VARIATION IN THE CHEMICAL COMPOSITION OF SOILS. 15 
tively large quantities of easily soluble compounds have to be washed 
from a relatively small quantity of a nearly insoluble compound, 
on appreciable quantity of the latter may be lost before all of the 
former have been removed. As an example, in the fusion analysis 
of a soil after removal of the silica there remains an acid solution of 
all the bases with a large excess of sodium chloride. In the pre- 
cipitation of calcium as calcium oxalate, it is necessary, of course, 
to remove this excess of chlorides, and calcium oxalate being slightly 
soluble, some is lost and may or may not be recovered later in the 
analysis. The personal factor of judgment comes in at this point 
and will explain many discrepancies. 
The limits of error stated apply to analyses of soils of average 
composition by the fusion method. It is not meant, for instance, 
that the limit of error in all CaO determination is 0.10. for in the 
case of a dilute acid extract of a soil where the only excess of salts 
introduced is ammonium salts, the error allowed should be much 
smaller. 
One of the reasons for drawing attention to this matter here is 
the practice in some quarters of publishing soil analyses in terms of 
pounds per acre. Such results are presumably based on an average 
of several analyses, but the variations of such analyses usually do not 
appear. 
Assuming that an acre of soil 6 inches deep weighs 1,750,000 
pounds, the limits of error stated when calculated to pounds per acre 
6 inches deep would be as follows : 
CaO limit 0.10 per cent or 1,750 pounds. 
K 2 limit 0.05 per cent or 875 pounds. 
P 2 5 limit 0.04 per cent or 700 pounds. 
Stated in another way it would appear as follows: Suppose a 
soil on analysis gave K 2 0, 0.50; CaO, 0.30; P 2 5 , 0.08, allowing the 
above limits of error, the composition might be : 
Per cent. 
K 2 0. 525 or 0. 475 
CaO 0. 35 or 0. 25 
PA 0. 10 or 0. 06 
Or in pounds per acre: 
K 2 9, 187. 5 or 8, 312. 5 
CaO ; 6,125 or 4, 375 
P 2 5 1, 750 or 1, 050 
It is quite plain then that too much significance should not be 
attached to differences of a few hundred pounds of any ingredient 
when the composition of a soil is expressed in pounds per acre, not 
because the actual presence or absence of such an amount might not 
have an effect, but because the statement based on chemical analysis 
may depart that much from the actual fact. 
