Objective. Currently, the methods for assessing milk ejection in women include serial sampling of plasma oxytocin and measurement of intraductal pressure, both of which are invasive and may induce stress. We hypothesized that milk ejection would cause an increase in milk-duct diameter that could be observed non-invasively with ultrasound, and this could be used to investigate the physiology of milk ejection in women.Methods. One milk duct was scanned in the unsuckled breast in 2 groups of mothers: group BB (n=21) for the beginning of a breastfeed and group EB (n=24) for the entire breastfeed. A duct also was monitored for a 5-minute period on 2 separate days in the absence of factors that may induce milk ejection in group EB to provide a baseline duct diameter. Milk intake at a breastfeed was measured by test weighing.Results. A significant increase in milk-duct diameter was observed when milk ejection was sensed and/or the infant changed its swallowing pattern in both groups. Multiple increases and decreases (mean:2.5 per breastfeed; standard deviation: 1.5; n=62) in duct diameter occurred in group EB. Duct diameter remained relatively stable between breastfeeds (coefficient of variation: 1.4%-8.3%). Infant milk intake was positively related to the number of milk ejections (r(2)=.365; n=57).Conclusions. Ultrasound is an objective, noninvasive technique for detecting milk ejection by observing an increase in milk-duct diameter. However, this technique requires an experienced ultrasonographer, adequate imaging time, and surroundings conducive to breastfeeding. Multiple milk ejections were common during breastfeeding, although they were not sensed by mothers. The number of milk ejections influenced the amount of milk the infant consumed.