In October 1988, a project was implemented for assessing the malaria chemoprophylactic efficacy of weekly chloroquine (CQ) and daily proguanil (PROG) during pregnancy in Muheza-Tanzania. Resultant CQ and PROG-cohorts of infants were followed up for prompt diagnosis and treatment of malaria. Infections were primarily treated with 25 mg base amodiaquine/kg over 3 days. By September 1990, 49 and 60 infants from PROG and CQ cohorts respectively had completed one year follow up. Thirty-five (71%) infants of PROG and 44 (73%) for CQ-cohort were infected with malaria before 3 months of age. The one year mean infection episode rates were 7 (PROG-cohort) and 6.6 (CQ-cohort). Amodiaquine cleared 209 (80%) of PROG's total infections and 224 (81%) for CQ-cohort, and significantly reduced the infection load among clearance failures. Clearance failures had high pre-treatment parasite densities whilst post-treatment densities were higher in the CQ-cohort than PROG-cohort. Low malaria immunity and chloroquine's long residence time could explain these differences. We conclude that early infancy malaria is common and should always be suspected, looked for and adequately treated. Amodiaquine is better than chloroquine for malaria primary therapy during infancy and early childhood.