Concentration and isotopic composition (delta(13)C and delta(18)O) of ambient CO2 and water vapour were determined within a Quercus petraea canopy, Northumberland, UK. From continuous measurements made across a 36-h period from three heights within the forest canopy, we generated mixing lines (Keeling plots) for delta(a) (CO2)-C-13, delta(a) (COO)-O-18-O-16 and delta(a) (H2O)-O-18, to derive the isotopic composition of the signal being released from forest to atmosphere. These were compared directly with measurements of different respective pools within the forest system, i.e. delta(13)C of organic matter input for delta(a) (CO2)-C-13 delta(18)O Of exchangeable water for delta(a) (COO)-O-18-O-16 and transpired water vapour for delta(a) (H2O)-O-18. [CO2] and delta(a) (CO2)-C-13 showed strong coupling, where the released CO2 was, on average, 4 per mil enriched compared to the organic matter of plant material in the system? suggesting either fractionation of organic material before eventual release as soil-respired CO2, or temporal differences in ecosystem discrimination. delta(a) (COO)-O-18-O-16 was less well coupled to [CO2], probably due to the heterogeneity and transient nature of water pools (soil, leaf and moss) within the forest. Similarly, delta(a) (H2O)-O-18 was less coupled to [H2O], again reflecting the transient nature of water transpired to the forest, seen as uncoupling during times of large changes in vapour pressure deficit. The delta(18)O of transpired water vapour, inferred from both mixing lines at the canopy scale and direct measurement at the leaf level, approximated that of source water, confirming that an isotopic steady state held for the forest integrated over the daily cycle. This demonstrates that isotopic coupling of CO2 and water vapour within a forest canopy will depend on absolute differences in the isotopic composition of the respective pools involved in exchange and on the stability of each of these pools with time.