We investigated if concentrations of abscisic acid (ABA) and other solutes measured in the first few droplets of xylem sap from detopped root systems, are good estimates of those in the transpiration stream as it enters the shoot-base of whole plants. Xylem sap from root systems of pot-grown tomato plants (Lycopersicon esculentum Mill., cv. Ailsa Craig), at the seven-leaf stage, was obtained by placing root systems in chambers pressurized to 0.3 MPa with air. The first sample was taken from the cut-surface of the hypo-cotyl stump within 30 s of removing the shoot. ABA, sucrose and other osmolytes were more concentrated in the initial 100–200 mm3 of xylem sap than in subsequent samples. This suggested the sap was contaminated and not unchanged transpiration fluid. The effect was reproduced on the same plant, several times, by recutting the hypocotyl prior to reassembling the sap collecting set-up and repressurizing. Similar results were found with castor-oil plants (Ricinus communis L., cv. Gibsonii). However, neither release of ABA from the cut surface of the hypocotyl stump, nor the effects of pressure to the roots caused the contamination. Instead, small radial pressures exerted by a rubber sleeve attached to the hypocotyl stump, for collecting the sap, were responsible. The effect was reproduced by lightly squeezing the hypocotyl by hand. The possibility was examined that reliable estimates of ABA concentrations in transpiration stream fluid may be obtained from sap samples taken immediately after rejecting the initial, contaminated 200 mm3. However, ABA concentrations in these later samples were also unsatisfactory since they changed with rate of sap flow. The problem may be overcome by analysing sap induced to flow through detached root systems at rates close to those of whole-plant transpiration.