Low-energy n-16O elastic scattering phase shifts are calculated from potential matrix elements using nuclear-reaction theory. As in bound-state calculations, it is found to be essential to include second-order contributions. When this is done the theoretical calculations which contain essentially no adjustable parameters reproduce the experimental phase shifts reasonably well, the disagreement between theory and experiment being generally no larger than the spread in the theoretical results produced by different two-body forces. A nonlocality of the optical potential arises naturally out of the calculation. We are able to see the effect of this by comparing the internal wave function to that obtained using a Woods-Saxon well with parameters adjusted to give the same phase shift, i.e., we can investigate the Perey effect. The scattering wave function is found to be about 30 to 50% smaller in the potential region than the corresponding wave function calculated from the Woods-Saxon well.