In fission yeast, meiosis is initiated by transcriptional activation of the mei3+ gene under the combined influence of the four mating type genes. The mei3+ gene product acts as a meiotic inducer by binding to and inhibiting the ran1+ protein kinase. Inactivation of ran1+ kinase is both necessary and sufficient to allow meiotic differentiation. We describe a class of mutants which are unable to undergo both normal meiosis and meiosis induced by inactivation of ran1+. In addition to these defects, the cells are sterile and unable to enter stationary phase. We have determined that the mutants define two complementation groups, designated cgs1+ and cgs2+ (continues to grow in stationary). The wild type allele of each gene has been isolated and sequence analysis of cgs1+ shows that it encodes a protein homologous to the regulatory subunit of cyclic AMP dependent protein kinase (cAPK). Biochemical studies demonstrate that in cgs1-1 containing cells, cAPK activity is unregulated by cyclic AMP (cAMP). Sequence analysis of cgs2+ shows that the predicted protein it encodes shares homology with a phosphodiesterase from Dictyostelium discoideum and biochemical studies demonstrate that cells containing a mutant allele of cgs2+ have elevated levels of cAMP. Thus, both genes encode proteins that regulate the activity of cAPK. We have previously shown that cells overproducing ran1+ kinase are meiotically defective. Here, we provide direct evidence that the meiotic defect caused by either unregulated cAPK activity or unregulated ran1+ kinase activity is due to inability to induce transcription of the mei2+ gene, which is required for meiotic initiation. We propose that the switch from vegetative growth to meiosis in fission yeast requires inactivation of ran1+ kinase and is prevented by unregulated levels of cAPK.