Introduction
Xylem sap in plants is frequently transported under negative pressure (Dixon and Jolly, 1896; Jansen and Schenk 2015). Under conditions of low soil water content and/or high transpiration rates, the tensile force of xylem sap may increase considerably, which could lead to interruption of water transport in tracheary elements by large gas bubbles (embolism). Understanding the frequency and mechanism behind embolism formation in plant species is important because the amount of embolised conduits may affect the transport efficiency of water, and therefore photosynthesis (Zhu et al. , 2013; Martin‐StPaul et al. , 2017). There is strong and convincing evidence that drought-induced embolism formation occurs via bordered pits in cell walls of adjacent conduits (Zimmermann, 1983; Sperry & Tyree, 1988; Jansen et al. , 2018; Kaack et al. , 2019). It has frequently been assumed that once the pressure difference between sap-filled conduits (under negative pressure) and embolised ones (under atmospheric pressure) exceeds a certain threshold, embolism spreads from an embolised conduit to a neighbouring one via the mesoporous pit membranes of bordered pits (Choat et al. , 2008; Tixier et al. , 2014; Wason et al. , 2018; Avila et al. , submitted). Although embolism spreading from previously embolised conduits has been well presented in many textbooks and papers (Zimmermann, 1983; Crombie et al. , 1985; Choat et al. , 2016; Lamarque et al. , 2018), various basic questions about this process remain unclear (Kaack et al. , 2019). Since gas movement across pit membranes is based on two processes, namely mass flow and diffusion, we prefer the general term embolism spreading instead of air-seeding, which includes mass flow of gas across a pit membrane only.
An important question is whether spreading of embolism in xylem tissue is facilitated by the presence of pre-existing embolised conduits, which may occur in conduits from a previous growth ring or protoxylem (Kitinet al. 2004; Sano et al. , 2011; Hochberg et al. , 2016). If this would be correct, the mechanism behind embolism spreading may not dependent on xylem pressure only (Avila et al. , submitted). Embolised conduits could also occur when a herbivore or xylem feeding insect damages conduits, or when a plant organ experiences die-back, which may result in local embolism spreading. Artificial embolism spreading may occur when xylem tissue has been cut open to take embolism resistance measurements, because when a transpiring plant is cut in the air, the air-water meniscus is quickly pulled back into the conduit lumina until it stops at an interconduit pit membrane (Zimmermann, 1983). A widely used approach to evaluate embolism resistance is to measure the xylem water potential that corresponds to 50% loss of hydraulic conductance (Ψ50, MPa), while the xylem water potential corresponding to 50% of the total amount of gas that can be extracted from a dehydrated xylem tissue has been suggested as an alternative approach (Oliveira et al. , 2019, Pereira et al. , 2016, 2019). Both experimental approaches rely on cut plant organs either due to the requirements to measure hydraulic conductivity, or the gas diffusion kinetics of dehydrating samples. Moreover, dehydration of a cut branch or leaf can proceed much faster than dehydration of an intact plant (Cochard et al. , 2013; Hochberg et al. , 2017). Other methods, however, such as microCT observations and the optical method can be used to quantify embolism in a non-destructive way in intact plants (Brodribb et al. , 2016a, b; Choat et al. , 2016; Lamarque et al. , 2018).
The amount of embolism propagation could be limited by hydraulic segmentation of the conduit network, which represents a hydraulic constriction or bottleneck (Zimmermann, 1983; Tyree et al. , 1991; Levionnois et al. , 2020). In a broad sense, hydraulic segmentation has also been described as compartmentalisation, connectivity, sectoriality, or modularity, and may include narrow conduit dimensions and/or poorly interconnected conduits, which increase the resistance of the hydraulic pathway (Ellmore et al. , 2006; Loepfe et al. , 2007; Espino & Schenk, 2009).
In a few studies, considerable differences in embolism resistance have been reported between intact plants and xylem tissue in cut organs ofVitis vinifera and Laurus nobilis , with cut-open xylem potentially underestimating stem embolism resistance (Choat et al. , 2010, Torres-Ruiz et al. , 2015; Lamarque et al. , 2018). In a few species, however, the bench dehydration method, which is a widely applied method for hydraulic estimations of embolism resistance, was found to show no difference in embolism resistance between cut, dehydrating branches and dehydration of intact plants ofQuercus and Populus (Breda et al. , 1993; Tyreeet al. , 1992; Skelton et al. , 2018). While more species need to be studied to understand a possible artefact associated with embolism spreading from cut-open xylem, three explanations could be suggested for the observed discrepancy. First, it is possible that the cutting of conduits with sap under negative pressure introduces a cutting artefact, although artificially embolised conduits near stem ends can be removed before hydraulic measurements are made (Wheeleret al. , 2013; Torres-Ruiz et al. , 2015). A second explanation is that embolism spreading could be avoided by hydraulic segmentation, which may occur at the transition between organs, growth rings, and nodes (Sano et al. , 2011; Levionnois et al. , 2020). Indeed, vessels are known not to run completely randomly, but may end near nodes, side branches, stem-petiole transitions, and between the vascular bundles of the petiole and major veins (Salleo et al. , 1984; André et al. , 1999, André, 2005, Wolfe et al. , 2016). Thirdly, embolism may also occur in conduits that are not connected to embolised conduits, although the frequency of such de novo process is unclear (Brodersen et al. , 2013; Choat et al. , 2015, 2016).
In this paper, we aim to test to what extent cut-open angiosperm xylem has an effect on embolism spreading in leaves across a diverse selection of six temperate species. In the first experiment we aimed to investigate if embolism resistance of leaf xylem was affected by the proximity to cut-open conduits. We hypothesise that leaf xylem would be more vulnerable to embolism for detached leaves with a cut petiole compared to leaves attached to stem segments. However, not only the proximity to cut-open vessels, but also hydraulic segmentation at the stem-leaf, or the petiole-leaf transition could affect embolism spreading, and may prevent a potential artefact in measurements of embolism resistance near cut xylem tissue. We therefore included species with both deciduous and marcescent leaves (i.e. species that retain dead leaves on the plant), and diffuse porous and ring-porous wood, because hydraulic segmentation can be associated with leaf phenology and vessel dimensions. If pit membranes in bordered pits of vessels and tracheids would function as safety valves that avoid the spreading of embolism from embolised to functional conduits, it is possible that embolism spreading is reduced by the number of interconduit endwalls and/or the connectivity between conduits (Kaack et al. , 2019; Johnsonet al. , 2020). Xylem tissue that shows hydraulic segmentation could include many tracheids and/or narrow, fibriform vessels (Rančićet al. , 2010). Species that show little or no hydraulic segmentation, may not have these safety valves. Removal of leaves in seedlings of the ring-porous species Quercus robur , for instance, was found to result in potential embolism formation in the stem based on microCT observations (Choat et al. , 2016).
Whether or not embolism spreading depends directly on vessel dimensions was tested in a second experiment by cutting minor leaf veins. Drought-induced embolism is frequently reported to initiate in large vessels, while narrow and short vessels or tracheids embolise typically later at lower xylem water potentials (Scoffoni et al. , 2016; Klepsch et al. , 2018). These observations may give the impression that wide conduits are more vulnerable to embolism, although any functional explanation for such differential embolism resistance remains unclear. Indeed, pit membrane thickness, which is a major determinant of vulnerability to embolism (Li et al. , 2016; Kaack et al. , 2019), is not related to conduit dimensions (Klepsch et al. , 2018; Wu et al. , 2020; Kotowska et al. , 2020). If the proximity of a gas source would determine embolism spreading, we expect that narrow and short vessels near cut minor veins would embolise before embolism occurs in the large vessels of major veins, which would make narrow vessels seemingly more vulnerable than wide ones.
Finally, we applied a methodological comparison of embolism resistance in leaf xylem between the optical method and the pneumatic method. If the pneumatic method would be subject to a potential artefact due to gas extraction from intact vessels that are neighbouring or connected to embolised, cut conduits, this method could systematically underestimate embolism resistance compared to the optical method. The pneumatic method, which estimates the changing gas volume in intact vessels during dehydration, showed a good agreement with hydraulic methods applied to stem segments (Pereira et al. , 2016, Zhang et al. , 2018). While direct comparison of the pneumatic and optical method to detached leaves of Eucalyptus camaldulensis suggested no significant difference for this species (Pereira et al. , 2019), a larger number of species should be tested to generalise this finding.
The three complementary sets of experiments will contribute to a better understanding of the driving forces behind embolism spreading in xylem tissue.