154 ROYAL SOCIETY OF CANADA 



red cells, the poetulaitied moeaic structure of Nathansohn may occur in 

 red bdood cells at least, 



Y\ie vi'iT fact that we are driven to postulate a mosaic membrane in 

 order to apply Overton's theory, indicates how inadequate the latter is in 

 explanation of oemotic phenomena. That lipoid substances can facilitate 

 the exchang-e between the exterior and the interior of a ceU may be ad- 

 mitted, but tliat it is tlie only factor or the most important factor is less 

 acceptable, and one must trust to a larger conception of osmosis in order 

 to be in a position to understand it« phenomena. 



This larger conception may be gleaned from tlie results olf the 

 studies of Kahlenberg.^ In order to understand rightly the significance 

 of his obiservations, one must, to a œrtiain extent, disregard the distinc- 

 tion between colloids and crystaUoids, which luas been held valid since 

 Graham's time. We must recognize that a, subsitanoe which acts as a 

 colloid to water is not necessarily a colloid to every other solvent. In- 

 deed, tJiere is evidence tJiat in the case of some of the typical colloidal 

 solutions in addition to the ultramicroscopic suspension particles present 

 there is a quantity of the same material in the solution side by side with 

 the particles. This Hardy hav^ shown in the case of agar and gelatine, 

 and the phenomena assodated with all other c-olloids, except those of the 

 inorganic class, seem to indicate tliat there are two phases in all such 

 suspensions, namely, the solid-water phase, which enters into the con- 

 stitution of the suspension particles, and the water-solid phase which 

 forms a true solution. ^Vhen the latter is eliminated there is brought 

 about a conversion of some of the solid-water phase into the water-solid 

 phase until equilibrium is established. It iis, therefore, not correct to 

 ppeak of colloids as wholly suspensions. 



On the other hand the ver^' fact that a membrane does not allow 

 the pa.«:sage of a so-called colloiid through it, is no evidence that the latter 

 is wholly a s\ispension, for, if the composition of the membrane is rubber 

 it will keep back even sodium chloride, that is, a crysitalloid. This shows 

 how artificial is the distinctlion which we obtain by using one kind of 

 membrane as a means of distinction between sus|>ension and solution. 

 If we wish to imderstand the phenomena of osmosis, we must discard all 

 the concepts which we have gained from the results obtained with the 

 ordinary dtialysing membrane formed of parchment. It would be just 

 A» reasonable to classify all substances into two groups in .their relation 

 to a rubber membrane, for neither rubber nor parchment membranes 

 exist in nature, and those membranes which do exist are not of the kind 

 that make a distinction between the so-called crystalloids and colloids. 



'Journ. of Physical CheTnistr>-, Vol. 10. p. 141, 1908. 



