The results of an investigation into the relationship between surface sediment subfossil chi- ronomid distribution and water quality are presented. Data from 30 lakes in the middle and lower reaches of the Yangtze River indicate that the nutrient gradi- ent was the major factor affecting the distribution of chironomids across these sites. Canonical corre- spondence analysis (CCA) revealed that of 12 sum- mer water environmental variables, total Phosphorus was most important, accounting for 20.1% of the variance in the chironomid data. This was significant enough to allow the development of quantitative in- ference models. A TP inference model was devel- oped using weighted averaging (WA), partial least squares (PLS) and weighted averaging partial least squares (WA-PLS). An optimal two-component WA-PLS model provided a high jack-knifed coeffi- cient of prediction for conductivity r 2jack = 0.76, with a low root mean squared error of prediction (RMSEPjack = 0.13). Using this model it is possible to produce long-term quantitative records of past water quality for lacustrine sites across the middle and lower reaches of the Yangtze River, which has important implications for future lake management and eco- logical restoration. The results of an investigation into the relationship between surface sediment subfossil chi- ronomid distribution and water quality are presented. Data from 30 lakes in the middle and lower reaches of the Yangtze River indicate that the nutrient gradi- ent the major factor affecting the distribution of chironomids across these sites. Canonical corre- spondence analysis (CCA) revealed that of 12 sum- mer water environmental variables, total Phosphorus was most important, accounting for 20.1% of the variance in the chironomid data. This was significant enough to allow the development of quantitative in- ference models. A TP inference model was devel- oped using weighted averaging (WA), partial least squares (PLS) and weighted averaging partial least squares (WA-PLS) provided a high jack-knifed coeffi- cient of prediction for conductivity r 2jack = 0.76, with a low root mean squared error of prediction (RMSEPjack = 0.13). Using this model it is possible to produce long-term quantitative records of past water quality for lacustrine sites across the middle and lower reaches of the Yangtze River, which has important implications for future lake management and eco-logical restoration.