The response of dairy cows to omitting one milking weekly was investigated in two successive experiments conducted with Holstein cows milked twice daily. Experiment 1 compared the lactational performances and udder changes in eight cows (31.2 L/d, 201 d in milk) in the 5 wk before and the 5 wk after introducing the suppression of one milking weekly. Milk yield was recorded daily and milk composition twice weekly. Milk partitioning in the udder (alveolar and cisternal milk) and cisternal size (ultrasonography), 8 h after milking, were also measured at the start and the end of the experiment. Although daily milk yield decreased 32% during the experiment (10 wk), linear regression analysis revealed a loss of milk yield of 1.1 L/d (3.7%) as a consequence of the omission of one milking weekly. Milk composition, lactation persistency, and somatic cell count (SCC) were unaffected by milking omission. Milk partitioning in the udder decreased by 38% in alveolar milk volume and showed a tendency to decrease in cisternal milk volume (15%) and cisternal size (7%), as a result of milking omission and advancing lactation. Loss in total milk yield was negatively related with cisternal milk volume (r = -0.77) and cisternal size (r = -0.70) indicating smaller losses in the udders with large cisterns. In Experiment 2, five cows (21.0 L/d, 227 d in milk) previously adapted to the milking omission schedule were used to study the daily effects of milking omission on milk yield, milk composition and udder health during 10 wk. Milk yield and milk composition were approximately constant but SCC increased with lactation stage. The omission of one milking caused an important decrease in milk yield, fat content and SCC on the omission day and a compensatory increase over the following 2 d, but milk protein and lactose did not vary. All variables reached the average weekly value three days after the milking omission (six milkings). In conclusion, under the conditions used, omitting one milking weekly slightly reduced milk yield and did not affect milk composition when healthy cows were used. Milk losses by milking omission depend on udder cistern characteristics; evaluating cistern size by ultrasonography may be a useful tool for choosing cows that are better adapted to a reduced milking frequency.