Two different methods have been used to fatigue test four bone cements. Each method has been used previously, but the results have not been compared. One method tests at least 10 samples over a single stress range in tension only and uses Weibull analysis to calculate the median number of cycles to failure and the Weibull modulus. The second test regime uses fewer specimens at various stress levels tested in fully reversed tension-compression, and generates a stress versus number of cycles to failure (S-N) or Wöhler curve. Data from specimens where the fracture surface contains pores greater than 1mm across is rejected. The single stress level test is quicker to perform however, provides only tensile fatigue data, but the material tested includes pores, thus more physiological cement is tested. The multiple stress level testing regime uses combined tension and compression loading and multiple stress levels, thus more physiological loading, but excludes specimens where the defects are greater than 1mm across, so is less representative of cement in vivo. The fatigue lives between the cements were up to a factor 15 different for the single stress level tension only tests, while they were only a factor of 2 different in the fully reversed tension-compression testing. The single stress level results are more effected by surface flaws, thus the differences found using the multiple stress levels are more indicative of differences in the fatigue lives of the cements. However, the single stress level tests are quicker, so are useful for initial screening.