Limnol. Oceanogr., 44(8), 1999, 1928–1935

Scaled chrysophytes and planktonic diatoms are used to infer changes in lake water pH, specific conductivity, trophic score, and total nitrogen in 23 Connecticut waterbodies over the last 100 yr, and the changes are correlated with quantified changes in land use in the surrounding watersheds. In general, there was good agreement between the changes inferred from both organismal groups in this suite of lakes. Significant correlations were observed between chemical conditions inferred from organisms in surface sediments and present-day land uses, especially the percentages of the watersheds that are forest or residential land cover types. Approximately 20% of the waterbodies have significantly increased in pH since 1890, and none of the lakes have significantly declined in pH despite the fact that this region receives significant amounts of acidic deposition. These findings support previous work, indicating that the pH of Connecticut lakes has not declined over the recent past. One fourth of the lakes have significantly increased in specific conductivity, especially those situated in watersheds that have become highly residential in nature. Inferred specific conductivity has more than doubled in six of the lakes during the last century. Six of the lakes have become significantly more eutrophic, while only one lake has become more oligotrophic since 1890. The majority of the lakes situated in watersheds that have remained over ;80% forested have not significantly changed, whereas those that have become over ;25% residential have experienced the greatest amount of change. The potential influences of watershed-based alkalinity generation, winter road salt use, and implications of these findings in lake management are discussed. One of the primary reasons that the effects of environmental stresses on aquatic ecosystems are difficult to assess is the fact that historical lake water data are often lacking (Brenner et al. 1993; Siver et al. 1996). As a result of the paucity of background data, it is often impossible to estimate the degree and rate of change in the chemical and biological structure of waterbodies and to distinguish between natural changes and those caused by anthropogenic stresses. However, recent advances in paleolimnological techniques, especially formation of biological-based inference models, have resulted in the ability to reconstruct historical lake water conditions with a high degree of resolution (Bennion et al. 1996). Most inference models are based on organisms that are differentially distributed over the environmental gradient of interest. The point along a given gradient where a species is most abundant and the degree of spread of the taxon along the gradient are referred to as the optimum and tolerance, respectively (Birks et al. 1990). The optima and tolerances of all taxa along a specific gradient are often estimated using weighted averaging where the abundances of each organism in surface sediments of a suite of lakes are correlated with contemporary lake water chemical conditions (Birks et al. 1990). The analyses result in relationships referred to as inference models (Birks et al. 1990). Once formed, the inference models are then applied to older sections of cores in order to estimate the historical chemical conditions. Such a paleolimnological method has been successfully used to con-