Multivariate statistical techniques were used to examine the relationships between surface‐sediment cladoceran assemblages and 28 physical and chemical variables in 53 small subarctic lakes from northern Fennoscandia. The lakes were distributed along a steep eco‐climatic gradient, spanning boreal corniferous forest to treeless tundra. In general, the sites were small, oligotrophic, and bathymetrically simple, with little or no disturbance in their catchments. From the initial 53 localities, only 36 contained a sufficient number of cladoceran remains for reasonable quantification. From these, a total of 29 cladoceran taxa representing 19 genera were identified, comprising predominantly littoral chydorid species. A constrained redundancy analysis (RDA) and associated Monte Carlo permutation tests indicated that maximum lake depth, sediment organic content, epilimnetic summer temperature, lake perimeter, and lake catchment area made statistically significant (p le; 0.05) contributions to explaining the variance in the cladoceran taxon data. These five variables together accounted for 67.7% of the explained variance, and made a unique contribution of 26.8% to the total variance: all physical determinants independently captured 33.2% of the total variance. The significance of the most powerful explanatory variables is discussed in the paper in detail, and autecological information regarding the most common cladoceran taxa is given.To assess the potential of cladoceran assemblages in environmental reconstruction, quantitative inference models for mean July water and air temperatures were developed for the cladoceran assemblage using partial least squares (PLS) regression. The final prediction model yielded a root mean squared error of prediction (RMSEP). as assessed by jackknifing, of 1.19°C for Cladocera‐water temperature data‐set, whereas the cladoceran assemblages showed only very weak relationships to mean July air temperature. The overall results emphasize the role of physical factors in regulating species abundance and distributions in these environmentally sensitive ecotonal lakes.