This paper examines several methods for calculating heating and cooling degree-days to any base by relying on temperature statistics such as the monthly mean temperature and the standard deviation of daily average temperature. It is found that the method developed by Schoenau and Kehrig works best, and is usually able to estimate monthly heating and cooling degree-days to within 3°C/day or 5.4°F/day (in a rootmean- square error sense) of their true value. The method clearly outperforms other widely used methods, such as the one derived by Erbs et al. (1983). The paper also presents an extension of the Schoenau and Kehrig method to the case when only monthly mean temperatures are known.