The axial and radial shrinkage of bovine rod outer segments, monitored by near-infrared scattering changes (P-signal), is investigated in dependence on the intensity of the activating flash. Suspensions of axially oriented and randomly oriented rod outer segments were measured. In the latter case, axial and radial effects are superimposed to another. The following results are obtained: 1. The axial signal (Pa, tau approximately 10 ms) and the radial signal (Pr, tau = 40-100 ms), simultaneously measured on axially oriented rod outer segments, are similarly saturated with a half-saturation at a rhodopsin turnover of 3.5%. 2. For the saturation of the signal amplitude, measured on randomly oriented rod outer segments, a good fit is obtained by: P(rho) approximately 1 - e-beta rho, rho: relative rhodopsin turnover by the flash; beta is found in the range 23 less than or equal to beta less than or equal to 27 in all measurements. 3. The kinetics of the signal, also measured on the isotropic sample, depends on the rhodopsin turnover, the apparent time constant becoming faster with increasing turnover, The distortion of the signal cannot be fitted by a sum of exponentials with a fixed set of time constants. The signals from the isotropic sample are fitted by a phenomenological model. It introduces three first order processes concatenated in series; the first step is assumed as a rhodopsin transition inducing the two further processes. The distortion of the signals with increasing rho is then described assuming a rho-dependent quenching of this induction, according to the measured amplitude saturation. The time constants remain thereby unchanged. The fit yields the values in 2/k = 4, 11, and 45 ms with mean square deviations of 20%.