the osmotic pressure, which is about 00034 at 18^ and of Pi and P2, which for the common salts is about 0*024. Accord- per cent, per degree at 18°, and this has been experimentally confirmed.
The friction coefficient of non-conductors cannot be determined by an electrical method; but its value can be estimated from the diffusion constant. From Graham's results, Nernst has calculated the following values : —
��DlffnsiDg Btttotance.
�Formul ».
CO(NH,),
�Namber , MoIecnUr of Atoms. ' weight.
�Coefficient of fHcilon.
�Urea ....
�2,500x100 kg.
�Chloral hydrate
�CC1,CH(0H),
�3joUtf ,j
�Mannitol . . .
�C«H,,0,
�Cane sugar . .
�CjjH-jO,,
n(C,lJ,oO,)
�Gam arable . .
�very great
�Tannic acid . .
� �f)
�Egg albumen .
�Caramel . . .
� �'>
��The friction evidently increases with the molecular weight of the substance.
According to a calculation made by Euler (^7), the friction of a gram-molecule is approximately proportional to the square root of the molecular weight of the substance examined. This gives very good values for substances of known molecular weight, provided that the molecular volume does not vary too much. If we apply this method to calculate the molecular weight, M, of the four colloids examined by Graham, we find the following values : —
Caramel 13,200.
These results are particularly interesting because we have no method free from objection for determining the
�� �
