236 PROCEEDINGS OF THE AMERICAN ACADEMY 



is a compound of the same class, — a conclusion which harmonizes with 

 the theory long held by many chemists, that hydrogen gas is the vapor 

 of a very volatile metal. This element, however, when combined with 

 palladium, is in a peculiarly active state, which sustains somewhat the 

 same relation to the familiar gas that ozone bears to ordinary oxygen. 

 Hence Graham distinguished this condition of hydrogen by the term 

 " Hydrogenium." Shortly before his death a medal was struck at the 

 Royal Mint from the hydrogen palladium alloy in honor of its discov- 

 ery ; but although this discovery attracted public attention chiefly on 

 account of the singular chemical relations of hydrogen, which it brought 

 so prominently to notice, it will be remembered in the history of science 

 rather as the beautiful termination of a life-long investigation, of which 

 the medal was the appropriate seal. 



Simultaneously with the experiments on gases, whose results we have 

 endeavored to present in the preceding pages, Graham carried forward 

 a parallel line of investigation of an allied class of phenomena, which 

 may be regarded as the manifestations of molecular motion in liquid 

 bodies. The phenomena of diffusion reappear in liquids, and Graham 

 carefully observed the times in which equal weights of various salts 

 dissolved in water diffused from an open-mouth bottle into a large vol- 

 ume of pure water, in which the bottle was immersed. He was not, 

 however, able to correlate the results of these experiments by such a 

 simple law as that which obtains with gases. It appeared, nevertheless, 

 that the rate of diffusion differs very greatly for the different soluble 

 salts, having some relation to the chemical composition of the salt which 

 he was unable to discover. But he found it possible to divide the 

 salts into groups ot equi-diffusive substances, and he showed that the 

 rates of diffusion of the several groups bear to one another simple nu- 

 merical ratios. 



More important results were obtained from the study of a class of 

 phenomena corresponding to the transpiration of gases through india- 

 rubber or metallic septa. These phenomena, as manifested in the trans- 

 fer of liquids and of salts in solution through bladder, or a similar mem- 

 brane, had previously been frequently studied under the names of exos- 

 mose and endosmose, but to Graham we owe the first satisfactory 

 explanation. As in the case of gases, he referred these effects to the 

 influence of chemical force, combination taking place on one sm-face of 

 the membrane, and the compound breaking up on the other, the differ- 

 ence depending, as in the previous instance, on the influence of mass. 



