SUMMARY. 1. Profundal macroinvertebrates and water chemistry of sixty‐eight nutrient poor, poorly buffered lakes were sampled in 1983 or 1986. Assemblages of profundal zoobenthos were classified using two‐way indicator species analysis (TWINSPAN), ordinated by canonical correspondence analysis (CCA), and related to physico‐chemical factors using CCA. discriminant analysis, and regression.2. 90% of zoobenthos (TWINSPAN) groups were classified correctly using discriminant analysis. Depth (Function 1), soluble reactive Si(OH)4 and HCO3, ‐(2), and phytoplankton biovolume and pH (3) were the strongest correlates with the three discriminant functions.3. Regression analysis showed that depth, KmnO4, consumption, and phytoplankton biovolume were the best estimators of zoobenthos biomass in the profundal.4. Multivariate analyses showed species assemblages amongst the profundal zoobenthos to be good indicators of lake type, particularly depth, pH, and phytoplankton biovolume.