52 



Professor Thomson referred to the experiments of Graham, 

 late Master of the Mint, and others, in which the phenomena of 

 diffusion of liquids and of gases had been extensively investigated. 

 The subject was one of great importance, not only on account of 

 its intimate bearing on some of the most profound questions of 

 physical research, and especially on questions as to the atomic 

 constitution of matter, but also on account of its practical applica- 

 tions in affording the new and valuable methods for chemical 

 analysis brought forward by Graham, and known as the processes 

 of diffusion and dialysis. If two fluids capable of mingling with 

 one another be placed in contact, a process of diffusion of each 

 into the other will take place, even when the contact is made 

 under conditions such as perfectly to avoid the production of cir- 

 culating currents. To prevent currents, the chief thing to be done 

 is to place the lighter fluid on the top of the heavier one. Thus, 

 if water be placed in the lower half of a vessel, and alcohol, which 

 is a lighter fluid, be placed gently, without commotion, on the top 

 of it, so as to occupy the upper half of the vessel, the two fluids 

 will diffuse into each other ; or, if sulphuric acid — that is, oil of 

 vitriol — be put into the bottom of a vessel, and water, which is much 

 lighter, be put on the top of it, diffusion will ensue. The same would 

 happen if syrup of sugar was put at the bottom, and water was placed 

 above it, or if any kind of salt dissolved in water was put below 

 and water above. Diffusion takes place also between two gases. 

 If the heavier one be placed below, and the lighter one above, the 

 difference of density will tend to make each remain where it has 

 been placed ; but still the two will diffuse into each other till they 

 are perfectly commingled. The diffusion of gases into one another 

 proceeds very quickly, but that of liquids very slowly. Substances 

 of a gelatinous character, of which glue is an example, and to 

 which Graham has given the name colloids, will diffuse only with 

 extreme slowness through one another, while salts and various 

 other crystallizable substances, called by Graham crystalloids, will 

 diffuse quickly through them, but at various rates. This fact forms 

 the basis of the process of dialysis, which consists in allowing dif- 

 fusion of mixed substances to occur through a gelatinous or colloid 



