Effect of pre-existing embolism on stem vulnerability
By tracking embolism formation through a cycle of dehydration followed by rehydration, and subsequent dehydration to complete desiccation we show that the presence of pre-existing embolism in the xylem can lower an apparent optical P 50. Given that the presence of embolism in conduits is the most likely source of air seeding for future embolism events should Ψ decline (Knipfer et al. 2015; Choat et al. 2015b; Torres-Ruiz et al. 2016), we suspected that the presence of pre-existing embolism would make xylem more vulnerable. Across our diverse sample of species, the more embolism initially present in the xylem, up to 60% embolism by area, the more negative the apparentP50 of the remaining conduits became (Figure 4). Nevertheless, a temporal effect on embolism spreading can also be interpreted to contribute to the more negativeP 50r values, as the slope of vulnerability curves may illustrate the relative speed of embolism propagation, rather than absolute embolism resistance per se. Hence, it would be interesting to compare normal vulnerability curves with curves subject to multiple cycles of rehydration and dehydration, which is likely to occur to plants in the field.
The impact of pre-existing embolism on the reliability of vulnerability curves is a widely discussed topic (Martin-StPaul et al. 2014). So much so that flushing stems to remove embolism prior to measurements of hydraulic conductivity on centrifuge rotors is common practice. One explanation given for this practice is to counter the chance that pre-existing embolism spreads rapidly through the conduits resulting in artefacts of highly vulnerable P 50 values in a stem (Cochard et al. 2013; Hacke et al. 2015). Our results suggest that the presence of pre-existing embolism will occur in the most vulnerable conduits first, thus P 50 values, determined optically on xylem in which some conduits are already embolized, may appear more resistant to embolism formation than if measurements were conducted on xylem in which all of the conduits were water-filled. This mirrors the results of a report of a comparison of vulnerability curves measured using a centrifuge in short stem segments of long-vessel species that had and had not been flushed (Hacke et al. 2015). While we found that pre-existing embolism altered apparentP 50 in some species, the impact on apparentP 50 was generally a reduction of no more than 20%. In this study we never measured an R -shaped curve using the optical method, despite initiating all experiments at high water potentials.
Besides accurate water potential measurements at the individual conduit level, a key unknown in our study is whether the differences in apparent embolism resistance driven by pre-existing embolism, translate to similarities in hydraulic vulnerability or drought tolerance. Given the strong correlations between mean embolism resistance of xylem between optical, microCT, pneumatic and hydraulic methods, when each is performed carefully (Sergent et al. 2020; Gauthey et al. 2020; Pereira et al. 2021), we hypothesize that there would be strong congruence with hydraulic vulnerability. However, given the considerable declines in hydraulic conductivity that would occur when a substantial percentage of xylem is embolized (Brodribb and Cochard 2009), the effect of this reduction, without reduced canopy transpiration, would accelerate declines in Ψ during drought, and may not enhance survival.