Search

· The 18O composition of plant cellulose is often used to reconstruct past climate and plant function. However, uncertainty remains regarding the estimation of the leaf sucrose 18O signal and its subsequent attenuation by 18O exchange with source water during cellulose synthesis.· We grew Lolium perenne at three CO2 concentrations (200, 400 or 800 mmol mol-1) and two relative humidity (RH) levels (50% or 75%), and determined 18O enrichment of leaf sucrose (Δ18OSucrose), bulk leaf water (Δ18OLW), leaf cellulose (Δ18OCellulose) and water at the site of cellulose synthesis (Δ18OCelSynW). · Δ18OCellulose correlated with Δ18OSucrose (R2=0.87) but not with Δ18OLW (R2=0.04), due to a variable 18O discrepancy (range 2.0-9.0‰) between sucrose synthesis water (Δ18OSucSynW, estimated from Δ18OSucrose) and bulk leaf water. The discrepancy resulted mainly from an RH effect. The proportion of oxygen in cellulose that exchanged with medium water during cellulose formation (pex), was near-constant when referenced to Δ18OSucSynW (pex-SucSynW = 0.52±0.02 SE), but varied when related to bulk leaf water (pex-LW = –0.01 to 0.46). · We conclude that previously reported RH-dependent variations of pex-LW in grasses are related to a discrepancy between Δ18OSucSynW and Δ18OLW that may result from spatial heterogeneity in 18O gradients of leaf water and photosynthetic sucrose synthesis.