11 
slowly turns brown. By this test 1 part of formaldehyde in 2o0.0<X» 
parts of milk is readily detected before the milk sour<. After souring, 
the limit of delicacy proves to be about 1 part in 50.000. Various 
aldehydes when introduced into milk give color reactions under the 
above treatment, but formaldehyde alone gives the violet coloration, 
which is perfectly distinguishable and unmistakable. 
According to Leach this seems to be a most satisfactory test. 
The following quantitative methods are generally considered to be 
the best : 
Romijn's (17) cyanide method. The formaldehyde solution is added 
to an excess of potassium cyanide solution, and this mixture is at once 
poured into an excess of silver nitrate solution, which contains enough 
nitric acid to maintain an acid reaction throughout. The formalde- 
hyde combines quantitatively with the cyanide to form the compound 
CH'O.KCA, the excess of the cyanide reacts with the silver nitrate to 
form silver cyanide, and the excess of the silver nitrate left in the 
solution is titrated with ammonium sulphocyanate. 
Romijn's (IS) iodometric method. To the formaldehyde solution are 
added '25 c. c. of decinormal iodine solution, followed by enough strong 
sodium hydroxide solution to make the mixture pale yellow. Allow 
to stand ten minutes (some authorities say longer), and then add con- 
centrated hydrochloric acid until the reaction is sKghtly acid. Titrate 
for free iodine with sodium thiosulphate. Each two atoms of iodine 
consumed represents one molecule of formaldehyde. 
Blank and Einkenbeiner's (19). To a carefully weighed amount of 
the formaldehyde solution are added normal sodium hydroxide solu 
tion. followed at once, drop by drop, by neutral hydrogen peroxide. 
When the reaction is completed, the excess of alkali is titrated with 
normal sulphuric acid. Modifications suggested by .Schoorl (22). 
Legler‘s (20). To the formaldehyde solution is added an excess of 
standard ammonia solution. In the gravimetric method the resulting 
hexamethylenetetramine is evaporated to dryness and weighed. In 
the volumetric the excess of ammonia is titrated with standard acid. 
Eschweiler (21) pointed out that when litmus or phenolphthalein is the 
indicator the hexamethylenetetramine itself is neutral: hence six 
molecules of formaldehyde need four molecules of ammonia for neu- 
tralization: but with methyl-orange, cochineal. Congo-red. or tropeo- 
lin the monoacid compound of hexamethylenetetramine is neutral: 
hence six molecules of formaldehyde need only three of ammonia for 
neutralization. 
The qualitative tests of Hehner (10). Lebbin (15). and Rimini (IH) 
are generally recommended as being about the best. It is generally 
best to distill the substance and to apply the tests to the distillate. It 
must be borne in mind that the presence of a trace of formaldehyde 
in food substances or elsewhere may not be the result of intentional 
