We used multiple linear regression analysis to investigate relationships between late-summer epilimnion thickness, transparency, lake area, acidity and summer weather conditions in a large ($n = 116$) multi-year data set for 9 small Boreal Shield lakes. Dissolved organic carbon (DOC) was the best individual predictor of late summer epilimnion thickness ($r;{2} = 0.69$). Total chlorophyll~$a$, the number of days between ice-out and late-summer stratification, and lake area collectively explained an additional 14% of the variation in epilimnion thickness. The three attributes of summer weather that we examined, mean daily temperature, mean daily wind speed, and mean daily hours of bright sunshine, did not add to the predictive ability of our regression model. Lake acidity also did not add directly to the predictive ability of the model, likely because DOC concentrations already reflected the effects of pH. Our study supports an increasing body of evidence indicating that the dominant effects of climate change on lake thermal structure in small lakes will be through effects on processes that affect lake transparency.