Sixteen Friesian cows, 55-150 days postpartum, were allotted to two groups of eight cows, according to calving date, lactation number and daily milk yield, and assigned randomly to one of two diets in a crossover design experiment. The two experimental diets were concentrate containing 0 or 1.2% sodium carbonate, plus sorghum silage. The concentrate was offered individually, in two equal meals daily at 0.5 kg kg-1 of milk produced, and the sorghum silage was offered ad libitum once daily on a group basis. The experimental period lasted from 14 June to 8 August 1989, and the cows were housed in an open lot, without shade. Dry matter intake, actual milk yield, milk protein content and yield, as well as content of milk lactose and milk solids excluding fat were not significantly affected. In contrast, sodium carbonate supplementation increased milk fat content (4.17 vs. 3.74%; P < 0.01), milk fat yield (1.10 vs. 0.97 kg day-1; P < 0.05), 4% fat-corrected milk yield (27.0 vs. 24.9 kg day-1; P < 0.05) and milk total solids content (13.17 vs. 12.71%; P < 0.05). No significant differences were observed in blood plasma concentrations of glucose, total protein, urea, acetate, beta-hydroxybutyrate, free fatty acids, sodium and potassium. In contrast, the plasma concentration of triglycerides was higher (35.4 vs. 27.8 mg per 100 ml; P < 0.01) when the cows were fed on the diet containing sodium carbonate.