The first 10 natural frequencies agree within 0.1% with those given by Johnsen and Dey (1978). The power spectral density of the displacement at the tip of the cantilever is shown in Figure 3. For all nodes the values at the eigenfrequencies compare well with the results of Johnsen and Dey. For nodes close to the built-in end of the cantilever, discrepancies appear at higher frequencies. These differences are attributed to the use of a beam-column element in Abaqus (element type B23) that uses the axial strain as an internal degree of freedom in the element, so some axial modes appear at higher frequencies. The element used by Johnsen and Dey does not have these same modes. The differences are not important because they could be eliminated by using a finer mesh if the high frequency response close to the base of the cantilever must be predicted accurately.