Predicting the effects of climate change and atmospheric deposition on water quality requires predicting the effects of landscape form on export of substances downstream. In this paper, we present dissolved organic carbon (DOC), total phosphorus (TP), and iron (Fe) export data (1980--1992) for 20 relatively undisturbed, forested catchments draining into seven lakes in central Ontario and develop regression models of chemical export as functions of landscape composition. The extent of wetlands was correlated with export of DOC and TP; the proportion of the catchment covered by peatlands accounted for 78% of the variance in a regression model of long-term average DOC export and 76% of the variance in a model of color export. Peatland coverage and Fe export together explained 76% of the variance in a long-term average TP export model, which is consistent with published experimental evidence that Fe facilitates P complexation with DOC in surface waters. The long-term average Fe export model was not significant when all 20 catchments were included. However, Fe export from the 14 catchments with thin tills was a function only of peatland coverage (R2=0.71), suggesting that Fe export is dependent to a large extent upon either the export of organic material or the reducing conditions. The long-term export models worked well when export of a substance was dominated by peatlands but was not very sensitive to the influence of mineral soils. The long-term average TP/DOC ratio was remarkably constant among most whole-lake catchments, ranging from 1.4 to 2.0 mg P/g C, the exception being the catchment of anthropogenically acidified Plastic Lake with a ratio of 0.8. Fe export to Plastic lake was also enriched relative to TP export compared with the other lakes. Therefore TP export to Plastic Lake may be limited by some mechanism related to acidification.¿ 1997 American Geophysical Union