In free living conditions, 24 breastfed infants, aged 6 to 10 months, were given successively five experimental gruels to study the effect of energy density (ED) and sweetness (sweet taste) on energy intakes (EI). Four gruels (G0, G1, G9 and G20) were prepared with experimental flours which were composed of the same local ingredients and which contained different levels of sucrose. The fifth gruel (GC) was prepared with an industrial flour. G0 had an average ED of 45 kcal/100 g (189 kJ/100 g) and the other gruels an average ED of 110 kcal/100 g (461 kJ/100 g). Although the sugar contents of the flours were 1% for G1, 9% for G0 and G9 and 20% for G20, because of flour composition and gruel dry matter content, the gruel G1 had the same sweetness as G0, G20 the same sweetness as GC and G9 a sweetness between that of G1 and G20. The results show that the amounts of G0 consumed were significantly higher than those of high ED gruels (7.84 for G0 vs 6.12, 5.63, 4.46, 4.72 g/kg body weight/meal, respectively for G20, G9, G1 and GC, P < 0.05). However, EI from high ED gruels were significantly (P < 0.001) higher than those from G0 (6.65, 6.10, 4.86, 4.83 kcal/kg/meal, respectively for G20, G9, G1 and GC vs 3.46 for G0). Energy intakes from G9 and G20 gruels were not significantly different but were significantly higher than those from GC and G1 (P < 0.001). So, consumption of sweet gruels with high ED and composed of local ingredients increased, at least by 76%, the EI from gruels in comparison with those from low ED gruels, but the amounts consumed by the infants remained too low to cover more than 15% of their daily total energy needs.