Journal of Plant Research

, 122:611

Leaf-level plasticity of Salix gordejevii in fixed dunes compared with lowlands in Hunshandake Sandland, North China


  • Hua Su
    • State Key Laboratory of Vegetation and Environmental Change, Institute of Botany Chinese Academy of Sciences
    • Graduate University of Chinese Academy of Sciences
    • State Key Laboratory of Vegetation and Environmental Change, Institute of Botany Chinese Academy of Sciences
  • Zhenjiang Lan
    • State Key Laboratory of Vegetation and Environmental Change, Institute of Botany Chinese Academy of Sciences
    • Graduate University of Chinese Academy of Sciences
  • Hong Xu
    • State Key Laboratory of Vegetation and Environmental Change, Institute of Botany Chinese Academy of Sciences
  • Wei Liu
    • Zeal Quest Scientific Technology Co. Ltd
  • Bingxue Wang
    • State Key Laboratory of Vegetation and Environmental Change, Institute of Botany Chinese Academy of Sciences
    • Graduate University of Chinese Academy of Sciences
  • Dilip Kumar Biswas
    • State Key Laboratory of Vegetation and Environmental Change, Institute of Botany Chinese Academy of Sciences
    • State Key Laboratory of Vegetation and Environmental Change, Institute of Botany Chinese Academy of Sciences
Regular Paper

DOI: 10.1007/s10265-009-0249-1

Cite this article as:
Su, H., Li, Y., Lan, Z. et al. J Plant Res (2009) 122: 611. doi:10.1007/s10265-009-0249-1


To cope with adverse environments, the majority of indigenous plants in arid regions possess adaptive plasticity after long-term evolution. Leaf-level morphology, anatomy, biochemical properties, diurnal water potential and gas exchange of Salix gordejevii distributed in fixed dunes and lowlands in Hunshandake Sandland, China, were compared. Compared to plants growing in lowlands, individuals of S. gordejevii in fixed dunes displayed much smaller leaf area (0.26 vs 0.70 cm2) and thicker leaves (leaf total thickness 148.59 vs 123.44 μm), together with heavier crust wax, denser hairs, and more compacted epidermal cells. Moreover, those growing in fixed dunes displayed stronger drought-resistance properties as evidenced by higher levels of proline (3.68 vs 0.20 mg g−1 DW) and soluble sugar (17.24 vs 14.49%). Furthermore, S. gordejevii in fixed dunes demonstrated lower water potential and lower light compensation point (28.8 vs 51.9 μmol m−2 s−1). Our findings suggest that morphological and/or anatomical plasticity in leaves has had great adaptive value for Salix in responding to deteriorating environments. The evidence provided here may facilitate the prediction of plant adaptation in community succession in sandy habitats.


AnatomyDiurnal gas exchangeHunshandake SandlandMorphologyPhenotypic plasticitySalix gordejevii


In arid and/or semiarid areas, abiotic factors such as extreme drought (Gutterman 2002), intensive solar irradiation (Aranda et al. 2001; Jiang and Zhu 2001), and high temperature (Sanchez-Blanco et al. 1998), are prominent in limiting the distribution and abundance of plants. Most plants cannot escape those stresses so have to tolerate large diurnal and seasonal environmental fluctuations in situ. However, the majority of indigenous plants distributed throughout arid or semi-arid regions, possess excellent adaptation strategies through long-term evolution. Perennial plants are examples of plants that respond to environment changes primarily by phenotypic plasticity (Rubio De Casas et al. 2007). The plastic nature of plant modular construction allows morphological and/or physiological integration of foliage organs between several habitats (Bazzaz 1991). Such plasticity in indigenous perennial plants may result in tremendous niche differentiation (Bell and Galloway 2007; Burghardt et al. 2008; Cordell et al. 1998; Donovan et al. 2007).

At the individual level, upon expression of phenotypic plasticity, foliage organs usually realize their potential for discriminating the environmental qualities most suitable for growth (Rubio De Casas et al. 2007). Small and thick leaves are produced by plants as a response to drought (Shields 1950). Leaves from high-light microsites often possess higher area-based photosynthetic capacity, accompanied by elevated leaf thickness and decreased specific leaf area (Mishio et al. 2007; Sims et al. 1998). Daily assimilation rate of a C3 forb significantly increased under experimental warming (Niu et al. 2008). Currently, the evolution of plasticity in complex environments is becoming an issue central to the prediction of plant responses to environmental changes caused either by global warming or human activity (Valladares et al. 2007). Nevertheless, research integrating ecophysiology and morphology as well as knowledge of the anatomy of desert shrubs in their natural habitats remains inadequate (Xu et al. 2007).

Hunshandake Sandland is located in a typical semiarid region in the northern hemisphere, and is characterized by higher rainfall and a warmer environment than other regions of this type (Jiang 2003). The habitat in Hunshandake Sandland is highly heterogeneous, including wetlands, lowlands, shifting dunes, semi-fixed dunes and fixed dunes (Niu et al. 2006). Recently, because of the rapid increase in human-induced forcing, e.g., overgrazing and persistent groundwater abstraction (Zheng et al. 2006)—activities that are ever more common in semi-arid and arid areas (Gates et al. 2008; Boulton et al. 2003)—more and more fixed dunes have become active in Hunshandake Sandland, increasing the area of shifting and semi-fixed dunes to as much as 7,120 km2 (Adiya 2006), and triggering environmental deterioration (Yang et al. 2007). Considering that indigenous plants in Hunshandake sand dunes generally experience aridity and other environmental stresses during the hot summers, we hypothesized that we might be able to distinguish morphological and physiological plasticity in the leaves of the plants that enables them to function in such a deteriorating local environment.

This aim of this research was to test the above mentioned hypothesis through a field investigation on a C3 shrub: Salix gordejevii Y. Li Chang et Skv.—a pioneer species colonizing the infertile sandy dunes of Hunshandake Sandland (Ren et al. 2001; Yuan et al. 2005). We compared the leaf morphology, anatomy, diurnal water potential and diurnal photosynthetic characteristics of S. gordejevii plants growing in lowlands and fixed dunes, exploring phenotypic plasticity and physiological mechanisms.

Materials and methods

Study area

The present research was conducted at an experimental site at the Hunshandake Sandy Ecosystem Research Station of the Institute of Botany, Chinese Academy of Sciences (42°54′39.2′′N, 116°01′07.5′′E), which is based in the middle of Xilingel league of the Inner Mongolia Autonomous Region of China (Fig. 1). Hunshandake Sandland, just 180 km north of Beijing, is one of the four largest sandlands in China. The prevailing climate is of the temperate arid or semiarid type. According to meteorological data provided by Zhenglanqi Meteorological Station, the mean annual temperature is 1.7°C, with the mean maximum (July) and minimum (January) temperatures being 20.7 and −16.6°C, respectively. The frostless period lasts approximately 100 days from June to August. The mean annual precipitation varies from 200 to 400 mm, 60% of which is concentrated in the growth period. The sandland contains shallow groundwater and a large number of groundwater-fed lakes, contrasting with the general trend of depletion in surrounding locations. Regarded as sparse-elm grassland, a special ecosystem type less described in China (Jiang et al. 2006), the vegetation consists largely of low, open shrubs dominated by Salix gordejevii Y. L. Chang et Skv., S. microstachya Turz. and Betula fruticosa Pall., with a few tree species such as Ulmus pumila L. var. sabulosa and Malus baccata (Linn.) Borkh. The herbaceous layer is dominated by species such as Leymus Chinensis (Trin.) Tzvel., Agropyron cristatum (L.) Gaertn, Artemisia frigida Willd., Cleistogenes squarrosa (Trin.) Keng and Potentilla acaulis L., etc.
Fig. 1

Geographical location of the experimental site (black dot), the shaded area showing the location of Hunshandake Sandland

The plant

Salix gordejevii, an indigenous shrub, distributes widely in the studied area, co-existing with Hedysarum fruticosum Pall. var. mongolicum (Turcz.) Turcz. ex B. Fedtsch. in semi-fixed or fixed dunes. It can endure drought, resist wind erosion and bear sand coverage. The mature plant is about 1–3 m in height, with its vertical and horizontal roots extending up to 3.5 and >20 m, respectively (Ren et al. 2001). It starts to expand leaves in mid-May and grows vigorously until July. S. gordejevii is recognized as one of the most important pioneer species in the primary stage of sand vegetation succession (Yuan et al. 2005). In 2001, we set up a huge experimental plot (26.7 km2) in Hunshandake Sandland in order to observe degraded ecosystem restoration. We have since observed the generation of S. gordejevii and labelled them every year since 2001. In July 2007, we selected three 4-year-old S. gordejevii naturally distributed in both lowlands (2.3 ± 0.4 m in height, Fig. 2a) and the adjacent fixed dunes (1.5 ± 0.3 m in height, Fig. 2b) for sampling and measurement. S. gordejevii growing in two habitats may belong to different ecotypes with the same genotypes, with variations in their leaf traits possibly being due to phenotypic plasticity associated with different habitats (Li et al. 2005).
Fig. 2

Illustrative images of Salix gordejevii growing in lowlands (a) and fixed dunes (b)

Basic soil properties

Three plots (2 m × 2 m) were selected at both lowland and fixed dune sites. Plot images at a resolution of 1,600 × 1,200 pixels were taken using a digital camera (Canon PowerShot A640, Canon, Tokyo, Japan) at the same locations at which spectral irradiance measurements were taken. Present vegetation coverage was estimated according to the method of Lukina et al. (1999). Six wells (three wells per site) were constructed for groundwater observation. The water table has been monitored every month since 2001. Three soil cores (3.5 cm diameter) were collected randomly with a soil core in each habitat. The soil sampling depth was 30 cm. Each repetition represented a composite of three soil sub-samples. Soil samples without roots were dried at 105°C for 48 h and weighed. We calculated soil moisture as follows: soil moisture = (fresh weight − dry weight)/dry weight. Air-dried soil samples were coarsely ground to pass through a 1 mm sieve for determination of pH value, and contents of organic matter, nitrogen (N), phosphorus (P) and potassium (K) according to the method of Smith (1983).

Leaf morphology

Thirty leaves from the middle of twigs were sampled from each lowland and fixed dune plant. Leaf length, leaf width and twig length was determined using a ruler; leaf inclination was investigated using a clinometer (PM-5/360PC, Suunto, Vantaa, Finland) and all weights were determined using a portable centesimal electronic balance (TD10002, Jinnuo, Yuyao, China). Individual leaves were photographed with a 2 cm willow branch as the reference object, and areas were then calculated using Image J software (1.38x, Wayne Rasband, National Institutes of Health, The leaf area index was the ratio of total upper leaf surface of plant divided by the surface area of the land on which the plant grows. Leaf mass per unit area was calculated as the ratio of leaf weight to leaf area (mg mm−2).

Leaf anatomy

For leaf anatomy analysis, samples were collected early in the morning to avoid dehydration of the tissue and the diurnal swelling/shrinkage cycling (McBurney 1992). Ten randomly chosen leaves were taken from three plants each growing in lowlands and fixed dunes. Leaf strips (1–2 cm width) midway between the leaf edge and the mid-vein were cut with a sharp blade in the field and immediately fixed in FAA [ethanol:acetic acid:formaldehyde (18:1:1)]. Leaf segments of approximately 3 × 5 mm adjacent to the leaf margin were taken from the fixed strips, dehydrated in ethanol and incubated in an ethanol–isoamyl acetate mixture for 1 h. The sections were further dried before being coated with gold. The mounted specimens were examined and photographed with a scanning electron microscope (Hitachi S-570 SEM, Hitachi, Tokyo, Japan) operated at 12 kV. The thickness of leaves, mesophyll tissues, and epidermis were measured for thickness. Both abaxial and adaxial stomata were counted for the calculation of the ratio upper/lower epidermis stomatal density (SDupper/SDlower). Each individual leaf was taken as a replication.

Leaf biochemical properties

Fifty expanded leaves from the middle of annual twigs were harvested and immediately dried in an electronic oven at 65°C for 48 h and weighed. Nitrogen concentration was determined colorimetrically by the Kjeldahl acid-digestion method using a Technicon Auto Analyzer (Pulse Instrumentations, Saskatoon, Canada) after extraction with sulfuric acid (Bremner and Mulvaney 1982). Leaf chlorophyll was extracted with 80% acetone and analyzed at 470, 646 and 663 nm with spectrophotometer (UV-2450, Shimadzu, Kyoto, Japan). The extinction coefficients and the equations originated by Lichtenthaler (1987) were applied to calculate chlorophyll content. Proline content was determined photometrically as previously described (Bates et al. 1973; Rohde et al. 2004). Soluble sugars were extracted and analyzed following the method described by Rohde et al. (2004).

Leaf ecophysiology

Leaf water potential (Ψleaf) was determined using an automated eight-channel water potential datalogger (PsyPro, Wescor, Logan, UT) in situ in three fully expanded leaves on the middle of annual twigs per plant from the middle of a shoot at each sampling time at 2 h intervals from 0800 hours to 1800 hours.

Gas exchange was measured on three fully expanded sun-lit leaves on the annual twigs in situ using a portable gas exchange fluorescence system (GFS-3000, Walz, Effeltrich, Germany). Such measurements were taken only on cloudless days with natural irradiance and prevalent environmental conditions. Gas exchange of selected leaves was monitored at 2 h intervals from 0600 hours to 1800 hours with temporal photosynthetic photon flux density (PPFD). Air flow rate was set at 700 μmol min−1 and CO2 concentration in the leaf cuvette was maintained at 340 μmol mol−1, which was approximately the same CO2 concentration found near the plant canopy. Intercellular CO2 concentration (Ci), transpiration rate (E), stomatal conductance (gs), and ambient CO2 concentration (Ca) were recorded simultaneously with the assimilation rate (A). Microenvironmental information, such as PPFD, air temperature, leaf temperature and (air-to-leaf-) vapor pressure deficit (VPD), were also recorded automatically by the instrument. Stomatal limitation value (Ls) was calculated as: Ls = 1 − Ci/(Ca − Γ) ≈ 1 − (Ci/Ca) (Γ is CO2 compensation point; Berry and Downton 1982). Water use efficiency (WUE) was calculated as A/E (Wullschleger and Oosterhuis 1989). The maximum photochemical efficiency in leaves (Fv/Fm) was measured at 1000 hours following 30 min dark adaptation.

A-light curve

On cloudless days, A-light curves were determined on three fully expanded sun-lit leaves on the annual twigs using a portable gas exchange fluorescence system (GFS-3000, Walz). The measurements were conducted at around 1000 hours in situ. To measure the A-light curves, we set the air flow rate at 700 μmol min−1 and CO2 concentration at 340 μmol mol−1, i.e., the same as for gas exchange measurements. The leaf selected for light curve measurement was allowed to equilibrate in the leaf chamber for about 20 min, and the built-in temperature controller was used to keep the temperature stable at 24 ± 1.0°C. The relative humidity in the leaf chamber was controlled at 30 ± 0.5%, which was approximately the same air relative humidity near the plant canopy. A and gs in response to PPFDs of 1,500, 1,300, 1,000, 700, 300, 100, 50 and 10 μmol m−2 s−1 at the leaf surface level were recorded. PPFDs were controlled using the light sensor in the upper chamber side. For A-light measurement, each PPFD step lasted 3–5 min before data were recorded. The same measurements were repeated at 7- to 10-day intervals as replicates, using the middle third of each leaf, on the same set of leaves. The data obtained were analyzed and logarithm fit curves were performed using a Sigmaplot program (Version 10.0, Sigmaplot, San Jose, CA) to obtain maximum assimilation rate (Amax) and light compensation point (ALCP).

Statistical analysis

The experiment consisted of randomized complete block design with three plants per replicate. Statistical analyses of data were performed using analysis of variance (ANOVA) in the General Linear Model process of SPSS (Version 13.0, SPSS, Chicago, IL). The effects of habitats on S. gordejevii leaf traits in lowlands and fixed dunes were analyzed using one-way ANOVA on the measured variables. The Tukey–Kramer method was applied to assess pair-wise comparisons among treatment means. Differences between treatments were considered significant if P < 0.05.


Soil properties

The detailed soil backgrounds of both lowlands and fixed dunes are shown in Table 1. Fixed dunes had more adverse soil conditions than lowlands, with significantly lower vegetation coverage, deeper groundwater, and drier and more infertile soil than lowlands.
Table 1

Sites description and soil characteristics



Fixed dune

P valuea

Vegetation coverage (%)

85 ± 5

20 ± 8


Annual average water table (m)

0.98 ± 0.15

6.64 ± 0.12


Average soil moisture (%)

15.39 ± 0.05

4.54 ± 0.01


Soil pH

9.09 ± 0.02

7.51 ± 0.01


Soil organic matter (g kg−1)

5.99 ± 0.23

1.46 ± 0.17


Soil N (g kg−1)

1.36 ± 0.03

0.29 ± 0.01


Soil P (g kg−1)

0.19 ± 0.04

0.05 ± 0.01


Soil K (g kg−1)

17.9 ± 0.05

15.4 ± 0.04


Values are means ± SE (n = 3)

aDifference significant if P ≤ 0.05

Leaf morphology

As shown in Table 2, the average leaf area of the plants in fixed dunes was significantly (P < 0.05) smaller than those in lowlands, either shorter in length or narrower in width. As calculated by the leaf width/length ratio, leaves of S. gordejevii in fixed dunes were more rounded than those in lowlands. Leaf area index in lowlands was twofold larger than in fixed dunes. Twig length and mean leaf inclination of S. gordejevii in lowlands were considerably longer and larger, respectively, than those in fixed dunes. However, leaf mass per area displayed no significant difference between plants growing in lowlands and fixed dunes.
Table 2

Leaf morphology, anatomy, biochemical properties and photochemical efficiency of PSII (Fv/Fm) of Salix gordejevii growing in lowlands and fixed dunes. Values are means ± SE




Fixed dune

P valuea

Morphology (n = 30)

Leaf length (cm)

2.81 ± 0.13

1.41 ± 0.16


Leaf width (cm)

0.29 ± 0.02

0.16 ± 0.04


Leaf area (cm2)

0.70 ± 0.04

0.26 ± 0.08


Leaf area index

0.88 ± 0.07

0.45 ± 0.07


Twig length (cm)

31.02 ± 0.11

23.07 ± 0.09


Mean leaf inclination (°)

60.04 ± 12.71

46.37 ± 13.12


Leaf mass per unit area (mg mm−2)

2.3 ± 0.04

1.9 ± 0.01


Anatomy (n = 10)

Stomatal density ratio (upper/lower)

0.64 ± 0.02

0.57 ± 0.01


Leaf thickness (μm)

123.44 ± 2.88

148.59 ± 4.94


Upper epidermis thickness (μm)

9.35 ± 1.62

15.87 ± 2.43


Lower epidermis thickness (μm)

7.38 ± 1.16

10.61 ± 1.47


Palisade parenchyma thickness (μm)

107.10 ± 2.99

122.71 ± 6.34


Biochemical properties (n = 3)

Leaf N (mg g−1)

4.76 ± 0.03

4.21 ± 0.02


Proline (mg g−1 DW)

0.20 ± 0.04

3.68 ± 0.10


Soluble sugar (%)

14.49 ± 1.65

17.24 ± 1.75


Chlorophyll (mg g−1 DW)

0.90 ± 0.02

0.97 ± 0.05


(n = 3)


0.60 ± 0.08

0.75 ± 0.02


aDifference significant if P ≤ 0.05

Leaf anatomy

Although both upper and lower leaf surfaces of S. gordejevii were covered with wax and hairs, leaf hairs in fixed dunes were denser (Fig. 3b, d) than in lowlands (Fig. 3a, c). Waxes belonged to two different types. Granules were found in leaves of S. gordejevii distributing in lowlands (Fig. 3e), whereas crusts were found in fixed dune plants (Fig. 3f).
Fig. 3a–f

Representative scanning electron microscopy (SEM) images of hairs and wax of S. gordejevii leaves growing in lowland and fixed dunes. a, b Hairs, upper surface: lowland (a), fixed dune (b). c, d Hairs, lower surface: lowland (c), fixed dune (d). e, f Wax, upper surface: lowland (e), fixed dune (f)

Ultrastructural observations on mesophyll cells revealed remarkable intraspecific differences between the leaves of S. gordejevii from lowlands and fixed dunes (Table 2; Fig. 4). Leaves of plants in fixed dunes developed much thicker mesophyll and palisade parenchyma than those in lowlands. Both upper and lower epidermis in leaves of fixed dunes were thicker than those of lowland plants. Also, stomatal density in the upper with respect to the lower leaf surface was significantly (P < 0.05) higher in lowlands than in fixed dunes. Furthermore, we found that leaves of S. gordejevii in Hunshandake Sandland had no spongy parenchyma, and were characterized as isobilateral leaves with total-palisade (Fig. 4).
Fig. 4

Representative SEM images of transverse section of S. gordejevii leaves growing in lowland (a) and fixed dunes (b)

Biochemical properties and photochemical efficiency

The leaf proline and soluble sugar contents of S. gordejevii in lowlands were significantly (P < 0.05) lower than those in fixed dunes. There was no difference in chlorophyll content in leaves between the two habitats (Table 2). The values of the maximum photochemical efficiency in dark-adapted leaves (Fv/Fm) in lowlands around 1000 hours differed significantly (P < 0.05) from that in fixed dunes (0.60 vs 0.75).

Leaf water potential and water use efficiency

In general, plants from both habitats showed a similar trend of leaf water potential (Ψleaf) over time. Lowland plants had higher Ψleaf than fixed dune plants until 1400 hours when both plants displayed the lowest Ψleaf value (about −1.8 MPa). Ψleaf then increased in the afternoon. However, there was no significant difference in Ψleaf between plants from the two kinds of habitats during the afternoon period (Fig. 5a). There was no significant difference in changes in WUE until 1600 hours. However, at later times, leaf WUE of fixed dune plants became significant higher than that of lowland plants (Fig. 5b).
Fig. 5

Diurnal variation in leaf water potential (Ψleaf) (a) and water use efficiency (WUE) (b) of S. gordejevii growing in lowland and fixed dunes (mean ± SE, n = 3)

Gas exchange and A-light curves

The diurnal PPFDs of both lowland and fixed dune plants showed a one-peak (1200 hours) curve type, with no differences being noted (Fig. 6e). The same trends were observed for air temperature (Fig. 6f). However, after 0800 hours, leaf temperatures of lowland plants were higher than air temperatures by 0.3–7.7°C, while the corresponding difference in fixed dune plants was as large as 1.5–8.1°C. The biggest gaps from both habitats appeared at 1200 hours (Fig. 6g). The peak value of air-to-leaf-vapor pressure deficit (VPD) of lowland plants occurred at 1200 hours, while that of fixed dune plants was at 1400 hours (Fig. 6h).
Fig. 6a–h

Diurnal variation in gas exchange of S. gordejevii growing in lowland and fixed dune (means ± SE, n = 3) with the microenvironments of days when the measurements were taken. a Photosynthetic rate (A). b Intercellular CO2 concentration (Ci). c Stomatal conductance (gs). d Stomatal limitation value (Ls). e Photosynthetic photon flux density (PPFD). f Air temperature. g Leaf temperature. h (Air-to-leaf-) vapor pressure deficit (VPD)

Assimilation rate (A) and stomatal conductance (gs) of S. gordejevii in both habitats showed double peaks (around 0800 hours and 1600 hours), with different peak values (Fig. 6a, c). For the first peak, A value of fixed dunes was 5.7 μmol m−2 s−1, 93% higher than that in fixed dunes (2.9 μmol m−2 s−1). For the second peak, however, the relatively decrease was only 36% (Fig. 6a). In terms of diurnal intercellular CO2 concentration (Ci), the trend in plants from both habitats appeared similar, the lowest value appearing at 1000 hours then climbing (Fig. 6b). On the contrary, the diurnal stomatal limitation of plants (Ls) of S. gordejevii in both lowland and fixed dune exhibited the opposite trend in diurnal Ci, the highest value appearing at 1000 hours then declining (Fig. 6d).

A of S. gordejevii increased continually with increasing PPFD, until PPFD reached around 1,500 μmol m−2 s−1, when A was then maintained at a constant maximum (Fig. 7a). The same trends appeared in gs versus PPFD (Fig. 7b). Significant differences were observed in maximum assimilation rate (Amax) and light compensation point (ALCP) in S. gordejevii in different habitats. The Amax of S. gordejevii in lowlands and fixed dunes was 10.8 and 7.8 μmol m−2 s−1, respectively. Lowland habitats nearly doubled their ALCP (51.9 μmol m−2 s−1) compared with fixed dunes (28.8 μmol m−2 s−1) (Fig. 7).
Fig. 7

Relationship between a photosynthetic rate (A), and b stomatal conductance (gs) of S. gordejevii and photosynthetic photon flux density (PPFD) (mean ± SE, n = 9)


Our data on leaf-level morphology, anatomy, biochemistry and eco-physiology showed that S. gordejevii growing in fixed dunes has developed a range of adaptive mechanisms to cope with environmental stresses (e.g., deeper groundwater, lower soil moisture, less soil organic matter and less soil N; Table 1). For instance, significantly smaller and more rounded leaves with denser hairs, and heavier crust wax and smaller inclination to the twig (Table 2) of leaves in fixed dunes are all indicates of adaptation (Ehleringer et al. 1981; Marchi et al. 2008). However, there was no difference in the leaf mass per unit leaf area of S. gordejevii between lowlands and fixed dunes. This evidence indicates that an ecotype of S. gordejevii has adapted well to fixed dunes by minimizing and thickening its leaves but keeping relative higher individual mass per leaf.

As a manifestation of the modular nature, plasticity was expressed concurrently with the morphological and anatomical characteristics. For S. gordejevii, both upper and lower epidermis were composed of a layer of approximately similar cells, i.e., rectangular and closely arranged (Fig. 4). There are numerous stomata on both upper and lower epidermis (Table 2). The leaf had no spongy parenchyma, belonging to the isobilateral leaf type with total-palisade. However, compared to leaves in lowlands, those in fixed dunes became much thicker (Table 2; Fig. 4), and the cells of the palisade parenchyma were smaller and more compact (Fig. 4). Further anatomical research revealed that the leaves of S. gordejevii had developed main veins in poor microhabitats, which can transport more water and nutrients (Cui et al. 2006). All these adaptations in fixed dunes usually result in a greater fraction of leaf tissues in mesophyll at the expense of epidermis and mechanical tissues (Garnier et al. 1999), thereby possibly decreasing water loss and increasing foliage photosynthetic potential.

Besides morphological/anatomical adaptations, the long-lived leaves have to develop biochemical mechanisms targeting at maintaining proper osmotic potential or dissipating excess radiant energy to survive through periods of hostile conditions (Demmig-Adams 1996). Proline and soluble sugar contents were elevated in S. gordejevii in fixed dunes (Table 2), indicating that plants growing in such locations might develop higher osmotic adjustment capacity, as these metabolites are regarded as the main organic solutes synthesized and accumulated for osmotic adjustment under environmental stresses (Hincha 2006; Hessini et al. 2008). Higher leaf N concentration of S. gordejevii was observed in lowlands compared to fixed dunes (Table 2), suggesting relatively higher soil N availability (Hendricks et al. 2000) in lowlands. Nevertheless, S. gordejevii could minimize N loss in habitats with lower N availability (Yuan et al. 2005).

The maximum photochemical efficiency in dark-adapted leaves ratio (Fv/Fm) reflects the potential quantum efficiency of PSII, which is normally used as a sensitive indicator of plant photosynthetic performance (Krause and Weis 1991). Low values (<0.8) of Fv/Fm around 1000 hours in S. gordejevii in both habitats suggested that a depression of PSII and photoinhibition occurred (Maxwell and Johnson 2000).

As an important indicator of drought tolerance, leaf water potential (Ψleaf) is an innate physiological trait that characterizes a plant’s attempt to minimize the difference in water potential between the plant body and its environment (Gutterman 2002; Korn et al. 2008). In our study, S. gordejevii in fixed dunes differed markedly from lowland plants in its ability to maintain relatively high Ψleaf before noon (Fig. 5). This was also reflected in the drier conditions of fixed dunes compared to lowlands, and the better adaptation of S. gordejevii in fixed dunes. Small and thick leaves ensure greater carbon gain over transpiration losses under the prolonged hot and dry periods that prevail in most arid regions (Chaves et al. 2002). The improved light interception (Fig. 6a) and higher WUE (Fig. 5b) in fixed dunes permitted by this increased thickness maybe counteracted by the tendency to become much hotter than ambient air when stomata close, and to restrict latent heat exchange under drought conditions (Monclus et al. 2006). However, leaf temperature of S. gordejevii in fixed dunes did not rise much above air temperature (Fig. 6), possibly because the small size of the leaves allowed for increased heat dissipation through convection/conduction (Chaves et al. 2002).

S. gordejevii growing in lowlands and fixed dunes demonstrated significantly different A values in diurnal photosynthesis, although with a similar two-peak pattern (Fig. 6). However, S. gordejevii reached the first peak around 0800 hours (Fig. 6a), 2 h earlier than that of most mesophyte species, whose first peak normally appears around 1000 hours (e.g., Zhang and Gao 1999). This result indicated that this species could quickly utilize radiant energy and water resources before the environmental stress occurs. The same phenomenon was seen in S. psammophila C. Wang et Ch. Y. Yang, H. fruticosum and Artemisia Ordosica Kraschen (e.g., Jiang and Zhu 2001). A depression of stomatal conductance (gs) occurred for longer periods of the day (from 1000 hours to 1400 hours) (Fig. 6c), followed by parallel decreases in A, indicating that closure of stomata might be the main factor underlying A midday depression (Jiang and Zhu 2001); other reasons might include serious noon water stress (Xu 2002), as evidenced by the opposite direction of Ci and Ls, or photoinhibition (Xu and Shen 1997), as shown by the low values of Fv/Fm. The above differences in photosynthetic rate in S. gordejevii in two distinct habitats might result from differences in morphological and/or anatomical features of leaves (Garnier et al. 1999; Nobel 1983; Sharkey 1985).

Saturating light intensity of leaf photosynthesis was much lower than sunlight intensity (about 2,100 μmol m−2 s−1) around noon (Fig. 7a), which further indicated serious photoinhibition in the field (Xu and Shen 1997). In addition, the lower light compensation point (ALCP) of leaves in fixed dunes might contribute to their effective carbon assimilation, using diffuse radiation in the dawn and dusk, which coincides with a diurnal photosynthetic performance.

It is well-documented that environmental factors induce photosynthetic machinery switching (Niu et al. 2006; Reiskind et al. 1997; Winter and Smith 1996) and acclimation for many species in laboratory experiments (Atkin et al. 2000; Yamori et al. 2005) and field measurements (Zhou et al. 2007; Niu et al. 2008). Obviously, photosynthesis of S. gordejevii here did not conform to the former phenomenon but might be accounted for by acclimation influenced by environmental variation.


In dry environments, Salix gordejevii has developed a range of distinct adaptive features related to drought-resistance, e.g., thickening and minimizing leaves with crust wax and dense hairs, accumulating osmotic adjustment substances, and modulating water potential quickly, etc. S. gordejevii growing in fixed dunes could realize their full potential, maintaining higher photosynthesis throughout the whole day, although a midday depression induced by photoinhibition did occur. Our field investigations will facilitate the prediction of community succession in sandy environments such as Hunshandake Sandland. The research here also lends support to the idea that ecosystems that are not damaged too seriously hold the ability to restore themselves through natural processes.


This research was supported financially by the “973” Project of China (No: 2007CB106804), the Key Innovative Project of Chinese Academy of Sciences (No: KZCX2-XB2-01), National Key Technologies R&D Programs of China (No: 2008BAD0B05 and 2006BAC01A12) and “Zealquest Scientific Foundation”. We sincerely thank Nasen Wuritu, Huhe Tuge, Benying Su and Benfu Li for their invaluable help in the field. We thank Guanglei Wu, Yong Li and Caihong Li from Shandong Agricultural University for their help with leaf proline and chlorophyll measurements.

Copyright information

© The Botanical Society of Japan and Springer 2009