In vitro studies have demonstrated that lead selectively and reversibly depresses the rod photoreceptor component of the electroretinogram (ERG). To determine if low-level lead exposure during early postnatal development produced long-term selective rod deficits, we examined rod and cone ERG functions and cyclic GMP and cyclic AMP metabolism in adult control and lead-exposed rats. A-wave and b-wave voltage-log intensity and latency-log intensity functions, generated from single-flash ERGs in fully dark-adapted rats, revealed that low-level lead exposure during early postnatal development caused a 23- and 18% decrease in maximum amplitude, a 1.0- and 0.5 log unit decrease in absolute sensitivity and a mean latency increase of 47- and 29%, respectively. Additional ERG experiments, using scotopically balanced stimuli and scotopic and photopic flicker fusion frequency functions, also demonstrated selective rod deficits. Cone ERGs, elicited by 30-Hz white flashes in the presence of a white background adapting light, were similar in control and lead-exposed rats. Lead exposure during early postnatal development caused cGMP levels in dark-adapted and light-adapted retinas to increase 40- and 25%, respectively, above controls whereas cyclic AMP levels remained unchanged. Light-activated cyclic GMP phosphodiesterase (cGMP-PDE) was inhibited 40% while guanylate cyclase activity was unchanged. The retinal lead concentration was 10(-6) M at the end of exposure (day 21) while at the time of ERG testing and biochemical analysis it was 10(-7) M. In vitro studies with adult control retinas incubated with 10(-9)-10(-4) M lead revealed a dose-response inhibition (10-40%) of cGMP-PDE between 10(-6)- and 10(-4) M lead and stimulation of guanylate cyclase (20-158%) only above 10(-4) M lead, indicating that cGMP-PDE is more sensitive to the direct effects of lead than the synthetic cGMP enzyme. These in vitro cyclic nucleotide metabolism results are similar to those we observed in vivo and both are consistent with the observed ERG changes. The selective rod-mediated amplitude, sensitivity and temporal deficits and the lack of effect on the cone ERGs clearly demonstrate that low-level lead exposure during early postnatal development causes a long-term selective disruption of rat rod photoreceptors. The relevance and applicability of these data to subclinical pediatric lead poisoning has yet to be established.