The ¹⁸O enrichment (Δ ¹⁸O) of leaf water affects the Δ ¹⁸O of photosynthetic products such as sucrose, generating an isotopic archive of plant function and past climate. However, uncertainty remains regarding how leaf water compartmentation between photosynthetic and non-photosynthetic tissue affects the relationship between Δ ¹⁸O of bulk leaf water (Δ ¹⁸O _LW) and leaf sucrose (Δ ¹⁸O _Sucrose). We grew Lolium perenne (a C ₃ grass) in mesocosm-scale, replicated experiments with daytime relative humidity (RH 50 or 75%) and CO ₂ level (200, 400 or 800 μmol mol ^-1) as factors, and determined Δ ¹⁸O _LW, Δ ¹⁸O _Sucrose and morpho-physiological leaf parameters, including transpiration ( E _leaf), stomatal conductance ( g _s) and mesophyll conductance to CO ₂ ( g _m). The Δ ¹⁸O of photosynthetic medium water (Δ ¹⁸O _SSW) was estimated from Δ ¹⁸O _Sucrose and the equilibrium fractionation between water and carbonyl groups (ε _bio). Δ ¹⁸O _SSW was well predicted by theoretical estimates of leaf water at the evaporative site (Δ ¹⁸O _e) with adjustments that correlated with gas exchange parameters ( g _s or total conductance to CO ₂). Isotopic mass balance and published work indicated that non-photosynthetic tissue water was a large fraction (~0.53) of bulk leaf water. Δ ¹⁸O _LW was a poor proxy for Δ ¹⁸O _Sucrose, mainly due to opposite Δ ¹⁸O responses of non-photosynthetic tissue water (Δ ¹⁸O _non-SSW) relative to Δ ¹⁸O _SSW, driven by atmospheric conditions.