[Vol. 6 



76 



ANNALS OF THE MISSOURI BOTANICAL GARDEN 



The liquid was decanted, the permutit covered with ammonia- 

 free water, and this decanted and the action repeated. Then 

 5 cc. 10 per cent NaOH were added, the flasks shaken, about 

 60 cc. of water poured in, and the flasks shaken again to mix 

 thoroughly, thus preventing the alkali being too concentrated 



for the Nessler reagent. Ten cubic centimeters of the Ness- 

 ler reagent, prepared by the method of Folin and Denis ( '16), 

 were added, the flasks shaken, allowed to stand 15 minutes and 

 made up to the mark, then read in a Kober ('17) colorimeter. 

 An attempt was made to have the standard contain the same 

 amount of ammonia as the solutions analyzed, but the other 



ed in such extremely dilute solutions were so 

 great that exactness in this respect was not attempted. Arti- 

 ficial light was used, also the long cups, giving approximately 

 80 mm. working depth. A rheostat, by which the light could 

 be dimmed, was found to be an advantage in working with 

 such dilute solutions, as the eye distinguishes more readily 

 between solutions which appear dark-colored than between 

 very light ones. The same applies to the micro-titration 



method about to be described. 



arowos. — The formol 



errors 



Determination of carboxyl and phenol 

 titration method as developed by Sorensen ('07) and by 

 Sorensen and Jessen-Hansen ('08) and extensively used since 

 for the determination of the ca 



tried 



th no better success th 



boxyl of amino-acids was 

 a Sorensen obtained with 



tyrosin, and after attempting several modifications it was 

 abandoned. Upon the addition of the formalin solution 

 (prepared by bubbling the formaldehyde given off from 

 paraformaldehyde during heating, into double distilled 

 water) the indicator, thymolsulphonphthalein, if present in 

 the usual amount (5 drops), was quickly decolorized, espe- 

 cially in the presence of the first drop of alkali added. Upon 

 the addition of more indicator, the solution finally turned a 

 dirty, purplish red, very different from that obtained with 

 this indicator at any known hydrogen ion concentration, so 

 that comparison was impossible. The pigment present did 

 not seem to have any color changes, at least in the region 

 in which one must work with tyrosin, thus making thymol- 



