658 
Analyses of Ashes of Plants. 
In this table we cannot fail to observe a certain degree of con- 
nection between the specific gravity and the weight per bushel in 
many cases. Thus specimens 2, 9, 21, 23, and 31 correspond in 
having low specific gravities and low weights per bushel, whilst 
several specimens of high specific gravity give also a good weight 
per bushel, as specimens 28, 38, and 46. But on the other hand, 
in the greater number of instances the relation existing between 
these two circumstances is anything but clear : thus in several 
samples of the same weight per bushel, the specific gravity is 
widely different; we need only instance specimens 21 and 30. A 
glance at the table will supply many other such examples. 
The weight per bushel of all the specimens was not experimen- 
tally ascertained — although the first 1 1 from the College farm were 
in the majority of cases weighed. Specimens 29 to 50 were de- 
termined by ourselves. Our method we detail, as it indicates the 
amount of confidence to be attached to the results obtained. 
A very nicely shaped glass flask, with a taper neck, capable of 
containing about 900 or 1000 grains of wheat, was selected ; it 
was filled repeatedly with the same wheat, various trials being 
made with the view of ascertaining in what way the nearest ap- 
proach to the same quantity in each trial could be secured. 
It was found that the plan affording most constant results was 
to allow the wheat to flow gently into the bottle without shaking 
it. In this way the quantity filling the bottle in several experi- 
ments did not ordinarily deviate to the extent of more than five or 
six grains in each weighing, whilst the difference was usually much 
less. A discrepancy in weight of 10 grains on 1000 would give a 
difference of abcut half a pound on a bushel. 
For each of these wheats six or eight weighings were made, and 
the mean taken as the true contents of the vessel. It is obvious 
that by this plan we obtain the relative weight per bushel of dif- 
ferent samples, and could immediately deduce the absolute weight 
were we provided with a standard of comparison — a specimen, that 
is, of known weight per bushel. Such a standard was obtained by 
comparing three specimens of wheat supplied to us by Mr. Farmer 
of the Enfield Mills, and of which the weight per bushel had been 
ascertained. We discovered that whilst two of these specimens 
agreed, when tested by our method, with the stated weight, the third 
exhibited a certain deviation which induced us to reject it as in- 
correct. 
By these two specimens we have decided the weight per bushel 
of all the others mentioned. We believe our method, which «as 
adopted on account of the smallncss of the samples, has afforded 
us results to within a quarter of a pound to the bushel, if not abso- 
lutely, at least relatively correct, and this is all we need in deter- 
mining lb^•» relation of specific gravity and weight per bushel; at 
