Seedlings of Betula pendula Roth were grown with their root systems separated between two soil compartments. Four treatments were imposed: (i) adequate irrigation in both compartments (WW, controls); (ii) adequate irrigation in one compartment and drought in the other compartment (WD); (iii) drought in both compartments (DD); and (iv) half of the root system severed and the remainder kept well-watered (root excision, RE). Predawn leaf water potential, stomatal conductance, soil-to-leaf specific hydraulic conductance, and root and leaf growth decreased in DD-treated seedlings, which also displayed severe leaf shedding (30% loss in leaf area). The DD treatment also resulted in increased concentrations of abscisic acid (ABA) and its glucose ester in the xylem sap of roots and shoots compared to concentrations in control seedlings (about 200 versus 20 nM). Despite the difference in xylem sap concentrations, total ABA flux to the shoots was similar in the two treatments (1–2 pmol ABA m–2 leaf area s–1) as a result of reduced transpiration in the DD-treated seedlings. Compared with root growth in control plants, root growth increased in the RE-treated plants and decreased in the drying compartment of the WD treatment; however, the RE and WD treatments only slightly reduced leaf expansion, and had no detectable effects on shoot water relations or ABA concentrations of the root and shoot xylem sap. We conclude that short-term soil water depletion affecting only 50% of the root system does not cause a measurable stress response in birch shoots, despite root growth cessation in the fraction of drying soil.