548 
Oil the Food of Plants. 
The liquid, holding' the finely divided matter in suspension, is then 
put on a -weighed filter, and the proportion of that substance, after proper 
drying, ascertained; but here a practical difficulty occurs : the mixture 
often refuses to filter, the finely divided substance closing the pores of 
the paper, so as greatly to protract that process, while what licpiid does 
come through passes muddy. This inconvenience is obviated by boiling 
the whole in a glass flask for an hour, when filtration tal<es place more 
readily. Even after this I have found, when operating on clayey soils, 
that the filtered liquid cannot be obtained clear, in which case it should 
be evaporated down in a porcelain vessel to a small bulk and again put 
upon a little weighed filter. In this manner the difficulty is got over, 
but a slight error is involved by the action of the boiling Avater upon 
the humus, whereby a small portion is taken up and the solution 
rendered yellowish, but this quantity is very small, as the examples to 
be adduced will show. When tlie suspended finely divided matter of 
clay soils is suff'ered to stand, even for many days, although the greater 
part settles down, the liquid remains turbid, and if examined for salts in 
this state, would certainly cause a larger error than that mentioned 
above. 
The clear yellowish liquid so obtained, Avhich usually amounts to 
about a pint, holds in solution all the soluble saline constituents of the 
soil, including the gy])sum, unless, indeed, that substance is present in 
very large proportion. It is evaporated to dryness without ebullition in 
a small platinum capsule, and the residue weighed, ignited to redness 
until the charcoal is burned away, and again weighed. 
There now remains to be examined the finely divided substance. 
The first thing to be done is to determine the proportion of organic 
matter ; to do this exactly would be a very difficult task. After much 
thought I am inclined to the opinion that the old plan, that of calcina- 
tion, is likely to give, when properly conducted, a result quite as good 
as that which might be obtained by a more complicated process. A 
convenient portion of the substance, well dried at 300^ and in fine 
powder, is heated to redness in an open platinum crucible until all 
blackness disappears, when the loss of weiglit after cooling gives the 
quantity of organic matter; a little too high from the w^ater disengaged 
from the clay at this higli temperature. So far as appears to me this 
error can hardly rise to 2 per cent. 
The calcined substance is next mixed with four times its weight of 
dry carbonate of soda, and fused in the same platinum crucible at a good 
red heat for 15 to 20 minutes; the melted mass is softened by water 
and then treated, with the precautions proper to these processes, with 
excess of hydrochloric and a little nitric acid, the whole evaporated to 
dryness, acidulated water added, and the insoluble silica collected on a 
filter, washed, ignited, and weighed. 
To the solution and wash-water, concentrated by evaporation, am- 
monia is added, which precipitates alumina, oxide of iron, and the 
little earthy ])hosphates w hich may exist in the substance ; the quantity 
of the latter is however so small, that in this analysis it may be safely 
neglected. The precipitate, well washed, is then heated with dilute 
caustic potash, which dissolves the alumina and leaves the oxide of iron, 
