Abstract: The problems and solutions of applying the bender element method to determine the maximum shear modulus G_max) of sand are investigated by conducting resonant columnRC) and bender elementBE) tests on four dry sands obtained from Germany. The study indicates that the start-to-start method is more reliable to identify the time of shear wave propagation in specimens compared to other methods. In BE tests, the excitation frequency of input waves has influence on the tested G_max, and this influence is dependent on soil types and reduced by increasing confining pressure. A comparison between RC and BE tests shows that G_max by both methods may be well expressed as a linear equation; for the Berlin sand and Braunschweig coarse sand, there exists a threshold G_max below which G_max by BE tests G_maxBE)) is greater than that by RC tests G_maxRC)), however beyond which the G_maxBE) is smaller than G_maxRC) and the distance between both increases with increasing stiffness soil. In practice, it is suggested that the input voltage with proper high frequency is preferred for BE tests, as compared with RC tests.