Ann,
 
Part of the problem with setting up said controled experiment is finding the
unfertilized field.  Estimates range as high as 90% of the area in New
England having been plowed, and possibly fertilized -- the other 10% being
welands and ledge.  Fortunately, most soil amending chemicals, particularly
the post 1940's species, are designed for optimum uptake by plants, not by
the soil.  Thus, the farmers' need to fertilize yearly -- good for Dow, but
not necessarilly for the rest of the environment.  With this short life-time,
applied fertilizers do little to mask the chemicals which due to their longer
life-time (not being readily extracted by intensive agriculture) have become
accluded into the soil, and thereby not readilly available to crops.
 However, again it depends on context.  For example, calcium and other earth
metals will quickly flush from a sandy acidic soil in a area of
moderate-to-heavy rainfall; organic carbon will be quickly recycled in warm
moist, non-histic (wetland) soils; and phosporous in low pH/high eH soils
will all be unlike signitures reguardless of their cultural source.  The
problem of interference from agricultural fertilizers becomes problematic (as
a general rule only) when the archaeological signiture being observed is in
such small quantities where it cannot be descerned from the background
(affected) soil, or the signiture being observed has been randomly spread by
cultures for so long that it has become uniform -- eg in European
agricultural fields.
 
I strongly support more use and research into the pedo-chemical analyses of
archaeological sites, but caution its mass application without first due
consideration of the unique context of the soil species being studied.  For
example, phospates which were the big craze in the early 80's, and in many
areas continues in favor, rarely works in northern portions of New England,
and I would suppose much of Canada, norther Europe, Asia, and the southern
couterparts of Australasia, Africa, and South America.  Spodisols, common to
these areas, tend to have low pH accompanied by high eH values.  Under these
conditions, the phosphate ion is quickly accluded by the Fe-3 ion in the
soil.  Fluctuating perched ground water reduces the Fe-3 to Fe-2 which is
water soluable.  The iron along with its phosphate are litterly flushed from
the soil on a seasonal (or more often) basis, and therefor seldom available
in quantities sufficient for measuring.  While a common problem for most
soils here, specific soil species are not so affected, and thus such analyses
are appropriate and provide us a valuable tool.
 
Douglas