The “Paradox of the Plankton” was the response of Hutchinson (1961) to the apparent violation of the Principle of Competitive Exclusion in phytoplankton communities. One explanation for phytoplankton associations was that communities exist in a perpetual nonequilibrium and competition is never allowed to proceed to exclusion. However, current ecological theory argues that disturbance does not preclude the importance of niche differences among species, including those that involve competitive processes. I collected phytoplankton and limnological data over the winter mixing period in two monomictic reservoirs to determine if niche differences between species are relevant during mixing. I used ordination to assess the importance of environmental parameters to species abundance patterns, and then evaluated the significance of changes in abundance along ecological gradients versus neutral theory, which assumes niche differences are irrelevant to community assembly. Two flagellate species displayed different abundance patterns, where one was more abundant during deep mixing and high nutrient concentrations, while the other was favored under slight water column stability and relatively lower phosphorus concentrations. Also, the abundance of four large diatom species was distinct along axes that were described by light availability and mixing intensity. The abundance of all species was significantly outside of the predictions of neutral theory. These results emphasize that non-neutral niche differences are relevant to community assembly and suggest that competitive processes cannot be ignored during disturbance.

Hutchinson also noted that the paradox may be "specious" and the lake environment is less homogenous than it appears. I used subsets of an extensive phytoplankton dataset to evaluate the effect of local sampling scale on the recognition of phytoplankton species abundance maxima and co-occurrence. Reduction of temporal or spatial scale resulted in considerable departures in the identification of abundance maxima for some species and created the illusion of coincident peaks in abundance between species. In addition, the use of presence/absence data and reduced sample scales created the appearance of co-occurrence between two species that in reality exhibited overlapping, but distinct, distributions. These results suggest caution in using phytoplankton associations as ecological indicators.