Paleolimnological methods can be used to identify baseline limnological conditions and to assess anthropogenic changes in lakes that lack historical limnological data. We studied a suite of Florida lakes and developed regression models that predict limnetic total P and Chiafrom diatom assemblages and nutrient accumulation rates in surface sediments. These models were applied to a210Pb-dated sediment core from Lake Parker, Polk County, Florida, and a pollen-dated profile from Lake Francis, Highlands County, Florida, to infer historical water quality. Our results show that Lake Parker displayed statistically significant increases in limnetic total P and Chiaconcentrations since the 1920s. Lake Francis has been eutrophic for more than 2 millennia, but shows a statistically significant decrease in limnetic total P during the present century. This study demonstrates that lakes may be naturally productive because of edaphic influences and may respond differently to human disturbance. We conclude that paleolimnological estimates of predisturbance nutrient/Chiaconcentrations can help define target conditions for water quality restoration programs.