The reactivation of rhodopsin after photoregeneration from metarhodopsin in the UV-sensitive cells of the median eye of Limulus was examined by means of extracellular electroretinogram (ERG) measurements. Absorbed photons convert the transducing rhodopsin (Rt) to metarhodopsin, which is thermostable and can be reconverted by another photon to non-transducing rhodopsin (Rn). The amplitude of the ERG is assumed to correlate linearly with the amount of Rt under otherwise constant conditions. The results demonstrate that the reactivation of Rn recorded in vivo in the intact animal is much faster than that in the excised eye [half period: 4 min (in vivo), 24 min (excised eye)]. In the excised eye the ERG amplitudes recover over a sigmoidal time course; however, in vivo the kinetics often appear to be exponential. The in vivo kinetics were measured by several defined molar ratios of Rn to rhodopsin plus metarhodopsin [Rn/(R + M)]. These were adjusted by different durations of pre-illumination with UV light, which is preferentially absorbed by rhodopsin. The in vivo kinetics were fitted by a single exponential function. At very high molar ratios of Rn (> 90%) the kinetics become sigmoidal even in the in vivo experiments. The half-life of the in vivo kinetics depends linearly on the initial molar Rn/(R + M) ratio [half-life: 4-12 min with a rise of 0.1 min/% inactivated metarhodopsin (Mn), 20 degrees C]. The results are consistent with the assumption of multiple-step dephosphorylation being the rate-limiting step in regeneration of rhodopsin in the dark.