To assess the safety, immunogenicity, and lot stability of the whole cell/recombinant B subunit cholera vaccine, 2 lots manufactured in June 1991 and February 1992 were tested in January 1995. Two oral doses of vaccine or placebo given 2 weeks apart were given with buffer to 216 Peruvian adults and children. Symptoms were elicited for 3 days after each dose. Serum and plasma specimens obtained from each volunteer before vaccination and 10-14 days after the second dose were tested for vibriocidal and anti-cholera toxin antibodies. The vaccine was well-tolerated. Nearly half of the 100 vaccinees had pre-vaccination vibriocidal titers > or = 1:40. Elevated titers were observed in 22% of 37 children 2-5 years of age compared with 66% of 63 vaccinees 6-65 years (P < 0.001). A > or =2-fold serum vibriocidal response was observed in 55% of 100 vaccinees and 6% of 32 placebo recipients. An elevated pre-vaccination titer (< or =1:40) did not change the proportion of vaccinees who responded with a > or =2-fold increase in vibriocidal titer (51% versus 59%, difference not significant), but did change the proportion responding with a > or =4-fold increase (41% versus 22%; P < 0.05). The vibriocidal seroconversion rate was lowest in children 2-5 years old despite low pre-vaccination titers. Two-fold or greater serum antitoxic responses in IgA and IgG were observed in >90% of the vaccinees; > or =4-fold responses were seen in 65-70% of the vaccinees with a 6-8-fold increase over baseline. Plasma specimens were as good as sera for determining anti-toxic antibodies by ELISA, but were less satisfactory for determining vibriocidal antibody titers.