157 



The ash results are for the most part very good and concordant, with the exception 

 of three, which were probably burned down hard without leaching. It is very evi- 

 dent that leaching with hot water after thorough charring is necessary in the presence 

 of sucrose. 



The sucrose results are probably as good as one could expect to obtain, but inasmuch 

 as this is not a normal but an added constituent of the milk it is best determined by 

 difference. 



The results on fat, one of the most important constituents of the milk, are far from 

 bein- satisfactory. The referee believes that these discrepancies are due to three 

 Fir-t. lack of experience with this kind of material. Second, a lack of 

 (I. -tail in irivini: tin -do ub le extraction method in Circular 32, and, third, faulty instruc- 

 tions in directing a 40 per cent solution to be used for this method. This solution, 

 as shown by the tables below, is too concentrated to get the best results when as large 

 an amount as 40 per cent cane sugar is present. This degree of concentration makes 

 such a thick layer ..f -u<;ar on the paper coil that quite a large proportion of the fat 

 on the paperafterthe first extraction, and then soaking in water causes a mechan- 

 ical lo- ..f fat when i hi- -ii-ar j s dissolved off. Such loss was proven by examination 

 of ih" -IL'ht .-cum ri.-iii'j n the water, which, under the microscope, showed the pres- 

 i'liles. When the work is carefully done, however, and dry ether is 

 this 1.*.- i- only small, amounting t<> one or two tenths of 1 percent. Our results 

 show that OH hiirh re-ult- can be obtained with 1 gram in a 10 or 20 per cent solution 

 in .1 -ini:le extraction of fourteen hours as with a more concentrated 40 per cent solu- 

 tion .uid a l.uMe extraction. The highest and most concordant results, however, 

 btained with a 10 or 20 percent solution and double extraction. 



Mr. (iei-1. r. tht- originator of the double-extraction method, stated, under date of 



January, 1908, that he had no changes to suggest from those given in his original 



i published in the Journal oHhe American Chemical Society in 1900 except 



that the time of each extraction should be extended to seven or eight hours instead 



of lour or five and that strictly dry ether probably is the best solvent on account of 



n-tain boiling point. In Mr. Geisler's paper he emphasizes the fact that the 



milk should be evenly distributed over the surface of the paper coil; also that the 



et h r r-hould be anhydrous to prevent the paper coil from becoming soggy. The 



method <>f drying on asbestos or paper in tubes is not so desirable as on strips of paper 



on account of the l,,m; heating required to dry out completely. 



The modified Habcock method, in the hands of men who have had much experi- 

 en< with it, usually gives very good results, but it is not to be considered as accurate 

 as the gravimetric, and as the reading is multiplied by three every error is increased 

 threefold. When the milk was fresh we were able to get very clear separations with 

 the copper sulphate in the centrifuge, and the sugar solution could be easily decanted 

 with. .111 lo>s of fat. the curd and fat remaining in a hard mass at the bottom of the 

 bottle, but in te.-ting some cans later, which had stood in the laboratory for three or 

 four month-, the -eparation was not so complete, and it was necessary to pass the solu- 

 tion through a tiller, washing the particles of curd back into the bottle to obtain all 

 the fat. and in several instances the results were low. This operation made the 

 method somewhat longer and more tedious. 



The results reported by the Gottlieb method are not very satisfactory and the ref- 

 : .11111,1 but little time to test it. Only two tests were made and they gave by 

 direct weight over 9 percent fat, but it was found that some proteid or foreign matter 

 was clinging to the bottom of the flasks after the fat was dissolved out with hot ben- 

 /ine. After weighing and deducing this weight from the original the results agreed 

 very closely with the gravimetric. The writer believes that by observing proper 

 precautions this method can be made very reliable, particularly for sweetened con- 

 densed milk. Small separatory funnels are more desirable than the long tubes recom- 

 mended in the method. 



