The plots of degradation versus displacement amplitude shown above have been obtained for a particular relative density and cycle number. Figure 11 shows the effects of increasing cycle numbers on the degradation factor for dense sand with a displacement amplitude of 1 mm. As the number of cycles increases, further reductions in the degradation factor occur, although these reductions are only significant for the higher normal stiffness. As has been mentioned above, the shear and normal stresses drop fairly rapidly during cycling in the CNS tests, and after 500 cycles the normal stress had dropped to very low values for both normal stiffnesses. Nevertheless, clearly there is still significant capacity when the specimens are sheared monotonically to failure. Also shown in Fig. 11 are data reported by Lee (1987) showing the effect of cycle number on the degradation factor in model pile tests with an estimated normal stiffness of K 5000 kPa/mm. The pile data initially show the same tendas evident from the CNS tests that increasing normal stiffness results in lower degradation factors but large cycle numbers appear to suggest a threshold below which the degradation factor will not drop The CNS tests. in contrast, provide no evidence of a threshold as compressive strains continue during cycling even when the normal stress has dropped to 3 kPa, and it might be expected that these strains will lead to further reduction in the degradation factor, It has previously been shown (Chan and Hanna 1980) that piles may fail after many thousands of load-controlled cycles, a response that could be predicted from the results of the CNS tests.