CHEMISTRY. 373 



tlie rod used, wliicli is very important. Trials with liO samples of natural 

 and artiticial butter are reported, using rods of different sizes. Six 

 grams of tiitered fat (liruUeuses cc.) was oxidized with 8 drops of fum 

 ing nitric acid, and the disli tlieu heated for an hour at 21^ C. This was 

 done by placing the dish upou water kept at this temperature and sub- 

 merging it as soon as the contents had hardened. The author found 

 the temperature at which this heating was done to be a very imi)ortaut 

 factor, since a variation of 1" or even less affected the results materially. 



From the trials reported it appears that a rod 12 mm. in diameter 

 gives results higher than BruUe's, and one 8 mm. lower results, while 

 with one 10 mm. the results more nearly approached Brulle's. The 

 results were not encouraging to the author, and he is inclined to doubt 

 that 10 per cent of margarin can be detected with certainty, as claimed 

 by Brulle. 



Trials of Kbnig and Hart's method of butter analysis, C. T. 

 MoRNER {Upsala lillcareJ\f6rhau<U.^ 1891 ; abs. in Kord. Mejeri Ti(lii.,9 

 {1891), pp. 435, 120). — The author investigated the applicability of 

 Konig and Hart's method of butter analysis' for Swedish butters. The 

 limits for pure butter fat found by Konig and Hart were 200 to 238, 

 average 221 mg., while for oleomargarin the ligures were 22, beef suet 

 3(5, leaf lard 17, mutton suet 5. The author examined 20 samples of 

 genuine Swedish butter, manufactured during the months of March 

 and April. The following baryta numbers were found : Minimum 180, 

 maximum 228, average 200.7 mg. — f. w. woll. 



A comparison of methods for the determination of starch, W. 

 E. Stone {Jour. Amer. Chcm. >SV>6\, 10 (1891), No. 11, pp. 7;2G-73o). — It 

 is urged at the outset that instead of determining the carbohydrates 

 (except fiber) in feeding stuffs by difference and grouping them under 

 the head of nitrogen-free extract, "a food analysis . . . should include 

 deternuuations of the pentosans, sugars, and starches separately." In 

 the comparison of methods pure starch, dried potato, wheat flour, corn 

 meal, wheat bran, hay, wheat middlings, cotton-seed meal, and a mix- 

 ture of starch, sugar, and dextrin were used. The methods tried were: 

 (1) Inversion by hydrochloric acid and titration with Fehling's solution 

 (Sachsse); (2) inversion with nitric acid and polarization (Guichard); 

 (3) a modification of the i)receding, first dissolving the starch with 

 oxalic acid and then inverting; (4) inversion with salicylic acid and 

 polarization (Baudry) ; and (5) precipitation with barium hydrate and 

 determination of the excess of the latter by titration with standard 

 acid (von Asboth). 



On pure starch these methods gave satisfactorily concordant results, 

 but on the other materials the results by different methods were " more 

 or less discordant, and in some cases quite unexplainable." 



"For instance, the hay and cotton-seed meal when boiled with water do not give 

 iodin reaction for starch, yet each of these special methods for the determination of 



' Ztschr, analyt. Chem., 30, p. 292. 

 10101— No. 5 3 



