which was also 

 passed through the filter. As little of the precipitate as possible 

 was brought on the filter. The filter was then dried and burned, 

 the residue dissolved in a little nitric acid, this acid poured into the 

 dish to dissolve the suboxide of copper, and the solution evaporated 

 with a little sulphuric acid until all nitrous fumes were expelled. 

 The solution was then diluted with water and the copper deposited 

 electrolytically in a small platinum dish. Rodewald and ToUens 

 (Berichte, xi. 2076) have shown, that, when milk-sugar is treated 

 with Fehling's solution as above described, the weight of copper 

 multiplied by 0.763 equals the weight of milk-sugar present. They 

 worked with asbestos filters, and certainly the paper filter does re- 

 tain a very little of the copper in the Fehling's solution ; but a blank 

 test showed that the filter used in these examinations of condensed 

 milk reti.ined only 0.0009 °f a gram of copper ; so that the above 

 factor, 0.763, was used in calculating the results. 



Cane-Sicgar was obtained in every instance by deducting the 

 weight of the remaining solids (milk solids) from the total solid resi- 

 due of the dried milk. 



General Observations. 



The following tables, I. and II., give the results of analysis of the 

 milks according to the method just described. The last column, 

 headed ' Times condensed,' indicates the number of volumes of the 

 original milk that were condensed to one volume. The figures in 

 this column are obtained by dividing the figures representing the 

 percentage of milk solids by 12.5, which is assumed as the average 

 percentage of solids in the original milks. Hehner (Analyst, iv. p. 

 44) has calculated the condensation by dividing the percentage of 

 milk solids not fat by 9.3, the assumed percentage of such solids in 

 cow's milk. The figures thus obtained would differ in the case of 

 our samples by less than 0.2 per cent in any instance. The per- 

 centage of fat in the original milk is naturally obtained by dividing 

 the figures representing the percentage of fat in the condensed milk 

 by those representing the condensation. Although No. 2i cannot 

 be regarded as made from a milk originally very rich in fat, yet 

 there is nothing to indicate that any of the samples were made from 

 skimmed milk. 



A word or two seems proper with reference to the proportions 

 of fat and caseine. The average percentage of caseine in cow's 

 milk is variously given by different authorities, but is probably 

 about 0.4 per cent greater than that of fat, as the writer has calcu- 

 lated from figures representing a very large number of analyses 

 given in the ' First Report of the New York State Dairy Commis- 

 sioner,' p. 58. It was stated, moreover, by Wigner {A?ialyst, iv. p. 

 48) that some of the caseine was decomposed during the conden- 

 sation of milk with sugar, and it would therefore seem that the per- 

 centage of caseine in average condensed milk should at all events 

 not greatly exceed that of the fat. In the writer's analyses it falls 

 slightly below. The caseine and albumen reported in many of the 

 analyses quoted in the New York daiiy commissioner's second re- 

 port, pp. 152-154, are very largely in excess of the fat; exceeding 

 it, for instance, in four out of many cases by the following figures 

 respectively : 8.1, 6.07, 8.24, 7.72 per cent. 



There seems to be but one explanation of such a result, and that 

 is, that the condensed milks were made from partially skimmed 

 milk, without regard to the fact that the percentage of fat actually 

 present in the condensed milk may not be below the average. 



Hehner {loc. cit.) found two samples of e certain brand -of con- 



