Leaf rosette closure in the alpine rock species Saxifraga paniculata Mill.: significance for survival of drought and heat under high irradiation

Authors


Correspondence: G. Neuner. Fax: 0043 512 5072715; e-mail: Gilbert.Neuner@uibk.ac.at

ABSTRACT

The significance of leaf rosette closure for survival of drought and heat under high irradiation on alpine rock sites was investigated in the cushion forming rosette plant, Saxifraga paniculata Mill. With decreasing water content the leaves fold over the rosette centre reducing reversibly the evaporative leaf surface area by 80%. Internal water redistribution driven by an osmotic gradient from older to younger leaves occurs. The oldest leaves dry out to promote the survival of the individual. Leaf temperatures above 45 °C (which match heat tolerance limits 45–57 °C; LT50) co-occurred with low substrate water potentials (less than – 0·5 MPa) on 11·3% of summer days. Shading by leaves can be crucial to surviving high temperatures as it keeps the rosette centre up to 10 °C colder. Mutual shading prevented sustained drought-induced photoinhibition in upper leaf surfaces at relative water contents below 60%. In exposed lower leaf surfaces restoration of photosystem II took several days. Leaf temperatures above 40 °C (21·3% of summer days) induced photoinhibition in situ. Periods with sufficient water supply can be fully utilized as rehydration is fast (< 12 h) and exposes the upper leaf surfaces that showed only minor photoinhibition. By reversible leaf rosette closure environmental extremes that otherwise could exceed tolerance are efficiently avoided.

INTRODUCTION

Alpine rock habitats are extreme environments for phanerogamic plant life. Severe drought damage to a range of alpine rock species has been reported during particularly dry summers ( Wetter, Schade & Firbas 1917). During water shortages leaf temperatures of up to 54 °C have been measured (Sempervivum montanum;Larcher & Wagner 1983). Leaf temperatures can occasionally even exceed the heat tolerance limit of compact, low rosette plants on bare soil or rock sites (see Körner & Larcher 1988). Heat and drought are known to sensitize photosynthesis to photo-inhibition (see Powles 1984) that in turn can affect carbon acquisition and survival (see Long, Humphries & Falkowski 1994).

Saxifraga paniculata Mill. is representative of the alpine rock vegetation ( Schröter 1926). The species grows in compact acaulescent leaf rosettes that often form loose cushions. It survives even on shallow rock hollows filled with only a thin substrate layer. Field observations indicated that S. paniculata is capable of reversible leaf cupping during drought that causes a successive closure of the leaf rosette. We aimed to examine the role of this structural alteration of the leaf rosettes of S. paniculata for survival of the rock site-specific environmental extremes.

Leaf cupping and mutual shading of the leaves during drought reduces the incident radiation and could result in more favourable leaf temperatures. Beneficial thermal effects of structural alterations of leaf rosettes have been reported for giant rosette plants that perform nyctinastic leaf movements ( Hedberg 1964; Smith 1974). Folding of the leaves insulates the apical bud, protecting it from low subzero night-time air temperatures that regularly can occur at night in tropical alpine climates ( Beck et al. 1982 ). For S. paniculata the closure of the leaf rosette could provide thermal insulation to young leaves from otherwise critically high daytime temperatures.

Reduction of incident radiation by leaf cupping and mutual shading of the leaves could also be an efficient way to minimize photoinhibition during high temperature and water stress (see Powles 1984). During drought photo-inhibition may be only a minor important component for carbon acquisition as earlier stomatal closure will override the direct effects of photoinhibition ( Boyer, Armond & Sharp 1987). The speed of recovery from drought-induced photo-inhibition could be, however, very important to the long-term carbon balance, particularly, at the rock site under conditions of frequent stress exposure.

Leaf cupping during severe drought ends in a ball-like structure of the leaf rosette of S. paniculata that is covered by often already dead old leaves. For several species water movement from older to younger leaves are reported as an important feature of drought adaptation ( Schäfer & Lüttge 1987; Tüffers, Martin & vonWillert 1995). In S. paniculata coverage by dead old leaves could be indicative for such a water movement to young leaves during drought. Shifting water to the more protected leaves in the covered rosette centre could help to conserve water during severe drought periods.

In this study we investigated the potential role that leaf cupping and concomitant successive closure of the acaulescent leaf rosettes of S. paniculata play in surviving the environmental extremes of alpine rock sites during summer. In field investigations the environmental conditions were monitored at five natural sites to assess the frequency and the extent of environmental extremes. The results were complemented by single factor experiments under controlled environmental conditions.

MATERIALS AND METHODS

Plant material and experimental sites

Saxifraga paniculata Mill. is an evergreen acaulescent rosette plant that forms loose cushions. The species is a typical rock plant of the Northern Hemisphere. Saxifraga paniculata can be found from lowlands up to the alpine and nival zone (3415 m; Reisigl & Pitschmann 1958).

Field investigations were conducted at five sites between 1950 and 2200 m a.s.l. in the Central Alps to the South of Innsbruck. All sites are typical habitats of S. paniculata, namely shallow hollows on big rocks, with only a thin layer of organic substrate (up to 5 cm thick). Precipitation is the only water source. Water storage capacity of the thin substrate layer is very low.

Photosynthetic photon flux density (PPFD; μmol photons m−2 s−1), temperatures, soil moisture and precipitation data were collected on CR 10 data loggers (Campbell Scientific, Logan, UT, USA). Leaf temperatures were monitored using type T fine wire thermocouples. The solder junctions (diameter 0·2 mm) were fixed to the leaf surface using a special adhesive which is permeable to vapour and air (Transpore; BM-Austria GmbH., Vienna, Austria). Temperatures in the rosette centre were determined by inserting the solder junction of the thermocouple between the innermost, youngest leaves. Air temperatures and temperatures within the thin substrate layer were monitored using thermistors (107 temperature probe; Campbell Scientific). The air temperature sensors were mounted inside an unaspirated radiation shield 2 m above ground. Precipitation was measured with a tipping bucket raingauge (ARG100; Campbell Scientific). Soil moisture was registered at two microsites at 1950 m with gypsum blocks (cylindrical soil moisture block 227, Campbell Scientific). The data loggers were programmed to sample every 30 s, and averages were calculated every 30 min.

Plants for laboratory experiments

For experiments under controlled conditions, cushions of S. paniculata were collected from several sites (between 1700 and 2500 m a.s.l.) in the Central Alps. Individual vegetative leaf rosettes were carefully separated from the cushion and potted into plastic containers (7 cm × 7 cm × 10 cm) that were filled with potting soil mixed in equal parts with sand. The plants were cultivated under the natural environmental conditions in the Botanical Garden of the University of Innsbruck at 600 m and on Mt. Patscherkofel at 1950 m. The single leaf rosettes started soon to form cushions. In the experiments with potted plants only such plants were used that had been in cultivation for at least 6 months.

Documentation of leaf cupping and leaf rosette closure

Leaf cupping was documented by measuring the distance between each single leaf tip and the rosette centre. The mean value of all leaf tip to rosette centre distances is a measure of the aperture of the leaf rosette. This value can only be used to monitor changes in the same individual over time as the rosette diameter and leaf length can vary to a large extent between individuals. Therefore, we compared individuals using the relative rosette aperture, i.e. the mean value of all distances from the leaf tip to the rosette centre expressed as a percentage of the mean leaf length of all leaves of the rosette.

Individual leaf rosettes were photographed together with a ruler. The photos were then digitized with a scanner (HP Scanjet 4p; Hewlett-Packard Company, Greeley, CO, USA). For determination of distances in the leaf rosette the images were further processed with image analysing software (O PTIMAS; Optimas Corp., Seattle, WA, USA). After calibration of the image and definition of the rosette centre as a reference point all leaf tips were marked on the image with mouse clicks. The distances were then determined automatically. At the end of the experiment all leaves were scanned, leaf length was determined with the line morphometry procedure of O PTIMAS and relative rosette aperture was calculated.

Desiccation treatment

Desiccation experiments were conducted under controlled conditions in a growth cabinet (Ecophyt VEPHL 5/1350; Heraeus Vötsch, Balingen, Germany): Air temperature was maintained at 15 °C and relative air humidity was 50 ± 5%. Air flow was 0·2 m s−1. Plants were either desiccated in darkness or under the following light conditions: 200, 400 or 1200 μmol photons m−2 s−1.

Water content

Fresh weight (FW) was determined either separately for each single leaf or for whole leaf rosettes. Leaves or leaf rosettes were dried for at least 1 week at 80 °C, then weighed again to determine their dry weight (DW). Water content was calculated as (FW – DW)/DW.

At several stages of dehydration the actual weight and dry weight was determined for each single leaf of the rosettes beginning with the oldest (outermost) one. A typical leaf rosette consisted of about 40 up to 65 leaves. Five leaves in succession were combined to a leaf age class. Mean values were then computed for each leaf age class using values from three rosettes (n = 15).

Critical water content

For determination of the critical water content (LD10) leaves at different stages of desiccation were collected and placed in a closed plastic box with water for rehydration. After black necrosis caused by dehydration damage had completely developed, the leaves were scanned with a conventional scanner (HP Scanjet 4p; Hewlett-Packard Company). The percentage of the leaf area that had been damaged was determined with image analysing software O PTIMAS (Optimas Corp.) and plotted against water content. By fitting a classic logistic function to the data (P.-Fit; Biosoft, Ferguson, MO, USA) the critical water content (LD10, the water content at 10% desiccation damage) could be determined graphically from the curve.

Water potential

For determination of actual osmotic water potential (Ψo(act)), leaves of different age from dehydrated and well-watered leaf rosettes were collected and tightly wrapped in aluminium foil. Immediately afterwards the leaves were killed by immersion in liquid nitrogen (– 196 °C). The value of Ψo(act) of killed leaves was then measured in calibrated C-52 sample chamber hygrometers connected to a HR-33T Dew Point Microvoltmeter (Wescor, Logan, UT, USA).

In vivo chlorophyll fluorescence measurements

In vivo chlorophyll fluorescence was measured with a Mini-PAM chlorophyll fluorometer system (H. Walz, Effeltrich, Germany). After a dark adaptation period of 30 min minimum fluorescence, F0, was determined. Maximum fluorescence, FM, was determined during a subsequent saturating light pulse. Potential efficiency of photosystem II (PSII), FV/FM, was calculated as (FMF0)/FM. During desiccation experiments chlorophyll fluorescence of single leaves (upper (adaxial) or lower (abaxial) leaf surface) was measured with the Miniature Fiberoptics probe (Mini-PAM/F1; H. Walz) that allows small spot measurements with an active diameter of 2 mm. Chlorophyll fluorescence measurements at the individual level (rosette surface) were conducted with the standard Fiberoptics probe (Mini-PAM/F; H. Walz) with an active diameter of 6 mm.

Statistical data analysis

Differences in water content and Ψo(act) with leaf age class and stage of dehydration were investigated using SPSS software (SPSS Inc., Chicago, IL, USA). After the data had passed the Kolmogorov–Smirnov test, the significance of difference was determined by analysis of variance and the Bonferroni test.

RESULTS

Leaf rosette closure during desiccation

The leaves of S. paniculata perform hydronastic movements as the water content decreases. By leaf cupping the leaf tips move radially closer to the centre of the rosette markedly reducing the relative rosette aperture ( Fig. 1). Severely dehydrated leaf rosettes have a ball-like structure with the exposed leaf surface reversibly reduced to the surface of this ball, i.e. a surface reduction of 80%. Leaf cupping and leaf rosette closure occurred independent from the light environment. It was observed during dehydration in complete darkness and under various light intensities up to 1200 μmol photons m−2 s−1. Exposure to heat (+ 40 °C) did not affect rosette aperture when the plant was supplied with sufficient water and kept in darkness (data not shown). Rosette leaf cupping was completely reversible and opening of the leaf rosette occurred particularly fast. Under the experimental conditions leaf rosettes opened within less than 1 d after re-watering.

Figure 1.

Relative rosette aperture (%) during a drying cycle at different water contents of leaf rosettes of potted S. paniculata cushions. Desiccation proceeded under controlled conditions at 15 °C (●) in darkness or under the following irradiations: (○) 200 (∇) 400 or (▵) 1200 μmol photons m−2s−1. Water uptake and opening of the closed rosettes occurred within 1 d in response to water supply even though the leaf rosettes had been dehydrated close to the critical water content (LD10 = 1·3). Values are means of five leaf rosettes.

Rosette leaf cupping was also observed at the field sites in response to natural drought periods ( Fig. 2). After 6 d without precipitation the substrate started to dry out. Substrate water potential was below – 1·5 MPa 11 d after the last rain. Rosette aperture had decreased by 38% during this natural drought. Closed leaf rosettes in the field opened as fast as under controlled conditions. Two days after sufficient rain rosette aperture was completely restored to that before the drought period.

Figure 2.

Mean distance of the leaf tips from the rosette centre (●) measured at three different dates in response to a natural drought period in summer 1996. The period without precipitation (grey bars) caused a decrease of (○) substrate water potential below – 1·5 MPa and a reduction of the rosette aperture by 38% within 3 d.

After 1 month without precipitation in October 1995 living, closed leaf rosettes of S. paniculata were covered by already dead leaves at several alpine rock sites. In response to water supply they readily opened within less than 1 d, exposing the unaffected leaves of the rosette centre.

Benefits of leaf rosette closure for drought survival

Young leaves in the leaf rosette centre are better protected from water loss during desiccation ( Fig. 3). Water loss differed significantly with leaf age (P < 0·01). Under the water-saturated conditions the oldest leaves in the leaf rosette have a significantly higher water content than the younger leaf age classes ( Fig. 3a). During desiccation these leaves lose significantly more water than younger leaves ( Figs 3b–d). The reduced water loss of young leaves in the centre of the leaf rosette is besides their lesser exposure achieved by an osmotic gradient: in dehydrated leaf rosettes young leaves have a significantly (P < 0·01) more negative actual osmotic water potential (Ψo(act)) than old leaves ( Table 1).

Figure 3.

Water content of different leaf age classes of leaf rosettes at different stages of desiccation during a drying cycle under controlled conditions at 15 °C air temperature and moderate irradiation of 400 μmol photons m−2 s−1: (a) before and (b) 6, (c) 14 and (d) 26 d after initiation of the drying cycle. Water loss in the oldest leaves is significantly higher (P < 0·01) than in younger leaves. The oldest leaves are the first to be drought injured. Values ± SD are means of five leaves per leaf age class and three individual leaf rosettes.

Table 1.  In well-watered Saxifraga paniculata plants actual osmotic water potential (Ψo(act)) is independent from leaf age but in plants dehydrated close to the critical water content (< 2·0) it is significantly lower in young leaves of the rosette centre than in older leaves. Significance of the difference between means of Ψo(act) of leaves of different age was tested by analysis of variance and the Bonferroni test
Ψo(act)± SE
 Well-wateredDehydrated
  1. a,b, different letters indicate a significant difference at P < 0·01.

Young leaves– 1·7 ± 0·4a– 3·0 ± 0·1a
– 1·6 ± 0·1a– 2·6 ± 0·2ab
 – 1·7 ± 0·0a– 2·3 ± 0·1b
– 1·4 ± 0·1a– 2·3 ± 0·1b
Old leaves– 1·5 ± 0·1a– 2·3 ± 0·0b

The oldest and outermost leaves of the rosette are the first to be drought injured at a LD10 value of 1·3 . Such desiccated old leaves, even if they are already dead, still tightly cover the living leaves in the rosette centre. Stomata are closed at this stage of dehydration (data not shown). In addition to the reduction in the area of evaporative surface by leaf rosette closure coverage by dead leaves can help to protect the living young leaves from further cuticular water loss.

Heat load under natural conditions

During summer, daily air temperature maxima (half an hour means) at 1950 m (a.s.l.) ranged between 2·3 and 26·3 °C and occurred most frequently (44·2%) in the range between 15 and 20 °C ( Fig. 4a). During daytime the leaves of S. paniculata were significantly warmer than air ( Fig. 4b). When relative frequencies are determined from data of five natural rock sites and three different years (1996–98), daily leaf temperature maxima higher than 35 °C were observed on 32·8% of all summer days. Actual leaf temperatures were even up to 36 °C higher than air. The heat tolerance limit (LT50) of the species is between 45 and 57 °C ( Table 2). On 11·3% of all summer days leaf temperature maxima were higher than 45 °C. In summer 1996 at one particular site the rosettes survived leaf temperatures of 55 °C, although a single rosette died at 56 °C.

Figure 4.

(a) Relative frequency (%) of daily air temperature maxima (5 °C temperature classes; grey bars) at 1950 m (a.s.l.) determined for the summer period (June–July–August) of three successive years (1996–98). (b) Relative frequency (%) of daily leaf temperature maxima (5 °C temperature classes; black bars) of S. paniculata determined for the summer period (June–July–August). Temperature records are from five different microsites between 1950 and 2600 m and different years (1996–98). Mean substrate water potential (MPa) (hatched bars ± SE) was significantly lower during periods when high leaf temperatures were recorded.

Table 2.  Heat tolerance (LT50 at °C) of Saxifraga paniculata leaves from different authors
LT50Author
  1. *after heat hardening, **minimum/***maximum heat tolerance limit measured at field sites.

44·7Kjelvik (1976)
45·5/46·5Sapper (1935)
46·3Gauslaa (1984)
50·7/56·0*Neuner, unpublished
52·0/54·0Biebl & Mair (1969)
47·5**/57·0***Buchner, unpublished

Heat and drought co-occurred at the sites. Leaf temperatures critical to survival (> 45 °C) were measured only during drought periods when mean substrate water potentials dropped below – 0·5 MPa (see Fig. 4b).

Benefits of leaf rosette closure with respect to critical heat load

Leaf rosette closure cools the rosette centre during drought. Leaves in the shaded centre of closed rosettes remained up to 5·3 °C colder than the fully exposed covering old leaves ( Fig. 5a). The cooling effect increased with increasing leaf temperatures. A direct comparison of leaf temperatures in open and closed leaf rosettes revealed that shaded leaves in the centre of closed leaf rosettes remained even up to 9·9 °C colder than leaves in the centre of open leaf rosettes during exposure to the same environmental conditions at 1950 m ( Fig. 5b). This difference was only observed when due to dry soil and stomatal closure latent heat loss was low in open leaf rosettes. Similar results were obtained when leaf cupping was artificially prevented during dehydration: the rosette centre remained 5·1 °C warmer than that of closed, similarly dehydrated leaf rosettes during a clear summer day at 1950 m. Considering the naturally occurring leaf temperature maxima that are frequently (11·3%) in the range of the heat tolerance limit of the species, shading by leaves can protect the young leaves and the rosette centre from an otherwise critical heat load.

Figure 5.

Leaf temperatures during daytime (PPFD > 20 μmol photons m−2s−1) in the shaded centre of closed rosettes compared (a) to leaf temperatures on the exposed surface and (b) to leaf temperatures in the exposed centre of open rosettes. Temperature measurements were conducted on S. paniculata leaf rosettes exposed to similar natural environmental conditions at 1950 m in summer 1995. Latent heat loss in open leaf rosettes was comparable to that of the closed rosettes due to dry soil and closed stomata.

Benefits of leaf cupping and rosette closure to avoid photoinhibition

As the leaves move during desiccation the angle of each single leaf in the rosette gets successively steeper creating a lower angle of incidence and so reducing light absorption during stress exposure. At lower water contents the leaves turn upside down exposing the lower (abaxial) side of the leaves. This shades the photosynthetically active palisade parenchyma layer of the upper leaf surface and protects it from severe photoinhibition at low water contents ( Fig. 6a). At water contents below 2·5, after about 40% of the tissue water had been lost, sensitivity of PSII to light exposure was significantly enhanced even under the moderate irradiation of 400 μmol photons m−2 s−1. In lower (abaxial) sides of the leaves potential efficiency of PSII, FV/FM, decreased more drastically down to a value of 0·46 compared with the upper leaf sides (0·71). Complete rehydration occurred fast within a 12 h period ( Fig. 6b). Concomitant opening of the rosette caused a rapid rise in the PSII efficiency of the rosette surface to match that measured on the upper (adaxial) leaf surface. For the lower (abaxial) leaf surfaces PSII efficiency did not return to predesiccation values for several days.

Figure 6.

(a) Photochemical efficiency of photosystem II (FV/FM) in response to water content measured on the (▵) upper leaf surface, (●) the rosette surface and on the (∇) lower leaf surface of S. paniculata leaves during a controlled desiccation treatment under 400 μmol photons m−2 s−1. (b) Photochemical efficiency of photosystem II (FV/FM) during recovery from a severe drought stress that came close to the critical water content measured on the (▵) upper (adaxial) leaf surface, (●) the rosette surface and on the (∇) lower (abaxial) leaf surface of S. paniculata leaves. The FV/FM-values are means of 10 measurements ± SD. Water content of the leaf rosettes during recovery is indicated by black bars.

Leaf temperatures higher than 40 °C induced immediate photoinhibition in leaf rosettes of S. paniculata in situ (Neuner & Braun, unpublished). Such temperatures were frequent, occurring on 21·3% of days in summer at the natural sites. Shading of the leaves in the closed leaf rosettes during drought is additionally beneficial to mitigate or even avoid heat-induced photoinhibition.

DISCUSSION

By leaf rosette closure the outermost leaves efficiently shade and cool the leaves in the rosette centre of S. paniculata during drought. At the natural rock sites critically high leaf temperatures in the range of the heat tolerance limit of S. paniculata (45–57 °C) frequently occurred. The species may, thus, depend on shading as an avoidance mechanism. The observed death of a single rosette at a field site at 56 °C stresses the significance of cooling by shading through leaf rosette closure for survival. In addition to avoidance of critical overheating S. paniculata tolerates considerably higher temperatures (up to 57 °C LT50) than other alpine species (44–55 °C; see Gauslaa 1984; Larcher 1995). Occasional exposure to temperatures at or beyond the heat tolerance limit at natural alpine and arctic sites has been proposed to be at least possible at certain sites ( Larcher & Wagner 1976; Körner & Cochrane 1983; Gauslaa 1984). However, the frequency of 11·3% of summer days with which S. paniculata was found to experience such critical temperatures appears to be particularly high.

Plant stature is, under given conditions of ambient temperature, radiation, wind and moisture, the major determinant of plant temperature. The acaulescent rosette life form is common to cold areas, probably due to beneficial effects of stature on temperatures ( Körner & Larcher 1988). Our results corroborate the warming effect of the rosette life form with leaf to air (2 m) temperature differences of up to 36 °C. The effect of rosette structure on heating of the rosette centre is well documented for giant rosette plants ( Smith 1974; Larcher 1975; Hedberg & Hedberg 1979; Meinzer, Goldstein & Rundel 1985; Melcher et al. 1994 ): the incoming solar radiation is focused to the rosette centre and attenuation of wind inside the rosette increases the boundary layer resulting in considerable daytime heating of the apical bud. The same applies to the small acaulescent rosettes of S. paniculata that, due to the cushion growth, additionally exhibit a considerable canopy boundary layer. The heating effect of open rosettes could be detrimental during drought but is abolished by leaf movements in S. paniculata.

Leaf temperature maxima higher than 45 °C always co-occurred with low substrate water potentials at the alpine rock site. With decreasing water content the stomata close, thereby reducing latent heat loss at a time when, during clear weather, the air temperatures and solar radiation and consequently also leaf temperatures are highest (e.g. Forseth & Ehleringer 1980). The observation that the top 5 cm of alpine soils are regularly drying out during periods of bright weather in summer ( Körner 1994) is in accordance to our results of repeated droughts on alpine rock sites. Nevertheless, plants in the alpine zone of high mountain systems appear to be not particularly limited by water supply in most parts of the world, since precipitation tends to increase while vapour pressure deficit and the length of the snow-free season decrease and roots usually penetrate into deeper mostly wet soil layers ( Körner 1994). In the European Alps severe droughts affecting survival during the main growing season may, thus, be restricted to special sites such as alpine rock sites with shallow substrate layers that can dry out quickly.

By leaf rosette closure during drought a reversible reduction (up to 80%) in the area of evaporative surface is achieved. This, together with a reduction (60%) in the amount of radiation absorbed, are well described mechanisms that can help to reduce water loss after stomatal closure in species that show drought tolerance at high tissue water potential ( Jones, Turner & Osmond 1981). Similar surface reductions can only be achieved by leaf shedding in other species ( Larcher 1995). The fast reversibility appears to be crucial for S. paniculata as drought at the alpine rock site does not occur seasonally but instead repeatedly during the growing season.

In S. paniculata the outermost, oldest leaves are sacrificed to promote survival of the individual during drought. The significantly higher water loss of these leaves at the beginning of a drought cannot only be attributed to their stronger exposure. It may be attributable to internal water redistribution driven by the observed osmotic gradient from older to younger leaves. This delays critical water losses in young leaves that in addition are perfectly protected from further evaporative water losses by leaf rosette closure. Water movement from older to younger leaves is a well-described feature of drought adaptation ( Schäfer & Lüttge 1987; Tüffers et al. 1995 ). Water movement between different tissues has also been reported: a reduced water loss from chlorenchyma tissue relative to loss from water storage tissue is maintained by osmotic gradients and can help to prolong photosynthesis during drought ( Barcikowski & Nobel 1984; Schmidt & Kaiser 1987; Schulte & Nobel 1989). A similar osmotic gradient exists in leaves of well-watered leaf rosettes of S. paniculata with a more negative osmotic potential in the mesophyll of the leaf tip (– 1·7 ± 0·1 MPa) than in the water storage tissue of the leaf base (– 0·8 ± 0·1 MPa) that diminishes during desiccation (Neuner & Taschler, unpublished). The osmotic gradient within leaves could be helpful to delay photosynthetic losses at the beginning of drought periods.

Leaf rosette closure is an efficient way of reducing excess light absorption of the photosynthetically active tissue in addition to mitigating the primary causes of photo-inhibition, i.e. heat and drought. The potential role of leaf movements in avoiding photoinhibition has been discussed for a number of species (e.g. Hirata et al. 1983 ; Ludlow & Björkman 1984; Ryel & Beyschlag 1995; stem curlingLebkuecher & Eickmeier 1993). Photoinhibition can reduce carbon gain and thus affect establishment of certain species in special environments ( Ball, Hodges & Laughlin 1991). During water stress stomatal closure occurs even before PSII is affected, thus drought-induced photo-inhibition did not appear to significantly reduce carbon gain ( Boyer et al. 1987 ). After drought, photoinhibition could affect carbon gain as photoinhibition persisted for several days in the lower leaf surfaces of S. paniculata. However, as during rehydration the unaffected upper leaf surfaces of S. paniculata get exposed, persistent drought-induced photo-inhibition may not play a role for carbon gain in this species.

The sensitivity of PSII to light (400 μmol photons m−2 s−1) was enhanced only after severe desiccation of the leaf rosette (below relative water contents of 60%). A possible explanation is that at the onset of water shortages the assimilatory tissue in S. paniculata leaves remains largely unaffected as it is initially supplied by water from the water storage tissue of the leaf base (Neuner and Taschler, unpublished). Different responses to drought are reported depending on the species and the experimental conditions either minor ( Gamon & Pearcy 1990; Epron, Dreyer & Bréda 1992; Valentini et al. 1995 ; Epron 1997) or more major influence on PSII efficiency were found ( Björkman & Powles 1984; Ludlow & Björkman 1984; Cornic et al. 1989 ; Valladares & Pearcy 1997).

Under field conditions drought is often meshed with temperature extremes and the influence of water stress on PSII is often hardly distinguishable from direct temperature effects ( Neuner, Ambach & Aichner 1999). Photosystem II is particularly sensitive to heat ( Weis & Berry 1988). The heat tolerance of PSII is enhanced by desiccation in many species ( Havaux 1992; Epron 1997; Valladares & Pearcy 1997; Arora, Pitchay & Bearce 1998) and also in S. paniculata (Neuner, unpublished). As heat always co-occurred during drought this must be of particular ecological significance. Despite the enhanced heat tolerance of PSII during drought, leaf temperatures higher than 40 °C induced photoinhibition in S. paniculata leaves under field conditions (Neuner & Braun, unpublished). When the frequency of such temperatures (21·3%) is considered, heat-induced photoinhibition could be an important contributor to photosynthetic losses at alpine rock sites in summer.

On alpine rock sites drought is a short-term periodic event that can occur repeatedly. At the beginning of drought water movement to the assimilatory tissue and to younger leaves may help to prolong photosynthesis. After drought the time necessary for a complete restoration of a functioning structure must be considered critical to survival. Prerequisites are that leaves are not shed during drought and that a fast rehydration is possible. By avoidance of photoinhibitory damage during drought in upper leaf surfaces photosynthesis is restored as fast as rehydration proceeds. After drought, this fast restoration ensures that periods with sufficient water supply can be fully utilized. This adds to the benefits of leaf rosette closure during drought that allows the avoidance of environmental extremes that otherwise would come close to the tolerance limits of the species.

ACKNOWLEDGMENTS

Part of this study was supported by the Austrian Fonds zur Förderung der wissenschaftlichen Forschung FWF-Projekt P12446. We wish to thank Prof. Walter Larcher for critical and helpful comments on the manuscript.

Ancillary