Objective: To determine if supplementation with corn-soya blend (CSB) or lipidbased nutrient supplement (LNS) improved the weight gain of moderately underweight infants and children when provided through the national health service. Design: A randomised, controlled, assessor-blinded clinical trial. Infants and children were randomised to receive for 12 weeks an average daily ration of 71 g CSB or 43 g LNS, providing 1188 kJ and 920 kJ, respectively, or no supplement (control). Main outcome was weight gain. Secondary outcomes included changes in anthropometric indices and incidence of serious adverse events. Intentionto-treat analyses were used. Setting: Kukalanga, Koche, Katema and Jalasi health centres in Mangochi District, rural Malawi. Subjects: Underweight (weight-for-age Z-score ,22) infants and children aged 6-15 months (n 299). Results: Mean weight gain was 630 g, 680 g and 750 g in control, CSB and LNS groups, respectively (P 5 0?21). When adjusted for baseline age, children receiving LNS gained on average 90 g more weight (P 5 0?185) and their weight-for-length Z-score increased 0?22 more (P 5 0?049) compared with those receiving no supplementation. No statistically significant differences were observed between the CSB and control groups in mean weight and length gain. Conclusions: LNS supplementation provided during the lean season via through the national health service was associated with a modest increase in weight. However, the effect size was lower than that previously reported under more controlled research settings.