Two toxins from Bordetella pertussis, pertussis toxin and Bordetella adenylate cyclase, cause profound disruptions of cAMP metabolism in mammalian cells. While the role of each toxin in the whooping cough syndrome is unknown, it is highly likely that together they confer on the organism an important proliferative advantage. Intact B. pertussis cells express large amounts of adenylate cyclase activity on their exterior surface. In a presently unknown fashion, this enzyme can enter human phagocytes, elevate cellular cAMP and impair host defense. We reasoned that this unusual enzyme might serve to signal the presence of B. pertussis in nasopharyngeal swabs from infected persons. Here we report a series of in vitro experiments which confirm the feasibility of such an approach. We find that calcium alginate swabs containing as few as 100 B. pertussis organisms produce readily detectable amounts of cAMP in our assay. Nasopharyngeal secretions swabbed from healthy volunteers, however produced no detectable cAMP alone and did not affect the production of cAMP by B. pertussis. We have also tested pure cultures of four common bacterial pathogens (Haemophilus influenzae, Streptococcus pyogenes, Staphylococcus aureus, and Escherichia coli) which may be co-isolated in clinical whooping cough and find that none interferes with the detection of B. pertussis. We conclude that the unique adenylate cyclase of B. pertussis might be a valuable diagnostic device.