Author's reply Dr Randolph's elegant, yet simple, analysis demonstrates the value of having closed-form solutions to engineering problems. Such solutions not only provide a check on corresponding solutions from numerical analyses, but also help to identify the dimensionless parameters which control behaviour. Hence, Dr Randolph has found that greater generality is achieved if one uses a stiffness ratio defined in terms of the rate of increase of skin friction k, rather than the rate of increase of soil modulus N,. The use of the latter definition is indeed appropriate when considering settlements in the elastic (working) range and can also be advantageous when cyclic degradation of skin friction is being incorporated into the analysis (Poulos, 198 1). There is little doubt that, although pile compressibility contributes to the length effect on skin friction, there are other factors which may also be significant, particularly strain-softening at the pileesoil interface. For example, model tests by Chandler & Martins (1982) indicate that reductions of up to 50% in the skin friction can occur following relatively large displacements. This effect would result in smaller values of the reduction rates i, than are indicated by the present analysis, and an analysis incorporating this strain-softening behaviour would give better overall agreement with the available data.