746 



THE AMERICAN NATURALIST [Vol. XLV 



( 1 ) crystallization; (5) 



its. 



Let us see what of real value these characters possess : 



1. If color is any criterion, all colored objects must con- 

 tain azo dyes, inasmuch as azo compounds occur in almost 

 every possible shade of color, 



2. If the data which Tower gives on page 11 (1903) be 

 referred to we find 



f Black 



These are not the characteristics of the azo compounds, 

 for azo compounds are soluble in alcohol and in many 

 cases are very soluble in water. The "amidoazo" com- 

 pounds are soluble in weak acids and the oxyazo com- 

 pounds are very soluble in weak alkalis. The characters 

 which Tower gives are, however, exactly those char- 

 acters which are possessed by the melanins. 



3. Tower has made a great mistake when he states 

 that azo compounds are reduced to colorless bodies and 

 then reoxidized to colored diazo, oxyazo or amidoazo 

 compounds by mild oxidizing agents. When azo com- 

 pounds are reduced by tin and hydrochloric acid the azo 

 union is broken, each nitrogen atom yielding an amino 

 group, and the colored compound can not be regenerated 

 by mild oxidation, but the original diazotization and 

 coupling must be repeated before a color results. For 

 example we reduce Congo red and we get benzidine and 

 "naphthionic acid" neither of which gives a color on 

 mild oxidation. (See Mulliken (1910), p. 26, footnote.) 



CH.— N==N— C lD H,(SO,H) (NH 2 ) 

 1 +2H, 

 C,H 4 — N=N — C 10 H 8 ( SO a H ) (NH 2 ) 



= ^ + 2C 10 H S ( SO,H ) ( NH 2 ] 



C„H— NH 2 

 } 8 H 4 — NH 2 H 



There are, to be sure, dye stuffs (for example, indigo), 

 which do reduce to colorless compounds, or leuco bases, 



