In a prospective randomised trial, children of hepatitis B carriers who were 3 months to 11 years of age and negative for all hepatitis B viral markers were randomised to receive (a) two doses of 5 micrograms of hepatitis B recombinant DNA yeast vaccine (H-B-VAX II) at 0 and 1 month (Group I), (b) three doses of 5 micrograms of H-B-VAX II at 0, 1 and 6 months (Group II) and (c) three doses of 10 micrograms of plasma-derived hepatitis B vaccine (H-B-VAX) (Group III). In the 141 children receiving H-B-VAX II, i.e. in Group I and Group II, side-effects were few and transient. Fever occurred in 10% of the vaccinees. Anti-HBs response rate was 96-100% by 8 months. None of the vaccinees became positive for anti-HBc. The geometric mean of the anti-HBs titre at 8 months was higher for Group II than for Group I (P less than 0.001). The practical and theoretical necessity for this higher titre of anti-HBs in the prevention of future infection by the hepatitis B virus is discussed. Both the two-dose regime and the three-dose regime of 5 micrograms of H-B-VAX II were safe, effective and immunogenic in children.