The optical method and Pneumatron showed strong similarity in measuring embolism resistance
We found a strong agreement between the optical method and the Pneumatron, with no significant difference between either P12, P50 or P88 values obtained from the optical method and the Pneumatron in five deciduous species (Figure 4, Table S2). This finding is in line with the results from Pereira et al. (2019) for Eucalyptus camaldulensis , and supports the hypothesis that the pneumatic method relies largely on gas extraction from embolism events in intact vessels (Jansen et al. , In press). The fast and straightforward approach of taking pneumatic methods on small samples such as individual leaves make this method also suitable for field observations and further work at the intraspecific and intratree level.
While a high coefficient of determination (R2=0.91) indicated good agreement between P50 values obtained from two methods (Fig. 5b), the vulnerability curves with the pneumatic method were less steep than with the optical method for four species studied, except for L. tulipifera . A somewhat weaker, but still significant correlation was found between the P12values based on both methods (Fig. 5a). The xylem area selected to apply the optical method was deliberately chosen in the upper part of the leaf blade, where vessels in the leaf veins are separated from cut vessels at the petiole end by at least one and most likely various intervessel walls, as shown based on the maximum vessel length in petioles. Since the Pneumatron extracts gas from the petiole end, the tight similarity in embolism resistance between both methods suggests that the gas extracted with the Pneumatron comes from the intact vessels that are equally affected by the presence of gas from the cut-open vessels, or close to those that are observed with the optical method (Fig. 6). It is possible that there could be an overlapping area with embolism measured in intact vessels based on both methods (Fig. 6). Yet, it is unclear over how many end walls the Pneumatron is able to extract gas. We speculate that this number of end walls depends at least partly on the pit membrane thickness, the complete or partial hydration of the pit membrane, and whether or not porous medium characteristics of interconduit pit membranes change during dehydration, since these would determine gas diffusion considerably (Crombie et al. , 1985; Zhanget al. , 2020; Kaack et al. , 2019).
It is possible that the pressure and humidity inside an embolised vessel affects the process of embolism spreading, with a more efficient propagation of embolism spreading from a cut vessel to an intact one than between two intact ones. This may explain why the percentage of gas discharged by the Pneumatron was slightly higher than the percentage of the cumulative embolised pixels during early stages of dehydration in four out of five species studied (Figure 3, Figure 5A), and a similar trend was found in Pereira et al. (2019). However, the opposite was found for P. avium , which may have slower gas diffusion due to its thick pit membranes as compared to the other species. If intact vessels become embolised, but gas diffusion across pit membranes is slow, atmospheric pressure will not be quickly reached in a recently embolised vessel (Wang et al. , 2015a, b). Interestingly, the amount of gas extracted with the Pneumatron is considerably lower than the amount of gas that would be available in intact vessels with average dimensions under atmospheric pressure.
In conclusion, our work revealed the potential impact of cut-open conduits on embolism resistance in leaf xylem, and we found that embolism spreading during dehydration may happen at a less negative water potential when there is a direct connection to a gas source. This process could be minimised by hydraulic segmentation, with narrow and short conduits slowing down gas diffusion due to their confined and/or poorly interconnected nature.