Milk accumulation in the udder decreases milk secretion and this effect is explained as well by the effects of the quantity of milk stored in the udder as by the duration and repetition of periods of milk stasis. This experiment aimed to better understand the underlying mechanisms of decreased milk yield in response to the specific effects of the quantity of milk stored in the udder, independent from storage duration, on milk yield and composition. Sixteen Holstein cows were assigned to 4 blocks of 4 cows in a 4 × 4 Latin square design using 7-d periods, with a 4-d sampling period and a 3-d washout period. Cows were milked twice daily at approximately 0700 and 1630 h throughout the trial. Treatments consisted of 4 degrees of milk removal (100, 70, 40, and 0%) applied at one morning milking, designated M0. Effects of the quantity accumulated were studied in relation to udder distension, via measurements of the total distance between the ends of the 4 teats, and cisternal capacity, via the evaluation of cisternal area by ultrasonographic scan at 1 and 9 h after M0. The effect of the quantity accumulated was also evaluated in relation to mammary epithelium permeability by determining plasma lactose concentrations 1 h before and 4, 7, and 10 h after M0. Leaving milk in the udder at M0 decreased milk production during the M0-M1 interval in a negative curvilinear manner. As a result, M0+M1 milk yield decreased or tended to decrease significantly by -1.3, -5.3, and -12.8 kg for the 70, 40, and 0% treatments compared with the 100% treatment (41.7 ± 1.26 kg/d), respectively. Negative carry-over effects on milk yield were observed until the M3 milking only for the 40 and 0% treatments, and no differences were observed between the effects of these treatments. The total distance between teats increased significantly but to decreasing degrees during the M0-M1 interval. For the 40 and 0% treatments, cisternal area, which was increased 1 h after M0 milking, exhibited no further increase during the M0-M1 interval, suggesting cisternal distension was close to maximum. Simultaneously, lactose concentrations increased in blood plasma for only these 2 treatments, and this increase occurred earlier for the 0% treatment. It was also observed that cows presenting the earliest increases in plasma lactose concentrations during milk accumulation lost more milk in response to extended milking intervals.