[HARDING] THE "NINHYDRIN" REACTION 41 



chloride solution containing 0-05 mgs. of nitrogen was placed in a 

 test tube previously thoroughly washed with ammonia-free water. 

 This was then poured out of the test tube and the tube allowed to 

 drain. 0-2 c.c. of one per cent, "ninhydrin" was then placed in the 

 tube and 2-0 c.c. of Schuchardt's crystallisable glycerol, which prev- 

 iously had only given the faintest reaction by itself, added, and the 

 mixture boiled over a free flame for thirty seconds, thus duplicating 

 the conditions given by Halle, Loewenstein and Pribram. A strong 

 blue color developed. The high- temperature of the boiling glycerol 

 had evidently caused the dissociation of the trace of ammonium 

 chloride and thus caused a reaction which the glycerol, being a reducing 

 agent in high concentration, had assisted. The presence of oxygen 

 is not necessary for the production of the "glycerol color" as claimed 

 by Halle, Loewenstein and Pribram, nor do their experiments prove 

 such a fact. By carrying out the reaction in vacuo those authors 

 merely reduced the temperature of the reaction and so failed to de- 

 compose the nitrogenous impurity present in the glycerol. The in- 

 tensification of the blue color on the addition of alkali is due to the 

 action of the alkali on the hydrindantin which would naturally be 

 formed in excess. This intensification of the "glycerol color" we have 

 been easily able to repeat and we find that the intensified color is 

 destroyed by shaking with air, thus corresponding to the color given 

 by hydrindantin and potassium hydroxide. We therefore cannot 

 see in the work of Halle, Loewenstein and Pribram any new classes of 

 blue color reactions of triketohydrindene hydrate. 



CONCLUSIONS. 



In reviewing the investigations on the "ninhydrin" reaction, it is 

 clear that under accurately defined conditions of concentration and 

 temperature the test can become specific for amino acids, for it is these 

 bodies alone which respond readily and clearly in very dilute solution. 

 In testing for amino acids we would recommend that not more than one 

 tenth of a milligram of nitrogen be present in every c.c. of the liquid 

 to be tested and that one c.c. of it be heated in a boiling water bath 

 with one cc. of one per cent, solution of "ninhydrin" for a period of 

 fifteen minutes. The procedure which is sometimes adopted of 

 heating to dryness in an evaporating basin the liquid to be tested 

 with the reagent is certainly to be condemned. 



In the presence of pyridine, ammonium salts primary amines ( ?) 

 and amino acids, react strongly with triketohydrindene hydrate. 

 Thus the method of Harding and MacLean of determining colori- 

 metrically amino acid a-nitrogen is at present inapplicable to the 



