118 
Some Field Experiments on the 
On carefully examining and comparing the foregoing Tables, 
together with the plan of the field, it will be noticed that all 
the plots on (he western side of the field gave better results 
than their duplicate neighbours on the eastern side, showing 
that this part of the land must have been naturally more pro- 
ductive. The differences, however, are not on the whole great, 
and as they pervade the series, the mean result of each pair of 
duplicates may be regarded as correctly showing the effect of 
the manure applied. 
Taking, in the first place, the plots that received no dung 
and no artificials, we find that the soil, unassisted by manure 
of any kind, produced less than 2^ tons of roots per acre. 
If we subtract this weight from the mean yields of the other 
plots, we shall get the following results in the shape of increase 
due to the manures employed : — 
Increase of Moots lycr acre due to Manures over and above the Yield of 
Unmanured Land, 
Manure per acre. Tons. cwts. qrs. lbs. 
10 tons dung ., 8 73 7 
10 tons dung and 5 cwts. supeiiilio-sphate .. 8 13 1 21 
10 tons dnng and G| cwts. ground coprolites .. 10 1 2 7 
5 cwts. superphosphate only 8 13 1 21 
6i cwts. ground coprolites only !) 18 1 25 
It is singular that while the dung alone gave an increase of 
8J tons nearly, and while a larger increase was given by the 
artificials used alone, yet where the dung was used in addition to 
the artificials, it will appear to have had no effect. By an odd 
coincidence, the mean of the superphosphate plots happens, 
indeed, to work out at precisely the same figures both on the 
dunged and on the undunged acre ; and there is practically no 
difference between the mean of the coprolite plots with and 
without dung. 
That the ground coprolites gave in these experiments better 
results than the superphosphate, occasioned me some surprise, 
but there was no doubt as to what would be the ultimate result 
of the weighing as soon as the roots were pulled ; for those 
from the coprolite plots were cleaner, larger, and rounder, than 
any of the others. It is to be noted that the proportion of 
phosphate actually applied per acre was more than twice as 
much in the 6J cwts. of coprolites as in the 5 cwts. of super- 
phosphate ; but I was inclined to attribute the less satisfactory 
effect of the superphosphate to the natural poverty of the soil in 
carbonate of lime. The total proportion of lime in the soil was 
not much more than one-half per cent. ; and this appears to 
exist in the form of silicate, since the dry soil does not eflcr- 
