In Section I, a model combining algal mortality due to selective zooplankton grazing with the Monod model of phytoplankton growth was used to predict phytoplankton species composition along a gradient of phosphorus (P) concentration and zooplankton grazing pressure. Model predictions were compared to the results of a mesocosm experiment of phytoplankton species composition along a range of P concentrations of 5 to 115 g/l and potential grazing loss rates of 0.001 to 0.27 per day. The effect of the P gradient was measured by monitoring phytoplankton and zooplankton species composition and biomass, and physical and chemical parameters for 7 wk in 12 fiberglass tanks (5500 L) filled with lake water and associated plankton. P concentrations were manipulated so that the tanks were evenly distributed along the gradient.

Using laboratory derived Monod growth constants for P and literature values of phytoplankton selectivity coefficients, zooplankton filtering rates, and zooplankton assimilation efficiencies, the model predicted (1) the existence of three alternate states, (2) breakpoints between the alternate states which are similar to the P concentrations defining Vollenweider's lake trophic states, and (3) that phytoplankton species composition along the P gradient is dependent on a trade-off between a species ability to compete for P and their susceptibility to zooplankton grazing. Results from the mesocosm study generally supported these predictions and suggest that (1) the trade-off between competitive ability for P and edibility is a primary factor structuring phytoplankton communities, (2) multiple states exist along an environmentally realistic P gradient, and (3) these multiple states are consistent with the long-standing P-based lake trophic classification.

In Section II, the effects of a reduction in total-P concentration on the water quality and plankton community structure in a 86-ha hypereutrophic sandpit lake with high internal P loading were assessed by dosing an isolated 4.6-ha section of the lake with 34,065 L (dose = 10 mg Al L-1) of liquid aluminum sulfate (alum). During the three summers following treatment, hypolimnetic total dissolved-P, epilimnetic total-P, and eplimnetic total nitrogen were decreased by 97, 74, and 36%, respectively, in the treated section. Secchi depth was 134% greater in the treated area. Alum treatment also increased the volume of usable fish habitat by 22%, as the depth of the 3.0 mg/l dissolved oxygen isocline was 52% deeper in the treated portion. Total phytoplankton biovolume decreased by 40% and chl a concentration by 65% in the treated area. Although cyanophytes continued to dominate in the treated area, there was a shift in relative abundance from cyanophytes to bacillariophytes and chlorophytes, especially during the second summer after treatment. This shift generally supports the modeling predictions and experimental results of the first section. Overall, alum was extremely effective in controlling sediment phosphorus release rates and lowering water column P concentrations and thus improving water clarity, reducing phytoplankton biomass, shifting phytoplankton species composition from cyanophyte dominance toward bacillariophytes and chlorophytes, increasing daphnid biomass, and increasing usable fish habitat.