Accurate sampling of pore waters and sediments in gravel river beds is prob
lematic because cobbles preclude most coring techniques, while the angulari
ty of sand grains destroys the membranes of standard pore water samplers. I
n the Clark Fork River, the recipient of over 100 years of mining activity,
a modified single tube core-freezing device was used to profile bed sedime
nt (similar to 65% cobbles and 35% gravel substrate), and a novel pore wate
r sampler was developed to collect interstitial water. In the sediment, the
< 63 um fraction comprised <4 wt % yet contained approximately an order of
magnitude higher metal concentrations compared to the 63 mum-2 mm fraction
(constituting similar to 20% of the mass). However, on a mass basis the sa
nd fraction contained 60% of the metals, compared to 40% in the clay/silt f
raction in the gravels, while in point bars the distribution was similar to
50:50. The metals occur predominantly in sulfides frequently armored with
an oxide rim and other sparingly soluble phases that may explain the low po
re water metal concentrations, These data demonstrate that consideration of
multiple particle size cutoffs is necessary to accurately characterize flu
vial bed sediment metal conditions and that the form of the metal is import
ant in understanding metal solubility in the benthos.