226 
PROFESSOR GRAHAM ON OSMOTIC FORCE. 
experiment by the action of diffusion. When pernitrate of iron, a salt of this class, 
is placed in the osmometer, a rapid decomposition of the salt occurs in the mem- 
brane ; the nitric acid escaping - , from its high diffusibility, into the water of the jar, 
and leaving a basic salt on the inner surface of the membrane. Here then, as with 
the preceding class of osmotic bodies, the osmose of the water is towards the basic 
side of the membrane. 
But the most curious circumstance, with reference to this empirical generalization, 
is observed in the magnesian class of salts. The barytic subdivision of this class, 
including all the soluble salts of baryta, strontia and lime, appear to be entirely 
unosmotic, or they oscillate between a small positive and small negative osmose. 
Such salts are neutral in their reaction, and further, have no disposition whatever 
to form subsalts. The salts of the earth, magnesia itself, offer the same characters. 
But in the salts of certain other oxides of the magnesian group an intensely osmotic 
character is developed, particularly in the salts of copper, protoxide of lead and 
protoxide of tin, with the exception of the soluble sulphates of these bases. Now 
those named are the members of the magnesian class most apt to break up into free 
acid and a basic salt. Like the aluminous salts, therefore, they are capable of 
investing the inner surface of the membrane with basicity, the necessary condition 
of positive osmose. Nitrate of uranium does not require to form a subsalt, as it is 
already constitutionally basic. The osmotic peculiarity of metaphosphoric acid, 
formerly referred to, also harmonizes with the same view. 
Neutral monobasic salts of the alkaline metals, such as the chlorides of potassium 
and sodium and the nitrates of potash, soda and silver, which possess a strict and 
unalterable neutrality, appear to have little or no true osmotic action. The salts 
named, together with the neutral magnesian sulphates and certain neutral organic 
substances, such as alcohol and sugar, give occasion, it is true, to an increase in the 
fluid of the osmometer, but only to the moderate extent which the exchange of 
diffusion-volumes might be supposed to produce. The comparative diffusibility of 
all these substances is well known, with the exception, unfortunately, of that of water 
itself, which 1 could only deduce by an indirect method in my previous inquiries re- 
specting liquid diffusion. As salts generally appeared to diffuse in water four times 
more rapidly than they did in alcohol, the diffusibility of water was then assumed as 
probably four times greater than that of alcohol, and consequently five or six times 
greater than that of sugar or sulphate of magnesia. Diffusion is thus made to 
account for the substances last named being replaced in the osmometer by five or 
six times their weight of Avater. This “diffusion-osmose” appears to follow in its 
amount the proportion of salt in solution, with a certain degree of regularity. The 
“chemical osmose” of substances, on the other hand, is found of high intensity with 
small quantities of the substance, such as 1 per cent, or even OT per cent., and to 
augment very slowly with increased proportions of the substance in solution. 
A small proportion of common salt accompanying carbonate of potash has been 
