AbstractIsotopic signatures of xylem water in different tree compartments such as roots, boles and branches, differ due to physiological and physical processes occurring inside trees. Accordingly, we hypothesised that the extent of such differences among the isotopic signatures of tree compartments is coherent with the distance travelled by the water inside trees and to its residence time. To test this, we compared the O–H isotopic composition of xylem water collected using an in‐situ water extraction method from roots, boles and branches of Fagus sylvatica trees growing on three geomorphological units of the Weierbach experimental catchment, Luxembourg. There was progressive 18O and 2H enrichment in xylem waters along the root–branch flow path for all the studied trees. Three explanations could be considered for this progressive enrichment: internal fractionation by xylem–phloem water exchange, chemical reactions of metabolic pathways and variable ages of water retained in the xylem, reflecting historical variation in isotopic composition of uptake water. Support for the hypothesis of isotopic fractionation linked to xylem–phloem water exchange and chemical reactions is that enrichment was generally consistent with the distance travelled by the water and to its residence time inside the trees. However, the relative enrichment of 2H and 18O was not consistent along the flow path, with Δ2H/Δ18O ~7.5 from the soil into the roots and bole and ~4.7–6.5 for pathways that included smaller branches. This contrast suggests different processes controlling above‐ground isotopic enrichment. In particular, the slope of ~7.5 in the lower tree is also consistent with variation in tree water uptake varying along the local meteoric water line, with water in the roots being closer to the composition of rainfall close to the time of sampling and water in the bole being closer to the composition of rainfall from the previous summer. The timing of root and bole sampling in early spring, just before leaf‐out, means sap flow was very slow and makes it plausible that varying‐age water was present in the tree at that time. We also compared the O–H isotopic composition of those samples with the ones of the potential water sources to identify the origin of the water uptaken. The latter varied during the 3 years of sampling, with a preferential uptake from near‐surface waters. Our results suggest multiple biochemical, physiological and physical processes may play fundamental roles in the isotopic composition of xylem water within trees.