Pneumatic measurements
An automatic Pneumatron apparatus was applied to measure gas diffusion kinetics of desiccating leaves (Pereira et al. , 2020a, Jansenet al. , In press). The principle of this apparatus was similar to the manual approach of Pereira et al. (2016) and Zhang et al. (2018), but with a much higher accuracy and temporal resolution. When samples for pneumatic measurements were prepared, the cut-open conduits embolised intentionally, which happened quickly by shaving the sample ends with a fresh razor blade in air.
Pneumatic measurements were taken every 15 min. For this a 40 kPa of absolute pressure was created with a vacuum pump, which extracted gas from a petiole end via a solenoid valve. The amount of gas extracted in a discharge tube with known volume could be measured with a pressure sensor. The vacuum pump reached 40 kPa (i.e., the initial pressure Pi) within less than a second. Pressure data were recorded in a SD card at a time interval of 500ms. The final pressure (Pf) was taken after 30s. According to the ideal gas law, the moles of air extracted from vessels (Δn) could be calculated following the equation below:
Δn = nf-ni =\(\frac{(Pf-Pi)\times V}{\text{RT}}\) (Eqn 1)
where ni and nf represented the moles of air molecules at the initial and final pressure, V was the fixed volume of the discharge tube (1.1 mL), R was the gas constant (8.134 kPa L mol-1 K-1), and T was the room temperature in the lab (around 25°C). Since a small V is needed to increase the measurement precision when a tiny amount of air is sucked from plant tissue, which is the case for detached leaves (Pereiraet al. , 2020a, Jansen et al. , In press), V was estimated as the maximum gas volume that could be extracted when leaves were fully dehydrated (ADmax, see below) divided by 510.2 (Pereiraet al. , 2020a). The volume of air discharged (AD, µL) from vessels could then be calculated based on the ideal gas law, with Patm being the atmospheric pressure: AD = 106×Δn R T/ Patm (Eqn 2),
Finally, the Percentage of Air Discharged (PAD, %) was calculated:
PAD = 100×(AD-ADmin)/(ADmax-ADmin) (Eqn 3)
where ADmin was the minimum volume of air discharged when the leaf was well hydrated, and ADmax was the maximum volume of air discharged when the leaf was strongly dehydrated.
Vulnerability curves were generated by plotting PAD or PEP against the corresponding leaf water potential (Ψ), with a fitting by the following equation (Pammenter and Vander Willigen, 1998):
PAD or PEP =100/ (1+exp (S /25) (Ψ-P50)) (Eqn 4)
S represented the slope of the fitted curve, and P50 represented the water potential at 50% of air discharged, or 50% of the total embolised pixels of the leaf area scanned. Values of P12 (water potential at 12% of air discharged or embolised pixels) and P88 (water potential at 88% of air discharged or embolised pixels) were calculated following the equations by Domec and Gartner (2001):
P12 = 2/(S /25) +P50 (Eqn 5)
P88 = -2/(S /25) +P50 (Eqn 6)