200 



EXPERIMENT STATION RECORD. 



the 3 other methods were: Babcock i 0.095 per cent, Gerber i 0.067 

 per cent, Thorner ± 0.07(> per cent. 



The probable error of determination of eacli of the 4 methods given 

 was calculated as follows : 



Error in different milk tests. 



Average error of a single determination. 

 Probable error 



Gravi- 

 metric. 



Per cent. 

 ±0.0344 

 ± .0232 



Babcock. 



Per cent. 

 ±0.0299 

 ± .0202 



Gerber. 



Thorner. 



Per cent. , Per cent. 

 ±0.0211 I ±0.0241 

 ± .0144 I ± .0162 



The author considers the difficulty of accurately reading oft" the 

 column of fat the main source of error in the rapid milk tests. — f. w. 



WOLL. 



Contribution to the determination of milk sugar in milk and 

 milk products, B. A. Van Ketel {Nederl. Tjdschr. Fharm., s, p. 15 J; 

 abs. in Cliem. Ceuthl., 1896, II, No. 2, p. 134). — The anthor states that 

 in the determination of milk sugar and other liquids it is necessary to 

 first remove the proteids and fat, which is best done by adding phenol 

 and lead acetate. To 50 cc. of milk are added 4 cc. of phenolum lique- 

 faetum and 10 cc. of a 10 per cent solution of lead acetate, the mixture 

 thoronghly shaken, filtered, and washed until the filtrate and wash 

 water aggregate 100 cc. If the milk sugar is to be determined by Feh- 

 ling's solution the lead is removed with a few drops of sodium sulphate 

 solution. 



Further notes on the detection of formalin, H. D. Eichmond and 

 L. K. BOSELEY [Analyst, 21 (ISi)O), April, i).'J2). — The authorsfouud that 

 the amnion iacal silver nitrate reaction, while delicate, was so general as 

 to be misleading, and that Schift's reagent was delicate but must be per- 

 formed in slightly acid solution. 



In making Hehuer's test with sulphuric acid the authors dilute the 

 milk with an equal volume of water and use 90 to 91 per cent sulphuric 

 acid. A violet ring is formed when formic aldehyde is present, and a 

 greenish tinge when none is present. Trillat's dimethylauilin reaction 

 and riochl's test are mentioned. — b. w. kilgore. 



The detection of formalin, A. Hehnek {Analyst, 21 {1896), April, p. 

 94). — The author does not consider SchifPs reagent a satisfactory test for 

 formic aldehyde, as it reacts in the same way with other aldehydes and 

 may give indication of the presence of formic aldehyde when none is 

 present on account of the oxidation of the sulphurous acid by the oxy- 

 gen in the liquid tested. It should only be used as a confirmatory test. 



The sulphuric acid test previously mentioned by the author,^ it is 

 said, will readily detect 1 part of formic aldehyde in 200,000 parts of 

 milk. If milk containing formic aldehyde is allowed to float upon 



1 Analyst, 20 (1895), July, p. 154. 



