The comments of Heppner and colleagues and Dr. Barry are useful and contribute to a dialogue on the important subject of malaria prophylaxis.

We appreciate Heppner and colleagues' emphasis on points that we also emphasized: the risk of giving primaquine to G6PD-deficient persons and the higher efficacy of the drug for P. falciparum than for P. vivax. We also appreciate the comments on the question of mefloquine in pregnancy. When we stated that "the risk of mefloquine prophylaxis  ... has not been clinically evaluated," we might have added the word "thoroughly" before "clinically evaluated." The risk has been and is being examined, and reasonable persons can disagree on the drug's safety in pregnancy (1, 2). We chose not to take sides and simply implied that final evaluation remains.

Follow-up time should ideally be sufficient to recognize both P. falciparum and P. vivax that survive in both the liver and blood. The mean of these time periods can be estimated to be 12 days (P. falciparum in the liver), more than 4 weeks (P. vivax in the liver), and up to 4 weeks (P. falciparum and P. vivax in the blood). The follow-up time that we used, 4 weeks, is routinely used (3) to rule out all events except relapse of P. vivax from the liver. In terms of lack of immunity, our study sample was identical to the U.S. traveling population, and our end point of clinical malaria (symptoms plus parasitemia) is identical to the presentation that, as Dr. Barry states, will bring a U.S. traveller to the doctor. We were interested in whether all low-level parasitemias in nonimmune persons do ultimately result in higher parasitemias and more symptoms. This question can be addressed only in a prospective study such as ours. A paragraph in the original manuscript, cut from the final version because of space constraints, stated: "Of the 32 patients in whom prophylaxis failed, 17 had asymptomatic parasitemia of at least 500 parasites/µL on 1 or more days before meeting the definition of failure (symptomatic parasitemia or >1500 parasites/µL). These 17 patients consisted of both P. falciparum patients plus 2 of the 6 P. vivax patients in the primaquine group, and 5 of the 11 P. falciparum patients plus 8 of the 13 P. vivax patients in the control group. In addition, 3 patients had demonstrated asymptomatic parasitemia on one occasion but became aparasitemic upon further follow-up. Parasite counts of 600, 715, and 730 per µL (all P. vivax) were recorded on weeks 4, 6, and 9, respectively, in placebo patients, but each of these three patients became aparasitemic 1 day later."