314 Dr. J. H. Long on the 



quantity of water in vessels upon whose bottoms the phials 

 stood. By carefully removing these from time to time and 

 analyzing the contents of the vessels, the amount of substance 

 diffused was found. In these experiments Graham dispensed 

 with a membrane entirely, and thus avoided a source of error 

 which had robbed the investigations of his predecessors of 

 much of their scientific value. In the two papers read before 

 the Eoyal Society in 1850, he established among others the 

 following facts : — 



1. The velocity of diffusion is in general different for each 

 substance in solution. 



2. The amounts of salt diffused in the same time from solu- 

 tions of the same substance, but of different concentrations, 

 are very nearly proportional to these concentrations. 



3. The amount of salt diffused from a given solution in- 

 creases with the temperature of the solution. 



From his investigations Graham was led also to believe 

 that a relation exists between the boiling-points of solutions 

 and their rates of diffusion, and that no simple relation seems 

 to exist between the molecular weight of a salt and its rate of 

 diffusion when dissolved. These and other points of Graham's 

 work will be referred to again later, as will also the results 

 he obtained by means of the method of jar-diffusion, as he 

 termed it*. 



The chief characteristic of the earlier Graham experiments 

 (those of 1850 I mean) is that they were concerned, not so 

 much with an investigation of the theory of diffusion (that is, 

 with the purely physical side of the question), as with the rela- 

 tion of various substances to each other as regards diffusion, 

 an inquiry more chemical than physical in its nature. This 

 is not the case, however, with an investigation published by 

 Fick in 1855f. He regarded the work of Graham as possess- 

 ing a qualitative rather than a quantitative value, and remarked 

 that it gave no clue to the fundamental law of the phenomenon. 

 He showed that liquid diffusion may be compared to the con- 

 duction of heat — that is, that the spread of salt-particles 

 through water is in many respects analogous to the spread of 

 heat in a conducting body, and that similar formulas to those 

 established by Fourier for the latter case may be applied in 

 the former. 



Fick described several methods by means of which the law 

 of diffusion may be determined experimentally, one of which 

 is the following. A vertical glass cylinder open at both ends 

 was fitted tightly below into a vessel containing saturated salt- 



* Phil. Trans. 1861, p. 138. 



t Pogg. Ann. xciv. p. 59. [Also Phil. Mag. [4] vol. x. p. 30.] 



