Figure 2. (A, B) Representative trace of the percentage of embolized xylem area (blue) and stem water potential (Ψstem) (black) plotted against time in a branch of the angiosperm Rhododendron hirsutum . The vertical dashed lines indicate the period between rehydration and the point at which the Ψstem on the second dehydration had declined to the Ψstem when the branch was first rehydrated. (B) The relationship between the area (%) of pre-existing embolism at rehydration and the relative area (%) of xylem embolised on the second cycle of dehydration occurring at a more hydrated Ψ than when the branch was rehydrated. An exponential relationship is plotted (P<0.0001, R2=0.69). A mean (n = 3, ± SE) is shown for T. californica . All traces used to assemble this dataset are included in Supplementary Figure S1. Species abbreviations are Tc T. californica ; Ar Agathis robusta ;Ts.c Tsuga canadensis ; Dw Drimys winteri ; Rh R. hirsutum ; Iv Ilex verticillata ; Fr Ficus religiosa ;Ti.c Tilia cordata ; Lb Lindera benzoin.
When branches, regardless of whether they had xylem comprised of vessels or only tracheids, were rehydrated at a Ψ more hydrated thanP 50, there were very few if any embolism events observed on the second cycle of dehydration at a Ψ more hydrated than when the stem was rehydrated (Figure 2B; Supplementary Figure S2). Once xylem had experienced around 60% of area embolized then an increasing percentage of the remaining xylem on the second cycle of dehydration was observed to embolize at a Ψ that was more hydrated than when the stem was rehydrated (Figure 2B; Figure 3). In three individual stems ofR. hirsutum that were rehydrated at different percentages of embolized xylem area we found on the second cycle of dehydration that there was little or no embolism formation (1.5% or 0% of total xylem area) prior to stem dehydration to the Ψ at which rehydration occurred when either 10.4% or 49.5% of total xylem area was embolized, respectively (Figure 3). This contrasts with a branch that was rehydrated when 81.7% of xylem area had embolized, in this branch on the second cycle of dehydration embolism began to form at -0.56 MPa (initial embolism events on the first dehydration occurred at a mean Ψ of -0.96 ±0.05 (SE) MPa) and continued to accumulate linearly with declining Ψ (with a slope of 3.7% remaining embolized xylem area bar-1) such that when the stem had reached the Ψ at which rehydration occurred more than 56% of the remaining xylem area had embolized (equivalent to 10% of the total xylem area of the stem) (Figure 3).