Abstract Dehydrated lucerne of low (L: 0.53), normal (N: 0.55) and high (H: 0.73) in vivo dry matter (DM) digestibility were treated with ammonia or urea to study the effects on in situ and pepsin-cellulase DM digestibilities, water solubility and nitrogen content (Experiments 1, 2, 4) and on cell wall composition and degradability (Experiment 3). (1) N lucerne was treated with 30 g NH 3 kg −1 DM for 1 to 12 weeks at 30°C and 2 to 6 days at 80°C; (2) L, N and H lucerne were treated with increasing ammonia levels: 15 to 100 g kg −1 DM for 3 weeks at 30°C and 4 days at 80°C; (3) L, N and H lucerne were treated with 60 g NH 3 kg −1 DM for 3 weeks at 30°C and 4 days at 80°C; (4) L, N and H lucerne were treated with 60 g urea kg −1 DM without addition of urease for 3 and 6 weeks at 30°C. All treatments were carried out at 40% humidity. In situ and pepsin-cellulase DM digestibilities increased significantly ( P < 0.05) with the duration of treatment (up to 3 weeks at 30°C and 4 days at 80°C) and with the level of ammonia ( P < 0.01) (up to 30 g kg −1 DM). The greatest improvements (similar at both temperatures) were for L, N and H of 7.3, 7.2 and 3.9 points for in situ and of 10.6, 11.3 and 6.3 points for cellulase digestibilities, respectively. Water solubility also increased with duration of treatment and level of ammonia ( P < 0.01) and was greater at 80°C than at 30°C. Urea treatment significantly improved ( P < 0.01) digestibilities and water solubility but the doubling of treatment duration had no influence. The degree of ureolysis was only 50 to 60%. Ammonia and urea treatments considerably increased ( P < 0.01) nitrogen content. Treatment with 60 g NH 3 kg −1 DM induced a decrease in ethanol insoluble residue content, which was significant ( P < 0.01 for L and N, P < 0.05 for H) at 80°C but not at 30°C, and was greater for L and N than for H (about 12 and 5 points, respectively). This decrease was essentially due to solubilisation of hemicelluloses (− 15%) and uronic acids (− 26%). Thus, at 30°C, the chemical solubility of the cell wall was lower than at 80°C for the same total increase in microbial degradation. This result indicates that other phenomena are involved, such as an increase in cell wall porosity and consequently improved accessibility of cell wall polysaccharides to glycolytic enzymes.