The Columbia River estuary is dominated by a detrital food web that feeds p
rimarily on river-borne particulate organic material in the estuarine turbi
dity maxima (ETM). Approximately 90 % of the bacterial activity and most of
the extracellular enzyme activity in the water column of the estuary were
associated with particles captured by a 3 mum filter. Earlier studies deter
mined that the relationship between particle-attached bacterial activity an
d suspended particulate mass (SPM) was variable, suggesting that some parti
cles supported a larger fraction of bacterial activity. In order to charact
erize these bacterially-active particles, suspended particulate material fr
om the estuary was fractionated by size and in situ settling velocity, and
analyzed for H-3-thymidine incorporation rate, total particle mass and part
iculate organic carbon (POC) concentration. The location and movement of ba
cterially-active particles in the estuary was traced by measuring H-3-thymi
dine incorporation rate and SPM in near-bottom depth profiles collected at
4 or 5 time points during ETM resuspension events. The smallest particle si
ze fraction (3 to 10 mum) supported 87 % (SE = 13) of bacterial activity, b
ut contained only 38 % (SE = 5) of total POC and 38 % (SE = 6) of total par
ticle mass. However, when particles were separated by in situ settling velo
city using a method that preserved the integrity of aggregated particles, t
he settling velocity of the majority of bacterially-active particles varied
from < 0.07 mm s(-1) in some samples to >0.75 mm s(-1). Microscopic analys
is of bacterially-active particles revealed that they were aggregates of sm
aller (<10 <mu>m) particles held together by a transparent matrix. This stu
dy shows that small, slow-settling particles host most of the bacterial act
ivity in the Columbia River estuary, and suggests that they contain the mos
t rapidly consumed pool of organic matter in the system. Furthermore, it de
monstrates that the configuration of these particles may be highly dynamic
as they are regularly aggregated and disaggregated in ETM. In the estuary,
these particles often appeared in the water column early during developing
flood tides prior to the appearance of the most turbid part of ETM, and som
etimes remained in the water column after most ETM particles had settled ba
ck to the bed. This cycle places bacterially-active particles in the water
column longer than most ETM particles, making them more likely to be washed
out of the estuary, but also making them more available to suspension-feed
ing detritivores. Formation of large, fast-settling macroaggregates during
the late stages of flood and ebb tides and during slack tides is probably t
he mechanism by which bacterially-active particles are trapped in the ETM,
and may be essential to maintaining the estuarine community of particle-att
ached bacteria.