Figure 1. The relationship between cumulative embolised xylem area in stems and stem water potential (Ψstem) during a cycle of dehydration, rehydration and dehydration in the angiospermIlex verticillata (A and B) and the gymnosperm Agathis robusta (C and D), with vessel and tracheid-based xylem, respectively. In A and C the solid red lines indicate Ψstem at which 50% of the xylem area was embolized (P 50) on the initial dehydration, and the Ψstem at which 50% of the remaining stem xylem area was embolized (P 50r) on the second dehydration. The dashed red line represents the beginning of second period of dehydration. The solid blue line represents the embolized area when rehydration occurred. In B and D cumulative embolized xylem area as a percentage of total embolized area (blue) on the first cycle of dehydration is shown alongside the cumulative embolized xylem area as a percentage of remaining un-embolized xylem area on the second cycle of dehydration (red). Horizontal red lines indicate when 50% of the total or remaining xylem area was embolized. Temporal traces of water potential and embolism accumulation through time can be found in Supplementary Figures S1A and S1B.
On the second period of dehydration, embolism occurred in the remaining conduits that had not embolised prior to rehydration. Differences were observed in the Ψ at which these embolism events first formed on the second dehydration that influenced the apparent embolism resistance of the remaining water-filled xylem conduits (Supplementary Figure S1). In most species embolism events on the second dehydration often did not form until Ψ had declined to values similar to those measured when the stem was rehydrated (Figure 1B; Supplementary Figure S2). In I. verticillata the impact of a lack of embolism formation at relatively high Ψ compared to when the first conduits embolized when initially dehydrated, meant that apparent P50(P50r ) was 0.69 MPa lower in the stem with 50% of xylem embolized, than P50 in the stem when no xylem was embolized (Figure 1A and B). In contrast in the conifer species assessed, embolism was often observed to form at the Ψ at which embolism initially started to occur on the first dehydration (Figure 1C and D; Supplementary Figure S1). In A. robusta this resulted in an apparent P50r that was only 0.06 MPa lower in the stem with 50% of xylem embolized, than P50in the stem with no pre-exisiting embolism (Figure 1C and D).
Differences between species in the nature of xylem embolism formation during rehydration and on the second cycle of dehydration were also apparent when cumulative embolism observations were plotted against time (Figure 2). In a representative trace from a stem of the angiosperm species Rhododendron hirsutum with vessel-based xylem, embolism first started to occur once Ψ reached -1.0 MPa, and progressively continued until -1.54 MPa, when approximately 50% of the total xylem area was embolized (Figure 2A). After rehydration, which occurred at -1.54 MPa, there was a complete and fast cessation of embolism accumulation (Figure 2A). During the second cycle of dehydration, new embolism events were only observed to form in the xylem after Ψ had declined to -1.54 MPa, which occurred 10 h after rehydration (Figure 2A). There was a similar slope in the decline of water potential through time on the first and second dehydration, however, once embolism began to form on the second dehydration the slope of water potential decline over time declined.