Ancestral stomatal control results in a canalization of fern and lycophyte adaptation to drought


Author for correspondence:

Timothy J. Brodribb

Tel: +61 4 6226 1707



  • Little is known about how a predominantly passive hydraulic stomatal control in ferns and lycophytes might impact water use under stress. Ferns and lycophytes occupy a diverse array of habitats, from deserts to rainforest canopies, raising the question of whether stomatal behaviour is the same under all ecological strategies and imposes ecological or functional constraints on ferns and lycophytes.
  • We examined the stomatal response of a diverse sample of fern and lycophyte species to both soil and atmospheric water stress, assessing the foliar level of the hormone abscisic acid (ABA) over drought and recovery and the critical leaf water potential (Ψl) at which photosynthesis in droughted leaves failed to recover.
  • The stomata of all ferns and lycophytes showed very predictable responses to soil and atmospheric water deficit via Ψl, while stomatal closure was poorly correlated with changes in ABA. We found that all ferns closed stomata at very low levels of water stress and their survival afterwards was limited only by their capacitance and desiccation tolerance.
  • Ferns and lycophytes have constrained stomatal responses to soil and atmospheric water deficit as a consequence of a predominantly passive stomatal regulation. This results in a monotypic strategy in ferns and lycophytes under water stress.


Despite a relative stasis in stomatal morphology over > 400 million yr, the vegetative tissue bearing stomata has evolved immense diversity and complexity in form and function, from the small bifurcating axes of the oldest fossilized stomatal-bearing plants (Bateman et al., 1998; Edwards et al., 1998) to highly productive modern angiosperm leaves (Brodribb et al., 2009). In conjunction with this morphological evolution, the functional behaviour of stomatal opening and closure has similarly radiated in complexity, with significant implications for productivity and water use (Doi et al., 2006; Doi & Shimazaki, 2008; Brodribb et al., 2009; Brodribb & McAdam, 2011; Haworth et al., 2011, 2013; McAdam & Brodribb, 2012a,b).

Extensive research has investigated the controls of stomatal aperture in angiosperms (Raschke, 1975; Cowan & Farquhar, 1977; Damour et al., 2010), with two processes responsible for regulating stomatal aperture, the passive control of guard cell turgor by leaf water status (Buckley & Mott, 2002; Buckley, 2005) and the active control of guard cell osmotic potential by the transport of ions across cell membranes (Schroeder et al., 2001; Shimazaki et al., 2007; Lawson, 2009). While much work has focused on the stomatal behaviour of angiosperms, our current understanding of the evolution of stomatal control comes from relatively recent investigations into functional stomatal behaviour in modern representatives of extant lineages of vascular plants (Brodribb & McAdam, 2011; McAdam & Brodribb, 2012b) as well as molecular investigations into the function of key genetic components essential for stomatal signalling (Chater et al., 2011; Ruszala et al., 2011). Particular focus in the investigation of the evolution of stomatal control has been placed on comparing the stomata of the well-researched seed plants with nonseed plant groups. While complex metabolically driven stomatal control predominates in the ecologically dominant and successful seed plants (Ache et al., 2010), fern and lycophyte stomata appear to be overwhelmingly regulated by a passive response to leaf water status in the light (Brodribb & McAdam, 2011; McAdam & Brodribb, 2012a). The predominance of a passive response of fern and lycophyte stomata occurs despite the presence and function of key genetic components in these lineages that are essential for effective metabolic stomatal signalling (Chater et al., 2011; Ruszala et al., 2011). While these two lines of investigation appear contradictory, it seems most likely that, while the basal lineages of land plants are in possession of an operational genetic framework to elicit metabolic stomatal responses, the passive control of stomata by water balance predominates in the regulation of stomatal behaviour. The most conclusive evidence for active control of fern and lycophyte stomata is a stomatal response to red light that originates in the guard cells (Doi & Shimazaki, 2008; McAdam & Brodribb, 2012b); there is also inconclusive evidence of stomatal responses to CO2 (Mansfield & Willmer, 1969; Doi & Shimazaki, 2008; Brodribb et al., 2009; Ruszala et al., 2011). The behaviour of fern and lycophyte stomata in the light is highly predictable, with guard cells passively linked to changes in the turgor pressure of the leaf (Brodribb & McAdam, 2011; Brodersen et al., 2012; McAdam & Brodribb, 2012a). It has been suggested (McAdam & Brodribb, 2012b) that an evolutionary transition in stomatal control towards a predominantly metabolic regulation of stomatal aperture occurred during the radiations of seed plant lineages in drier late-Palaeozoic environments (DiMichele & Aronson, 1992) and was instrumental in ensuring the competitive success and ecological dominance of seed plants and the demise of fern- and lycophyte-dominated forests into the Mesozoic. Indeed, during the Palaeozoic, lycophytes and ferns enjoyed a rich period of ecological dominance (Phillips et al., 1985) and over their evolutionary history have included morphological representatives of all terrestrial life forms currently represented by extant seed plants (Rothwell, 1996).

Despite a reputation as mesic relicts, modern ferns and lycophytes are certainly not restricted to ever-wet environments. Evidence from the fossil record and extant taxa indicates that ferns and lycophytes have repeatedly radiated into xeric, cold and variable habitats (DiMichele & Phillips, 2002; Hietz, 2010), with the ability to survive desiccation widely represented (Hietz, 2010). The epiphytic growth habit, which exposes individuals to fluctuating water availabilities (Hietz & Briones, 1998; Watkins et al., 2007b; Watkins & Cardelús, 2009), is an ecological diversification widely represented in ferns (Schneider et al., 2004; Schuettpelz & Pryer, 2009) and the lycophyte genus Huperzia (Wikström & Kenrick, 2000; Wikström, 2001), concurrent with the rise of angiosperm-dominated forests. A number of extant fern genera are highly competitive and invasive, successfully out-competing seed plants and dominating landscapes and vegetation types both in modern forests (e.g. Pteridium (Marrs & Watt, 2006) and Lygodium (Pemberton & Ferriter, 1998)) and throughout the Mesozoic (Wing et al., 1993). Associated with this ecological diversity within the ferns and lycophytes is a substantial variability in maximum photosynthetic rates (A) and stomatal conductances (gs) (McAdam & Brodribb, 2012b) as well as a recent suggestion of highly variable maximum xylem hydraulic resistances and vulnerability in the stipe of different species (Pittermann et al., 2011; Brodersen et al., 2012). Ecological diversity, geological persistence and the capacity for physiological and morphological variability observed in ferns and lycophytes raise the question: do all ferns and lycophytes have the same predominance of passive control of leaf hydration by stomata and, if so, what are the adaptations adopted by ferns and lycophytes that allow them to survive with this stomatal control mechanism (Watkins & Cardelús, 2012)?

To answer these questions we examined the physiological responses to water stress of a diversity of fern and lycophyte species, including mesophytic terrestrial species, epiphytes and a desiccation-tolerant species. In particular, we quantified the sensitivity of stomata to soil drought, critical water potentials at which plants died, and the role of ABA in the response to drought. Our results provide a standardized comparison of the diversity of water management strategies in ferns and lycophytes.

Materials and Methods

Species examined

To observe the diversity of physiological responses of ferns and lycophytes to water deficit, six fern species and a lycophyte were specifically selected to span a wide ecological range of spore-bearing vascular plants, including mesophytic, drought deciduous, terrestrial species, epiphytes and a desiccation-tolerant species (Table 1).

Table 1. Species used to assess the physiological diversity of ferns and lycophytes in response to water deficit, and their distinctive morphological features and ecologies (see Supporting Information Fig. S1 for images of each species)
Terrestrial ferns
Adiantum capillus-veneris L.PteridaceaeRhizomatous terrestrial, mesophyticTemperate to tropical moist, shaded habitats
Dicksonia antarctica Labill.DicksoniaceaeTree fern, mesophyticTemperate rainforest understorey
Pteridium esculentum (G. Forst.)    CockayneDennstaedtiaceaeSubterranean rhizome, drought deciduousCosmopolitan, full sun to open canopied habitats
Epiphytic ferns
Pyrrosia lingua (Thunb.) Farw.PolypodiaceaeEpiphytic, rhizomatousSubtropical canopy dwelling
Rumohra adiantiformis (G. Forst.)   ChingDryopteridaceaeEpiphytic/lithophytic, rhizomatousTemperate closed or open canopy
Desiccation-tolerant fern
Cheilanthes myriophylla Desv.PteridaceaeRhizomatous terrestrial, mesophyticDesiccation-tolerant from rock crevices
Selaginella kraussiana (Kuntze) A. BraunSelaginellaceaeTerrestrialSubtropical, ever-wet habitats

All species examined were represented by three identical-aged individuals grown either from spores or from rhizomes collected in the field. Plants were grown in 1.3-l pots containing an 8 : 2 : 1 mix of composted pine bark, coarse river sand and peat moss with added slow-release fertilizer, and housed in the glasshouses of the School of Plant Science, University of Tasmania, Hobart, Australia. No individual had experienced previous drought stress. Plants were grown under natural light conditions supplemented and extended to a 16-h photoperiod by sodium vapour lamps, ensuring a minimum 300 μmol quanta m−2 s−1 at the leaf surface throughout the day period (maximum light intensity on cloudless days did not exceed 1100 μmol quanta m−2 s−1 at leaf level). Temperatures in the glasshouse were maintained at 22°C during the day and 15°C at night. All plants were watered daily and fertilized with liquid nutrients weekly (unless undergoing drought).

Drought, leaf gas exchange, water potential and foliar ABA level

Drought was initiated in three individuals per species by withholding water. Over the course of the drought, once a day (reducing to once a week after 10 d in more drought-tolerant species), between 12:00 h and 13:00 h, plants were transported to the laboratory where leaf gas exchange, leaf water potential (Ψl) and foliar ABA level were quantified on each individual. Leaf gas exchange was measured using an infrared gas analyser (Li-6400; Li-Cor, Lincoln, NE, USA) on photosynthetic tissue from a single leaf or stem (cuvette conditions: leaf temperature 22°C, vapour pressure difference (VPD) maintained between 1.1 and 1.2 kPa, 390 μmol mol−1 CO2, light intensity 1000 μmol quanta m−2 s−1, and flow rate 500 ml min−1), and tissue was allowed to equilibrate to chamber conditions until stability was reached (c. 20 min). On the same leaf or stem, Ψl was quantified on tissue excised and immediately wrapped in damp paper towel using a Scholander pressure chamber and microscope to precisely measure the xylem balance pressure without the loss of any water from the leaf tissue. Foliar ABA was extracted, purified and physicochemically quantified using ultra-high-performance liquid chromatography tandem mass spectrometry (UPLC-MS/MS) with an added internal standard (a phytohormone quantification method first used by Kojima et al., 2009) according to the methods of McAdam & Brodribb (2012a), with the following modifications for improving sample purification for UPLC-MS/MS analysis. Following extraction and reduction under vacuum at 35°C to < 1 ml, the sample was taken up in 3 × 1-ml washes of weak aqueous sodium hydroxide (pH 8). These washes were loaded on to a solid-phase extraction 600-mg SAX cartridge (Maxi-Clean™; Grace Davison Discovery Sciences, Deerfield, IL, USA), preconditioned with 10 ml of weak aqueous sodium hydroxide (pH 8). The loaded sample on the cartridge was washed with 10 ml of methanol and ABA was eluted with 15 ml of 2% acetic acid in methanol (v/v).

Assessing recovery using rehydrated leaves and plants

Following leaf gas exchange measurements, a pinnule, leaf or stem in closest proximity to the tissue sampled for gas exchange was excised in water and bagged. Over the course of the imposed drought, no individual suffered defoliation in excess of 20% of the original leaf area. The excised tissue was left overnight rehydrating through the cut end of the rachis, petiole or stem to determine the viability of photosynthesis upon rehydration. Rehydrating individual fronds or leaves simulated the effect of soil rewetting without having to rewater whole plants, thereby allowing the three individuals to be tracked through an entire drought cycle. The following morning, leaf gas exchange was measured on the rehydrated tissue, while the cut end remained under water; this measurement was performed to determine the Ψl at which different species sustained photosynthetic damage from water stress. Leaf recovery was defined quantitatively as the recovery of photosynthesis following rehydration of tissue overnight; leaf death was defined as the Ψl at which photosynthesis failed to recover above 0.5 μmol m−2 s−1 following overnight rehydration (A0). Once photosynthesis in rehydrated leaves failed to recover in individual plants rehydration of leaves did not continue, in Adiantum capillus-veneris only, drought continued until plants reached −6 MPa. Following leaf gas exchange measurement, the leaf area in the chamber was marked and the leaf or branch was taken for immediate Ψl measurement to assess effective rehydration of the tissue (> −0.2 MPa). In the desiccation-tolerant species Cheilanthes myriophylla, A recovered in rehydrated leaves when the plants were droughted beyond the limit of the pressure chamber (−10 MPa); at this point, all three individuals were rewatered and leaf gas exchange, Ψl and ABA concentration were quantified daily until gs had fully recovered to predrought levels. In the two epiphytic fern species, Rumohra adiantiformis and Pyrrosia lingua, the three droughted individuals were rewatered at different stages to observe the response of gs, A, Ψl and foliar ABA concentration over recovery. The first individual was rewatered on the first day on which stomata closed (stomatal closure was defined as stomatal conductance < 20% that of initial fully hydrated stomatal conductance), and the second and third individuals were rewatered over the extended period that followed when leaves were losing water but not declining in Ψl or showing signs of leaf death.

Turgor loss point and volume of available water until leaf death

Pressure–volume (PV) analysis was performed on at least five foliage samples, each from different plants, for each species to determine turgor loss point (Ψtlp) (Tyree & Hammel, 1972) as well as the volume of water in the leaf that was available until plants reached lethal relative water content (RWC). The night before measurements, foliage was bagged and excised under water to ensure that Ψl was high (> −0.05 MPa). First leaf weight (± 0.0001 g; Mettler-Toledo, Melbourne, Australia) followed immediately by Ψl was periodically measured over gradual desiccation in the laboratory; care was taken to ensure that no water was lost from the petiole of the stem during Ψl assessment. PV curves were constructed by plotting Ψl against RWC of an aggregation of points from each of the five leaves and Ψtlp determined by the inflection point of the relationship (Supporting Information Fig. S2); RWC was determined according to the following equation:

display math(Eqn 1)

(FW, the fresh weight (mass) of tissue; DW, the dry weight (mass) of tissue; TW, the weight (mass) of fully hydrated, turgid tissue.)

In all species, the volume of water available, as capacitance, between stomatal closure and leaf death was calculated by extrapolating from the PV relationship the RWC at which leaves died by using the Ψl at which A in rehydrated leaves failed to recover. In the desiccation-tolerant Cmyriophylla, three additional individuals were droughted to determine an accurate RWC and time until A in rehydrated leaves failed to recover and leaves died. The FW of leaves at this lethal RWC was then used to calculate the volume of water available until leaf death according to the following formula:

display math(Eqn 2)

(WW, the weight (mass) of leaf water at 100% RWC (g); FWdeath, the fresh leaf weight (mass) at the lowest recoverable RWC of leaves (g); DW, the dry leaf weight (mass) (g); M, the molar mass of water (g mol−1); LA, leaf area (m2).)

As a consequence of slow equilibration of Ψl in droughted leaves of the two epiphytic fern species (R. adiantiformis and P. lingua), PV curves were constructed using a modified method to observe the relationship between RWC and Ψl over the extended period of desiccation. Following the inflection point of the relationship between RWC and Ψl, leaves were sealed in a bag containing damp paper towel to maintain high humidity and allowed to equilibrate overnight. The following morning leaf weight and Ψl were concurrently measured and the leaves were then allowed to desiccate on the bench for c. 10 h before leaf weight and Ψl were again measured. The leaf was then bagged and allowed to equilibrate overnight. This cycle was continued for 5 d or until Ψl in equilibrated leaves failed to recover.

Stomatal response to vapour pressure difference

The response of stomata to atmospheric water stress (as opposed to soil water stress) was investigated by examining the response of gs to step-wise transitions in VPD in well-watered individuals. Three individuals of each species were acclimated to laboratory conditions overnight and the following day the response of gs to a sequence of VPD transitions (1–2–1 kPa) was measured on photosynthetic tissue using an infrared gas analyser (Li-6400; Li-Cor); leaf cuvette conditions were maintained at 22°C, and VPD was regulated at the required kPa by adjusting the humidity in the inlet air by bubbling the incoming air through water and adjusting the amount passing through a desiccant column containing calcium sulphate (390 μmol mol−1 CO2, light intensity 1000 μmol quanta m−2 s−1 and flow rate 500 ml min−1). All conditions in the chamber, including A and gs, were automatically logged every minute. Leaves were acclimated for 20 min following each VPD step.


Stomatal response to leaf water potential

The stomata of a diverse sample of fern and lycophyte species closed over a relatively small window of Ψl between −0.6 and −2.1 MPa when droughted (Fig. 1). A significant linear relationship between Ψl at stomatal closure and Ψtlp was observed when all species were compared (< 0.05; R2 = 0.616; data not shown). During the imposition of water stress and recovery from stress, gs of all fern and lycophyte species was strongly influenced by Ψl, with no hysteresis in gs observed in individuals that were rehydrated (Fig. 1).

Figure 1.

The relationship between stomatal conductance and leaf water potential pooled from three individuals of three mesophytic terrestrial fern species (a–c) and a lycophyte (d), two epiphytic fern species (e, f) and a desiccation-tolerant fern species (g) when droughted by withholding water (circles). Solid vertical lines represent the turgor loss point (Ψtlp) (see Supporting Information Fig. S2) and dashed red vertical lines represent the leaf water potential (Ψl) at which the assimilation rate did not recover in rehydrated leaves (A0). Square symbols represent measurements made on individuals that were rehydrated.

A small margin between stomatal closure and leaf death in ferns and lycophytes

Terrestrial fern and lycophyte species typically had a small Ψl margin between stomatal closure and leaf death from water stress, except for the dessication-tolerant species C. myriophylla, which could survive extremely low Ψl (Figs 1, 2). In the mesophytic lycophyte Selaginella kraussiana, the Ψl margin between stomatal closure and leaf death was only 0.01 MPa, while in the terrestrial fern species this margin ranged from 0.19 MPa in A. capillus-veneris to 0.4 MPa in Pteridum esculentum (Fig. 2). The two epiphytic fern species had slightly larger safety margins between stomatal closure and irreversible photosynthetic damage, with 0.42 and 0.92 MPa for Plingua and R. adiantiformis, respectively (Fig. 1).

Figure 2.

The assimilation rate of excised leaves rehydrated overnight to full hydration (>-0.2 MPa) from known droughted leaf water potentials, in three mesophytic terrestrial drought deciduous ferns (a–c), a lycophyte (d), two epiphytic ferns (e, f) and a desiccation-tolerant fern (g). Solid vertical lines represent the leaf water potential at which stomata close (defined as < 20% of initial stomatal conductance).

There was very little difference between the Ψl values when drought-stressed leaves died in all fern species (c. 0.8 MPa lower in the epiphytic species; Fig. 1). In terrestrial fern and lycophyte species, > 80% of leaves died as a result of drought at a Ψl between −1.2 MPa (Acapillus-veneris and Skraussiana) and −2.2 MPa (P. esculentum), while the leaves of the epiphyte Radiantiformis perished when droughted to –2.2 MPa and > 50% of leaves of Plingua died when droughted to −3 MPa (Figs 1, 2). Importantly, there was no evidence of a feedback in stomatal control between A and gs, whereby reduced or damaged A as a result of drought caused stomata to remain closed on rehydration; in recovered leaves with depressed photosynthesis, gs remained relatively high (Fig. S3). This resulted in highly inefficient loss of water in plants recovering from drought damage.

The stomatal response of ferns and lycophytes to drought is not mediated by ABA

In all species, ABA levels were augmented in response to water deficit, but in most species this augmentation did not correlate with changes in gs (Fig. 3), instead occurring after stomatal closure and Ψtlp; following Ψtlp, the levels of ABA in the leaves of the terrestrial and desiccation-tolerant species increased to levels exceeding 1000 ng g−1 FW (Fig. 4). Similar relationships between gs, Ψl and ABA level were observed when ABA levels were expressed in terms of DW calculated from the relationship between RWC and Ψl determined from PV curves (Figs S4, S5). Only two species showed an overlap between the phase of ABA rise and stomatal closure. In Plingua, this relationship occurred over extremely low foliar ABA levels (< 20 ng g−1 FW) (Fig. 3) and in Radiantiformis the recovery of gs when droughted plants were rewatered was not significantly influenced by the ABA level in the leaf (Fig. 1).

Figure 3.

The relationship between foliar abscisic acid (ABA) level (ng g−1 fresh weight) and stomatal conductance before stomatal closure in three mesophytic terrestrial fern species (a–c), a lycophyte (d), two epiphytic fern species (e, f) and a desiccation-tolerant fern species (g) when droughted (circles) and recovered (squares). Significant regressions (solid lines) and R2 are shown where present.

Figure 4.

The relationship between foliar abscisic acid (ABA) level (ng g−1 fresh weight) and leaf water potential (–MPa) showing a clear relationship between water stress and ABA levels, pooled from three individuals of three mesophytic terrestrial fern species (a–c), a lycophyte (d), two epiphytic fern species (e, f) and a desiccation-tolerant fern species (g) when droughted by withholding water (circles). Solid vertical lines represent the leaf water potential (Ψl) when stomata closed (see Fig. 1) and dashed red vertical lines represent Ψl at which the assimilation rate did not recover in rehydrated leaves (A0). Square symbols each represent measurements made on individuals that were rehydrated.

Predictable responses of fern and lycophyte stomata to atmospheric humidity

All fern and lycophyte stomata were highly sensitive to atmospheric humidity (Fig. 5). Following an increase in VPD from 1 to 2 kPa, the fern and lycophyte stomata closed following a single exponential decay function, reaching a new steady-state gs in < 20 min, with no hydropassive, wrong-way response (Fig. 5). The degree of stomatal closure in response to this increase in VPD varied between species, with the stomata of C. myriophylla closing by only 30.95 ± 3.65% while the stomata of S. kraussiana closed by 66.26 ± 4.37% (Table 2). Stomatal responses to a subsequent decrease in VPD from 2 to 1 kPa resulted in stomata reopening that followed the same exponential trajectory with no evidence of hydropassive closure (Fig. 5, Table 2). The stomatal responses of ferns and lycophytes to changes in atmospheric humidity were highly predictable (Fig. S6).

Table 2. The mean percentage reduction in stomatal conductance (gs) following an increase in vapour pressure deficit from 1 to 2 kPa and the mean gs (as a percentage of an initial gs at 1 kPa) after a decrease in vapour pressure difference (VPD) from 2 to 1 kPa in seven fern and lycophyte species (n = 3)
Species% stomatal closure (1–2 kPa)± SE% gs of initial (2–1 kPa)± SE
Adiantum capillus-veneris 47.504.68105.453.81
Cheilanthes myriophylla 30.953.6599.850.63
Dicksonia antarctica 33.533.82100.303.26
Pteridium esculentum 35.375.31102.652.60
Pyrrosia lingua 32.895.7098.041.02
Rumohra adiantiformis 42.609.87106.751.53
Selaginella kraussiana 66.264.37100.231.85
Figure 5.

Representative responses of stomatal conductance (mol m−2 s−1) to step-wise transitions in vapour pressure deficit (kPa) in six fern species (a–f) and a lycophyte (g).

A limited diversity of fern and lycophyte strategy in response to water stress

There was a large variability in the minimum RWC that the leaves of fern and lycophyte species could survive (Fig. 6). An increase in desiccation tolerance increased the volume of internal leaf water available to the plant during drought (Fig. 6). The leaves of terrestrial fern and lycophyte species were unable to survive desiccation to an RWC < 85% and had only a minimal volume of water available before leaf death, ranging from 46.4 mmol H2O m−2 in Acapillus-veneris to 492.1 mmol H2O m−2 in Skraussiana (Fig. 6). The leaves of the least resistant species, Dicksonia antarctica, were unable to survive an RWC below 97.2%, while the leaves of the lycophyte Skraussiana survived to an RWC of only 87.3% (Fig. 6). The limited tolerance of leaf desiccation was a trait characteristic of the mesophytic terrestrial species only (Fig. 6). The leaves of the two epiphytic fern species were characterized by a tolerance of low RWC, with leaves of Radiantiformis surviving to an RWC of 41.7% and those of Plingua to 26.7%; this tolerance of low RWC resulted in a large volume of available water in the leaf before death as a result of drought (Fig. 6). The desiccation-tolerant fern species Cmyriophylla was able to survive extremely low RWCs (c. 10%) and, like the two epiphytic species, had a large volume of internal leaf water (3824.8 mmol H2O m−2 ± 44.07) available before leaf death (Fig. 6). The size of the internal leaf water buffer was a product of leaf volume and desiccation tolerance and was highly correlated with the duration of drought that could be tolerated (= 0.001; R2 = 0.902; Fig. 6).

Figure 6.

In a wide diversity of fern and lycophyte species, the tolerance of leaves to minimal relative water contents influenced the volume of available leaf water each species could use before leaf death (a). There was a significant relationship between available leaf water until leaf death and the number of days between stomatal closure and leaf death in ferns and lycophytes (P=0.001; days to leaf death = 67.34 × (1-e(−0.0004 × available leaf water)); R2=0.902; n = 7 species) (b). Each symbol represents the mean ± SE (n = 5) for an individual species (Ac, Adiantum capillus-veneris; Cm, Cheilanthes myriophylla; Da, Dicksonia antarctica; Pe, Pteridium esculentum; Pl, Pyrrosia lingua; Ra, Rumohra adiantiformis; Sk, Selaginella kraussiana).


Predominantly passive stomatal control in ferns and lycophytes

The stomatal behaviour among fern and lycophyte species with diverse growth habits and ecologies was found to be remarkably conservative, with the stomata of all species predominantly controlled by a passive response to leaf water status (Brodribb & McAdam, 2011). Such uncomplicated regulation of plant hydration appears to result in a very simple drought adaptation strategy whereby ferns and lycophytes modify water storage and desiccation tolerance as a means of extending drought survival time (Fig. 6). This canalized adaptive pathway in response to drought must have significant ramifications for the ability of ferns and lycophytes to compete with seed plants, which have complex stomatal control and diverse water management strategies (Tardieu & Simonneau, 1998). The complexity of a largely metabolic regulation of stomatal aperture in seed plants can be recognized by the variability in stomatal responses to changes in VPD, from highly sensitive plants that elicit ‘feed-forward’ responses to increases in VPD, to plants that are relatively insensitive (Farquhar, 1978; Franks & Farquhar, 1999; Mott & Peak, 2012). This diversity in response is unlike the predictable responses of fern and lycophyte stomata to changes in VPD which are regulated by a passive response to leaf water status (Fig. 5) (Brodribb & McAdam, 2011; McAdam & Brodribb, 2012a).

Physiological and morphological adaptations alone are essential for ferns and lycophytes to survive dry environments

Many epiphytic fern and lycophyte species are exposed to dry and/or fluctuating environments, which raises the question of how these plants compete with seed plants when their stomata are predominantly regulated passively by leaf water status (Watkins & Cardelús, 2012). Only a small amount is known about the physiological adaptations of epiphytic fern species. These include a lower leaf hydraulic conductance compared with terrestrial ferns (Watkins et al., 2010), desiccation-tolerant gametophytes (Watkins et al., 2007a) and occasionally Crassulacean acid metabolism (CAM) (Ong et al., 1986; Holtum & Winter, 1999). Interestingly, the adaptation of CAM in tropical epiphytic ferns is not associated with an enhanced tolerance to long periods of drought as seen in angiosperm CAM species; rather, CAM in epiphytic ferns improves carbon and water balance over short (diurnal) periods of water stress (Ong et al., 1986).

While we found that the function of stomata in a diversity of fern and lycophyte species was the same, species from drier environments (epiphytes and desiccation-tolerant species) were able to survive for a long period of time with stomata closed before leaves died, unlike the mesophytic terrestrial species (Fig. 6). This increased tolerance of epiphytic fern species to drought stress was a result of the combination of a physiological tolerance of low RWC and anatomical modification of leaves to increase the amount of water available before leaf death during drought (Fig. 6). Epiphytic ferns are capable of developmentally adapting a reliance on either of these two strategies to survive water stress, with Asplenium auritum being desiccation-tolerant as a young sporophyte and maturing into a drought-avoiding, high-capacitance plant as size increases (Testo & Watkins, 2012). Epiphytic ferns are known to have large water storage capabilities in leaves (as in the genus Pyrrosia with specialized hydrenchyma cells in the leaf (Ong et al., 1992)) as well as water-storing rhizomes and stems (Dubuisson et al., 2009). Available water, as defined here by the leaf capacitance multiplied by the viable range of leaf water content after stomatal closure, gives a volume of water that can supply the slow leakage of water from plants with closed stomata assuming hydraulic isolation from the soil (Linton & Nobel, 1999). The significant relationship we found between the time to leaf death and leaf available water highlights the fundamental importance of the interaction between the storage and partitioning of leaf water and the time to leaf death as a result of drought. Epiphytic ferns are known to maintain a Ψl low enough to close stomata for an extended period of time before leaves die (Ong et al., 1992). The combination of these two strategies, a physiological tolerance of low RWC and partitioning and isolation of large volumes of water from the stomata, is a common feature of epiphytic ferns and lycophytes, which are largely characterized by succulent leaves or rhizomes and can survive in environments with highly sporadic water availability (Watkins & Cardelús, 2012). This apparent dependence on water storage and desiccation tolerance contrasts with seed plants which, with a predominantly metabolic regulation of stomatal control, have the opportunity of modifying stomatal behaviour through hormones such as ABA during drought stress (Wilkinson & Davies, 2002).

High physiological tolerance of leaves to low RWC is common in (fern) epiphytes and exemplified by desiccation-tolerant and poikilohydric species (Hietz, 2010). Within ferns, desiccation tolerance is widely represented, occurring in most major lineages, with estimates ranging from 5 to 10% of fern species being desiccation-tolerant or poikilohydric, while in the lycophytes two of the three extant families have desiccation-tolerant species (Hietz, 2010). The over-representation of desiccation tolerance in ferns and lycophytes relative to seed plant lineages (Oliver et al., 2000) is probably a reflection of the limited options of water management strategy provided by a predominantly passive stomatal control. The stomatal responses of desiccation-tolerant and epiphytic ferns are the same as those of mesophytic terrestrial ferns.

Implications for the evolution of land plants

In ferns and lycophytes, the passive stomatal control of leaf hydration still provides an efficient means of regulating leaf hydration in response to changes in water stress (Brodribb & McAdam, 2011) (Fig. 5). Yet this mechanism of stomatal control in ferns and lycophytes does not offer plants the ability to entertain a diversity of responses to soil water content through changes in metabolic regulation of stomata like seed plants (Tardieu & Simonneau, 1998). It is possible that the evolution of an increased regulation of stomata by ABA in seed plants may have been driven by the selective pressure for a stomatal strategy that could enhance the survival of plants over both short- and long-term periods of soil water stress; this would offer an advantage over the changes in morphology or desiccation tolerance required by ferns and lycophytes (Fig. 6). This might have been particularly advantageous during the drying climate of the early Permian (DiMichele & Aronson, 1992). A heightened sensitivity of stomata to increased ABA concentrations, augmented by drought, offers seed plants a dynamic, stomatal-mediated response to episodes of soil drought that occurs over relatively brief periods (hours to days). The predominance of a metabolic regulation of stomata particularly by ABA in seed plants additionally allows the leaves of species that do not have physiological and morphological adaptations to dry environments to survive periods of soil drought by modifying the sensitivity of stomata to small changes in Ψl (Umezawa et al., 2004; Fujita et al., 2005), unlike the mesophytic terrestrial ferns and lycophytes (Fig. 1). Adopting the strategy of ABA-regulated stomatal control in seed plants is also likely to be important both for optimizing water use during fluctuating soil water availability (Chaves et al., 2003) and during the recovery from stress in facilitating the repair of xylem tissue after drought (Lovisolo et al., 2008).

The response of fern and lycophyte stomata to drought stress suggests that stomatal closure occurs with a relatively small Ψl margin before leaf death in these species (Fig. 1). Interestingly, ferns and lycophytes close their stomata with a similar Ψl margin before leaves lose hydraulic conductivity (Brodribb & Holbrook, 2004), suggesting that cavitation-induced losses in hydraulic conductivity probably lead to plant death during drought. Further investigation is required to develop a more comprehensive understanding of the role of xylem and leaf hydraulics in influencing the survival and ecology of ferns and lycophytes.

Suggested roles for ABA in droughted ferns and lycophytes

While the augmentation of ABA levels occurs in fern and lycophyte leaves during drought (Fig. 4), this increase is not typically associated with the closing of stomata (Fig. 3), which is traditionally held to be the primary function of this phytohormone (Wilkinson & Davies, 2002). Stomatal closure during drought in the fern and lycophyte species observed in this study occurred over a range of Ψl that would elicit significant passive stomatal closure provided guard cell turgor at 80% maximum aperture was 2 MPa, as observed in the fern Nephrolepis exalata and lycophyte Huperzia prolifera (Franks & Farquhar, 2007). Rather than increasing during the passive stomatal closure during drought, ABA levels in ferns and lycophytes instead increased largely after plants had been stressed beyond Ψtlp (Fig. 4). This could be a consequence of ABA in the leaf being unbound from a fettered state when the leaf loses turgor; this free ABA then drives the release of much higher levels of ABA in the leaf (Georgopoulou & Milborrow, 2012). There are a number of functional roles for ABA beyond active stomatal control that have been suggested in ferns and lycophytes, not all associated with the sporophyte, such as the regulation of sex expression in the gametophyte (Banks et al., 1993). In most land plants, including ferns and lycophytes, ABA enhances the survival of tissues grown in vitro and is instrumental in initiating tissue desiccation tolerance (Bagniewska-Zadworna et al., 2007). In the diversity of droughted fern and lycophyte species we examined, the level of ABA in the leaf dramatically increased when leaves were dehydrated beyond the critical Ψl at which point photosynthesis was unable to recover and leaves died (Fig. 4). It is possible that the augmentation of ABA in ferns and lycophyte occurring beyond Ψtlp and the recovery of photosynthesis reflect a role of ABA in initiating leaf senescence; this may explain why the concentrations of foliar ABA within species of ferns and lycophytes were highly variable during drought. High concentrations of ABA in seed-plant leaves have been shown to initiate leaf senescence pathways following stress (Lee et al., 2011). Senescence of droughted leaves of ferns and lycophytes would be particularly important because the stomata of ferns and lycophytes reopen on leaf hydration, regardless of photosynthetic damage (Fig. S3), constituting a substantial transpiration cost without photosynthetic benefit (McAdam & Brodribb, 2012b). It would be highly advantageous to initiate leaf senescence and shed leaves that have incurred photosynthetic damage following drought stress, through an increase in ABA level. Leaf senescence triggered by ABA in droughted ferns and lycophytes would additionally prevent hydraulic damage to the rhizome in recovering plants. This role for ABA may explain the prevalence of drought deciduousness and the ability of many terrestrial, mesophytic and epiphytic fern species to resprout from a protected meristem following drought (Hietz, 2010).


The stomata of all ferns and lycophytes studied showed a predominantly passive regulation of stomata by leaf hydration, regardless of ecology or morphology. While this mechanism of stomatal control means that ferns and lycophytes have an efficient regulation of water loss under fluctuations in vapour pressure deficit and also close in response to soil drought, it results in canalized adaptation when it comes to surviving extended periods of water stress. While Ψl is sufficient to close the stomata of ferns and lycophytes when water-stressed, ferns and lycophytes must make significant morphological and physiological adaptations over generational time-scales in order to adapt and survive for extended periods of drought or in water-limiting environments.


We thank John Ross and Noel Davies for advice and assistance in the extraction, purification and quantification of ABA samples, as well as three anonymous reviewers and the editor for their insightful improvements to the manuscript. This work was supported by Australian Research Council grants DP0878177 and FT100100237 (T.J.B.).