86 ANNALS OF THE MISSOURI BOTANICAL GARDEN 



[Vol. 6 



A study of the curve A, plotted from an average of many 

 determinations, will show that one-fourth of the carboxyl 

 has disappeared, and none of the phenol. Of all pos- 

 sible ways of combining tyrosin molecules, we may dis- 

 regard the possibility of linkage through the two amino 

 groups, as such compounds are notoriously unstable. Link- 

 age of the carboxyl with the amino group, as in polypeptid 

 formation, is unlikely, as the melanin is notoriously stable, 

 resisting hydrolysis as .very few proteins do. A distinct 

 break will be noted in the carboxyl portion of the curve, the 

 portion nearest neutrality dissociating at practically the same 

 point as tyrosin and very probably representing unattacked 

 tyrosin. It will be noted, too, that exactly the same amount 

 of the phenol group dissociating at the same point as the 

 phenol group of tyrosin is shown. Deducting these amounts 

 as unused tyrosin, we see that the ratio of the carboxyl to 

 the phenol group is 1:2. This would point to linkage not 

 being through the phenol group, and not wholly through the 

 carboxyls. This leaves as the only other possibility linkage 

 of the amino group with the carboxyl, although probably not 

 as polypeptids, for the amino groups disappear during the 

 process without showing up as ammonia. Von Furth and 

 Schneider ('01) have pointed out that skatol and indol are 

 given off when melanin is melted with pure sodium hydroxide. 

 This would indicate the possible formation of a heterocyclic 

 ring with the amino group, but we are hardly in a position to 

 discuss an exact formula for it at this time. 



The probability that both phenol groups belong to the 

 same ring of the compound, which would then be considered 

 related to compounds like homogentisic acid, reported so fre- 

 quently as resulting from the tyrosin metabolism of alkap- 

 tonurics, also from seeds (Bertel, '02, Czapek, '02, Czapek 

 and Bertel, '06), is remote. An inspection of the dissociation 

 constants of the di-hydroxy-benzene compounds shows that 

 the phenol groups in such cases tend to be dissociated farther 

 toward the alkaline side of neutrality, although it is barely 

 possible that the melanin molecule is so large that these 

 groups are not dissociated at such low hydrogen ion concen- 



