Apparent bias in river water inoculum following centrifugation

We collected four measures of viable bacterial concentration (heterotrophic plate count, total coliform, fecal coliform, and Escherichia coli) and three measures of well color development in Biolog GN2 microtiter plates from water samples that were collected on two or three separate occasions from a fixed site on 19 different streams throughout Oregon. Our goal was to determine whether concentrating the water sample with centrifugation prior to analysis would change the in situ composition of the culturable bacterial assemblage. Each sample was split and one subsample was centrifuged while the other subsample served as a control. A shift in the proportion of each group of culturable bacteria toward more fecal coliform bacteria was observed following centrifugation. In samples with the lowest initial heterotrophic bacterial densities (under 50CFU/100ml), the observed concentration following centrifugation was much lower than expected. However, samples that had high initial heterotrophic bacterial densities (over 1000CFU/100ml) had concentrations at or above expected values following centrifugation, but were biased toward a higher proportion of coliform bacteria. Bacteria in centrifuged subsamples utilized more sole-carbon substrates on Biolog GN2 microtiter plates and showed a shorter lag time prior to tetrazolium color development than their uncentrifuged counterparts. Future research that focuses on characterizing and accounting for the bias associated with centrifugation of water samples held for less than 24h is recommended.