116 MORRIS LOEB 



composition of the roseo- from the purpureo-cobalts. His re- 

 searches were the natural foundation of Werner's theory, 

 which has gained general recognition during the past fifteen 

 years. 



Analogy to cobalt-ammines appeared to exist in a com- 

 pound obtained by Fremy in 1844, by the action of ammoni- 

 um chloride upon potassium osmiate, and Gibbs proved this, 

 in a brief note published with Genth in 1857, by showing 

 that chlorine could be replaced by other negative radicals 

 without altering the Os : NHs ratio. In his final paper, in 

 1881, he named these compounds the osmyl-tetrammine 

 series. But he was deflected from the continued study of the 

 ammines of the platinum group, which he had evidently pro- 

 posed to himself in 1858, by the interest which the separa- 

 tion and complete characterization of these metals them- 

 selves had excited. Working chiefly with refractory California 

 ores, he found it necessary to develop new methods of attack, 

 and his work may well be placed by the side of that of 

 Claus and St. Clair-Deville. 



Meanwhile, the platino-ammines were fully studied by 

 other observers, and Gibbs rather devoted his attention to the 

 behavior of the platinum group to acids. In 1877 he found that 

 the oxides of these metals would unite with the tungstates, to 

 form the salts of complex acids, analogous to the silico-tung- 

 states of Marignac, and this was the starting-point for his 

 great researches on the complex acids, which virtually mo- 

 nopolized the remainder of his experimental activity. Begin- 

 ning with attempts at systematizing the straggling data on 

 silico-tungstates, phospho-tungstates, etc., recorded by other 

 observers, through the preparation of parallel compounds, he 

 was led to draw one element after another into the compli- 

 cated molecules that gather around the tungstic or molybdic 

 nucleus. With a large corps of assistants at his disposal, the 



