Introduction
Water is transported under negative pressure through a plant in a complex network of xylem cells (Dixon and Joly 1895). Being under negative pressure means the water column is at constant risk of interruption by large bubbles (Tyree and Sperry 1989). When tension increases, such as when soil water is limiting or when evaporative demand exceeds hydraulic conductivity, there is an increasing likelihood of embolism formation, which blocks the internal flow of water (Urli et al. 2013; Brodribb and Cochard 2009). The primary source of gas leading to embolism formation is believed to be air entrance through the nanoscale pores of pit membranes from neighbouring gas-filled conduits (Choat et al. 2015b; Tyree and Sperry 1989; Guan et al. 2021; Kaack et al. 2021). If water stress is not relieved, more xylem conduits will experience embolism, leading to progressive declines in hydraulic conductivity, and eventual failure of the hydraulic system, with consequences for photosynthetic performance and potential dieback of organs (Urli et al. 2013; Brodribb and Cochard 2009; Adams et al. 2017; Cardoso et al. 2020a).
A suite of adaptations have evolved to reduce the likelihood of embolism spread when hydrated xylem is exposed to negative pressure, with species native to seasonally dry environments having xylem that is highly resistant to embolism formation (Choat et al. 2012). These adaptations range from gross anatomical differences in xylem conduits, like reduced conduit diameter, length, interconnectivity, or increased cell-wall thickness (Hacke et al. 2001; Blackman et al. 2010; Jacobsen et al. 2019; Scoffoni et al. 2017b; Schumann et al. 2019), through to micro-anatomical variation such as increased pit membrane thickness (Li et al. 2016; Kaack et al. 2019), or in conifers, an increased torus overlap (Bouche et al. 2014). Considerable focus has been placed on establishing relationships between xylem anatomy and emergent traits that surmise embolism resistance, such as the water potential at which 50% of the xylem is embolised (P50 ) (Choat et al. 2018; Brodribb et al. 2020b). Less emphasis has been placed on understanding if individual xylem conduits have specific thresholds at which embolism will form (Jacobsen et al. 2019).
Assuming that most declines in xylem hydraulic conductance during drought are due to embolism then the gradual decline in hydraulic conductance as a drought progresses (Sergent et al. 2020), suggests that there might be a range of water potentials at which embolism events will occur across a population of xylem conduits. This range of water potential is typically quantified by the slope of vulnerability curves, with a narrow range and a steep slope characterising vulnerable species, and a wider range and flatter slope more embolism resistant species (Kaack et al. 2021). Methods that are capable of visualizing individual embolism events support the idea that individual embolism events will occur over a wide range of water potentials in many species in both stem and leaf xylem (Jacobsen et al. 2019; Venturas et al. 2016; Johnson et al. 2020; Knipfer et al. 2015). In stems the first embolism events are often observed near the primary xylem (Choat et al. 2015a), or in some cases the largest volume vessels (Jacobsen et al. 2019; Johnson et al. 2020; Knipfer et al. 2015), suggesting these conduits embolise first. In leaves, the first embolism events are almost always observed in the midrib and proceed, as leaf water potential declines, through the increasingly higher orders of veins (Skelton et al. 2017; Brodribb et al. 2016a; Scoffoni et al. 2017b).
The apparently wide range of water potentials at which embolism formation will occur could also be explained by a temporal aspect of embolism spread, because propagation is largely known to occur from an embolised to a non-embolised conduit, suggesting that pre-existing embolism or the availability of gas affects the actual spreading process (Guan et al. 2021; Wason et al. 2021). Once initial embolism has formed in xylem conduits, extrinsic factors beyond the anatomy of a single conduit, such as the proximity of water-filled conduits to embolised ones, cut-open conduits, and hydraulic segmentation can influence the likelihood of embolism formation (Choat et al. 2010; Knipfer et al. 2015; Choat et al. 2015b; Torres-Ruiz et al. 2016; Lamarque et al. 2018; Guan et al. 2021). Gas-filled conduits may act as a source of embolism propagation if a plant experiences a subsequent drought, thereby rendering neighbouring water-filled conduits more vulnerable. Pit membranes in the bordered pits between xylem conduits are believed to prevent this spread of air between conduits (Choat et al. 2008). In conifers with torus and margo pit membranes, aspiration of the torus prevents the spread of air from embolized tracheids into neighbouring sap-filled tracheids (Liese and Bauch 1967; Hacke et al. 2004; Pittermann et al. 2005). In angiosperms, the multiple pore constrictions in pit membranes with a given thickness fulfil a similar function, with the most narrow pore constriction determining mass flow of gas (Kaack et al. 2021).
The pattern of embolism events in leaves suggest that the proximity to air-filled neighbours influences the vulnerability of individual conduits (Guan et al. 2021). While often initiated in the midrib, embolism can display an unpredictable pattern of progression across the network of leaf veins, with large areas, comprised of many hundreds of xylem conduits, often observed embolizing simultaneously (Brodribb et al. 2016b). In detached leaves in which all of the xylem conduits in a petiole are embolized, the seeding of embolism throughout the leaf venation network is accelerated, occurring at higher water potentials than in an intact leaf on dehydration (Guan et al. 2021).
The relative importance of individual conduit traits versus the hydration status of surrounding conduits on determining the embolism resistance of an individual xylem conduit remains largely unknown. There is evidence suggesting gas-filled conduits make water filled neighbours more vulnerable (Knipfer et al. 2015; Choat et al. 2015b; Torres-Ruiz et al. 2016), yet pit membranes greatly reduce the spread of gas into water-filled conduits from embolized neighbours (Choat et al. 2008). To what extent does the presence of pre-existing embolism influence the vulnerability of the remaining xylem? Here we designed an experiment in which to test four hypotheses related to this question. The first hypothesis is that individual conduits have a specific water potential threshold at which an embolism will occur, and that pit membranes reduce the likelihood of embolism spread between gas and water-filled conduits under a narrow range of pressure differences. Challenges in testing this hypothesis are the lack of accurate water potential measurements at the individual conduit level (Bouda et al. 2019), and the uncoupling of pressure-driven embolism propagation from temporal effects. The second hypothesis is that pre-existing embolism changes the apparent embolism resistance of remaining individual water-filled conduits. If each conduit has a high fidelity to an individual threshold at which embolism will occur, then the most vulnerable conduits will consistently experience embolism earliest in drought, such that if drought is abated and refilling does not occur, then the relative embolism resistance of the remaining xylem may appear to be higher than the conduits with a low embolism resistance. Additionally, we hypothesize that on rehydration no refilling of embolized conduits will occur and that leaf xylem behaves in a similar way to stem xylem.
To test these hypotheses, we selected five angiosperm species with xylem consisting of vessels and one vessel-less angiosperm species, all with uniform pit membranes, and three conifers with tracheid-based xylem separated by torus-margo pit membranes. Embolism was observed using the optical method, which is non-destructive, and permits a clear visualization of embolism events, through a cycle of dehydration to a variable degree of embolism, rehydration, and then subsequent dehydration to complete desiccation.