PROCEEDINGS OF SECTION B. 349 



however, be given to the preparation of the reagent (" Allen's Commercial 

 Organic Analysis," vol. i., third edition, p. 219). The test is also of 

 greater delicacy than found by Hehner, distinct reactions being ob- 

 tained with a solution containing one part formalin in 2,000,000, and 

 with the distillate from milk containing one part in 1,000,000. Dis- 

 tillates from sour milks may give a faint reaction in absence of formalin, 

 but the coloration is dispelled on acidulation with a few drops of dilute 

 sulphuric acid. This test is serviceable for the quantitative estimation 

 of formaldehyde. Fifty ccs. of milk are acidified with Ice. dilute 

 sulphuric acid (1 : 3), and on distillation from fresh milk approximately 

 50 per rent, of the contained formaldehyde is collected in the first 

 25ccs. of distillate. The distillate is made up to SOccs. in a Nessler 

 cylinder, and Ice. of SchifE's reagent added. After half an hour's stand- 

 ing, comparison of color is made ynih. formaldehyde solutions of known 

 strength similarly treated. 



Fury's Test (Abstr. ii., J.C.S., 1904, p. 687) is preferred to Leach's 

 test. The test is modified by the use of dilute sulphuric acid (1 : 2) in 

 lieu of acid (1 : 1) ; the boiling is done in a casserole, and the observa- 

 tion of color in test tubes. Pure milks give a delicate flesh tint ; milks 

 containing one part formaldehyde in 500,000 showing a distinct pui'ple 

 shade. With sour milks the reaction is less sharp, and of a brownish 

 tint. 



Lehbin's Resorcinol Test (J.C.S., Abstr., 1897, vol. ii., p. 606) 

 detects one part formaldehyde in 300,000 (c./., Pilhashy, J. Amer. C.S., 

 1900, p. 133). With more" than 1 : 100,000 the color is permanent for 

 a few hours. The reaction is not given by other aldehydes. 



Phloroglucin Test (L. Vanino, J.S.C., i., 1899, p. 403) is sensitive 

 to 1 : 100,000, the color being transient. 



Nessler' s Reagent gives a sharp and characteristic reaction, but the 

 test is not suitable for colorimetric estimation. 



Phenifhdyrazine Test (Arnold and Mentzel, " Analyst," 1902, 

 p. 227) can detect 1 : 10,000 in milk, even after keeping four weeks. 



The subjects of the disappearance of formaldehyde from milk on 

 keeping and the recovery of the aldehyde from milk by distillation 

 have been investigated by Leonard and Smith (" Analyst," 1897, p. 5) 

 and B. H. Smith (J. Amer. C.S., 1903, p. 1036). R. Steenegger (" Land- 

 ■svirthechaftliches Jahrbuch der Schweiz," 1905) shows that combination 

 of the aldehyde with the Amino groups of the milk proteids occurs. 



Conclusions. — (1) For the qualitative detection of formaldehyde in 

 fresh milk, Eury's or Bonnet's morphine test may be applied to the 

 samples direct, and tests with Schiff's reagent, resorcin, phloroglucin, 

 gallic acid, and phenylhydrazine to the distillate. (2) The qualitative 

 detection of formaldehyde in old and sour milks presents no difficulties. 



(3) For the estimation of minute traces of formaldehyde in fresh milk, 

 colorimetric methods — either Bonnet's, with the original samples, or 

 Schiff's reagent, with the distillate — may be used with advantage. 



(4) In milks kept any length of time these methods will indicate less 

 quantity of formaldehyde than originally present, depending mainly 

 on the length of keeping. 



