Herb Hydraulics: Inter- and Intraspecific Variation in Three Ranunculus Species^,

Small herbaceous species are vulnerable to water stress but adjusted to their habitat conditions and internally coordinated to avoid hydraulic failure.

Abstract

The requirements of the water transport system of small herbaceous species differ considerably from those of woody species. Despite their ecological importance for many biomes, knowledge regarding herb hydraulics remains very limited. We compared key hydraulic features (vulnerability to drought-induced hydraulic decline, pressure-volume relations, onset of cellular damage, in situ variation of water potential, and stomatal conductance) of three Ranunculus species differing in their soil humidity preferences and ecological amplitude. All species were very vulnerable to water stress (50% reduction in whole-leaf hydraulic conductance [k_leaf] at −0.2 to −0.8 MPa). In species with narrow ecological amplitude, the drought-exposed Ranunculus bulbosus was less vulnerable to desiccation (analyzed via loss of k_leaf and turgor loss point) than the humid-habitat Ranunculus lanuginosus. Accordingly, water stress-exposed plants from the broad-amplitude Ranunculus acris revealed tendencies toward lower vulnerability to water stress (e.g. osmotic potential at full turgor, cell damage, and stomatal closure) than conspecific plants from the humid site. We show that small herbs can adjust to their habitat conditions on interspecific and intraspecific levels in various hydraulic parameters. The coordination of hydraulic thresholds (50% and 88% loss of k_leaf, turgor loss point, and minimum in situ water potential) enabled the study species to avoid hydraulic failure and damage to living cells. Reversible recovery of hydraulic conductance, desiccation-tolerant seeds, or rhizomes may allow them to prioritize toward a more efficient but vulnerable water transport system while avoiding the severe effects that water stress poses on woody species.

The resistance of terrestrial plant species to water stress is an important determinant of their spatial distribution (Engelbrecht et al., 2007), and in a globally changing climate, resistance to water stress-induced damage to the water transport system is an essential parameter for the prediction of species survival and future changes in biodiversity and species composition. The hydraulic vulnerability to drought-induced embolism in woody species is adjusted to their environmental conditions on a global scale (Choat et al., 2012), but increased hydraulic safety entails several tradeoffs, such as transport efficiency, as plants can reduce the risk of hydraulic failure by the formation of smaller and hydraulically less efficient conduits (Tyree et al., 1994; Wagner et al., 1998, Gleason et al., 2016).

While many studies have dealt with the hydraulics of woody species (Maherali et al., 2004; Choat et al., 2012), the hydraulic characteristics of small herbaceous species are largely understudied, and our knowledge regarding their water transport system, including vulnerability to drought-induced loss of conductivity, remains very limited (Kocacinar and Sage, 2003; Brodribb and Holbrook, 2004; Iovi et al., 2009; Holloway-Phillips and Brodribb, 2011a; Tixier et al., 2013). Yet, herbs play an important ecological role in many biomes, such as grasslands and the alpine zone (Billings and Mooney, 1968; Gilliam, 2007; Scholz et al., 2010), and represent a significant percentage of crop plants (Monfreda et al., 2008); therefore, they are interesting from an ecophysiological viewpoint.

The requirements of the water transport system of herbs differ in several ways from those of woody species (Mencuccini, 2003). First, transport distances and supported leaf areas are much smaller in herbs, and reversible extraxylary limitations (Brodribb and Holbrook, 2005; Kim and Steudle, 2007) may have a stronger effect on hydraulic conductivity. Second, herbs often do not feature an essential, lignified main axis designed for long-term functioning, in which hydraulic failure may mean whole-plant mortality. Third, small plants may more easily restore water transport capacity in the case of embolism formation, because, due to smaller size, a water potential (Ψ) at or near zero can be reached quickly under favorable conditions due to positive root pressure (Tyree and Sperry, 1989; Tyree and Ewers, 1991; Stiller and Sperry, 2002; Tyree and Zimmermann, 2002; Ganthaler and Mayr, 2015). Thus, herbs may not be as threatened by drought-induced failure as taller woody species.

Hydraulic failure occurs when vulnerability thresholds are exceeded under water stress and the xylem tension is suddenly released as water is replaced by air, effectively blocking water transport in the affected conduits (Tyree and Zimmermann, 2002). Vulnerability to drought-induced embolism can thus be analyzed hydraulically by measuring decreases in conductance during increasing water stress (Sperry et al., 1988a; Cochard, 2002; Cochard et al., 2013), by noninvasive imaging of xylem water content (Choat et al., 2010; Brodersen et al., 2013; Cochard et al., 2015), or indirectly by monitoring of ultrasonic acoustic emissions (UE) which occur during the spontaneous energy release at embolism formation (Milburn and Johnson, 1966; Mayr and Rosner, 2011; Ponomarenko et al., 2014). However, the usability of UE analysis to noninvasively study hydraulic vulnerability remains under discussion (Sandford and Grace, 1985; Kikuta, 2003; Wolkerstorfer et al., 2012; Cochard et al., 2013; Rosner, 2015).

On the leaf level, desiccation may induce turgor loss, decreases in mesophyll conductance (Kim and Steudle, 2007; Scoffoni et al., 2014), or cell collapse (Blackman et al., 2010; Holloway-Phillips and Brodribb, 2011a), and limit hydraulic conductance, gas exchange, and thus photosynthesis before embolism occurs. When water stress increases further, living cells may incur damage resulting in tissue and leaf mortality.

In this study, we compared key hydraulic features (vulnerability to drought-induced loss of conductance as measured via hydraulic flow and xylem staining, pressure-volume relations, and the onset of cellular damage) of three species from the genus Ranunculus, which is distributed almost worldwide and contains species adapted to dry and humid habitats as well as species with broad ecological amplitude. We hypothesized that herbaceous species would be more vulnerable to water stress than woody species but also would show interspecific and intraspecific adjustments in hydraulic vulnerability based on the water availability of their respective habitats. Namely, the hydraulics of species with narrow ecological amplitude were hypothesized to reflect their respective habitat conditions, and broad-amplitude species should show adequate intraspecific variation to enable growth in dry and humid habitats.

RESULTS

Volumetric soil water content (SWC) at field capacity was 40.4% (dry site), 51% (humid site; Ranunculus acris), and 30.4% (humid site; Ranunculus lanuginosus). Relative SWC after 7 d of rain-free weather was above 90% at the humid sites and 43% to 58% at the dry site (Table I). The low field capacity SWC of the second humid site was due to the soil composition (thin layer of humus and litter on top of gravel substrate), yet the SWC remained constant due to the humid conditions inside the gorge. The reason for a relative soil water content (SWCrel) of 100% or greater at this site may lie in water seeping through the slope from uphill soils. Ecological indicator values for the study species and cooccurring species at the sampling sites supported the observed differences in water availability: averaged humidity values were 4.4 at the dry site and notably higher at 5.5 (R. acris) and 5.7 (R. lanuginosus) at the humid sites (Table I). At the dry site, Ranunculus bulbosus (humidity index of 3; indicator for dry sites) and R. acris (humidity index of 6; indicator for humid sites but broad amplitude) were accompanied by dry-habitat species such as Bromus erectus, Knautia arvensis, and Lotus corniculatus. In contrast, at the humid sites, R. lanuginosus (humidity index of 6) and R. acris were cooccurring with other species typical for moist habitats, including Polygonum bistorta, Rumex conglomeratus, Urtica dioica, and Solanum dulcamara (Supplemental Table S1).

Table I. SWC_rel (measured in summer 2014 and 2015; n = 10), averaged humidity value (n = 10–13), k_{leaf max} (n = 5), TLP (n = 5), diurnal Ψ_min (n = 3 [Fig. 4; Supplemental Fig. S2]), and diurnal g_{s max} (n = 5 [Fig. 4; Supplemental Fig. S2]).

Values are means ± se. Letters indicate significant differences in each row (P < 0.05).

Species	SWCrel (2014)	SWCrel (2015)	Humidity Value	TLP	kleaf max	Ψmin	gs max
	%	%		MPa	mmol m−2 s−1 MPa−1	MPa	mmol m−2 s−1
R. bulbosus (dry)	57.57 ± 4.03 a	43.29 ± 2.42 a	4.40 ± 0.34 a	−1.43 ± 0.02 a	28.00 ± 0.78 a	−1.70 ± 0.03 a	537.7 ± 105.3 a
R. lanuginosus (humid)	102.72 ± 6.14 b	98.34 ± 2.31 b	5.69 ± 0.26 b	−1.14 ± 0.05 b	27.77 ± 1.12 a	−0.42 ± 0.12 b	492.1 ± 86.1 a
R. acris (dry)	57.57 ± 4.03 a	43.29 ± 2.42 a	4.40 ± 0.34 a	−1.72 ± 0.08 c	23.40 ± 0.42 a	−2.20 ± 0.06 c	756.6 ± 195.9 a
R. acris (humid)	93.40 ± 1.87 b	90.59 ± 1.30 c	5.50 ± 0.23 b	−1.63 ± 0.05 a,c	24.06 ± 0.78 a	−1.48 ± 0.15 a	885.0 ± 37.3 a

The turgor loss point (TLP) was least negative in R. lanuginosus at a leaf water potential (Ψ_leaf) of −1.1 ± 0.1 MPa and was reached at −1.4 ± 0.0 MPa in R. bulbosus and at −1.6 ± 0.1 to −1.7 ± 0.1 MPa in R. acris (Table I). The osmotic potential (Π₀) followed a similar trend at −0.9 ± 0.1 MPa (R. lanuginosus), −1.3 ± 0.0 MPa (R. bulbosus), and −1.4 ± 0.1 to −1.5 ± 0.1 MPa (R. acris). In contrast, the modulus of elasticity was lowest (i.e. cells were more elastic) at 9.6 ± 1.9 MPa in R. lanuginosus and 1.5 to 2.5 times higher in the other species (R. bulbosus, 16.7 ± 1.3 MPa; and R. acris, 20.4 ± 4.4 to 24.9 ± 3.6 MPa).

Maximum whole-leaf hydraulic conductance (k_{leaf max}) was between 23.4 and 28 mmol m⁻² s⁻¹ MPa⁻¹ and did not differ between species (Table I). Hydraulically measured water potential at 50% loss of leaf hydraulic conductance (P_k50) was least negative in R. lanuginosus (−0.2 MPa) and more negative in R. acris and R. bulbosus (−0.6 to −0.8 MPa; Table II; Fig. 1). Similarly, water potential at 88% loss of leaf hydraulic conductance (P_k88) was highest in R. lanuginosus (−0.4 MPa) and significantly more negative in the other two species (−1.5 to −3.3 MPa; Table II; Fig. 1).

Table II. Water potential at 50 and 88% loss of leaf hydraulic conductance (P_k50 and P_k88; MPa), and 12% of cellular lysis (P_12CL; MPa).

Thresholds were computed after fitting the reparameterized Weibull function (Ogle et al., 2009) to vulnerability curves. Values are means (MPa) and 95% confidence intervals. Letters indicate significant differences in each row (P < 0.05).

Species	Leaf Hydraulic Vulnerability		Electrolyte Leakage
	Pk50	Pk88	P12CL
R. bulbosus (dry)	−0.57 (−0.63, −0.48) a	−1.53 (<−1.55, −1.31) a	−1.11 (−1.53, −0.84) a
R. lanuginosus (humid)	−0.15 (−0.19, −0.10) b	−0.44 (−0.60, −0.35) b	−1.16 (−1.47, −0.88) a
R. acris (dry)	−0.82 (−1.08, −0.54) a	−3.27 (<−3.27, −2.64) c	−2.02 (−2.43, −1.64) b
R. acris (humid)	−0.73 (−0.93, −0.50) a	−2.84 (<−3.15, −2.25) c	−2.02 (−2.29, −1.62) b

Figure 1.

k_leaf versus Ψ in R. bulbosus, R. lanuginosus, and dry and humid R. acris populations measured with the non-steady state leaf rehydration technique. Vertical lines indicate fitted hydraulic vulnerability parameters P_k50 (solid) and P_k88 (dashed). Curves were extrapolated using the same equation and parameters used for vulnerability curves (Eq. 2).

Conductive xylem staining in R. acris (dry) showed the first notable signs of embolism in minor leaf veins and vascular bundles at −2 MPa, with significant losses of even first-order leaf veins by −2.7 MPa (Fig. 2).

Figure 2.

Stained leaf laminas and petiole cross sections of R. acris (dry) indicating loss of hydraulic conductance at increasing levels of dehydration. Conducting elements appear red, while embolized conduits remain unstained. A, Leaf lamina undersides and petioles. Bars = 20 mm. B, Leaf lamina details. Bars = 5 mm. C, Petiole cross sections approximately 3 cm from the laminas. Bars = 200 µm. D, Vascular bundle details from cross sections. Bars = 50 µm.

Significant cellular lysis started at more negative Ψ than P_k50, with water potential at 12% of cellular lysis (P_12CL) between −1.1 and −2 MPa (Table II; Fig. 3), whereas thresholds for cell damage did not correspond to humidity conditions. In contrast, the onset of UE was observed at less negative Ψ than hydraulic vulnerability thresholds, with peak activity at −1.1 MPa or greater (Supplemental Table S2; Supplemental Fig. S1).

Figure 3.

Cellular lysis (percentage of maximum after autoclaving) versus Ψ in R. bulbosus, R. lanuginosus, and dry and humid R. acris populations. Vertical lines (dashed) indicate the onset of damage to living cells (P_12CL). Curves were extrapolated using the same equation and parameters used for vulnerability curves (Eq. 2).

In situ, species reached minimum leaf water potential (Ψ_min) of −0.4 ± 0.1 to −2.2 ± 0.1 MPa, whereby Ψ_min was lower at the dry site (Table I; Fig. 4). In contrast, stomatal conductance (g_s) exhibited considerable variability at more negative Ψ. Maximum stomatal conductance (g_{s max}) was between 492.1 ± 86.1 and 885 ± 37.3 mmol m⁻² s⁻¹ and did not reflect soil humidity differences between sites (Table I; Fig. 4). The progression of Ψ_leaf and g_s in the field (illustrated in Fig. 4 for R. acris) revealed that g_s was reduced in all species when Ψ_min was reached. Theoretical whole-leaf hydraulic conductance (k_leaf) as calculated from measured Ψ_leaf and transpiration (E; based on measured g_s and climate data from the closest weather station at approximately 3 km distance) agreed well with k_leaf predicted from vulnerability curves. However, vapor pressure deficit and transpirational demand were expected to be lower in the lowest canopy layer than calculated here (Wohlfahrt et al., 2010). Both Ψ_leaf and g_s were compared on representative, sunny days, but environmental conditions (wind or prolonged drought stress) may lead to more extreme values. Because of windy (R. bulbosus) or shady (R. lanuginosus; due to its understory gorge habitat) conditions during in situ measurements (Supplemental Fig. S2), diurnal variation data for only R. acris were used for further interpretation.

Figure 4.

Diurnal variation of Ψ (MPa; black circles, solid black lines) and g_s (mmol m⁻² s⁻¹; red open circles, dashed red lines) in R. acris populations on dry and humid sites. Vertical lines (dashed) indicate meteorological dawn, and yellow areas indicate direct sunlight. Values are means ± se.

DISCUSSION

We found that all studied species were generally vulnerable to dehydration, but differences between species/populations reflected the humidity conditions of their respective habitats.

Sites differed in their water availability, as the SWC_rel between dry and humid sites differed significantly after only 7 d without precipitation. At the dry site, SWC_rel decreased to less than 60% of field capacity after only 1 week of rain-free weather, whereas the humid sites experienced hardly any reduction in soil humidity. The low SWC at the humid site of R. lanuginosus, which was due to soil composition, was compensated by the humid conditions inside the gorge, so that SWC_rel did not decrease during the rain-free period. Species composition and ecological indicator values supported the distinction between dry and humid sites (Table I; Supplemental Table S1).

Herb Hydraulics

k_leaf is known to vary widely across species and plant functional types, with an average k_leaf of 11.5 ± 0.9 mmol m⁻² s⁻¹ MPa⁻¹ across 107 woody and herbaceous species and with individual species ranging between 0.8 and 49 mmol m⁻² s⁻¹ MPa⁻¹ (Sack and Holbrook, 2006). Leaf hydraulic efficiency in herbaceous crops was above average at 24.1 ± 2.2 mmol m⁻² s⁻¹ MPa⁻¹ (Sack and Holbrook, 2006), and other, noncultivated herbaceous angiosperms in the literature had values between 2.7 and 22 mmol m⁻² s⁻¹ MPa⁻¹ (Brodribb and Holbrook, 2004; Bunce, 2006; Galmes et al., 2007). Our study species were within the range known for herbaceous plants and, therefore, showed much higher hydraulic efficiency than woody species (Table I; Fig. 1).

Leaf hydraulic vulnerability studies on small herbaceous species found 50% reduction in k_leaf between −1 and −1.8 MPa in perennial grasses (Brodribb and Holbrook, 2004; Holloway-Phillips and Brodribb, 2011a, 2011b). Woody species generally feature leaves that are less vulnerable to water stress, with an average P_k50 of −1.8 ± 0.1 MPa across 81 woody angiosperm species when measured with comparable methods (Nolf et al., 2015). Although some of these species also had very vulnerable leaves at P_k50 of −1 MPa or greater (Hao et al., 2008; Chen et al., 2009; Blackman et al., 2012; Guyot et al., 2012), the majority of previously studied species exhibited P_k50 of −1.5 MPa or less and down to −4.3 MPa (Sack and Holbrook, 2006; Blackman et al., 2010; Johnson et al., 2012; Nardini et al., 2012; Bucci et al., 2013). Accordingly, stems were typically also more vulnerable to drought-induced hydraulic failure in herbs (−1 to −3.8 MPa; Kocacinar and Sage, 2003; Rosenthal et al., 2010; Tixier et al., 2013; Nolf et al., 2014) than in woody species (−0.1 to −14.1 MPa, and −3.4 ± 0.1 MPa across 480 species; Choat et al., 2012).

Leaf hydraulic analysis in our study showed that all three Ranunculus spp. were very vulnerable to water stress, with P_k50 of −0.8 MPa or greater in k_leaf across all species (Table II; Fig. 1). However, xylem staining indicated that early hydraulic conductivity losses at moderate Ψ were due to extraxylary, likely reversible, effects rather than xylem embolism. Stained petiole sections and leaves revealed loss of conductivity due to embolism and suggested an estimated 50% loss of xylem hydraulic conductivity due to embolism of −2 MPa or less in R. acris (dry; Fig. 2). Thus, despite the fact that plants were still very vulnerable to drought, it is important to stress that the leaf hydraulics method employed measured bulk hydraulic conductivity, including both xylary and extraxylary (e.g. mesophyll) pathways, and reflected the total effect on symplastic and apoplastic transport. Interestingly, stained cross sections also indicated an all-or-nothing effect, where, in most cases, individual vascular bundles lost all conductivity at once (Fig. 2C).

Compared with the staining results, ultrasonic activity was observed at much less negative Ψ than expected (Supplemental Table S2; Supplemental Fig. S1). We believe that other signal sources unrelated to water transport (e.g. parenchyma and sclerenchyma) were responsible for a considerable portion of emissions, skewing the acoustic analysis. Thus, interpretation of UE analyses on herbs is difficult and requires further methodical optimization.

With daily Ψ_min being 0.3 to 1.4 MPa lower than P_k50, at least two of our species would incur significant water transport reduction during the course of the day (Tables I and II; Stiller et al., 2003), limiting transpiration and, thus, according to Ohm’s law, preventing a further decrease of Ψ. However, the staining experiment suggests that embolism plays only a minor role in hydraulic decline at the observed in situ Ψ (Fig. 2); therefore, water transport is limited mainly by reversible extraxylary effects (e.g. turgor loss or cell collapse and associated mesophyll conductance losses; Brodribb and Holbrook, 2005; Kim and Steudle, 2007; Blackman et al., 2010; Holloway-Phillips and Brodribb, 2011a; Scoffoni et al., 2014).

Like significant hydraulic transport reductions, the TLP is another critical point plants need to avoid so they can maintain photosynthesis, transport of solutes, and growth (Frensch and Hsiao, 1994; Kramer and Boyer, 1995; Brodribb and Holbrook, 2003; Bartlett et al., 2012). In our study species, the TLP was lower than hydraulic P_k50, indicating that all studied species would face notable reductions in transport capacity under limited water supply before leaves would lose turgor (Tables I and II). This indicates that the impairment of whole-leaf water transport is the first critical event during water stress. Pressure-volume analysis revealed that TLPs in this study were determined mainly by the Π₀, whereas a mediating, opposite effect caused by the observed trends in bulk elasticity was small. Only in R. lanuginosus, the high elasticity may have compensated for the species’ weak water-holding capability (high Π₀) to maintain turgor longer.

Electrolyte leakage analysis showed that desiccation-induced damage to living cells (P_12CL) would only begin long after significant hydraulic decline (P_k50; Table II; Figs. 1 and 3). Since leaves were benchtop dehydrated for up to several hours, we would expect an even higher apparent vulnerability due to continuous exposure to low Ψ in these measurements, compared with shorter in situ stress peaks followed by stomatal regulation and Ψ recovery. Thus, living cells were well protected from desiccation-induced cell damage, as found previously by other authors (Vilagrosa et al., 2010; Beikircher et al., 2013). Interestingly, plants from dry habitats did not appear more resistant to cell damage than plants from humid sites (Table II; Fig. 3). While we sampled only fully developed, mature leaves, we note that R. bulbosus required measurement 1 to 2 months earlier than other species due to their summer dormancy (see “Study Sites and Plant Material”).

Interspecific Differences

Study species with narrow ecological amplitude (R. bulbosus and R. lanuginosus) showed significant differences in some key hydraulic parameters (TLP, P_k50, and P_k88), which were correlated with their respective soil humidity conditions. The corresponding trend in daily Ψ_min further indicated that Ψ_leaf was regulated near P_k88 in both species. In contrast, k_{leaf max}, daily g_{s max}, and P_12CL revealed no corresponding trend, which suggests that these parameters are more evolutionarily conserved.

The hydraulics of R. acris, with its broad ecological amplitude, showed characteristics closer to the drought-adapted species (e.g. P_k50 and P_k88; Table II) regarding hydraulic measurements and did not reveal the intermediate values we expected. The fact that P_k50, P_k88, and P_12CL were more negative in R. acris than in R. bulbosus may suggest that, although R. acris also occurs frequently in humid sites, it is generally well adapted for dry habitats. Another possible explanation is that, by dying back to perennial tubers in early summer, R. bulbosus effectively avoids periods of intense water stress in summer and autumn (Coles, 1973; Sarukhan, 1974; Larcher, 1995).

Intraspecific Acclimation

Within-species variation in hydraulic parameters was found previously for woody species (Matzner et al., 2001; Cochard et al., 2007; Beikircher and Mayr, 2009; Charra-Vaskou et al., 2012) and herbaceous species (Holste et al., 2006; Tixier et al., 2013; Nolf et al., 2014). R. acris was chosen for its wide distribution and broad ecological amplitude and showed trends for intraspecific variation. While most of the species’ hydraulic traits indicated adjustment to dry habitats when compared with other species (Tables I and II; Figs. 1, 3, and 4), plants from the dry site tended to have more negative vulnerability thresholds (P_k50 and P_k88), TLP, and daily Ψ_min than conspecific plants from the humid site. Thus, we suggest that the intraspecific variability in hydraulic parameters of R. acris allows the species to retain fitness over a range of humidity conditions (Rejmanek, 1999) and is a major factor determining its wide distribution, as found for other herbaceous species (Hao et al., 2013; Nolf et al., 2014). Interestingly, the onset of desiccation-induced cellular damage (P_12CL; Table II; Fig. 3) was constant across sites, indicating that this parameter was not subject to adjustment.

CONCLUSION

Overall, all of our study species were very vulnerable to water stress, even R. bulbosus, which is a typical species of dry meadows. However, the first effect of drought in these species was found to be a loss of k_leaf at moderate Ψ based on extraxylary pathways rather than embolism formation. This may be an effective, easily reversible way for plants to avoid critical decreases in Ψ and subsequent hydraulic failure.

Yet, also in these herbs, differences in vulnerability to water stress were observed on interspecific and intraspecific levels: in narrow-amplitude species, some key hydraulic parameters were correlated to the humidity conditions of their respective habitats (TLP, daily Ψ_min, P_k50, and P_k88), while other parameters (k_{leaf max}, g_{s max}, and P_12CL) did not reveal such trends. In contrast, the broad-amplitude species showed correlations or trends in intraspecific adjustment to different habitat conditions. Due to the ecological importance of herbaceous species in many ecosystems, better knowledge regarding their hydraulics will be essential to improve evaluations of water stress vulnerability on vegetation and ecosystem levels.

MATERIALS AND METHODS

Study Sites and Plant Material

The study was conducted at three sites in Innsbruck, Austria (47.267°N, 11.393°E), with 896.5 mm of precipitation and an average air temperature of 8.5°C in the 30-year mean (ZAMG Zentralanstalt für Meteorologie und Geodynamik, 2002). Three species of Ranunculus, which occur frequently within the study area, were chosen for analysis due to their differing habitat conditions. Ranunculus bulbosus is a typical species of dry meadows and nutrient-poor grasslands and grows up to 15 to 40 cm tall (Hegi, 1974; Fischer et al., 2008). This species dies back to perennial bulb-like corms in July and remains dormant until the following spring (Coles, 1973; Sarukhan, 1974). Sample plants were taken from a south-exposed, nonintensively managed meadow (Hötting). In contrast, Ranunculus lanuginosus is found in humid, shaded environments such as riparian forests or gorges and grows up to 50 to 70 cm (Hegi, 1974; Fischer et al., 2008). Plants were collected in a small gorge (Sillschlucht) at the south end of the city. The third species, Ranunculus acris, occurs in a range of humidity conditions from dry to humid meadows and reaches intermediate heights of 30 to 60 cm (Hegi, 1974; Polatschek, 2000; Fischer et al., 2008). For this species, a dry population cooccurring with the sampled R. bulbosus population and a humid population occurring on a separate high-humidity site (Völs) were used for analysis.

Soil humidity, species composition, and diurnal variation of g_s and Ψ_leaf were recorded on site in late spring and summer (May to August) in two consecutive years. For all other analyses, whole plants including major roots were collected from the field sites between May and August, immediately placed in dark plastic bags to limit water loss, and transported to the laboratory within 30 min. Plants were rehydrated overnight with their roots in water and covered with a dark plastic bag. Measurements were made on healthy, fully developed leaves including petioles (pressure-volume relations, leaf hydraulic vulnerability, and conductive xylem staining), leaf petioles with intact laminas (acoustic emissions), and leaf discs cut from the leaf laminas (electrolyte leakage).

Soil Humidity

Soil humidity was analyzed in situ in 2 years during dry periods with the Hydrosense II Soil Moisture Measurement System (Campbell Scientific), whereby volumetric SWC was measured 24 h after strong rainfall (SWC_{field capacity}) and after 7 d without rain (SWC_dry) in 10 samples per site and date. The SWC_rel (percentage of field capacity) after 7 d without rain was calculated as:

(1)

Additionally, we recorded species composition at the three sites and characterized water availability using averaged ecological indicator values (humidity value; Ellenberg et al., 1992; Schaffers and Sykora, 2000).

Ψ Determination

Ψ_leaf was measured with a pressure chamber (Model 1000 Pressure Chamber; PMS Instrument) by applying pressure to an excised leaf sealed in the chamber, with the petiole protruding out through a rubber gasket (Scholander et al., 1965). Air pressure inside the chamber was increased slowly until the xylem sap meniscus became visible at the cutting site.

Pressure-Volume Analysis

Pressure-volume analyses (Tyree and Hammel, 1972; Sack and Pasquet-Kok, 2011; Bartlett et al., 2012) were carried out on five leaves per species/population to determine TLP, Π₀ at full turgor, and modulus of elasticity. Leaves were fully saturated before dehydration and regular measurements of Ψ_leaf and fresh weight. Relative water content was calculated from saturated, fresh, and dry weights and plotted against the inverse of Ψ_leaf to find the TLP (transition point between the linear and curved portions of the graph), Π₀ (negative inverse of the linear graph portion’s intercept with the y axis), and modulus of elasticity (slope of the curve above and including the TLP). Parameters were calculated individually per leaf and then averaged per species/population.

Leaf Hydraulic Vulnerability

Declines in k_leaf due to increasing desiccation were analyzed with a modified non-steady-state leaf rehydration technique (Brodribb and Cochard, 2009; Blackman and Brodribb, 2011) and benchtop dehydration. For 39 to 45 leaves per species/population, we measured the decay of initial water uptake into the leaf from a small water container placed on a digital balance (CP64; Sartorius), which was connected to the leaf petiole via hydraulic tubing (see below).

At increasing levels of dehydration, plants were equilibrated in dark plastic bags and plant Ψ was determined on two leaves per plant. If Ψ between both leaves differed by 10% or less, the decay of initial flow was measured by excising a third leaf under water (including approximately 3 cm of the petiole), immediately connecting the leaf to hydraulic tubing, and covering the leaf with a moist paper towel to prevent transpiration. Thus, hydraulic flow was measured on non-light-acclimated, nontranspiring leaves. Flow rates (k_leaf [mmol m⁻² s⁻¹ MPa⁻¹]) were determined by linear regression of decreasing mass readings (water uptake) during the first 10 s and adjusted for temperature and leaf area. k_{leaf max} was calculated as the average of five maximum conductances observed per species/population. No cutting artifact (Wheeler et al., 2013) was expected for leaves excised under water (Scoffoni and Sack, 2015).

The reparameterized Weibull function (Ogle et al., 2009) was fitted to plots of relative k_leaf versus Ψ using the fitplc package for R version 3.1.1 (Duursma, 2014; R Core Team, 2014):

(2)

where k is the expected hydraulic conductance at X% loss of conductance, k_{leaf max} is the k at full saturation (i.e. k at 0 MPa), P is the positive-valued Ψ (P = −Ψ), P_x is the Ψ at X% loss of conductance, and S_x is the slope of the vulnerability curve at P = P_x. P_k50 and P_k88 were defined as above.

Staining of Conducting Xylem Elements

To confirm the observed high levels of leaf hydraulic vulnerability, we stained conducting xylem elements in leaves of R. acris from the dry site. At increasing levels of dehydration, whole plants were equilibrated in dark plastic bags and plant Ψ was determined on one leaf per plant. Two additional leaves of the same plant were excised under water and immediately placed in a beaker containing (1) 1% (w/v) phloxine B (modified after Nardini et al., 2003) to reveal ongoing water transport in the lamina and (2) 0.05% (w/v) safranin (modified after Sperry et al., 1988b) to stain water-conducting conduits in the petioles. Samples were exposed to sunlight for about 10 min. Transpiration was favored by partially removing the cuticle and confirmed with an SC-1 Leaf Porometer (Decagon Devices). We analyzed images of the lamina taken with a digital camera and petiole cross sections (approximately 3 cm from the lamina) taken with a light microscope (Olympus BX41; Olympus Austria) and interfaced digital camera (ProgRes CT3; Jenoptik). To prevent artifacts by passive dye diffusion, only samples with unstained parenchyma and phloem were used.

Electrolyte Leakage

The vulnerability of living tissues to dehydration (Morin et al., 2007; Hu et al., 2010; Vilagrosa et al., 2010; Beikircher et al., 2013) was analyzed by benchtop dehydrating plants and comparing the leakage of electrolytes from 52 to 82 leaves per species at increasing levels of dehydration. After Ψ determination, three to 10 circular discs (0.6 cm in diameter) were cut from each leaf by use of a cork borer while avoiding major leaf veins. Leaf discs were pooled for each sample leaf, submerged in 15 mL of distilled water in individual centrifuge tubes, shaken for 24 h on a horizontal shaker (ST5 Bidimensional Shaker; CAT) at 5°C, followed by electrolytic conductivity measurement at room temperature (after 1 h of temperature equilibration) using a four-electrode conductivity sensor (Tetracon 325; WTW). Samples were then autoclaved at 120°C for 30 min and shaken overnight at room temperature, before final electrolytic conductivity was measured.

Relative electrolyte leakage (REL) was calculated as:

(3)

where C₁ and C₂ are the initial and final electrolytic conductivities, respectively. REL was then standardized to the percentage of cellular lysis by relating each sample’s REL to the REL of the control (REL_c; Morin et al., 2007); in this case, samples were measured at Ψ of −0.2 MPa or greater:

(4)

Vulnerability thresholds were again calculated by curve fitting to Equation 2, where k is the percentage of intact cells (k = 100 − the percentage of cellular lysis). The onset of damage to living cells (P_12CL) was defined as above.

Diurnal Variation of g_s and Ψ_leaf

We monitored the diurnal progression of g_s (using an SC-1 Leaf Porometer; Decagon Devices) and corresponding Ψ_leaf between sunrise and sunset on one clear, sunny day per species following a period of rainfall. For each data point, g_s was measured on the lower leaf side of four to five undamaged, fully developed leaves of different plants, whereas Ψ_leaf was determined on three undamaged, fully developed leaves of different plants.

Statistical Analyses

Data for curve fitting were pooled per species/population and method before further analysis. Vulnerability thresholds were statistically tested for differences using bootstrapped confidence intervals (n = 999 bootstrap estimates). Other data were normally distributed (QQ plots and Shapiro-Wilk test) and tested for differences using Student’s t test (homoscedastic data as determined by the robust Brown-Forsythe Levene-type test) or Welch’s two-sample t test (heteroscedastic data). All analyses were done in R version 3.1.1 (R Core Team, 2014) at a significance level of P < 0.05, whereas P was Bonferroni corrected for multiple comparisons. Values are presented as means ± se or mean (lower confidence interval, upper confidence interval).

Supplemental Data

The following supplemental materials are available.

• Supplemental Figure S1. Relative cumulative UE versus Ψ in R. bulbosus, R. lanuginosus, and dry and humid R. acris populations.

• Supplemental Figure S2. Diurnal variation of Ψ and g_s in R. bulbosus and R. lanuginosus.

• Supplemental Table S1. Species composition and ecological indicator values.

• Supplemental Table S2. Water potential at 50 and 88% of total cumulative ultrasonic emissions above -3 MPa.

Glossary

Ψ

water potential

UE

ultrasonic acoustic emissions

SWC

soil water content

SWC_rel

relative soil water content

TLP

turgor loss point

Ψ_leaf

leaf water potential

Π₀

osmotic potential

k_{leaf max}

maximum whole-leaf hydraulic conductance

P_k50

water potential at 50% loss of leaf hydraulic conductance

P_k88

water potential at 88% loss of leaf hydraulic conductance

P_12CL

water potential at 12% of cellular lysis

Ψ_min

minimum leaf water potential

g_s

stomatal conductance

g_{s max}

maximum stomatal conductance

k_leaf

whole-leaf hydraulic conductance