402 PROFESSOR W. C. ROBERTS-AUSTEN ON THE DIFFUSION OF METALS. 
pressure is small. This effect appears to be connected with the dilution of the gold 
below a concentration of about 0‘005 gramme-atom per litre of lead, corresponding, 
at 450°, to an osmotic pressure of about one-third of an atmosphere. 
The results given in Table E show that the diffusion of metals is not increased 
nearly so rapidly by a rise in temperature of 50° as the ordinary aqueous diffusion of 
salts is. This is probably due to the fact that the resistance presented by metals to 
diffusing metallic molecules is not much reduced by heat, and it may be that the 
molecules of the diffusing substance are not so liable to disruption by a rise in 
temperature. 
It may be well to wait until more results have been obtained before attempting to 
deduce evidence as to the molecular condition of the metals composing these alloys, 
though, in this respect, metallic diffusion presents several advantages over the 
diffusion of salt solutions, the latter being very limited as regards the choice of a 
solvent and range of temperature at which experiments could be made, and, more¬ 
over, if a salt is dissociated into its ions its diffusion rates will be modified. The 
fact that when water is used, both the salt and the solvent are chemical compounds, 
renders their diffusivities less directly significant than those of metals, because, with 
the exception of gaseous elements, molten metals present the simplest possible case 
which can occur, as they represent the diffusion of one element into another. 
A few general deductions may, however, be drawn. 
It will be seen that gold diffuses more rapidly in bismuth and in tin than it does 
in the heavier metal lead. It has also been observed (though Table E does not 
embody the results) that platinum diffuses faster in bismuth than in lead. The 
diffusion of platinum and of gold is increased in about equal ratio by the sub¬ 
stitution of bismuth for lead as a solvent. On the other hand, platinum diffuses 
much more slowly in lead than gold does, although their atomic weights and their 
densities do not greatly differ. Rhodium, another metal of the platinum group, 
diffuses in lead nearly as fast as gold does, but if allowance be made for the smaller 
atomic weight, it will be found to agree fairly well with platinum. This would 
point to the conclusion that the platinum metals are molecularly more complex than 
either gold or silver, as a complex molecule exerts less osmotic pressure and diffuses 
more slowly than a comparatively simple one. 
The early workers on diffusion of salts used water as a solvent; Kawalki, # 
however, has recently given tables of diffusions of salts in both water and alcohol. 
He found that there is a fairly constant ratio between the diffusivities of salts in the 
two solvents. Experiments seem to show that this is also true of two metallic 
solvents such as lead and bismuth. 
Calculations of the Older Results .—The results obtained from the earlier experi¬ 
ments (which were conducted in U-tubes) have,as already stated, been omitted, because 
the great difference of temperature between the top and bottom of the tubes made 
* ' Wied. Ann vol. 52, 1894, p. 300. 
