The within-day, between-day and month-to-month variability of light attenuation and microalgal chlorophyll a (chl a ) and photosynthetic response was measured in San Antonio Bay, Texas, USA. Waters were shallow (<2 m) and turbid (attenuation coefficients of 0.7 to 15.3 m"), yet daily rates of primary production (0.1 to 2.5 g C d-') were comparable to those in much deeper and clearer estuaries. Chl a concentrations in the sediment (459 to 7837 mg m-3 in the surficial millimeter) were much higher than those In the water column ( 4 to 48 mg m-'). The benthic assemblages were photosynthetically competent, but dally benthic primary productivity was low (0.00 to 0.09 g C m-' d-'; an average of 2'Lb of productivity in the water colun~n) because of the very low irradiances at the sedimentwater interface. The high rates of production by the suspended mlci-oalgae were largely due to high chlorophyll-specific light-saturdted rates of photosynthesis. P:,:' 13.0 to 24.4 g C (g chl a)-' h -' ] , which were correlated positively wlth temperat~lre and inversely with the mean irradiance in the water column. The between-day and temperature-~ndependent variation in P$! was also correlated with F,,Chl-l, a n index of the proportion of functional photosystem 11 reaction centers. In turn, within-day variability in F,,Chl-l was inversely correlated with the mean irradiance in the water column in 8 of 10 observations, a pattern consistent with photoinhibition. A depression of P : ! ' caused by photoinhibition throughout the water column may therefore b e responsible for the inverse trend of P$' with mean irradiance in the water column. Short-term (h) variability in suspended chl a and turbidity was high (coefficient of variation = 13 to 7 5 % ) , but estimates of daily productivity could be predicted with reasonable fidelity (mean error 2 7 % ) from a single midday determination of chl a, the photosynthesis versus lrrad~a n c e response and the attenuation coefficient, along with daily incident radiation. The predictive power of a single observation was due to coherence in the variation of chl a, P: :' and the attenuation coefficient In the water column-the decrease in mean irradiance caused by resuspension was compensated for by concomitant Increases in Pc,hland suspended chl a Between-day variability in productivity of 15 to 5 2 % approached month-to-month differences, so the optimal use of resources in monitoring productivity would be to take slngle samples daily.