1. Preincubation of luteal membranes with human choriogonadotropin results in the formation of an activated state of adenylate cyclase which is not reversed by washing and which is limited only by the absence of guanine nucleotides, whereas preincubation with GTP yields only a partially activated adenylate cyclase which requires the presence of both GTP and human choriogonadotropin during assay to demonstrate maximal activity. 2. Preincubation of luteal membranes with GTP and human choriogonadotropin does not lead to a synergistic increase in wash-resistant activity. 3. Luteal membranes that had been preincubated with GTP and hormone exhibited a decreasing rate of cyclic AMP synthesis during the adenylate cyclase assay incubation; addition of GTP during the assay incubation reversed the decrease. 4. Membranes that had been preincubated in the absence of guanine nucleotide and hormone showed a `burst' phase of cyclic AMP synthesis when GTP was present in the assay incubation and a `lag' phase with p[NH]ppG (guanosine 5′-[β,γ-imido]triphosphate) present in the assay. The presence of human choriogonadotropin with either nucleotide in the assay incubation eliminated the curvatures in plots observed with guanine nucleotides alone. 5. Luteal adenylate cyclase was persistently activated by preincubation with p[NH]ppG alone or in combination with human choriogonadotropin; the activation caused by p[NH]ppG alone was still increasing after 70min of preincubation, whereas that caused by p[NH]ppG in the presence of hormone was essentially complete within 10min of preincubation. 6. Luteal adenylate cyclase that had been partially preactivated by preincubation with p[NH]ppG was slightly increased in activity by the inclusion of further p[NH]ppG in the adenylate cyclase assay incubation, but more so with p[NH]ppG and hormone. Human choriogonadotropin alone caused no further increase in the activity of the partially stimulated preparation unless p[NH]ppG was also added to the assay incubation. 7. GTP decreased the activity of adenylate cyclase in membranes that had been partially preactivated in the presence of p[NH]ppG; the decrease in activity was greater when GTP and hormone were present simultaneously in the assay. 8. The results indicate that stable activation states of adenylate cyclase can be induced by preincubation of luteal membranes in vitro with human choriogonadotropin or p[NH]ppG, and that in the presence of p[NH]ppG the hormone may accelerate events subsequent to guanine nucleotide binding. Stable activation of luteal adenylate cyclase by prior exposure to GTP is not achieved. The involvement of GTPase activity and of hormone-promoted guanine nucleotide exchange in the modulation of luteal adenylate cyclase activity is discussed.