The Diffusion Process. 63 
to diffusion the broken up megass, which does not, like 
the cane, require cutting.' 
Now though this argument is very plausible, and 
though it has been advanced by men whose opinion is 
entitled to every consideration, I venture to say it is 
entirely fallacious and unscientific. From what I have 
already said about the treatment of the cane in the 
diffusion battery, it is evident that the material to be 
diffused must be saturated with water ; consequently 
that any bibulous body like megass must absorb a large 
quantity of water, and so dilute its juice, before diffusion 
proper commences. Thus even a theoretically perfe6l 
diffusion of megass will give only a dilute juice, while 
that of the cane yields a juice almost as sweet as that of 
the cane. I say almost, because there is an absorption 
of water even with the cane. 
I venture to think that these considerations alone 
show that the diffusion of megass is unsound and im- 
practicable ; apart from the question of the advisability 
of twice handling the raw material, and the fa6l that it 
was found by trial in Guadeloupe that the megass must 
be sliced preparatory to diffusion, and that in the juice 
so obtained a large amount of crystallisable sugar was 
inverted. Nor need I show here, as I might, that the 
extraction of sugar from the megass by this means, 
though it is a diffusion, is a very different phenomenon 
from the apparently similar extraction from the cane ; 
where there comes into play the natural law of ' osmosis/ 
or passage of soluble bodies through organic septa, 
and not merely the mixture of fluids not so separated. 
Hitherto we have only considered the extraction of 
the sugar from the cane in the form of a juice : it will 
