The main aim of this work was to investigate previously published repo
rts that an acidic mixture of 0.67 M phosphoric acid, 20% dimethyl sul
phoxide, 2 M urea, 20 mM EDTA, 0.75 mM adenosine and 32.5 mM Zwitterge
nt (termed the PA reagent) extracted up to 7 times more adenosine 5'-t
riphosphate (ATP) from soil than a reagent consisting of 0.5 M trichlo
roacetic acid (TCA), 0.25 M phosphate and 0.10 M paraquat (termed the
TCA reagent). Several extraction tests were performed using both fresh
and air-dried soils at different soil-to-extractant ratios. The ATP c
oncentrations in the soil extracts were determined by the fire-fly luc
iferin-luciferase system. In 5 fresh U.K. arable soils the amounts of
ATP extracted, corrected for incomplete extraction by measurement of t
he recovery of an internal standard of added ATP (the spike), ranged f
rom 1.29 to 7.80 and from 1.01 to 5.24 nmol ATP g(-1) soil, for the TC
A and PA reagent, respectively. In air-dried soils the range was from
0.35 to 1.11 and from 0.35 to 1.24 nmol ATP g(-1) soil, respectively,
for the TCA and PA reagent. Except for an acid soil, the amount of nat
ive ATP extracted from the soils by the two reagents and uncorrected f
or incomplete extraction was not significantly different within soils
when a 1:5 soil-to-extractant ratio was used. However, the percentage
recovery of added ATP as a spike was larger at higher soil-to-solution
ratios with the TCA reagent but not with the PA reagent. The recoveri
es of the spike with the TCA reagent were always lower than with PA re
agent. However, soil ATP, corrected for incomplete extraction, was alw
ays greater with the TCA reagent than with PA. The use of [C-14]ATP as
a spike showed that no appreciable dephosphorylation of ATP added as
the spike occurred with either reagents. The main difference between P
A and TCA reagents was that with PA a constant and high recovery of sp
ike ATP (ca. 90%) was obtained at all soil-to-solution ratios tested,
while TCA gave higher recoveries of spike ATP at higher soil-to-soluti
on ratios. Paradoxically, the amounts of native ATP extracted at each
soil-to-solution ratio were generally similar between PA and TCA, alth
ough increasing with increasing soil-to-solution ratio. Correction for
incomplete extraction of the spike with the TCA reagent gave very sim
ilar soil ATP concentrations while this was not the case with the PA r
eagent. We suggest that this is the main difference between the two re
agents. We can find no evidence to support other work which suggested
that the PA reagent extracts more ATP from soil than the TCA reagent.