This paper introduces an experimental methodology for estimating satur
ation concentrations of oxygen in surface waters and wastewater that e
mploys commonly available analytic techniques (polarographic probe and
Winkler titration) for dissolved oxygen (DO) analysis. Because DO ana
lysis techniques can contain significant measurement bias, a protocol
intended to minimize the estimate uncertainty is proposed. The protoco
l's chief attribute is its ability to estimate solubility indirectly,
i.e. it avoids the complication of equilibrating samples with the atmo
sphere and uses measurements with unsaturated water to project a value
of solubility. Using this protocol, saturation concentrations in Miss
issippi River water collected in Minnesota were measured. The ratio of
the saturation concentration of Mississippi River water to pure disti
lled water varied seasonally in 1993 and 1994, with slightly higher ra
tios in the winter months and lower values in the late summer and fall
. The lowest measured ratio was more than 8% below the equivalent clea
n water value. (C) 1997 Elsevier Science Ltd.