Figure 3. Cumulative embolized xylem area as a percentage of total embolized area on the first cycle of dehydration (blue) in three stems of Rhododendron hirsutum that were rehydrated at varying degrees of embolized xylem area. The water potential at rehydration is marked by a vertical dashed line. Cumulative embolized xylem area as a percentage of the remaining un-embolized xylem area on the second cycle of dehydration (red) is also shown. Temporal traces of water potential and embolism accumulation through time can be found in Figure 2, Supplementary Figure S1F and S1G. Similar traces for all other species are found in Supplementary Figure S2.
Differences across species in the Ψ that triggered embolism during a second cycle of dehydration had a considerable impact on whether pre-existing embolism in the xylem altered apparentP50 . In many cases, pre-existing embolism caused a minor increase in apparent embolism resistance, resulting in more negative values of apparent P50(P 50r), which was P50calculated from data collected only from embolism events that occurred on the second cycle of dehydration (Figure 4). We did note that in angiosperms the higher the percentage of pre-existing embolism in the xylem, the higher the percent change in apparentP 50, with P 50rparticularly differing from P50 changing once 40% of xylem was embolized (Figure 4). In conifers, the degree of pre-existing embolism did not appear to alter apparentP50 as much as in angiosperm species (Figure 4). In some conifer stems, with pre-existing embolism, apparentP 50 was slightly more negative, while in others apparent P 50 was less negative or even unchanged, irrespective of the degree of pre-existing embolism in the xylem (Supplementary Figure S1).